Datasets:

ChicagoHAI

/

CaseSumm

like 10

Follow

ChicagoHAI 19

582.US.2016_15-1503

Petitioners—Timothy Catlett, Russell Overton, Levy Rouse, Kelvin Smith, Charles and Christopher Turner, and Clifton Yarborough—and several others were indicted for the kidnaping, robbery, and murder of Catherine Fuller. At trial, the Government advanced the theory that Fuller was attacked by a large group of individuals. Its evidentiary centerpiece consisted of the testimony of Calvin Alston and Harry Bennett, who confessed to participating in a group attack and cooperated with the Government in return for leniency. Several other Government witnesses corroborated aspects of Alston’s and Bennett’s testimony. Melvin Montgomery testified that he was in a park among a group of people, heard someone say they were “going to get that one,” saw petitioner Overton pointing to Fuller, and saw several persons, including some petitioners, cross the street in her direction. Maurice Thomas testified that he saw the attack, identified some petitioners as participants, and later overheard petitioner Catlett say that they “had to kill her.” Carrie Eleby and Linda Jacobs testified that they heard screams coming from an alley where a “gang of boys” was beating someone near a garage, approached the group, and saw some petitioners participating in the attack. Finally, the Government played a videotape of petitioner Yarborough’s statement to detectives, describing how he was part of a large group that carried out the attack. None of the defendants rebutted the prosecution witnesses’ claims that Fuller was killed in a group attack. The seven petitioners were convicted. Long after their convictions became final, petitioners discovered that the Government had withheld evidence from the defense at the time of trial. In postconviction proceedings, they argued that seven specific pieces of withheld evidence were both favorable to the defense and material to their guilt under Brady v. Maryland, 373 U. S. 83 . This evidence included the identity of a man seen running into the alley after the murder and stopping near the garage where Fuller’s body had already been found; the statement of a passerby who claimed to hear groans coming from a closed garage; and evidence tending to impeach witnesses Eleby, Jacobs, and Thomas. The D. C. Superior Court rejected petitioners’ Brady claims, finding that the withheld evidence was not material. The D. C. Court of Appeals affirmed. Held: The withheld evidence is not material under Brady. Pp. 9–14. (a) The Government does not contest petitioners’ claim that the withheld evidence was “favorable to the defense.” Petitioners and the Government, however, do contest the materiality of the undisclosed Brady information. Such “evidence is ‘material’ . . . when there is a reasonable probability that, had the evidence been disclosed, the result of the proceeding would have been different.” Cone v. Bell, 556 U. S. 449 –470. “A ‘reasonable probability’ of a different result” is one in which the suppressed evidence “ ‘undermines confidence in the outcome of the trial.’ ” Kyles v. Whitley, 514 U. S. 419 . To make that determination, this Court “evaluate[s]” the withheld evidence “in the context of the entire record.” United States v. Agurs, 427 U. S. 97 . Pp. 9–11. (b) Petitioners’ main argument is that, had they known about the withheld evidence, they could have challenged the Government’s basic group attack theory by raising an alternative theory, namely, that a single perpetrator (or two at most) had attacked Fuller. Considering the withheld evidence “in the context of the entire record,” Agurs, supra, at 112, that evidence is too little, too weak, or too distant from the main evidentiary points to meet Brady’s standards. A group attack was the very cornerstone of the Government’s case, and virtually every witness to the crime agreed that Fuller was killed by a large group of perpetrators. It is not reasonably probable that the withheld evidence could have led to a different result at trial. Petitioners’ problem is that their current alternative theory would have had to persuade the jury that both Alston and Bennett falsely confessed to being active participants in a group attack that never occurred; that Yarborough falsely implicated himself in that group attack and yet gave a highly similar account of how it occurred; that Thomas, an otherwise disinterested witness, wholly fabricated his story; that both Eleby and Jacobs likewise testified to witnessing a group attack that did not occur; and that Montgomery in fact did not see petitioners and others, as a group, identify Fuller as a target and leave together to rob her. As for the undisclosed impeachment evidence, the record shows that it was largely cumulative of impeachment evidence petitioners already had and used at trial. This is not to suggest that impeachment evidence is immaterial with respect to a witness who has already been impeached with other evidence, see Wearry v. Cain, 577 U. S. ___, ___–___. But in the context of this trial, with respect to these witnesses, the cumulative effect of the withheld evidence is insufficient to undermine confidence in the jury’s verdict, see Smith v. Cain, 565 U. S. 73 –76. Pp. 11–14. 116 A. 3d 894, affirmed. Breyer, J., delivered the opinion of the Court, in which Roberts, C. J., and Kennedy, Thomas, Alito, and Sotomayor, JJ., joined. Kagan, J., filed a dissenting opinion, in which Ginsburg, J., joined. Gorsuch, J., took no part in the consideration or decision of the cases.Notes 1 Together with No. 15–1504, Overton v. United States, also on certiorari to the same court.

In Brady v. Maryland, 373 U. S. 83 (1963) , this Court held that the government violates the Constitution’s Due Process Clause “if it withholds evidence that is favorable to the defense and material to the defendant’s guilt or punishment.” Smith v. Cain, 565 U. S. 73, 75 (2012) (emphasis added) (summarizing Brady holding). In 1985 the seven petitioners in these cases were tried together in the Superior Court for the District of Columbia for the kidnaping, armed robbery, and murder of Catherine Fuller. Long after petitioners’ convictions became final, it emerged that the Government possessed certain evidence that it failed to disclose to the defense. The only question before us here is whether that withheld evidence was “material” under Brady. The D. C. Superior Court, after a 16-day evidentiary hearing, determined that the withheld evidence was not material. Catlett v. United States, Crim. No. 8617–FEL–84 etc. (Aug. 6, 2012), App. to Pet. for Cert. in No. 15–1503, pp. 84a, n. 4, 81a–131a. The D. C. Court of Appeals reviewed the record, reached the same conclusion, and affirmed the Superior Court. 116 A. 3d 894 (2015). After reviewing the record, we reach the same conclusion as did the lower courts. I In these fact-intensive cases, we set out here only a basic description of the record facts along with our reasons for reaching our conclusion. We refer those who wish more detail to the opinions of the lower courts. App. to Pet. for Cert. in No. 15–1503, at 81a–131a; 116 A. 3d 894. A The Trial On March 22, 1985, a grand jury indicted the seven petitioners—Timothy Catlett, Russell Overton, Levy Rouse, Kelvin Smith, Charles Turner, Christopher Turner, and Clifton Yarborough—and several others for the kidnaping, robbery, and murder of Catherine Fuller. The evidence produced at their joint trial showed that on October 1, 1984, at around 4:30 p.m., Catherine Fuller left her home to go shopping. At around 6 p.m., William Freeman, a street vendor, found Fuller’s body inside an alley garage between Eighth and Ninth Street N. E., just a few blocks from Fuller’s home. See Appendix, infra (showing a map of the area in which the murder was committed). Fuller had been robbed, severely beaten, and sodomized with an object that caused extensive internal injuries. The Government advanced the theory at trial that Fuller had been attacked in the alley by a large group of individuals, including petitioners; codefendants Steve Webb, Alfonso Harris, and Felicia Ruffin; as well as by Calvin Alston and Harry Bennett. The Government’s evidentiary centerpiece consisted of testimony by Alston and Bennett, who confessed to participating in the offense and who cooperated with the Government in return for leniency. Although the testimony of Alston and Bennett diverged on minor details, it was consistent in stating that, and describing how, Fuller was attacked by a siz-able group of individuals, including petitioners and they themselves. Alston testified that at about 4:10 p.m. on the day of the murder, he arrived in a park located on H Street between Eighth and Ninth Streets. He said he found a group of people gathered there. It included petitioners Levy Rouse, Russell Overton, Christopher Turner, Charles Turner, Kelvin Smith, Clifton Yarborough, and Timothy Catlett, as well as several codefendants and others. Those in the group were talking and singing while Catlett was banging out a beat. Alston suggested “getting paid” by robbing someone. App. A467. Catlett, Overton, Rouse, Smith, Charles Turner, Christopher Turner, Yarborough, and several others agreed. Alston pointed at Catherine Fuller, who was walking on the other side of H Street near the corner of H and Eighth Streets. Those in the group said they were “game for getting paid.” Id., at A471–A472. Alston, Rouse, Yarborough, and Charles Turner crossed H Street moving toward Eighth Street and followed Fuller down Eighth Street. The rest of the group crossed H Street and moved toward Ninth Street. When Alston’s group approached Fuller, Charles Turner shoved her into an alley that runs between Eighth and Ninth Streets. Charles Turner, Rouse, and Alston began punching Fuller. They were soon joined by Christopher Turner, Smith, and others. All of them continued to hit and kick Fuller until she fell to the ground. Rouse and Charles Turner then carried Fuller to the center of the alley and dropped her in front of a garage located at the point where the alley joins another, perpendicular alley that runs toward I Street. Someone dragged Fuller into the garage. Alston, Rouse, Charles Turner, Overton, Yarborough, and Catlett followed. Others stood outside. Members of the group tore Fuller’s clothes off and struggled over her change purse. Overton and Charles Turner then held Fuller’s legs, and Alston, Catlett, Harris, and Yarborough stood around her while Rouse sodomized her with a foot-long pipe. Shortly after, the group dispersed and left the alley. Harry Bennett’s testimony was similar. Bennett also described a group attack. He said that he had gone to the H Street park, where he saw Rouse, Overton, Christopher Turner, Smith, Catlett, and others gathered. Alston was talking to the group about “[g]etting paid” and said “let’s go get that lady.” Id., at A368–A370. At that point Alston, Rouse, Overton, and Webb crossed H Street and approached Fuller, while Catlett, Christopher Turner, Charles Turner, and Harris followed in a separate group. Bennett added that he himself went to the corner of Eighth and H Streets to watch for police. He then went into the alley and joined the group in kicking and beating Fuller. He testified that at least 12 people were there, with some beating Fuller and others watching or picking up her jewelry. Overton then dragged Fuller into the garage, and Bennett, Rouse, Christopher Turner, Charles Turner, Catlett, Smith, Harris, and Webb followed, as did some “girls.” Id., at A402–A405. Alston and Steve Webb held Fuller’s legs, and Rouse sodomized her with a pole. The group then dispersed from the garage and alley. The Government presented several other witnesses who corroborated aspects of Alston’s and Bennett’s testimony, including the fact that Fuller was attacked by a group. Melvin Montgomery testified that he was in the H Street park on the afternoon of the murder. He saw Overton, Catlett, Rouse, Charles Turner, and others gathered there. The group was being noisy and singing a song about needing money. Somebody then said they were “going to get that one,” and Montgomery saw that Overton was pointing to a woman standing on the corner of Eighth Street. Id., at 77–79. Overton, Catlett, Rouse, Charles Turner, and others crossed H Street. Some headed toward Eighth Street while others went toward Ninth Street. Montgomery did not follow them. Maurice Thomas, then 14 years old, testified that he witnessed the attack itself. Thomas lived in the neighborhood and knew many of the defendants. As he was walking home, he glanced down the Eighth Street alley and saw a group surrounding Fuller. Thomas saw Catlett pat Fuller down and then hit her. He then saw everyone in the group join in hitting her. Thomas said he knew Catlett, Yarborough, Rouse, Charles Turner, Christopher Turner, and Smith and recognized them in the group. Thomas heard Fuller calling for help. He ran home where he found his aunt, who told him not to tell anyone what he saw. Later that day, Thomas saw Catlett at a corner store, and heard Catlett say to someone that they “had to kill her” because “she spotted someone he was with.” Id., at 127–128. On the afternoon of the murder, Carrie Eleby and Linda Jacobs were looking for petitioner Smith, who was Eleby’s boyfriend, near the corner of H and Eighth Streets. They heard screams coming from where a “gang of boys” was beating somebody near the garage in the alley. Id., at A539–A541. Eleby and Jacobs approached the group. Eleby recognized Christopher Turner, Smith, Catlett, Rouse, Overton, Alston, and Webb kicking Fuller while Yarborough stood nearby. Both Eleby and Jacobs testified that they saw Rouse sodomize Fuller with a pole. Eleby added that Overton held Fuller’s legs. Finally, the Government played a videotape of a recorded statement that Yarborough, one of the petitioners, had given to detectives on December 9, 1984, approximately two months after the murder. Names were redacted. The video shows Yarborough describing in detail how he was part of a large group that forced Fuller into the alley, jointly robbed and assaulted her, and dragged her into the garage. None of the defendants testified, nor did any of them try, through witnesses or other evidence, to rebut the prosecution witnesses’ claim that Fuller was killed in a group attack. Rather, each petitioner pursued what was essentially a “not me, maybe them” defense, namely, that he was not part of the group that attacked Fuller. Each tried to establish this defense by impeaching witnesses who had placed that particular petitioner at the scene. Some, for example, provided evidence that Eleby and Jacobs had used PCP the day of Fuller’s murder. Some also tried to establish alibis for the time of Fuller’s death. The jury convicted all seven petitioners, along with codefendant Steve Webb (who subsequently died). The jury acquitted codefendants Alfonso Harris and Felicia Ruffin. On direct appeal, the D. C. Court of Appeals affirmed petitioners’ convictions, though it remanded for resentencing. 545 A. 2d 1202, 1219 (1988). The trial court resentenced petitioners to the same amount of prison time. App. to Pet. for Cert. in No. 15–1503, at 82a, n. 2. B The Brady Claims Beginning in 2010, petitioners pursued postconviction proceedings in which they sought to vacate their convictions or to be granted a new trial. App. to Pet. for Cert. in No. 15–1503, at 84a, n. 4. After petitioners’ convictions became final, it emerged that the Government possessed certain evidence that it had withheld from the defense at the time of trial. Petitioners discovered other withheld evidence in their review of the trial prosecutor’s case file, which the Government turned over to petitioners in the course of the postconviction proceedings. Among other postconviction claims, petitioners contended that the withheld evidence was both favorable and material, entitling them to relief under Brady. The D. C. Superior Court considered petitioners’ Brady claims as part of a 16-day evidentiary hearing. It rejected those claims, finding that “none of the undisclosed information was material.” App. to Pet. for Cert. in No. 15–1503, at 130a. The D. C. Court of Appeals affirmed. 116 A. 3d, at 901. It similarly concluded that the withheld evidence was not material under Brady. 116 A. 3d, at 913–926. At issue in those proceedings were the following seven specific pieces of evidence: 1. The identity of James McMillan. Freeman, the vendor who discovered Fuller’s body in the alley garage, testified at trial that, while he was waiting for police to arrive, he saw two men run into the alley and stop near the garage for about five minutes before running away when an officer approached. One of the men had a bulge under his coat. Early in the trial, codefendant Harris’ counsel had requested the identity of the two men to confirm that her client was not one of them. But the Government refused to disclose the men’s identity. In their postconviction review of the prosecutor’s files, petitioners learned that Freeman had identified the two men he saw in the alley as James McMillan and Gerald Merkerson. McMillan lived in a house which opens in the back onto a connecting alley. In the weeks following Fuller’s murder, but before petitioners’ trial, McMillan was arrested for beating and robbing two women in the neighborhood. Neither attack included a sexual assault. Separately, petitioners learned that seven years after petitioners’ trial, McMillan had robbed, sodomized, and murdered a young woman in an alley. 2. The interview with Willie Luchie. The prosecutor’s notes also recorded an undisclosed interview with Willie Luchie, who told the prosecutor that he and three others walked through the alley on their way to an H Street liquor store between 5:30 and 5:45 p.m. on the evening of the murder. As the group walked by the garage, Luchie “heard several groans” and “remembers the doors to the garage being closed.” App. 25. Another person in the group recalled “hear[ing] some moans,” while the other two persons did not recall hearing anything unusual. Id., at 27, 53; id., at A992. The group continued walking without looking into the garage or otherwise investigating the source of the sounds. They did not see McMillan or any other person in the alley when they passed through. 3. The interviews with Ammie Davis. Undisclosed notes written by a police officer and the prosecutor refer to two interviews with Ammie Davis, who had been arrested for disorderly conduct a few weeks after Fuller’s murder. Davis initially told a police investigator that she had seen another individual, James Blue, beat Fuller to death in the alley. Shortly thereafter, she said she only saw Blue grab Fuller and push her into the alley. Davis also said that a girlfriend, whom she did not name, accompanied her. She promised to call the investigator with more details, but she did not do so. About 9 months later (after petitioners were indicted but approximately 11 weeks before their trial), a prosecutor learned of the investigator’s notes and interviewed Davis. The prosecutor’s notes state that Davis did not provide any more details, except to say that the girlfriend who accompanied her was nicknamed “ ‘Shorty.’ ” Id., at 267–268. About two months later, which was shortly before petitioners’ trial, Blue murdered Davis in an unrelated drug dispute. During the postconviction evidentiary hearing, the prosecutor who interviewed Davis testified that he did not disclose Davis’ statement because she acted “playful” and “not serious” during the interview and he found her to be “totally incredible.” Id., at 269–272. Additionally, the prosecutor stated that he knew Davis had previously falsely accused Blue of a different murder, and on another occasion had falsely accused a different individual of a different murder. 4. Impeachment of Kaye Porter and Carrie Eleby. Kaye Porter accompanied Eleby during an initial interview with homicide detectives. Porter agreed with Eleby that she had also heard Alston state that he was involved in robbing Fuller. An undisclosed prosecutorial note states that in a later interview with detectives, Porter stated that she did not actually recall hearing Alston’s statement and just went along with what Eleby said. The note also states that Eleby likewise admitted that she had lied about Porter being present during Alston’s statement and had asked Porter to support her. 5. Impeachment of Carrie Eleby. A prosecutor’s un-disclosed note revealed that Eleby said she had beenhigh on PCP during a January 9, 1985, meeting with investigators. 6. Impeachment of Linda Jacobs. An undisclosed note of an interview with Linda Jacobs said that the detective had “question[ed] her hard,” and that she had “vacillated” about what she saw. Id., at A1009. The prosecutor recalled that the detective “kept raising his voice” and was “smacking his hand on the desk” during the interview. Id., at A2298–A2299. 7. Impeachment of Maurice Thomas. An undisclosed note of an interview with Maurice Thomas’ aunt stated that she “does not recall Maurice ever telling her anything such as this.” Id., at A1010; see id., at 295–296. II A The Government does not contest petitioners’ claim that the withheld evidence was “favorable to the accused, either because it is exculpatory, or because it is impeaching.” Strickler v. Greene, 527 U. S. 263 –282 (1999). Neither does the Government contest petitioners’ claim that it “suppressed” the evidence, “either willfully or inadvertently.” Id., at 282. It does, as it must, concede that the Brady rule’s “ ‘overriding concern [is] with the justice of the finding of guilt,’ ” United States v. Bagley, 473 U. S. 667, 678 (1985) (quoting United States v. Agurs, 427 U. S. 97, 112 (1976) ), and that the Government’s “ ‘interest . . . in a criminal prosecution is not that it shall win a case, but that justice shall be done,’ ” Kyles v. Whitley, 514 U. S. 419, 439 (1995) (quoting Berger v. United States, 295 U. S. 78, 88 (1935) ). Consistent with these principles, the Government assured the Court at oral argument that subsequent to petitioners’ trial, it has adopted a “generous policy of discovery” in criminal cases under which it discloses any “information that a defendant might wish to use.” Tr. of Oral Arg. 47–48. As we have recognized, and as the Government agrees, ibid., “[t]his is as it should be.” Kyles, supra, at 439 (explaining that a “ ‘prudent prosecutor[’s]’ ” better course is to take care to disclose any evidence favorable to the defendant (quoting Agurs, supra, at 108)). Petitioners and the Government, however, do contest the materiality of the undisclosed Brady information. “[E]vidence is ‘material’ within the meaning of Brady when there is a reasonable probability that, had the evidence been disclosed, the result of the proceeding would have been different.” Cone v. Bell, 556 U. S. 449 –470 (2009) (citing Bagley, supra, at 682). “A ‘reasonable probability’ of a different result” is one in which the suppressed evidence “ ‘undermines confidence in the outcome of the trial.’ ” Kyles, supra, at 434 (quoting Bagley, supra, at 678). In other words, petitioners here are entitled to a new trial only if they “establis[h] the prejudice necessary to satisfy the ‘materiality’ inquiry.” Strickler, supra, at 282. Consequently, the issue before us here is legally simple but factually complex. We must examine the trial record, “evaluat[e]” the withheld evidence “in the context of the entire record,” Agurs, supra, at 112, and determine in light of that examination whether “there is a reasonable prob-ability that, had the evidence been disclosed, the result of the proceeding would have been different.” Cone, supra, at 470 (citing Bagley, supra, at 682). Having done so, we agree with the lower courts that there was no such reasonable probability. B Petitioners’ main argument is that, had they known about McMillan’s identity and Luchie’s statement, they could have challenged the Government’s basic theory that Fuller was killed in a group attack. Petitioners contend that they could have raised an alternative theory, namely, that a single perpetrator (or two at most) had attacked Fuller. According to petitioners, the groans that Luchie and his companion heard when they walked through the alley between 5:30 and 5:45 p.m. suggest that the attack was taking place inside the garage at that moment. The added facts that the garage was small and that Luchie’s group saw no one in the alley could bolster a “single attacker” theory. Freeman’s recollection that one garage door was open when he found Fuller’s body at around 6 p.m., combined with Luchie’s recollection that both doors were shut around 5:30 or 5:45 p.m., could suggest that one or two perpetrators were in the garage when Luchie walked by but left before Freeman arrived. McMillan’s identity as one of the men Freeman saw enter the alley after Freeman discovered Fuller’s body would have revealed McMillan’s criminal convictions in the months before petitioners’ trial. Petitioners argue that together, this evidence would have permitted the defense to knit together a theory that the group attack did not occur at all—and that it was actually McMillan, alone or with an accomplice, who murdered Fuller. They add that they could have used the investigators’ failure to follow up on Ammie Davis’ claim about James Blue, and the various pieces of withheld impeachment evidence, to suggest that an incomplete investigation had ended up accusing the wrong persons. Considering the withheld evidence “in the context of the entire record,” however, Agurs, supra, at 112, we conclude that it is too little, too weak, or too distant from the main evidentiary points to meet Brady’s standards. As petitioners recognize, McMillan’s guilt (or that of any other single, or near single, perpetrator) is inconsistent with petitioners’ guilt only if there was no group attack. But a group attack was the very cornerstone of the Government’s case. The witnesses may have differed on minor details, but virtually every witness to the crime itself agreed as to a main theme: that Fuller was killed by a large group of perpetrators. The evidence at trial was such that, even though petitioners knew that Freeman saw two men enter the alley after he discovered Fuller’s body, that one appeared to have a bulky object hidden under his coat, and that both ran when the police arrived, none of the petitioners attempted to mount a defense that implicated those men as alternative perpetrators acting alone. Is it reasonably probable that adding McMillan’s identity, and Luchie’s ambiguous statement that he heard groans but saw no one, could have led to a different result at trial? We conclude that it is not. The problem for petitioners is that their current alternative theory would have had to persuade the jury that both Alston and Bennett falsely confessed to being active participants in a group attack that never occurred; that Yarborough falsely implicated himself in that group attack and, through coordinated effort or coincidence, gave a highly similar account of how it occurred; that Thomas, a disinterested witness who recognized petitioners when he happened upon the attack and heard Catlett refer to it later that night, wholly fabricated his story; that both Eleby and Jacobs likewise testified to witnessing a group attack that did not occur; and that Montgomery in fact did not see petitioners and others, as a group, identify Fuller as a target and leave the park to rob her. With respect to the undisclosed impeachment evidence, the record shows that it was largely cumulative of impeachment evidence petitioners already had and used at trial. For example, the jury heard multiple times about Eleby’s frequent PCP use, including Eleby’s own testimony that she and Jacobs had smoked PCP shortly before they witnessed Fuller’s attack. In this context, it would not have surprised the jury to learn that Eleby used PCP on yet another occasion. Porter was a minor witness who was also impeached at trial with evidence about changes in her testimony over time, leaving little added significance to the note that she changed her mind about having agreed with Eleby’s claims. The jury was also well aware of Jacobs’ vacillation, as she was impeached on the stand with her shifting stories about what she witnessed. Knowledge that a detective raised his voice during an interview with her would have added little more. Nor do we see how the note about the statement by Thomas’ aunt could have mattered much, given the facts that neither side chose to call the aunt as a witness and that the jury already knew, from Thomas’ testimony, that his aunt had told him not to tell anyone what he saw. As for James Blue, petitioners argue that the investigators’ delay in following up on Ammie Davis’ statement could have led the jury to doubt the thoroughness of the investigation. But this likelihood is seriously undercut by notes about Davis’ demeanor and lack of detail, and by her prior false accusations that Blue committed a different murder and that yet another person committed yet a different murder. We of course do not suggest that impeachment evidence is immaterial with respect to a witness who has already been impeached with other evidence. See Wearry v. Cain, 577 U. S. ___, ___–___ (2016) (per curiam) (slip op., at 7–9). We conclude only that in the context of this trial, with respect to these witnesses, the cumulative effect of the withheld evidence is insufficient to “ ‘undermine confidence’ ” in the jury’s verdict, Smith, 565 U. S., at 75–76 (quoting Kyles, 514 U. S., at 434; brackets omitted). III On the basis of our review of the record, we agree with the lower courts that there is not a “reasonable probability” that the withheld evidence would have changed the outcome of petitioners’ trial, id., at 434 (internal quotation marks omitted). The judgment of the D. C. Court of Appeals, accordingly, is affirmed. It is so ordered. Justice Gorsuch took no part in the consideration or decision of these cases. APPENDIX

581.US.2016_15-423

The Foreign Sovereign Immunities Act (FSIA) shields foreign states from suits in United States Courts, 28 U. S. C. §1604, with specified exceptions. The expropriation exception applies to “any case . . . in which rights in property taken in violation of international law are in issue and that property . . . is owned or operated by an agency or instrumentality of the foreign state . . . engaged in a commercial activity in the United States.” §1605(a)(3). A wholly owned Venezuelan subsidiary (Subsidiary) of an American company (Parent) has long supplied oil rigs to oil development entities that were part of the Venezuelan Government. The American Parent and its Venezuelan Subsidiary (plaintiffs) filed suit in federal court against those entities (Venezuela), claiming that Venezuela had unlawfully expropriated the Subsidiary’s rigs by nationalizing them. Venezuela moved to dismiss the case on the ground that its sovereign immunity deprived the District Court of jurisdiction. Plaintiffs argued that the case falls within the expropriation exception, but Venezuela claimed that international law did not cover the expropriation of property belonging to a country’s nationals like the Subsidiary and that the American Parent did not have property rights in the Subsidiary’s assets. The District Court agreed as to the Subsidiary, dismissing its claim on jurisdictional grounds. But it rejected the claim that the Parent had no rights in the Subsidiary’s property. The District of Columbia Circuit reversed in part and affirmed in part, finding that both claims fell within the exception. With respect to the Subsidiary’s claim, it concluded that a sovereign’s taking of its own nationals’ property would violate international law if the expropriation unreasonably discriminated based on a company’s shareholders’ nationality. With respect to the Parent’s claim, it held that the exception applied because the Parent had raised its rights in a nonfrivolous way. The court decided only whether the plaintiffs might have a nonfrivolous expropriation claim, making clear that, under its standard, a nonfrivolous argument would be sufficient to bring a case within the scope of the exception. Given the factual stipulations, the court concluded, the Subsidiary had satisfied that standard for purposes of surviving a motion to dismiss. Held: The nonfrivolous-argument standard is not consistent with the FSIA. A case falls within the scope of the expropriation exception only if the property in which the party claims to hold rights was indeed “property taken in violation of international law.” A court should decide the foreign sovereign’s immunity defense “[a]t the threshold” of the action, Verlinden B. V. v. Central Bank of Nigeria, 461 U. S. 480 , resolving any factual disputes as near to the outset of the case as is reasonably possible. Pp. 6–16. (a) The expropriation exception grants jurisdiction only where there is a legally valid claim that a certain kind of right is at issue (property rights) and that the relevant property was taken in a certain way (in violation of international law). Simply making a nonfrivolous argument to that effect is not sufficient. This reading is supported by the provision’s language, which applies in a “case. . . in which rights in property taken in violation of international law are in issue.” Such language would normally foresee a judicial decision about the jurisdictional matter. This interpretation is supported by precedent. See, e.g., Permanent Mission of India to United Nations v. City of New York, 551 U. S. 193 –202. It is also supported by a basic objective of the FSIA, which is to follow international law principles, namely, that granting foreign sovereigns immunity from suit both recognizes the “absolute independence of every sovereign authority” and helps to “induc[e]” each nation state, as a matter of “international comity,” to “respect the independence and dignity of every other,” Berizzi Brothers Co. v. S. S. Pesaro, 271 U. S. 562 . Nothing in the FSIA’s history suggests that Congress intended a radical departure from these principles in codifying the mid-20th-century doctrine of “restrictive” sovereign immunity, which denies immunity in cases “arising out of a foreign state’s strictly commercial acts,” but applies immunity in “suits involving the foreign sovereign’s public acts,” Verlinden, supra, at 487. It is thus not surprising that the expropriation exception on its face emphasizes conformity with international law, requiring both a commercial connection with the United States and a taking of property “in violation of international law.” A “nonfrivolous-argument” reading of the exception would undermine the objectives embedded in the statute’s language, history, and structure. It could also embroil a foreign sovereign in an American lawsuit for some time by adopting a standard limited only by the bounds of a lawyer’s (nonfrivolous) imagination. And it could cause friction with other nations, leading to reciprocal actions against this country. Pp. 6–12. (b) Plaintiffs’ arguments to the contrary are unpersuasive. They suggest that the expropriation exception should be treated similarly to 28 U. S. C. §1331’s “arising under” jurisdiction, which applies if a plaintiff can make a nonfrivolous argument that a federal law provides the relief sought—even if, in fact, it does not, Bell v. Hood, 327 U. S. 678 685. But §1331 differs from the exception in language and concerns. Section 1331 often simply determines which court doors—federal or state—are open, and neither it nor related jurisdictional sections seek to provide a sovereign foreign nation with immunity—the FSIA’s basic objective. Nor does the text of §1331 suggest that consistency with international law is of particular importance. Plaintiffs also claim that the nonfrivolous-argument approach will work little harm since the matter could be resolved by motion practice before the sovereign bears the expense of a full trial. But resolving a claim pursuant to Federal Rule of Civil Procedure 12(b)(6) or summary judgment under Rule 56 may impose increased burdens of time and expense upon the foreign nation. And a district court’s decision that there is a “violation of international law” as a matter of jurisdiction may be immediately appealable as a collateral order, while the same decision made pursuant to a Rule 12(b)(6) or Rule 56 motion would be a decision on the “merits” not subject to immediate appeal. Moreover, the Circuit would part with its nonfrivolous-argument standard where a “violation of international law” is not an element of the claim to be decided on the merits. This bifurcated approach is difficult to reconcile with the statute’s language, history, or purpose; and it creates needless complexity for judges and lawyers, domestic and foreign. Pp. 12–16. 784 F. 3d 804, vacated and remanded. Breyer, J., delivered the opinion of the Court, in which all other Members joined, except Gorsuch, J., who took no part in the consideration or decision of the case.

The Foreign Sovereign Immunities Act of 1976 (FSIA or Act), provides, with specified exceptions, that a “foreign state shall be immune from the jurisdiction of the courts of the United States and of the States . . . .” 28 U. S. C. §1604. One of the jurisdictional exceptions—the expropriation exception—says that “[a] foreign state shall not be immune from the jurisdiction of courts of the United States or of the States in any case . . . (3) in which rights in property taken in violation of international law are in issue and that property . . . is owned or operated by an agency orinstrumentality of the foreign state . . . engaged ina commercial activity in the United States.” §1605(a)(3). The question here concerns the phrase “case . . . in which rights in property taken in violation of international law are in issue.” Does this phrase mean that, to defeat sovereign immu-nity, a party need only make a “nonfrivolous” argument that the case falls within the scope of the exception? Once made, does the existence of that nonfrivolous argument mean that the court retains jurisdiction over the case until the court decides, say, the merits of the case? Or does a more rigorous jurisdictional standard apply? To put the question more generally: What happens in a case where the party seeking to rely on the expropriation exception makes a nonfrivolous, but ultimately incorrect, claim that his property was taken in violation of international law? In our view, a party’s nonfrivolous, but ultimately incorrect, argument that property was taken in violation of international law is insufficient to confer jurisdiction. Rather, state and federal courts can maintain jurisdiction to hear the merits of a case only if they find that the property in which the party claims to hold rights was indeed “property taken in violation of international law.” Put differently, the relevant factual allegations must make out a legally valid claim that a certain kind of right is at issue (property rights) and that the relevant property was taken in a certain way (in violation of international law). A good argument to that effect is not sufficient. But a court normally need not resolve, as a jurisdictional matter, disputes about whether a party actually held rights in that prop-erty; those questions remain for the merits phase of the litigation. Moreover, where jurisdictional questions turn upon further factual development, the trial judge may take evidence and resolve relevant factual disputes. But, consistent with foreign sovereign immunity’s basic objective, namely, to free a foreign sovereign from suit, the court should normally resolve those factual disputes and reach a decision about immunity as near to the outset of the case as is reasonably possible. See Verlinden B. V. v. Central Bank of Nigeria, 461 U. S. 480 –494 (1983). I Since the mid-1970’s a wholly owned Venezuela-incorporated subsidiary (Subsidiary) of an American company (Parent) supplied oil rigs to oil development entities that were part of the Venezuelan Government. In 2011 the American Parent company and its Venezuelan Subsidiary (the respondents here) brought this lawsuit in federal court against those foreign government entities. (The entities go by their initials, PDVSA, but we shall normally refer to them as “Venezuela” or the “Venezuelan Government.”) The American Parent and the Venezuelan Subsidiary claimed that the Venezuelan Government had unlawfully expropriated the Subsidiary’s oil rigs. And they sought compensation. According to stipulated facts, by early 2010 the Venezuelan Government had failed to pay more than $10 million that it owed the Subsidiary. At that point the government sent troops to the equipment yard where the rigs were stored, prevented the Subsidiary from removing the rigs, and issued a “ ‘Decree of Expropriation’ ” nationalizing the rigs. App. 72–74. Subsequently, the president of the oil development entities led a rally at the Subsidiary’s offices, where he referred to the Venezuelan Subsidiary as an “ ‘American company’ ” with “ ‘foreign gentlemen investors.’ ” Id., at 54. Venezuela asked the court to dismiss the case on the ground that Venezuela possessed sovereign immunity and that the court consequently lacked “jurisdiction” to hear the case. See 28 U. S. C. §1604; Fed. Rules Civ. Proc. 12(b)(1) and (b)(2); Verlinden, supra, at 485, n. 5 (explaining that a court lacks “subject-matter” and “personal” jurisdiction over a foreign sovereign unless an FSIA exception applies). The companies replied that the case falls within the expropriation exception. Venezuela in turn argued that the Subsidiary’s expropriation claim did not satisfy the exception because “ ‘international law does not cover expropriations of property belonging to a country’s own nationals’ ”; the taking was not “ ‘in violation of international law,’ ” and the exception thus does not apply. Record in No. 11–cv–01735 (D DC), Doc. 22, p. 13. Venezuela further argued that the American Parent’s nationality makes no difference because, “as a corporate parent, [it] does not own [the Subsidiary’s] assets.” Id., Doc. 24,at 12. The parties agreed that the District Court should then decide whether the exception applies, and it should do so on the basis of governing law, taking all of the plaintiffs’ well-pleaded allegations as true and construing the complaint in the light most favorable to the plaintiffs. App. 119. The court decided, in relevant part, that the exception did not apply to the Venezuelan Subsidiary’s claim because the Subsidiary was a national of Venezuela. See 971 F. Supp. 2d 49, 57–61 (2013). The court concluded that Venezuela consequently possessed sovereign immu-nity, and it dismissed the Subsidiary’s claim on jurisdictional grounds. It rejected, however, Venezuela’s argument that the Parent had no rights in property in the Subsidiary. It concluded that Venezuela’s “actions have deprived [the Parent], individually, of its essential and unique rights as sole shareholder . . . by dismantling its voting power, destroying its ownership, and frustrating its control over the company.” Id., at 73. The Venezuelan Subsidiary appealed the dismissal of its expropriation claim, and Venezuela appealed the court’s refusal to dismiss the Parent’s claim. The Court of Appeals for the District of Columbia Circuit reversed in part and affirmed in part the District Court’s conclusions. It decided that both the Subsidiary’s and the Parent’s claims fell within the exception. With respect to the Subsidiary’s claim, the court agreed that a sovereign’s taking of its own nationals’ property normally does not violate international law. But, the court said, there is an “exception” to this rule. And that exception applies when a sovereign’s expropriation unreasonably discriminates on the basis of a company’s shareholders’ nationality, 784 F. 3d 804, 812 (CADC 2015) (citing Banco Nacional de Cuba v. Sabbatino, 307 F. 2d 845 (CA2 1962)). That exception, it added, might apply here, in which case the expropriation would violate international law, the FSIA’s expropriation exception would apply, and the federal courts would possess jurisdiction over the case. 784 F. 3d, at 813. With respect to the Parent’s expropriation claim, the court agreed with the District Court that the expropriation exception applied because the Parent had “ ‘put its rights in property in issue in a non-frivolous way.’ ” Id., at 816. For present purposes, it is important to keep in mind that the Court of Appeals did not decide (on the basis of the stipulated facts) that the plaintiffs’ allegations are sufficient to show their property was taken in violation of international law. It decided instead that the plaintiffs might have such a claim. And it made clear the legal standard that it would apply. It said that, in deciding whether the expropriation exception applies, it would set an “exceptionally low bar.” Id., at 812. Any possible, i.e., “ ‘non-frivolous,’ ” ibid., claim of expropriation is sufficient, in the Court of Appeals’ view, to bring a case within the scope of the FSIA’s exception. In particular: If a plaintiff alleges facts and claims that permit the plaintiff to make an expropriation claim that is not “ ‘wholly insubstantial or frivolous,’ ” then the exception permits the suit and the sovereign loses its immunity. Ibid. (emphasis added). Given the factual stipulations, the Court of Appeals did not suggest further factfinding on this jurisdictional issue but, rather, decided that the Subsidiary had “satisfied this Circuit’s forgiving standard for surviving a motion to dismiss in an FSIA case.” Id., at 813. Venezuela filed a petition for certiorari asking us to decide whether the Court of Appeals had applied the correct standard in deciding that the companies had met the expropriation exception’s requirements. We agreed to do so. II Foreign sovereign immunity is jurisdictional in this case because explicit statutory language makes it so. See §1604 (“[A] foreign state shall be immune from the jurisdiction of the courts of the United States and of the States except as provided” by the FSIA’s exceptions); §1605(a) (“A foreign state shall not be immune from the jurisdiction” of federal and state courts if the exception at issue here is satisfied). Given the parties’ stipulations as to all relevant facts, our inquiry poses a “ ‘pure question of statutory construction,’ ” Republic of Austria v. Altmann, 541 U. S. 677, 701 (2004) .In our view, the expropriation exception grants jurisdiction only where there is a valid claim that “property” has been “taken in violation of international law.” §1605(a)(3). A nonfrivolous argument to that effect is insufficient. For one thing, the provision’s language, while ambiguous, supports such a reading. It says that there is jurisdiction in a “case . . . in which rights in property taken in violation of international law are in issue.” Ibid. Such language would normally foresee a judicial decision about the jurisdictional matter. And that matter is whether a certain kind of “right” is “at issue,” namely, a property right taken in violation of international law. To take a purely hypothetical example, a party might assert a claim to a house in a foreign country. If the foreign country nationalized the house and, when sued, asserted sovereign immunity, then the claiming party would as a jurisdictional matter prove that he claimed “property” (which a house obviously is) and also that the property was “taken in violation of international law.” He need not show as a jurisdictional matter that he, rather than someone else, owned the house. That question is part of the merits of the case and remains “at issue.” We recognize that merits and jurisdiction will sometimes come intertwined. Suppose that the party asserted a claim to architectural plans for the house. It might be necessary to decide whether the law recognizes the kind of right that he asserts, or whether it is a right in “property” that was “taken in violation of international law.” Perhaps that is the only serious issue in the case. If so, the court must still answer the jurisdictional question. If to do so, it must inevitably decide some, or all, of the merits issues, so be it. Our reading of the statute is consistent with its language. The case is one which the existence of “rights” remains “at issue” until the court decides the merits of the case. But whether the rights asserted are rights of a certain kind, namely, rights in “property taken in violation of international law,” is a jurisdictional matter that the court must typically decide at the outset of the case, or as close to the outset as is reasonably possible. Precedent offers a degree of support for our interpretation. In Permanent Mission of India to United Nations v. City of New York, 551 U. S. 193 (2007) , we interpreted a different FSIA exception for cases “in which . . . rights in immovable property situated in the United States are in issue.” §1605(a)(4). We held that there was jurisdiction over the case because the plaintiff’s lawsuit to enforce a tax lien “directly implicate[d]” the property rights described by the FSIA exception. See id., at 200–201. We did not simply rely upon a finding that the plaintiff had made a nonfrivolous argument that the exception applied. For another thing, one of the FSIA’s basic objectives, as shown by its history, supports this reading. The Act for the most part embodies basic principles of international law long followed both in the United States and elsewhere. See Schooner Exchange v. McFaddon, 7 Cranch 116, 136–137 (1812); see also Verlinden, 461 U. S., at 493 (explaining that the Act “comprehensively regulat[es] the amen-ability of foreign nations to suit in the United States”). Our courts have understood, as international law itself understands, foreign nation states to be “independent sovereign” entities. To grant those sovereign entities an immunity from suit in our courts both recognizes the “absolute independence of every sovereign authority” and helps to “ ‘induc[e]’ ” each nation state, as a matter of “ ‘international comity,’ ” to “ ‘respect the independence and dignity of every other,’ ” including our own. Berizzi Brothers Co. v. S. S. Pesaro, 271 U. S. 562, 575 (1926) (quoting The Parlement Belge, [1880] 5 P. D. 197, 214–215 (appeal taken from Admiralty Div.)). In the mid-20th century, we, like many other nations, began to treat nations acting in a commercial capacity like other commercial entities. See Permanent Mission, supra, at 199–200. And we consequently began to limit our recognition of sovereign immunity, denying that immunity in cases “arising out of a foreign state’s strictly commercial acts,” but continuing to apply that doctrine in “suits involving the foreign sovereign’s public acts,” Verlinden, 461 U. S., at 487 (emphasis added). At first, our courts, aware of the expertise of the Executive Branch in matters of foreign affairs, relied heavily upon the advice of that branch when deciding just when and how this “restrictive” sovereign immunity doctrine applied. Ibid. See also H. R. Rep. No. 94–1487, pp. 8–9 (1976) (similar). But in 1976, Congress, at the urging of the Department of State and Department of Justice, began to codify the doctrine. The resulting statute, the FSIA, “starts from a premise of immunity and then creates exceptions to the general principle.” Id., at 17; Verlinden, supra, at 493. Almost all the exceptions involve commerce or immovable property located in the United States. E.g., §§1605(a)(2) and (4); see also §1602 (expressing the finding that “[u]nder international law, states are not immune from the jurisdiction of foreign courts insofar as their commercial activities are concerned”). The statute thereby creates a doctrine that by and large continues to reflect basic principles of international law, in particular those principles embodied in what jurists refer to as the “restrictive” theory of sovereign immunity. See, e.g., Restatement (Third) of Foreign Relations Law of the United States §451, and Comment a (1986) (describing the restrictive theory of immunity); United Nations General Assembly, Convention on Jurisdictional Immunities of States and Their Property, Res. 59/38, Arts. 5, 10–12 (Dec. 2, 2004) (adopting a restrictive theory of immunity and withdrawing immunity for loss of property where, among other requirements, “the act or omission occurred in whole or in part in the territory of th[e] other State”); United Nations General Assembly, Report of the Ad Hoc Committee on Jurisdictional Immunities of States and Their Property, Supp. A/59/22 No. 1, pp. 7–11 (Mar. 1–5, 2004) (same). We have found nothing in the history of the statute that suggests Congress intended a radical departure from these basic principles. To the contrary, the State Department, which helped to draft the FSIA’s language (and to whose views on sovereign immunity this Court, like Congress, has paid special attention, Altmann, 541 U. S., at 696), told Congress that the Act was “drafted keeping in mind what we believe to be the general state of the law internationally, so that we conform fairly closely . . . to our accepted international standards,” Hearing on H. R. 3493 before the Subcommittee on Claims and Governmental Relations of the House of Representatives Committee on the Judiciary, 93d Cong., 1st Sess., 18 (1973). The Department added that, by doing so, we would diminish the likelihood that other nations would each go their own way, thereby “subject[ing]” the United States “abroad” to more claims “than we permit in this country . . . .” Ibid. It is consequently not surprising to find that the expropriation exception on its face emphasizes conformity with international law by requiring not only a commercial connection with the United States but also a taking of property “in violation of international law.” We emphasize this point, embedded in the statute’s language, history, and structure, because doing so reveals a basic objective of our sovereign immunity doctrine, which a “nonfrivolous-argument” reading of the expropriation exception would undermine. A sovereign’s taking or regulating of its own nationals’ property within its own territory is often just the kind of foreign sovereign’s public act (a “jure imperii”) that the restrictive theory of sovereign immunity ordinarily leaves immune from suit. See Permanent Mission, 551 U. S., at 199 (describing the FSIA’s distinction between public acts, or jure imperii, and purely commercial ones); Restatement (Third) of Foreign Relations Law of the United States §712, at 196 (noting that, under international law, a state is responsible for a “taking of the property of a national of another state” (emphasis added)). See also Restatement (Fourth) of Foreign Relations Law of the United States §455, Reporter’s Note 12, p. 9 (Tent. Draft No. 2, Mar. 22, 2016) (noting that “[n]o provision comparable” to the exception “has yet been adopted in the domestic immunity statutes of other countries” and that expropriations are considered acts jure imperii); United States v. Belmont, 301 U. S. 324, 332 (1937) ; B. Cheng & G. Schwarzberger, General Principles of Law as Applied by International Courts and Tribunals 37–38 (1953) (collecting cases describing “the power of the sovereign State to expropriate” (internal quotation marks omitted)); Jurisdictional Immunities of the State (Germany v. Italy), 2012 I. C. J. 99, 123–125, ¶¶56–60 (Judgt. of Feb. 3) (noting consistent state practice in respect to the distinction between public and commercial acts and describing an international law of immunity recognizing such a difference); Altmann, supra, at 708 (Breyer, J., concurring) (describing the French Court of Appeals’ decision about whether a King who has abdicated the throne is “ ‘entitled to claim . . . immunity’ ” as “ ‘Hea[d] of State’ ” when his sovereign status at the time of suit was in doubt (quoting Ex-King Farouk of Egypt v. Christian Dior, 84 Clunet 717, 24 I. L. R. 228, 229 (CA Paris 1957))). To be sure, there are fair arguments to be made that a sovereign’s taking of its own nationals’ property sometimes amounts to an expropriation that violates international law, and the expropriation exception provides that the general principle of immunity for these otherwise public acts should give way. But such arguments are about whether such an expropriation does violate international law. To find jurisdiction only where a taking does violate international law is thus consistent with basic international law and the related statutory objectives and principles that we have mentioned. But to find jurisdiction where a taking does not violate international law (e.g., where there is a nonfrivolous but ultimately incorrect argument that the taking violates international law) is inconsistent with those objectives. And it is difficult to understand why Congress would have wanted that result. Moreover, the “nonfrivolous-argument” interpretation would, in many cases, embroil the foreign sovereign in an American lawsuit for an increased period of time. It would substitute for a more workable standard (“violation of international law”) a standard limited only by the bounds of a lawyer’s (nonfrivolous) imagination. It would create increased complexity in respect to a jurisdictional matter where clarity is particularly important. Hertz Corp. v. Friend, 559 U. S. 77 –95 (2010). And clarity is doubly important here where foreign nations and foreign lawyers must understand our law. Finally, the Solicitor General and the Department of State also warn us that the nonfrivolous-argument interpretation would “affron[t]” other nations, producing friction in our relations with those nations and leading some to reciprocate by granting their courts permission to embroil the United States in “expensive and difficult litigation, based on legally insufficient assertions that sovereign immunity should be vitiated.” Brief for United States as Amicus Curiae 21–22. (At any given time the Department of Justice’s Office of Foreign Litigation represents the United States in about 1,000 cases in 100 courts around the world. Ibid.) See also National City Bank of N. Y. v. Republic of China, 348 U. S. 356, 362 (1955) (noting that our grant of immunity to foreign sovereigns dovetails with our own interest in receiving similar treatment). III The plaintiffs make two important arguments to the contrary. First, they point to the federal statute that gives federal courts jurisdiction over cases “arising under the Constitution, laws, or treaties of the United States,” 28 U. S. C. §1331. They note that in Bell v. Hood, 327 U. S. 678 (1946) , this Court held that the “arising under” statute confers jurisdiction if a plaintiff can make a nonfrivolous argument that a federal law provides the relief he seeks—even if, in fact, it does not. See id., at 685 (jurisdiction exists where, if the “Constitution and laws of the United States are given one construction,” a claim will be “sustained,” but if the laws are given a different construction, the claim “will be defeated”). And the plaintiffs say we should treat the expropriation exception similarly. Section 1331, however, uses different language from the expropriation exception (“arising under”) and focuses on different concerns. Section 1331 often simply determines which court’s doors are open (federal or state). Cf. Mims v. Arrow Financial Services, LLC, 565 U. S. 368 –379 (2012). Unlike the FSIA, neither that jurisdictional section nor related jurisdictional sections seeks to provide a sovereign foreign nation (or any party) with immunity—the basic FSIA objective. See Dole Food Co. v. Patrickson, 538 U. S. 468, 479 (2003) (FSIA’s objective is to give “protection from the inconvenience of suit as a gesture of comity”); Republic of Philippines v. Pimentel, 553 U. S. 851, 866 (2008) . And unlike the expropriation exception, the “arising under” statute’s language does not suggest that consistency with international law is of particular importance. Moreover, this Court has interpreted other jurisdictional statutes differently. Where jurisdiction depends on diversity of citizenship, for example, courts will look to see whether the parties are in fact diverse, not simply whether they are arguably so. See Indianapolis v. Chase Nat. Bank, 314 U. S. 63, 69 (1941) ; McNutt v. General Motors Acceptance Corp., 298 U. S. 178, 189 (1936) ; see also 13E C. Wright, A. Miller, & E. Cooper, Federal Practice and Procedure §3611 (2009). We do not believe either jurisdictional analogy ( 28 U. S. C. §1331 or §1332) is particularly helpful, but the expropriation exception’s substantive goals suggest that the diversity jurisdiction example provides a marginally closer analogy. Second, the plaintiffs argue that the nonfrivolous-argument approach will work little harm. They say that a court faced with an arguable, but ultimately incorrect, claim of jurisdiction can simply decide the same question—say, whether there was a “violation of international law”—as part of its decision on the merits. Thus a foreign sovereign defendant (in court because a plaintiff has made a nonfrivolous but incorrect argument that its property was taken in violation of international law) can simply move for judgment on the merits under Rule 12(b)(6), which provides for judgment where a plaintiff does not “state a claim upon which relief can be granted.” Or the defendant could move for summary judgment under Rule 56. In a word, the defendant may not need to undergo a full trial and judgment, remaining in court until the bitter end. These alternatives, however, have their own problems. For one thing, they will sometimes mean increased delay, imposing increased burdens of time and expense upon the foreign nation. For another, where a district court decides that there is a “violation of international law” as a matter of jurisdiction, then (according to the Courts of Appeals) the losing sovereign nation can immediately appeal the decision as a collateral order. But the same decision made to dispose of, say, a Rule 12(b)(6) motion or a Rule 56 motion would not be a “collateral order.” It would be a decision on the “merits.” And the foreign sovereign would not enjoy a right to take an immediate appeal. See Coopers & Lybrand v. Livesay, 437 U. S. 463, 468 (1978) (permitting interlocutory appeal of a collateral order that “resolve[s] an important issue completely separate from the merits of the action”); Will v. Hallock, 546 U. S. 345, 349 (2006) (same). See also Intel Corp. v. Commonwealth Scientific, 455 F. 3d 1364, 1366 (CA Fed. 2006) (permitting collateral appeal of an FSIA jurisdictional decision denying immunity); Rubin v. Islamic Republic of Iran, 637 F. 3d 783, 785 (CA7 2011) (same); Compania Mexicana de Aviacion v. Central Dist. of Cal., 859 F. 2d 1354, 1356 (CA9 1988) (per curiam) (same); Foremost-McKesson v. Islamic Republic of Iran, 905 F. 2d 438, 443 (CADC 1990) (same). Moreover, what is a court to do in a case where a “violation of international law,” while a jurisdictional prerequisite, is not an element of the claim to be decided on the merits? The Circuit has suggested that they arise when the plaintiffs’ claim is not an “expropriation claim,” but rather a simple common-law claim of conversion, restitution, or breach of contract, the merits of which do not involve the merits of international law. See Simon v. Republic of Hungary, 812 F. 3d 127, 141–142 (2016). The Circuit has recognized that there are such cases, id., at 141, and a cursory survey of the principal district courts in which these cases are brought confirms the reality of the problem. See, e.g., Philipp v. Federal Republic of Ger-many, ___ F. Supp. 3d ___, 2017 WL 1207408 (DC, Mar. 31, 2017) (deciding whether the expropriation exception is satisfied where the complaint pleads only common-law or statutory claims for relief); De Csepel v. Hungary, 169 F. Supp. 3d 143 (DC 2016) (similar); Pablo Star Ltd. v. Welsh Government, 170 F. Supp. 3d 597 (SDNY 2016) (similar); Chettri v. Nepal, 2014 WL 4354668 (SDNY, Sept. 2, 2014) (similar); Order Granting Defendants’ Motion To Dismiss in Lu v. Central Bank of Republic ofChina, No. 2:12–cv–317 (CD Cal., June 13, 2013) (similar); Orkin v. Swiss Confederation, 770 F. Supp. 2d 612 (SDNY 2011) (similar); Hammerstein v. Federal Republic of Germany, 2011 WL 9975796 (EDNY, Aug. 1, 2011) (similar); Cassirer v. Kingdom of Spain, 461 F. Supp. 2d 1157 (CD Cal. 2006) (similar). Indeed, cases in which the jurisdictional inquiry does not overlap with the elements of a plaintiff’s claims have been the norm in cases arising under other exceptions to the FSIA. E.g., Republic of Argentina v. Weltover, Inc., 504 U. S. 607, 610 (1992) (deciding whether a plaintiffs’ breach-of-contractclaim satisfied the jurisdictional requirements of the commercial-activity exception, §1605(a)(2)). To address the problem raised by these cases in which the “jurisdictional and merits inquiries” are not fully “overlap[ping],” the Circuit has held that a district court is not to apply its nonfrivolous-argument standard in such cases. Simon, 812 F. 3d, at 141. Rather, a court is to ask “whether the plaintiffs’ allegations satisfy the jurisdictional standard.” Ibid. We can understand why the Circuit has departed from its nonfrivolous-argument standard in these latter cases. For, unless it did so, how could a foreign nation ever obtain a decision on the merits of the nonfrivolous argument that a plaintiff has advanced? But what in the statutory provision suggests that sometimes courts should, but sometimes they should not, simply look to the existence of a nonfrivolous argument when they decide whether the requirements of the expropriation exception are satisfied? It is difficult, if not impossible, to reconcile this bifurcated approach with the statute’s language. It receives little, if any, support from the statute’s history or purpose. And, it creates added complexity, making it more difficult for judges and lawyers, domestic and foreign, to understand the intricacies of the law. IV We conclude that the nonfrivolous-argument standard is not consistent with the statute. Where, as here, the facts are not in dispute, those facts bring the case within the scope of the expropriation exception only if they do show (and not just arguably show) a taking of property in violation of international law. Simply making a nonfrivolous argument to that effect is not sufficient. Moreover, as we have previously stated, a court should decide the foreign sovereign’s immunity defense “[a]t the threshold” of the action. Verlinden, 461 U. S., at 493. As we have said, given the parties’ stipulations as to all relevant facts, the question before us is purely a legal one and can be resolved at the outset of the case. If a decision about the matter requires resolution of factual disputes, the court will have to resolve those disputes, but it should do so as near to the outset of the case as is reasonably possible. * * * The judgment of the Court of Appeals is vacated, and the case is remanded for further proceedings consistent with this opinion. It is so ordered. Justice Gorsuch took no part in the consideration or decision of this case.

581.US.2016_16-254

Petitioner Water Splash sued respondent Menon, a former employee, in a Texas state court, alleging that she had begun working for a competitor while still employed by Water Splash. Because Menon resided in Canada, Water Splash obtained permission to effect service by mail. After Menon declined to answer or otherwise enter an appearance, the trial court issued a default judgment for Water Splash. That court subsequently denied Menon’s motion to set aside the judgment on the ground that she had not been properly served. On appeal, Menon argued that service by mail does not comport with the requirements of the Convention on the Service Abroad of Judicial and Extrajudicial Documents in Civil and Commercial Matters (Hague Service Convention), which seeks to simplify, standardize, and generally improve the process of serving documents abroad, specifying certain approved methods of service and preempting “inconsistent methods of service” wherever it applies, Volkswagenwerk Aktiengesellschaft v. Schlunk, 486 U. S. 694 . The Texas Court of Appeals agreed with Menon, holding that the Convention prohibited service of process by mail. Article 10, the provision at issue, consists of Articles 10(b) and 10(c), which plainly address permissible methods of “service,” and Article 10(a), which provides that the Convention will not interfere with “the freedom to send judicial documents, by postal channels, directly to persons abroad,” but does not expressly refer to “service.” Held: The Hague Service Convention does not prohibit service of process by mail. Pp. 4–12. (a) This Court begins its analysis by looking to the treaty’s text and the context in which its words are used. See Schlunk, 486 U. S., at 699. The key word in Article 10(a)—“send”—is a broad term, and there is no apparent reason why it would exclude the transmission of documents for the purpose of service. The structure of the Convention strongly counsels against such an exclusion. The Convention’s preamble and Article 1 limit the scope of the Convention to service of documents abroad, and its full title includes the phrase “Service Abroad.” This Court has also held that the scope of the Convention is limited to service of documents. Id., at 701. It would thus be quite strange if Article 10(a)—apparently alone among the Convention’s provisions—concerned something other than service of documents. Indeed, such a reading would render Article 10(a) superfluous. Article 10’s function is to ensure that, generally, the Convention “shall not interfere” with the activities described in 10(a), 10(b), and 10(c). But since Article 1 already “eliminates [the] possibility” that the Convention would apply to any communications that “do not culminate in service,” id., at 701, in order for Article 10(a) to do any work, it must pertain to sending documents for the purposes of service. Menon’s attempt to avoid this superfluity problem by suggesting that Article 10(a) applies not to service of process but only to the service of “post-answer judicial documents” lacks any plausible textual footing in Article 10. If the drafters wished to limit Article 10(a) to a particular subset of documents, they could have said so—as they did, e.g., in Article 15, which refers to “a writ of summons or an equivalent document.” Instead, Article 10(a) uses the term “judicial documents”—the same term featured in 10(b) and 10(c). And the ordinary meaning of the word “send” is broad enough to cover the transmission of any judicial documents. Accordingly, the text and structure of the Convention indicate that Article 10(a) encompasses service by mail. Pp. 4–6. (b) The main counterargument—that Article 10(a)’s phrase “send judicial documents” should mean something different than the phrase “effect service of judicial documents” in Article 10(b) and Article 10(c)—is unpersuasive. First, it must contend with the compelling structural considerations strongly suggesting that Article 10(a) pertains to service of documents. Second, reading the word “send” as a broad concept that includes, but is not limited to, service is probably more plausible than interpreting the word to exclude service, and it does not create the same superfluity problem. Third, the French version of the Convention, which is “equally authentic” to the English version, Schlunk, supra, at 699, uses the word “adresser,” which has consistently been understood to mean service or notice. At best, Menon’s argument creates an ambiguity as to Article 10(a)’s meaning. The Court thus turns to additional tools of treaty interpretation, which comfortably resolve any lingering ambiguity in Water Splash’s favor. Pp. 7–8. (c) Three extratextual sources are especially helpful in ascertaining Article 10(a)’s meaning. First, the Convention’s drafting history strongly suggests that the drafters understood that service by postal channels was permissible. Second, in the half-century since the Convention was adopted, the Executive Branch has consistently maintained that the Hague Service Convention allows service by mail. Finally, other signatories to the Convention have consistently adopted Water Splash’s view. Pp. 8–12. (d) The fact that Article 10(a) encompasses service by mail does not mean that it affirmatively authorizes such service. Rather, service by mail is permissible if the receiving state has not objected to service by mail and if such service is authorized under otherwise-applicable law. Because the Court of Appeals concluded that the Convention prohibited service by mail, it did not consider whether Texas law authorizes the methods of service used by Water Splash. That and any other remaining issues are left to be considered on remand to the extent they are properly preserved. P. 12. 472 S. W. 3d 28, vacated and remanded. Alito, J., delivered the opinion of the Court, in which all other Members joined, except Gorsuch, J., who took no part in the consideration or decision of the case.

This case concerns the scope of the Convention on the Service Abroad of Judicial and Extrajudicial Documents in Civil and Commercial Matters, Nov. 15, 1965 (Hague Service Convention), 20 U. S. T. 361, T. I. A. S. No. 6638. The purpose of that multilateral treaty is to simplify, standardize, and generally improve the process of serving documents abroad. Preamble, ibid.; see Volkswagenwerk Aktiengesellschaft v. Schlunk, 486 U. S. 694, 698 (1988) . To that end, the Hague Service Convention specifies certain approved methods of service and “pre-empts inconsistent methods of service” wherever it applies. Id., at 699. Today we address a question that has divided the lower courts: whether the Convention prohibits service by mail. We hold that it does not. I A Petitioner Water Splash is a corporation that produces aquatic playground systems. Respondent Menon is a former employee of Water Splash. In 2013, Water Splash sued Menon in state court in Texas, alleging that she had begun working for a competitor while still employed by Water Splash. 472 S. W. 3d 28, 30 (Tex. App. 2015). Water Splash asserted several causes of action, including unfair competition, conversion, and tortious interference with business relations. Because Menon resided in Can-ada, Water Splash sought and obtained permission to effect service by mail. Ibid. After Menon declined to answer or otherwise enter an appearance, the trial court issued a default judgment in favor of Water Splash. Menon moved to set aside the judgment on the ground that she had not been properly served, but the trial court denied the motion. Ibid. Menon appealed, arguing that service by mail does not “comport with the requirements of the Hague Service Convention.” Ibid. The Texas Court of Appeals majority sided with Menon and held that the Convention prohibits service of process by mail. Id., at 32. Justice Christopher dissented. Id., at 34. The Court of Appeals declined to review the matter en banc, App. 95–96, and the Texas Supreme Court denied discretionary review, id., at 97–98. The disagreement between the panel majority and Justice Christopher tracks a broader conflict among courts as to whether the Convention permits service through postal channels. Compare, e.g., Bankston v. Toyota Motor Corp., 889 F. 2d 172, 173–174 (CA8 1989) (holding that the Convention prohibits service by mail), and Nuovo Pignone, SpA v. Storman Asia M/V, 310 F. 3d 374, 385 (CA5 2002) (same), with, e.g., Brockmeyer v. May, 383 F. 3d 798, 802 (CA9 2004) (holding that the Convention allows service by mail), and Ackermann v. Levine, 788 F. 2d 830, 838–840 (CA2 1986) (same). We granted certiorari to resolve that conflict. 580 U. S. ___ (2016). B The “primary innovation” of the Hague Service Convention—set out in Articles 2–7—is that it “requires each state to establish a central authority to receive requests for service of documents from other countries.” Schlunk, supra, at 698. When a central authority receives an appropriate request, it must serve the documents or arrange for their service, Art. 5, and then provide a certificate of service, Art. 6. Submitting a request to a central authority is not, however, the only method of service approved by the Convention. For example, Article 8 permits service through diplomatic and consular agents; Article 11 provides that any two states can agree to methods of service not otherwise specified in the Convention; and Article 19 clarifies that the Convention does not preempt any internal laws of its signatories that permit service from abroad via methods not otherwise allowed by the Convention. At issue in this case is Article 10 of the Convention, the English text of which reads as follows: “Provided the State of destination does not object, the present Convention shall not interfere with— “(a) the freedom to send judicial documents, by postal channels, directly to persons abroad, “(b) the freedom of judicial officers, officials or other competent persons of the State of origin to effect service of judicial documents directly through the judicial officers, officials or other competent persons of the State of destination, “(c) the freedom of any person interested in a judicial proceeding to effect service of judicial documents directly through the judicial officers, officials or other competent persons of the State of destination.” 20 U. S. T., at 363. Articles 10(b) and 10(c), by their plain terms, address additional methods of service that are permitted by the Convention (unless the receiving state objects). By contrast, Article 10(a) does not expressly refer to “service.” The question in this case is whether, despite this textual difference, the Article 10(a) phrase “send judicial documents” encompasses sending documents for the purposes of service. II A In interpreting treaties, “we begin with the text of the treaty and the context in which the written words are used.” Schlunk, 486 U. S., at 699 (internal quotation marks omitted). For present purposes, the key word in Article 10(a) is “send.” This is a broad term,[1] and there is no apparent reason why it would exclude the transmission of documents for a particular purpose (namely, service). Moreover, the structure of the Hague Service Convention strongly counsels against such a reading. The key structural point is that the scope of the Convention is limited to service of documents. Several elements of the Convention indicate as much. First, the preamble states that the Convention is intended “to ensure that judicial and extrajudicial documents to be served abroad shall be brought to the notice of the addressee in sufficient time.” (Emphasis added.) And Article 1 defines the Convention’s scope by stating that the Convention “shall apply in all cases, in civil or commercial matters, where there is occasion to transmit a judicial or extrajudicial document for service abroad.” (Emphasis added.) Even the Convention’s full title reflects that the Convention concerns “Service Abroad.” We have also held as much. Schlunk, 486 U. S., at 701 (stating that the Convention “applies only to documents transmitted for service abroad”). As we explained, a preliminary draft of Article 1 was criticized “because it suggested that the Convention could apply to transmissions abroad that do not culminate in service.” Ibid. The final version of Article 1, however, “eliminates this possibility.” Ibid. The wording of Article 1 makes clear that the Convention “applies only when there is both transmission of a document from the requesting state to the receiving state, and service upon the person for whom it is intended.” Ibid. In short, the text of the Convention reveals, and we have explicitly held, that the scope of the Convention is limited to service of documents. In light of that, it would be quite strange if Article 10(a)—apparently alone among the Convention’s provisions—concerned something other than service of documents. Indeed, under that reading, Article 10(a) would be superfluous. The function of Article 10 is to ensure that, absent objection from the receiving state, the Convention “shall not interfere” with the activities described in 10(a), 10(b) and 10(c). But Article 1 already “eliminates [the] possibility” that the Convention would apply to any communications that “do not culminate in service,” id., at 701, so it is hard to imagine how the Convention could interfere with any non-service communications. Accordingly, in order for Article 10(a) to do any work, it must pertain to sending documents for the purposes of service. Menon attempts to avoid this superfluity problem by suggesting that Article 10(a) does refer to serving documents—but only some documents. Specifically, she makes a distinction between two categories of service. According to Menon, Article 10(a) does not apply to service of process (which we have defined as “a formal delivery of documents that is legally sufficient to charge the defendant with notice of a pending action,” id., at 700)). But Article 10(a) does apply, Menon suggests, to the service of “post-answer judicial documents” (that is, any additional documents which may have to be served later in the litigation). Brief for Respondent 30–31. The problem with this argument is that it lacks any plausible textual footing in Article 10.[2] If the drafters wished to limit Article 10(a) to a particular subset of documents, they presumably would have said so—as they did, for example, in Article 15, which refers to “a writ of summons or an equivalent document.” Instead, Article 10(a) uses the term “judicial documents”—the same term that is featured in 10(b) and 10(c). Accord-ingly, the notion that Article 10(a) governs a different set of documents than 10(b) or 10(c) is hard to fathom. And it certainly derives no support from the use of the word “send,” whose ordinary meaning is broad enough to cover the transmission of any judicial documents (including litigation-initiating documents). Nothing about the word “send” suggests that Article 10(a) is narrower than 10(b) and 10(c), let alone that Article 10(a) is somehow limited to “post-answer” documents. Ultimately, Menon wishes to read the phrase “send judicial documents” as “serve a subset of judicial documents.” That is an entirely atextual reading, and Menon offers no sustained argument in support of it. Therefore, the only way to escape the conclusion that Article 10(a) includes service of process is to assert that it does not cover service of documents at all—and, as shown above, that reading is structurally implausible and renders Article 10(a) superfluous. B The text and structure of the Hague Service Convention, then, strongly suggest that Article 10(a) pertains to service of documents. The only significant counterargument is that, unlike many other provisions in the Convention, Article 10(a) does not include the word “service” or any of its variants. The Article 10(a) phrase “send judicial documents,” the argument goes, should mean something different than the phrase “effect service of judicial documents” in the other two subparts of Article 10. This argument does not win the day for several reasons. First, it must contend with the compelling structural considerations discussed above. See Air France v. Saks, 470 U. S. 392, 397 (1985) (treaty interpretation must take account of the “context in which the written words are used”); cf. University of Tex. Southwestern Medical Center v. Nassar, 570 U. S. ___, ___ (2013) (slip op., at 13) (“Just as Congress’ choice of words is presumed to be deliberate, so too are its structural choices”). Second, the argument fails on its own terms. Assume for a second that the word “send” must mean something other than “serve.” That would not imply that Article 10(a) must exclude service. Instead, “send[ing]” could be a broader concept that includes service but is not limited to it. That reading of the word “send” is probably more plausible than interpreting it to exclude service, and it does not create the same superfluity problem.[3] Third, it must be remembered that the French version of the Convention is “equally authentic” to the English version. Schlunk, 486 U. S., at 699. Menon does not seri-ously engage with the Convention’s French text. But the word “adresser”—the French counterpart to the word “send” in Article 10(a)—“has been consistently interpreted as meaning service or notice.” Hague Conference on Private Int’l Law, Practical Handbook on the Operation of the Service Convention ¶279, p. 91 (4th ed. 2016). In short, the most that could possibly be said for this argument is that it creates an ambiguity as to Article 10(a)’s meaning. And when a treaty provision is ambiguous, the Court “may look beyond the written words to the history of the treaty, the negotiations, and the practical construction adopted by the parties.” Schlunk, supra, at 700 (internal quotation marks omitted). As discussed below, these traditional tools of treaty interpretation comfortably resolve any lingering ambiguity in Water Splash’s favor. III Three extratextual sources are especially helpful in ascertaining Article 10(a)’s meaning: the Convention’s drafting history, the views of the Executive, and the views of other signatories. Drafting history has often been used in treaty interpretation. See Medellín v. Texas, 552 U. S. 491, 507 (2008) ; Saks, supra, at 400; see also Schlunk, supra, at 700 (analyzing the negotiating history of the Hague Service Convention). Here, the Convention’s drafting history strongly suggests that Article 10(a) allows service through postal channels. Philip W. Amram was the member of the United States delegation who was most closely involved in the drafting of the Convention. See S. Exec. Rep. No. 6, 90th Cong., 1st Sess. 5 (App.) (1967) (S. Exec. Rep.) (statement of State Department Deputy Legal Adviser Richard D. Kearney). A few months before the Convention was signed, he published an article describing and summarizing it. In that article, he stated that “Article 10 permits direct service by mail . . . unless [the receiving] state objects to such service.” The Proposed International Convention on the Service of Documents Abroad, 51 A. B. A. J. 650, 653 (1965).[4] Along similar lines, the Rapporteur’s report on a draft version of Article 10—which did not materially differ from the final version—stated that the “provision of paragraph 1 also permits service . . . by telegram” and that the drafters “did not accept the proposal that postal channels be limited to registered mail.” 1 Ristau §4–3–5(a), at 149. In other words, it was clearly understood that service by postal channels was permissible, and the only question was whether it should be limited to registered mail. The Court also gives “great weight” to “the Executive Branch’s interpretation of a treaty.” Abbott v. Abbott, 560 U. S. 1, 15 (2010) (internal quotation marks omitted). In the half century since the Convention was adopted, the Executive has consistently maintained that the Hague Service Convention allows service by mail. When President Johnson transmitted the Convention to the Senate for its advice and consent, he included a report by Secretary of State Dean Rusk. That report stated that “Article 10 permits direct service by mail . . . unless [the receiving] state objects to such service.” Convention on the Service Abroad of Judicial and Extrajudicial Documents in Civil or Commercial Matters: Message From the President of the United States, S. Exec. Doc. C, 90th Cong., 1st Sess., 5 (1967). In 1989, the Eighth Circuit issued Bankston, the first Federal Court of Appeals decision holding that the Hague Service Convention prohibits service by mail. 889 F. 2d, at 174. The State Department expressed its disagreement with Bankston in a letter addressed to the Administrative Office of the U. S. Courts and the National Center for State Courts. See Notice of Other Documents (1), United States Department of State Opinion Regarding the Bankston Case and Service by Mail to Japan Under the Hague Service Convention, 30 I. L. M. 260, 260–261 (1991) (excerpts of Mar. 14, 1990, letter). The letter stated that “Bankston is incorrect to the extent that it suggests that the Hague Convention does not permit as a method of service of process the sending of a copy of a summons and complaint by registered mail to a defendant in a foreign country.” Id., at 261. The State Department takes the same position on its website.[5] Finally, this Court has given “considerable weight” to the views of other parties to a treaty. Abbott, 560 U. S., at 16 (internal quotation marks omitted); see Lozano v. Montoya Alvarez, 572 U. S. ___, ___ (2014) (slip op., at 9) (noting the importance of “read[ing] the treaty in a manner consistent with the shared expectations of the contracting parties” (internal quotation marks omitted)). And other signatories to the Convention have consistently adopted Water Splash’s view. Multiple foreign courts have held that the Hague Service Convention allows for service by mail.[6] In addition, several of the Convention’s signatories have either objected, or declined to object, to service by mail under Article 10, thereby acknowledging that Article 10 encompasses service by mail.[7] Finally, several Special Commissions—comprising numerous contracting States—have expressly stated that the Convention does not prohibit service by mail.[8] By contrast, Menon identifies no evidence that any signatory has ever rejected Water Splash’s view. * * * In short, the traditional tools of treaty interpretation unmistakably demonstrate that Article 10(a) encompasses service by mail. To be clear, this does not mean that the Convention affirmatively authorizes service by mail. Article 10(a) simply provides that, as long as the receiving state does not object, the Convention does not “interfere with . . . the freedom” to serve documents through postal channels. In other words, in cases governed by the Hague Service Convention, service by mail is permissible if two conditions are met: first, the receiving state has not objected to service by mail; and second, service by mail is authorized under otherwise-applicable law. See Brockmeyer, 383 F. 3d, at 803–804. Because the Court of Appeals concluded that the Convention prohibited service by mail outright, it had no occasion to consider whether Texas law authorizes the methods of service used by Water Splash. We leave that question, and any other remaining issues, to be considered on remand to the extent they are properly preserved. For these reasons, we vacate the judgment of the Court of Appeals, and we remand the case for further proceedings not inconsistent with this opinion. It is so ordered. Justice Gorsuch took no part in the consideration or decision of this case.Notes 1 See Black’s Law Dictionary 1568 (10th ed. 2014) (defining “send,” in part, as “[t]o cause to be moved or conveyed from a present location to another place; esp., to deposit (a writing or notice) in the mail”). 2 The argument also assumes that the scope of the Convention is not limited to service of process (otherwise, Article 10(a) would be superfluous even under Menon’s reading). Schlunk can be read to suggest that this assumption is wrong. 486 U. S., at 700–701; see 1 B. Ristau, International Judicial Assistance §4–1–4(2), p. 112 (1990 rev. ed.) (Ristau) (stating that the English term “service” in the Convention “means the formal delivery of a legal document to the addressee in such a manner as to legally charge him with notice of the institution of a legal proceeding”). For the purposes of this discussion, we will assume, arguendo, that Menon’s assumption is correct. 3 Another plausible explanation for the distinct terminology of Article 10(a) is that it is the only provision in the Convention that specifically contemplates direct service, without the use of an intermediary. See Brief for United States as Amicus Curiae 13 (“[I]n contrast to Article 10(a), all other methods of service identified in the Convention require the affirmative engagement of an intermediary to effect ‘service’ ”). The use of the word “send” may simply have been intended to reflect that distinction. 4 Two years later, Amram testified to the same effect before the Senate Foreign Relations Committee. S. Exec. Rep., at 13 (stating that service by central authority “is not obligatory,” and that other available techniques included “direct service by mail”). 5 Dept. of State, Legal Considerations: International Judicial Assistance: Service of Process (stating that “[s]ervice by registered . . . mail . . . is an option in many countries in the world,” but that it “should . . . not be used in the countries party to the Hague Service Conven-tion that objected to the method described in Article 10(a) (postal channels)”), online at https://travel.state.gov/content/travel/en/legal-considerations/judicial/service-of-process.html (all Internet materials as last visited May 19, 2017). 6 See, e.g., Wang v. Lin, [2016] 132 O. R. 3d 48, 61 (Can. Ont. Sup. Ct. J.); Crystal Decisions (U. K.), Ltd. v. Vedatech Corp., EWHC (Ch) 1872 (2004), 2004 WL 1959749 ¶21 (High Court, Eng.); R. v. Re Recognition of an Italian Judgt., 2000 WL 33541696, ¶4 (D. F. Thes. 2000); Case C–412/97, ED Srl v. Italo Fenocchio, 1999 E. C. R. I–3845, 3877–3878, ¶6 [2000] 3 C. M. L. R. 855; see also Brockmeyer v. May, 383 F. 3d 798, 802 (CA9 2004) (noting that foreign courts are “essentially unanimous” in the view “that the meaning of ‘send’ in Article 10(a) includes ‘serve’ ”). 7 Canada, for example, has stated that it “does not object to service by postal channels.” By contrast, the Czech Republic has adopted Czechoslovakia’s position that “judicial documents may not be served . . . through postal channels.” Dutch Govt. Treaty Database: Convention on the Service Abroad of Judicial and Extrajudicial Documents in Civil or Commercial Matters: Parties With Reservations, Declarationsand Objections, (entries for Canada and the Czech Republic) on-line at https://treatydatabase.overheid.nl/en/Verdrag/Details/004235_b; see also, e.g., ibid. (entries for Latvia, Australia, and Slovenia). In addition, some states have objected to all of the channels of transmission listed in Article 10, referring to them collectively with the term “service.” See, e.g., ibid. (entries for Bulgaria, Hungary, Kuwait, and Turkey). 8 Hague Conference on Private International Law, Conclusions and Recommendations Adopted by the Special Commission on the Practical Operation of the Hague Apostille, Evidence and Service Conventions ¶55, p. 11 (Oct. 28–Nov. 4, 2003) (“reaffirm[ing]” the Special Commission’s “clear understanding that the term ‘send’ in Article 10(a) is to be understood as meaning ‘service’ through postal channels”), online at https://assets.hcch.net/upload/wop/lse_concl_e.pdf; Hague Conference on Private International Law, Report on the Work of the Special Commission of April 1989 on the Operation of the Hague Conventions of 15 November 1965 on the Service Abroad of Judicial and Extrajudicial Documents in Civil or Commercial Matters and of 18 March 1970 on the Taking of Evidence Abroad in Civil or Commercial Matters ¶16, p. 5 (Apr. 1989) (criticizing “certain courts in the United States” which “had concluded that service of process abroad by mail was not permitted under the Convention”), online at https://assets.hcch.net/upload/scrpt89e_20.pdf; Report on the Work of the Special Commission on the Operation of the Hague Convention on the Service Abroad of Judicial and Extrajudicial Documents in Civil or Commercial Matters, Nov. 21–25 1977, 17 I. L. M. 312, 326 (1978) (observing that “most of the States made no objection to the service of judicial documents coming from abroad directly by mail in their territory” (emphasis added)).

582.US.2016_16-240

When petitioner was tried in a Massachusetts trial court, the courtroom could not accommodate all the potential jurors. As a result, for two days of jury selection, an officer of the court excluded from the courtroom any member of the public who was not a potential juror, including petitioner’s mother and her minister. Defense counsel neither objected to the closure at trial nor raised the issue on direct review. Petitioner was convicted of murder and a related charge. Five years later, he filed a motion for a new trial in state court, arguing, as relevant here, that his attorney had provided ineffective assistance by failing to object to the courtroom closure. The trial court ruled that he was not entitled to relief. The Massachusetts Supreme Judicial Court affirmed in relevant part. Although it recognized that the violation of the right to public trial was a structural error, it rejected petitioner’s ineffective-assistance claim because he had not shown prejudice. Held: 1. In the context of a public-trial violation during jury selection, where the error is neither preserved nor raised on direct review but is raised later via an ineffective-assistance-of-counsel claim, the defendant must demonstrate prejudice to secure a new trial. Pp. 5–14. (a) This case requires an examination of the proper application of the doctrines of structural error and ineffective assistance of counsel. They are intertwined, because the reasons an error is deemed structural may influence the proper standard used to evaluate an ineffective-assistance claim premised on the failure to object to that error. Pp. 5–10. (1) Generally, a constitutional error that “did not contribute to the verdict obtained” is deemed harmless, which means the defendant is not entitled to reversal. Chapman v. California, 386 U. S. 18 . However, a structural error, which “affect[s] the framework within which the trial proceeds,” Arizona v. Fulminante, 499 U. S. 279 , defies harmless error analysis, id., at 309. Thus, when a structural error is objected to and then raised on direct review, the defendant is entitled to relief without any inquiry into harm. There appear to be at least three broad rationales for finding an error to be structural. One is when the right at issue does not protect the defendant from erroneous conviction but instead protects some other interest—like the defendant’s right to conduct his own defense—where harm is irrelevant to the basis underlying the right. See United States v. Gonzalez-Lopez, 548 U. S. 140 , n. 4. Another is when the error’s effects are simply too hard to measure—e.g., when a defendant is denied the right to select his or her own attorney—making it almost impossible for the government to show that the error was “harmless beyond a reasonable doubt,” Chapman, supra, at 24. Finally, some errors always result in fundamental unfairness, e.g., when an indigent defendant is denied an attorney, see Gideon v. Wainwright, 372 U. S. 335 –345. For purposes of this case, a critical point is that an error can count as structural even if it does not lead to fundamental unfairness in every case. See Gonzalez-Lopez, supra, at 149, n. 4. Pp. 5–7. (2) While a public-trial violation counts as structural error, it does not always lead to fundamental unfairness. This Court’s opinions teach that courtroom closure is to be avoided, but that there are some circumstances when it is justified. See Waller v. Georgia, 467 U. S. 39 ; Presley v. Georgia, 558 U. S. 209 –216. The fact that the public-trial right is subject to exceptions suggests that not every public-trial violation results in fundamental unfairness. Indeed, the Court has said that a public-trial violation is structural because of the “difficulty of assessing the effect of the error.” Gonzalez-Lopez, supra, at 149, n. 4. The public-trial right also furthers interests other than protecting the defendant against unjust conviction, including the rights of the press and of the public at large. See, e.g., Press-Enterprise Co. v. Superior Court of Cal., Riverside Cty., 464 U. S. 501 –510. Thus, an unlawful closure could take place and yet the trial will still be fundamentally fair from the defendant’s standpoint. Pp. 7–10. (b) The proper remedy for addressing the violation of the right to a public trial depends on when the objection was raised. If an objection is made at trial and the issue is raised on direct appeal, the defendant generally is entitled to “automatic reversal” regardless of the error’s actual “effect on the outcome.” Neder v. United States, 527 U. S. 1 . If, however, the defendant does not preserve a structural error on direct review but raises it later in the context of an ineffective-assistance claim, the defendant generally bears the burden to show deficient performance and that the attorney’s error “prejudiced the defense.” Strickland v. Washington, 466 U. S. 668 . To demonstrate prejudice in most cases, the defendant must show “a reasonable probability that . . . the result of the proceeding would have been different” but for attorney error. Id., at 694. For the analytical purposes of this case, the Court will assume, as petitioner has requested, that even if there is no showing of a reasonable probability of a different outcome, relief still must be granted if the defendant shows that attorney errors rendered the trial fundamentally unfair. Not every public-trial violation will lead to a fundamentally unfair trial. And the failure to object to that violation does not always deprive the defendant of a reasonable probability of a different outcome. Thus, a defendant raising a public-trial violation via an ineffective-assistance claim must show either a reasonable probability of a different outcome in his or her case or, as assumed here, that the particular violation was so serious as to render the trial fundamentally unfair. Neither this reasoning nor the holding here calls into question the Court’s precedents deeming certain errors structural and requiring reversal because of fundamental unfairness, see Sullivan v. Louisiana, 508 U. S., at 278–279; Tumey v. Ohio, 273 U. S. 510 ; Vasquez v. Hillery, 474 U. S., at 261–264, or those granting automatic relief to defendants who prevailed on claims of race or gender discrimination in jury selection, e.g., Batson v. Kentucky, 476 U. S. 79 . The errors in each of these cases were preserved and then raised on direct appeal. The reason for placing the burden on the petitioner here, however, derives both from the nature of the error and the difference between a public-trial violation preserved and then raised on direct review and a public-trial violation raised as an ineffective-assistance claim. When a defendant objects to a courtroom closure, the trial court can either order the courtroom opened or explain the reasons for keeping it closed, but when a defendant first raises the closure in an ineffective-assistance claim, the trial court has no chance to cure the violation. The costs and uncertainties of a new trial are also greater because more time will have elapsed in most cases. And the finality interest is more at risk. See Strickland, supra, at 693–694. These differences justify a different standard for evaluating a structural error depending on whether it is raised on direct review or in an ineffective-assistance claim. Pp. 10–14. 2. Because petitioner has not shown a reasonable probability of a different outcome but for counsel’s failure to object or that counsel’s shortcomings led to a fundamentally unfair trial, he is not entitled to a new trial. Although potential jurors might have behaved differently had petitioner’s family or the public been present, petitioner has offered no evidence suggesting a reasonable probability of a different outcome but for counsel’s failure to object. He has also failed to demonstrate fundamental unfairness. His mother and her minister were indeed excluded during jury selection. But his trial was not conducted in secret or in a remote place; closure was limited to the jury voir dire; the courtroom remained open during the evidentiary phase of the trial; the closure decision apparently was made by court officers, not the judge; venire members who did not become jurors observed the proceedings; and the record of the proceedings indicates no basis for concern, other than the closure itself. There was no showing, furthermore, that the potential harms flowing from a courtroom closure came to pass in this case, e.g., misbehavior by the prosecutor, judge, or any other party. Thus, even though this case comes here on the assumption that the closure was a Sixth Amendment violation, the violation here did not pervade the whole trial or lead to basic unfairness. Pp. 14–16. 474 Mass. 787, 54 N. E. 3d 495, affirmed. Kennedy, J., delivered the opinion of the Court, in which Roberts, C. J., and Thomas, Ginsburg, Sotomayor, and Gorsuch, JJ., joined. Thomas, J., filed a concurring opinion, in which Gorsuch, J., joined. Alito, J., filed an opinion concurring in the judgment, in which Gorsuch, J., joined. Breyer, J., filed a dissenting opinion, in which Kagan, J., joined.

During petitioner’s trial on state criminal charges, the courtroom was occupied by potential jurors and closed to the public for two days of the jury selection process. Defense counsel neither objected to the closure at trial nor raised the issue on direct review. And the case comes to the Court on the assumption that, in failing to object, defense counsel provided ineffective assistance. In the direct review context, the underlying constitutional violation—the courtroom closure—has been treated by this Court as a structural error, i.e., an error entitling the defendant to automatic reversal without any inquiry into prejudice. The question is whether invalidation of the conviction is required here as well, or if the prejudice inquiry is altered when the structural error is raised in the context of an ineffective-assistance-of-counsel claim. I In 2003, a 15-year-old boy was shot and killed in Boston. A witness saw a young man fleeing the scene of the crime and saw him pull out a pistol. A baseball hat fell off of his head. The police recovered the hat, which featured a distinctive airbrushed Detroit Tigers logo on either side. The hat’s distinctive markings linked it to 16-year-old Kentel Weaver. He is the petitioner here. DNA obtained from the hat matched petitioner’s DNA. Two weeks after the crime, the police went to petitioner’s house to question him. He admitted losing his hat around the time of the shooting but denied being involved. Petitioner’s mother was not so sure. Later, she questioned petitioner herself. She asked whether he had been at the scene of the shooting, and he said he had been there. But when she asked if he was the shooter, or if he knew who the shooter was, petitioner put his head down and said nothing. Believing his response to be an admission of guilt, she insisted that petitioner go to the police station to confess. He did. Petitioner was indicted in Massachusetts state court for first-degree murder and the unlicensed possession of a handgun. He pleaded not guilty and proceeded to trial. The pool of potential jury members was large, some 60 to 100 people. The assigned courtroom could accommodate only 50 or 60 in the courtroom seating. As a result, the trial judge brought all potential jurors into the courtroom so that he could introduce the case and ask certain preliminary questions of the entire venire panel. Many of the potential jurors did not have seats and had to stand in the courtroom. After the preliminary questions, the potential jurors who had been standing were moved outside the courtroom to wait during the individual questioning of the other potential jurors. The judge acknowledged that the hallway was not “the most comfortable place to wait” and thanked the potential jurors for their patience. 2 Tr. II–103 (Apr. 10, 2006). The judge noted that there was simply not space in the courtroom for everybody. As all of the seats in the courtroom were occupied by the venire panel, an officer of the court excluded from the courtroom any member of the public who was not a potential juror. So when petitioner’s mother and her minister came to the courtroom to observe the two days of jury selection, they were turned away. All this occurred before the Court’s decision in Presley v. Georgia, 558 U. S. 209 (2010) (per curiam). Presley made it clear that the public-trial right extends to jury selection as well as to other portions of the trial. Id., at 213–215. Before Presley, Massachusetts courts would often close courtrooms to the public during jury selection, in particular during murder trials. In this case petitioner’s mother told defense counsel about the closure at some point during jury selection. But counsel “believed that a courtroom closure for [ jury selection] was constitutional.” Crim. No. 2003–11293 (Super. Ct. Mass., Feb. 22, 2013), App. to Pet. for Cert. 49a. As a result, he “did not discuss the matter” with petitioner, or tell him “that his right to a public trial included the [ jury voir dire],” or object to the closure. Ibid. During the ensuing trial, the government presented strong evidence of petitioner’s guilt. Its case consisted of the incriminating details outlined above, including petitioner’s confession to the police. The jury convicted petitioner on both counts. The court sentenced him to life in prison on the murder charge and to about a year in prison on the gun-possession charge. Five years later, petitioner filed a motion for a new trial in Massachusetts state court. As relevant here, he argued that his attorney had provided ineffective assistance by failing to object to the courtroom closure. After an evidentiary hearing, the trial court recognized a violation of the right to a public trial based on the following findings: The courtroom had been closed; the closure was neither de minimis nor trivial; the closure was unjustified; and the closure was full rather than partial (meaning that all members of the public, rather than only some of them, had been excluded from the courtroom). The trial court further determined that defense counsel failed to object because of “serious incompetency, inefficiency, or inattention.” Id., at 63a (quoting Massachusetts v. Chleikh, 82 Mass. App. 718, 722, 978 N. E. 2d 96, 100 (2012)). On the other hand, petitioner had not “offered any evidence or legal argument establishing prejudice.” App. to Pet. for Cert. 64a. For that reason, the court held that petitioner was not entitled to relief. Petitioner appealed the denial of the motion for a new trial to the Massachusetts Supreme Judicial Court. The court consolidated that appeal with petitioner’s direct appeal. As noted, there had been no objection to the closure at trial; and the issue was not raised in the direct appeal. The Supreme Judicial Court then affirmed in relevant part. Although it recognized that “[a] violation of the Sixth Amendment right to a public trial constitutes structural error,” the court stated that petitioner had “failed to show that trial counsel’s conduct caused prejudice warranting a new trial.” 474 Mass. 787, 814, 54 N. E. 3d 495, 520 (2016). On this reasoning, the court rejected petitioner’s claim of ineffective assistance of counsel. There is disagreement among the Federal Courts of Appeals and some state courts of last resort about whether a defendant must demonstrate prejudice in a case like this one—in which a structural error is neither preserved nor raised on direct review but is raised later via a claim alleging ineffective assistance of counsel. Some courts have held that, when a defendant shows that his attorney unreasonably failed to object to a structural error, the defendant is entitled to a new trial without further inquiry. See, e.g., Johnson v. Sherry, 586 F. 3d 439, 447 (CA6 2009); Owens v. United States, 483 F. 3d 48, 64–65 (CA1 2007); Littlejohn v. United States, 73 A. 3d 1034, 1043–1044 (D. C. 2013); State v. Lamere, 327 Mont. 115, 125, 112 P. 3d 1005, 1013 (2005). Other courts have held that the defendant is entitled to relief only if he or she can show prejudice. See, e.g., Purvis v. Crosby, 451 F. 3d 734, 738 (CA11 2006); United States v. Gomez, 705 F. 3d 68, 79–80 (CA2 2013); Reid v. State, 286 Ga. 484, 487, 690 S. E. 2d 177, 180–181 (2010). This Court granted certio-rari to resolve that disagreement. 580 U. S. ___ (2017). The Court does so specifically and only in the context of trial counsel’s failure to object to the closure of the courtroom during jury selection. II This case requires a discussion, and the proper application, of two doctrines: structural error and ineffective assistance of counsel. The two doctrines are intertwined; for the reasons an error is deemed structural may influence the proper standard used to evaluate an ineffective-assistance claim premised on the failure to object to that error. A The concept of structural error can be discussed first. In Chapman v. California, 386 U. S. 18 (1967) , this Court “adopted the general rule that a constitutional error does not automatically require reversal of a conviction.” Ari-zona v. Fulminante, 499 U. S. 279, 306 (1991) (citing Chap-man, supra). If the government can show “beyond a reasonable doubt that the error complained of did not contribute to the verdict obtained,” the Court held, then the error is deemed harmless and the defendant is not entitled to reversal. Id., at 24. The Court recognized, however, that some errors should not be deemed harmless beyond a reasonable doubt. Id., at 23, n. 8. These errors came to be known as structural errors. See Fulminante, 499 U. S., at 309–310. The purpose of the structural error doctrine is to ensure insistence on certain basic, constitutional guarantees that should define the framework of any criminal trial. Thus, the defining feature of a structural error is that it “affect[s] the framework within which the trial proceeds,” rather than being “simply an error in the trial process itself.” Id., at 310. For the same reason, a structural error “def[ies] analysis by harmless error standards.” Id., at 309 (internal quotation marks omitted). The precise reason why a particular error is not amen-able to that kind of analysis—and thus the precise reason why the Court has deemed it structural—varies in a significant way from error to error. There appear to be at least three broad rationales. First, an error has been deemed structural in some instances if the right at issue is not designed to protect the defendant from erroneous conviction but instead protects some other interest. This is true of the defendant’s right to conduct his own defense, which, when exercised, “usu-ally increases the likelihood of a trial outcome unfavorable to the defendant.” McKaskle v. Wiggins, 465 U. S. 168 , n. 8 (1984). That right is based on the fundamental legal principle that a defendant must be allowed to make his own choices about the proper way to protect his own liberty. See Faretta v. California, 422 U. S. 806, 834 (1975) . Because harm is irrelevant to the basis underlying the right, the Court has deemed a violation of that right structural error. See United States v. Gonzalez-Lopez, 548 U. S. 140 , n. 4 (2006). Second, an error has been deemed structural if the effects of the error are simply too hard to measure. For example, when a defendant is denied the right to select his or her own attorney, the precise “effect of the violation cannot be ascertained.” Ibid. (quoting Vasquez v. Hillery, 474 U. S. 254, 263 (1986) ). Because the government will, as a result, find it almost impossible to show that the error was “harmless beyond a reasonable doubt,” Chapman, supra, at 24, the efficiency costs of letting the government try to make the showing are unjustified. Third, an error has been deemed structural if the error always results in fundamental unfairness. For example, if an indigent defendant is denied an attorney or if the judge fails to give a reasonable-doubt instruction, the resulting trial is always a fundamentally unfair one. See Gideon v. Wainwright, 372 U. S. 335 –345 (1963) (right to an attorney); Sullivan v. Louisiana, 508 U. S. 275 (1993) (right to a reasonable-doubt instruction). It therefore would be futile for the government to try to show harmlessness. These categories are not rigid. In a particular case, more than one of these rationales may be part of the explanation for why an error is deemed to be structural. See e.g., id., at 280–282. For these purposes, however, one point is critical: An error can count as structural even if the error does not lead to fundamental unfairness in every case. See Gonzalez-Lopez, supra, at 149, n. 4 (rejecting as “inconsistent with the reasoning of our precedents” the idea that structural errors “always or necessarily render a trial fundamentally unfair and unreliable” (emphasis deleted)). B As noted above, a violation of the right to a public trial is a structural error. See supra, at 1, 4. It is relevant to determine why that is so. In particular, the question is whether a public-trial violation counts as structural because it always leads to fundamental unfairness or for some other reason. In Waller v. Georgia, 467 U. S. 39 (1984) , the state court prohibited the public from viewing a weeklong suppression hearing out of concern for the privacy of persons other than those on trial. See id., at 41–43. Although it recognized that there would be instances where closure was justified, this Court noted that “such circumstances will be rare” and that the closure in question was unjustified. Id., at 45, 48. Still, the Court did not order a new trial. Id., at 49–50. Instead it ordered a new suppression hearing that was open to the public. Id., at 50. If the same evidence was found admissible in that renewed pretrial proceeding, the Court held, no new trial as to guilt would be neces-sary. Ibid. This was despite the structural aspect of the violation. Some 25 years after the Waller decision, the Court issued its per curiam ruling in Presley v. Georgia. 558 U. S. 209 . In that case, as here, the courtroom was closed to the public during jury voir dire. Id., at 210. Unlike here, however, there was a trial objection to the closure, and the issue was raised on direct appeal. Id., at 210–211. On review of the State Supreme Court’s decision allowing the closure, this Court expressed concern that the state court’s reasoning would allow the courtroom to be closed during jury selection “whenever the trial judge decides, for whatever reason, that he or she would prefer to fill the courtroom with potential jurors rather than spectators.” Id., at 215 (internal quotation marks omitted). Although the Court expressly noted that courtroom closure may be ordered in some circumstances, the Court also stated that it was “still incumbent upon” the trial court “to consider all reasonable alternatives to closure.” Id., at 215–216. These opinions teach that courtroom closure is to be avoided, but that there are some circumstances when it is justified. The problems that may be encountered by trial courts in deciding whether some closures are necessary, or even in deciding which members of the public should be admitted when seats are scarce, are difficult ones. For example, there are often preliminary instructions that a judge may want to give to the venire as a whole, rather than repeating those instructions (perhaps with unintentional differences) to several groups of potential jurors. On the other hand, various constituencies of the public—the family of the accused, the family of the victim, members of the press, and other persons—all have their own interests in observing the selection of jurors. How best to manage these problems is not a topic discussed at length in any decision or commentary the Court has found. So although the public-trial right is structural, it is subject to exceptions. See Simonson, The Criminal Court Audience in a Post-Trial World, 127 Harv. L. Rev. 2173, 2219–2222 (2014) (discussing situations in which a trial court may order a courtroom closure). Though these cases should be rare, a judge may deprive a defendant of his right to an open courtroom by making proper factual findings in support of the decision to do so. See Waller, supra, at 45. The fact that the public-trial right is subject to these exceptions suggests that not every public-trial violation results in fundamental unfairness. A public-trial violation can occur, moreover, as it did in Presley, simply because the trial court omits to make the proper findings before closing the courtroom, even if those findings might have been fully supported by the evidence. See 558 U. S., at 215. It would be unconvincing to deem a trial fundamentally unfair just because a judge omitted to announce factual findings before making an otherwise valid decision to order the courtroom temporarily closed. As a result, it would be likewise unconvincing if the Court had said that a public-trial violation always leads to a fundamentally unfair trial. Indeed, the Court has not said that a public-trial violation renders a trial fundamentally unfair in every case. In the two cases in which the Court has discussed the reasons for classifying a public-trial violation as structural error, the Court has said that a public-trial violation is structural for a different reason: because of the “difficulty of assessing the effect of the error.” Gonzalez-Lopez, 548 U. S., at 149, n. 4; see also Waller, supra, at 49, n. 9. The public-trial right also protects some interests that do not belong to the defendant. After all, the right to an open courtroom protects the rights of the public at large, and the press, as well as the rights of the accused. See, e.g., Press-Enterprise Co. v. Superior Court of Cal., Riverside Cty., 464 U. S. 501 –510 (1984); Richmond Newspapers, Inc. v. Virginia, 448 U. S. 555 –573 (1980). So one other factor leading to the classification of structural error is that the public-trial right furthers interests other than protecting the defendant against unjust conviction. These precepts confirm the conclusion the Court now reaches that, while the public-trial right is important for fundamental reasons, in some cases an unlawful closure might take place and yet the trial still will be fundamentally fair from the defendant’s standpoint. III The Court now turns to the proper remedy for addressing the violation of a structural right, and in particular the right to a public trial. Despite its name, the term “structural error” carries with it no talismanic significance as a doctrinal matter. It means only that the government is not entitled to deprive the defendant of a new trial by showing that the error was “harmless beyond a reasonable doubt.” Chapman, 386 U. S., at 24. Thus, in the case of a structural error where there is an objection at trial and the issue is raised on direct appeal, the defendant gener-ally is entitled to “automatic reversal” regardless of the error’s actual “effect on the outcome.” Neder v. United States, 527 U. S. 1, 7 (1999) . The question then becomes what showing is necessary when the defendant does not preserve a structural error on direct review but raises it later in the context of an ineffective-assistance-of-counsel claim. To obtain relief on the basis of ineffective assistance of counsel, the defendant as a general rule bears the burden to meet two standards. First, the defendant must show deficient performance—that the attorney’s error was “so serious that counsel was not functioning as the ‘counsel’ guaranteed the defendant by the Sixth Amendment.” Strickland v. Washington, 466 U. S. 668, 687 (1984) . Second, the defendant must show that the attorney’s error “prejudiced the defense.” Ibid. The prejudice showing is in most cases a necessary part of a Strickland claim. The reason is that a defendant has a right to effective representation, not a right to an attorney who performs his duties “mistake-free.” Gonzalez-Lopez, 548 U. S., at 147. As a rule, therefore, a “violation of the Sixth Amendment right to effective representation is not ‘complete’ until the defendant is prejudiced.” Ibid. (emphasis deleted); see also Premo v. Moore, 562 U. S. 115, 128 (2011) ; Lockhart v. Fretwell, 506 U. S. 364, 370 (1993) . That said, the concept of prejudice is defined in different ways depending on the context in which it appears. In the ordinary Strickland case, prejudice means “a reasonable probability that, but for counsel’s unprofessional errors, the result of the proceeding would have been different.” 466 U. S., at 694. But the Strickland Court cautioned that the prejudice inquiry is not meant to be applied in a “mechanical” fashion. Id., at 696. For when a court is evaluating an ineffective-assistance claim, the ultimate inquiry must concentrate on “the fundamental fairness of the proceeding.” Ibid. Petitioner therefore argues that under a proper interpretation of Strickland, even if there is no showing of a reasonable probability of a different outcome, relief still must be granted if the convicted person shows that attorney errors rendered the trial fundamentally unfair. For the analytical purposes of this case, the Court will assume that petitioner’s interpretation of Strickland is the correct one. In light of the Court’s ultimate holding, however, the Court need not decide that question here. As explained above, not every public-trial violation will in fact lead to a fundamentally unfair trial. See supra, at 10. Nor can it be said that the failure to object to a public-trial violation always deprives the defendant of a reason-able probability of a different outcome. Thus, when a de-fendant raises a public-trial violation via an ineffective-assistance-of-counsel claim, Strickland prejudice is not shown automatically. Instead, the burden is on the defendant to show either a reasonable probability of a different outcome in his or her case or, as the Court has assumed for these purposes, see supra, at 11, to show that the particular public-trial violation was so serious as to render his or her trial fundamentally unfair. Neither the reasoning nor the holding here calls into question the Court’s precedents determining that certain errors are deemed structural and require reversal because they cause fundamental unfairness, either to the defendant in the specific case or by pervasive undermining of the systemic requirements of a fair and open judicial process. See Murray, A Contextual Approach to Harmless Error Review, 130 Harv. L. Rev. 1791, 1813, 1822 (2017) (noting that the “eclectic normative objectives of criminal procedure” go beyond protecting a defendant from erroneous conviction and include ensuring “ ‘that the administration of justice should reasonably appear to be disinterested’ ” (quoting Liljeberg v. Health Services Acquisition Corp., 486 U. S. 847 –870 (1988))). Those precedents include Sullivan v. Louisiana, 508 U. S., at 278–279 (failure to give a reasonable-doubt instruction); Tumey v. Ohio, 273 U. S. 510, 535 (1927) (biased judge); and Vasquez v. Hillery, 474 U. S., at 261–264 (exclusion of grand jurors on the basis of race). See Neder, supra, at 8 (describing each of these errors as structural). This Court, in addition, has granted automatic relief to defendants who prevailed on claims alleging race or gender discrimination in the selection of the petit jury, see Batson v. Kentucky, 476 U. S. 79, 100 (1986) ; J. E. B. v. Alabama ex rel. T. B., 511 U. S. 127 –146 (1994), though the Court has yet to label those errors structural in express terms, see, e.g., Neder, supra, at 8. The errors in those cases necessitated automatic reversal after they were preserved and then raised on direct appeal. And this opinion does not address whether the result should be any different if the errors were raised instead in an ineffective-assistance claim on collateral review. The reason for placing the burden on the petitioner in this case, however, derives both from the nature of the error, see supra, at 11–12, and the difference between a public-trial violation preserved and then raised on direct review and a public-trial violation raised as an ineffective-assistance-of-counsel claim. As explained above, when a defendant objects to a courtroom closure, the trial court can either order the courtroom opened or explain the reasons for keeping it closed. See supra, at 8–9. When a defendant first raises the closure in an ineffective-assistance claim, however, the trial court is deprived of the chance to cure the violation either by opening the courtroom or by explaining the reasons for closure. Furthermore, when state or federal courts adjudicate errors objected to during trial and then raised on direct review, the systemic costs of remedying the error are diminished to some extent. That is because, if a new trial is ordered on direct review, there may be a reasonable chance that not too much time will have elapsed for witness memories still to be accurate and physical evidence not to be lost. There are also advantages of direct judicial supervision. Reviewing courts, in the regular course of the appellate process, can give instruction to the trial courts in a familiar context that allows for elaboration of the relevant principles based on review of an adequate record. For instance, in this case, the factors and circumstances that might justify a temporary closure are best considered in the regular appellate process and not in the context of a later proceeding, with its added time delays. When an ineffective-assistance-of-counsel claim is raised in postconviction proceedings, the costs and uncertainties of a new trial are greater because more time will have elapsed in most cases. The finality interest is more at risk, see Strickland, 466 U. S., at 693–694 (noting the “profound importance of finality in criminal proceedings”), and direct review often has given at least one opportunity for an appellate review of trial proceedings. These differences justify a different standard for evaluating a structural error depending on whether it is raised on direct review or raised instead in a claim alleging ineffective assistance of counsel. In sum, “[a]n ineffective-assistance claim can function as a way to escape rules of waiver and forfeiture and raise issues not presented at trial,” thus undermining the final-ity of jury verdicts. Harrington v. Richter, 562 U. S. 86, 105 (2011) . For this reason, the rules governing ineffective-assistance claims “must be applied with scrupulous care.” Premo, 562 U. S., at 122. IV The final inquiry concerns the ineffective-assistance claim in this case. Although the case comes on the assumption that petitioner has shown deficient performance by counsel, he has not shown prejudice in the ordinary sense, i.e., a reasonable probability that the jury would not have convicted him if his attorney had objected to the closure. It is of course possible that potential jurors might have behaved differently if petitioner’s family had been present. And it is true that the presence of the public might have had some bearing on juror reaction. But here petitioner offered no “evidence or legal argument establishing prejudice” in the sense of a reasonable probability of a different outcome but for counsel’s failure to object. App. to Pet. for Cert. 64a; see Strickland, 466 U. S., at 694. In other circumstances a different result might obtain. If, for instance, defense counsel errs in failing to object when the government’s main witness testifies in secret, then the defendant might be able to show prejudice with little more detail. See ibid. Even in those circumstances, however, the burden would remain on the defendant to make the prejudice showing, id., at 694, 696, because a public-trial violation does not always lead to a fundamentally unfair trial, see supra, at 10. In light of the above assumption that prejudice can be shown by a demonstration of fundamental unfairness, see supra, at 11, the remaining question is whether petitioner has shown that counsel’s failure to object rendered the trial fundamentally unfair. See Strickland, supra, at 696. The Court concludes that petitioner has not made the showing. Although petitioner’s mother and her minister were indeed excluded from the courtroom for two days during jury selection, petitioner’s trial was not conducted in secret or in a remote place. Cf. In re Oliver, 333 U. S. 257 , n. 22 (1948). The closure was limited to the jury voir dire; the courtroom remained open during the evidentiary phase of the trial; the closure decision apparently was made by court officers rather than the judge; there were many members of the venire who did not become jurors but who did observe the proceedings; and there was a record made of the proceedings that does not indicate any basis for concern, other than the closure itself. There has been no showing, furthermore, that the potential harms flowing from a courtroom closure came to pass in this case. For example, there is no suggestion that any juror lied during voir dire; no suggestion of misbehavior by the prosecutor, judge, or any other party; and no suggestion that any of the participants failed to approach their duties with the neutrality and serious purpose that our system demands. It is true that this case comes here on the assumption that the closure was a Sixth Amendment violation. And it must be recognized that open trials ensure respect for the justice system and allow the press and the public to judge the proceedings that occur in our Nation’s courts. Even so, the violation here did not pervade the whole trial or lead to basic unfairness. In sum, petitioner has not shown a reasonable probability of a different outcome but for counsel’s failure to object, and he has not shown that counsel’s shortcomings led to a fundamentally unfair trial. He is not entitled to a new trial. * * * In the criminal justice system, the constant, indeed unending, duty of the judiciary is to seek and to find the proper balance between the necessity for fair and just trials and the importance of finality of judgments. When a structural error is preserved and raised on direct review, the balance is in the defendant’s favor, and a new trial generally will be granted as a matter of right. When a structural error is raised in the context of an ineffective-assistance claim, however, finality concerns are far more pronounced. For this reason, and in light of the other circumstances present in this case, petitioner must show prejudice in order to obtain a new trial. As explained above, he has not made the required showing. The judgment of the Massachusetts Supreme Judicial Court is affirmed. It is so ordered.

585.US.2017_17-586

In 2011, the Texas Legislature adopted a new congressional districting plan and new districting maps for the two houses of the State Legislature to account for population growth revealed in the 2010 census. To do so, Texas had to comply with a complicated legal regime. The Equal Protection Clause of the Fourteenth Amendment forbids “racial gerrymandering,” that is, intentionally assigning citizens to a district on the basis of race without sufficient justification. Shaw v. Reno, 509 U. S. 630, 641. But other legal requirements tend to require that state legislatures consider race in drawing districts. Like all States, Texas is subject to §2 of the Voting Rights Act of 1965 (VRA), which is violated when a state districting plan provides “less opportunity” for racial minorities “to elect representatives of their choice,” League of United Latin American Citizens v. Perry, 548 U. S. 399, 425. And at the time, Texas was also subject to §5, which barred it from making any districting changes unless it could prove that they did not result in retrogression with respect to the ability of racial minorities to elect the candidates of their choice, Alabama Legislative Black Caucus v. Alabama, 575 U. S. ___, ___. In an effort to harmonize these conflicting demands, the Court has assumed that compliance with the VRA is a compelling State interest for Fourteenth Amendment purposes, see, e.g., Bethune-Hill v. Virginia State Bd. of Elections, 580 U. S. ___, ___; and a State’s consideration of race in making a districting decision is narrowly tailored if the State has “good reasons” for believing that its decision is necessary in order to comply with the VRA, Cooper v. Harris, 581 U. S. ___, ___. The Texas Legislature’s 2011 plans were immediately tied up in litigation and never used. The case was assigned to a three-judge court (Texas court). Texas also submitted the plans for preclearance to the District Court for the District of Columbia (D. C. court). The Texas court drew up interim plans for the State’s rapidly approaching primaries, giving no deference to the Legislature’s plans. Texas challenged the court-ordered plans in this Court, which reversed and remanded with instructions for the Texas court to start with the Texas Legislature’s 2011 plans but to make adjustments as required by the Constitution and the VRA. The Texas court then adopted new interim plans. After the D. C. court denied preclearance of the 2011 plans, Texas used the Texas court’s interim plans for the 2012 elections. In 2013, the Legislature repealed the 2011 plans and enacted the Texas court’s plans (with minor modifications). After Shelby County v. Holder, 570 U. S. 529, was decided, Texas, no longer covered by §5, obtained a vacatur of the D. C. court’s preclearance order. But the Texas court did not dismiss the case against the 2011 plans as moot. Instead, it allowed the plaintiffs to amend their complaint to challenge the 2013 plans and held that their challenges to the 2011 plans were live. Texas conducted its 2014 and 2016 elections under the 2013 plans. In 2017, the Texas court found defects in several of the districts in the 2011 federal congressional and State House plans (the State Senate plan is not at issue here). Subsequently, it also invalidated multiple Congressional (CD) and House (HD) Districts in the 2013 plans, holding that the Legislature failed to cure the “taint” of discriminatory intent allegedly harbored by the 2011 Legislature. And the court relied on that finding to invalidate several challenged 2013 districts. The court also held that three districts—CD27, HD32, and HD34—were invalid under §2 of the VRA because they had the effect of depriving Latinos of the equal opportunity to elect their candidates of choice. And it found that HD90 was a racial gerrymander based on changes made by the 2013 Legislature. It gave the state attorney general three days to tell the court whether the Legislature would remedy the violations; and if the Legislature did not intend to adopt new plans, the court would hold remedial hearings. Held: 1. This Court has jurisdiction to review the orders at issue. Pp. 11–21. (a) The Texas court’s orders fall within 28 U. S. C. §1253, which gives the Court jurisdiction to hear an appeal from an order of a three-judge district court “granting or denying . . . an interlocutory or permanent injunction.” The Texas court did not call its orders “injunctions,” but where an order has the “practical effect” of granting or denying an injunction, it should be treated as such for purposes of appellate jurisdiction. Carson v. American Brands, Inc., 450 U. S. 79, 83. Pp. 11–16. (b) The text of the orders and the context in which they were issued make clear that they qualify as interlocutory injunctions under §1253. The orders were unequivocal that the current legislative plans “violate §2 and the Fourteenth Amendment” and that these violations “must be remedied.” And the short time frame the attorney general was given to act is strong evidence that the court did not intend to allow the elections to go ahead under the plans it had just condemned. The unmistakable import of these actions is that the court intended to have new plans ready for use in this year’s elections. Texas also had reason to fear that if it tried to conduct elections under those plans, the court would infer an evil motive and perhaps subject the State to the strictures of preclearance under §3(c) of the VRA. These cases differ from Gunn v. University Comm. to End War in Viet Nam, 399 U. S. 383, where the order did not have the same practical effect as an injunction. Nor does it matter that the remedy is not yet known. The issue here is whether this year’s elections can be held under the plans enacted by the Legislature, not whether any particular remedies should ultimately be ordered if it is determined that the current plans are flawed. Section 1253 must be strictly, but sensibly construed, and here the District Court’s orders, for all intents and purposes, constituted injunctions. Thus, §1253 provides jurisdiction. Pp. 16–21. 2. The Texas court erred in requiring the State to show that the 2013 Legislature purged the “taint” that the court attributed to the defunct and never-used plans enacted by a prior legislature in 2011. Pp. 21–32. (a) Whenever a challenger claims that a state law was enacted with discriminatory intent, the burden of proof lies with the challenger, not the State. Reno v. Bossier Parish School Bd., 520 U. S. 471, 481. In redistricting cases, the “good faith of [the] state legislature must be presumed.” Miller v. Johnson, 515 U. S. 900, 915. The allocation of the burden of proof and the presumption of legislative good faith are not changed by a finding of past discrimination, which is but “one evidentiary source” relevant to the question of intent. Arlington Heights v. Metropolitan Housing Development Corp., 429 U. S. 252, 267. Here, the 2011 plans were repealed, and not reenacted, by the 2013 Legislature. Nor did it use criteria that arguably carried forward the effects of the 2011 Legislature’s discriminatory intent. Instead, it enacted, with only small changes, the Texas court plans developed pursuant to this Court’s instructions. The Texas court contravened these basic burden of proof principles, referring, e.g., to the need to “cure” the earlier Legislature’s “taint” and concluding that the Legislature had engaged in no deliberative process to do so. This fundamentally flawed approach must be reversed. Pp. 21–25. (b) Both the 2011 Legislature’s intent and the court’s interim plans are relevant to the extent that they give rise to—or tend to refute—inferences about the 2013 Legislature’s intent, but they must be weighed together with other relevant direct and circumstantial evidence of the Legislature’s intent. But when this evidence is taken into account, the evidence in the record is plainly insufficient to prove that the 2013 Legislature acted in bad faith and engaged in intentional discrimination. Pp. 25–32. 3. Once the Texas court’s intent finding is reversed, there remain only four districts that were invalidated on alternative grounds. The Texas court’s holding as to the three districts in which it relied on §2’s “effects” test are reversed, but its holding that HD90 is a racial gerrymander is affirmed. Pp. 32–41. (a) To make out a §2 “effects” claim, a plaintiff must establish the three “Gingles factors”: (1) a geographically compact minority population sufficient to constitute a majority in a single-member district, (2) political cohesion among the members of the minority group, and (3) bloc voting by the majority to defeat the minority’s preferred candidate. Thornburg v. Gingles, 478 U. S. 30, 48–51. A plaintiff who makes that showing must then prove that, under the totality of the circumstances, the district lines dilute the votes of the members of the minority group. Pp. 33–39. (1) The Texas court held that CD27 violates §2 because it has the effect of diluting the votes of Nueces County Latino voters, who, the court concluded, should have been included in a Latino opportunity district rather than CD27, which is not such a district. Plaintiffs, however, could not show that an additional Latino opportunity district could be created in that part of Texas. Pp. 33–35. (2) The Texas court similarly erred in holding that HD32 and HD34, which make up the entirety of Nueces County, violate §2. The 2013 plan created two districts that lie wholly within the county: HD34 is a Latino opportunity district, but HD32 is not. The court’s findings show that these two districts do not violate §2, and it is hard to see how the ultimate Gingles vote dilution standard could be met if the alternative plan would not enhance the ability of minority voters to elect the candidates of their choice. Pp. 35–38. (b) HD90 is an impermissible racial gerrymander. HD90 was not copied from the Texas court’s interim plans. Instead, the 2013 legislature substantially modified that district. In 2011, the Legislature, responding to pressure from counsel to one of the plaintiff groups, increased the district’s Latino population in an effort to make it a Latino opportunity district. It also moved the city of Como, which is predominantly African-American, out of the district. When Como residents and their Texas House representative objected, the Legislature moved Como back. But that decreased the Latino population, so the Legislature moved more Latinos into the district. Texas argues that its use of race as the predominant factor in HD90’s design was permissible because it had “good reasons to believe” that this was necessary to satisfy §2, Bethune-Hill, 580 U. S., at ___. But it is the State’s burden to prove narrow tailoring, and Texas did not do so on the record here. Pp. 38–41. No. 17–586, 274 F. Supp. 3d 624, reversed; No. 17–626, 267 F. Supp. 3d 750, reversed in part and affirmed in part; and cases remanded. Alito, J., delivered the opinion of the Court, in which Roberts, C. J., and Kennedy, Thomas, and Gorsuch, JJ., joined. Thomas, J., filed a concurring opinion, in which Gorsuch, J., joined. Sotomayor, J., filed a dissenting opinion, in which Ginsburg, Breyer, and Kagan, JJ., joined. Notes 1 Together with No. 17–626, Abbott, Governor of Texas, et al. v. Perez et al., also on appeal from the same court.

Before us for review are orders of a three-judge court in the Western District of Texas effectively directing the State not to conduct this year’s elections using districting plans that the court itself adopted some years earlier. The court developed those plans for use in the 2012 elections pursuant to our directions in Perry v. Perez, 565 U. S. 388 (2012) (per curiam). We instructed the three-judge court to start with the plans adopted by the Texas Legislature in 2011 but to make adjustments as required by the Constitution and the Voting Rights Act. Id., at 392–396. After those plans were used in 2012, the Texas Legislature enacted them (with only minor modifications) in 2013, and the plans were used again in both 2014 and 2016. Last year, however, the three-judge court reversed its prior analysis and held that some of the districts in those plans are unlawful. After reviewing the repealed 2011 plans, which had never been used, the court found that they were tainted by discriminatory intent and that the 2013 Legislature had not “cured” that “taint.” We now hold that the three-judge court committed a fundamental legal error. It was the challengers’ burden to show that the 2013 Legislature acted with discriminatory intent when it enacted plans that the court itself had produced. The 2013 Legislature was not obligated to show that it had “cured” the unlawful intent that the court attributed to the 2011 Legislature. Thus, the essential pillar of the three-judge court’s reasoning was critically flawed. When the congressional and state legislative districts are reviewed under the proper legal standards, all but one of them, we conclude, are lawful. I A The 2010 decennial census revealed that the population of Texas had grown by more than 20% and the State was therefore apportioned four additional seats in the United States House of Representatives. C. J. S. 369a.[1] To accommodate this new allocation and the population changes shown by the census, the Legislature adopted a new congressional districting plan, as well as new districting maps for the two houses of the State Legislature. Redistricting is never easy, and the task was especially complicated in Texas in 2011. Not only was the Legislature required to draw districts that were substantially equal in population, see Perry, supra, at 391–392; Rey- nolds v. Sims, 377 U. S. 533 (1964); Wesberry v. Sanders, 376 U. S. 1 (1964), and to comply with special state-law districting rules,[2] but federal law imposed complex and delicately balanced requirements regarding the consideration of race. Then, as now, federal law restricted the use of race in making districting decisions. The Equal Protection Clause forbids “racial gerrymandering,” that is, intentionally assigning citizens to a district on the basis of race without sufficient justification. Shaw v. Reno, 509 U. S. 630, 641 (1993). It also prohibits intentional “vote dilution”—“invidiously . . . minimiz[ing] or cancel[ing] out the voting potential of racial or ethnic minorities.” Mobile v. Bolden, 446 U. S. 55, 66–67 (1980) (plurality opinion). While the Equal Protection Clause imposes these important restrictions, its application in the field of districting is complicated. For one thing, because a voter’s race sometimes correlates closely with political party preference, see Cooper v. Harris, 581 U. S. ___, ___–___ (2017) (slip op., at 19–20); Easley v. Cromartie, 532 U. S. 234, 243 (2001), it may be very difficult for a court to determine whether a districting decision was based on race or party preference. Here, the three-judge court found that the two factors were virtually indistinguishable.[3] At the same time that the Equal Protection Clause restricts the consideration of race in the districting process, compliance with the Voting Rights Act of 1965, 79Stat. 437, as amended, 52 U. S. C. §10301 et seq. (VRA), pulls in the opposite direction: It often insists that districts be created precisely because of race. Two provisions of the VRA exert such demands, and in 2011, Texas was subject to both. At that time, Texas was covered by §5 of the VRA[4] and was thus barred from making any districting changes unless it could prove that they did not result in “retrogression” with respect to the ability of racial minorities to elect the candidates of their choice. Alabama Legislative Black Caucus v. Alabama, 575 U. S. ___, ___ (2015) (slip op., at 3). That showing obviously demanded consideration of race. On top of this, Texas was (and still is) required to comply with §2 of the VRA. A State violates §2 if its districting plan provides “ ‘less opportunity’ ” for racial minorities “ ‘to elect representatives of their choice.’ ” League of United Latin American Citizens v. Perry, 548 U. S. 399, 425 (2006) (LULAC). In a series of cases tracing back to Thornburg v. Gingles, 478 U. S. 30 (1986), we have interpreted this standard to mean that, under certain circumstance, States must draw “opportunity” districts in which minority groups form “effective majorit[ies],” LULAC, supra, at 426. Since the Equal Protection Clause restricts consideration of race and the VRA demands consideration of race, a legislature attempting to produce a lawful districting plan is vulnerable to “ ‘competing hazards of liability.’ ” Bush v. Vera, 517 U. S. 952, 977 (1996) (plurality opinion). In an effort to harmonize these conflicting demands, we have assumed that compliance with the VRA may justify the consideration of race in a way that would not otherwise be allowed. In technical terms, we have assumed that complying with the VRA is a compelling state interest, see, e.g., Bethune-Hill v. Virginia State Bd. of Elections, 580 U. S. ___, ___ (2017) (slip op., at 13); Shaw v. Hunt, 517 U. S. 899, 915 (1996), and that a State’s consideration of race in making a districting decision is narrowly tailored and thus satisfies strict scrutiny if the State has “ ‘good reasons’ ” for believing that its decision is necessary in order to comply with the VRA. Cooper, supra, at ___ (slip op., at 3). B Facing this legal obstacle course, the Texas Legislature in 2011 adopted new districting plans, but those plans were immediately tied up in litigation and were never used. Several plaintiff groups quickly filed challenges in the District Court for the Western District of Texas, arguing that some of the districts in the new plans were racial gerrymanders, some were based on intentional vote dilution, and some had the effect of depriving minorities of the equal opportunity to elect the candidates of their choice. This case was assigned to a three-judge court, as required by 28 U. S. C. §2284(a). (We will call this court “the Texas court” or simply “the District Court.”) The situation was further complicated by the requirement that Texas obtain preclearance of its new plans. To do this, Texas filed for a declaratory judgment in the District Court for the District of Columbia. See Texas v. United States, 887 F. Supp. 2d 133 (2012). (We will call this court “the D. C. court.”) By early 2012, the D. C. court had not yet issued a decision, and Texas needed usable plans for its rapidly approaching primaries. Accordingly, the Texas court drew up interim plans for that purpose. Perez v. Perry, 835 F. Supp. 2d 209 (2011). In creating those plans, the majority of the Texas court thought that it was not “required to give any deference to the Legislature’s enacted plan.” Id., at 213. Instead, it based its plans on what it called “neutral principles that advance the interest of the collective public good.” Id., at 212.[5] Texas challenged those court-ordered plans in this Court, and we reversed. Perry v. Perez, 565 U. S. 388 (2012) (per curiam). Noting that “[r]edistricting is ‘primarily the duty and responsibility of the State,’ ” we held that the Texas court should have respected the legislative judgments embodied in the 2011 plans to the extent allowed by the Constitution and the VRA. Id., at 392–399. We remanded the case with very specific instructions. The Texas court was told to start with the plans adopted by the Legislature but to modify those plans as needed so as “not to incorporate . . . any legal defects.” Id., at 394. With respect to claims under the Constitution or §2 of the VRA, the District Court was told to change a district if the plaintiffs were likely to succeed on the merits of their challenge. Ibid. And with respect to §5 claims, the court was instructed to make whatever changes were needed to obviate any legal claim that was “not insubstantial.”[6] Id., at 395. Thus, our instructions, in an abundance of caution, demanded changes in the challenged 2011 plans without proof that those changes were actually required by either the Constitution or the VRA. On remand, the Texas court ordered additional briefing and heard two more days of argument. App. 29a, 35a–50a; Order in Civ. No. 11–cv–00360, Doc. No. 616. It issued two opinions, totaling more than 70 pages, and analyzed disputed districts in detail. C. J. S. 367a–423a; H. J. S. 300a–315a. While stressing the preliminary nature of its determinations, see C. J. S. 368a; H. J. S. 314a–315a, the court found that some districts required change and that others were lawful, C. J. S. 367a–423a; H. J. S. 300a–315a. The court then adopted plans for the State’s congressional districts and for both houses of the State Legislature. (The plan for the State Senate is not at issue.) Both the congressional plan and the plan for the Texas House departed significantly from the State’s 2011 plans. At least 8 of the 36 congressional districts were markedly altered, and 21 districts in the plan for the Texas House were “substantially” changed. H. J. S. 314a; C. J. S. 397a–408a. In August 2012, the D. C. court denied preclearance of the plans adopted by the Legislature in 2011, see Texas v. United States, supra, so the State conducted the 2012 elections under the interim plans devised by the Texas court. At the same time, Texas filed an appeal in this Court contesting the decision of the D. C. court,[7] but that appeal ultimately died for two reasons. First, the 2011 plans were repealed. The Texas attorney general urged the Legislature to pass new redistricting plans, C. J. S. 429a, and in his view, the “best way to remedy the violations found by the D. C. court” was to “adopt the [Texas court’s] interim plans as the State’s permanent redistricting maps.” Id., at 432a. Doing so, he said, would “confirm the legislature’s intent” to adopt “a redistricting plan that fully comports with the law.” Id., at 429a. The Governor called a special session to do just that, and the Legislature complied. One of the legislative sponsors, Senator Seliger, explained that, although “ ‘the Texas Legislature remains confident that the legislatively-drawn maps adopted in 2011 are fair and legal . . . , there remain several outstanding legal questions regarding these maps that undermine the stability and predictability of the electoral process in Texas.’ ” 274 F. Supp. 3d 624, 649, n. 40 (2017). Counsel for one of the plaintiff groups, the Mexican American Legal Defense and Education Fund (MALDEF), testified in favor of the plans. C. J. S. 436a–439a. The 2013 Legislature then repealed the 2011 plans and enacted the Texas court’s interim plans with just a few minor changes. The federal congressional plan was not altered at all, and only small modifications were made to the plan for the Texas House. C. J. S. Findings 231a–232a. On the day after the Legislature passed the new plans and the day before the Governor signed them, this Court issued its decision in Shelby County v. Holder, 570 U. S. 529 (2013), which invalidated the coverage formula in §4 of the Voting Rights Act. Now no longer subject to §5, Texas obtained a vacatur of the D. C. court’s order on preclearance. 274 F. Supp. 3d, at 634–635, and n. 11. With the never-effective 2011 plans now repealed and any preclearance issues overcome by events, the State argued in the Texas court that the plaintiffs’ case against the 2011 plans was moot. In September 2013, the Texas court allowed the plaintiffs to amend their complaints to challenge the 2013 plans, but the court held that their challenges to the 2011 plans were still alive, reasoning that the repeal of the 2011 plans represented the “voluntary cessation” of allegedly unconstitutional conduct.[8] Texas conducted its 2014 and 2016 elections under the plans that had been preliminarily approved by the Texas court and subsequently adopted (with only minor changes) by the Legislature in 2013. But in March and April 2017, after multiple trials, the Texas court issued a pair of rulings on the defunct 2011 plans. The court reaffirmed the conclusions it had reached in 2012 about defects in the 2011 plans, and it went further. Contrary to its earlier decision, it held that Congressional District (CD) 35 is an impermissible racial gerrymander and that CD27 violates §2 of the Voting Rights Act because it has the effect of diluting the electoral opportunities of Latino voters. C. J. S. 181a, 193a–194a. Previously, the court had provided detailed reasons for rejecting the very arguments that it now accepted. Id., at 409a–423a. Similarly, the court held that multiple districts in the plan for the Texas House were the result of intentional vote dilution. These included districts in the counties of Nueces (House District (HD) 32, HD34), Bell (HD54, HD55), and Dallas (HD103, HD104, HD105). H. J. S. 275a–276a.[9] In August 2017, having ruled on the repealed 2011 plans, the Texas court finally turned its attention to the plans then in effect—i.e., the plans that had been developed by the court, adopted by the Legislature in 2013, and used in both the 2014 and 2016 elections. The court invalidated the districts in those plans that correspond to districts in the 2011 plan that it had just held to be unlawful, i.e., CD27, CD35, HD32, HD34, HD54, HD55, HD103, HD104, and HD105. See 274 F. Supp. 3d 624 (2017) (No. 17–586) and 267 F. Supp. 3d 750 (2017) (No. 17–626). In reaching these conclusions, the court pointed to the discriminatory intent allegedly harbored by the 2011 Legislature, and it attributed this same intent to the 2013 Legislature because it had failed to “engage in a deliberative process to ensure that the 2013 plans cured any taint from the 2011 plans.” 274 F. Supp. 3d, at 645–652; 267 F. Supp. 3d, at 757. The court saw “no indication that the Legislature looked to see whether any discriminatory taint remained in the plans.” 274 F. Supp. 3d, at 649. And it faulted the State because it “did not accept [findings of the D. C. court] and instead appealed to the Supreme Court.” Ibid. Seeing no evidence that the State had undergone “a change of heart,” the court concluded that the Legislature’s “decision to adopt the [District Court’s] plans” was a “litigation strategy designed to insulate the 2011 or 2013 plans from further challenge, regardless of their legal infirmities.” Id., at 649–650. Finally, summarizing its analysis, the court reiterated that the 2011 Legislature’s “discriminatory taint was not removed by the [2013] Legislature’s enactment of the Court’s interim plans, because the Legislature engaged in no deliberative process to remove any such taint, and in fact intended any such taint to be maintained but be safe from remedy.” Id., at 686. The Texas court’s decisions about CD35 and all but three of the Texas House districts were based entirely on its finding that the 2013 Legislature had not purged its predecessor’s discriminatory intent. However, the court also held that three districts—CD27, HD32, and HD34—were invalid under §2 of the Voting Rights Act because they had the effect of depriving Latinos of the equal opportunity to elect their candidates of choice. Id., at 682–686; 267 F. Supp. 3d, at 775–783. And the court found independent proof that HD90 was a racial gerrymander. Id., at 788–794. The court held that violations in all these districts “must be remedied.” 274 F. Supp. 3d, at 686; see also 267 F. Supp. 3d, at 795 (describing State House district violations that “must be remedied”). Mindful that October 1 was the deadline for the Texas Secretary of State to provide voter registration templates to the State’s counties, App. 380a–381a, the court took steps to bring about prompt remedial action. In two orders issued on August 15 and 24, the Texas attorney general was instructed to advise the court, within three days, “whether the Legislature intends to take up redistricting in an effort to cure these violations.” 274 F. Supp. 3d, at 686; 267 F. Supp. 3d, at 795. If the Legislature chose not to do so, the court warned, it would “hold a hearing to consider remedial plans.” Ibid. After the Governor made clear that the State would not act, the court ordered the parties to proceed with a hearing on the congressional plan on September 5, as well as a hearing on the plan for the Texas House on September 6. 274 F. Supp. 3d, at 686; 267 F. Supp. 3d, at 795; App. 134a–136a; Defendants’ Opposed Motion To Stay Order on Plan C235 Pending Appeal or Final Judgment in Civ. No. 11–cv–00360, Doc. 1538, pp. 3–4; Defendants’ Opposed Motion To Stay Order on Plan H358 Pending Appeal or Final Judgment, Doc. No. 1550, pp. 4–5. Texas applied for stays of both orders, but the District Court denied the applications. App. 134a–136a. Texas then asked this Court to stay the orders, and we granted that relief. After receiving jurisdictional statements, we postponed consideration of jurisdiction and set the cases for consolidated argument. 583 U. S. ___ (2018). II Before reaching the merits of these appeals, we must assure ourselves that we have jurisdiction to review the orders at issue. Appellants claim that the orders amount to injunctions and are therefore appealable to this Court under 28 U. S. C. §1253. Appellees disagree, contending that the orders do not qualify as injunctions. We hold that we have jurisdiction because the orders were effectively injunctions in that they barred Texas from using the districting plans now in effect to conduct this year’s elections. A The Judiciary Act of 1789, 1Stat. 73, “established the general principle that only final decisions of the federal district courts would be reviewable on appeal.” Carson v. American Brands, Inc., 450 U. S. 79, 83 (1981) (emphasis deleted). But because “rigid application of this principle was found to create undue hardship in some cases,” Congress created exceptions. Ibid. Two are relevant here. We have jurisdiction under 28 U. S. C. §1253 to hear an appeal from an order of a three-judge district court “granting or denying . . . an interlocutory or permanent injunction.” Similarly, §1292(a)(1) gives the courts of appeals jurisdiction over “[i]nterlocutory orders of the district courts” “granting, continuing, modifying, refusing or dissolving injunctions,” “except where a direct review may be had in the Supreme Court.” The orders in these cases fall within §1253. To be sure, the District Court did not call its orders “injunctions”—in fact, it disclaimed the term, App. 134a–136a—but the label attached to an order is not dispositive. We have previously made clear that where an order has the “practical effect” of granting or denying an injunction, it should be treated as such for purposes of appellate jurisdiction. Carson, supra, at 83; see also Gulfstream Aerospace Corp. v. Mayacamas Corp., 485 U. S. 271, 287–288 (1988). We applied this test in Carson, holding that an order that declined to enter a consent decree prohibiting certain conduct could be appealed under §1292(a)(1) because it was the practical equivalent of an order denying an injunction and threatened serious and perhaps irreparable harm if not immediately reviewed. 450 U. S., at 83–84, 86–90. This “practical effect” rule serves a valuable purpose. If an interlocutory injunction is improperly granted or denied, much harm can occur before the final decision in the district court. Lawful and important conduct may be barred, and unlawful and harmful conduct may be allowed to continue. Recognizing this, Congress authorized interlocutory appellate review of such orders. But if the availability of interlocutory review depended on the district court’s use of the term “injunction” or some other particular language, Congress’s scheme could be frustrated. The harms that Congress wanted to avoid could occur so long as the district court was careful about its terminology. The “practical effect” inquiry prevents such manipulation. In analogous contexts, we have not allowed district courts to “shield [their] orders from appellate review” by avoiding the label “injunction.” Sampson v. Murray, 415 U. S. 61, 87 (1974). For instance, in Sampson, we held that an order labeled a temporary restraining order (which is not appealable under §1292(a)(1)) should be treated as a “preliminary injunction” (which is appealable) since the order had the same practical effect as a preliminary injunction. Id., at 86–88. Appellees and the dissent contend that the “practical effect” approach should be confined to §1292(a)(1), but we see no good reason why it should not apply to §1253 as well. Appellees note that we “narrowly constru[e]” §1253, Goldstein v. Cox, 396 U. S. 471, 478 (1970), but we also construe §1292(a)(1) “narrowly,” Carson, supra, at 84. In addition, the relevant language in the two provisions is nearly identical;[10] both provisions serve the same purpose; and we have previously called them “analogous.” Goldstein, supra, at 475. The provisions are also textually interlocked. Section 1292(a)(1) does not apply where “direct review may be had in the Supreme Court,” i.e., where §1253 applies. If the “practical effects” test applied under §1292(a)(1) but not §1253, the consequences would be unfortunate and strange. We would have to identify the magic language needed for an order to qualify as an order granting or denying an injunction, and that standard would hardly constitute the sort of “[s]imple” rule that the dissent prizes. See post, at 14 (opinion of Sotomayor, J.). Then, having developed that standard, we would have to apply it in any case in which a party took an appeal to us from an order of a three-judge court that clearly had the practical effect of an injunction. If we concluded that the magic-words test was not met, the order would appear to be appealable to one of the courts of appeals under §1292(a)(1). In the language of that provision, the order would be an “orde[r] of [a] district cour[t] of the United States . . . granting [an] injunctio[n].” And because this Court would lack jurisdiction under §1253, the appeal would not fall within §1292(1)’s exception for cases “where a direct review may be had in the Supreme Court.” Having taken pains to provide for review in this Court, and not in the courts of appeals, of three-judge court orders granting injunctions Congress surely did not intend to produce that result.[11] Appellees argue that an order denying an injunction (the situation in Carson) and an order granting an injunction (the situation here) should be treated differently, Brief for Appellees in No. 17–586, p. 27, but they offer no convincing reason for doing so. No authority supports their argument. The language of §§1253 and 1292(a)(1) makes no such distinction, and we have stated that the “practical effect” analysis applies to the “granting or denying” of injunctions. Gulfstream, supra, at 287–288. In addition, appellees’ suggested distinction would put appellate courts in an awkward position. Suppose that a district court granted an injunction that was narrower than the one requested by the moving party. Would an appellate court (whether this Court or a court of appeals) have jurisdiction to rule on only part of that decision? Suppose the appellate court concluded that the district court was correct in refusing to give the movant all the injunctive relief it sought because the movant’s entire claim was doomed to fail. Would the appellate court be limited to holding only that the lower court properly denied the relief that was withheld? The rule advocated by the appellees would needlessly complicate appellate review.[12] Finally, appellees point in passing to Rule 65(d) of the Federal Rules of Civil Procedure, which requires that an injunction “state its terms specifically” and “describe in reasonable detail . . . the act or acts restrained or required.” Rules 65(d)(1)(B), (C); see Brief for Appellees in No. 17–586, at 27. But as explained in Gunn v. University Comm. to End War in Viet Nam, 399 U. S. 383, 389, n. 4 (1970), we have never suggested that a failure to meet the specificity requirements of Rule 65(d) would “deprive the Court of jurisdiction under §1253.” A contrary holding would be perverse. Rule 65(b) protects the party against which an injunction is issued by requiring clear notice as to what that party must do or refrain from doing. Where a vague injunction does not comply with Rule 65(b), the aggrieved party has a particularly strong need for appellate review. It would be odd to hold that there can be no appeal in such a circumstance. For these reasons, we hold that we have jurisdiction under §1253 to hear an appeal from an order that has the same practical effect as one granting or denying an injunction. B With these principles settled, we conclude that the orders in these cases qualify as interlocutory injunctions under §1253. The text of the orders and the context in which they were issued make this clear. The orders are unequivocal that the current legislative plans “violate §2 and the Fourteenth Amendment” and that these violations “must be remedied.” 274 F. Supp. 3d, at 686; see also, e.g., 267 F. Supp. 3d, at 795 (“[V]iolations found by this Court in its Order on [the State House plan] now require a remedy”); ibid. (“In Bell County, the intentional discrimination previously found by the Court must be remedied”); ibid. (“In Dallas County, the intentional discrimination previously found by the Court must be remedied”). We do not suggest that this language alone is sufficient to show that the orders had the practical effect of enjoining use of the current plans in this year’s elections, but the court did not stop with these pronouncements. As we have noted, the orders required the Texas attorney general to inform the court within three days whether the Legislature would remedy the violations, and the orders stated that if the Legislature did not intend to adopt new plans, the court would hold remedial hearings. The short time given the Legislature to respond is strong evidence that the three-judge court did not intend to allow the elections to go ahead under the plans it had just condemned. The Legislature was not in session, so in order to take up the task of redistricting, the Governor would have been required to convene a special session—which is no small matter. And, when the Governor declined to call a special session, the court moved ahead with its scheduled hearings and invited the parties to continue preparing for them even after this Court administratively stayed the August 15 order. The import of these actions is unmistakable: The court intended to have new plans ready for use in this year’s elections. Nothing in the record even hints that the court contemplated the possibility of allowing the elections to proceed under the 2013 plans. What is more, Texas had reason to believe that it would risk deleterious consequences if it defied the court and attempted to conduct the elections under the plans that the court had found to be based on intentional racial discrimination. In the very orders at issue, the court inferred discriminatory intent from Texas’s choice to appeal the D. C. court’s preclearance decision rather than immediately taking steps to bring its plans into compliance with that decision. 274 F. Supp. 3d, at 649; see Part III, infra. Reading such an order, Texas had reason to fear that if it tried to conduct elections under plans that the court had found to be racially discriminatory, the court would infer an evil motive and perhaps subject the State once again to the strictures of preclearance under §3(c) of the Voting Rights Act.[13] This is a remedy that the plaintiffs hoped to obtain, see, e.g., App. 177a, and that the District Court seemed inclined to consider, see C. J. S. 122a–123a (declining to declare moot the challenges to the long-since-repealed 2011 plans because “there remains the possibility of declaratory and equitable relief under §3(c)”). Contending that the orders here do not qualify under §1253, appellees analogize this case to Gunn, 399 U. S. 383, but there is no relevant similarity. In Gunn, anti-war protesters were charged with violating a Texas “disturbing-the-peace statute,” id., at 384, and they challenged the constitutionality of the statute in federal court. After the state charges were dismissed, the District Court issued a “discursive” opinion “expressing the view that [the statute was] constitutionally invalid.” Id., at 386–387. But the court then refrained from going any further, “pending the next session, special or general, of the Texas legislature, at which time the State of Texas may, if it so desires, enact such disturbing-the-peace statute as will meet constitutional requirements.” University Comm. to End War in Viet Nam v. Gunn, 289 F. Supp. 469, 475 (WD Tex. 1968). The defendants appealed to this Court, and at the time of our decision two years later, neither the Legislature nor the District Court had taken any further action. We therefore held that we lacked jurisdiction under §1253. The District Court order in that case did not have the same practical effect as an injunction. Indeed, it had no practical effect whatsoever and is thus entirely different from the orders now before us.[14] Appellees suggest that appellate jurisdiction is lacking in this case because we do not know at this point “what a remedy would entail, who it would affect, and when it would be implemented.” Brief for Appellees in No. 17–586, at 27. The dissent makes a similar argument with respect to two of the Texas House districts. Post, at 13–14.[15] But the issue here is whether this year’s elections can be held under the plans enacted by the Legislature, not whether any particular remedies would have ulti- mately been ordered by the District Court. Appellees and the dissent also fret that this Court will be inundated with redistricting appeals if we accept jurisdiction here, Brief for Appellees in No. 17–626, p. 34; post, at 15–16, and n. 8, but there is no reason to fear such a flood. Because §1253 expressly authorizes “interlocutory” appeals, there is no question that there can be more than one appeal in a case challenging a redistricting plan. District courts sometimes expressly enjoin the use of districting plans before moving on to the remedial phase. See, e.g., Whitford v. Gill, No. 3:15–cv–421, Doc. No. 190 (WD Wis., Feb. 22, 2017); Harris v. McCrory, No. 1:13–cv–949, Doc. No. 143 (MDNC, Feb. 5, 2016). But appeals from such orders have not overwhelmed our docket. Our holding here will affect only a small category of additional cases.[16] It should go without saying that our decision does not mean that a State can always appeal a district court order holding a redistricting plan unlawful. A finding on liability cannot be appealed unless an injunction is granted or denied, and in some cases a district court may see no need for interlocutory relief. If a plan is found to be unlawful long before the next scheduled election, a court may defer any injunctive relief until the case is completed. And if a plan is found to be unlawful very close to the election date, the only reasonable option may be to use the plan one last time. We appreciate our obligation to heed the limits of our jurisdiction, and we reiterate that §1253 must be strictly construed. But it also must be sensibly construed, and here the District Court’s orders, for all intents and purposes, constituted injunctions barring the State from conducting this year’s elections pursuant to a statute enacted by the Legislature. Unless that statute is unconstitutional, this would seriously and irreparably harm[17] the State, and only an interlocutory appeal can protect that State interest. See Carson, supra, at 89–90. As a result, §1253 provides jurisdiction. III We now turn to the merits of the appeal. The primary question is whether the Texas court erred when it required the State to show that the 2013 Legislature somehow purged the “taint” that the court attributed to the defunct and never-used plans enacted by a prior legislature in 2011. A Whenever a challenger claims that a state law was enacted with discriminatory intent, the burden of proof lies with the challenger, not the State. Reno v. Bossier Parish School Bd., 520 U. S. 471, 481 (1997). This rule takes on special significance in districting cases. Redistricting “is primarily the duty and responsibility of the State,” and “[f]ederal-court review of districting legislation represents a serious intrusion on the most vital of local functions.” Miller v. Johnson, 515 U. S. 900, 915 (1995) (internal quotation marks omitted). “[I]n assessing the sufficiency of a challenge to a districting plan,” a court “must be sensitive to the complex interplay of forces that enter a legislature’s redistricting calculus.” Id., at 915–916. And the “good faith of [the] state legislature must be presumed.” Id., at 915. The allocation of the burden of proof and the presumption of legislative good faith are not changed by a finding of past discrimination. “[P]ast discrimination cannot, in the manner of original sin, condemn governmental action that is not itself unlawful.” Mobile, 446 U. S., at 74 (plurality opinion). The “ultimate question remains whether a discriminatory intent has been proved in a given case.” Ibid. The “historical background” of a legislative enactment is “one evidentiary source” relevant to the question of intent. Arlington Heights v. Metropolitan Housing Development Corp., 429 U. S. 252, 267 (1977). But we have never suggested that past discrimination flips the evidentiary burden on its head. Neither the District Court nor appellees have pointed to any authority that would justify shifting the burden. The appellees rely primarily on Hunter v. Underwood, 471 U. S. 222 (1985), but that case addressed a very different situation. Hunter involved an equal protection challenge to an article of the Alabama Constitution adopted in 1901 at a constitutional convention avowedly dedicated to the establishment of white supremacy. Id., at 228–230. The article disenfranchised anyone convicted of any crime on a long list that included many minor offenses. Id., at 226–227. The court below found that the article had been adopted with discriminatory intent, and this Court accepted that conclusion. Id., at 229. The article was never repealed, but over the years, the list of disqualifying offenses had been pruned, and the State argued that what remained was facially constitutional. Id., at 232–233. This Court rejected that argument because the amendments did not alter the intent with which the article, including the parts that remained, had been adopted. Id., at 233. But the Court specifically declined to address the question whether the then-existing version would have been valid if “[re]enacted today.” Ibid. In these cases, we do not confront a situation like the one in Hunter. Nor is this a case in which a law originally enacted with discriminatory intent is later reenacted by a different legislature. The 2013 Texas Legislature did not reenact the plan previously passed by its 2011 predecessor. Nor did it use criteria that arguably carried forward the effects of any discriminatory intent on the part of the 2011 Legislature. Instead, it enacted, with only very small changes, plans that had been developed by the Texas court pursuant to instructions from this Court “not to incorporate . . . any legal defects.” Perry, 565 U. S., at 394. Under these circumstances, there can be no doubt about what matters: It is the intent of the 2013 Legislature. And it was the plaintiffs’ burden to overcome the presumption of legislative good faith and show that the 2013 Legislature acted with invidious intent. The Texas court contravened these basic principles. Instead of holding the plaintiffs to their burden of overcoming the presumption of good faith and proving discriminatory intent, it reversed the burden of proof. It imposed on the State the obligation of proving that the 2013 Legislature had experienced a true “change of heart” and had “engage[d] in a deliberative process to ensure that the 2013 plans cured any taint from the 2011 plans.” 274 F. Supp. 3d, at 649. The Texas court’s references to the need to “cure” the earlier Legislature’s “taint” cannot be dismissed as stray comments. On the contrary, they were central to the court’s analysis. The court referred repeatedly to the 2013 Legislature’s duty to expiate its predecessor’s bad intent, and when the court summarized its analysis, it drove the point home. It stated: “The discriminatory taint [from the 2011 plans] was not removed by the Legislature’s enactment of the Court’s interim plans, because the Legislature engaged in no deliberative process to remove any such taint, and in fact intended any such taint to be maintained but be safe from remedy.” Id., at 686.[18] The dissent labors to explain away all these references to the 2013’s Legislature’s supposed duty to purge its predecessor’s allegedly discriminatory intent, but the dissent loses track of its own argument and characterizes the District Court’s reasoning exactly as we have. Indeed, the dissent criticizes us on page 33 of its opinion for saying precisely the same thing that it said 11 pages earlier. On page 33, the dissent states: “[T]he majority quotes the orders as requiring proof that the Legislature “ ‘engage[d] in a deliberative process to ensure that the 2013 plans cured any taint from the 2011 plans.’ ” But the District Court did not put the burden on Texas to make that affirmative showing.” Post, at 33–34 (quoting supra, at 23, in turn quoting 274 F. Supp. 3d, at 649; citations omitted). But earlier, the dissent itself describes the District Court’s analysis as follows: “Despite knowing of the discrimination in its 2011 maps, ‘the Legislature did not engage in a deliberative process to ensure that the 2013 plans cured any taint from the 2011 plans.’ ” Post, at 22 (quoting 274 F. Supp. 3d, at 649). And this is not just a single slip of the pen. The dissent writes that the District Court was required “to assess how the 2013 Legislature addressed the known discrimination that motivated” the districts approved by that Court in 2012. Post, at 31. The dissent quotes the District Court’s statement that “ ‘there is no indication that the Legislature looked to see whether any discriminatory taint remained in the plans.’ ” Post, at 24 (quoting 274 F. Supp. 3d, at 649). And there is also this: “Texas was just ‘not truly interested in fixing any remaining discrimination in [its 2011 maps].’ ” Post, at 22 (quoting 274 F. Supp. 3d, at 651, n. 45). The District Court’s true mode of analysis is so obvious that the dissent cannot help but repeat it. And that approach was fundamentally flawed and demands reversal. While a district court’s finding of fact on the question of discriminatory intent is reviewed for clear error, see Cromartie, 532 U. S., at 242, whether the court applied the correct burden of proof is a question of law subject to plenary review, U. S. Bank N. A. v. Village at Lakeridge, LLC, 583 U. S. ___, ___ (2018) (slip op., at 5); Highmark Inc. v. Allcare Health Management System, Inc., 572 U. S. ___, ___ (2014) (slip op., at 4). And when a finding of fact is based on the application of an incorrect burden of proof, the finding cannot stand. Bose Corp. v. Consumers Union of United States, Inc., 466 U. S. 485, 501 (1984) (“An appellate cour[t has] power to correct errors of law, including those that . . . infect . . . a finding of fact that is predicated on a misunderstanding of the governing rule of law”). B In holding that the District Court disregarded the presumption of legislative good faith and improperly reversed the burden of proof, we do not suggest either that the intent of the 2011 Legislature is irrelevant or that the plans enacted in 2013 are unassailable because they were previously adopted on an interim basis by the Texas court. Rather, both the intent of the 2011 Legislature and the court’s adoption of the interim plans are relevant to the extent that they naturally give rise to—or tend to refute—inferences regarding the intent of the 2013 Legislature. They must be weighed together with any other direct and circumstantial evidence of that Legislature’s intent. But when all the relevant evidence in the record is taken into account, it is plainly insufficient to prove that the 2013 Legislature acted in bad faith and engaged in intentional discrimination.[19] See, e.g., Ricci v. DeStefano, 557 U. S. 557, 585 (2009); McCleskey v. Zant, 499 U. S. 467, 497 (1991). There is thus no need for any further prolongation of this already protracted litigation. The only direct evidence brought to our attention suggests that the 2013 Legislature’s intent was legitimate. It wanted to bring the litigation about the State’s districting plans to an end as expeditiously as possible. The attorney general advised the Legislature that the best way to do this was to adopt the interim, court-issued plans. The sponsor of the 2013 plans voiced the same objective, and the Legislature then adopted the court-approved plans. On its face, this explanation of the Legislature’s intent is entirely reasonable and certainly legitimate. The Legislature had reason to know that any new plans it devised were likely to be attacked by one group of plaintiffs or another. (The plaintiffs’ conflicting positions with regard to some of the districts in the plans now before us bear this out.) Litigating districting cases is expensive and time consuming, and until the districts to be used in the next election are firmly established, a degree of uncertainty clouds the electoral process. Wishing to minimize these effects is understandable and proper. The court below discounted this direct evidence, but its reasons for doing so are not sound. The court stated that the “strategy” of the 2013 Legislature was to “insulate [the plans] from further challenge, regardless of [the plans’] legal infirmities.” 274 F. Supp. 3d, at 650; see also id., at 651, n. 45. But there is no evidence that the Legislature’s aim was to gain acceptance of plans that it knew were unlawful.[20] Indeed, there is no evidence that the Legislature thought that the plans were invalid—and as we will explain, the Legislature had sound reasons to believe just the opposite.[21] The District Court found it significant that the Legislature must have realized that enacting the interim plans would not “end the litigation,” because it knew that at least some plaintiffs would pursue their challenges anyway. Id., at 651, n. 45. But even if, as seems likely, the Legislature did not think that all the plaintiffs would immediately abandon all their claims, it does not follow that the Legislature was insincere in stating that it adopted the court-approved plan with the aim of bringing the litigation to a close. It was reasonable for the Legislature to think that approving the court-approved plans might at least reduce objections and thus simplify and expedite the conclusion of the litigation.[22] That MALDEF, counsel for one of the plaintiff groups, testified in favor of the plans is evidence that the Legislature’s objective was reasonable. C. J. S. 436a–439a. Not only does the direct evidence suggest that the 2013 Legislature lacked discriminatory intent, but the circumstantial evidence points overwhelmingly to the same conclusion. Consider the situation when the Legislature adopted the court-approved interim plans. First, the Texas court had adopted those plans, and no one would claim that the court acted with invidious intent when it did so. Second, the Texas court approved those plans only after reviewing them and modifying them as required to comply with our instructions. Not one of the judges on that court expressed the view that the plans were unlawful. Third, we had directed the Texas court to make changes in response to any claims under the Equal Protection Clause and §2 of the Voting Rights Act if those claims were merely likely to prevail. Perry, 565 U. S., at 394. And the Texas court was told to accommodate any claim under §5 of the VRA unless it was “insubstantial.” Id., at 395. Fourth, the Texas court had made a careful analysis of all the claims, had provided a detailed examination of individual districts, and had modified many districts. Its work was anything but slapdash. All these facts gave the Legislature good reason to believe that the court-approved interim plans were legally sound. Is there any evidence from which a contrary inference can reasonably be drawn? Appellees stress the preliminary nature of the Texas court’s approval of the interim plans, and as we have said, that fact is relevant. But in light of our instructions to the Texas court and the care with which the interim plans were developed, the court’s approval still gave the Legislature a sound basis for thinking that the interim plans satisfied all legal requirements. The court below and the dissent infer bad faith because the Legislature “pushed the redistricting bills through quickly in a special session.” 274 F. Supp. 3d, at 649. But we do not see how the brevity of the legislative process can give rise to an inference of bad faith—and certainly not an inference that is strong enough to overcome the presumption of legislative good faith (a concept to which the dissent pays only the briefest lip service, post, at 21). The “special session” was necessary because the regular session had ended. As explained, the Legislature had good reason to believe that the interim plans were sound, and the adoption of those already-completed plans did not require a prolonged process. After all, part of the reason for adopting those plans was to avoid the time and expense of starting from scratch and leaving the electoral process in limbo while that occurred.[23] The District Court and the dissent also err when they charge that Representative Darby, the chair of the Texas House Redistricting Committee at the time in question, “ ‘willfully ignored those who pointed out deficiencies’ ” in the plans. Post, at 23 (quoting 274 F. Supp. 3d, at 651, n. 45). This accusation is not only misleading, it misses the point. The Legislature adopted the interim plans in large part because they had the preliminary approval of the District Court, and Darby was open about the fact that he wanted to minimize amendments to the plans for that reason. See, e.g., Joint Exh. 17.3, pp. S1–S2. That Darby generally hoped to minimize amendments—so that the plans would remain legally compliant—hardly shows that he, or the Legislature, acted with discriminatory intent. In any event, it is misleading to characterize this attitude as “willfu[l] ignor[ance].” The record shows that, al- though Darby hoped to minimize amendments, he did not categorically refuse to consider changes. This is illustrated by his support for an amendment to HD90, which was offered by the then-incumbent, Democrat Lon Burnam, precisely because it fixed an objection raised by the Mexican-American Legal Caucus (MALC) that the district’s Latino population was too low. 267 F. Supp. 3d, at 790.[24] The Texas court faulted the 2013 Legislature for failing to take into account the problems with the 2011 plans that the D. C. court identified in denying preclearance, ibid., but the basis for that criticism is hard to understand. One of the 2013 Legislature’s principal reasons for adopting the court-approved plans was to fix the problems identified by the D. C. court. The attorney general advised the Legislature to adopt the interim plans because he thought that was the “best way to remedy the violations found by the D. C. court.” C. J. S. 432a. Chairman Darby similarly stated that the 2013 plans fixed the errors found by the D. C. court, Tr. 1498, 1584–1585 (July 14, 2017), as did Senator Seliger, Joint Exh. 26.2, p. A–5. There is nothing to suggest that the Legislature proceeded in bad faith—or even that it acted unreasonably—in pursuing this strategy. Recall that we instructed the Texas court, in developing the interim plans, to remedy any §5 claim that was “not insubstantial.” Perry, 565 U. S., at 395. And that is just what the interim plans, which the Legislature later enacted, attempted to do. For instance, the D. C. court held that the congressional plan had one too few “ability to elect” districts for Latinos, largely because of changes to CD23, Texas, 887 F. Supp. 2d, at 156–159; the interim plan (and, by extension, the 2013 plan) amended CD23, C. J. S. 397a–399a. Similarly, in the plan for the Texas House, the D. C. court found §5 retrogression with respect to HD35, HD117, and HD149, Texas, supra, at 167–175, and all of those districts were changed in the 2013 plans, H. J. S. 305a–307a, 312a. Although the D. C. court found that the 2011 Legislature acted with discriminatory intent in framing the congressional plan, that finding was based on evidence about districts that the interim plan later changed. The D. C. court was concerned about the intent reflected in the drawing of CDs 9, 18, and 30, but all those districts were amended by the Texas court. Texas, supra, at 159–160; C. J. S. 406a–408a. With respect to the plan for the Texas House, the D. C. court made no intent findings, but its areas of concern were generally addressed by the Texas court and the 2013 plans. Compare Texas, supra, at 178 (noting evidence of unlawful intent in HD117), with H. J. S. 307a (amending HD117).[25] It is indicative of the District Court’s mistaken approach that it inferred bad faith from Texas’s decision to take an appeal to this Court from the D. C. court’s decision denying preclearance. See 274 F. Supp. 3d, at 649 (“Defendants did not accept [these findings] and instead appealed to the Supreme Court”). Congress gave the State the right to appeal, and no bad motive can be inferred from its decision to make use of this right—unless of course the State had no reasonable grounds for appeal. Before our decision in Shelby County mooted Texas’s appeal to this Court from the D. C. court’s preclearance decision, Texas filed a jurisdictional statement claiming that the D. C. court made numerous errors, but the Texas court made no attempt to show that Texas’s arguments were frivolous. As a final note, appellees assert that the 2013 Legislature should have either defended the 2011 plans in litigation or gone back to the drawing board and devised en- tirely new plans, Brief for Appellees in No. 17–626, at 45, but there is no reason why the Legislature’s options should be limited in this way. It was entirely permissible for the Legislature to favor a legitimate option that promised to simplify and reduce the burden of litigation. That the Legislature chose this course is not proof of discriminatory intent. IV Once the Texas court’s intent finding is reversed, there remain only four districts that were invalidated on alternative grounds. For three of these districts, the District Court relied on the “effects” test of §2. We reverse as to each of these, but we affirm the District Court’s final holding that HD90 is a racial gerrymander. A To make out a §2 “effects” claim, a plaintiff must establish the three so-called “Gingles factors.” These are (1) a geographically compact minority population sufficient to constitute a majority in a single-member district, (2) political cohesion among the members of the minority group, and (3) bloc voting by the majority to defeat the minority’s preferred candidate. Gingles, 478 U. S., at 48–51; LULAC, 548 U. S., at 425. If a plaintiff makes that showing, it must then go on to prove that, under the totality of the circumstances, the district lines dilute the votes of the members of the minority group. Id., at 425–426. 1 The Texas court held that CD27 violates §2 of the VRA because it has the effect of diluting the votes of Latino voters in Nueces County. C. J. S. 191a. CD27 is anchored in Nueces County (home to Corpus Christi) and follows the Gulf of Mexico to the northeast before taking a turn inland to the northwest in the direction of Austin. Nueces County contains a Latino population of roughly 200,000 (a little less than one-third the size of an ideal Texas congressional district), and the court held that the Nueces County Latinos should have been included in a Latino opportunity district, rather than CD27, which is not such a district. The court found that an area centered on Nueces County satisfies the Gingles factors and that, under the totality of the circumstances, the placement of the Nueces County Latinos in CD27 deprives them of the equal opportunity to elect candidates of their choice. C. J. S. 181a–195a. The problem with this holding is that plaintiffs could not establish a violation of §2 of the VRA without showing that there is a “ ‘possibility of creating more than the existing number of reasonably compact’ ” opportunity districts. LULAC, supra, at 430. And as the Texas court itself found, the geography and demographics of south and west Texas do not permit the creation of any more than the seven Latino opportunity districts that exist under the current plan. 274 F. Supp. 3d, at 684, and n. 85. Attempting to get around this problem, the Texas court relied on our decision in LULAC, but it misapplied our holding. In LULAC, we held that the State should have created six proper Latino opportunity districts but instead drew only five. 548 U. S., at 435. Although the State claimed that the plan actually included a sixth opportunity district, that district failed to satisfy the Gingles factors. 548 U. S., at 430. We held that a “State’s creation of an opportunity district for those without a §2 right offers no excuse for its failure to provide an opportunity district for those with a §2 right.” Ibid. Here, the Texas court concluded that Texas committed the same violation as in LULAC: It created “an opportu-nity district for those without a §2 right” (the Latinos in CD35), while failing to create such a district “for those with a §2 right” (the Latinos of Nueces County). Ibid. This holding is based on a flawed analysis of CD35. CD35 lies to the north of CD27 and runs along I–35 from San Antonio up to Austin, the center of Travis County. In the District Court’s view, the Latinos of CD35 do not have a §2 right because one of the Gingles factors, majority bloc voting, is not present. The Court reached this conclusion because the non-Latino voters of Travis County tend to favor the same candidates as the great majority of Latinos. There are two serious problems with the District Court’s analysis. First, the Court took the wrong approach in evaluating the presence of majority bloc voting in CD35. The Court looked at only one, small part of the district, the portion that falls within Travis County. 274 F. Supp. 3d, at 683; C. J. S. 175a–176a. But Travis County makes up only 21% of the district. We have made clear that redistricting analysis must take place at the district level. Bethune-Hill, 580 U. S., at ___ (slip op., at 12). In failing to perform that district-level analysis, the District Court went astray. Second, here, unlike in LULAC, the 2013 Legislature had “good reasons” to believe that the district at issue (here CD35) was a viable Latino opportunity district that satisfied the Gingles factors. CD35 was based on a concept proposed by MALDEF, C. J. S. Findings 315a–316a, and the Latino Redistricting Task Force (a plaintiff group) argued that the district is mandated by §2. C. J. S. 174a. The only Gingles factor disputed by the court was majority bloc voting, and there is ample evidence that this factor is met. Indeed, the court found that majority bloc voting exists throughout the State. C. J. S. Findings 467a. In addition, the District Court extensively analyzed CD35 in 2012 and determined that it was likely not a racial gerrymander and that even if it was, it likely satisfied strict scrutiny. C. J. S. 415a. In other words, the 2013 Legislature justifiably thought that it had placed a viable opportunity district along the I–35 corridor. 2 The District Court similarly erred in holding that HD32 and HD34 violate §2. These districts make up the entirety of Nueces County, which has a population that is almost exactly equal to twice the population of an ideal Texas House district. (It can fit 2.0295 ideal districts. H. J. S. Findings 91a.) In 2010, Latinos made up approximately 56% of the voting age population of the county. Ibid. The 2013 plan created two districts that lie wholly within the county; one, HD34, is a Latino opportunity district, but the other, HD32, is not. 267 F. Supp. 3d, at 767. Findings made by the court below show that these two districts do not violate §2 of the Voting Rights Act. Under Gingles, the ultimate question is whether a districting decision dilutes the votes of minority voters, see LULAC, supra, at 425–426, and it is hard to see how this standard could be met if the alternative to the districting decision at issue would not enhance the ability of minority voters to elect the candidates of their choice. The only plaintiff that pressed a §2 claim with respect to HD32 and HD34 was MALC, 267 F. Supp. 3d, at 767, and as the District Court recognized, that group’s own expert determined that it was not possible to divide Nueces County into more than one performing Latino district. In his analysis, the expert relied on Nueces County election returns for statewide elections between 2010 and 2016. Id., at 775–776. Based on this data, he calculated that when both HD32 and HD34 were maintained as Latino-majority districts, one performed for Latinos in only 7 out of 35 relevant elections, and the other did so in none of the 35 elections. Ibid. In order to create two performing districts in that area, it was necessary, he found, to break county lines in multiple places, id., at 778, but the District Court held that “breaking the County Line Rule” in the Texas Constitution, see Art. III, §26, to “remove Anglos and incorporate even more Hispanics to improve electoral outcomes goes beyond what §2 requires.” 267 F. Supp. 3d, at 783. So if Texas could not create two performing districts in Nueces County and did not have to break county lines, the logical result is that Texas did not dilute the Latino vote. The court refused to accept this conclusion, but its reasons for doing so cannot stand up. As an initial matter, the court thought that the two districts would have to be redrawn based on its finding regarding the intent of the 2013 Legislature,[26] and it therefore deferred a final decision on the §2 issue and advised the plaintiffs to consider at the remedial phase of the case whether they preferred to have two districts that might not perform or just one safe district. Id., at 783. The court’s decision cannot be sustained on this ground, since its finding of discriminatory intent is erroneous. The only other reason provided by the court was the observation that MALC “failed to show” that two majority-Latino districts in Nueces County would not perform. Id., at 782. This observation twisted the burden of proof beyond recognition. It suggested that a plaintiff might succeed on its §2 claim because its expert failed to show that the necessary factual basis for the claim could not be established.[27] Courts cannot find §2 effects violations on the basis of uncertainty. In any event, if even the District Court remains unsure how to draw these districts to comply with §2 (after six years of litigation, almost a dozen trials, and numerous opinions), the Legislature surely had the “ ‘broad discretion’ ” to comply as it reasonably saw fit in 2013, LULAC, supra, at 429. The dissent charges us with ignoring the District Court’s “ ‘intensely local appraisal” ’ of Nueces County, post, at 44, but almost none of the “findings” that the District Court made with respect to HD32 and HD34 referred to present local conditions, and none cast any significant light on the question whether another opportunity district is possible at the present time. For instance, what the dissent describes as Texas’s “long ‘history of voting-related discrimination,’ ” ibid., in no way undermines—or even has any logical bearing on—the conclusions reached by MALC’s expert about whether Latino voters would have a real opportunity to elect the candidates of their choice if the county was divided into two districts with narrow majorities of Latino citizens of voting age. The same is true with respect to the District Court’s findings regarding racially polarized voting in the county and Latinos’ “continuing pattern of disadvantage relative to” non-Latinos. 267 F. Supp. 3d, at 779 (internal quotation marks omitted). Perhaps recognizing as much, both the District Court and the dissent point to the anticipated future growth in the percentage of eligible voters of Latino descent, but the districts now at issue would not necessar-ily be used beyond 2020, after which time the 2020 census would likely require redistricting once again. B HD90 is a district in Tarrant County that, unlike the other districts at issue in this appeal, was not copied from the District Court’s interim plans. Instead, the 2013 Legislature substantially modified the district developed by the District Court, and the District Court held that the 2013 Legislature’s creation is an invalid racial gerrymander. 267 F. Supp. 3d, at 794. In drawing HD90, the Legislature was pulled in opposite directions by competing groups. In 2011, the Legislature, responding to pressure from MALDEF, increased the Latino population of the district in an effort to make it a Latino opportunity district. H. J. S. Findings 258a–262a. In the process of doing so, the Legislature moved the community of Como, which is predominantly African-American, out of the district. But Como residents and the member of the Texas House who represented the district, Lon Burnam, objected, and in 2013, the Legislature moved Como back into the district. 267 F. Supp. 3d, at 788–789. That change was opposed by MALC because it decreased the Latino population below 50%. App. 398a–399a. So the Legislature moved Latinos into the district to bring the Latino population back above 50%. 267 F. Supp. 3d, at 789–790. In light of these maneuvers, Texas does not dispute that race was the predominant factor in the design of HD90, but it argues that this was permissible because it had “ ‘good reasons to believe’ ” that this was necessary to satisfy §2 of the Voting Rights Act.” Bethune-Hill, 580 U. S., at ___ (slip op., at 14). Texas offers two pieces of evidence to support its claim. The first—that one of the plaintiffs, MALC, demanded as much—is insufficient. A group that wants a State to create a district with a particular design may come to have an overly expansive understanding of what §2 demands. So one group’s demands alone cannot be enough. The other item of evidence consists of the results of the Democratic primaries in 2012 and 2014. In 2012, Representative Burnham, who was not the Latino candidate of choice, narrowly defeated a Latino challenger by 159 votes. And in 2014, the present representative, Ramon Romero, Jr., beat Burnam by 110 votes. See Brief for Appellants 70. These election returns may be suggestive, but standing alone, they were not enough to give the State good reason to conclude that it had to alter the district’s lines solely on the basis of race. And putting these two evidentiary items together helps, but it is simply too thin a reed to support the drastic decision to draw lines in this way. We have previously rejected proffers of evidence that were at least as strong as Texas’s here. For example, in Cooper, 581 U. S., at ___ (slip op., at 11), we analyzed North Carolina’s justification for deliberately moving “African-American voters” into a district to “ensure . . . the district’s racial composition” in the face of its expansion in size. North Carolina argued that its race-based decisions were necessary to comply with §2, but the State could point to “no meaningful legislative inquiry” into “whether a new, enlarged” district, “created without a focus on race, . . . could lead to §2 liability.” Id., at ___ (slip op., at 15). North Carolina pointed to two expert reports on “voting patterns throughout the State,” but we rejected that evidence as insufficient. Id., at ___, n. 5 (slip op., at 15, n. 5). Here, Texas has pointed to no actual “legislative inquiry” that would establish the need for its manipulation of the racial makeup of the district. By contrast, where we have accepted a State’s “good reasons” for using race in drawing district lines, the State made a strong showing of a pre-enactment analysis with justifiable conclusions. In Bethune-Hill, the State established that the primary mapdrawer “discussed the district with incumbents from other majority-minority districts[,] . . . considered turnout rates, the results of the recent contested primary and general elections,” and the district’s large prison population. 580 U. S., at ___ (slip op., at 15). The State established that it had performed a “functional analysis,” and acted to achieve an “informed bipartisan consensus.” Id., at ___ (slip op., at 14). Texas’s showing here is not equivalent. Perhaps Texas could have made a stronger showing, but it is the State’s burden to prove narrow tailoring, and it did not do so on the record before us. We hold that HD90 is an impermissible racial gerrymander. On remand, the District Court will have to consider what if any remedy is appropriate at this time. * * * Except with respect to one Texas House district, we hold that the court below erred in effectively enjoining the use of the districting maps adopted by the Legislature in 2013. We therefore reverse with respect to No. 17–586; reverse in part and affirm in part with respect to No. 17–626; and remand for proceedings consistent with this opinion. It is so ordered. Notes 1 There are several appendixes in these cases. We use “App.” to refer to the joint appendix filed at the merits stage. We use “C. J. S.” and “H. J. S.” to refer to the appendixes attached to Texas’s jurisdictional statements in No. 17–586 and No. 17–626, respectively. We use “C. J. S. Findings” and “H. J. S. Findings” to refer to appellees’ supplemental appendixes in No. 17–586 and No. 17–626. 2 See, e.g., Tex. Const., Art. III, §25 (Senate), §26 (House). 3 The court found: “[I]t is difficult to differentiate an intent to affect Democrats from an intent to affect minority voters. Making minorities worse off will likely make Democrats worse off, and vice versa.” C. J. S. Findings 467a (citation omitted). “This correlation is so strong that [an expert] assessed whether districts were minority opportunity districts by looking at Democratic results/wins (noting that in Texas, minority candidates of choice means Democrats).” Ibid. 4 See Shelby County v. Holder, 570 U. S. 529 (2013). 5 Judge Smith dissented, arguing that the majority had produced a “runaway plan” that “award[ed] judgment on the pleadings in favor of one side—a slam-dunk victory for the plaintiffs.” Perez v. Perry, 835 F. Supp. 2d 209, 218 (WD Tex. 2011). 6 The Texas court was given more leeway to make changes to districts challenged under §5 because it would have been inappropriate for that court to address the “merits of §5 challenges,” a task committed by statute to the District Court for the District of Columbia. Perez, 565 U. S., at 394. 7 Notice of Appeal in Texas v. United States, Civ. No. 11–cv–1303, Doc. No. 234 (D DC, Aug. 31, 2012). 8 We express no view on the correctness of this holding. 9 Judge Smith again dissented, on both mootness and the merits. On mootness, Judge Smith explained that, “[s]ix years later, we are still enveloped in litigation over plans that have never been used and will never be implemented.” C. J. S. 349a. On the merits, Judge Smith argued that the majority erroneously inferred a “complex, widespread conspiracy of scheming and plotting, by various legislators and staff, carefully designed to obscure the alleged race-based motive,” when the intent was in fact partisan. H. J. S. 294a; C. J. S. 351a. 10 In relevant part, §1253 applies to “an order granting . . . an interlocutory . . . injunction.” Section 1292(a)(1) applies to “[i]nterlocutory orders . . . granting . . . injunctions.” Although the similarity is obvious, the dissent perceives some unspecified substantive difference. 11 The dissent sees nothing strange about such a result because we held in Mitchell v. Donovan, 398 U. S. 427 (1970) (per curiam), that we lacked jurisdiction under §1253 to hear an appeal from a three-judge court order denying a declaratory judgment. The decision in Donovan was based on the plain language of §1253, which says nothing about orders granting or denying declaratory judgments. By contrast, §1253 gives us jurisdiction to hear appeals from orders granting or denying injunctions. The same goes for Rockefeller v. Catholic Medical Center of Brooklyn & Queens, Inc., 397 U. S. 820 (1970) (per curiam), also cited by the dissent. In that case, the District Court issued a declaratory judgment, not an injunction. Again, the text of §1253 says nothing about declaratory judgments. 12 The inquiry required by the practical effects test is no more difficult when the question is whether an injunction was effectively granted than it is when the question is whether an injunction was effectively denied. Lower courts have had “no problem concluding that [certain orders have] the practical effect of granting an injunction.” I. A. M. Nat. Pension Fund Benefit Plan A v. Cooper Industries, Inc., 789 F. 2d 21, 24 (CADC 1986); see also Andrew v. American Import Center, 110 A. 3d 626, 634 (D. C. 2015) (“[G]ranting a stay pending arbitration does have the ‘practical effect’ of enjoining the party opposing arbitration”). 13 Section 3(c) provides that if “the court finds that violations of the fourteenth or fifteenth amendment justifying equitable relief,” the court “shall retain jurisdiction for such period as it may deem appropriate and during such period no voting” practice shall go into effect unless first precleared by the court or the United States Attorney General. 52 U. S. C. §10302(c). 14 The other authority cited by the dissent is a footnote in Whitcomb v. Chavis, 403 U. S. 124 (1971), a case that came to us in an exceedingly complicated procedural posture. In Whitcomb, the District Court held in August 1969 that Indiana’s legislative districting scheme was unconstitutional, but the court made it clear that it would take no further action for two months. See Chavis v. Whitcomb, 305 F. Supp. 1364, 1392 (SD Ind. 1969). The Governor nevertheless appealed to this Court, but by the time we ruled, the Governor had taken another appeal from a later order, entered in December 1969, prohibiting the use of Indiana’s current plans and requiring the use of court-created plans in the 1970 elections. See 403 U. S., at 139; Juris. Statement in Whitcomb v. Chavis, O. T. 1970, No. 92, pp. 1–3. And to further complicate matters, by the time we reviewed the case, the Indiana Legislature had enacted new plans. Whitcomb, 403 U. S., at 140. This Court entertained the later appeal and reversed, but the Court dismissed the earlier—and by then, entirely superfluous—appeal, stating that, at the time when it was issued, “no judgment had been entered and no injunction had been granted or denied.” Id., at 138, n. 19. But that cursory conclusion has little relevance here, where the District Court’s orders were far more specific, immediate, and likely to demand compliance. 15 While we think it clear that the District Court effectively enjoined the use of these districts as currently configured for this year’s elections, even if the Court had not done so, that would not affect our jurisdiction to review the Court’s order with respect to all otherdistricts. 16 The dissent cites exactly two cases (Gunn and Whitcomb) decided during the past half-century in which a party attempted to take an appeal to this Court from a three-judge court order holding a state statute unconstitutional but declining to issue an injunction. 17 The dissent argues that we give “short shrift” to the irreparable harm question, post, at 16, but the inability to enforce its duly enacted plans clearly inflicts irreparable harm on the State, see, e.g., Maryland v. King, 567 U. S. 1301 (2012) (Roberts, C. J., in chambers). 18 The dissent attempts to rehabilitate this statement by focusing on the last part of this sentence, in which the District Court stated that the Legislature “intended [the] taint to be maintained but safe from remedy.” Post, at 34–35. In making this argument, the dissent, like the District Court, refuses to heed the presumption of legislative good faith and the allocation of the burden of proving intentional discrimination. We do not dispute that the District Court purportedly found that the 2013 Legislature acted with discriminatory intent. The problem is that, in making that finding, it relied overwhelmingly on what it perceived to be the 2013 Legislature’s duty to show that it had purged the bad intent of its predecessor. 19 The dissent is simply wrong in claiming over and over that we have not thoroughly examined the record. See post, at 19, 27, 28, 30, 31, 35, 39–40, 43, 47. The dissent seems to think that the repetition of these charges somehow makes them true. It does not. On the contrary, it betrays the substantive weakness of the dissent’s argument. 20 The dissent and the District Court attach much meaning to the attorney general’s use of the term “insulate” when he advised the Legislature to adopt the District Court’s plans to avoid further legal challenge. Setting aside that the word “insulate” is a common term used to describe minimizing legal concerns, the context of the letter makes clear that the attorney general was trying to make the point that adopting these plans was the best method of obtaining legal compliance, not the start of a grand conspiracy to trick the District Court. Indeed, if his plan was to dupe the District Court, shouting it to the world in a public letter was an odd way to go about it. 21 In any event, the Texas court was simply wrong that Texas believed its plans would be free from any legal challenge. 274 F. Supp. 3d 624, 651 (2017). Texas consistently acknowledged that effects claims would continue to be available and responded in detail to those arguments in both the District Court and this Court. See Brief for Appellants 64; Defendants’ Post-Trial Brief, Doc. No. 1526, p. 53. Moreover, Texas has not argued that intentional discrimination claims are unavailable; it has instead argued that intent must be assessed with respect to the 2013 Legislature, the Legislature that actually enacted the plans at issue. 22 The 2013 Legislature had no reason to believe that the District Court would spend four years examining moot plans before reversing its own previous decisions by imputing the intent of the 2011 Legislature to the 2013 Legislature. At the very least, the 2013 Legislature had good reason to believe that adopting the court-approved plans would lessen the time, expense, and complexity of further litigation (even if that belief turned out to be wrong). 23 Moreover, in criticizing the Legislature for moving too quickly, the dissent downplays the significant time and effort that went into consideration of the 2013 plans. Legislative committees held multiple field hearings in four cities, Tr. 1507 (July 14, 2017), and the legislative actors spent significant time considering the legislation, as well as accepting and rejecting amendments, see, e.g., Joint Exh. 17.3, p. S29; Joint Exh. 24.4, p. 21. 24 The dissent tries to minimize the relevance of this amendment by arguing that it turned HD90 into a racial gerrymander. See post, at 23, n. 12. But again this is misleading. The Legislature adopted changes to HD90 at the behest of minority groups, not out of a desire to discriminate. See Part IV–B, infra. That is, Darby was too solicitous of changes with respect to HD90. 25 In assessing the significance of the D. C. court’s evaluation of intent, it is important not to forget that the burden of proof in a preclearance proceeding was on the State. Texas v. United States, 887 F. Supp. 2d 133, 151 (DC 2012). Particularly where race and partisanship can so often be confused, see supra, at 3, and n. 3, the burden of proof may be crucial. 26 The District Court also purported to find a violation of the “one person, one vote” principle in Nueces County, 267 F. Supp. 3d 750, 783 (2017); H. J. S. 254a–255a, but that finding was in actuality a restatement of its racial discrimination finding. The population deviations from the ideal are quite small (0.34% in HD32 and 3.29% in HD34, id., at 254a), and the District Court relied solely on the “evidence of the use of race in drawing the lines in Nueces County” to find a one person, one vote violation. Id., at 255a; see also id., at 254a (“[T]he State intentionally discriminated against minority voters by overpopulating minority districts and underpopulating Anglo districts”). Even assuming that a court could find a one person, one vote violation on the basis of such a small deviation, cf. Brown v. Thomson, 462 U. S. 835, 842–843 (1983) (noting that deviations under 10% are generally insufficient to show invidious discrimination), the District Court erred in relying on its unsound finding regarding racial discrimination. Moreover, plaintiffs rejected any separate one person, one vote claims before the District Court, Tr. 22 (July 10, 2017), and they have not mentioned such a claim as a separate theory in their briefing in this Court. 27 The District Court’s belief that simple Latino majorities in Nueces County might be sufficient to create opportunity districts—and that Texas should have known as much—conflicts with other parts of its decision. With respect to numerous other districts, the District Court chided Texas for focusing on bare numbers and not considering real opportunity to elect. See, e.g., C. J. S. 134a (“[T]he court rejects [the] bright-line rule that any HCVAP-majority district is by definition a Latino opportunity district” because it “may still lack real electoral opportunity” (internal quotation marks omitted)); H. J. S. 121a (Texas “increase[d the Latino population] while simultaneously ensuring that election success rates remained minimally improved”).

585.US.2017_16-1220

Petitioners, U. S.-based purchasers of vitamin C (U. S. purchasers), filed a class-action suit, alleging that four Chinese corporations that manufacture and export the nutrient (Chinese sellers), including the two respondents here, had agreed to fix the price and quantity of vitamin C exported to the United States, in violation of §1 of the Sherman Act. The Chinese sellers moved to dismiss the complaint on the ground that Chinese law required them to fix the price and quantity of vitamin C exports, thus shielding them from liability under U. S. antitrust law. The Ministry of Commerce of the People’s Republic of China (Ministry) filed an amicus brief in support of the motion, explaining that it is the administrative authority authorized to regulate foreign trade, and stating that the alleged conspiracy in restraint of trade was actually a pricing regime mandated by the Chinese Government. The U. S. purchasers countered that the Ministry had identified no law or regulation ordering the Chinese sellers’ price agreement, highlighted a publication announcing that the Chinese sellers had agreed to control the quantity and rate of exports without government intervention, and presented supporting expert testimony. The District Court denied the Chinese sellers’ motion in relevant part, concluding that it did not regard the Ministry’s statements as “conclusive,” particularly in light of the U. S. purchasers’ evidence. When the Chinese sellers subsequently moved for summary judgment, the Ministry submitted another statement, reiterating its stance, and the U. S. purchasers pointed to China’s statement to the World Trade Organization that it ended its export administration of vitamin C in 2002. The court denied this motion as well. The case was then tried to a jury, which returned a verdict for the U. S. purchasers. The Second Circuit reversed, holding that the District Court erred by denying the Chinese sellers’ motion to dismiss the complaint. When a foreign government whose law is in contention submits an official statement on the meaning and interpretation of its domestic law, the court concluded, federal courts are “bound to defer” to the foreign government’s construction of its own law, whenever that construction is “reasonable.” Inspecting only the Ministry’s brief and the sources cited therein, the court found the Ministry’s account of Chinese law “reasonable.” Held: A federal court determining foreign law under Federal Rule of Civil Procedure 44.1 should accord respectful consideration to a foreign government’s submission, but the court is not bound to accord conclusive effect to the foreign government’s statements. Rule 44.1 fundamentally changed the mode of determining foreign law in federal courts. Before adoption of the rule in 1966, a foreign nation’s laws had to be “proved as facts.” Talbot v. Seeman, 1 Cranch 1, 38. Rule 44.1, in contrast, specifies that a court’s determination of foreign law “must be treated as a ruling on a question of law.” And in ascertaining foreign law, courts are not limited to materials submitted by the parties, but “may consider any relevant material or source.” Appellate review, as is true of domestic law determinations, is de novo. The purpose of these changes was to align, to the extent possible, the process for determining alien law and the process for determining domestic law. Neither Rule 44.1 nor any other rule or statute addresses the weight a federal court determining foreign law should give to the views presented by a foreign government. In the spirit of “international comity,” Société Nationale Industrielle Aérospatiale v. United States Dist. Court for Southern Dist. of Iowa, 482 U. S. 522, 543, and n. 27, a federal court should carefully consider a foreign state’s views about the meaning of its own laws. The appropriate weight in each case, however, will depend upon the circumstances; a federal court is neither bound to adopt the foreign government’s characterization nor required to ignore other relevant materials. No single formula or rule will fit all cases, but relevant considerations include the statement’s clarity, thoroughness, and support; its context and purpose; the transparency of the foreign legal system; the role and authority of the entity or official offering the statement; and the statement’s consistency with the foreign government’s past positions. Judged in this light, the Second Circuit’s unyielding rule is inconsistent with Rule 44.1 and, tellingly, with this Court’s treatment of analogous submissions from States of the United States. If the relevant state law is established by a decision of “the State’s highest court,” that decision is “binding on the federal courts,” Wainwright v. Goode, 464 U. S. 78, 84, but views of the State’s attorney general, while attracting “respectful consideration,” do not garner controlling weight, Arizonans for Official English v. Arizona, 520 U. S. 43, 76–77, n. 30. Furthermore, because the Second Circuit riveted its attention on the Ministry’s submission, it did not address evidence submitted by the U. S. purchasers. The court also misperceived the pre-Rule 44.1 decision of United States v. Pink, 315 U. S. 203. Under the particular circumstances of that case, this Court found conclusive a declaration from the government of the Russian Socialist Federal Soviet Republic on the extraterritorial effect of a decree nationalizing assets: The declaration was obtained by the United States through official “diplomatic channels,” id., at 218; there was no indication that the declaration was inconsistent with the Russian Government’s past statements; and the declaration was consistent with expert evidence in point. The Second Circuit expressed concern about reciprocity, but the United States has not historically argued that foreign courts are bound to accept its characterizations or precluded from considering other relevant sources. International practice is also inconsistent with the Second Circuit’s rigid rule. Pp. 7–12. 837 F. 3d 175, vacated and remanded. Ginsburg, J., delivered the opinion for a unanimous Court.

When foreign law is relevant to a case instituted in a federal court, and the foreign government whose law is in contention submits an official statement on the meaning and interpretation of its domestic law, may the federal court look beyond that official statement? The Court of Appeals for the Second Circuit answered generally “no,” ruling that federal courts are “bound to defer” to a foreign government’s construction of its own law, whenever that construction is “reasonable.” In re Vitamin C Antitrust Litigation, 837 F. 3d 175, 189 (2016). We hold otherwise. A federal court should accord respectful consideration to a foreign government’s submission, but is not bound to accord conclusive effect to the foreign government’s statements. Instead, Federal Rule of Civil Procedure 44.1 instructs that, in determining foreign law, “the court may consider any relevant material or source . . . whether or not submitted by a party.” As “[t]he court’s determination must be treated as a ruling on a question of law,” Fed. Rule Civ. Proc. 44.1, the court “may engage in its own research and consider any relevant material thus found,” Advisory Committee’s 1966 Note on Fed. Rule Civ. Proc. 44.1, 28 U. S. C. App., p. 892 (hereinafter Advisory Committee’s Note). Because the Second Circuit ordered dismissal of this case on the ground that the foreign government’s statements could not be gainsaid, we vacate that court’s judgment and remand the case for further consideration. I Petitioners, U. S.-based purchasers of vitamin C (hereinafter U. S. purchasers), filed a class-action suit against four Chinese corporations that manufacture and export the nutrient (hereinafter Chinese sellers). The U. S. purchasers alleged that the Chinese sellers, two of whom are respondents here, had agreed to fix the price and quantity of vitamin C exported to the United States from China, in violation of §1 of the Sherman Act, 15 U. S. C. §1. More particularly, the U. S. purchasers stated that the Chinese sellers had formed a cartel “facilitated by the efforts of their trade association,” the Chamber of Commerce of Medicines and Health Products Importers and Exporters (Chamber). Complaint in No. 1:05–CV–453, Docket No. 1, ¶43. The Judicial Panel on Multidistrict Litigation consolidated the instant case and related suits for pretrial proceedings in the United States District Court for the Eastern District of New York. The Chinese sellers moved to dismiss the U. S. purchasers’ complaint on the ground that Chinese law required them to fix the price and quantity of vitamin C exports. Therefore, the Chinese sellers urged, they are shielded from liability under U. S. antitrust law by the act of state doctrine, the foreign sovereign compulsion doctrine, and principles of international comity. The Ministry of Commerce of the People’s Republic of China (Ministry) filed a brief as amicus curiae in support of the Chinese sellers’ motion. The Ministry’s brief stated that the Ministry is “the highest administrative authority in China authorized to regulate foreign trade,” App. to Pet. for Cert. 190a; that the Chamber is “an entity under the Ministry’s direct and active supervision” and is authorized to regulate vitamin C exports, id., at 196a; and that the conspiracy in restraint of trade alleged by the U. S. purchasers was in fact “a regulatory pricing regime mandated by the government of China,” id., at 197a.[1] In response, the U. S. purchasers disputed that Chinese law required the Chinese sellers to engage in price fixing. Among other things, the U. S. purchasers noted that the Ministry had not identified any written law or regulation expressly ordering the Chinese sellers’ price agreement.[2] They also highlighted a Chamber announcement that the manufacturers “were able to reach a self-regulated agreement . . . whereby they would voluntarily control the quantity and pace of exports . . . without any government intervention.” App. 109. In addition, the U. S. purchasers presented expert testimony that the Chinese Government’s authorization of a Vitamin C Subcommittee within the Chamber did not necessarily mean that the subcommittee’s price fixing was mandated by law. The District Court denied the Chinese sellers’ motion to dismiss the complaint in relevant part. In re Vitamin C Antitrust Litigation, 584 F. Supp. 2d 546, 559 (EDNY 2008). That court acknowledged that the Ministry’s amicus brief was “entitled to substantial deference.” Id., at 557. The court, however, did not regard the Ministry’s statements as “conclusive,” emphasizing particularly that the U. S. purchasers had submitted evidence suggesting that the price fixing was voluntary. Ibid. The record, the District Court determined, was “too ambiguous to foreclose further inquiry into the voluntariness of [the Chinese sellers’] actions.” Id., at 559. After further discovery, focused on whether Chinese law compelled the Chinese sellers to enter into a price-fixing agreement, the Chinese sellers moved for summary judgment. See In re Vitamin C Antitrust Litigation, 810 F. Supp. 2d 522, 525–526 (EDNY 2011). The Ministry submitted an additional statement, reiterating that “the Ministry specifically charged the Chamber . . . with the authority and responsibility . . . for regulating, through consultation, the price of vitamin C manufactured for export.” App. 133. The Chinese sellers tendered expert testimony in accord with the Ministry’s account, which stressed that the Ministry’s “interpretation of its own regulations and policies carries decisive weight under Chinese law.” Id., at 142. The U. S. purchasers, in response, cited further materials supporting their opposing view, including China’s statement to the World Trade Organization (WTO) that it “gave up export administration of . . . vitamin C” in 2002. 810 F. Supp. 2d, at 532 (internal quotation marks omitted). Denying the Chinese sellers’ motion for summary judgment, the District Court held that Chinese law did not require the sellers to fix the price or quantity of vitamin C exports. Id., at 525. The case was then tried to a jury, which returned a verdict for the U. S. purchasers. The jury found that the Chinese sellers had agreed to fix the prices and quantities of vitamin C exports, see App. to Pet. for Cert. 276a–279a, and further found that the Chinese sellers were not “actually compelled” by China to enter into those agreements, id., at 278a. In accord with the jury’s verdict, the District Court entered judgment for the U. S. purchasers, awarding some $147 million in treble damages and enjoining the Chinese sellers from further violations of the Sherman Act. The Court of Appeals for the Second Circuit reversed, holding that the District Court erred in denying the Chinese sellers’ motion to dismiss the complaint. In re Vitamin C Antitrust Litigation, 837 F. 3d 175, 178, 195–196 (2016). The Court of Appeals determined that the propri- ety of dismissal hinged on whether the Chinese sellers could adhere to both Chinese law and U. S. antitrust law. See id., at 186. That question, in turn, depended on “the amount of deference” owed to the Ministry’s characterization of Chinese law. Ibid. Cognizant of “competing authority” on this question, ibid., the Court of Appeals settled on a highly deferential rule: “[W]hen a foreign government, acting through counsel or otherwise, directly participates in U. S. court proceedings by providing a [statement] regarding the construction and effect of [the foreign government’s] laws and regulations, which is reasonable under the circumstances presented, a U. S. court is bound to defer to those statements,” id., at 189. The appeals court “note[d] that[,] if the Chinese Government had not appeared in this litigation, the [D]istrict [C]ourt’s careful and thorough treatment of the evidence before it in analyzing what Chinese law required at both the motion to dismiss and summary judgment stages would have been entirely appropriate.” Id., at 191, n. 10. Applying its highly deferential rule, the Court of Appeals concluded that the Ministry’s account of Chinese law was “reasonable.” In so concluding, the Court of Appeals inspected only the Ministry’s brief and sources cited therein. Id., at 189–190. Because it thought that “a U. S. court [must] not embark on a challenge to a foreign government’s official representation,” id., at 189, the Court of Appeals disregarded the submissions made by the U. S. purchasers casting doubt on the Ministry’s account of Chinese law, id., at 189–190. Based solely on the Ministry’s statements, the Court of Appeals held that “Chinese law required [the Chinese sellers] to engage in activities in China that constituted antitrust violations here in the United States.” Ibid. We granted certiorari to resolve a Circuit conflict over this question: Is a federal court determining foreign law under Rule 44.1 required to treat as conclusive a submission from the foreign government describing its own law? 583 U. S. ___ (2018).[3] II At common law, the content of foreign law relevant to a dispute was treated “as a question of fact.” Miller, Federal Rule 44.1 and the “Fact” Approach to Determining Foreign Law: Death Knell for a Die-Hard Doctrine, 65 Mich. L. Rev. 613, 617–619 (1967) (Miller). In 1801, this Court endorsed the common-law rule, instructing that “the laws of a foreign nation” must be “proved as facts.” Talbot v. Seeman, 1 Cranch 1, 38 (1801); see, e.g., Church v. Hubbart, 2 Cranch 187, 236 (1804) (“Foreign laws are well understood to be facts.”). Ranking questions of foreign law as questions of fact, however, “had a number of undesir- able practical consequences.” 9A C. Wright & A. Miller, Federal Practice and Procedure §2441, p. 324 (3d ed. 2008) (Wright & Miller). Foreign law “had to be raised in the pleadings” and proved “in accordance with the rules of evidence.” Ibid. Appellate review was deferential and limited to the record made in the trial court. Ibid.; see also Miller 623. Federal Rule of Civil Procedure 44.1, adopted in 1966, fundamentally changed the mode of determining foreign law in federal courts. The Rule specifies that a court’s determination of foreign law “must be treated as a ruling on a question of law,” rather than as a finding of fact.[4] Correspondingly, in ascertaining foreign law, courts are not limited to materials submitted by the parties; instead, they “may consider any relevant material or source . . . , whether or not . . . admissible under the Federal Rules of Evidence.” Ibid. Appellate review, as is true of domestic law determinations, is de novo. Advisory Committee’s Note, at 892. Rule 44.1 frees courts “to reexamine and amplify material . . . presented by counsel in partisan fashion or in insufficient detail.” Ibid. The “obvious” purpose of the changes Rule 44.1 ordered was “to make the process of determining alien law identical with the method of ascertaining domestic law to the extent that it is possible to do so.” Wright & Miller §2444, at 338–342. Federal courts deciding questions of foreign law under Rule 44.1 are sometimes provided with the views of the relevant foreign government, as they were in this case through the amicus brief of the Ministry. See supra, at 2–3. As the Court of Appeals correctly observed, Rule 44.1 does not address the weight a federal court determining foreign law should give to the views presented by the foreign government. See 837 F. 3d, at 187. Nor does any other rule or statute. In the spirit of “international comity,” Société Nationale Industrielle Aérospatiale v. United States Dist. Court for Southern Dist. of Iowa, 482 U. S. 522, 543, and n. 27 (1987), a federal court should carefully consider a foreign state’s views about the meaning of its own laws. See United States v. McNab, 331 F. 3d 1228, 1241 (CA11 2003); cf. Bodum USA, Inc. v. La Cafetière, Inc., 621 F. 3d 624, 638–639 (CA7 2010) (Wood, J., concurring). But the appropriate weight in each case will depend upon the circumstances; a federal court is neither bound to adopt the foreign government’s characterization nor required to ignore other relevant materials. When a foreign government makes conflicting statements, see supra, at 5, or, as here, offers an account in the context of litigation, there may be cause for caution in evaluating the foreign government’s submission. Given the world’s many and diverse legal systems, and the range of circumstances in which a foreign government’s views may be presented, no single formula or rule will fit all cases in which a foreign government describes its own law. Relevant considerations include the statement’s clarity, thoroughness, and support; its context and purpose; the transparency of the foreign legal system; the role and authority of the entity or official offering the statement; and the statement’s consistency with the foreign government’s past positions. Judged in this light, the Court of Appeals erred in deeming the Ministry’s submission binding, so long as facially reasonable. That unyielding rule is inconsistent with Rule 44.1 (determination of an issue of foreign law “must be treated as a ruling on a question of law”; court may consider “any relevant material or source”) and, tellingly, with this Court’s treatment of analogous submissions from States of the United States. If the relevant state law is established by a decision of “the State’s highest court,” that decision is “binding on the federal courts.” Wainwright v. Goode, 464 U. S. 78, 84 (1983) (per curiam); see Mullaney v. Wilbur, 421 U. S. 684, 691 (1975). But views of the State’s attorney general, while attracting “respectful consideration,” do not garner controlling weight. Arizonans for Official English v. Arizona, 520 U. S. 43, 76–77, n. 30 (1997); see, e.g., Virginia v. American Booksellers Assn., Inc., 484 U. S. 383, 393–396 (1988). Furthermore, because the Court of Appeals riveted its attention on the Ministry’s submission, it did not address other evidence, including, for example, China’s statement to the WTO that China had “g[i]ve[n] up export administration . . . of vitamin C” at the end of 2001. 810 F. Supp. 2d, at 532 (internal quotation marks omitted).[5] The Court of Appeals also misperceived this Court’s decision in United States v. Pink, 315 U. S. 203 (1942). See 837 F. 3d, at 186–187, 189. Pink, properly comprehended, is not compelling authority for the attribution of controlling weight to the Ministry’s brief. We note, first, that Pink was a pre-Rule 44.1 decision. Second, Pink arose in unusual circumstances. Pink was an action brought by the United States to recover assets of the U. S. branch of a Russian insurance company that had been nationalized in 1918, after the Russian revolution. 315 U. S., at 210–211. In 1933, the Soviet Government assigned the nationalized assets located in this country to the United States. Id., at 211–212. The disposition of the case turned on the extraterritorial effect of the nationalization decree—specifically, whether the decree reached assets of the Russian insurance company located in the United States, or was instead limited to property in Russia. Id., at 213–215, 217. To support the position that the decree reached all of the company’s assets, the United States obtained an “official declaration of the Commissariat for Justice” of the Russian Socialist Federal Soviet Republic. Id., at 218. The declaration certified that the nationalization decree reached “the funds and property of former insurance companies . . . irrespective of whether [they were] situated within the territorial limits of [Russia] or abroad.” Id., at 220 (internal quotation marks omitted). This Court determined that “the evidence supported [a] finding” that “the Commissariat for Justice ha[d] power to interpret existing Russian law.” Ibid. “That being true,” the Court concluded, the “official declaration [wa]s conclusive so far as the intended extraterritorial effect of the Russian decree [wa]s concerned.” Ibid. This Court’s treatment of the Commissariat’s submission as conclusive rested on a document obtained by the United States, through official “diplomatic channels.” Id., at 218. There was no indication that the declaration was inconsistent with the Soviet Union’s past statements. Indeed, the Court emphasized that the declaration was consistent with expert evidence in point. See ibid. That the Commissariat’s declaration was deemed “conclusive” in the circumstances Pink presented scarcely suggests that all submissions by a foreign government are entitled to the same weight. The Court of Appeals also reasoned that a foreign government’s characterization of its own laws should be afforded “the same respect and treatment that we would expect our government to receive in comparable matters.” 837 F. 3d, at 189. The concern for reciprocity is sound, but it does not warrant the Court of Appeals’ judgment. Indeed, the United States, historically, has not argued that foreign courts are bound to accept its characterizations or precluded from considering other relevant sources.[6] The understanding that a government’s expressed view of its own law is ordinarily entitled to substantial but not conclusive weight is also consistent with two international treaties that establish formal mechanisms by which one government may obtain from another an official statement characterizing its laws. Those treaties specify that “[t]he information given in the reply shall not bind the judicial authority from which the request emanated.” European Convention on Information on Foreign Law, Art. 8, June 7, 1968, 720 U. N. T. S. 154; see Inter-American Convention on Proof of and Information on Foreign Law, Art. 6, May 8, 1979, O. A. S. T. S. 1439 U. N. T. S. 111 (similar). Al- though the United States is not a party to those treaties, they reflect an international practice inconsistent with the Court of Appeals’ “binding, if reasonable” resolution. * * * Because the Court of Appeals concluded that the District Court was bound to defer to the Ministry’s brief, the court did not consider the shortcomings the District Court identified in the Ministry’s position or other aspects of “the [D]istrict [C]ourt’s careful and thorough treatment of the evidence before it.” 837 F. 3d, at 191, n. 10. The correct interpretation of Chinese law is not before this Court, and we take no position on it. But the materials identified by the District Court were at least relevant to the weight the Ministry’s submissions should receive and to the question whether Chinese law required the Chinese sellers’ conduct. We therefore vacate the judgment of the Court of Appeals and remand the case for renewed consideration consistent with this opinion. It is so ordered. Notes 1 The Ministry told the District Court: For much of the 20th century, China allowed only state-owned entities to export products. App. to Pet. for Cert. 198a. When China started to allow private enterprises to obtain export licenses, the Ministry established the Chamber to regulate exports under the Ministry’s authority and direction. Ibid. In 1997, the Ministry authorized the establishment of the Chamber’s Vitamin C Subcommittee. Id., at 202a. That year, the Ministry promulgated a regulation authorizing and requiring the subcommittee to limit the production of vitamin C for export and to set export prices. Id., at 202a–204a. Under the regulation delineating this “Export Licensing System,” the Ministry issued export licenses only to manufacturers whose export volume and price complied with the output quota and price coordinated by the Vitamin C Subcommittee. Id., at 204a. In 2002, the Ministry replaced the Export Licensing System with a “Verification and Chop System.” Id., at 208a. As set forth in a 2002 Ministry Notice, the Chamber itself—instead of the Ministry—would inspect each export contract and certify its compliance with the coordinated quotas and price by affixing a special seal, known as a “chop.” Id., at 208a–209a. China’s Customs would allow export only if the exporter presented its contract bearing the Chamber’s “chop.” Id., at 209a. According to the Ministry, it was implicit in this arrangement that vitamin C exporters would remain under an obligation to fix prices and volumes. Id., at 208a. The effect of China’s regime on the Chinese sellers’ liability under the Sherman Act, we note, is not an issue before the Court today. 2 The complaint, the U. S. purchasers emphasized, was directed only at conduct occurring after December 2001. As they understood the Ministry’s 2002 Notice, see supra, at 3, n. 1, vitamin C exporters could have lawfully opted out of price fixing. Beyond that, the Vitamin C Subcommittee had replaced its 1997 Charter with a new 2002 Charter, App. 182–197, which eliminated the 1997 Charter’s requirement that subcommittee members “[s]trictly execute” the “coordinated price” set by the Chamber, compare id., at 85, with id., at 185, and granted members an express “[r]igh[t]” to “freely resign from the Subcommittee,” id., at 186. 3 Compare In re Vitamin C Antitrust Litigation, 837 F.3d 175 (CA2 2016) (case below), with In re Oil Spill by Amoco Cadiz, 954 F.2d 1279, 1311–1313 (CA7 1992) (adopting French Government’s interpretation of French law, but only after considering all of the circumstances, including the French Government’s statements in other contexts); United States v. McNab, 331 F.3d 1228, 1239–1242 (CA11 2003) (noting Honduran Government’s shift in position on the question of Honduran law and determining that the original position stated the proper interpretation); McKesson HBOC, Inc. v. Islamic Republic of Iran, 271 F.3d 1101, 1108–1109 (CADC 2001), vacated in part on other grounds, 320 F.3d 280 (CADC 2003) (declining to adopt the view of Iranian law advanced by Iranian Government because it was not supported by the affidavits submitted by Iran’s experts). 4 Federal Rule of Criminal Procedure 26.1 establishes “substantially the same” rule for criminal cases. Advisory Committee’s 1966 Note on Fed. Rule Crim. Proc. 26.1, 18 U. S. C. App., p. 709. 5 The Court of Appeals additionally mischaracterized the Ministry’s brief as a “sworn evidentiary proffer.” 837 F. 3d, at 189. In so describing the Ministry’s submission, the Court of Appeals overlooked that a court’s resolution of an issue of foreign law “must be treated as a ruling on a question of law.” Fed. Rule Civ. Proc. 44.1. The Ministry’s brief, while a probative source for resolving the legal question at hand, was not an attestation to facts. 6 The Chinese sellers assert, see Supp. Brief for Respondents 7–8, that the United States sought a greater degree of deference in a 2002 submission to a World Trade Organization panel. In fact, the submission acknowledged that “the Panel is not bound to accept the interpretation [of U. S. law] presented by the United States.” Brief for United States as Amicus Curiae 29, n. 6 (quoting Second Written Submission of the United States of America, United States—Section 129(c)(1) of the Uruguay Round Agreements Act, WT/DS221 ¶11 (Mar. 8, 2002)).

583.US.2017_16-460

Federal district courts may exercise supplemental jurisdiction over state claims not otherwise within their adjudicatory authority if those claims are “part of the same case or controversy” as the federal claims the plaintiff asserts. 28 U. S. C. §1367(a). When a district court dismisses all claims independently qualifying for the exercise of federal jurisdiction, it ordinarily also dismisses all related state claims. See §1367(c)(3). Section 1367(d) provides that the “period of limitations for” refiling in state court a state claim so dismissed “shall be tolled while the claim is pending [in federal court] and for a period of 30 days after it is dismissed unless State law provides for a longer tolling period.” When petitioner Artis filed a federal-court suit against respondent District of Columbia (District), alleging a federal employment-discrimination claim and three allied claims under D. C. law, nearly two years remained on the applicable statute of limitations for the D. C.-law violations. Two and a half years later, the Federal District Court ruled against Artis on her sole federal claim and dismissed the D. C.-law claims under §1367(c). Fifty-nine days after the dismissal, Artis refiled her state-law claims in the D. C. Superior Court, but that court dismissed them as time barred. The D. C. Court of Appeals affirmed, holding that §1367(d) accorded Artis only a 30-day grace period to refile in state court and rejecting her argument that the word “tolled” in §1367(d) means that the limitations period is suspended during the pendency of the federal suit. Held: 1. Section 1367(d)’s instruction to “toll” a state limitations period means to hold it in abeyance, i.e., to stop the clock. Pp. 7–16. (a) Statutes that shelter from time bars claims earlier commenced in another forum generally employ one of two means. First, the period of limitations may be “tolled,” i.e., suspended, while the claim is pending elsewhere; the time clock starts running again when the tolling period ends, picking up where it left off. A legislature may instead elect simply to provide a grace period, permitting the statute of limitations to run while the claim is pending in another forum and averting the risk of a time bar by according the plaintiff a fixed period in which to refile. The District has identified no federal statute in which a grace-period meaning has been ascribed to the word “tolled” or any word similarly rooted. And the one case in which this Court used tolling language to describe a grace period, see Hardin v. Straub, 490 U. S. 536 , is a feather on the scale against the weight of decisions in which “tolling” a statute of limitations signals stopping the clock. Pp. 7–11. (b) Considering first the ordinary meaning of the statutory language, §1367(d) is phrased as a tolling provision. It suspends the statute of limitations both while the claim is pending in federal court and for 30 days postdismissal. Artis’ interpretation is a natural fit with this language. The District, in contrast, reads “tolled” to mean to remove, temporarily, the bar that would ordinarily accompany the expiration of the limitations period. But the District offers no reason to home in only on the word “tolled” itself and ignore information about the verb’s ordinary meaning gained from its grammatical object, “period of limitations.” That object sheds light on what it means to “be tolled.” The District’s reading also tenders a strained interpretation of the phrase “period of limitations”; makes the first portion of the tolling period, the duration of the claim’s pendency in federal court, superfluous; and could yield an absurdity, permitting a plaintiff to refile in state court even if the limitations period on her claim had expired before she filed in federal court. Pp. 11–13. (c) The D. C. Court of Appeals erred in concluding that Congress adopted an American Law Institute (ALI) recommendation to allow refiling in state court only for 30 days after a dismissal. The ALI provision, like §1367(d), established a 30-day federal floor on the time allowed for refiling, but it did not provide for tolling “while the [state] claim is pending” in federal court. Pp. 13–14. (d) The 30-day provision casts no large shadow on Artis’ stop-the-clock interpretation. The provision accounts for cases in which a plaintiff commenced a federal action close to the expiration date of the relevant state statute of limitations, by giving such a plaintiff breathing space to refile in state court. Adding a brief span of days to the tolling period is not unusual in stop-the-clock statutes. See, e.g., 46 U. S. C. §53911. Section 1367(d)’s proviso “unless State law provides for a longer tolling period” could similarly aid a plaintiff who filed in federal court just short of the expiration of the state limitations period. Pp. 14–16. 2. The stop-the-clock interpretation of §1367(d) does not present a serious constitutional problem. In Jinks v. Richland County, 538 U. S. 456 , the Court rejected an argument that §1367(d) impermissibly exceeds Congress’ authority under the Necessary and Proper Clause. Id., at 464–465. The District contends that a stop-the-clock prescription serves “no federal purpose” that could not be served by a grace-period prescription. But both devices are standard, off-the-shelf means of accounting for the fact that a claim was timely pressed in another forum. Requiring Congress to choose one over the other would impose a tighter constraint on Congress’ discretion than this Court has countenanced. A concern that a stop-the-clock prescription entails a greater imposition on the States than a grace-period prescription may also be more theoretical than real. Finally, a stop-the-clock rule like §1367(d) is suited to the primary purposes of limitations statutes: “ ‘preventing surprises’ ” to defendants and “ ‘barring a plaintiff who has slept on his rights.’ ” American Pipe & Constr. Co. v. Utah, 414 U. S. 538 . Pp. 16–19. 135 A. 3d 334, reversed and remanded. Ginsburg, J., delivered the opinion of the Court, in which Roberts, C. J., and Breyer, Sotomayor, and Kagan, JJ., joined. Gorsuch, J., filed a dissenting opinion, in which Kennedy, Thomas, and Alito, JJ., joined.

The Supplemental Jurisdiction statute, 28 U. S. C. §1367, enables federal district courts to entertain claims not otherwise within their adjudicatory authority when those claims “are so related to claims . . . within [federal-court competence] that they form part of the same case or controversy.” §1367(a). Included within this supplemental jurisdiction are state claims brought along with federal claims arising from the same episode. When district courts dismiss all claims independently qualifying for the exercise of federal jurisdiction, they ordinarily dismiss as well all related state claims. See §1367(c)(3). A district court may also dismiss the related state claims if there is a good reason to decline jurisdiction. See §1367(c)(1), (2), and (4). This case concerns the time within which state claims so dismissed may be refiled in state court. Section 1367(d), addressing that issue, provides: “The period of limitations for any [state] claim [ joined with a claim within federal-court competence] shall be tolled while the claim is pending [in federal court] and for a period of 30 days after it is dismissed unless State law provides for a longer tolling period.” The question presented: Does the word “tolled,” as used in §1367(d), mean the state limitations period is suspended during the pendency of the federal suit; or does “tolled” mean that, although the state limitations period continues to run, a plaintiff is accorded a grace period of 30 days to refile in state court post dismissal of the federal case? Petitioner urges the first, or stop-the-clock, reading. Respondent urges, and the District of Columbia Court of Appeals adopted, the second, or grace-period, reading. In the case before us, plaintiff-petitioner Stephanie C. Artis refiled her state-law claims in state court 59 days after dismissal of her federal suit.[1] Reading §1367(d) as a grace-period prescription, her complaint would be time barred. Reading §1367(d) as stopping the limitations clock during the pendency of the federal-court suit, her complaint would be timely. We hold that §1367(d)’s instruction to “toll” a state limitations period means to hold it in abeyance, i.e., to stop the clock. Because the D. C. Court of Appeals held that §1367(d) did not stop the D. C. Code’s limitations clock, but merely provided a 30-day grace period for refiling in D. C. Superior Court, we reverse the D. C. Court of Appeals’ judgment. I A Section 1367, which Congress added to Title 28 as part of the Judicial Improvements Act of 1990, 104Stat. 5089, codifies the court-developed pendent and ancillary jurisdiction doctrines under the label “supplemental jurisdiction.” See Exxon Mobil Corp. v. Allapattah Services, Inc., 545 U. S. 546 –558 (2005) (describing the development of pendent and ancillary jurisdiction doctrines and subsequent enactment of §1367); id., at 579–584 (Ginsburg, J., dissenting) (same). The House Report accompanying the Act explains that Congress sought to clarify the scope of federal courts’ authority to hear claims within their supplemental jurisdiction, appreciating that “[s]upplemental jurisdiction has enabled federal courts and litigants to . . . deal economically—in single rather than multiple litigation—with related matters.” H. R. Rep. No. 101–734, p. 28 (1990) (H. R. Rep.). Section 1367(a) provides, in relevant part, that a district court with original jurisdiction over a claim “shall have supplemental jurisdiction over all other claims . . . form[ing] part of the same case or controversy.” “[N]ot every claim within the same ‘case or controversy’ as the claim within the federal courts’ original jurisdiction will be decided by the federal court.” Jinks v. Richland County, 538 U. S. 456, 459 (2003) . Section 1367(c) states: “The district courts may decline to exercise supplemental jurisdiction over a claim under subsection (a) if— “(1) the claim raises a novel or complex issue of State law, “(2) the claim substantially predominates over the claim or claims over which the district court has original jurisdiction, “(3) the district court has dismissed all claims over which it has original jurisdiction, or “(4) in exceptional circumstances, there are other compelling reasons for declining jurisdiction.” If a district court declines to exercise jurisdiction over a claim asserted under §1367(a) and the plaintiff wishes to continue pursuing it, she must refile the claim in state court. If the state court would hold the claim time barred, however, then, absent a curative provision, the district court’s dismissal of the state-law claim without prejudice would be tantamount to a dismissal with prejudice. See, e.g., Carnegie-Mellon Univ. v. Cohill, 484 U. S. 343, 352 (1988) (under the doctrine of pendent jurisdiction, if the statute of limitations on state-law claims expires before the federal court “relinquish[es] jurisdiction[,] . . . a dismissal will foreclose the plaintiff from litigating his claims”). To prevent that result, §1367(d) supplies “a tolling rule that must be applied by state courts.” Jinks, 538 U. S., at 459. Section 1367(d) provides: “The period of limitations for any claim asserted under subsection (a), and for any other claim in the same action that is voluntarily dismissed at the same time as or after the dismissal of the claim under subsection (a), shall be tolled while the claim is pending and for a period of 30 days after it is dismissed unless State law provides for a longer tolling period.” This case requires us to determine how §1367(d)’s tolling rule operates. B Petitioner Artis worked as a health inspector for respondent, the District of Columbia (the “District”). In November 2010, Artis was told she would lose her job. Thirteen months later, Artis sued the District in the United States District Court for the District of Columbia, alleging that she had suffered employment discrimination in violation of Title VII of the Civil Rights Act of 1964, 78Stat. 253, as amended, 42 U. S. C. §2000e et seq. She also asserted three allied claims under D. C. law: retaliation in violation of the District of Columbia Whistleblower Act, D. C. Code §1–615.54 (2001); termination in violation of the District of Columbia False Claims Act, §2–381.04; and wrongful termination against public policy, a common-law claim. Artis alleged that she had been subjected to gender discrimination by her supervisor, and thereafter encountered retaliation for reporting the supervisor’s unlawful activities. See Artis v. District of Columbia, 51 F. Supp. 3d 135, 137 (2014). On June 27, 2014, the District Court granted the District’s motion for summary judgment on the Title VII claim. Having dismissed Artis’ sole federal claim, the District Court, pursuant to §1367(c)(3), declined to exercise supplemental jurisdiction over her remaining state-law claims. “Artis will not be prejudiced,” the court noted, “because 28 U. S. C. §1367(d) provides for a tolling of the statute of limitations during the period the case was here and for at least 30 days thereafter.” Id., at 142. Fifty-nine days after the dismissal of her federal action, Artis refiled her state-law claims in the D. C. Superior Court, the appropriate local court. The Superior Court granted the District’s motion to dismiss, holding that Artis’ claims were time barred, because they were filed 29 days too late. See App. to Pet. for Cert. 14a. When Artis first asserted her state-law claims in the District Court, nearly two years remained on the applicable three-year statute of limitations.[2] But two and a half years passed before the federal court relinquished jurisdiction. Unless §1367(d) paused the limitations clock during that time, Artis would have had only 30 days to refile. The Superior Court rejected Artis’ stop-the-clock reading of §1367(d), reasoning that Artis could have protected her state-law claims by “pursuing [them] in a state court while the federal court proceeding [was] pending.” Ibid. In tension with that explanation, the court noted that duplicative filings in federal and state court are “generally disfavored . . . as ‘wasteful’ and . . . ‘against [the interests of] judicial efficiency.’ ” Id., at 14a, n. 1 (quoting Stevens v. Arco Management of Wash. D.C., Inc., 751 A. 2d 995, 1002 (D. C. 2000); alteration in original). The D. C. Court of Appeals affirmed. That court began by observing that two “competing approaches [to §1367(d)] have evolved nationally”: the stop-the-clock reading and the grace-period reading. 135 A. 3d 334, 337 (2016).[3] Without further comment on §1367(d)’s text, the D. C. Court of Appeals turned to the legislative history. Section 1367(d)’s purpose, the court noted, was “to prevent the loss of claims to statutes of limitations where state law might fail to toll the running of the period of limitations while a supplemental claim was pending in federal court.” Id., at 338 (quoting H. R. Rep., at 30; internal quotation marks omitted). Following the lead of the California Supreme Court, the D. C. Court of Appeals determined that Congress had intended to implement a 1969 recommendation by the American Law Institute (ALI) to allow refiling in state court “within 30 days after dismissal.” 135 A. 3d, at 338 (quoting Los Angeles v. County of Kern, 59 Cal. 4th 618, 629, 328 P. 3d 56, 63 (2014)). The D. C. Court of Appeals also concluded that the grace-period approach “better accommodates federalism concerns,” by trenching significantly less on state statutes of limitations than the stop-the-clock approach. 135 A. 3d, at 338–339. Construing §1367(d) as affording only a 30-day grace period, the court commented, was “consistent with [its] presumption favoring narrow interpretations of federal preemption of state law.” Id., at 339. To resolve the division of opinion among State Supreme Courts on the proper construction of §1367(d), see supra, at 6, n. 3, we granted certiorari. 580 U. S. ___ (2017). II A As just indicated, statutes that shelter from time bars claims earlier commenced in another forum generally employ one of two means. First, the period (or statute) of limitations may be “tolled” while the claim is pending elsewhere.[4] Ordinarily, “tolled,” in the context of a time prescription like §1367(d), means that the limitations period is suspended (stops running) while the claim is sub judice elsewhere, then starts running again when the tolling period ends, picking up where it left off. See Black’s Law Dictionary 1488 (6th ed. 1990) (“toll,” when paired with the grammatical object “statute of limitations,” means “to suspend or stop temporarily”). This dictionary definition captures the rule generally applied in federal courts. See, e.g., Chardon v. Fumero Soto, 462 U. S. 650, 652, n. 1 (1983) (Court’s opinion “use[d] the word ‘tolling’ to mean that, during the relevant period, the statute of limitations ceases to run”).[5] Our decisions employ the terms “toll” and “suspend” interchangeably. For example, in American Pipe & Constr. Co. v. Utah, 414 U.S. 538 (1974) , we characterized as a “tolling” prescription a rule “suspend[ing] the applicable statute of limitations,” id., at 554; accordingly, we applied the rule to stop the limitations clock, id., at 560–561.[6] We have similarly comprehended what tolling means in decisions on equitable tolling. See, e.g., CTS Corp. v. Waldburger, 573 U. S. ___, ___ (2014) (slip op., at 7) (describing equitable tolling as “a doctrine that pauses the running of, or ‘tolls’ a statute of limitations” (some internal quotation marks omitted)); United States v. Ibarra, 502 U. S. 1, 4, n. 2 (1991) ( per curiam) (“Principles of equitable tolling usually dictate that when a time bar has been suspended and then begins to run again upon a later event, the time remaining on the clock is calculated by subtracting from the full limitations period whatever time ran before the clock was stopped.”). In lieu of “tolling” or “suspending” a limitations period by pausing its progression, a legislature might elect sim- ply to provide a grace period. When that mode is adopted, the statute of limitations continues to run while the claim is pending in another forum. But the risk of a time bar is averted by according the plaintiff a fixed period in which to refile. A federal statute of that genre is 28 U. S. C. §2415. That provision prescribes a six-year limitations period for suits seeking money damages from the United States for breach of contract. §2415(a). The statute further provides: “In the event that any action . . . is timely brought and is thereafter dismissed without prejudice, the action may be recommenced within one year after such dismissal, regardless of whether the action would otherwise then be barred by this section.” §2415(e).[7] Many States have enacted similar grace-period provisions. See App. to Brief for National Conference of State Legislatures et al. as Amici Curiae 1a–25a. For example, Georgia law provides: “When any case has been commenced in either a state or federal court within the applicable statute of limitations and the plaintiff discontinues or dismisses the same, it may be recommenced in a court of this state or in a federal court either within the original applicable period of limitations or within six months after the discontinuance or dismissal, whichever is later . . . .” Ga. Code Ann. §9–2–61(a) (2007). Tellingly, the District has not identified any federal statute in which a grace-period meaning has been ascribed to the word “tolled” or any word similarly rooted. Nor has the dissent, for all its mighty strivings, identified even one federal statute that fits its bill, i.e., a federal statute that says “tolled” but means something other than “suspended,” or “paused,” or “stopped.” From what statutory text, then, does the dissent start? See post, at 5.[8] Turning from statutory texts to judicial decisions, only once did an opinion of this Court employ tolling language to describe a grace period: Hardin v. Straub, 490 U. S. 536 (1989) . In Hardin, we held that, in 42 U. S. C. §1983 suits, federal courts should give effect to state statutes sheltering claims from time bars during periods of a plaintiff’s legal disability. We there characterized a state statute providing a one-year grace period as “tolling” or “suspend[ing]” the limitations period “until one year after the disability has been removed.” 490 U. S., at 537. This atypical use of “tolling” or “suspending” to mean something other than stopping the clock on a limitations period is a feather on the scale against the weight of decisions in which “tolling” a statute of limitations signals stopping the clock. B In determining the meaning of a statutory provision, “we look first to its language, giving the words used their ordinary meaning.” Moskal v. United States, 498 U. S. 103, 108 (1990) (citation and internal quotation marks omitted). Section 1367(d) is phrased as a tolling provision. It suspends the statute of limitations for two adjacent time periods: while the claim is pending in federal court and for 30 days postdismissal. Artis urges that the phrase “shall be tolled” in §1367(d) has the same meaning it does in the statutes cited supra, at 7, n. 4. That is, the limitations clock stops the day the claim is filed in federal court and, 30 days postdismissal, restarts from the point at which it had stopped. The District reads “tolled” for §1367(d)’s purposes differently. To “toll,” the District urges, means to “remove or take away an effect.” Brief for Respondent 12–13. To “toll” a limitations period, then, would mean to “remov[e] the bar that ordinarily would accompany its expiration.” Id., at 14.[9] “[T]here is nothing special,” the District maintains, “about tolling limitations periods versus tolling any other fact, right, or consequence.” Id., at 13. But the District offers no reason why, in interpreting “tolled” as used in §1367(d), we should home in only on the word itself, ignoring the information about the verb’s ordinary meaning gained from its grammatical object. Just as when the object of “tolled” is “bell” or “highway traveler,” the object “period of limitations” sheds light on what it means to “be tolled.” The District’s reading, largely embraced by the dissent, is problematic for other reasons as well. First, it tenders a strained interpretation of the phrase “period of limitations.” In the District’s view, “period of limitations” means “the effect of the period of limitations as a time bar.” See id., at 18 (“Section 1367(d) . . . provides that ‘the period of limitations’—here its effect as a time bar—‘shall be [removed or taken away] while the claim is pending [in federal court] and for a period of 30 days after it is dismissed.’ ” (alterations in original)). Second, the first portion of the tolling period, the duration of the claim’s pendency in federal court, becomes superfluous under the District’s construction. The “effect” of the limitations period as a time bar, on the District’s reading, becomes operative only after the case has been dismissed. That being so, what need would there be to remove anything while the claim is pending in federal court? Furthermore, the District’s reading could yield an absurdity: It could permit a plaintiff to refile in state court even if the limitations period on her claim had expired before she filed in federal court. To avoid that result, the District’s proposed construction of “tolled” as “removed” could not mean simply “removed.” Instead, “removed” would require qualification to express “removed, unless the period of limitations expired before the claim was filed in federal court.” In sum, the District’s interpretation maps poorly onto the language of §1367(d), while Artis’ interpretation is a natural fit. C The D. C. Court of Appeals adopted the District’s grace-period construction primarily because it was convinced that in drafting §1367(d), Congress embraced an ALI recommendation. 135 A. 3d, at 338. Two decades before the enactment of §1367(d), the ALI, in its 1969 Study of the Division of Jurisdiction Between State and Federal Courts, did recommend a 30-day grace period for refiling certain claims. The ALI proposed the following statutory language: “If any claim in an action timely commenced in a federal court is dismissed for lack of jurisdiction over the subject matter of the claim, a new action on the same claim brought in another court shall not be barred by a statute of limitations that would not have barred the original action had it been commenced in that court, if such new action is brought in a proper court, federal or State, within thirty days after dismissal of the original claim has become final or within such longer period as may be available under applicable State law.” ALI, Study of the Division of Jurisdiction Between State and Federal Courts §1386(b), p. 65 (1969) (ALI Study). Congress, however, did not adopt the ALI’s grace-period formulation. Instead, it ordered tolling of the state limitations period “while the claim is pending” in federal court. Although the provision the ALI proposed, like §1367(d), established a 30-day federal floor on the time allowed for refiling, it did not provide for tolling the period of limitations while a claim is pending.[10] True, the House Report contained a citation to the ALI Study, but only in reference to a different provision, 28 U. S. C. §1391 (the general venue statute). There, Congress noted that its approach was “taken from the ALI Study.” H. R. Rep., at 23. Had Congress similarly embraced the ALI’s grace-period formulation in §1367(d), one might expect the House Report to have said as much.[11] D The District asks us to zero in on §1367(d)’s “express inclusion” of the “period of 30 days after the claim is dismissed” within the tolling period. Brief for Respondent 20 (internal quotation marks omitted). Under Artis’ stop-the-clock interpretation, the District contends, “the inclusion of 30 days within the tolling period would be relegated to insignificance in the mine-run of cases.” Id., at 21 (citation and internal quotation marks omitted). In §1367(d), Congress did provide for tolling not only while the claim is pending in federal court, but also for 30 days thereafter. Including the 30 days within §1367(d)’s tolling period accounts for cases in which a federal action is commenced close to the expiration date of the relevant state statute of limitations. In such a case, the added days give the plaintiff breathing space to refile in state court. Adding a brief span of days to the tolling period is not unusual in stop-the-clock statutes. In this respect, §1367(d) closely resembles 46 U. S. C. §53911, which provides, in a subsection titled “Tolling of limitations period,” that if a plaintiff submits a claim for war-related vessel damage to the Secretary of Transportation, “the running of the limitations period for bringing a civil action is suspended until the Secretary denies the claim, and for 60 days thereafter.” §53911(d). Numerous other statutes similarly append a fixed number of days to an initial tolling period. See, e.g., 22 U. S. C. §1631k(c) (“Statutes of limitations on assessments . . . shall be suspended with respect to any vested property . . . while vested and for six months thereafter. . . .”); 26 U. S. C. §6213(f )(1) (“In any case under title 11 of the United States Code, the running of the time prescribed by subsection (a) for filing a petition in the Tax Court with respect to any deficiency shall be suspended for the period during which the debtor is prohibited by reason of such case from filing a petition in the Tax Court with respect to such deficiency, and for 60 days thereafter.”); §6503(a)(1) (“The running of the period of limitations provided in section 6501 or 6502 . . . shall . . . be suspended for the period during which the Secretary is prohibited from making the assessment . . . and for 60 days thereafter.”); 50 U. S. C. §4000(c) (“The running of a statute of limitations against the collection of tax deferred under this section . . . shall be suspended for the period of military service of the servicemember and for an addi- tional period of 270 days thereafter.”). Thus, the “30 days” provision casts no large shadow on Artis’ interpretation. Section 1367(d)’s proviso, “unless State law provides for a longer tolling period,” could similarly aid a plaintiff who filed in federal court just short of the expiration of the state limitations period. She would have the benefit of §1367(d)’s 30-days-to-refile prescription, or such longer time as state law prescribes.[12] It may be that, in most cases, the state-law tolling period will not be longer than §1367(d)’s. But in some cases it undoubtedly will. For example, Indiana permits a plaintiff to refile within three years of dismissal. See Ind. Code §34–11–8–1 (2017). And Louisiana provides that after dismissal the limitations period “runs anew.” La. Civ. Code Ann., Arts. 3462, 3466 (West 2007). III Satisfied that Artis’ text-based arguments overwhelm the District’s, we turn to the District’s contention that the stop-the-clock interpretation of §1367(d) raises a significant constitutional question: Does the statute exceed Congress’ authority under the Necessary and Proper Clause, Art. I, §8, cl. 18, because its connection to Congress’ enumerated powers is too attenuated or because it is too great an incursion on the States’ domain? Brief for Respondent 46–49. To avoid constitutional doubt, the District urges, we should adopt its reading. “[W]here an alternative interpretation of [a] statute is fairly possible,” the District reminds, we have construed legislation in a manner that “avoid[s] [serious constitutional] problems” raised by “an otherwise acceptable construction.” INS v. St. Cyr, 533 U. S. 289 –300 (2001) (internal quotation marks omitted). But even if we regarded the District’s reading of §1367(d) as “fairly possible,” our precedent would undermine the proposition that §1367(d) presents a serious constitutional problem. See Jinks, 538 U. S., at 461–465. In Jinks, we unanimously rejected an argument that §1367(d) impermissibly exceeds Congress’ enumerated powers.[13] Section 1367(d), we held, “is necessary and proper for carrying into execution Congress’s power ‘[t]o constitute Tribunals inferior to the supreme Court,’ . . . and to assure that those tribunals may fairly and effi- ciently exercise ‘[t]he judicial Power of the United States.’ ” Id., at 462 (quoting U. S. Const., Art. I, §8, cl. 9, and Art. III, §1). In two principal ways, we explained, §1367(d) is “conducive to the due administration of justice in federal court.” 538 U. S., at 462 (internal quotation marks omitted). First, “it provides an alternative to the unsatisfactory options that federal judges faced when they decided whether to retain jurisdiction over supplemental state-law claims that might be time barred in state court.” Ibid. Section 1367(d) thus “unquestionably promotes fair and efficient operation of the federal courts.” Id., at 463. Second, §1367(d) “eliminates a serious impediment to access to the federal courts on the part of plaintiffs pursuing federal- and state-law claims” arising from the same episode. Ibid. With tolling available, a plaintiff disinclined to litigate simultaneously in two forums is no longer impelled to choose between forgoing either her federal claims or her state claims. Moreover, we were persuaded that §1367(d) was “plainly adapted” to Congress’ exercise of its enumerated power: there was no cause to suspect that Congress had enacted §1367(d) as a “ ‘pretext’ for ‘the accomplishment of objects not entrusted to [it],’ ”; nor was there reason to believe that the connection between §1367(d) and Congress’ authority over the federal courts was too attenuated. Id., at 464 (quoting McCulloch v. Maryland, 4 Wheat. 316, 423 (1819)). Our decision in Jinks also rejected the argument that §1367(d) was not “proper” because it violates principles of state sovereignty by prescribing a procedural rule for state courts’ adjudication of purely state-law claims. 538 U. S., at 464–465. “Assuming [without deciding] that a principled dichotomy can be drawn, for purposes of determining whether an Act of Congress is ‘proper,’ between federal laws that regulate state-court ‘procedure’ and laws that change the ‘substance’ of state-law rights of action,” we concluded that the tolling of state limitations periods “falls on the [permissible] ‘substantive’ side of the line.” Ibid. The District’s contention that a stop-the-clock prescription serves “no federal purpose” that could not be served by a grace-period prescription is unavailing. Brief for Respondent 49. Both devices are standard, off-the-shelf means of accounting for the fact that a claim was timely pressed in another forum. Requiring Congress to choose one over the other would impose a tighter constraint on Congress’ discretion than we have ever countenanced. The concern that a stop-the-clock prescription entails a greater imposition on the States than a grace-period prescription, moreover, may be more theoretical than real. Consider the alternative suggested by the D. C. Superior Court. Plaintiffs situated as Artis was could simply file two actions and ask the state court to hold the suit filed there in abeyance pending disposition of the federal suit. See supra, at 6. Were the dissent’s position to prevail, cautious plaintiffs would surely take up the D. C. Superior Court’s suggestion. How it genuinely advances federalism concerns to drive plaintiffs to resort to wasteful, inefficient duplication to preserve their state-law claims is far from apparent. See, e.g., Stevens, 751 A. 2d, at 1002 (it “work[s] against judicial efficiency . . . to compel prudent federal litigants who present state claims to file duplicative and wasteful protective suits in state court”). We do not gainsay that statutes of limitations are “fundamental to a well-ordered judicial system.” Board of Regents of Univ. of State of N. Y. v. Tomanio, 446 U. S. 478, 487 (1980) . We note in this regard, however, that a stop-the-clock rule is suited to the primary purposes of limitations statutes: “preventing surprises” to defendants and “barring a plaintiff who has slept on his rights.” American Pipe & Constr. Co. v. Utah, 414 U. S. 538, 554 (1974) (internal quotation marks omitted). Whenever §1367(d) applies, the defendant will have notice of the plaintiff’s claims within the state-prescribed limitations period. Likewise, the plaintiff will not have slept on her rights. She will have timely asserted those rights, endeavoring to pursue them in one litigation. * * * For the reasons stated, we resist unsettling the usual understanding of the word “tolled” as it appears in legislative time prescriptions and court decisions thereon. The judgment of the D. C. Court of Appeals is therefore reversed, and the case is remanded for further proceedings not inconsistent with this opinion. It is so ordered. Notes 1 The nonfederal claims Artis asserted arose under the D. C. Code and common law; on dismissal of her federal-court suit, she refiled those claims in D. C. Superior Court. For the purpose at hand, District of Columbia law and courts are treated as state law and courts. See 28 U. S. C. §1367(e) (“As used in this section, the term ‘State’ includes the District of Columbia, the Commonwealth of Puerto Rico, and any territory or possession of the United States.”). 2 The D. C. False Claims Act and the tort of wrongful termination each have a three-year statute of limitations that started to run on the day Artis lost her job in November 2010. See D. C. Code §2–381.04(c) (2001) (D. C. False Claims Act); Stephenson v. American Dental Assn., 789 A. 2d 1248, 1249, 1252 (D. C. 2002) (tort of wrongful termination governed by D. C.’s catchall three-year limitations period and claim accrues on the date when plaintiff has unequivocal notice of termination). Artis’ whistleblower claim had a one-year limitations period, which began to accrue when Artis “first bec[a]m[e] aware” that she had been terminated for reporting her supervisor’s misconduct. D. C. Code §1–615.54(a)(2). The parties dispute the date the whistleblower claim accrued. See Brief for Petitioner 10, n. 2; Brief for Respondent 8, n. 2. 3 The high courts of Maryland and Minnesota, along with the Sixth Circuit, have held that §1367(d)’s tolling rule pauses the clock on the statute of limitations until 30 days after the state-law claim is dismissed. See In re Vertrue Inc. Marketing & Sales Practices Litigation, 719 F. 3d 474, 481 (CA6 2013); Goodman v. Best Buy, Inc., 777 N. W. 2d 755, 759–760 (Minn. 2010); Turner v. Kight, 406 Md. 167, 180–182, 957 A. 2d 984, 992–993 (2008). In addition to the D. C. Court of Appeals, the high courts of California and the Northern Mariana Islands have held that §1367(d) provides only a 30-day grace period for the refiling of otherwise time-barred claims. See Los Angeles v. County of Kern, 59 Cal. 4th 618, 622, 328 P. 3d 56, 58 (2014); Juan v. Commonwealth, 2001 MP 18, 6 N. Mar. I. 322, 327 (2001). 4 Among illustrations: 21 U. S. C. §1604 (allowing suits to proceed against certain biomaterial providers and providing that “[a]ny applicable statute of limitations shall toll during the period from the time a claimant files a petition with the Secretary under this paragraph until such time as either (i) the Secretary issues a final decision on the petition, or (ii) the petition is withdrawn,” §1604(b)(3)(C)); 28 U. S. C. §1332 (permitting the removal of “mass actions” to federal court and providing that “[t]he limitations periods on any claims asserted in a mass action that is removed to Federal court pursuant to this subsection shall be deemed tolled during the period that the action is pending in Federal court,” §1332(d)(11)(D)); 42 U. S. C. §233 (providing a remedy against the United States for certain injuries caused by employees of the Public Health Service, and stating that “[t]he time limit for filing a claim under this subsection . . . shall be tolled during the pendency of a[n] [administrative] request for benefits,” §233(p)(3)(A)(ii)). See also Wis. Stat. §893.15(3) (2011–2012) (“A Wisconsin law limiting the time for commencement of an action on a Wisconsin cause of action is tolled from the period of commencement of the action in a non-Wisconsin forum until the time of its final disposition in that forum.”). The dissent maintains that “stop clock examples [from the U. S. Code] often involve situations where some disability prevents the plaintiff from proceeding to court.” Post, at 12, n. 7. Plainly, however, the several statutes just set out do not fit that description: They do not involve “disabilities.” Instead, like §1367(d), they involve claims earlier commenced in another forum. 5 As we recognized in Chardon v. Fumero Soto, 462 U. S. 650 (1983) , there may be different ways of “calculating the amount of time avail-able to file suit after tolling has ended.” Id., at 652, n. 1 (emphasis added). In addition to the “common-law” stop-the-clock effect, id., at 655, under which the plaintiff must file within the amount of time left in the limitations period, a statute might either provide for the limitations period to be “renewed,” so that “the plaintiff has the benefit of a new period as long as the original,” or “establish a fixed period such as six months or one year during which the plaintiff may file suit, without regard to the length of the original limitations period or the amount of time left when tolling began.” Id., at 652, n. 1. Notably, under each of the “tolling effect[s]” enumerated in Chardon, ibid., the word “tolled” means that the progression of the limitations clock is stopped for the duration of “tolling.” 6 The dissent’s notion that federal tolling periods may be understood as grace periods, not stop-the-clock periods, see post, at 7–8, is entirely imaginative. 7 Also illustrative, the Equal Credit Opportunity Act prescribes a five-year limitations period for certain suits. 15 U. S. C. §1691e(f ). Where a government agency has brought a timely suit, however, an individual may bring an action “not later than one year after the commencement of that proceeding or action.” Ibid. 8 Reasons of history, context, and policy, the dissent maintains, would have made it sensible for Congress to have written a grace-period statute. See post, at 4–5. But “[t]he controlling principle in this case is the basic and unexceptional rule that courts must give effect to the clear meaning of statutes as written[,] . . . giving each word its ordinary, contemporary, common meaning.” Star Athletica, L. L. C. v. Varsity Brands, Inc., 580 U. S. ___, ___ (2017) (slip op., at 6) (internal quotation marks omitted). 9 This is indeed a definition sometimes used in reference to a right. See, e.g., Ricard v. Williams, 7 Wheat. 59, 120 (1822) (“[A]n adverse possession . . . toll[s] the right of entry of the heirs, and, consequently, extinguish[es], by the lapse of time, their right of action for the land.”). See also Black’s Law Dictionary 1488 (6th ed. 1990) (“toll” can mean “bar, defeat, or take away; thus, to toll the entry means to deny or take away the right of entry”). The dissent, also relying on this sense of the word “toll,” cites Chardon as support for the proposition that §1367(d)’s tolling instruction is ambiguous. See post, at 3; supra, at 8, n. 5. But, importantly, the grace-period statutes noted in Chardon, 462 U. S., at 660, n. 13, were precise about their operation. Chardon provides no support for the notion that a statute’s instruction that a “period of limitations shall be tolled” plausibly could mean that the limitations clock continues to run but its effect as a bar is removed during the tolling. See post, at 2–3. 10 The District emphasizes that the Reporter’s note accompanying the ALI’s proposed statute stated: “[A]ny governing statute of limitations is tolled by the commencement of an action in a federal court, and for at least thirty days following dismissal . . . in any case in which the dismissal was for lack of jurisdiction.” ALI Study 66. The similarity between this language and §1367(d), the District argues, rebuts any argument that Congress did not adopt the ALI’s recommendation. We are unpersuaded. The District offers no explanation why, if Congress wanted to follow the substance of the ALI’s grace-period recommendation, it would neither cite the ALI Study in the legislative history of §1367(d), see infra this page, nor adopt the precise language of either the proposed statute or the Reporter’s note. The ALI Study, moreover, cautions that the Reporter’s notes reflect “the Reporter’s work alone,” not a position taken by the Institute. ALI Study, p. x. 11 The dissent offers a history lesson on the ancient common-law principle of “journey’s account,” see post, at 5–6, and n. 4, but nothing suggests that the 101st Congress had any such ancient law in mind when it drafted §1367(d). Cf. post, at 9. More likely, Congress was mindful that “suspension” during the pendency of other litigation is “the common-law rule.” Chardon, 462 U. S., at 655. 12 The dissent, post, at 8–9, conjures up absurdities not presented by this case, for the District of Columbia has no law of the kind the dissent describes. All agree that the phrase “unless State law provides for a longer tolling period” leaves room for a more generous state-law regime. The dissent posits a comparison between the duration of the federal suit, plus 30 days, and a state-law grace period. But of course, as the dissent recognizes, post, at 9, the more natural comparison is between the amount of time a plaintiff has left to refile, given the benefit of §1367(d)’s tolling rule, and the amount of time she would have to refile under the applicable state law. Should the extraordinary circumstances the dissent envisions in fact exist in a given case, the comparison the dissent makes would be far from inevitable. 13 The dissent refers to an “understanding,” post, at 14, by the Court in Jinks v. Richland County, 538 U. S. 456 (2003) , that §1367(d) accords only a 30-day “window” for refiling in state court. Scattered characterizations in the Jinks briefing might be seen as conveying that understanding. See post, at 14, n. 9. The opinion itself, however, contains nary a hint of any such understanding. And indeed, one year earlier, we described §1367(d) as having the effect of stopping the clock, i.e., “toll[ing] the state statute of limitations for 30 days in addition to however long the claim had been pending in federal court.” Raygor v. Regents of Univ. of Minn., 534 U.S. 533, 542 (2002) .

584.US.2017_16-6795

Petitioner Ayestas was convicted of murder and sentenced to death in a Texas state court. He secured new counsel, but his conviction and sentence were affirmed on appeal. A third legal team sought, unsuccessfully, state habeas relief, claiming trial-level ineffective assistance of counsel but not counsel’s failure to investigate petitioner’s mental health and alcohol and drug abuse during the trial’s penalty phase. His fourth set of attorneys did raise that failure in a federal habeas petition, but because the claim had never been raised in state court, the District Court held, it was barred by procedural default. That decision was vacated and remanded for reconsideration in light of Martinez v. Ryan, 566 U. S. 1 —where this Court held that an Arizona prisoner seeking federal habeas relief could overcome the procedural default of a trial-level ineffective-assistance-of-counsel claim by showing that the claim is substantial and that state habeas counsel was also ineffective in failing to raise the claim in a state habeas proceeding—and Trevino v. Thaler, 569 U. S. 413 —which extended that holding to Texas prisoners. Petitioner filed an ex parte motion asking the District Court for funding to develop his claim that both his trial and state habeas counsel were ineffective, relying on 18 U. S. C. §3599(f), which provides, in relevant part, that a district court “may authorize” funding for “investigative, expert, or other services . . . reasonably necessary for the representation of the defendant.” The court found his claim precluded by procedural default and thus denied his funding request. The Fifth Circuit also rejected the funding claim under its precedent: that a §3599(f) funding applicant must show that he has a “substantial need” for investigative or other services, and that funding may be denied when an applicant fails to present “a viable constitutional claim that is not procedurally barred.” 817 F. 3d 888, 895–896. Held: 1. The District Court’s denial of petitioner’s funding request was a judicial decision subject to appellate review under the standard jurisdictional provisions. Pp. 7–14. (a) Title 28 U. S. C. §§1291, 2253, and 1254 confer jurisdiction to review decisions made by a district court in a judicial capacity. “Administrative” decisions—about, e.g., facilities, personnel, equipment, supplies, and rules of procedure—are “not subject to [this Court’s] review,” Hohn v. United States, 524 U. S. 236 , but the District Court’s ruling here does not remotely resemble such decisions. Petitioner’s request was made by motion in his federal habeas proceeding, which is indisputably a judicial proceeding. And resolution of the funding question requires the application of a legal standard—whether the funding is “reasonably necessary” for effective representation—that demands an evaluation of petitioner’s prospects of obtaining habeas relief. Pp. 8–10. (b) Respondent’s arguments in support of her claim that §3599’s funding requests are nonadversarial and administrative are unpersuasive. First, that the requests can be decided ex parte does not make the proceeding nonadversarial. The habeas proceeding here was clearly adversarial. And petitioner and respondent plainly have adverse interests on the funding question and have therefore squared off as adversaries. The mere fact that a §3599 funding request may sometimes be made ex parte is thus hardly dispositive. Second, nothing in §3599 even hints that the funding decisions may be revised by the Director of the Administrative Office of the Courts. Lower court cases that appear to have accepted Administrative Office review of certain Criminal Justice Act (CJA) payments, even if a proper interpretation of the CJA, are inapposite. Finally, the fact that §3599(g)(2) requires funding in excess of the generally applicable statutory cap to be approved by the circuit’s chief judge or another designated circuit judge, instead of by a panel of three, does not make the proceeding administrative. If Congress wishes to make certain rulings reviewable by a single circuit judge, the Constitution does not stand in the way. Pp. 10–14. 2. The Fifth Circuit did not apply the correct legal standard in affirming the denial of petitioner’s funding request. Section 3599 authorizes funding for the “reasonably necessary” services of experts, investigators, and the like. But the Fifth Circuit’s requirement that applicants show a “substantial need” for the services is arguably a more demanding standard. Section 3599 appears to use the term “necessary” to mean something less than essential. Because it makes little sense to refer to something as being “reasonably essential,” the Court concludes that the statutory phrase calls for the district court to determine, in its discretion, whether a reasonable attorney would regard the services as sufficiently important, guided by considerations detailed in the opinion. The term “substantial” in the Fifth Circuit’s test, however, suggests a heavier burden. And that court exacerbated the difference by also requiring a funding applicant to present “a viable constitutional claim that is not procedurally barred.” That rule that is too restrictive after Trevino, see 569 U. S. at 429, because, in cases where funding stands a credible chance of enabling a habeas petitioner to overcome the procedural default obstacle, it may be error for a district court to refuse funding. That being said, district courts were given broad discretion in assessing funding requests when Congress changed the phrase “shall authorize” in §3599’s predecessor statute, see 21 U. S. C. §848(q)(9), to “may authorize” in §3599(f). A funding applicant must not be expected to prove that he will be able to win relief if given the services, but the “reasonably necessary” test does require an assessment of the likely utility of the services requested. Respondent’s alternative ground for affirmance—that funding is never “reasonably necessary” where a habeas petitioner seeks to present a procedurally defaulted ineffective-assistance-of-trial-counsel claim that depends on facts outside the state-court record—remains open for the Fifth Circuit to consider on remand. Pp. 14–19. 817 F. 3d 888, vacated and remanded. Alito, J., delivered the opinion for a unanimous Court. Sotomayor, J., filed a concurring opinion, in which Ginsburg, J., joined.

Petitioner Carlos Ayestas, who was convicted of murder and sentenced to death in a Texas court, argues that he was wrongfully denied funding for investigative services needed to prove his entitlement to federal habeas relief. Petitioner moved for funding under 18 U. S. C. §3599(f ), which makes funds available if they are “reasonably necessary,” but petitioner’s motion was denied. We hold that the lower courts applied the wrong legal standard, and we therefore vacate the judgment below and remand for further proceedings. I A In 1997, petitioner was convicted of capital murder in a Texas court. Evidence at trial showed that he and two accomplices invaded the home of a 67-year-old Houston woman, Santiaga Paneque, bound her with duct tape and electrical cord, beat and strangled her, and then made off with a stash of her belongings. The jury also heard testimony from Henry Nuila regarding an incident that occurred about two weeks after the murder. Petitioner was drunk at the time, and he revealed to Nuila that he had recently murdered a woman in Houston. Petitioner then brandished an Uzi machinegun and threatened to murder Nuila if he did not help petitioner kill his two accomplices. Fortunately for Nuila, petitioner kept talking until he eventually passed out; Nuila then called the police, who arrested petitioner, still in possession of the gun. After the jury found petitioner guilty, it was asked to determine whether he should be sentenced to death or to life in prison. In order to impose a death sentence, Texas law required the jury to answer the following three questions. First, would petitioner pose a continuing threat to society? Second, had he personally caused the death of the victim, intended to kill her, or anticipated that she would be killed? Third, in light of all the evidence surrounding the crime and petitioner’s background, were there sufficient mitigating circumstances to warrant a sentence of life without parole instead of death? Tex. Code Crim. Proc. Ann., Art. 37.071, §§2(b), (e) (Vernon Cum. Supp. 2017). Only if the jury gave a unanimous yes to the first two questions, and a unanimous no to the third question, could a death sentence be imposed; otherwise, petitioner would receive a sentence of life without parole. See §§2(d)(2), (f )(2), (g). In asking the jury to impose a death sentence, the prosecution supplemented the trial record with evidence of petitioner’s criminal record and his encounter with a man named Candelario Martinez a few days after the murder. Martinez told the jury that he was standing in a hotel parking lot waiting for a friend when petitioner approached and began to make small talk. Before long, petitioner pulled out a machinegun and forced Martinez into a room where two of petitioner’s compatriots were holding Martinez’s friend at knifepoint. Ordered to lie down on the bathroom floor and await his execution, Martinez begged for his life while petitioner and his cohorts haggled about who would carry out the killing. Finally, petitioner relented, but he threatened to kill Martinez and his family if he contacted the police. Petitioner then stole Martinez’s truck. Petitioner’s trial counsel presented very little mitigation evidence. This was due, at least in part, to petitioner’s steadfast refusal for many months to allow his lawyers to contact his family members, who were living in Honduras and might have testified about his character and upbringing. Petitioner gave in on the eve of trial, and at that point, according to the state habeas courts, his lawyers “made every effort to contact [his] family.” App. 171. They repeatedly contacted petitioner’s family members and urged them to attend the trial; they requested that the U. S. Embassy in Honduras facilitate family members’ travel to the United States; and they met in person with the Honduran Consulate to seek assistance. But these efforts were to no avail. Petitioner’s sister told his legal team that the family would not leave Honduras because the journey would create economic hardship and because their father was ill and had killed one of their neighbors. A defense attorney who spoke to petitioner’s mother testified that she seemed unconcerned about her son’s situation. In general, the state habeas courts found, petitioner “did nothing to assist counsel’s efforts to contact his family and did not want them contacted by the consulate or counsel.” Id., at 174. In the end, the only mitigation evidence introduced by petitioner’s trial counsel consisted of three letters from petitioner’s English instructor. The letters, each two sentences long, described petitioner as “a serious and attentive student who is progressing well in English.” Ibid. The jury unanimously concluded that petitioner should be sentenced to death, and a capital sentence was imposed. Petitioner secured new counsel to handle his appeal, and his conviction and sentence were affirmed by the Texas Court of Criminal Appeals in 1998. Ayestas v. State, No. 72,928, App. 115. Petitioner did not seek review at that time from this Court. B While petitioner’s direct appeal was still pending, a third legal team filed a habeas petition on his behalf in state court. This petition included several claims of trial-level ineffective assistance of counsel, but the petition did not assert that trial counsel were ineffective for failing to investigate petitioner’s mental health and abuse of alcohol and drugs. Petitioner’s quest for state habeas relief ended unsuccessfully in 2008. Ex parte Ayestas, No. WR–69,674–01 (Tex. Ct. Crim. App., Sept. 10, 2008), 2008 WL 4151814 (per curiam) (unpublished). In 2009, represented by a fourth set of attorneys, petitioner filed a federal habeas petition under 28 U. S. C. §2254, and this time he did allege that his right to the effective assistance of counsel at trial was violated because his attorneys failed to conduct an adequate search for mitigation evidence. As relevant here, petitioner argued that trial counsel overlooked evidence that he was mentally ill and had a history of drug and alcohol abuse. Ayestas v. Thaler, Civ. Action No. H–09–2999 (SD Tex., Jan. 26, 2011), 2011 WL 285138, *4. Petitioner alleged that he had a history of substance abuse, and he noted that he had been diagnosed with schizophrenia while the state habeas proceeding was still pending. See Pet. for Writ of Habeas Corpus in Ayestas v. Quarterman, No. 4:09–cv–2999 (SD Tex.), Doc. 1, pp. 21–23. Petitioner claimed that trial counsel’s deficient performance caused prejudice because there was a reasonable chance that an adequate investigation would have produced mitigation evidence that would have persuaded the jury to spare his life. Among the obstacles standing between petitioner and federal habeas relief, however, was the fact that he never raised this trial-level ineffective-assistance-of-counsel claim in state court. The District Court therefore held that the claim was barred by procedural default, Ayestas v. Thaler, 2011 WL 285138, *4–*7, and the Fifth Circuit affirmed, Ayestas v. Thaler, 462 Fed. Appx. 474, 482 (2012) (per curiam). Petitioner sought review in this Court, and we vacated the decision below and remanded for reconsideration in light of two of our subsequent decisions, Martinez v. Ryan, 566 U. S. 1 (2012) , and Trevino v. Thaler, 569 U. S. 413 (2013) . Ayestas v. Thaler, 569 U. S. 1015 (2013) . Martinez held that an Arizona prisoner seeking federal habeas relief could overcome the procedural default of a trial-level ineffective-assistance-of-counsel claim by showing that the claim is substantial and that state habeas counsel was also ineffective in failing to raise the claim in a state habeas proceeding. 566 U. S., at 14. Trevino extended that holding to Texas prisoners, 569 U. S., at 416–417, and on remand, petitioner argued that he fell within Trevino because effective state habeas counsel would have uncovered evidence showing that trial counsels’ investigative efforts were deficient. To assist in developing these claims, petitioner filed an ex parte motion asking the District Court for $20,016 in funding to conduct a search for evidence supporting his petition. He relied on 18 U. S. C. §3599(f ), which provides in relevant part as follows: “Upon a finding that investigative, expert, or other services are reasonably necessary for the representation of the defendant, whether in connection with issues relating to guilt or the sentence, the court may authorize the defendant’s attorneys to obtain such services on behalf of the defendant and, if so authorized, shall order the payment of fees and expenses therefor.” Petitioner averred that the funds would be used to conduct an investigation that would show that his trial counsel and his state habeas counsel were ineffective. Accordingly, he claimed, the investigation would establish both that his trial-level ineffective-assistance-of-counsel claim was not barred by procedural default and that he was entitled to resentencing based on the denial of his Sixth Amendment right to the effective assistance of trial counsel. The District Court refused the funding request and ultimately denied petitioner’s habeas petition. Ayestas v. Stephens, Civ. Action No. H–09–2999, (SD Tex., Nov. 18, 2014), 2014 WL 6606498, *6–*7. On the merits of petitioner’s new ineffective-assistance-of-trial-counsel claim, the District Court held that petitioner failed both prongs of the Strickland test. See Strickland v. Washington, 466 U. S. 668 (1984) . Noting that most of the evidence bearing on petitioner’s mental health had emerged only after he was sentenced, the court concluded that petitioner’s trial lawyers were not deficient in failing to find such evidence in time for the sentencing proceeding. 2014 WL 6606498, *5. In addition, the court found that state habeas counsel did not render deficient performance by failing to investigate petitioner’s history of substance abuse, and that, in any event, petitioner was not prejudiced at the sentencing phase of the trial or during the state habeas proceedings because the potential mitigation evidence at issue would not have made a difference to the jury in light of “the extremely brutal nature of [the] crime and [petitioner’s] history of criminal violence.” Ibid. With respect to funding, the District Court pointed to Fifth Circuit case law holding that a §3599(f ) funding applicant cannot show that investigative services are “ ‘reasonably necessary’ ” unless the applicant can show that he has a “ ‘substantial need’ ” for those services. Id., at *6. In addition, the court noted that “[t]he Fifth Circuit upholds the denial of funding” when, among other things, “a petitioner has . . . failed to supplement his funding request with a viable constitutional claim that is not procedurally barred.” Ibid. (internal quotation marks omitted). Given its holding that petitioner’s new ineffective-assistance-of-counsel claim was precluded by procedural default, this rule also doomed his request for funding. The District Court denied petitioner’s habeas petition and refused to grant him a certificate of appealability (COA). Id., at *7. On appeal, the Fifth Circuit held that a COA was not needed for review of the funding issue, but it rejected that claim for essentially the same reasons as the District Court, citing both the “substantial need” test and the rule that funding may be denied when a funding applicant fails to present “a viable constitutional claim that is not procedurally barred.” Ayestas v. Stephens, 817 F. 3d 888, 895–896 (2016) (internal quotation marks omitted). With respect to petitioner’s other claims, including his claim of ineffective assistance of trial counsel, the Fifth Circuit refused to issue a COA. Id., at 898. C We granted certiorari to decide whether the lower courts applied the correct legal standard in denying the funding request. 581 U. S. ___ (2017). II Before we reach that question, however, we must consider a jurisdictional argument advanced by respondent, the Director of the Texas Department of Criminal Justice.[1] Respondent contends that the District Court’s denial of petitioner’s funding request was an administrative, not a judicial, decision and therefore falls outside the scope of the jurisdictional provisions on which petitioner relied in seeking review in the Court of Appeals and in this Court. A When the District Court denied petitioner’s funding request and his habeas petition, he took an appeal to the Fifth Circuit under 28 U. S. C. §§1291 and 2253, which grant the courts of appeals jurisdiction to review final “decisions” and “orders” of a district court.[2] And when the Fifth Circuit affirmed, petitioner sought review in this Court under §1254, which gives us jurisdiction to review “[c]ases” in the courts of appeals.[3] As respondent correctly notes, these provisions confer jurisdiction to review decisions made by a district court in a judicial capacity. But we have recognized that not all decisions made by a fed- eral court are “judicial” in nature; some decisions are prop- erly understood to be “administrative,” and in that case they are “not subject to our review.” Hohn v. United States, 524 U. S. 236, 245 (1998) . The need for federal judges to make many administrative decisions is obvious. The Federal Judiciary, while tiny in comparison to the Executive Branch, is nevertheless a large and complex institution, with an annual budget exceeding $7 billion and more than 32,000 employees. See Administrative Office of the U. S. Courts, The Judiciary FY 2018 Congressional Budget Summary Revised 9–10 (June 2017). Administering this operation requires many “decisions” in the ordinary sense of the term—decisions about such things as facilities, personnel, equipment, supplies, and rules of procedure. In re Application for Exemption from Electronic Pub. Access Fees by Jennifer Gollan and Shane Shifflett, 728 F. 3d 1033, 1037 (CA9 2013). It would be absurd to suggest that every “final decision” on any such matter is appealable under §1291 or reviewable in this Court under §1254. See Hohn, supra; 15A C. Wright, A. Miller, & E. Cooper, Federal Practice and Procedure §3903, pp. 134–135 (2d ed. 1992). Such administrative decisions are not the kind of decisions or orders—i.e., decisions or orders made in a judicial capacity—to which the relevant jurisdictional provisions apply. Respondent argues that the denial of petitioner’s funding request was just such an administrative decision, but the District Court’s ruling does not remotely resemble the sort of administrative decisions noted above. Petitioner’s request was made by motion in his federal habeas proceeding, which is indisputably a judicial proceeding. And as we will explain, resolution of the funding question requires the application of a legal standard—whether the funding is “reasonably necessary” for effective representation—that demands an evaluation of petitioner’s prospects of obtaining habeas relief. We have never held that a ruling like that is administrative and thus not subject to appellate review under the standard jurisdictional provisions. Respondent claims that two factors support the conclusion that the funding decision was administrative, but her argument is unpersuasive. B Respondent first argues as follows: Judicial proceedings must be adversarial; 18 U. S. C. §3599(f ) funding adjudications are not adversarial because the statute allows requests to be decided ex parte; therefore, §3599(f ) funding adjudications are not judicial in nature. This reasoning is flawed. It is certainly true that cases and controversies in our legal system are adversarial in nature, e.g., Bond v. United States, 564 U. S. 211, 217 (2011) ; Aetna Life Ins. Co. v. Haworth, 300 U. S. 227 –241 (1937), but here, both the habeas proceeding as a whole and the adjudication of the specific issue of funding were adversarial. That the habeas proceeding was adversarial is beyond dispute. And on the funding question, petitioner and respondent plainly have adverse interests and have therefore squared off as adversaries. The motion for funding was formally noted as “opposed” on the Disrict Court’s docket. App. 341. That is not surprising: On one side, petitioner is seeking funding that he hopes will prevent his execution. On the other, respondent wants to enforce the judgment of the Texas courts and to do so without undue delay. Petitioner and respondent have vigorously litigated the funding question all the way to this Court. In arguing that the funding dispute is nonadversarial, respondent attaches too much importance to the fact that the request was made ex parte. As we have noted, the “ex parte nature of a proceeding has not been thought to imply that an act otherwise within a judge’s lawful jurisdiction was deprived of its judicial character.” Forrester v. White, 484 U. S. 219, 227 (1988) . In our adversary system, ex parte motions are disfavored, but they have their place. See, e.g., Hohn, supra, at 248 (application for COA); Dalia v. United States, 441 U. S. 238, 255 (1979) (application for a search warrant); 50 U. S. C. §1805(a) (application to conduct electronic surveillance for foreign intelligence); 18 U. S. C. §2518(3) (applications to intercept “wire, oral, or electronic communications”); 15 U. S. C. §1116(d)(1)(A) (application to seize certain goods and counterfeit marks involved in violations of the trademark laws); Fed. Rule Crim. Proc. 17(b) (application for witness subpoena); Fed. Rule Crim. Proc. 47(c) (generally recognizing ex parte motions and applications); Ullmann v. United States, 350 U. S. 422 –424, 434 (1956) (application for an order granting a witness immunity in exchange for self-incriminating testimony); United States v. Monsanto, 491 U. S. 600 –604 (1989) (motion to freeze defendant’s assets pending trial). Thus, the mere fact that a §3599 funding request may sometimes be made ex parte is hardly dispositive. See Hohn, 524 U. S., at 249; Tutun v. United States, 270 U. S. 568, 577 (1926) . C Respondent’s second argument is based on the vener- able principle “that Congress cannot vest review of the decisions of Article III courts in” entities other than “superior courts in the Article III hierarchy.” Plaut v. Spendthrift Farm, Inc., 514 U. S. 211 –219 (1995) (citing Hayburn’s Case, 2 Dall. 409 (1792)). Respondent claims that §3599 funding decisions may be revised by the Director of the Administrative Office of the Courts and that this shows that such decisions must be administrative. This argument, however, rests on a faulty premise. Nothing in §3599 even hints that review by the Director of the Administrative Office is allowed. Respondent’s argument rests in part on a handful of old lower court cases that appear to have accepted Administrative Office review of Criminal Justice Act of 1964 (CJA) payments that had been authorized by a District Court and approved by the chief judge of the relevant Circuit. See United States v. Aadal, 282 F. Supp. 664, 665 (SDNY 1968); United States v. Gast, 297 F. Supp. 620, 621–622 (Del. 1969); see also United States v. Hunter, 385 F. Supp. 358, 362 (DC 1974). The basis for these decisions was a provision of the CJA, 18 U. S. C. §3006A(h) (1964 ed.), stating that CJA payments “shall be made under the supervision of the Director of the Administrative Office of the United States Courts.”[4] It is not clear whether these decisions correctly interpreted the CJA,[5] but in any event, no similar language appears in §3599. And respondent has not identified a single instance in which the Director of the Administrative Office or any other nonjudicial officer has attempted to review or alter a §3599 decision. Moreover, attorneys’ requests for CJA funds are markedly different from the funding application at issue here. Attorneys appointed under the CJA typically submit those requests after the conclusion of the case, and the prosecution has no stake in the resolution of the matter. The judgment in the criminal case cannot be affected by a decision on compensation for services that have been completed, and any funds awarded come out of the budget of the Judiciary, not the Executive. See 18 U. S. C. §3006A(i) (2012 ed.). Thus, the adversaries in the criminal case are not pitted against each other. In this case, on the other hand, as we have explained, petitioner and respondent have strong adverse interests. For these reasons, we reject respondent’s argument that the adjudication of the funding issue is nonadversarial and administrative. Respondent, however, claims that the funding decision is administrative for an additional reason. “A §3599(f ) funding determination is properly deemed administrative,” she contends, “because it . . . may be revised outside the traditional Article III judicial hierarchy.” Brief for Respondent 23. The basis for this argument is a provision of §3599 stating that funding in excess of the generally applicable statutory cap of $7,500 must be approved by the chief judge of the circuit or another designated circuit judge. §3599(g)(2). If a funding decision is judicial and not administrative, respondent suggests, it could not be reviewed by a single circuit judge as opposed to a panel of three. This argument confuses what is familiar with what is constitutionally required. Nothing in the Constitution ties Congress to the typical structure of appellate review established by statute. If Congress wishes to make certain rulings reviewable by a single circuit judge, rather than a panel of three, the Constitution does not stand in the way. III Satisfied that we have jurisdiction, we turn to the question whether the Court of Appeals applied the correct legal standard when it affirmed the denial of petitioner’s funding request. Section 3599(a) authorizes federal courts to provide funding to a party who is facing the prospect of a death sentence and is “financially unable to obtain adequate representation or investigative, expert, or other reason- ably necessary services.” The statute applies to defendants in federal cases, §3599(a)(1), as well as to state and federal prisoners seeking collateral relief in federal court, §3599(a)(2). Here we are concerned not with legal representation but with services provided by experts, investigators, and the like. Such services must be “reasonably necessary for the representation of the [applicant]” in order to be eligible for funding. §3599(f ). If the statutory standard is met, a court “may authorize the [applicant’s] attorneys to obtain such services on [his] behalf.” Ibid. The Fifth Circuit has held that individuals seeking funding for such services must show that they have a “substantial need” for the services. 817 F. 3d, at 896; Allen v. Stephens, 805 F. 3d 617, 626 (2015); Ward v. Stephens, 777 F. 3d 250, 266, cert. denied, 577 U. S. ___ (2015). Petitioner contends that this interpretation is more demanding than the standard—“reasonably necessary”—set out in the statute. And although the difference between the two formulations may not be great, petitioner has a point. In the strictest sense of the term, something is “necessary” only if it is essential. See Webster’s Third New International Dictionary 1510 (1993) (something is necessary if it “must be by reason of the nature of things,” if it “cannot be otherwise by reason of inherent qualities”); 10 Oxford English Dictionary 275–276 (2d ed. 1989) (OED) (defining the adjective “necessary” to mean “essential”). But in ordinary speech, the term is often used more loosely to refer to something that is merely important or strongly desired. (“I need a vacation.” “I need to catch up with an old friend.”) The term is sometimes used in a similar way in the law. The term “necessary” in the Necessary and Proper Clause does not mean “absolutely necessary,” McCulloch v. Maryland, 4 Wheat. 316, 414–415 (1819), and a “necessary” business expense under the Internal Revenue Code, 26 U. S. C. §162(a), may be an expense that is merely helpful and appropriate, Commissioner v. Tellier, 383 U. S. 687, 689 (1966) . As Black’s Law Dictionary puts it, the term “may import absolute physical necessity or inevitability, or it may import that which is only convenient, useful, appropriate, suitable, proper, or conducive to the end sought.” Black’s Law Dictionary 928 (5th ed. 1979) (Black’s). Section 3599 appears to use the term “necessary” to mean something less than essential. The provision applies to services that are “reasonably necessary,” but it makes little sense to refer to something as being “reasonably essential.” What the statutory phrase calls for, we conclude, is a determination by the district court, in the exercise of its discretion, as to whether a reasonable attorney would regard the services as sufficiently important, guided by the considerations we set out more fully below. The Fifth Circuit’s test—“substantial need”—is arguably more demanding. We may assume that the term “need” is comparable to “necessary”—that is, that something is “needed” if it is “necessary.” But the term “substantial” suggests a heavier burden than the statutory term “reasonably.” Compare 13 OED 291 (defining “reasonably” to mean, among other things, “[s]ufficiently, suitably, fairly”; “[f ]airly or pretty well”) with 17 id., at 66–67 (defining “substantial,” with respect to “reasons, causes, evidence,” to mean “firmly or solidly established”); see also Black’s 1456 (10th ed. 2014) (defining “reasonable” to mean “[f ]air, proper, or moderate under the circumstances . . . See plausible”); id., at 1656 (defining “substantial” to mean, among other things, “[i]mportant, essential, and material”). The difference between “reasonably necessary” and “substantially need[ed]” may be small, but the Fifth Circuit exacerbated the problem by invoking precedent to the effect that a habeas petitioner seeking funding must present “a viable constitutional claim that is not procedurally barred.” 817 F. 3d, at 895 (internal quotation marks omitted). See also, e.g., Riley v. Dretke, 362 F. 3d 302, 307 (CA5 2004) (“A petitioner cannot show a substantial need when his claim is procedurally barred from review”); Allen, supra, at 638–639 (describing “ ‘our rule that a prisoner cannot show a substantial need for funds when his claim is procedurally barred from review’ ” (quoting Crutsinger v. Stephens, 576 Fed. Appx. 422, 431 (CA5 2014) (per curiam)); Ward, supra, at 266 (“The denial of funding will be upheld . . . when the constitutional claim is procedurally barred”). The Fifth Circuit adopted this rule before our decision in Trevino, but after Trevino, the rule is too restrictive. Trevino permits a Texas prisoner to overcome the failure to raise a substantial ineffective-assistance claim in state court by showing that state habeas counsel was ineffective, 569 U. S., at 429, and it is possible that investigation might enable a petitioner to carry that burden. In those cases in which funding stands a credible chance of enabling a habeas petitioner to overcome the obstacle of procedural default, it may be error for a district court to refuse funding. Congress has made it clear, however, that district courts have broad discretion in assessing requests for funding. Section 3599’s predecessor declared that district courts “shall authorize” funding for services deemed “reasonably necessary.” 21 U. S. C. §848(q)(9) (1988 ed.). Applying this provision, courts of appeals reviewed district court funding decisions for abuse of discretion. E.g., Bonin v. Calderon, 59 F. 3d 815, 837 (CA9 1995); In re Lindsey, 875 F. 2d 1502, 1507, n. 4 (CA11 1989); United States v. Alden, 767 F. 2d 314, 319 (CA7 1984). Then, as part of the Antiterrorism and Effective Death Penalty Act of 1996, 110Stat. 1226, Congress changed the verb from “shall” to “may,” and thus made it perfectly clear that determining whether funding is “reasonably necessary” is a decision as to which district courts enjoy broad discretion. See Kingdomware Technologies, Inc. v. United States, 579 U. S. ___, ___ (2016) (slip op., at 9). A natural consideration informing the exercise of that discretion is the likelihood that the contemplated services will help the applicant win relief. After all, the proposed services must be “reasonably necessary” for the applicant’s representation, and it would not be reasonable—in fact, it would be quite unreasonable—to think that services are necessary to the applicant’s representation if, realistically speaking, they stand little hope of helping him win relief. Proper application of the “reasonably necessary” standard thus requires courts to consider the potential merit of the claims that the applicant wants to pursue, the likelihood that the services will generate useful and admissible evidence, and the prospect that the applicant will be able to clear any procedural hurdles standing in the way. To be clear, a funding applicant must not be expected to prove that he will be able to win relief if given the services he seeks. But the “reasonably necessary” test requires an assessment of the likely utility of the services requested, and §3599(f ) cannot be read to guarantee that an applicant will have enough money to turn over every stone. Petitioner does not deny this. He agrees that an applicant must “articulat[e] specific reasons why the services are warranted”—which includes demonstrating that the underlying claim is at least “ ‘plausible’ ”—and he acknowledges that there may even be cases in which it would be within a court’s discretion to “deny funds after a finding of ‘reasonable necessity.’ ” Brief for Petitioner 43. These interpretive principles are consistent with the way in which §3599’s predecessors were read by the lower courts. See, e.g., Alden, supra, at 318–319 (explaining that it was “appropriate for the district court to satisfy itself that [the] defendant may have a plausible defense before granting the defendant’s . . . motion for psychiatric assistance to aid in that defense,” and that it is not proper to use the funding statute to subsidize a “ ‘fishing expedition’ ”); United States v. Hamlet, 480 F. 2d 556, 557 (CA5 1973) (per curiam) (upholding District Court’s refusal to fund psychiatric services based on the District Court’s conclusion that “the request for psychiatric services was . . . lacking in merit” because there was “no serious possibility that appellant was legally insane at any time pertinent to the crimes committed”). This abundance of precedent shows courts have plenty of experience making the determinations that §3599(f) contemplates. IV Perhaps anticipating that we might not accept the Fifth Circuit’s reading of §3599(f), respondent devotes a substantial portion of her brief to an alternative ground for affirmance that was neither presented nor passed on below. Respondent contends that whatever “reasonably necessary” means, funding is never “reasonably necessary” in a case like this one, where a habeas petitioner seeks to present a procedurally defaulted ineffective-assistance-of-trial-counsel claim that depends on facts outside the state-court record. Citing 28 U. S. C. §2254(e)(2), respondent contends that the fruits of any such investigation would be inadmissible in a federal habeas court. We decline to decide in the first instance whether respondent’s reading of §2254(e)(2) is correct. Petitioner agrees that the argument remains open for the Fifth Circuit to consider on remand. Tr. of Oral Arg. 6. * * * We conclude that the Fifth Circuit’s interpretation of §3599(f ) is not a permissible reading of the statute. We therefore vacate the judgment below and remand the case for further proceedings consistent with this opinion. It is so ordered. Notes 1 We also consider a jurisdictional issue not raised by the parties, namely, whether we have jurisdiction even though no COA has yet been issued. We do not have jurisdiction if jurisdiction was lacking in the Court of Appeals, and the jurisdiction of a court of appeals to entertain an appeal from a final order in a habeas proceeding is dependent on the issuance of a COA. See 28 U. S. C. §2253(c)(l); Gonzalez v. Thaler, 565 U. S. 134, 142 (2012) . In this case, petitioner appealed an order of the District Court that denied both his request for funding under 18 U. S. C. §3599 and his underlying habeas claims. The Court of Appeals denied a COA as to the merits of his request for habeas relief but held that a COA was not required insofar as petitioner challenged the District Court’s denial of funding under §3599. The Fifth Circuit relied on Harbison v. Bell, 556 U. S. 180 (2009) , in which a prisoner appealed from an order that denied counsel under §3599 for a state clemency proceeding but that did not address the merits of any habeas petition. This Court held that a COA was not required. Here, petitioner took his appeal from the final order in his habeas proceeding. The parties have not briefed whether that difference between Harbison and the present case is relevant or whether an appeal from a denial of a §3599 request for funding would fit within the COA framework, and we find it unnecessary to resolve the issue. Though we take no view on the merits, we will assume for the sake of argument that the Court of Appeals could not entertain petitioner’s §3599 claim without the issuance of a COA. We may review the denial of a COA by the lower courts. See, e.g., Miller-El v. Cockrell, 537 U. S. 322 –327 (2003). When the lower courts deny a COA and we conclude that their reason for doing so was flawed, we may reverse and remand so that the correct legal standard may be applied. See Slack v. McDaniel, 529 U. S. 473 –486, 489–490 (2000). We take that course here. As we will explain, the correctness of the rule applied by the District Court in denying the funding request was not only debatable; it was erroneous. 2 In relevant part §1291 declares that “[t]he courts of appeals . . . shall have jurisdiction of appeals from all final decisions of the district courts of the United States, the United States District Court for the District of the Canal Zone, the District Court of Guam, and the District Court of the Virgin Islands, except where a direct review may be had in the Supreme Court.” Similarly, §2253 provides, as relevant, that “[i]n a habeas corpus proceeding or a proceeding under section 2255 before a district judge, the final order shall be subject to review, on appeal, by the court of appeals for the circuit in which the proceeding is held.” §2253(a). 3 “Cases in the courts of appeals may be reviewed by the Supreme Court by . . . writ of certiorari granted upon the petition of any party to any civil or criminal case, before or after rendition of judgment or decree.” §1254(1). 4 This language now appears at 18 U. S. C. §3006A(i) (2012 ed.). 5 As far as we are aware, neither the Administrative Office nor any other nonjudicial entity currently claims the power to revise or reject a CJA compensation order issued by a court. Nothing in the CJA Guidelines suggests such a policy. See generally 7A Guide to Judiciary Policy (May 17, 2017).

584.US.2017_16-1371

Latasha Reed rented a car in New Jersey while petitioner Terrence Byrd waited outside the rental facility. Her signed agreement warned that permitting an unauthorized driver to drive the car would violate the agreement. Reed listed no additional drivers on the form, but she gave the keys to Byrd upon leaving the building. He stored personal belongings in the rental car’s trunk and then left alone for Pittsburgh, Pennsylvania. After stopping Byrd for a traffic infraction, Pennsylvania State Troopers learned that the car was rented, that Byrd was not listed as an authorized driver, and that Byrd had prior drug and weapons convictions. Byrd also stated he had a marijuana cigarette in the car. The troopers proceeded to search the car, discovering body armor and 49 bricks of heroin in the trunk. The evidence was turned over to federal authorities, who charged Byrd with federal drug and other crimes. The District Court denied Byrd’s motion to suppress the evidence as the fruit of an unlawful search, and the Third Circuit affirmed. Both courts concluded that, because Byrd was not listed on the rental agreement, he lacked a reasonable expectation of privacy in the car. Held: 1. The mere fact that a driver in lawful possession or control of a rental car is not listed on the rental agreement will not defeat his or her otherwise reasonable expectation of privacy. Pp. 6–13. (a) Reference to property concepts is instructive in “determining the presence or absence of the privacy interests protected by [the Fourth] Amendment.” Rakas v. Illinois, 439 U. S. 128, 144, n. 12. Pp. 6–7. (b) While a person need not always have a recognized common-law property interest in the place searched to be able to claim a reasonable expectation of privacy in it, see, e.g., Jones v. United States, 362 U. S. 257, 259, legitimate presence on the premises, standing alone, is insufficient because it “creates too broad a gauge for measurement of Fourth Amendment rights,” Rakas, 439 U. S., at 142. The Court has not set forth a single metric or exhaustive list of relevant considerations, but “[l]egitimation of expectations of privacy must have a source outside of the Fourth Amendment, either by reference to concepts of real or personal property law or to understandings that are recognized and permitted by society.” Id., at 144, n. 12. These concepts may be linked. “One of the main rights attaching to property is the right to exclude others,” and “one who owns or lawfully possesses or controls property will in all likelihood have a legitimate expectation of privacy by virtue of the right to exclude.” Ibid. This general property-based concept guides resolution of the instant case. Pp. 8–9. (c) The Government’s contention that drivers who are not listed on rental agreements always lack an expectation of privacy in the car rests on too restrictive a view of the Fourth Amendment’s protections. But Byrd’s proposal that a rental car’s sole occupant always has an expectation of privacy based on mere possession and control would, without qualification, include thieves or others who have no reasonable expectation of privacy. Pp. 9–13. (1) The Government bases its claim that an unauthorized driver has no privacy interest in the vehicle on a misreading of Rakas. There, the Court disclaimed any intent to hold that passengers cannot have an expectation of privacy in automobiles, but found that the passengers there had not claimed “any legitimate expectation of privacy in the areas of the car which were searched.” 439 U. S., at 150, n. 17. Byrd, in contrast, was the rental car’s driver and sole occupant. His situation is similar to the defendant in Jones, who had a reasonable expectation of privacy in his friend’s apartment because he “had complete dominion and control over the apartment and could exclude others from it.” Rakas, supra, at 149. The expectation of privacy that comes from lawful possession and control and the attendant right to exclude should not differ depending on whether a car is rented or owned by someone other than the person currently possessing it, much as it did not seem to matter whether the defendant’s friend in Jones owned or leased the apartment he permitted the defendant to use in his absence. Pp. 9–11. (2) The Government also contends that Byrd had no basis for claiming an expectation of privacy in the rental car because his driving of that car was so serious a breach of Reed’s rental agreement that the rental company would have considered the agreement “void” once he took the wheel. But the contract says only that the violation may result in coverage, not the agreement, being void and the renter’s being fully responsible for any loss or damage, and the Government fails to explain what bearing this breach of contract, standing alone, has on expectations of privacy in the car. Pp. 11–12. (3) Central, though, to reasonable expectations of privacy in these circumstances is the concept of lawful possession, for a “ ‘wrongful’ presence at the scene of a search would not enable a defendant to object to the legality of the search,” Rakas, supra, at 141, n. 9. Thus, a car thief would not have a reasonable expectation of privacy in a stolen car no matter the degree of possession and control. The Court leaves for remand the Government’s argument that one who intentionally uses a third party to procure a rental car by a fraudulent scheme for the purpose of committing a crime is no better situated than a car thief. Pp. 12–13. 2. Also left for remand is the Government’s argument that, even if Byrd had a right to object to the search, probable cause justified it in any event. The Third Circuit did not reach this question because it concluded, as an initial matter, that Byrd lacked a reasonable expectation of privacy in the rental car. That court has discretion as to the order in which the remanded questions are best addressed. Pp. 13–14. 679 Fed. Appx. 146, vacated and remanded. Kennedy, J., delivered the opinion for a unanimous Court. Thomas, J., filed a concurring opinion, in which Gorsuch, J., joined. Alito, J., filed a concurring opinion.

In September 2014, Pennsylvania State Troopers pulled over a car driven by petitioner Terrence Byrd. Byrd was the only person in the car. In the course of the traffic stop the troopers learned that the car was rented and that Byrd was not listed on the rental agreement as an authorized driver. For this reason, the troopers told Byrd they did not need his consent to search the car, including its trunk where he had stored personal effects. A search of the trunk uncovered body armor and 49 bricks of heroin. The evidence was turned over to federal authorities, who charged Byrd with distribution and possession of heroin with the intent to distribute in violation of 21 U. S. C. §841(a)(1) and possession of body armor by a prohibited person in violation of 18 U. S. C. §931(a)(1). Byrd moved to suppress the evidence as the fruit of an unlawful search. The United States District Court for the Middle District of Pennsylvania denied the motion, and the Court of Appeals for the Third Circuit affirmed. Both courts concluded that, because Byrd was not listed on the rental agreement, he lacked a reasonable expectation of privacy in the car. Based on this conclusion, it appears that both the District Court and Court of Appeals deemed it unnecessary to consider whether the troopers had probable cause to search the car. This Court granted certiorari to address the question whether a driver has a reasonable expectation of privacy in a rental car when he or she is not listed as an authorized driver on the rental agreement. The Court now holds that, as a general rule, someone in otherwise lawful possession and control of a rental car has a reasonable expectation of privacy in it even if the rental agreement does not list him or her as an authorized driver. The Court concludes a remand is necessary to address in the first instance the Government’s argument that this general rule is inapplicable because, in the circumstances here, Byrd had no greater expectation of privacy than a car thief. If that is so, our cases make clear he would lack a legitimate expectation of privacy. It is necessary to remand as well to determine whether, even if Byrd had a right to object to the search, probable cause justified it in any event. I On September 17, 2014, petitioner Terrence Byrd and Latasha Reed drove in Byrd’s Honda Accord to a Budget car-rental facility in Wayne, New Jersey. Byrd stayed in the parking lot in the Honda while Reed went to the Budget desk and rented a Ford Fusion. The agreement Reed signed required her to certify that she had a valid driver’s license and had not committed certain vehicle-related offenses within the previous three years. An addendum to the agreement, which Reed initialed, provides the following restriction on who may drive the rental car: “I understand that the only ones permitted to drive the vehicle other than the renter are the renter’s spouse, the renter’s co-employee (with the renter’s permission, while on company business), or a person who appears at the time of the rental and signs an Additional Driver Form. These other drivers must also be at least 25 years old and validly licensed. “PERMITTING AN UNAUTHORIZED DRIVER TO OPERATE THE VEHICLE IS A VIOLATION OF THE RENTAL AGREEMENT. THIS MAY RESULT IN ANY AND ALL COVERAGE OTHERWISE PROVIDED BY THE RENTAL AGREEMENT BEING VOID AND MY BEING FULLY RESPONSIBLE FOR ALL LOSS OR DAMAGE, INCLUDING LIABILITY TO THIRD PARTIES.” App. 19. In filling out the paperwork for the rental agreement, Reed did not list an additional driver. With the rental keys in hand, Reed returned to the parking lot and gave them to Byrd. The two then left the facility in separate cars—she in his Honda, he in the rental car. Byrd returned to his home in Patterson, New Jersey, and put his personal belongings in the trunk of the rental car. Later that afternoon, he departed in the car alone and headed toward Pittsburgh, Pennsylvania. After driving nearly three hours, or roughly half the distance to Pittsburgh, Byrd passed State Trooper David Long, who was parked in the median of Interstate 81 near Harrisburg, Pennsylvania. Long was suspicious of Byrd because he was driving with his hands at the “10 and 2” position on the steering wheel, sitting far back from the steering wheel, and driving a rental car. Long knew the Ford Fusion was a rental car because one of its windows contained a barcode. Based on these observations, he decided to follow Byrd and, a short time later, stopped him for a possible traffic infraction. When Long approached the passenger window of Byrd’s car to explain the basis for the stop and to ask for identification, Byrd was “visibly nervous” and “was shaking and had a hard time obtaining his driver’s license.” Id., at 37. He handed an interim license and the rental agreement to Long, stating that a friend had rented the car. Long returned to his vehicle to verify Byrd’s license and noticed Byrd was not listed as an additional driver on the rental agreement. Around this time another trooper, Travis Martin, arrived at the scene. While Long processed Byrd’s license, Martin conversed with Byrd, who again stated that a friend had rented the vehicle. After Martin walked back to Long’s patrol car, Long commented to Martin that Byrd was “not on the renter agreement,” to which Martin replied, “yeah, he has no expectation of privacy.” 3 App. to Brief for Appellant in No. 16–1509 (CA3), at 21:40. A computer search based on Byrd’s identification returned two different names. Further inquiry suggested the other name might be an alias and also revealed that Byrd had prior convictions for weapons and drug charges as well as an outstanding warrant in New Jersey for a probation violation. After learning that New Jersey did not want Byrd arrested for extradition, the troopers asked Byrd to step out of the vehicle and patted him down. Long asked Byrd if he had anything illegal in the car. When Byrd said he did not, the troopers asked for his consent to search the car. At that point Byrd said he had a “blunt” in the car and offered to retrieve it for them. The officers understood “blunt” to mean a marijuana cigarette. They declined to let him retrieve it and continued to seek his consent to search the car, though they stated they did not need consent because he was not listed on the rental agreement. The troopers then opened the passenger and driver doors and began a thorough search of the passenger compartment. Martin proceeded from there to search the car’s trunk, including by opening up and taking things out of a large cardboard box, where he found a laundry bag containing body armor. At this point, the troopers decided to detain Byrd. As Martin walked toward Byrd and said he would be placing him in handcuffs, Byrd began to run away. A third trooper who had arrived on the scene joined Long and Martin in pursuit. When the troopers caught up to Byrd, he surrendered and admitted there was heroin in the car. Back at the car, the troopers resumed their search of the laundry bag and found 49 bricks of heroin. In pretrial proceedings Byrd moved to suppress the evidence found in the trunk of the rental car, arguing that the search violated his Fourth Amendment rights. Al- though Long contended at a suppression hearing that the troopers had probable cause to search the car after Byrd stated it contained marijuana, the District Court denied Byrd’s motion on the ground that Byrd lacked “standing” to contest the search as an initial matter, 2015 WL 5038455, *2 (MD Pa., Aug. 26, 2015) (citing United States v. Kennedy, 638 F. 3d 159, 165 (CA3 2011)). Byrd later entered a conditional guilty plea, reserving the right to appeal the suppression ruling. The Court of Appeals affirmed in a brief summary opinion. 679 Fed. Appx. 146 (CA3 2017). As relevant here, the Court of Appeals recognized that a “circuit split exists as to whether the sole occupant of a rental vehicle has a Fourth Amendment expectation of privacy when that occupant is not named in the rental agreement”; but it noted that Circuit precedent already had “spoken as to this issue . . . and determined such a person has no expectation of privacy and therefore no standing to challenge a search of the vehicle.” Id., at 150 (citing Kennedy, supra, at 167–168). The Court of Appeals did not reach the probable-cause question. This Court granted Byrd’s petition for a writ of certio- rari, 582 U. S. ___ (2017), to address the conflict among the Courts of Appeals over whether an unauthorized driver has a reasonable expectation of privacy in a rental car. Compare United States v. Seeley, 331 F. 3d 471, 472 (CA5 2003) (per curiam); United States v. Wellons, 32 F. 3d 117, 119 (CA4 1994); United States v. Roper, 918 F. 2d 885, 887–888 (CA10 1990), with United States v. Smith, 263 F. 3d 571, 581–587 (CA6 2001); Kennedy, supra, at 165–168, and with United States v. Thomas, 447 F. 3d 1191, 1196–1199 (CA9 2006); United States v. Best, 135 F. 3d 1223, 1225 (CA8 1998). II Few protections are as essential to individual liberty as the right to be free from unreasonable searches and seizures. The Framers made that right explicit in the Bill of Rights following their experience with the indignities and invasions of privacy wrought by “general warrants and warrantless searches that had so alienated the colonists and had helped speed the movement for independence.” Chimel v. California, 395 U. S. 752, 761 (1969). Ever mindful of the Fourth Amendment and its history, the Court has viewed with disfavor practices that permit “police officers unbridled discretion to rummage at will among a person’s private effects.” Arizona v. Gant, 556 U. S. 332, 345 (2009). This concern attends the search of an automobile. See Delaware v. Prouse, 440 U. S. 648, 662 (1979). The Court has acknowledged, however, that there is a diminished expectation of privacy in automobiles, which often permits officers to dispense with obtaining a warrant before conducting a lawful search. See, e.g., California v. Acevedo, 500 U. S. 565, 579 (1991). Whether a warrant is required is a separate question from the one the Court addresses here, which is whether the person claiming a constitutional violation “has had his own Fourth Amendment rights infringed by the search and seizure which he seeks to challenge.” Rakas v. Illinois, 439 U. S. 128, 133 (1978). Answering that question requires examination of whether the person claiming the constitutional violation had a “legitimate expectation of privacy in the premises” searched. Id., at 143. “Expectations of privacy protected by the Fourth Amendment, of course, need not be based on a common-law interest in real or personal property, or on the invasion of such an interest.” Id., at 144, n. 12. Still, “property concepts” are instructive in “determining the presence or absence of the privacy interests protected by that Amendment.” Ibid. Indeed, more recent Fourth Amendment cases have clarified that the test most often associated with legitimate expectations of privacy, which was derived from the second Justice Harlan’s concurrence in Katz v. United States, 389 U. S. 347 (1967), supplements, rather than displaces, “the traditional property-based understanding of the Fourth Amendment.” Florida v. Jardines, 569 U. S. 1, 11 (2013). Perhaps in light of this clarification, Byrd now argues in the alternative that he had a common-law property interest in the rental car as a second bailee that would have provided him with a cognizable Fourth Amendment interest in the vehicle. But he did not raise this argument before the District Court or Court of Appeals, and those courts did not have occasion to address whether Byrd was a second bailee or what consequences might follow from that determination. In those courts he framed the question solely in terms of the Katz test noted above. Because this is “a court of review, not of first view,” Cutter v. Wilkinson, 544 U. S. 709, 718, n. 7 (2005), it is generally unwise to consider arguments in the first instance, and the Court declines to reach Byrd’s contention that he was a second bailee. Reference to property concepts, however, aids the Court in assessing the precise question here: Does a driver of a rental car have a reasonable expectation of privacy in the car when he or she is not listed as an authorized driver on the rental agreement? III A One who owns and possesses a car, like one who owns and possesses a house, almost always has a reasonable expectation of privacy in it. More difficult to define and delineate are the legitimate expectations of privacy of others. On the one hand, as noted above, it is by now well established that a person need not always have a recognized common-law property interest in the place searched to be able to claim a reasonable expectation of privacy in it. See Jones v. United States, 362 U. S. 257, 259 (1960); Katz, supra, at 352; Mancusi v. DeForte, 392 U. S. 364, 368 (1968); Minnesota v. Olson, 495 U. S. 91, 98 (1990). On the other hand, it is also clear that legitimate presence on the premises of the place searched, standing alone, is not enough to accord a reasonable expectation of privacy, because it “creates too broad a gauge for measurement of Fourth Amendment rights.” Rakas, 439 U. S., at 142; see also id., at 148 (“We would not wish to be understood as saying that legitimate presence on the premises is irrelevant to one’s expectation of privacy, but it cannot be deemed controlling”); Minnesota v. Carter, 525 U. S. 83, 91 (1998). Although the Court has not set forth a single metric or exhaustive list of considerations to resolve the circumstances in which a person can be said to have a reasonable expectation of privacy, it has explained that “[l]egitimation of expectations of privacy by law must have a source outside of the Fourth Amendment, either by reference to concepts of real or personal property law or to understandings that are recognized and permitted by society.” Rakas, 439 U. S., at 144, n. 12. The two concepts in cases like this one are often linked. “One of the main rights attaching to property is the right to exclude others,” and, in the main, “one who owns or lawfully possesses or controls property will in all likelihood have a legitimate expectation of privacy by virtue of the right to exclude.” Ibid. (citing 2 W. Blackstone, Commentaries on the Laws of England, ch. 1). This general property-based concept guides resolution of this case. B Here, the Government contends that drivers who are not listed on rental agreements always lack an expectation of privacy in the automobile based on the rental company’s lack of authorization alone. This per se rule rests on too restrictive a view of the Fourth Amendment’s protections. Byrd, by contrast, contends that the sole occupant of a rental car always has an expectation of privacy in it based on mere possession and control. There is more to recommend Byrd’s proposed rule than the Government’s; but, without qualification, it would include within its ambit thieves and others who, not least because of their lack of any property-based justification, would not have a reasonable expectation of privacy. 1 Stripped to its essentials, the Government’s position is that only authorized drivers of rental cars have expectations of privacy in those vehicles. This position is based on the following syllogism: Under Rakas, passengers do not have an expectation of privacy in an automobile glove compartment or like places; an unauthorized driver like Byrd would have been the passenger had the renter been driving; and the unauthorized driver cannot obtain greater protection when he takes the wheel and leaves the renter behind. The flaw in this syllogism is its major premise, for it is a misreading of Rakas. The Court in Rakas did not hold that passengers cannot have an expectation of privacy in automobiles. To the contrary, the Court disclaimed any intent to hold “that a passenger lawfully in an automobile may not invoke the exclusionary rule and challenge a search of that vehicle unless he happens to own or have a possessory interest in it.” 439 U. S., at 150, n. 17 (internal quotation marks omitted). The Court instead rejected the argument that legitimate presence alone was sufficient to assert a Fourth Amendment interest, which was fatal to the petitioners’ case there because they had “claimed only that they were ‘legitimately on [the] premises’ and did not claim that they had any legitimate expectation of privacy in the areas of the car which were searched.” Ibid. What is more, the Government’s syllogism is beside the point, because this case does not involve a passenger at all but instead the driver and sole occupant of a rental car. As Justice Powell observed in his concurring opinion in Rakas, a “distinction . . . may be made in some circumstances between the Fourth Amendment rights of passengers and the rights of an individual who has exclusive control of an automobile or of its locked compartments.” Id., at 154. This situation would be similar to the defendant in Jones, supra, who, as Rakas notes, had a reasonable expectation of privacy in his friend’s apartment because he “had complete dominion and control over the apartment and could exclude others from it,” 439 U. S., at 149. Justice Powell’s observation was also consistent with the majority’s explanation that “one who owns or lawfully possesses or controls property will in all likelihood have a legitimate expectation of privacy by virtue of [the] right to exclude,” id., at 144, n. 12, an explanation tied to the majority’s discussion of Jones. The Court sees no reason why the expectation of privacy that comes from lawful possession and control and the attendant right to exclude would differ depending on whether the car in question is rented or privately owned by someone other than the person in current possession of it, much as it did not seem to matter whether the friend of the defendant in Jones owned or leased the apartment he permitted the defendant to use in his absence. Both would have the expectation of privacy that comes with the right to exclude. Indeed, the Government conceded at oral argument that an unauthorized driver in sole possession of a rental car would be permitted to exclude third parties from it, such as a carjacker. Tr. of Oral Arg. 48–49. 2 The Government further stresses that Byrd’s driving the rental car violated the rental agreement that Reed signed, and it contends this violation meant Byrd could not have had any basis for claiming an expectation of privacy in the rental car at the time of the search. As anyone who has rented a car knows, car-rental agreements are filled with long lists of restrictions. Examples include prohibitions on driving the car on unpaved roads or driving while using a handheld cellphone. Few would contend that violating provisions like these has anything to do with a driver’s reasonable expectation of privacy in the rental car—as even the Government agrees. Brief for United States 32. Despite this concession, the Government argues that permitting an unauthorized driver to take the wheel of a rental car is a breach different in kind from these others, so serious that the rental company would consider the agreement “void” the moment an unauthorized driver takes the wheel. Id., at 4, 15, 16, 27. To begin with, that is not what the contract says. It states: “Permitting an unauthorized driver to operate the vehicle is a violation of the rental agreement. This may result in any and all coverage otherwise provided by the rental agreement being void and my being fully responsible for all loss or damage, including liability to third parties.” App. 24 (emphasis deleted). Putting the Government’s misreading of the contract aside, there may be countless innocuous reasons why an unauthorized driver might get behind the wheel of a rental car and drive it—perhaps the renter is drowsy or inebriated and the two think it safer for the friend to drive them to their destination. True, this constitutes a breach of the rental agreement, and perhaps a serious one, but the Government fails to explain what bearing this breach of contract, standing alone, has on expectations of privacy in the car. Stated in different terms, for Fourth Amendment purposes there is no meaningful difference between the authorized-driver provision and the other provisions the Government agrees do not eliminate an expectation of privacy, all of which concern risk allocation between private parties—violators might pay additional fees, lose insurance coverage, or assume liability for damage resulting from the breach. But that risk allocation has little to do with whether one would have a reasonable expectation of privacy in the rental car if, for example, he or she other- wise has lawful possession of and control over the car. 3 The central inquiry at this point turns on the concept of lawful possession, and this is where an important qualification of Byrd’s proposed rule comes into play. Rakas makes clear that “ ‘wrongful’ presence at the scene of a search would not enable a defendant to object to the legal- ity of the search.” 439 U. S., at 141, n. 9. “A burglar plying his trade in a summer cabin during the off season,” for example, “may have a thoroughly justified subjective expectation of privacy, but it is not one which the law recognizes as ‘legitimate.’ ” Id., at 143, n. 12. Likewise, “a person present in a stolen automobile at the time of the search may [not] object to the lawfulness of the search of the automobile.” Id., at 141, n. 9. No matter the degree of possession and control, the car thief would not have a reasonable expectation of privacy in a stolen car. On this point, in its merits brief, the Government asserts that, on the facts here, Byrd should have no greater expectation of privacy than a car thief because he intentionally used a third party as a strawman in a calculated plan to mislead the rental company from the very outset, all to aid him in committing a crime. This argument is premised on the Government’s inference that Byrd knew he would not have been able to rent the car on his own, because he would not have satisfied the rental company’s requirements based on his criminal record, and that he used Reed, who had no intention of using the car for her own purposes, to procure the car for him to transport heroin to Pittsburgh. It is unclear whether the Government’s allegations, if true, would constitute a criminal offense in the acquisition of the rental car under applicable law. And it may be that there is no reason that the law should distinguish between one who obtains a vehicle through subterfuge of the type the Government alleges occurred here and one who steals the car outright. The Government did not raise this argument in the District Court or the Court of Appeals, however. It relied instead on the sole fact that Byrd lacked authorization to drive the car. And it is unclear from the record whether the Government’s inferences paint an accurate picture of what occurred. Because it was not addressed in the District Court or Court of Appeals, the Court declines to reach this question. The proper course is to remand for the argument and potentially further factual development to be considered in the first instance by the Court of Appeals or by the District Court. IV The Government argued in its brief in opposition to certiorari that, even if Byrd had a Fourth Amendment interest in the rental car, the troopers had probable cause to believe it contained evidence of a crime when they initiated their search. If that were true, the troopers may have been permitted to conduct a warrantless search of the car in line with the Court’s cases concerning the automobile exception to the warrant requirement. See, e.g., Acevedo, 500 U. S., at 580. The Court of Appeals did not reach this question because it concluded, as an initial matter, that Byrd lacked a reasonable expectation of privacy in the rental car. It is worth noting that most courts analyzing the question presented in this case, including the Court of Appeals here, have described it as one of Fourth Amendment “standing,” a concept the Court has explained is not distinct from the merits and “is more properly subsumed under substantive Fourth Amendment doctrine.” Rakas, supra, at 139. The concept of standing in Fourth Amendment cases can be a useful shorthand for capturing the idea that a person must have a cognizable Fourth Amendment interest in the place searched before seeking relief for an unconstitutional search; but it should not be confused with Article III standing, which is jurisdictional and must be assessed before reaching the merits. Arizona Christian School Tuition Organization v. Winn, 563 U. S. 125, 129 (2011) (“To obtain a determination on the merits in federal court, parties seeking relief must show that they have standing under Article III of the Constitution”); see also Rakas, supra, at 138–140. Because Fourth Amendment standing is subsumed under substantive Fourth Amendment doctrine, it is not a jurisdictional question and hence need not be addressed before addressing other aspects of the merits of a Fourth Amendment claim. On remand, then, the Court of Appeals is not required to assess Byrd’s reason- able expectation of privacy in the rental car before, in its discretion, first addressing whether there was probable cause for the search, if it finds the latter argument has been preserved. V Though new, the fact pattern here continues a well-traveled path in this Court’s Fourth Amendment jurisprudence. Those cases support the proposition, and the Court now holds, that the mere fact that a driver in lawful possession or control of a rental car is not listed on the rental agreement will not defeat his or her otherwise reasonable expectation of privacy. The Court leaves for remand two of the Government’s arguments: that one who intention- ally uses a third party to procure a rental car by a fraudulent scheme for the purpose of committing a crime is no better situated than a car thief; and that probable cause justified the search in any event. The Court of Appeals has discretion as to the order in which these questions are best addressed. * * * The judgment of the Court of Appeals is vacated, and the case is remanded for further proceedings consistent with this opinion. It is so ordered.

585.US.2017_16-402

Cell phones perform their wide and growing variety of functions by continuously connecting to a set of radio antennas called “cell sites.” Each time a phone connects to a cell site, it generates a time-stamped record known as cell-site location information (CSLI). Wireless carriers collect and store this information for their own business purposes. Here, after the FBI identified the cell phone numbers of several robbery suspects, prosecutors were granted court orders to obtain the suspects’ cell phone records under the Stored Communications Act. Wireless carriers produced CSLI for petitioner Timothy Carpenter’s phone, and the Government was able to obtain 12,898 location points cataloging Carpenter’s movements over 127 days—an average of 101 data points per day. Carpenter moved to suppress the data, arguing that the Government’s seizure of the records without obtaining a warrant supported by probable cause violated the Fourth Amendment. The District Court denied the motion, and prosecutors used the records at trial to show that Carpenter’s phone was near four of the robbery locations at the time those robberies occurred. Carpenter was convicted. The Sixth Circuit affirmed, holding that Carpenter lacked a reasonable expectation of privacy in the location information collected by the FBI because he had shared that information with his wireless carriers. Held: 1. The Government’s acquisition of Carpenter’s cell-site records was a Fourth Amendment search. Pp. 4–18. (a) The Fourth Amendment protects not only property interests but certain expectations of privacy as well. Katz v. United States, 389 U. S. 347, 351. Thus, when an individual “seeks to preserve something as private,” and his expectation of privacy is “one that society is prepared to recognize as reasonable,” official intrusion into that sphere generally qualifies as a search and requires a warrant supported by probable cause. Smith v. Maryland, 442 U. S. 735, 740 (internal quotation marks and alterations omitted). The analysis regarding which expectations of privacy are entitled to protection is informed by historical understandings “of what was deemed an unreasonable search and seizure when [the Fourth Amendment] was adopted.” Carroll v. United States, 267 U. S. 132, 149. These Founding-era understandings continue to inform this Court when applying the Fourth Amendment to innovations in surveillance tools. See, e.g., Kyllo v. United States, 533 U. S. 27. Pp. 4–7. (b) The digital data at issue—personal location information maintained by a third party—does not fit neatly under existing precedents but lies at the intersection of two lines of cases. One set addresses a person’s expectation of privacy in his physical location and movements. See, e.g., United States v. Jones, 565 U. S. 400 (five Justices concluding that privacy concerns would be raised by GPS tracking). The other addresses a person’s expectation of privacy in information voluntarily turned over to third parties. See United States v. Miller, 425 U. S. 435 (no expectation of privacy in financial records held by a bank), and Smith, 442 U. S. 735 (no expectation of privacy in records of dialed telephone numbers conveyed to telephone company). Pp. 7–10. (c) Tracking a person’s past movements through CSLI partakes of many of the qualities of GPS monitoring considered in Jones—it is detailed, encyclopedic, and effortlessly compiled. At the same time, however, the fact that the individual continuously reveals his location to his wireless carrier implicates the third-party principle of Smith and Miller. Given the unique nature of cell-site records, this Court declines to extend Smith and Miller to cover them. Pp. 10–18. (1) A majority of the Court has already recognized that individuals have a reasonable expectation of privacy in the whole of their physical movements. Allowing government access to cell-site records—which “hold for many Americans the ‘privacies of life,’ ” Riley v. California, 573 U. S. ___, ___—contravenes that expectation. In fact, historical cell-site records present even greater privacy concerns than the GPS monitoring considered in Jones: They give the Government near perfect surveillance and allow it to travel back in time to retrace a person’s whereabouts, subject only to the five-year retention policies of most wireless carriers. The Government contends that CSLI data is less precise than GPS information, but it thought the data accurate enough here to highlight it during closing argument in Carpenter’s trial. At any rate, the rule the Court adopts “must take account of more sophisticated systems that are already in use or in development,” Kyllo, 533 U. S., at 36, and the accuracy of CSLI is rapidly approaching GPS-level precision. Pp. 12–15. (2) The Government contends that the third-party doctrine governs this case, because cell-site records, like the records in Smith and Miller, are “business records,” created and maintained by wireless carriers. But there is a world of difference between the limited types of personal information addressed in Smith and Miller and the exhaustive chronicle of location information casually collected by wireless carriers. The third-party doctrine partly stems from the notion that an individual has a reduced expectation of privacy in information knowingly shared with another. Smith and Miller, however, did not rely solely on the act of sharing. They also considered “the nature of the particular documents sought” and limitations on any “legitimate ‘expectation of privacy’ concerning their contents.” Miller, 425 U. S., at 442. In mechanically applying the third-party doctrine to this case the Government fails to appreciate the lack of comparable limitations on the revealing nature of CSLI. Nor does the second rationale for the third-party doctrine—voluntary exposure—hold up when it comes to CSLI. Cell phone location information is not truly “shared” as the term is normally understood. First, cell phones and the services they provide are “such a pervasive and insistent part of daily life” that carrying one is indispensable to participation in modern society. Riley, 573 U. S., at ___. Second, a cell phone logs a cell-site record by dint of its operation, without any affirmative act on the user’s part beyond powering up. Pp. 15–17. (d) This decision is narrow. It does not express a view on matters not before the Court; does not disturb the application of Smith and Miller or call into question conventional surveillance techniques and tools, such as security cameras; does not address other business records that might incidentally reveal location information; and does not consider other collection techniques involving foreign affairs or national security. Pp. 17–18. 2. The Government did not obtain a warrant supported by probable cause before acquiring Carpenter’s cell-site records. It acquired those records pursuant to a court order under the Stored Communications Act, which required the Government to show “reasonable grounds” for believing that the records were “relevant and material to an ongoing investigation.” 18 U. S. C. §2703(d). That showing falls well short of the probable cause required for a warrant. Consequently, an order issued under §2703(d) is not a permissible mechanism for accessing historical cell-site records. Not all orders compelling the production of documents will require a showing of probable cause. A warrant is required only in the rare case where the suspect has a legitimate privacy interest in records held by a third party. And even though the Government will generally need a warrant to access CSLI, case-specific exceptions—e.g., exigent circumstances—may support a warrantless search. Pp. 18–22. 819 F. 3d 880, reversed and remanded. Roberts, C. J., delivered the opinion of the Court, in which Ginsburg, Breyer, Sotomayor, and Kagan, JJ., joined. Kennedy, J., filed a dissenting opinion, in which Thomas and Alito, JJ., joined. Thomas, J., filed a dissenting opinion. Alito, J., filed a dissenting opinion, in which Thomas, J., joined. Gorsuch, J., filed a dissenting opinion.

This case presents the question whether the Government conducts a search under the Fourth Amendment when it accesses historical cell phone records that provide a comprehensive chronicle of the user’s past movements. I A There are 396 million cell phone service accounts in the United States—for a Nation of 326 million people. Cell phones perform their wide and growing variety of functions by connecting to a set of radio antennas called “cell sites.” Although cell sites are usually mounted on a tower, they can also be found on light posts, flagpoles, church steeples, or the sides of buildings. Cell sites typically have several directional antennas that divide the covered area into sectors. Cell phones continuously scan their environment looking for the best signal, which generally comes from the closest cell site. Most modern devices, such as smartphones, tap into the wireless network several times a minute whenever their signal is on, even if the owner is not using one of the phone’s features. Each time the phone connects to a cell site, it generates a time-stamped record known as cell-site location information (CSLI). The precision of this information depends on the size of the geographic area covered by the cell site. The greater the concentration of cell sites, the smaller the coverage area. As data usage from cell phones has increased, wireless carriers have installed more cell sites to handle the traffic. That has led to increasingly compact coverage areas, especially in urban areas. Wireless carriers collect and store CSLI for their own business purposes, including finding weak spots in their network and applying “roaming” charges when another carrier routes data through their cell sites. In addition, wireless carriers often sell aggregated location records to data brokers, without individual identifying information of the sort at issue here. While carriers have long retained CSLI for the start and end of incoming calls, in recent years phone companies have also collected location information from the transmission of text messages and routine data connections. Accordingly, modern cell phones generate increasingly vast amounts of increasingly precise CSLI. B In 2011, police officers arrested four men suspected of robbing a series of Radio Shack and (ironically enough) T-Mobile stores in Detroit. One of the men confessed that, over the previous four months, the group (along with a rotating cast of getaway drivers and lookouts) had robbed nine different stores in Michigan and Ohio. The suspect identified 15 accomplices who had participated in the heists and gave the FBI some of their cell phone numbers; the FBI then reviewed his call records to identify additional numbers that he had called around the time of the robberies. Based on that information, the prosecutors applied for court orders under the Stored Communications Act to obtain cell phone records for petitioner Timothy Carpenter and several other suspects. That statute, as amended in 1994, permits the Government to compel the disclosure of certain telecommunications records when it “offers specific and articulable facts showing that there are reasonable grounds to believe” that the records sought “are relevant and material to an ongoing criminal investigation.” 18 U. S. C. §2703(d). Federal Magistrate Judges issued two orders directing Carpenter’s wireless carriers—MetroPCS and Sprint—to disclose “cell/site sector [information] for [Carpenter’s] telephone[ ] at call origination and at call termination for incoming and outgoing calls” during the four-month period when the string of robberies occurred. App. to Pet. for Cert. 60a, 72a. The first order sought 152 days of cell-site records from MetroPCS, which produced records spanning 127 days. The second order requested seven days of CSLI from Sprint, which produced two days of records covering the period when Carpenter’s phone was “roaming” in northeastern Ohio. Altogether the Government obtained 12,898 location points cataloging Carpenter’s movements—an average of 101 data points per day. Carpenter was charged with six counts of robbery and an additional six counts of carrying a firearm during a federal crime of violence. See 18 U. S. C. §§924(c), 1951(a). Prior to trial, Carpenter moved to suppress the cell-site data provided by the wireless carriers. He argued that the Government’s seizure of the records violated the Fourth Amendment because they had been obtained without a warrant supported by probable cause. The District Court denied the motion. App. to Pet. for Cert. 38a–39a. At trial, seven of Carpenter’s confederates pegged him as the leader of the operation. In addition, FBI agent Christopher Hess offered expert testimony about the cell-site data. Hess explained that each time a cell phone taps into the wireless network, the carrier logs a time-stamped record of the cell site and particular sector that were used. With this information, Hess produced maps that placed Carpenter’s phone near four of the charged robberies. In the Government’s view, the location records clinched the case: They confirmed that Carpenter was “right where the . . . robbery was at the exact time of the robbery.” App. 131 (closing argument). Carpenter was convicted on all but one of the firearm counts and sentenced to more than 100 years in prison. The Court of Appeals for the Sixth Circuit affirmed. 819 F. 3d 880 (2016). The court held that Carpenter lacked a reasonable expectation of privacy in the location information collected by the FBI because he had shared that information with his wireless carriers. Given that cell phone users voluntarily convey cell-site data to their carriers as “a means of establishing communication,” the court concluded that the resulting business records are not entitled to Fourth Amendment protection. Id., at 888 (quoting Smith v. Maryland, 442 U. S. 735, 741 (1979)). We granted certiorari. 582 U. S. ___ (2017). II A The Fourth Amendment protects “[t]he right of the people to be secure in their persons, houses, papers, and effects, against unreasonable searches and seizures.” The “basic purpose of this Amendment,” our cases have recognized, “is to safeguard the privacy and security of individuals against arbitrary invasions by governmental officials.” Camara v. Municipal Court of City and County of San Francisco, 387 U. S. 523, 528 (1967). The Founding generation crafted the Fourth Amendment as a “response to the reviled ‘general warrants’ and ‘writs of assistance’ of the colonial era, which allowed British officers to rummage through homes in an unrestrained search for evidence of criminal activity.” Riley v. California, 573 U. S. ___, ___ (2014) (slip op., at 27). In fact, as John Adams recalled, the patriot James Otis’s 1761 speech condemning writs of assistance was “the first act of opposition to the arbitrary claims of Great Britain” and helped spark the Revolution itself. Id., at ___–___ (slip op., at 27–28) (quoting 10 Works of John Adams 248 (C. Adams ed. 1856)). For much of our history, Fourth Amendment search doctrine was “tied to common-law trespass” and focused on whether the Government “obtains information by physically intruding on a constitutionally protected area.” United States v. Jones, 565 U. S. 400, 405, 406, n. 3 (2012). More recently, the Court has recognized that “property rights are not the sole measure of Fourth Amendment violations.” Soldal v. Cook County, 506 U. S. 56, 64 (1992). In Katz v. United States, 389 U. S. 347, 351 (1967), we established that “the Fourth Amendment protects people, not places,” and expanded our conception of the Amendment to protect certain expectations of privacy as well. When an individual “seeks to preserve something as private,” and his expectation of privacy is “one that society is prepared to recognize as reasonable,” we have held that official intrusion into that private sphere generally qualifies as a search and requires a warrant supported by probable cause. Smith, 442 U. S., at 740 (internal quotation marks and alterations omitted). Although no single rubric definitively resolves which expectations of privacy are entitled to protection,[1] the analysis is informed by historical understandings “of what was deemed an unreasonable search and seizure when [the Fourth Amendment] was adopted.” Carroll v. United States, 267 U. S. 132, 149 (1925). On this score, our cases have recognized some basic guideposts. First, that the Amendment seeks to secure “the privacies of life” against “arbitrary power.” Boyd v. United States, 116 U. S. 616, 630 (1886). Second, and relatedly, that a central aim of the Framers was “to place obstacles in the way of a too permeating police surveillance.” United States v. Di Re, 332 U. S. 581, 595 (1948). We have kept this attention to Founding-era understandings in mind when applying the Fourth Amendment to innovations in surveillance tools. As technology has enhanced the Government’s capacity to encroach upon areas normally guarded from inquisitive eyes, this Court has sought to “assure[ ] preservation of that degree of privacy against government that existed when the Fourth Amendment was adopted.” Kyllo v. United States, 533 U. S. 27, 34 (2001). For that reason, we rejected in Kyllo a “mechanical interpretation” of the Fourth Amendment and held that use of a thermal imager to detect heat radiating from the side of the defendant’s home was a search. Id., at 35. Because any other conclusion would leave homeowners “at the mercy of advancing technology,” we determined that the Government—absent a warrant—could not capitalize on such new sense-enhancing technology to explore what was happening within the home. Ibid. Likewise in Riley, the Court recognized the “immense storage capacity” of modern cell phones in holding that police officers must generally obtain a warrant before searching the contents of a phone. 573 U. S., at ___ (slip op., at 17). We explained that while the general rule allowing warrantless searches incident to arrest “strikes the appropriate balance in the context of physical objects, neither of its rationales has much force with respect to” the vast store of sensitive information on a cell phone. Id., at ___ (slip op., at 9). B The case before us involves the Government’s acquisition of wireless carrier cell-site records revealing the location of Carpenter’s cell phone whenever it made or received calls. This sort of digital data—personal location information maintained by a third party—does not fit neatly under existing precedents. Instead, requests for cell-site records lie at the intersection of two lines of cases, both of which inform our understanding of the privacy interests at stake. The first set of cases addresses a person’s expectation of privacy in his physical location and movements. In United States v. Knotts, 460 U. S. 276 (1983), we considered the Government’s use of a “beeper” to aid in tracking a vehicle through traffic. Police officers in that case planted a beeper in a container of chloroform before it was purchased by one of Knotts’s co-conspirators. The officers (with intermittent aerial assistance) then followed the automobile carrying the container from Minneapolis to Knotts’s cabin in Wisconsin, relying on the beeper’s signal to help keep the vehicle in view. The Court concluded that the “augment[ed]” visual surveillance did not constitute a search because “[a] person traveling in an automobile on public thoroughfares has no reasonable expectation of privacy in his movements from one place to another.” Id., at 281, 282. Since the movements of the vehicle and its final destination had been “voluntarily conveyed to anyone who wanted to look,” Knotts could not assert a privacy interest in the information obtained. Id., at 281. This Court in Knotts, however, was careful to distinguish between the rudimentary tracking facilitated by the beeper and more sweeping modes of surveillance. The Court emphasized the “limited use which the government made of the signals from this particular beeper” during a discrete “automotive journey.” Id., at 284, 285. Significantly, the Court reserved the question whether “different constitutional principles may be applicable” if “twenty-four hour surveillance of any citizen of this country [were] possible.” Id., at 283–284. Three decades later, the Court considered more sophisticated surveillance of the sort envisioned in Knotts and found that different principles did indeed apply. In United States v. Jones, FBI agents installed a GPS tracking device on Jones’s vehicle and remotely monitored the vehicle’s movements for 28 days. The Court decided the case based on the Government’s physical trespass of the vehicle. 565 U. S., at 404–405. At the same time, five Justices agreed that related privacy concerns would be raised by, for example, “surreptitiously activating a stolen vehicle detection system” in Jones’s car to track Jones himself, or conducting GPS tracking of his cell phone. Id., at 426, 428 (Alito, J., concurring in judgment); id., at 415 (Sotomayor, J., concurring). Since GPS monitoring of a vehicle tracks “every movement” a person makes in that vehicle, the concurring Justices concluded that “longer term GPS monitoring in investigations of most offenses impinges on expectations of privacy”—regardless whether those movements were disclosed to the public at large. Id., at 430 (opinion of Alito, J.); id., at 415 (opinion of Sotomayor, J.).[2] In a second set of decisions, the Court has drawn a line between what a person keeps to himself and what he shares with others. We have previously held that “a person has no legitimate expectation of privacy in information he voluntarily turns over to third parties.” Smith, 442 U. S., at 743–744. That remains true “even if the information is revealed on the assumption that it will be used only for a limited purpose.” United States v. Miller, 425 U. S. 435, 443 (1976). As a result, the Government is typically free to obtain such information from the recipient without triggering Fourth Amendment protections. This third-party doctrine largely traces its roots to Miller. While investigating Miller for tax evasion, the Government subpoenaed his banks, seeking several months of canceled checks, deposit slips, and monthly statements. The Court rejected a Fourth Amendment challenge to the records collection. For one, Miller could “assert neither ownership nor possession” of the documents; they were “business records of the banks.” Id., at 440. For another, the nature of those records confirmed Miller’s limited expectation of privacy, because the checks were “not confidential communications but negotiable instruments to be used in commercial transactions,” and the bank statements contained information “exposed to [bank] employees in the ordinary course of business.” Id., at 442. The Court thus concluded that Miller had “take[n] the risk, in revealing his affairs to another, that the information [would] be conveyed by that person to the Government.” Id., at 443. Three years later, Smith applied the same principles in the context of information conveyed to a telephone com- pany. The Court ruled that the Government’s use of a pen register—a device that recorded the outgoing phone numbers dialed on a landline telephone—was not a search. Noting the pen register’s “limited capabilities,” the Court “doubt[ed] that people in general entertain any actual expectation of privacy in the numbers they dial.” 442 U. S., at 742. Telephone subscribers know, after all, that the numbers are used by the telephone company “for a variety of legitimate business purposes,” including routing calls. Id., at 743. And at any rate, the Court explained, such an expectation “is not one that society is prepared to recognize as reasonable.” Ibid. (internal quotation marks omitted). When Smith placed a call, he “voluntarily conveyed” the dialed numbers to the phone company by “expos[ing] that information to its equipment in the ordinary course of business.” Id., at 744 (internal quotation marks omitted). Once again, we held that the defendant “assumed the risk” that the company’s records “would be divulged to police.” Id., at 745. III The question we confront today is how to apply the Fourth Amendment to a new phenomenon: the ability to chronicle a person’s past movements through the record of his cell phone signals. Such tracking partakes of many of the qualities of the GPS monitoring we considered in Jones. Much like GPS tracking of a vehicle, cell phone location information is detailed, encyclopedic, and effortlessly compiled. At the same time, the fact that the individual continuously reveals his location to his wireless carrier implicates the third-party principle of Smith and Miller. But while the third-party doctrine applies to telephone numbers and bank records, it is not clear whether its logic extends to the qualitatively different category of cell-site records. After all, when Smith was decided in 1979, few could have imagined a society in which a phone goes wherever its owner goes, conveying to the wireless carrier not just dialed digits, but a detailed and comprehensive record of the person’s movements. We decline to extend Smith and Miller to cover these novel circumstances. Given the unique nature of cell phone location records, the fact that the information is held by a third party does not by itself overcome the user’s claim to Fourth Amendment protection. Whether the Government employs its own surveillance technology as in Jones or leverages the technology of a wireless carrier, we hold that an individual maintains a legitimate expectation of privacy in the record of his physical movements as captured through CSLI. The location information obtained from Carpenter’s wireless carriers was the product of a search.[3] A A person does not surrender all Fourth Amendment protection by venturing into the public sphere. To the contrary, “what [one] seeks to preserve as private, even in an area accessible to the public, may be constitutionally protected.” Katz, 389 U. S., at 351–352. A majority of this Court has already recognized that individuals have a reasonable expectation of privacy in the whole of their physical movements. Jones, 565 U. S., at 430 (Alito, J., concurring in judgment); id., at 415 (Sotomayor, J., concurring). Prior to the digital age, law enforcement might have pursued a suspect for a brief stretch, but doing so “for any extended period of time was difficult and costly and therefore rarely undertaken.” Id., at 429 (opinion of Alito, J.). For that reason, “society’s expectation has been that law enforcement agents and others would not—and indeed, in the main, simply could not—secretly monitor and catalogue every single movement of an individual’s car for a very long period.” Id., at 430. Allowing government access to cell-site records contravenes that expectation. Although such records are generated for commercial purposes, that distinction does not negate Carpenter’s anticipation of privacy in his physical location. Mapping a cell phone’s location over the course of 127 days provides an all-encompassing record of the holder’s whereabouts. As with GPS information, the time-stamped data provides an intimate window into a person’s life, revealing not only his particular movements, but through them his “familial, political, professional, religious, and sexual associations.” Id., at 415 (opinion of Sotomayor, J.). These location records “hold for many Americans the ‘privacies of life.’ ” Riley, 573 U. S., at ___ (slip op., at 28) (quoting Boyd, 116 U. S., at 630). And like GPS monitoring, cell phone tracking is remarkably easy, cheap, and efficient compared to traditional investigative tools. With just the click of a button, the Government can access each carrier’s deep repository of historical location information at practically no expense. In fact, historical cell-site records present even greater privacy concerns than the GPS monitoring of a vehicle we considered in Jones. Unlike the bugged container in Knotts or the car in Jones, a cell phone—almost a “feature of human anatomy,” Riley, 573 U. S., at ___ (slip op., at 9)—tracks nearly exactly the movements of its owner. While individuals regularly leave their vehicles, they compulsively carry cell phones with them all the time. A cell phone faithfully follows its owner beyond public thoroughfares and into private residences, doctor’s offices, political headquarters, and other potentially revealing locales. See id., at ___ (slip op., at 19) (noting that “nearly three-quarters of smart phone users report being within five feet of their phones most of the time, with 12% admitting that they even use their phones in the shower”); contrast Cardwell v. Lewis, 417 U. S. 583, 590 (1974) (plurality opinion) (“A car has little capacity for escaping public scrutiny.”). Accordingly, when the Government tracks the location of a cell phone it achieves near perfect surveillance, as if it had attached an ankle monitor to the phone’s user. Moreover, the retrospective quality of the data here gives police access to a category of information otherwise unknowable. In the past, attempts to reconstruct a person’s movements were limited by a dearth of records and the frailties of recollection. With access to CSLI, the Government can now travel back in time to retrace a person’s whereabouts, subject only to the retention polices of the wireless carriers, which currently maintain records for up to five years. Critically, because location information is continually logged for all of the 400 million devices in the United States—not just those belonging to persons who might happen to come under investigation—this newfound tracking capacity runs against everyone. Unlike with the GPS device in Jones, police need not even know in advance whether they want to follow a particular individual, or when. Whoever the suspect turns out to be, he has effectively been tailed every moment of every day for five years, and the police may—in the Government’s view—call upon the results of that surveillance without regard to the constraints of the Fourth Amendment. Only the few with- out cell phones could escape this tireless and absolute surveillance. The Government and Justice Kennedy contend, however, that the collection of CSLI should be permitted because the data is less precise than GPS information. Not to worry, they maintain, because the location records did “not on their own suffice to place [Carpenter] at the crime scene”; they placed him within a wedge-shaped sector ranging from one-eighth to four square miles. Brief for United States 24; see post, at 18–19. Yet the Court has already rejected the proposition that “inference insulates a search.” Kyllo, 533 U. S., at 36. From the 127 days of location data it received, the Government could, in combination with other information, deduce a detailed log of Carpenter’s movements, including when he was at the site of the robberies. And the Government thought the CSLI accurate enough to highlight it during the closing argument of his trial. App. 131. At any rate, the rule the Court adopts “must take account of more sophisticated systems that are already in use or in development.” Kyllo, 533 U. S., at 36. While the records in this case reflect the state of technology at the start of the decade, the accuracy of CSLI is rapidly approaching GPS-level precision. As the number of cell sites has proliferated, the geographic area covered by each cell sector has shrunk, particularly in urban areas. In addition, with new technology measuring the time and angle of signals hitting their towers, wireless carriers already have the capability to pinpoint a phone’s location within 50 meters. Brief for Electronic Frontier Foundation et al. as Amici Curiae 12 (describing triangulation methods that estimate a device’s location inside a given cell sector). Accordingly, when the Government accessed CSLI from the wireless carriers, it invaded Carpenter’s reason- able expectation of privacy in the whole of his physical movements. B The Government’s primary contention to the contrary is that the third-party doctrine governs this case. In its view, cell-site records are fair game because they are “business records” created and maintained by the wireless carriers. The Government (along with Justice Kennedy) recognizes that this case features new technology, but asserts that the legal question nonetheless turns on a garden-variety request for information from a third-party witness. Brief for United States 32–34; post, at 12–14. The Government’s position fails to contend with the seismic shifts in digital technology that made possible the tracking of not only Carpenter’s location but also everyone else’s, not for a short period but for years and years. Sprint Corporation and its competitors are not your typical witnesses. Unlike the nosy neighbor who keeps an eye on comings and goings, they are ever alert, and their memory is nearly infallible. There is a world of difference between the limited types of personal information addressed in Smith and Miller and the exhaustive chronicle of location information casually collected by wireless carriers today. The Government thus is not asking for a straightforward application of the third-party doctrine, but instead a significant extension of it to a distinct category of information. The third-party doctrine partly stems from the notion that an individual has a reduced expectation of privacy in information knowingly shared with another. But the fact of “diminished privacy interests does not mean that the Fourth Amendment falls out of the picture entirely.” Riley, 573 U. S., at ___ (slip op., at 16). Smith and Miller, after all, did not rely solely on the act of sharing. Instead, they considered “the nature of the particular documents sought” to determine whether “there is a legitimate ‘expectation of privacy’ concerning their contents.” Miller, 425 U. S., at 442. Smith pointed out the limited capabilities of a pen register; as explained in Riley, telephone call logs reveal little in the way of “identifying information.” Smith, 442 U. S., at 742; Riley, 573 U. S., at ___ (slip op., at 24). Miller likewise noted that checks were “not confidential communications but negotiable instruments to be used in commercial transactions.” 425 U. S., at 442. In mechanically applying the third-party doctrine to this case, the Government fails to appreciate that there are no comparable limitations on the revealing nature of CSLI. The Court has in fact already shown special solicitude for location information in the third-party context. In Knotts, the Court relied on Smith to hold that an individual has no reasonable expectation of privacy in public movements that he “voluntarily conveyed to anyone who wanted to look.” Knotts, 460 U. S., at 281; see id., at 283 (discussing Smith). But when confronted with more pervasive tracking, five Justices agreed that longer term GPS monitoring of even a vehicle traveling on public streets constitutes a search. Jones, 565 U. S., at 430 (Alito, J., concurring in judgment); id., at 415 (Sotomayor, J., concurring). Justice Gorsuch wonders why “someone’s location when using a phone” is sensitive, post, at 3, and Justice Kennedy assumes that a person’s discrete movements “are not particularly private,” post, at 17. Yet this case is not about “using a phone” or a person’s movement at a particular time. It is about a detailed chronicle of a person’s physical presence compiled every day, every moment, over several years. Such a chronicle implicates privacy concerns far beyond those considered in Smith and Miller. Neither does the second rationale underlying the third-party doctrine—voluntary exposure—hold up when it comes to CSLI. Cell phone location information is not truly “shared” as one normally understands the term. In the first place, cell phones and the services they provide are “such a pervasive and insistent part of daily life” that carrying one is indispensable to participation in modern society. Riley, 573 U. S., at ___ (slip op., at 9). Second, a cell phone logs a cell-site record by dint of its operation, without any affirmative act on the part of the user beyond powering up. Virtually any activity on the phone generates CSLI, including incoming calls, texts, or e-mails and countless other data connections that a phone automatically makes when checking for news, weather, or social media updates. Apart from disconnecting the phone from the network, there is no way to avoid leaving behind a trail of location data. As a result, in no meaningful sense does the user voluntarily “assume[ ] the risk” of turning over a comprehensive dossier of his physical movements. Smith, 442 U. S., at 745. We therefore decline to extend Smith and Miller to the collection of CSLI. Given the unique nature of cell phone location information, the fact that the Government obtained the information from a third party does not overcome Carpenter’s claim to Fourth Amendment protection. The Government’s acquisition of the cell-site records was a search within the meaning of the Fourth Amendment. * * * Our decision today is a narrow one. We do not express a view on matters not before us: real-time CSLI or “tower dumps” (a download of information on all the devices that connected to a particular cell site during a particular interval). We do not disturb the application of Smith and Miller or call into question conventional surveillance techniques and tools, such as security cameras. Nor do we address other business records that might incidentally reveal location information. Further, our opinion does not consider other collection techniques involving foreign affairs or national security. As Justice Frankfurter noted when considering new innovations in airplanes and radios, the Court must tread carefully in such cases, to ensure that we do not “embarrass the future.” Northwest Airlines, Inc. v. Minnesota, 322 U. S. 292, 300 (1944).[4] IV Having found that the acquisition of Carpenter’s CSLI was a search, we also conclude that the Government must generally obtain a warrant supported by probable cause before acquiring such records. Although the “ultimate measure of the constitutionality of a governmental search is ‘reasonableness,’ ” our cases establish that warrantless searches are typically unreasonable where “a search is undertaken by law enforcement officials to discover evidence of criminal wrongdoing.” Vernonia School Dist. 47J v. Acton, 515 U. S. 646, 652–653 (1995). Thus, “[i]n the absence of a warrant, a search is reasonable only if it falls within a specific exception to the warrant requirement.” Riley, 573 U. S., at ___ (slip op., at 5). The Government acquired the cell-site records pursuant to a court order issued under the Stored Communications Act, which required the Government to show “reasonable grounds” for believing that the records were “relevant and material to an ongoing investigation.” 18 U. S. C. §2703(d). That showing falls well short of the probable cause required for a warrant. The Court usually requires “some quantum of individualized suspicion” before a search or seizure may take place. United States v. Martinez-Fuerte, 428 U. S. 543, 560–561 (1976). Under the standard in the Stored Communications Act, however, law enforcement need only show that the cell-site evidence might be pertinent to an ongoing investigation—a “gigantic” departure from the probable cause rule, as the Government explained below. App. 34. Consequently, an order issued under Section 2703(d) of the Act is not a permissible mechanism for accessing historical cell-site records. Before compelling a wireless carrier to turn over a subscriber’s CSLI, the Government’s obligation is a familiar one—get a warrant. Justice Alito contends that the warrant requirement simply does not apply when the Government acquires records using compulsory process. Unlike an actual search, he says, subpoenas for documents do not involve the direct taking of evidence; they are at most a “constructive search” conducted by the target of the subpoena. Post, at 12. Given this lesser intrusion on personal privacy, Justice Alito argues that the compulsory production of records is not held to the same probable cause standard. In his view, this Court’s precedents set forth a categorical rule—separate and distinct from the third-party doctrine—subjecting subpoenas to lenient scrutiny without regard to the suspect’s expectation of privacy in the records. Post, at 8–19. But this Court has never held that the Government may subpoena third parties for records in which the suspect has a reasonable expectation of privacy. Almost all of the examples Justice Alito cites, see post, at 14–15, contemplated requests for evidence implicating diminished pri- vacy interests or for a corporation’s own books.[5] The lone exception, of course, is Miller, where the Court’s analysis of the third-party subpoena merged with the application of the third-party doctrine. 425 U. S., at 444 (concluding that Miller lacked the necessary privacy interest to contest the issuance of a subpoena to his bank). Justice Alito overlooks the critical issue. At some point, the dissent should recognize that CSLI is an entirely different species of business record—something that implicates basic Fourth Amendment concerns about arbitrary government power much more directly than corporate tax or payroll ledgers. When confronting new concerns wrought by digital technology, this Court has been careful not to uncritically extend existing precedents. See Riley, 573 U. S., at ___ (slip op., at 10) (“A search of the information on a cell phone bears little resemblance to the type of brief physical search considered [in prior precedents].”). If the choice to proceed by subpoena provided a categorical limitation on Fourth Amendment protection, no type of record would ever be protected by the warrant requirement. Under Justice Alito’s view, private letters, digital contents of a cell phone—any personal information reduced to document form, in fact—may be collected by subpoena for no reason other than “official curiosity.” United States v. Morton Salt Co., 338 U. S. 632, 652 (1950). Justice Kennedy declines to adopt the radical implications of this theory, leaving open the question whether the warrant requirement applies “when the Government obtains the modern-day equivalents of an individual’s own ‘papers’ or ‘effects,’ even when those papers or effects are held by a third party. ” Post, at 13 (citing United States v. Warshak, 631 F. 3d 266, 283–288 (CA6 2010)). That would be a sensible exception, because it would prevent the subpoena doctrine from overcoming any reasonable expectation of privacy. If the third-party doctrine does not apply to the “modern-day equivalents of an individual’s own ‘papers’ or ‘effects,’ ” then the clear implication is that the documents should receive full Fourth Amendment protection. We simply think that such protection should extend as well to a detailed log of a person’s movements over several years. This is certainly not to say that all orders compelling the production of documents will require a showing of probable cause. The Government will be able to use subpoenas to acquire records in the overwhelming majority of investigations. We hold only that a warrant is required in the rare case where the suspect has a legitimate privacy interest in records held by a third party. Further, even though the Government will generally need a warrant to access CSLI, case-specific exceptions may support a warrantless search of an individual’s cell-site records under certain circumstances. “One well-recognized exception applies when ‘ “the exigencies of the situation” make the needs of law enforcement so compelling that [a] warrantless search is objectively reasonable under the Fourth Amendment.’ ” Kentucky v. King, 563 U. S. 452, 460 (2011) (quoting Mincey v. Arizona, 437 U. S. 385, 394 (1978)). Such exigencies include the need to pursue a fleeing suspect, protect individuals who are threatened with imminent harm, or prevent the imminent destruction of evidence. 563 U. S., at 460, and n. 3. As a result, if law enforcement is confronted with an urgent situation, such fact-specific threats will likely justify the warrantless collection of CSLI. Lower courts, for instance, have approved warrantless searches related to bomb threats, active shootings, and child abductions. Our decision today does not call into doubt warrantless access to CSLI in such circumstances. While police must get a warrant when collecting CSLI to assist in the mine-run criminal investigation, the rule we set forth does not limit their ability to respond to an ongoing emergency. * * * As Justice Brandeis explained in his famous dissent, the Court is obligated—as “[s]ubtler and more far-reaching means of invading privacy have become available to the Government”—to ensure that the “progress of science” does not erode Fourth Amendment protections. Olmstead v. United States, 277 U. S. 438, 473–474 (1928). Here the progress of science has afforded law enforcement a powerful new tool to carry out its important responsibilities. At the same time, this tool risks Government encroachment of the sort the Framers, “after consulting the lessons of history,” drafted the Fourth Amendment to prevent. Di Re, 332 U. S., at 595. We decline to grant the state unrestricted access to a wireless carrier’s database of physical location information. In light of the deeply revealing nature of CSLI, its depth, breadth, and comprehensive reach, and the inescapable and automatic nature of its collection, the fact that such information is gathered by a third party does not make it any less deserving of Fourth Amendment protection. The Government’s acquisition of the cell-site records here was a search under that Amendment. The judgment of the Court of Appeals is reversed, and the case is remanded for further proceedings consistent with this opinion. It is so ordered. Notes 1 Justice Kennedy believes that there is such a rubric—the “property-based concepts” that Katz purported to move beyond. Post, at 3 (dissenting opinion). But while property rights are often informative, our cases by no means suggest that such an interest is “fundamental” or “dispositive” in determining which expectations of privacy are legitimate. Post, at 8–9. Justice Thomas (and to a large extent Justice Gorsuch) would have us abandon Katz and return to an exclusively property-based approach. Post, at 1–2, 17–21 (Thomas J., dissenting); post, at 6–9 (Gorsuch, J., dissenting). Katz of course “discredited” the “premise that property interests control,” 389 U. S., at 353, and we have repeatedly emphasized that privacy interests do not rise or fall with property rights, see, e.g., United States v. Jones, 565 U. S. 400, 411 (2012) (refusing to “make trespass the exclusive test”); Kyllo v. United States, 533 U. S. 27, 32 (2001) (“We have since decoupled violation of a person’s Fourth Amendment rights from trespassory violation of his property.”). Neither party has asked the Court to reconsider Katz in this case. 2 Justice Kennedy argues that this case is in a different category from Jones and the dragnet-type practices posited in Knotts because the disclosure of the cell-site records was subject to “judicial authorization.” Post, at 14–16. That line of argument conflates the threshold question whether a “search” has occurred with the separate matter of whether the search was reasonable. The subpoena process set forth in the Stored Communications Act does not determine a target’s expectation of privacy. And in any event, neither Jones nor Knotts purported to resolve the question of what authorization may be required to conduct such electronic surveillance techniques. But see Jones, 565 U. S., at 430 (Alito, J., concurring in judgment) (indicating that longer term GPS tracking may require a warrant). 3 The parties suggest as an alternative to their primary submissions that the acquisition of CSLI becomes a search only if it extends beyond a limited period. See Reply Brief 12 (proposing a 24-hour cutoff); Brief for United States 55–56 (suggesting a seven-day cutoff). As part of its argument, the Government treats the seven days of CSLI requested from Sprint as the pertinent period, even though Sprint produced only two days of records. Brief for United States 56. Contrary to Justice Kennedy’s assertion, post, at 19, we need not decide whether there is a limited period for which the Government may obtain an individual’s historical CSLI free from Fourth Amendment scrutiny, and if so, how long that period might be. It is sufficient for our purposes today to hold that accessing seven days of CSLI constitutes a Fourth Amendment search. 4 Justice Gorsuch faults us for not promulgating a complete code addressing the manifold situations that may be presented by this new technology—under a constitutional provision turning on what is “reasonable,” no less. Post, at 10–12. Like Justice Gorsuch, we “do not begin to claim all the answers today,” post, at 13, and therefore decide no more than the case before us. 5 See United States v. Dionisio, 410 U. S. 1, 14 (1973) (“No person can have a reasonable expectation that others will not know the sound of his voice”); Donovan v. Lone Steer, Inc., 464 U. S. 408, 411, 415 (1984) (payroll and sales records); California Bankers Assn. v. Shultz, 416 U. S. 21, 67 (1974) (Bank Secrecy Act reporting requirements); See v. Seattle, 387 U. S. 541, 544 (1967) (financial books and records); United States v. Powell, 379 U. S. 48, 49, 57 (1964) (corporate tax records); McPhaul v. United States, 364 U. S. 372, 374, 382 (1960) (books and records of an organization); United States v. Morton Salt Co., 338 U. S. 632, 634, 651–653 (1950) (Federal Trade Commission reporting requirement); Oklahoma Press Publishing Co. v. Walling, 327 U. S. 186, 189, 204–208 (1946) (payroll records); Hale v. Henkel, 201 U. S. 43, 45, 75 (1906) (corporate books and papers).

585.US.2017_17-5639

The Federal Sentencing Guidelines require a sentencing judge to first identify the recommended Guidelines sentencing range based on certain offender and offense characteristics. The judge might choose a penalty within that Guidelines range, or the judge may “depart” or “vary” from the Guidelines and select a sentence outside the range. See United States v. Booker, 543 U. S. 220, 258–265. Either way, the judge must take into account certain statutory sentencing factors, see 18 U. S. C. §3553(a), and must “state in open court the reasons for [imposing] the particular sentence,” §3553(c). But when it comes to how detailed that statement of reasons must be, “[t]he law leaves much . . . to the judge’s own professional judgment.” Rita v. United States, 551 U. S. 338, 356. The explanation need not be lengthy, especially where “a matter is . . . conceptually simple . . . and the record makes clear that the sentencing judge considered the evidence and arguments.” Id., at 359. Here, petitioner pleaded guilty to possessing methamphetamine with intent to distribute. The judge reviewed the Guidelines, determined the range to be 135 to 168 months, and imposed a sentence at the bottom of the range. The Sentencing Commission later lowered the relevant range to 108 to 135 months, and petitioner sought a sentence reduction under §3582(c)(2). Petitioner asked the judge to reduce his sentence to the bottom of the new range, but the judge reduced petitioner’s sentence to 114 months instead. The order was entered on a form certifying that the judge had “considered” petitioner’s “motion” and had “tak[en] into account” the §3553(a) factors and the relevant Guidelines policy statement. On appeal, petitioner argued the sentencing judge did not adequately explain why he rejected petitioner’s request for a 108-month sentence. The Court of Appeals affirmed. Held: Because the record as a whole demonstrates the judge had a reasoned basis for his decision, the judge’s explanation for petitioner’s sentence reduction was adequate. Pp. 5–10. (a) The Government argues petitioner was not entitled to an explanation at all because the statute governing sentence-modification motions does not expressly require a sentencing judge to state his reasons for imposing a particular sentence. See §3582(c)(2). It is unnecessary to go as far as the Government urges, however, because, even assuming the District Court had a duty to explain its reasons when modifying petitioner’s sentence, what the court did here was sufficient. Pp. 5–6. (b) Petitioner contends that a district court must explain its reasoning in greater detail when the court imposes a “disproportionate” sentence reduction—that is, when the court reduces the prisoner’s sentence to a different point in the amended Guidelines range than the court previously selected in the original Guidelines range. That argument is unconvincing. As a technical matter, determining “proportionality” may prove difficult when the sentence is somewhere in the middle of the range. More importantly, the choice among points on the Guidelines range often reflects the belief that the chosen sentence is the “right” sentence based on various factors, including those found in §3553(a). If the applicable Guidelines range is later reduced, it is unsurprising that the sentencing judge may choose a non-proportional point in the new range. Pp. 6–7. (c) Even assuming that a judge reducing a prisoner’s sentence must satisfy the same explanation requirement that applies at an original sentencing, the District Court’s explanation was adequate. At the original sentencing, petitioner asked for a downward variance from the Guidelines range, which the judge denied. The judge observed that petitioner’s sentence was high because of the destructiveness of methamphetamine and the quantity involved. The record from the original sentencing was before the judge—the same judge who imposed the original sentence—when he considered petitioner’s sentence-modification motion. By entering the form order, the judge certified that he had “considered” petitioner’s “motion” and had “tak[en] into account” the §3553(a) factors and the relevant Guidelines policy statement. Because the record as a whole suggests the judge originally believed that 135 months was an appropriately high sentence in light of petitioner’s offense conduct, it is unsurprising that he considered a sentence somewhat higher than the bottom of the reduced range to be appropriate as well. That is not to say that a disproportionate sentence reduction never may require a more detailed explanation. But given the simplicity of this case, the judge’s awareness of the arguments, his consideration of the relevant sentencing factors, and the intuitive reason why he picked a sentence above the very bottom of the new range, his explanation fell within the scope of lawful professional judgment that the law confers upon the sentencing judge. Pp. 7–10. 854 F. 3d 655, affirmed. Breyer, J., delivered the opinion of the Court, in which Roberts, C. J., and Thomas, Ginsburg, and Alito, JJ., joined. Kennedy, J., filed a dissenting opinion, in which Sotomayor and Kagan, JJ., joined. Gorsuch, J., took no part in the consideration or decision of the case.

This case concerns a criminal drug offender originally sentenced in accordance with the Federal Sentencing Guidelines. Subsequently, the Sentencing Commission lowered the applicable Guidelines sentencing range; the offender asked for a sentence reduction in light of the lowered range; and the District Judge reduced his original sentence from 135 months’ imprisonment to 114 months’. The offender, believing he should have obtained a yet greater reduction, argues that the District Judge did not adequately explain why he imposed a sentence of 114 months rather than a lower sentence. The Court of Appeals held that the judge’s explanation was adequate. And we agree with the Court of Appeals. I A The Sentencing Guidelines require a sentencing judge to consider certain listed characteristics of the offender and the offense for which he was convicted. Those characteristics (and certain other factors) bring the judge to a Guidelines table that sets forth a range of punishments, for example, 135 to 168 months’ imprisonment. A sentencing judge often will choose a specific penalty from a Guidelines range. But a judge also has the legal authority to impose a sentence outside the range either because he or she “departs” from the range (as is permitted by certain Guidelines rules) or because he or she chooses to “vary” from the Guidelines by not applying them at all. See United States v. Booker, 543 U. S. 220, 258–265 (2005) (holding the Sentencing Guidelines are advisory). The judge, however, must always take account of certain statutory factors. See 18 U. S. C. §3553(a) (requiring the judge to consider the “seriousness of the offense” and the need to “afford adequate deterrence,” among other factors). And, of particular relevance here, the judge “shall state in open court the reasons for [the] imposition of the particular sentence.” §3553(c). If the sentence is outside the Guidelines range (whether because of a “departure” or a “variance”), the judge must state “the specific reason for the imposition of a . . . different” sentence. §3553(c)(2). If the sentence is within the Guidelines range, and the Guidelines range exceeds 24 months, the judge must also state “the reason for imposing a sentence at a particular point within the range.” §3553(c)(1). B We here consider one aspect of the judge’s obligation to provide reasons. In an earlier case, we set forth the law that governs the explanation requirement at sentencing. In Rita v. United States, 551 U. S. 338 (2007), the offender sought a downward departure from the Guidelines. The record, we said, showed that the sentencing judge “listened to each argument[,] . . . considered the supporting evidence[,] . . . was fully aware of defendant’s various physical ailments[,]” imposed a sentence at the bottom of the Guidelines range, and, having considered the §3553(a) factors, said simply that the sentence was “ ‘appropriate.’ ” Id., at 358. We held that where “a matter is as conceptually simple as in the case at hand and the record makes clear that the sentencing judge considered the evidence and arguments, we do not believe the law requires the judge to write more extensively.” Id., at 359. We also discussed more generally the judge’s obligation to explain. We wrote that the statute calls “for the judge to ‘state’ his ‘reasons.’ And that requirement reflects sound judicial practice. Judicial decisions are reasoned decisions. Confidence in a judge’s use of reason underlies the public’s trust in the judicial institution. A public statement of those reasons helps provide the public with the assurance that creates that trust.” Id., at 356. But, we continued, “we cannot read the statute (or our precedent) as insisting upon a full opinion in every case. The appropriateness of brevity or length, conciseness or detail, when to write, what to say, depends upon circumstances. Sometimes a judicial opinion responds to every argument; sometimes it does not; sometimes a judge simply writes the word ‘granted’ or ‘denied’ on the face of a motion while relying upon context and the parties’ prior arguments to make the reasons clear. The law leaves much, in this respect, to the judge’s own professional judgment.” Ibid. At bottom, the sentencing judge need only “set forth enough to satisfy the appellate court that he has considered the parties’ arguments and has a reasoned basis for exercising his own legal decisionmaking authority.” Ibid. When a judge applies a sentence within the Guidelines range, he or she often does not need to provide a lengthy explanation. As we said in Rita, “[c]ircumstances may well make clear that the judge rests his decision upon the Commission’s own reasoning that the Guidelines sentence is a proper sentence (in terms of §3553(a) and other congressional mandates) in the typical case, and that the judge has found that the case before him is typical.” Id., at 357. We have followed this same reasoning in other sentencing cases, including Gall v. United States, 552 U. S. 38 (2007), which we decided the same year as Rita. Cf. Kimbrough v. United States, 552 U. S. 85, 109 (2007) (suggesting a district judge’s decision to vary from the Guidelines range may be entitled to greater respect when the judge finds a particular case “ ‘outside the “heartland” ’ ” of the Guidelines). Indeed, the case before us differs from the Guidelines cases that Rita describes in only one significant respect. It concerns a limited form of resentencing. C The relevant lower court proceedings are not complicated. In 2013, petitioner pleaded guilty to a federal crime, namely, possessing methamphetamine with the intent to distribute it. The judge reviewed the Guidelines, determined that the applicable range was 135 to 168 months’ imprisonment and imposed a sentence at the bottom of that range: 135 months. Pursuant to its statutory authority, the Sentencing Commission subsequently lowered the relevant Guidelines range from 135 to 168 months to 108 to 135 months. United States Sentencing Commission, Guidelines Manual App. C, Amdt. 782 (Supp. Nov. 2012–Nov. 2016) (USSG); see also 28 U. S. C. §994(o). Petitioner then sought and obtained a sentence modification. See 18 U. S. C. §3582(c)(2); USSG §1B1.10. He asked the judge to lower his sentence to the bottom of the new range, namely 108 months. But the judge instead lowered it to 114 months, not 108 months. The order was entered on a form issued by the Administrative Office of the United States Courts. The form certified the judge had “considered” petitioner’s motion and “tak[en] into account” the §3553(a) factors and the relevant Guidelines policy statement. App. 106–107 (under seal). Petitioner appealed, claiming that the judge did not adequately explain why he rejected petitioner’s 108-month request. The Court of Appeals rejected his argument. 854 F. 3d 655 (CA10 2017). In its view, “absent any indication the court failed to consider the §3553(a) factors, a district court . . . need not explain choosing a particular guidelines-range sentence.” Id., at 659. Petitioner sought certiorari, and we granted his petition. II A The Government, pointing out that this is a sentence-modification case, argues that this fact alone should secure it a virtually automatic victory. That is because, unlike an ordinary Guidelines sentencing case, the statute governing sentence-modification motions does not insist that the judge provide a “reason for imposing a sentence at a particular point within the range.” Compare §3553(c)(1) with §3582(c)(2). It adds that sentence modifications also differ procedurally from sentencing in that the offender is not entitled to be present in court at the time the reduced sentence is imposed. See Dillon v. United States, 560 U. S. 817, 828 (2010) (citing Fed. Rule Crim. Proc. 43(b)(4)). As we have said before, “Congress intended to authorize only a limited adjustment to an otherwise final sentence and not a plenary resentencing proceeding.” Dillon, supra, at 826. These procedural features, the Government asserts, mean that “the court has no duty” to provide an “on-the-record explanation” of its reasons. Brief for United States 12, 19. We need not go so far. Even assuming (purely for argument’s sake) district courts have equivalent duties when initially sentencing a defendant and when later modifying the sentence, what the District Court did here was sufficient. At the original sentencing, the judge “must adequately explain the chosen sentence to allow for meaningful appellate review.” 552 U. S., at 50; see also Rita, 551 U. S., at 356 (“The sentencing judge should set forth enough to satisfy the appellate court that he has considered the parties’ arguments and has a reasoned basis for exercising his own legal decisionmaking authority”). Just how much of an explanation this requires, however, depends, as we have said, upon the circumstances of the particular case. Id., at 356–357. In some cases, it may be sufficient for purposes of appellate review that the judge simply relied upon the record, while making clear that he or she has considered the parties’ arguments and taken account of the §3553(a) factors, among others. But in other cases, more explanation may be necessary (depending, perhaps, upon the legal arguments raised at sentencing, see id., at 357). That may be the case even when there is little evidence in the record affirmatively showing that the sentencing judge failed to consider the §3553(a) factors. If the court of appeals considers an explanation inadequate in a particular case, it can send the case back to the district court for a more complete explanation. Cf. Molina-Martinez v. United States, 578 U. S. ___, ___ (2016) (slip op., at 15) (“[A]ppellate courts retain broad discretion in determining whether a remand for resentencing is necessary”). B Petitioner argues that the judge should have explained more here because there is, or should be, some kind of presumption that the judge will choose a point within the new lower Guidelines range that is “proportional” to the point previously chosen in the older higher Guidelines range. We are not aware of any law or any convincing reason, however, suggesting that this is so. As a technical matter, determining just what “proportionality” means in this context would often prove difficult when the sentence is somewhere in the middle of the Guidelines range. The Sentencing Table calculates punishments according to a logarithmic scale. Take petitioner’s original and amended Guidelines ranges, for example. The original range was 135 to 168 months, a difference of 33 months. The amended range, by comparison, is 108 to 135 months, a difference of 27 months. And viewed logarithmically, what may seem the middle of a new lower range is not necessarily proportionate to what may seem the middle of the old higher range. Nothing in the Guidelines, or elsewhere, encourages arguments about such matters among lawyers or judges who are not experts in advanced mathematics. More importantly, the Guidelines ranges reflect to some degree what many, perhaps most, judges believed in the pre-Guidelines era was a proper sentence based upon the criminal behavior at issue and the characteristics of the offender. Thus, a judge’s choice among points on a range will often simply reflect the judge’s belief that the chosen sentence is the “right” sentence (or as close as possible to the “right” sentence) based on various factors, including those found in §3553(a). Insofar as that is so, it is unsurprising that changing the applicable range may lead a judge to choose a nonproportional point on the new range. We see nothing that favors the one or the other. So, as is true of most Guidelines sentences, the judge need not provide a lengthy explanation if the “context and the record” make clear that the judge had “a reasoned basis” for reducing the defendant’s sentence. Rita, supra, at 356, 359. C Turning to the facts of this case, we find that the District Court’s explanation satisfies the standard we used in Rita and Gall, assuming it applies to sentence modifications. In Rita, as we earlier said, we upheld as lawful a sentencing judge’s explanation that stated simply that the Guidelines sentence imposed was “ ‘appropriate.’ ” 551 U. S., at 358. We noted that, in respect to the brevity or length of the reasons the judge gives for imposing a particular Guidelines sentence, the “law leaves much” to “the judge’s own professional judgment.” Id., at 356. We pointed out that the sentencing judge in that case had “set forth enough to satisfy the appellate court that he ha[d] considered the parties’ arguments and ha[d] a reasoned basis for exercising his own legal decisionmaking author- ity.” Ibid. The same is true here. At petitioner’s original sentencing, he sought a variance from the Guidelines range (135 to 168 months) on the ground that his history and family circumstances warranted a lower sentence. The judge denied his request. In doing so, the judge noted that he had “consulted the sentencing factors of 18 U. S. C. 3553(a)(1).” He explained that the “reason the guideline sentence is high in this case, even the low end of 135 months, is because of the [drug] quantity.” He pointed out that petitioner had “distributed 1.7 kilograms of actual methamphetamine,” a “significant quantity.” And he said that “one of the other reasons that the penalty is severe in this case is because of methamphetamine.” He elaborated this latter point by stating that he had “been doing this a long time, and from what [he] gather[ed] and what [he had] seen, methamphetamine, it destroys individual lives, it destroys families, it can destroy communities.” App. 25. This record was before the judge when he considered petitioner’s request for a sentence modification. He was the same judge who had sentenced petitioner originally. Petitioner asked the judge to reduce his sentence to 108 months, the bottom of the new range, stressing various educational courses he had taken in prison. The Government pointed to his having also broken a moderately serious rule while in prison. The judge certified (on a form) that he had “considered” petitioner’s “motion” and had “tak[en] into account” the relevant Guidelines policy statements and the §3553(a) factors. Id., at 106–107 (under seal). He then reduced the sentence to 114 months. The record as a whole strongly suggests that the judge originally believed that, given petitioner’s conduct, 135 months was an appropriately high sentence. So it is unsurprising that the judge considered a sentence somewhat higher than the bottom of the reduced range to be appropriate. As in Rita, there was not much else for the judge to say. The dissent would have us ignore the record from the initial sentencing and consider only what the judge said when modifying petitioner’s sentence. See post, at 4-5 (opinion of Kennedy, J.). But, as we have made clear before, a sentence modification is “not a plenary resentencing proceeding.” Dillon, 560 U. S., at 826. We therefore need not turn a blind eye to what the judge said at petitioner’s initial sentencing. The dissent suggests the judge’s failure to grant petitioner a proportional reduction “limits the relevance of the initial sentencing proceeding.” Post, at 5. To the contrary, the record of the initial sentencing sheds light on why the court picked a point slightly above the bottom of the reduced Guidelines range when it modified petitioner’s sentence. Our decision is not (as the dissent claims) based on mere “speculation.” Post, at 7. Rather, we simply find the record as a whole satisfies us that the judge “considered the parties’ arguments and ha[d] a reasoned basis for exercising his own legal decisionmaking authority.” Rita, supra, at 356. This is not to say that a disproportionate sentence reduction never may require a more detailed explanation. It could be that, under different facts and a different record, the district court’s use of a barebones form order in response to a motion like petitioner’s would be inadequate. As we said above, the courts of appeals are well suited to request a more detailed explanation when necessary. See supra, at 6. The dissent asserts that appellate courts would not need to remand for further explanation if district courts provided an additional “short statement or check[ed] additional boxes” on the form order. Post, at 8. That may be so, and nothing in this decision prevents judges from saying more when, in their professional judgment, saying more is appropriate. Providing a more detailed statement of reasons often serves “a salutary purpose” separate and apart from facilitating appellate review. Rita, 551 U. S., at 357. But our task here is to decide the case before us. And given the simplicity of this case, the judge’s awareness of the arguments, his consideration of the relevant sentencing factors, and the intuitive reason why he picked a sentence above the very bottom of the new range, the judge’s explanation (minimal as it was) fell within the scope of the lawful professional judgment that the law confers upon the sentencing judge. See Id. at 356. The Court of Appeals concluded the same. Its judgment is therefore affirmed. It is so ordered. Justice Gorsuch took no part in the consideration or decision of this case.

584.US.2017_17-432

American Pipe & Constr. Co. v. Utah, 414 U. S. 538, established that the timely filing of a class action tolls the applicable statute of limitations for all persons encompassed by the class complaint and that members of a class that fails to gain certification can timely intervene as individual plaintiffs in the still-pending action, shorn of its class character. American Pipe’s tolling rule also applies to putative class members who, after denial of class certification, “prefer to bring an individual suit rather than intervene.” Crown, Cork & Seal Co. v. Parker, 462 U. S. 345, 350. The question presented in this case is whether American Pipe tolling applies not only to individual claims, but to successive class actions as well. This suit is the third class action brought on behalf of purchasers of petitioner China Agritech’s common stock, alleging materially identical violations of the Securities Exchange Act of 1934. The Act has both a two-year statute of limitations and a five-year statute of repose, 28 U. S. C. §1658(b). Here, the accrual date for purposes of the Act’s limitation period is February 3, 2011, and for the repose period, November 12, 2009. Theodore Dean, a China Agritech shareholder, filed the first class-action complaint on February 11, 2011. As required by the Private Securities Litigation Reform Act of 1995 (PSLRA), his counsel posted notice of the action and invited any member of the purported class to move to serve as lead plaintiff. Six shareholders sought lead-plaintiff status. On May 3, 2012, the District Court denied class certification; the action settled in September 2012, and the suit was dismissed. On October 4, Dean’s counsel filed a new complaint (Smyth), still timely, with a new set of plaintiffs. Eight shareholders sought lead-plaintiff appointment in response to the PSLRA notice, but the District Court again denied class certification. Thereafter, the Smyth plaintiffs settled their individual claims and dismissed their suit. Respondent Michael Resh, who did not seek lead-plaintiff status in the earlier actions, filed the present class action in 2014, a year and a half after the statute of limitations expired. The other respondents moved to intervene in the suit commenced by Resh, seeking lead-plaintiff status. The District Court dismissed the class complaint as untimely, holding that the Dean and Smyth actions did not toll the time to initiate class claims. The Ninth Circuit reversed, holding that the reasoning of American Pipe extends to successive class claims. Held: Upon denial of class certification, a putative class member may not, in lieu of promptly joining an existing suit or promptly filing an individual action, commence a class action anew beyond the time allowed by the applicable statute of limitations. Pp. 5–15. (a) American Pipe and Crown, Cork addressed only putative class members who wish to sue individually after a class-certification denial. The “efficiency and economy of litigation” that support tolling of individual claims, American Pipe, 414 U. S., at 553, do not support maintenance of untimely successive class actions such as the one brought by Resh. Economy of litigation favors delaying individual claims until after a class-certification denial. With class claims, on the other hand, efficiency favors early assertion of competing class representative claims. If class treatment is appropriate, and all would-be representatives have come forward, the district court can select the best plaintiff with knowledge of the full array of potential class representatives and class counsel. And if the class mechanism is not a viable option, the decision denying certification will be made at the outset of the case, litigated once for all would-be class representatives. Federal Rule of Procedure 23 evinces a preference for preclusion of untimely successive class actions by instructing that class certification should be resolved early on. The PSLRA, which governs this litigation, evinces a similar preference, this time embodied in legislation providing for early notice and lead-plaintiff procedures. There is little reason to allow plaintiffs who passed up opportunities to participate in the first (and second) round of class litigation to enter the fray several years after class proceedings first commenced. Class representatives who commence suit after expiration of the limitation period are unlikely to qualify as diligent in asserting claims and pursuing relief. See, e.g., McQuiggin v. Perkins, 569 U. S. 383, 391. And respondents’ proposed reading would allow extension of the statute of limitations time and again; as each class is denied certification, a new named plaintiff could file a class complaint that resuscitates the litigation. Endless tolling of a statute of limitations is not a result envisioned by American Pipe. Pp. 5–11. (b) If Resh’s suit meets the requirements of Rule 23(a) and (b), respondents assert, the suit should be permitted to proceed as a class action in keeping with Shady Grove Orthopedic Associates, P. A. v. Allstate Ins. Co., 559 U. S. 393. Shady Grove, however, addressed a case in which a Rule 23 class action could have been maintained absent a state law proscribing class actions, while Resh’s class action would be untimely unless saved by American Pipe’s tolling exception. Rule 23 itself does not address timeliness of claims or tolling and nothing in the Rule calls for the revival of class claims if individual claims are tolled. The clarification of American Pipe’s reach does not run afoul of the Rules Enabling Act by abridging or modifying a substantive right. Plaintiffs have no substantive right to bring claims outside the statute of limitations. Nor is the clarification likely to cause a substantial increase in the number of protective class-action filings. Several Courts of Appeals have already declined to read American Pipe to permit a successive class action filed outside the limitations period, and there is no showing that these Circuits have experienced a disproportionate number of duplicative, protective class-action filings. Multiple filings, moreover, could aid a district court in determining, early on, whether class treatment is warranted, and if so, who would be the best representative. The Federal Rules provide a range of mechanisms to aid district courts in overseeing complex litigation, but they offer no reason to permit plaintiffs to exhume failed class actions by filing new, untimely class claims. Pp. 11–15. 857 F. 3d 994, reversed and remanded. Ginsburg, J., delivered the opinion of the Court, in which Roberts, C. J., and Kennedy, Thomas, Breyer, Alito, Kagan, and Gorsuch, JJ., joined. Sotomayor, J., filed an opinion concurring in the judgment.

This case concerns the tolling rule first stated in American Pipe & Constr. Co. v. Utah, 414 U. S. 538 (1974). The Court held in American Pipe that the timely filing of a class action tolls the applicable statute of limitations for all persons encompassed by the class complaint. Where class-action status has been denied, the Court further ruled, members of the failed class could timely intervene as individual plaintiffs in the still-pending action, shorn of its class character. See id., at 544, 552–553. Later, in Crown, Cork & Seal Co. v. Parker, 462 U. S. 345 (1983), the Court clarified American Pipe’s tolling rule: The rule is not dependent on intervening in or joining an existing suit; it applies as well to putative class members who, after denial of class certification, “prefer to bring an individual suit rather than intervene . . . once the economies of a class action [are] no longer available.” 462 U. S., at 350, 353–354; see California Public Employees’ Retirement System v. ANZ Securities, Inc., 582 U. S. ___, ___ (2017) (slip op., at 13) (American Pipe “permitt[ed] a class action to splinter into individual suits”); Smith v. Bayer Corp., 564 U. S. 299, 313–314, n. 10 (2011) (under American Pipe tolling rule, “a putative member of an uncertified class may wait until after the court rules on the certification motion to file an individual claim or move to intervene in the [existing] suit”). The question presented in the case now before us: Upon denial of class certification, may a putative class member, in lieu of promptly joining an existing suit or promptly filing an individual action, commence a class action anew beyond the time allowed by the applicable statute of limitations? Our answer is no. American Pipe tolls the statute of limitations during the pendency of a putative class action, allowing unnamed class members to join the action individually or file individual claims if the class fails. But American Pipe does not permit the maintenance of a follow-on class action past expiration of the statute of limitations. I The instant suit is the third class action brought on behalf of purchasers of petitioner China Agritech’s common stock, alleging violations of the Securities Exchange Act of 1934, 48Stat. 881, as amended, 15 U. S. C. §78a et seq. In short, the successive complaints each make materially identical allegations that China Agritech engaged in fraud and misleading business practices, causing the company’s stock price to plummet when several reports brought the misconduct to light. See App. 60–100 (Resh complaint), 205–235 (Smyth complaint), 133–156 (Dean complaint). The Exchange Act has a two-year statute of limitations that begins to run upon discovery of the facts constituting the violation. 28 U. S. C. §1658(b). The Act also has a five-year statute of repose. Ibid.[1] The parties agree that the accrual date for purposes of the two-year limitation period is February 3, 2011, and for the five-year repose period, November 12, 2009. Brief for Respondents 8, n. 3. Theodore Dean, a China Agritech shareholder, filed the first class-action complaint on February 11, 2011, at the start of the two-year limitation period. As required by the Private Securities Litigation Reform Act of 1995 (PSLRA), 109Stat. 737, Dean’s counsel posted notice of the action in two “widely circulated national business-oriented publication[s],” 15 U. S. C. §78u–4(a)(3)(A)(i), and invited any member of the purported class to move to serve as lead plaintiff. App. 274–280. Six shareholders responded to the notice, seeking to be named lead plaintiffs; other shareholders who had filed their own class complaints dismissed them in view of the Dean action. On May 3, 2012, after several months of discovery and deferral of a lead-plaintiff ruling, the District Court denied class certification. The plaintiffs, the District Court determined, had failed to establish that China Agritech stock traded on an efficient market—a necessity for proving reliance on a classwide basis. App. 192. Dean’s counsel then published a notice informing shareholders of the certification denial and advising: “You must act yourself to protect your rights. You may protect your rights by joining in the current Action as a plaintiff or by filing your own action against China Agritech.” Id., at 281–282. The Dean action settled in September 2012, occasioning dismissal of the suit. See 857 F. 3d 994, 998 (CA9 2017). On October 4, 2012—within the two-year statute of limitations—Dean’s counsel filed a new complaint (Smyth) with a new set of plaintiffs and new efficient-market evidence. Eight shareholders responded to the PSLRA notice, seeking lead-plaintiff appointment. The District Court again denied class certification, this time on typicality and adequacy grounds. See App. 254. Thereafter, the Smyth plaintiffs settled their individual claims with the defendants and voluntarily dismissed their suit. Because the Smyth litigation was timely commenced, putative class members who promptly initiated individual suits in the wake of the class-action denial would have encountered no statute of limitations bar. Respondent Michael Resh, who had not sought lead-plaintiff status in either the Dean or Smyth proceedings and was represented by counsel who had not appeared in the earlier actions, filed the present suit on June 30, 2014, styling it a class action—a year and a half after the statute of limitations expired. The other respondents moved to intervene, seeking designation as lead plaintiffs; together with Resh, they filed an amended complaint. The District Court dismissed the class complaint as untimely, holding that the Dean and Smyth actions did not toll the time to initiate class claims. App. to Pet. for Cert. 36a. The Court of Appeals for the Ninth Circuit reversed: “[P]ermitting future class action named plaintiffs, who were unnamed class members in previously uncertified classes, to avail themselves of American Pipe tolling,” the court reasoned, “would advance the policy objectives that led the Supreme Court to permit tolling in the first place.” 857 F. 3d, at 1004. Applying American Pipe tolling to successive class actions, the Ninth Circuit added, would cause no unfair surprise to defendants and would promote economy of litigation by reducing incentives for filing protective class suits during the pendency of an initial certification motion. 857 F. 3d, at 1004. We granted certiorari, 583 U. S. ___ (2017), in view of a division of authority among the Courts of Appeals over whether otherwise-untimely successive class claims may be salvaged by American Pipe tolling. Compare the instant case and Phipps v. Wal-Mart Stores, Inc., 792 F. 3d 637, 652–653 (CA6 2015) (applying American Pipe tolling to successive class action), with, e.g., Basch v. Ground Round, Inc., 139 F. 3d 6, 11 (CA1 1998) (“Plaintiffs may not stack one class action on top of another and continue to toll the statute of limitations indefinitely.”); Griffin v. Singletary, 17 F. 3d 356, 359 (CA11 1994) (similar); Korwek v. Hunt, 827 F. 2d 874, 879 (CA2 1987) (American Pipe does not apply to successive class suits); Salazar-Calderon v. Presidio Valley Farmers Assn., 765 F. 2d 1334, 1351 (CA5 1985) (“Plaintiffs have no authority for their contention that putative class members may piggyback one class action onto another and thus toll the statute of limitations indefinitely, nor have we found any.”). See also Yang v. Odom, 392 F. 3d 97, 112 (CA3 2004) (American Pipe tolling does not apply to successive class actions where certification was previously denied due to a class defect, but does apply when certification was denied based on the putative representative’s deficiencies). II A American Pipe established that “the commencement of the original class suit tolls the running of the statute [of limitations] for all purported members of the class who make timely motions to intervene after the court has found the suit inappropriate for class action status.” 414 U. S., at 553. “A contrary rule,” the Court reasoned in American Pipe, “would deprive [Federal Rule of Civil Procedure] 23 class actions of the efficiency and economy of litigation which is a principal purpose of the procedure.” Ibid. This is so, the Court explained, because without tolling, “[p]otential class members would be induced to file protective motions to intervene or to join in the event that a class was later found unsuitable.” Ibid. In Crown, Cork, the Court further elaborated: Failure to extend the American Pipe rule “to class members filing separate actions,” in addition to those who move to intervene, would result in “a needless multiplicity of actions” filed by class members preserving their individual claims—“precisely the situation that Federal Rule of Civil Procedure 23 and the tolling rule of American Pipe were designed to avoid.” 462 U. S., at 351. American Pipe and Crown, Cork addressed only putative class members who wish to sue individually after a class-certification denial. See, e.g., American Pipe, 414 U. S., at 552 (addressing “privilege of intervening in an individual suit”); Crown, Cork, 462 U. S., at 349 (applying American Pipe to those who “file individual actions”); 462 U. S., at 352 (tolling benefits “class members who choose to file separate suits”). What about a putative class representative, like Resh, who brings his claims as a new class action after the statute of limitations has expired? Neither decision so much as hints that tolling extends to otherwise time-barred class claims. We hold that American Pipe does not permit a plaintiff who waits out the statute of limitations to piggyback on an earlier, timely filed class action. The “efficiency and economy of litigation” that support tolling of individual claims, American Pipe, 414 U. S., at 553, do not support maintenance of untimely successive class actions; any additional class filings should be made early on, soon after the commencement of the first action seeking class certification. American Pipe tolls the limitation period for individual claims because economy of litigation favors delaying those claims until after a class-certification denial. If certification is granted, the claims will proceed as a class and there would be no need for the assertion of any claim individually. If certification is denied, only then would it be necessary to pursue claims individually. With class claims, on the other hand, efficiency favors early assertion of competing class representative claims. If class treatment is appropriate, and all would-be representatives have come forward, the district court can select the best plaintiff with knowledge of the full array of po- tential class representatives and class counsel. And if the class mechanism is not a viable option for the claims, the decision denying certification will be made at the outset of the case, litigated once for all would-be class representatives.[2] Rule 23 evinces a preference for preclusion of untimely successive class actions by instructing that class certification should be resolved early on. See Fed. Rule Civ. Proc. 23(c)(1)(A). Indeed, Rule 23(c) was amended in 2003 to permit district courts to take account of multiple class-representative filings. Before the amendment, Rule 23(c) encouraged district courts to issue certification rulings “as soon as practicable.” The amendment changed the recommended timing target to “an early practicable time.” The alteration was made to allow greater leeway, more time for class discovery, and additional time to “explore designation of class counsel” and consider “additional [class counsel] applications rather than deny class certification,” thus “afford[ing] the best possible representation for the class.” Advisory Committee’s 2003 Note on subds. (c)(1)(A) and (g)(2)(A) of Fed. Rule Civ. Proc. 23, 28 U. S. C. App., pp. 815, 818; see Willging & Lee, From Class Actions to Multidistrict Consolidations: Aggregate Mass-Tort Litigation after Ortiz, 58 U. Kan. L. Rev. 775, 785 (2010) (2003 amendments “raised the standard for certifying a class from an early, conditional ruling to a later, relatively final decision” and “expand[ed] the opportunity for parties to engage in discovery prior to moving for class certification”). The PSLRA, which governs this litigation, evinces a similar preference, this time embodied in legislation, for grouping class-representative filings at the outset of litigation. See supra, at 3. When the Dean and Smyth timely commenced actions were first filed, counsel put any shareholder who might wish to serve as lead plaintiff on notice of the action. Several heeded the call—six in Dean and eight in Smyth. See 857 F. 3d, at 997–998. The PSLRA, by requiring notice of the commencement of a class action, aims to draw all potential lead plaintiffs into the suit so that the district court will have the full roster of contenders before deciding which contender to appoint.[3] See Brief for Securities Industry and Financial Markets Association as Amicus Curiae 12–13 (PSLRA “seeks to achieve Congress[’] goal of curbing duplicative . . . litigation by encouraging all interested parties to apply to serve as lead plaintiff at the early stages of the case [and] providing for the consolidation of similar class actions”). With notice and the opportunity to participate in the first (and second) round of class litigation, there is little reason to allow plaintiffs who passed up those opportunities to enter the fray several years after class proceedings first commenced. Ordinarily, to benefit from equitable tolling, plaintiffs must demonstrate that they have been diligent in pursuit of their claims. See, e.g., McQuiggin v. Perkins, 569 U. S. 383, 391 (2013); Menominee Tribe of Wis. v. United States, 577 U. S. ___, ___ (2016) (slip op., at 5). Even American Pipe, which did not analyze “criteria of the formal doctrine of equitable tolling in any direct manner,” ANZ, 582 U. S., at ___–___ (slip op., at 10–11), observed that tolling was permissible in the circumstances because plaintiffs who later intervened to pursue individual claims had not slept on their rights, American Pipe, 414 U. S., at 554–555. Those plaintiffs reasonably relied on the class representative, who sued timely, to protect their interests in their individual claims. See Crown, Cork, 462 U. S., at 350. A would-be class representative who commences suit after expiration of the limitation period, however, can hardly qualify as diligent in asserting claims and pursuing relief. Her interest in representing the class as lead plaintiff, therefore, would not be preserved by the prior plaintiff’s timely filed class suit. Respondents’ proposed reading would allow the statute of limitations to be extended time and again; as each class is denied certification, a new named plaintiff could file a class complaint that resuscitates the litigation. See Yang, 392 F. 3d, at 113 (Alito, J., concurring in part and dissenting in part) (tolling for successive class actions could allow “lawyers seeking to represent a plaintiff class [to] extend the statute of limitations almost indefinitely until they find a district court judge who is willing to certify the class”); Ewing Industries Corp. v. Bob Wines Nursery, Inc., 795 F. 3d 1324, 1326 (CA11 2015) (tolling for successive class actions allows plaintiffs “limitless bites at the apple”).[4] This prospect points up a further distinction between the individual-claim tolling established by American Pipe and tolling for successive class actions. The time to file individual actions once a class action ends is finite, extended only by the time the class suit was pending; the time for filing successive class suits, if tolling were allowed, could be limitless. Respondents’ claims happen to be governed by 28 U. S. C. §1658(b)(2)’s five-year statute of repose, so the time to file complaints has a finite end. Statutes of repose, however, are not ubiquitous. See Dekalb County Pension Fund v. Transocean Ltd., 817 F. 3d 393, 397 (CA2 2016). Most statutory schemes provide for a single limitation period without any outer limit to safeguard against serial relitigation. Endless tolling of a statute of limitations is not a result envisioned by American Pipe.[5] B Respondents emphasize that in Shady Grove Orthopedic Associates, P. A. v. Allstate Ins. Co., 559 U. S. 393 (2010), we said that “a class action may be maintained,” id., at 398 (internal quotation marks omitted), if the requirements of Rule 23(a) and (b) are satisfied, and “Rule 23 automatically applies in all civil actions and proceedings in the United States district courts,” id., at 400 (internal quotation marks omitted). See Brief for Respondents 21–23. If Resh’s suit meets the requirements of Rule 23(a) and (b), respondents assert, there is no reason why Resh’s suit cannot proceed as a class action. Shady Grove does not call for that outcome. In Shady Grove, the Court held that a federal diversity action could proceed under Rule 23 despite a state law prohibiting class treatment of suits seeking damages of the kind asserted in the Shady Grove complaint. 559 U. S., at 396, 416. Our opinion in Shady Grove addressed a case in which a Rule 23 class action could have been maintained absent a contrary state-law command. Id., at 396. Resh’s case presents the reverse situation: The class action would be untimely unless saved by American Pipe’s equitable-tolling exception to statutes of limitations. Rule 23 itself does not address timeliness of claims or tolling and nothing in the Rule calls for the revival of class claims if individual claims are tolled. In fact, as already explained, Rule 23 prescribes the opposite result. See supra, at 6–8. Today’s clarification of American Pipe’s reach does not run afoul of the Rules Enabling Act by causing a plaintiff’s attempted recourse to Rule 23 to abridge or modify a substantive right. See Brief for Respondents 23–26 (citing Tyson Foods, Inc. v. Bouaphakeo, 577 U. S. ___ (2016)). Plaintiffs have no substantive right to bring their claims outside the statute of limitations. That they may do so, in limited circumstances, is due to a judicially crafted tolling rule that itself does not abridge, enlarge, or modify any substantive right. American Pipe, 414 U. S., at 558. Without American Pipe, respondents would have no peg to seek tolling here; as we have explained, however, American Pipe does not provide for the extension of the statute of limitations sought by Resh for institution of an untimely third class suit. Respondents urge that American Pipe’s logic in fact supports their position because declining to toll the limitation period for successive class suits will lead to a “needless multiplicity” of protective class-action filings. Brief for Respondents 32–34. See also post, at 6–7 (expressing concern about duplicative and dueling class actions). But there is little reason to think that protective class filings will substantially increase. Several Courts of Appeals have already declined to read American Pipe to permit a successive class action filed outside the limitation period. See supra, at 5; 3 W. Rubenstein, Newberg on Class Actions §9:64, n. 5 (5th ed. 2012). These courts include the Second and Fifth Circuits (no strangers to class-action practice); both courts declined to entertain out-of-time class actions in the 1980’s. See Korwek, 827 F. 2d 874 (CA2 1987); Salazar-Calderon, 765 F. 2d 1334 (CA5 1985). Respondents and their amici make no showing that these Circuits have experienced a disproportionate number of duplicative, protective class-action filings. Amicus National Conference on Public Employee Retirement Systems cites examples of protective filings responding to courts’ disallowance of American Pipe tolling for statutes of repose, but those examples in fact suggest that protective class filings are uncommon. See Brief of the National Conference on Public Employee Retirement Systems as Amicus Curiae 7–8. Between dozens and hundreds of class plaintiffs filed protective individual claims while class-certification motions were pending in securities cases and the statute of repose was about to run out, placing a permanent bar against their claims. Ibid. But none of the plaintiffs appears to have filed a protective class action—even though, if the statute of repose expired and the pending class-certification motions were denied, there would be no further opportunity to assert class claims.[6] Nor do the incentives of class-action practice suggest that many more plaintiffs will file protective class claims as a result of our holding. Any plaintiff whose individual claim is worth litigating on its own rests secure in the knowledge that she can avail herself of American Pipe tolling if certification is denied to a first putative class. The plaintiff who seeks to preserve the ability to lead the class—whether because her claim is too small to make an individual suit worthwhile or because of an attendant financial benefit[7]—has every reason to file a class action early, and little reason to wait in the wings, giving another plaintiff first shot at representation. In any event, as previously explained, see supra, at 6–8, a multiplicity of class-action filings is not necessarily “needless.” Indeed, multiple filings may aid a district court in determining, early on, whether class treatment is warranted, and if so, which of the contenders would be the best representative. And sooner rather than later filings are just what Rule 23 encourages. See ibid. Multiple timely filings might not line up neatly; they could be filed in different districts, at different times—perhaps when briefing on class certification has already begun—or on behalf of only partially overlapping classes. See Wasserman, Dueling Class Actions, 80 B. U. L. Rev. 461, 464–465 (2000) (describing variety of “dueling” class filings). But district courts have ample tools at their disposal to manage the suits, including the ability to stay, consolidate, or transfer proceedings. District courts are increasingly familiar with overseeing such complex cases, given the surge in multidistrict litigation. See Cabraser & Issacharoff, The Participatory Class Action, 92 N. Y. U. L. Rev. 846, 850–851 (2017) (multidistrict litigation frequently combines individual suits and multiple putative class actions). The Federal Rules provide a range of mechanisms to aid courts in this endeavor. What the Rules do not offer is a reason to permit plaintiffs to exhume failed class actions by filing new, untimely class claims. * * * The watchwords of American Pipe are efficiency and economy of litigation, a principal purpose of Rule 23 as well. Extending American Pipe tolling to successive class actions does not serve that purpose. The contrary rule, allowing no tolling for out-of-time class actions, will propel putative class representatives to file suit well within the limitation period and seek certification promptly. For all the above-stated reasons, it is the rule we adopt today: Time to file a class action falls outside the bounds of American Pipe. Accordingly, the judgment of the Court of Appeals for the Ninth Circuit is reversed, and the case is remanded for further proceedings consistent with this opinion. It is so ordered. Notes 1 A statute of limitations “begin[s] to run when the cause of action accrues—that is, when the plaintiff can file suit and obtain relief.” California Public Employees’ Retirement System v. ANZ Securities, Inc., 582 U. S. ___, ___ (2017) (slip op., at 5) (internal quotation marks omitted). A statute of repose, by contrast, “begin[s] to run on the date of the last culpable act or omission of the defendant.” Ibid. (internal quotation marks omitted). 2 Encouraging early class filings will help ensure sufficient time remains under the statute of limitations, in the event that certification is denied for one of the actions or a portion of the class. Subclasses might be pleaded in one or more complaints and taken up if necessary; as class discovery proceeds and weaknesses in the class theory or adequacy of representation come to light, the lead complaint might be amended or a new plaintiff might intervene. See Brief of Plaintiffs in Post-Dukes Successor Class Actions as Amici Curiae 8–10 (describing regional subclasses asserted in the Dukes v. Wal-Mart litigation following this Court’s decision decertifying the nationwide class, Wal-Mart Stores, Inc. v. Dukes, 564 U. S. 338 (2011)); Pierce, Improving Predictability and Consistency in Class Action Tolling, 23 Geo. Mason L. Rev. 339, 349 (2016) (some Dukes plaintiffs moved to amend the original complaint to replead subclasses). 3 Although the Private Securities Litigation Reform Act of 1995 (PSLRA), 109Stat. 737, includes a presumption that the most adequate plaintiff is the one who moves first and has the largest financial interest in the case, see 15 U. S. C. §78u–4(a)(3)(B)(iii)(I), multiple potential lead plaintiffs have reason to apply for the role because there may not be an obvious candidate. Which plaintiff has the largest financial interest may not be immediately apparent; the statute does not define the term, and the size of a shareholder’s financial interest can depend on how many shares were purchased and sold, when, and at what price, as well as the order in which the losses are tallied. See, e.g., Cortina v. Anavex Life Sciences Corp., 2016 WL 1337305 (SDNY, Apr. 5, 2016). District courts often permit aggregation of plaintiffs into plaintiff groups, so even a small shareholder could apply for lead-plaintiff status, hoping to join with other shareholders to create a unit with the largest financial interest. See Choi & Thompson, Securities Litigation and Its Lawyers: Changes During the First Decade After the PSLRA, 106 Colum. L. Rev. 1489, 1507, 1521, 1530 (2006) (80% of securities class actions in post-PSLRA data sample had two or more co-lead counsel firms). Thus, it is a reasonable expectation that, in litigation governed by the PSLRA, a district court will have several competing candidates for lead plaintiff to choose among. 4 Respondents observe that in Smith v. Bayer Corp., 564 U. S. 299 (2011), we held that federal class-certification denials do not have preclusive effect in subsequent state-court suits, despite concerns about successive class actions. See Brief for Respondents 40–41. But in Smith, we were guided by “the fundamental nature of the general rule that only parties can be bound by prior judgments.” 564 U. S., at 313 (internal quotation marks omitted). The state-court plaintiffs were not parties to the federal-court litigation, hence they could not be bound by its holding—despite a “stron[g] argument” about the inefficiencies of serial class relitigation supporting the contrary position. Id., at 316. No such countervailing presumption favors Resh’s untimely third federal class suit. 5 Justice Sotomayor suggests that the Court might adopt a rule under which tolling “becomes unavailable for future class claims where class certification is denied for a reason that bears on the suitability of the claims for class treatment,” but not where “class certification is denied because of the deficiencies of the lead plaintiff as class representative.” Post, at 5; see Yang v. Odom, 392 F. 3d 97, 112 (CA3 2004) (embracing similar rule). But Rule 23 contains no instruction to give denials of class certification different effect based on the reason for the denial. And as the Advisory Committee Notes explain, affording district courts time to consider competing claims for class representation will advance the likelihood that lead plaintiff or class counsel deficiencies will be discovered and acted upon early in the litigation. See supra, at 7–8. Rule 23 and putative class members’ own interests in adequate representation, and the efficient adjudication thereof, weigh heavily against tolling for successive class actions. There is nothing inequitable in following these guides. See post, at 5, n. 2. 6 The Second Circuit Petrobras litigation, referenced in amicus’ brief, illustrates that multiple timely class filings do not sow unmanageable chaos. Five class actions were filed there and consolidated, along with individual claims, for pretrial purposes, including class-certification determination. See In re Petrobras Securities, 862 F. 3d 250, 258 (CA2 2017). 7 The class representative might receive a share of class recovery above and beyond her individual claim. See, e.g., Cook v. Niedert, 142 F. 3d 1004, 1016 (CA7 1998) (affirming class representative’s $25,000 incentive award).

583.US.2017_16-424

A federal grand jury indicted petitioner, Rodney Class, for possessing firearms in his locked jeep, which was parked on the grounds of the United States Capitol in Washington, D. C. See 40 U. S. C. §5104(e)(1) (“An individual . . . may not carry . . . on the Grounds or in any of the Capitol Buildings a firearm”). Appearing pro se, Class asked the District Court to dismiss the indictment. He alleged that the statute, §5104(e), violates the Second Amendment and the Due Process Clause. After the District Court dismissed both claims, Class pleaded guilty to “Possession of a Firearm on U. S. Capitol Grounds, in violation of 40 U. S. C. §5104(e).” App. 30. A written plea agreement set forth the terms of Class’ guilty plea, including several categories of rights that he agreed to waive. The agreement said nothing about the right to challenge on direct appeal the constitutionality of the statute of conviction. After conducting a hearing pursuant to Rule 11(b) of the Federal Rules of Criminal Procedure, the District Court accepted Class’ guilty plea and sentenced him. Soon thereafter, Class sought to raise his constitutional claims on direct appeal. The Court of Appeals held that Class could not do so because, by pleading guilty, he had waived his constitutional claims. Held: A guilty plea, by itself, does not bar a federal criminal defendant from challenging the constitutionality of his statute of conviction on direct appeal. Pp. 3–11. (a) This holding flows directly from this Court’s prior decisions. Fifty years ago, in Haynes v. United States, the Court addressed a similar claim challenging the constitutionality of a criminal statute. Justice Harlan’s opinion for the Court stated that the defendant’s “plea of guilty did not, of course, waive his previous [constitutional] claim.” 390 U. S. 85 , n. 2. That clear statement reflects an understanding of the nature of guilty pleas that stretches, in broad outline, nearly 150 years. Subsequent decisions have elaborated upon it. In Blackledge v. Perry, 417 U. S. 21 , the Court recognized that a guilty plea bars some “ ‘antecedent constitutional violations,’ ” related to events (such as grand jury proceedings) that “ ‘occu[r] prior to the entry of the guilty plea.’ ” Id., at 30 (quoting Tollett v. Henderson, 411 U. S. 258 –267). However, where the claim implicates “the very power of the State” to prosecute the defendant, a guilty plea cannot by itself bar it. 417 U. S., at 30. Likewise, in Menna v. New York, 423 U. S. 61 , the Court held that because the defendant’s claim was that “the State may not convict [him] no matter how validly his factual guilt is established,” his “guilty plea, therefore, [did] not bar the claim.” Id., at 63, n. 2. In more recent years, the Court has reaffirmed the Menna-Blackledge doctrine’s basic teaching that “ ‘a plea of guilty to a charge does not waive a claim that—judged on its face—the charge is one which the State may not constitutionally prosecute.’ ” United States v. Broce, 488 U. S. 563 (quoting Menna, supra, at 63, n. 2). Pp. 3–7. (b) In this case, Class neither expressly nor implicitly waived his constitutional claims by pleading guilty. As this Court understands them, the claims at issue here do not contradict the terms of the indictment or the written plea agreement and they can be resolved “on the basis of the existing record.” Broce, supra, at 575. Class challenges the Government’s power to criminalize his (admitted) conduct and thereby calls into question the Government’s power to “ ‘constitutionally prosecute’ ” him. Ibid. (quoting Menna, supra, at 61–62, n. 2). A guilty plea does not bar a direct appeal in these circumstances. Pp. 7–8. (c) Federal Rule of Criminal Procedure 11(a)(2), which governs “conditional” guilty pleas, cannot resolve this case. By its own terms, the Rule does not say whether it sets forth the exclusive procedure for a defendant to preserve a constitutional claim following a guilty plea. And the Rule’s drafters acknowledged that the “Supreme Court has held that certain kinds of constitutional objections may be raised after a plea of guilty” and specifically stated that Rule 11(a)(2) “has no application” to the “kinds of constitutional objections” that may be raised under the “Menna-Blackledge doctrine.” Advisory Committee’s Notes on 1983 Amendments to Fed. Rule Crim. Proc. 11, 18 U. S. C. App., p. 912. Because the applicability of the Menna-Blackledge doctrine is at issue here, Rule 11(a)(2) cannot resolve this case. Pp. 8–10. Reversed and remanded. Breyer, J., delivered the opinion of the Court, in which Roberts, C. J., and Ginsburg, Sotomayor, Kagan, and Gorsuch, JJ., joined. Alito, J., filed a dissenting opinion, in which Kennedy and Thomas, JJ., joined.

Does a guilty plea bar a criminal defendant from later appealing his conviction on the ground that the statute of conviction violates the Constitution? In our view, a guilty plea by itself does not bar that appeal. I In September 2013, a federal grand jury indicted petitioner, Rodney Class, for possessing firearms in his locked jeep, which was parked in a lot on the grounds of the United States Capitol in Washington, D. C. See 40 U. S. C. §5104(e)(1) (“An individual . . . may not carry . . . on the Grounds or in any of the Capitol Buildings a firearm”). Soon thereafter, Class, appearing pro se, asked the Federal District Court for the District of Columbia to dismiss the indictment. As relevant here, Class alleged that the statute, §5104(e), violates the Second Amendment. App. in No. 15–3015 (CADC), pp. 32–33. He also raised a due process claim, arguing that he was denied fair notice that weapons were banned in the parking lot. Id., at 39. Following a hearing, the District Court denied both claims. App. to Pet. for Cert. 9a. Several months later, Class pleaded guilty to “Possession of a Firearm on U. S. Capitol Grounds, in violation of 40 U. S. C. §5104(e).” App. 30. The Government agreed to drop related charges. Id., at 31. A written plea agreement set forth the terms of Class’ guilty plea, including several categories of rights that he expressly agreed to waive. Those express waivers included: (1) all defenses based upon the statute of limitations; (2) several specified trial rights; (3) the right to appeal a sentence at or below the judicially determined, maximum sentencing guideline range; (4) most collateral attacks on the conviction and sentence; and (5) various rights to request or receive information concerning the investigation and prosecution of his criminal case. Id., at 38–42. At the same time, the plea agreement expressly enumerated categories of claims that Class could raise on appeal, including claims based upon (1) newly discovered evidence; (2) ineffective assistance of counsel; and (3) certain statutes providing for sentence reductions. Id., at 41. Finally, the plea agreement stated under the heading “Complete Agreement”: “No agreements, promises, understandings, or representations have been made by the parties or their counsel other than those contained in writing herein, nor will any such agreements . . . be made unless committed to writing and signed . . . .” Id., at 45. The agreement said nothing about the right to raise on direct appeal a claim that the statute of conviction was unconstitutional. The District Court held a plea hearing during which it reviewed the terms of the plea agreement (with Class present and under oath) to ensure the validity of the plea. See Fed. Rule Crim. Proc. 11(b); United States v. Ruiz, 536 U. S. 622, 629 (2002) (defendant’s guilty plea must be “ ‘voluntary’ ” and “related waivers” must be made “ ‘knowing[ly], intelligent[ly], [and] with sufficient awareness of the relevant circumstances and likely consequences’ ”). After providing Class with the required information and warnings, the District Court accepted his guilty plea. Class was sentenced to 24 days imprisonment followed by 12 months of supervised release. Several days later, Class appealed his conviction to the Court of Appeals for the District of Columbia Circuit. Class was appointed an amicus to aid him in presenting his arguments. He repeated his constitutional claims, namely, that the statute violates the Second Amendment and the Due Process Clause because it fails to give fair notice of which areas fall within the Capitol Grounds where firearms are banned. The Court of Appeals held that Class could not raise his constitutional claims because, by pleading guilty, he had waived them. App. to Pet. for Cert. 1a–5a. Class filed a petition for certiorari in this Court asking us to decide whether in pleading guilty a criminal defendant inherently waives the right to challenge the constitutionality of his statute of conviction. We agreed to do so. II The question is whether a guilty plea by itself bars a federal criminal defendant from challenging the constitutionality of the statute of conviction on direct appeal. We hold that it does not. Class did not relinquish his right to appeal the District Court’s constitutional determinations simply by pleading guilty. As we shall explain, this holding flows directly from this Court’s prior decisions. Fifty years ago this Court directly addressed a similar claim (a claim that the statute of conviction was unconstitutional). And the Court stated that a defendant’s “plea of guilty did not . . . waive his previous [constitutional] claim.” Haynes v. United States, 390 U. S. 85 , n. 2 (1968). Though Justice Harlan’s opinion for the Court in Haynes offered little explanation for this statement, subsequent decisions offered a rationale that applies here. In Blackledge v. Perry, 417 U. S. 21 (1974) , North Carolina indicted and convicted Jimmy Seth Perry on a misdemeanor assault charge. When Perry exercised his right under a North Carolina statute to a de novo trial in a higher court, the State reindicted him, but this time the State charged a felony, which carried a heavier penalty, for the same conduct. Perry pleaded guilty. He then sought habeas relief on the grounds that the reindictment amounted to an unconstitutional vindictive prosecution. The State argued that Perry’s guilty plea barred him from raising his constitutional challenge. But this Court held that it did not. The Court noted that a guilty plea bars appeal of many claims, including some “ ‘antecedent constitutional violations’ ” related to events (say, grand jury proceedings) that had “ ‘occurred prior to the entry of the guilty plea.’ ” Id., at 30 (quoting Tollett v. Henderson, 411 U. S. 258 –267 (1973)). While Tollett claims were “of constitutional dimension,” the Court explained that “the nature of the underlying constitutional infirmity is markedly different” from a claim of vindictive prosecution, which implicates “the very power of the State” to prosecute the defendant. Blackledge, 417 U. S., at 30. Accordingly, the Court wrote that “the right” Perry “asserts and that we today accept is the right not to be haled into court at all upon the felony charge” since “[t]he very initiation of the proceedings” against Perry “operated to deprive him due process of law.” Id., at 30–31. A year and a half later, in Menna v. New York, 423 U. S. 61 (1975) (per curiam), this Court repeated what it had said and held in Blackledge. After Menna served a 30-day jail term for refusing to testify before the grand jury on November 7, 1968, the State of New York charged him once again for (what Menna argued was) the same crime. Menna pleaded guilty, but subsequently appealed arguing that the new charge violated the Double Jeopardy Clause. U. S. Const., Amdt. 5. The lower courts held that Menna’s constitutional claim had been “waived” by his guilty plea. This Court reversed. Citing Blackledge, supra, at 30, the Court held that “a plea of guilty to a charge does not waive a claim that—judged on its face—the charge is one which the State may not constitutionally prosecute.” Menna, 423 U. S., at 63, and n. 2. Menna’s claim amounted to a claim that “the State may not convict” him “no matter how validly his factual guilt is established.” Ibid. Menna’s “guilty plea, therefore, [did] not bar the claim.” Ibid. These holdings reflect an understanding of the nature of guilty pleas which, in broad outline, stretches back nearly 150 years. In 1869 Justice Ames wrote for the Supreme Judicial Court of Massachusetts: “The plea of guilty is, of course, a confession of all the facts charged in the indictment, and also of the evil intent imputed to the defendant. It is a waiver also of all merely technical and formal objections of which the defendant could have availed himself by any other plea or motion. But if the facts alleged and admitted do not constitute a crime against the laws of the Commonwealth, the defendant is entitled to be discharged.” Commonwealth v. Hinds, 101 Mass. 209, 210. Decisions of federal and state courts throughout the 19th and 20th centuries reflect a similar view of the nature of a guilty plea. See United States v. Ury, 106 F. 2d 28 (CA2 1939) (holding the “plea of guilty did not foreclose the appellant,” who argued that a statute was unconstitutional, “from the review he now seeks” (citing earlier cases)); Hocking Valley R. Co. v. United States, 210 F. 735 (CA6 1914) (holding that a defendant may raise the claim that, because the indictment did not charge an offense no crime has been committed, for it is “the settled rule that,” despite a guilty plea, a defendant “may urge” such a contention “in the reviewing court”); Carper v. State, 27 Ohio St. 572, 575 (1875) (same). We refer to these cases because it was against this background that Justice Harlan in his opinion for the Court made the statement to which we originally referred, namely, that a defendant’s “plea of guilty did not, of course, waive his previous [constitutional] claim.” Haynes, 390 U. S., at 87, n. 2 (citing Ury, supra, at 28). In more recent years, we have reaffirmed the Menna- Blackledge doctrine and refined its scope. In United States v. Broce, 488 U. S. 563 (1989) , the defendants pleaded guilty to two separate indictments in a single proceeding which “on their face” described two separate bid-rigging conspiracies. Id., at 576. They later sought to challenge their convictions on double jeopardy grounds, arguing that they had only admitted to one conspiracy. Citing Blackledge and Menna, this Court repeated that a guilty plea does not bar a claim on appeal “where on the face of the record the court had no power to enter the conviction or impose the sentence.” 488 U. S., at 569. However, because the defendants could not “prove their claim by relying on those indictments and the existing record” and “without contradicting those indictments,” this Court held that their claims were “foreclosed by the admissions inherent in their guilty pleas.” Id., at 576. Unlike the claims in Broce, Class’ constitutional claims here, as we understand them, do not contradict the terms of the indictment or the written plea agreement. They are consistent with Class’ knowing, voluntary, and intelligent admission that he did what the indictment alleged. Those claims can be “resolved without any need to venture beyond that record.” Id., at 575. Nor do Class’ claims focus upon case-related constitutional defects that “ ‘occurred prior to the entry of the guilty plea.’ ” Blackledge, 417 U. S., at 30. They could not, for example, “have been ‘cured’ through a new indictment by a properly selected grand jury.” Ibid. (citing Tollett, 411 U. S., at 267). Because the defendant has admitted the charges against him, a guilty plea makes the latter kind of constitutional claim “irrelevant to the constitutional validity of the conviction.” Haring v. Prosise, 462 U. S. 306, 321 (1983) . But the cases to which we have referred make clear that a defendant’s guilty plea does not make irrelevant the kind of constitutional claim Class seeks to make. In sum, the claims at issue here do not fall within any of the categories of claims that Class’ plea agreement forbids him to raise on direct appeal. They challenge the Government’s power to criminalize Class’ (admitted) conduct. They thereby call into question the Government’s power to “ ‘constitutionally prosecute’ ” him. Broce, supra, at 575 (quoting Menna, supra, at 61–62, n. 2). A guilty plea does not bar a direct appeal in these circumstances. III We are not convinced by the three basic arguments that the Government and the dissent make in reply. First, the Government contends that by entering a guilty plea, Class inherently relinquished his constitutional claims. The Government is correct that a guilty plea does implicitly waive some claims, including some constitutional claims. However, as we explained in Part II, supra, Class’ valid guilty plea does not, by itself, bar direct appeal of his constitutional claims in these circumstances. As an initial matter, a valid guilty plea “forgoes not only a fair trial, but also other accompanying constitutional guarantees.” Ruiz, 536 U. S., at 628–629. While those “simultaneously” relinquished rights include the privilege against compulsory self-incrimination, the jury trial right, and the right to confront accusers, McCarthy v. United States, 394 U. S. 459, 466 (1969) , they do not include “a waiver of the privileges which exist beyond the confines of the trial.” Mitchell v. United States, 526 U. S. 314, 324 (1999) . Here, Class’ statutory right directly to appeal his conviction “cannot in any way be characterized as part of the trial.” Lafler v. Cooper, 566 U. S. 156, 165 (2012) . A valid guilty plea also renders irrelevant—and thereby prevents the defendant from appealing—the constitutionality of case-related government conduct that takes place before the plea is entered. See, e.g., Haring, supra, at 320 (holding a valid guilty plea “results in the defendant’s loss of any meaningful opportunity he might otherwise have had to challenge the admissibility of evidence obtained in violation of the Fourth Amendment”). Neither can the defendant later complain that the indicting grand jury was unconstitutionally selected. Tollett, supra, at 266. But, as we have said, those kinds of claims are not at issue here. Finally, a valid guilty plea relinquishes any claim that would contradict the “admissions necessarily made upon entry of a voluntary plea of guilty.” Broce, supra, at 573–574. But the constitutional claim at issue here is consistent with Class’ admission that he engaged in the conduct alleged in the indictment. Unlike the defendants in Broce, Class’ challenge does not in any way deny that he engaged in the conduct to which he admitted. Instead, like the defendants in Blackledge and Menna, he seeks to raise a claim which, “ ‘judged on its face’ ” based upon the existing record, would extinguish the government’s power to “ ‘constitutionally prosecute’ ” the defendant if the claim were successful. Broce, supra, at 575 (quoting Menna, 423 U. S., at 62–63, and n. 2). Second, the Government and the dissent point to Rule 11(a)(2) of the Federal Rules of Criminal Procedure, which governs “conditional” guilty pleas. The Rule states: “Conditional Plea. With the consent of the court and the government, a defendant may enter a conditional plea of guilty or nolo contendere, reserving in writing the right to have an appellate court review an adverse determination of a specified pretrial motion. A defendant who prevails on appeal may then withdraw the plea.” The Government and the dissent argue that Rule 11(a)(2) means that “a defendant who pleads guilty cannot challenge his conviction on appeal on a forfeitable or waivable ground that he either failed to present to the district court or failed to reserve in writing.” Brief for United States 23; see also post, at 3–4, 17–18 (opinion of Alito, J.). They support this argument by pointing to the notes of the Advisory Committee that drafted the text of Rule 11(a)(2). See Advisory Committee’s Notes on 1983 Amendments to Fed. Rule Crim. Proc. 11, 18 U. S. C. App., p. 911 (hereinafter Advisory Committee’s Notes). In particular, the dissent points to the suggestion that an unconditional guilty plea constitutes a waiver of “nonjurisdictional defects,” while the Government points to the drafters’ statement that they intended the Rule’s “conditional plea procedure . . . to conserve prosecutorial and judicial resources and advance speedy trial objectives,” while ensuring “much needed uniformity in the federal system on this matter.” Ibid.; see United States v. Vonn, 535 U. S. 55 , n. 6 (2002) (approving of Advisory Committee’s Notes as relevant evidence of the drafters’ intent). The Government adds that its interpretation of the Rule furthers these basic purposes. And, the argument goes, just as defendants must use Rule 11(a)(2)’s procedures to preserve, for instance, Fourth Amendment unlawful search-and-seizure claims, so must they use it to preserve the constitutional claims at issue here. The problem with this argument is that, by its own terms, the Rule itself does not say whether it sets forth the exclusive procedure for a defendant to preserve a constitutional claim following a guilty plea. At the same time, the drafters’ notes acknowledge that the “Supreme Court has held that certain kinds of constitutional objections may be raised after a plea of guilty.” Advisory Committee’s Notes, at 912. The notes then specifically refer to the “Menna-Blackledge doctrine.” Ibid. They add that the Rule “should not be interpreted as either broadening or narrowing [that] doctrine or as establishing procedures for its application.” Ibid. And the notes state that Rule 11(a)(2) “has no application” to the “kinds of constitutional objections” that may be raised under that doctrine. Ibid. The applicability of the Menna-Blackledge doctrine is at issue in this case. Cf. Broce, 488 U. S., at 569 (acknowledging Menna and Blackledge as covering claims “where on the face of the record the court had no power to enter the conviction or impose the sentence”). We therefore hold that Rule 11(a)(2) cannot resolve this case. Third, the Government argues that Class “expressly waived” his right to appeal his constitutional claim. Brief for United States 15. The Government concedes that the written plea agreement, which sets forth the “Complete Agreement” between Class and the Government, see App. 45–46, does not contain this waiver. Id., at 48–49. Rather, the Government relies on the fact that during the Rule 11 plea colloquy, the District Court Judge stated that, under the written plea agreement, Class was “giving up [his] right to appeal [his] conviction.” Id., at 76. And Class agreed. We do not see why the District Court Judge’s statement should bar Class’ constitutional claims. It was made to ensure Class understood “the terms of any plea-agreement provision waiving the right to appeal or to collaterally attack the sentence.” Fed. Rule Crim. Proc. 11(b)(1)(N). It does not expressly refer to a waiver of the appeal right here at issue. And if it is interpreted as expressly including that appeal right, it was wrong, as the Government acknowledged at oral argument. See Tr. of Oral Arg. 35–36. Under these circumstances, Class’ acquiescence neither expressly nor implicitly waived his right to appeal his constitutional claims. * * * For these reasons, we hold that Rodney Class may pursue his constitutional claims on direct appeal. The contrary judgment of the Court of Appeals for the District of Columbia Circuit is reversed, and the case is remanded for further proceedings consistent with this opinion. It is so ordered.

584.US.2017_16-1027

During the investigation of two traffic incidents involving an orange and black motorcycle with an extended frame, Officer David Rhodes learned that the motorcycle likely was stolen and in the possession of petitioner Ryan Collins. Officer Rhodes discovered photographs on Collins’ Facebook profile of an orange and black motorcycle parked in the driveway of a house, drove to the house, and parked on the street. From there, he could see what appeared to be the motorcycle under a white tarp parked in the same location as the motorcycle in the photograph. Without a search warrant, Office Rhodes walked to the top of the driveway, removed the tarp, confirmed that the motorcycle was stolen by running the license plate and vehicle identification numbers, took a photograph of the uncovered motorcycle, replaced the tarp, and returned to his car to wait for Collins. When Collins returned, Officer Rhodes arrested him. The trial court denied Collins’ motion to suppress the evidence on the ground that Officer Rhodes violated the Fourth Amendment when he trespassed on the house’s curtilage to conduct a search, and Collins was convicted of receiving stolen property. The Virginia Court of Appeals affirmed. The State Supreme Court also affirmed, holding that the warrantless search was justified under the Fourth Amendment’s automobile exception. Held: The automobile exception does not permit the warrantless entry of a home or its curtilage in order to search a vehicle therein. Pp. 3–14. (a) This case arises at the intersection of two components of the Court’s Fourth Amendment jurisprudence: the automobile exception to the warrant requirement and the protection extended to the curtilage of a home. In announcing each of the automobile exception’s justifications—i.e., the “ready mobility of the automobile” and “the pervasive regulation of vehicles capable of traveling on the public highways,” California v. Carney, 471 U. S. 386, 390, 392—the Court emphasized that the rationales applied only to automobiles and not to houses, and therefore supported their different treatment as a constitutional matter. When these justifications are present, officers may search an automobile without a warrant so long as they have probable cause. Curtilage—“the area ‘immediately surrounding and associated with the home’ ”—is considered “ ‘part of the home itself for Fourth Amendment purposes.’ ” Florida v. Jardines, 569 U. S. 1, 6. Thus, when an officer physically intrudes on the curtilage to gather evidence, a Fourth Amendment search has occurred and is presumptively unreasonable absent a warrant. Pp. 3–6. (b) As an initial matter, the part of the driveway where Collins’ motorcycle was parked and subsequently searched is curtilage. When Officer Rhodes searched the motorcycle, it was parked inside a partially enclosed top portion of the driveway that abuts the house. Just like the front porch, side garden, or area “outside the front window,” that enclosure constitutes “an area adjacent to the home and ‘to which the activity of home life extends.’ ” Jardines, 569 U. S., at 6, 7. Because the scope of the automobile exception extends no further than the automobile itself, it did not justify Officer Rhodes’ invasion of the curtilage. Nothing in this Court’s case law suggests that the automobile exception gives an officer the right to enter a home or its curtilage to access a vehicle without a warrant. Such an expansion would both undervalue the core Fourth Amendment protection afforded to the home and its curtilage and “ ‘untether’ ” the exception “ ‘from the justifications underlying’ ” it. Riley v. California, 573 U. S. ___, ___. This Court has similarly declined to expand the scope of other exceptions to the warrant requirement. Thus, just as an officer must have a lawful right of access to any contraband he discovers in plain view in order to seize it without a warrant—see Horton v. California, 496 U. S. 128, 136–137—and just as an officer must have a lawful right of access in order to arrest a person in his home—see Payton v. New York, 445 U. S. 573, 587–590—so, too, an officer must have a lawful right of access to a vehicle in order to search it pursuant to the automobile exception. To allow otherwise would unmoor the exception from its justifications, render hollow the core Fourth Amendment protection the Constitution extends to the house and its curtilage, and transform what was meant to be an exception into a tool with far broader application. Pp. 6–11. (c) Contrary to Virginia’s claim, the automobile exception is not a categorical one that permits the warrantless search of a vehicle anytime, anywhere, including in a home or curtilage. Scher v. United States, 305 U. S. 251; Pennsylvania v. Labron, 518 U. S. 938, distinguished. Also unpersuasive is Virginia’s proposed bright line rule for an automobile exception that would not permit warrantless entry only of the house itself or another fixed structure, e.g., a garage, inside the curtilage. This Court has long been clear that curtilage is afforded constitutional protection, and creating a carveout for certain types of curtilage seems more likely to create confusion than does uniform application of the Court’s doctrine. Virginia’s rule also rests on a mistaken premise, for the ability to observe inside curtilage from a lawful vantage point is not the same as the right to enter curtilage without a warrant to search for information not otherwise accessible. Finally, Virginia’s rule automatically would grant constitutional rights to those persons with the financial means to afford residences with garages but deprive those persons without such resources of any individualized consideration as to whether the areas in which they store their vehicles qualify as curtilage. Pp. 11–14. 292 Va. 486, 790 S. E. 2d 611, reversed and remanded. Sotomayor, J., delivered the opinion of the Court, in which Roberts, C. J., and Kennedy, Thomas, Ginsburg, Breyer, Kagan, and Gorsuch, JJ., joined. Thomas, J., filed a concurring opinion. Alito, J., filed a dissenting opinion.

This case presents the question whether the automobile exception to the Fourth Amendment permits a police officer, uninvited and without a warrant, to enter the curtilage of a home in order to search a vehicle parked therein. It does not.IOfficer Matthew McCall of the Albemarle County Police Department in Virginia saw the driver of an orange and black motorcycle with an extended frame commit a traffic infraction. The driver eluded Officer McCall’s attempt to stop the motorcycle. A few weeks later, Officer David Rhodes of the same department saw an orange and black motorcycle traveling well over the speed limit, but the driver got away from him, too. The officers compared notes and concluded that the two incidents involved the same motorcyclist.Upon further investigation, the officers learned that the motorcycle likely was stolen and in the possession of petitioner Ryan Collins. After discovering photographs on Collins’ Facebook profile that featured an orange and black motorcycle parked at the top of the driveway of a house, Officer Rhodes tracked down the address of the house, drove there, and parked on the street. It was later established that Collins’ girlfriend lived in the house and that Collins stayed there a few nights per week.[1]From his parked position on the street, Officer Rhodes saw what appeared to be a motorcycle with an extended frame covered with a white tarp, parked at the same angle and in the same location on the driveway as in the Facebook photograph. Officer Rhodes, who did not have a warrant, exited his car and walked toward the house. He stopped to take a photograph of the covered motorcycle from the sidewalk, and then walked onto the residential property and up to the top of the driveway to where the motorcycle was parked. In order “to investigate further,” App. 80, Officer Rhodes pulled off the tarp, revealing a motorcycle that looked like the one from the speeding incident. He then ran a search of the license plate and vehicle identification numbers, which confirmed that the motorcycle was stolen. After gathering this information, Officer Rhodes took a photograph of the uncovered motorcycle, put the tarp back on, left the property, and returned to his car to wait for Collins.Shortly thereafter, Collins returned home. Officer Rhodes walked up to the front door of the house and knocked. Collins answered, agreed to speak with Officer Rhodes, and admitted that the motorcycle was his and that he had bought it without title. Officer Rhodes then arrested Collins.Collins was indicted by a Virginia grand jury for receiving stolen property. He filed a pretrial motion to suppress the evidence that Officer Rhodes had obtained as a result of the warrantless search of the motorcycle. Collins argued that Officer Rhodes had trespassed on the curtilage of the house to conduct an investigation in violation of the Fourth Amendment. The trial court denied the motion and Collins was convicted.The Court of Appeals of Virginia affirmed. It assumed that the motorcycle was parked in the curtilage of the home and held that Officer Rhodes had probable cause to believe that the motorcycle under the tarp was the same motorcycle that had evaded him in the past. It further concluded that Officer Rhodes’ actions were lawful under the Fourth Amendment even absent a warrant because “numerous exigencies justified both his entry onto the property and his moving the tarp to view the motorcycle and record its identification number.” 65 Va. App. 37, 46, 773 S. E. 2d 618, 623 (2015).The Supreme Court of Virginia affirmed on different reasoning. It explained that the case was most properly resolved with reference to the Fourth Amendment’s automobile exception. 292 Va. 486, 496–501, 790 S. E. 2d 611, 616–618 (2016). Under that framework, it held that Officer Rhodes had probable cause to believe that the motorcycle was contraband, and that the warrantless search therefore was justified. Id., at 498–499, 790 S. E. 2d, at 617.We granted certiorari, 582 U. S. ___ (2017), and now reverse.IIThe Fourth Amendment provides in relevant part that the “right of the people to be secure in their persons, houses, papers, and effects, against unreasonable searches and seizures, shall not be violated.” This case arises at the intersection of two components of the Court’s Fourth Amendment jurisprudence: the automobile exception to the warrant requirement and the protection extended to the curtilage of a home.A1The Court has held that the search of an automobile can be reasonable without a warrant. The Court first articulated the so-called automobile exception in Carroll v. United States,267 U. S. 132 (1925). In that case, law enforcement officers had probable cause to believe that a car they observed traveling on the road contained illegal liquor. They stopped and searched the car, discovered and seized the illegal liquor, and arrested the occupants. Id., at 134–136. The Court upheld the warrantless search and seizure, explaining that a “necessary difference” exists between searching “a store, dwelling house or other structure” and searching “a ship, motor boat, wagon or automobile” because a “vehicle can be quickly moved out of the locality or jurisdiction in which the warrant must be sought.” Id., at 153.The “ready mobility” of vehicles served as the core justification for the automobile exception for many years. California v. Carney,471 U. S. 386, 390 (1985) (citing, e.g., Cooper v. California,386 U. S. 58, 59 (1967); Chambers v. Maroney,399 U. S. 42, 51–52 (1970)). Later cases then introduced an additional rationale based on “the pervasive regulation of vehicles capable of traveling on the public highways.” Carney, 471 U. S., at 392. As the Court explained in South Dakota v. Opperman,428 U. S. 364 (1976):“Automobiles, unlike homes, are subjected to pervasive and continuing governmental regulation and controls, including periodic inspection and licensing requirements. As an everyday occurrence, police stop and examine vehicles when license plates or inspection stickers have expired, or if other violations, such as exhaust fumes or excessive noise, are noted, or if headlights or other safety equipment are not in proper working order.” Id., at 368.In announcing each of these two justifications, the Court took care to emphasize that the rationales applied only to automobiles and not to houses, and therefore supported “treating automobiles differently from houses” as a constitutional matter. Cady v. Dombrowski,413 U. S. 433, 441 (1973).When these justifications for the automobile exception “come into play,” officers may search an automobile without having obtained a warrant so long as they have probable cause to do so. Carney, 471 U. S., at 392–393.2Like the automobile exception, the Fourth Amendment’s protection of curtilage has long been black letter law. “[W]hen it comes to the Fourth Amendment, the home is first among equals.” Florida v. Jardines,569 U. S. 1, 6 (2013). “At the Amendment’s ‘very core’ stands ‘the right of a man to retreat into his own home and there be free from unreasonable governmental intrusion.’ ” Ibid. (quoting Silverman v. United States,365 U. S. 505, 511 (1961)). To give full practical effect to that right, the Court considers curtilage—“the area ‘immediately surrounding and associated with the home’ ”—to be “ ‘part of the home itself for Fourth Amendment purposes.’ ” Jardines, 569 U. S., at 6 (quoting Oliver v. United States,466 U. S. 170, 180 (1984)). “The protection afforded the curtilage is essentially a protection of families and personal privacy in an area intimately linked to the home, both physically and psychologically, where privacy expectations are most heightened.” California v. Ciraolo,476 U. S. 207, 212–213 (1986).When a law enforcement officer physically intrudes on the curtilage to gather evidence, a search within the meaning of the Fourth Amendment has occurred. Jardines, 569 U. S., at 11. Such conduct thus is presumptively unreasonable absent a warrant.B1With this background in mind, we turn to the application of these doctrines in the instant case. As an initial matter, we decide whether the part of the driveway where Collins’ motorcycle was parked and subsequently searched is curtilage.According to photographs in the record, the driveway runs alongside the front lawn and up a few yards past the front perimeter of the house. The top portion of the driveway that sits behind the front perimeter of the house is enclosed on two sides by a brick wall about the height of a car and on a third side by the house. A side door provides direct access between this partially enclosed section of the driveway and the house. A visitor endeavoring to reach the front door of the house would have to walk partway up the driveway, but would turn off before entering the enclosure and instead proceed up a set of steps leading to the front porch. When Officer Rhodes searched the motorcycle, it was parked inside this partially enclosed top portion of the driveway that abuts the house.The “ ‘conception defining the curtilage’ is . . . familiar enough that it is ‘easily understood from our daily experience.’ ” Jardines, 569 U. S., at 7 (quoting Oliver, 466 U. S., at 182, n. 12). Just like the front porch, side garden, or area “outside the front window,” Jardines, 569 U. S., at 6, the driveway enclosure where Officer Rhodes searched the motorcycle constitutes “an area adjacent to the home and ‘to which the activity of home life extends,’ ” and so is properly considered curtilage, id., at 7 (quoting Oliver, 466 U. S., at 182, n. 12).2In physically intruding on the curtilage of Collins’ home to search the motorcycle, Officer Rhodes not only invaded Collins’ Fourth Amendment interest in the item searched, i.e., the motorcycle, but also invaded Collins’ Fourth Amendment interest in the curtilage of his home. The question before the Court is whether the automobile exception justifies the invasion of the curtilage.[2] The answer is no.Applying the relevant legal principles to a slightly different factual scenario confirms that this is an easy case. Imagine a motorcycle parked inside the living room of a house, visible through a window to a passerby on the street. Imagine further that an officer has probable cause to believe that the motorcycle was involved in a traffic infraction. Can the officer, acting without a warrant, enter the house to search the motorcycle and confirm whether it is the right one? Surely not.The reason is that the scope of the automobile exception extends no further than the automobile itself. See, e.g., Pennsylvania v. Labron,518 U. S. 938, 940 (1996) (per curiam) (explaining that the automobile exception “permits police to search the vehicle”); Wyoming v. Houghton,526 U. S. 295, 300 (1999) (“[T]he Framers would have regarded as reasonable (if there was probable cause) the warrantless search of containers within an automobile”). Virginia asks the Court to expand the scope of the automobile exception to permit police to invade any space outside an automobile even if the Fourth Amendment protects that space. Nothing in our case law, however, suggests that the automobile exception gives an officer the right to enter a home or its curtilage to access a vehicle without a warrant. Expanding the scope of the automobile exception in this way would both undervalue the core Fourth Amendment protection afforded to the home and its curtilage and “ ‘untether’ ” the automobile exception “ ‘from the justifications underlying’ ” it. Riley v. California, 573 U. S. ___, ___ (2014) (slip op., at 10) (quoting Arizona v. Gant,556 U. S. 332, 343 (2009)).The Court already has declined to expand the scope of other exceptions to the warrant requirement to permit warrantless entry into the home. The reasoning behind those decisions applies equally well in this context. For instance, under the plain-view doctrine, “any valid warrantless seizure of incriminating evidence” requires that the officer “have a lawful right of access to the object itself.” Horton v. California,496 U. S. 128, 136–137 (1990); see also id., at 137, n. 7 (“ ‘[E]ven where the object is contraband, this Court has repeatedly stated and enforced the basic rule that the police may not enter and make a warrantless seizure’ ”); G. M. Leasing Corp. v. United States,429 U. S. 338, 354 (1977) (“It is one thing to seize without a warrant property resting in an open area . . . , and it is quite another thing to effect a warrantless seizure of property . . . situated on private premises to which access is not otherwise available for the seizing officer”). A plain-view seizure thus cannot be justified if it is effectuated “by unlawful trespass.” Soldal v. Cook County,506 U. S. 56, 66 (1992). Had Officer Rhodes seen illegal drugs through the window of Collins’ house, for example, assuming no other warrant exception applied, he could not have entered the house to seize them without first obtaining a warrant.Similarly, it is a “settled rule that warrantless arrests in public places are valid,” but, absent another exception such as exigent circumstances, officers may not enter a home to make an arrest without a warrant, even when they have probable cause. Payton v. New York,445 U. S. 573, 587–590 (1980). That is because being “ ‘arrested in the home involves not only the invasion attendant to all arrests but also an invasion of the sanctity of the home.’ ” Id., at 588–589 (quoting United States v. Reed, 572 F. 2d 412, 423 (CA2 1978)). Likewise, searching a vehicle parked in the curtilage involves not only the invasion of the Fourth Amendment interest in the vehicle but also an invasion of the sanctity of the curtilage.Just as an officer must have a lawful right of access to any contraband he discovers in plain view in order to seize it without a warrant, and just as an officer must have a lawful right of access in order to arrest a person in his home, so, too, an officer must have a lawful right of access to a vehicle in order to search it pursuant to the automobile exception. The automobile exception does not afford the necessary lawful right of access to search a vehicle parked within a home or its curtilage because it does not justify an intrusion on a person’s separate and substantial Fourth Amendment interest in his home and curtilage.As noted, the rationales underlying the automobile exception are specific to the nature of a vehicle and the ways in which it is distinct from a house. See Part II–A–1, supra. The rationales thus take account only of the balance between the intrusion on an individual’s Fourth Amendment interest in his vehicle and the governmental interests in an expedient search of that vehicle; they do not account for the distinct privacy interest in one’s home or curtilage. To allow an officer to rely on the automobile exception to gain entry into a house or its curtilage for the purpose of conducting a vehicle search would unmoor the exception from its justifications, render hollow the core Fourth Amendment protection the Constitution extends to the house and its curtilage, and transform what was meant to be an exception into a tool with far broader application. Indeed, its name alone should make all this clear enough: It is, after all, an exception for automobiles.[3]Given the centrality of the Fourth Amendment interest in the home and its curtilage and the disconnect between that interest and the justifications behind the automobile exception, we decline Virginia’s invitation to extend the automobile exception to permit a warrantless intrusion on a home or its curtilage.IIIAVirginia argues that this Court’s precedent indicates that the automobile exception is a categorical one that permits the warrantless search of a vehicle anytime, anywhere, including in a home or curtilage. Specifically, Virginia points to two decisions that it contends resolve this case in its favor. Neither is dispositive or persuasive.First, Virginia invokes Scher v. United States,305 U. S. 251 (1938). In that case, federal officers received a confidential tip that a particular car would be transporting bootleg liquor at a specified time and place. The officers identified and followed the car until the driver “turned into a garage a few feet back of his residence and within the curtilage.” Id., at 253. As the driver exited his car, an officer approached and stated that he had been informed that the car was carrying contraband. The driver acknowledged that there was liquor in the trunk, and the officer proceeded to open the trunk, find the liquor, arrest the driver, and seize both the car and the liquor. Id., at 253–254. Although the officer did not have a search warrant, the Court upheld the officer’s actions as reasonable. Id., at 255.Scher is inapposite. Whereas Collins’ motorcycle was parked and unattended when Officer Rhodes intruded on the curtilage to search it, the officers in Scher first encountered the vehicle when it was being driven on public streets, approached the curtilage of the home only when the driver turned into the garage, and searched the vehicle only after the driver admitted that it contained contraband. Scher by no means established a general rule that the automobile exception permits officers to enter a home or its curtilage absent a warrant. The Court’s brief analysis referenced Carroll, but only in the context of observing that, consistent with that case, the “officers properly could have stopped” and searched the car “just before [petitioner] entered the garage,” a proposition the petitioner did “not seriously controvert.” Scher, 305 U. S., at 254–255. The Court then explained that the officers did not lose their ability to stop and search the car when it entered “the open garage closely followed by the observing officer” because “[n]o search was made of the garage.” Id., at 255. It emphasized that “[e]xamination of the automobile accompanied an arrest, without objection and upon admission of probable guilt,” and cited two search-incident-to-arrest cases. Ibid. (citing Agnello v. United States,269 U. S. 20, 30 (1925); Wisniewski v. United States, 47 F. 2d 825, 826 (CA6 1931)). Scher’s reasoning thus was both case specific and imprecise, sounding in multiple doctrines, particularly, and perhaps most appropriately, hot pursuit. The decision is best regarded as a factbound one, and it certainly does not control this case.Second, Virginia points to Labron,518 U. S. 938, where the Court upheld under the automobile exception the warrantless search of an individual’s pickup truck that was parked in the driveway of his father-in-law’s farmhouse. Id., at 939–940; Commonwealth v. Kilgore, 544 Pa. 439, 444, 677 A.2d 311, 313 (1995). But Labron provides scant support for Virginia’s position. Unlike in this case, there was no indication that the individual who owned the truck in Labron had any Fourth Amendment interest in the farmhouse or its driveway, nor was there a determination that the driveway was curtilage.BAlternatively, Virginia urges the Court to adopt a more limited rule regarding the intersection of the automobile exception and the protection afforded to curtilage. Virginia would prefer that the Court draw a bright line and hold that the automobile exception does not permit warrantless entry into “the physical threshold of a house or a similar fixed, enclosed structure inside the curtilage like a garage.” Brief for Respondent 46. Requiring officers to make “case-by-case curtilage determinations,” Virginia reasons, unnecessarily complicates matters and “raises the potential for confusion and . . . error.” Id., at 46–47 (internal quotation marks omitted).The Court, though, has long been clear that curtilage is afforded constitutional protection. See Oliver, 466 U. S., at 180. As a result, officers regularly assess whether an area is curtilage before executing a search. Virginia provides no reason to conclude that this practice has proved to be unadministrable, either generally or in this context. Moreover, creating a carveout to the general rule that curtilage receives Fourth Amendment protection, such that certain types of curtilage would receive Fourth Amendment protection only for some purposes but not for others, seems far more likely to create confusion than does uniform application of the Court’s doctrine.In addition, Virginia’s proposed rule rests on a mistaken premise about the constitutional significance of visibility. The ability to observe inside curtilage from a lawful vantage point is not the same as the right to enter curtilage without a warrant for the purpose of conducting a search to obtain information not otherwise accessible. Cf. Cir- aolo, 476 U. S., at 213–214 (holding that “physically non- intrusive” warrantless aerial observation of the curtilage of a home did not violate the Fourth Amendment, and could form the basis for probable cause to support a warrant to search the curtilage). So long as it is curtilage, a parking patio or carport into which an officer can see from the street is no less entitled to protection from trespass and a warrantless search than a fully enclosed garage.Finally, Virginia’s proposed bright-line rule automatically would grant constitutional rights to those persons with the financial means to afford residences with garages in which to store their vehicles but deprive those persons without such resources of any individualized consideration as to whether the areas in which they store their vehicles qualify as curtilage. See United States v. Ross,456 U. S. 798, 822 (1982) (“[T]he most frail cottage in the kingdom is absolutely entitled to the same guarantees of privacy as the most majestic mansion”).IVFor the foregoing reasons, we conclude that the automobile exception does not permit an officer without a warrant to enter a home or its curtilage in order to search a vehicle therein. We leave for resolution on remand whether Officer Rhodes’ warrantless intrusion on the curtilage of Collins’ house may have been reasonable on a different basis, such as the exigent circumstances exception to the warrant requirement. The judgment of the Supreme Court of Virginia is therefore reversed, and the case is remanded for further proceedings not inconsistent with this opinion.It is so ordered.Notes 1 Virginia does not dispute that Collins has Fourth Amendment standing. See Minnesota v. Olson,495 U. S. 91, 96–100 (1990). 2 Helpfully, the parties have simplified matters somewhat by each making a concession. Petitioner concedes “for purposes of this appeal” that Officer Rhodes had probable cause to believe that the motorcycle was the one that had eluded him, Brief for Petitioner 5, n. 3, and Virginia concedes that “Officer Rhodes searched the motorcycle,” Brief for Respondent 12. 3 The dissent concedes that “the degree of the intrusion on privacy” is relevant in determining whether a warrant is required to search a motor vehicle “located on private property.” Post, at 5–6 (opinion of Alito, J.). Yet it puzzlingly asserts that the “privacy interests at stake” here are no greater than when a motor vehicle is searched “on public streets.” Post, at 3–4. “An ordinary person of common sense,” post,at 2, however, clearly would understand that the privacy interests atstake in one’s private residential property are far greater than on a public street. Contrary to the dissent’s suggestion, it is of no significance that the motorcycle was parked just a “short walk up the driveway.” Ibid. The driveway was private, not public, property, and the motorcycle was parked in the portion of the driveway beyond where a neighbor would venture, in an area “intimately linked to the home, . . . where privacy expectations are most heightened.” California v. Ciraolo,476 U. S. 207, 213 (1986). Nor does it matter that Officer Rhodes“did not damage any property,” post, at 2, for an officer’s care in conducting a search does not change the character of the place being searched. And, as we explain, see infra, at 13–14, it is not dispositive that Officer Rhodes did not “observe anything along the way” to the motorcycle “that he could not have seen from the street,” post, at 2. Law enforcement officers need not “shield their eyes when passing by a home on public thoroughfares,” Ciraolo, 476 U. S., at 213, but the ability visually to observe an area protected by the Fourth Amendment does not give officers the green light physically to intrude on it. See Florida v. Jardines,569 U. S. 1, 7–8 (2013). It certainly does not permit an officer physically to intrude on curtilage, remove a tarp to reveal license plate and vehicle identification numbers, and use those numbers to confirm that the defendant committed a crime. The dissent also mistakenly relies on a law enacted by the First Congress and mentioned in Carroll v. United States,267 U. S. 132, 150–151 (1925), that authorized the warrantless search of vessels. Post, at 4–5, n. 3. The dissent thinks it implicit in that statute that “officers could cross private property such as wharves in order to reach and board those vessels.” Ibid. Even if it were so that a police officer could have entered a private wharf to search a vessel, that would not prove he could enter the curtilage of a home to do so. To the contrary, whereas the statute relied upon in Carroll authorized warrantless searches of vessels, it expressly required warrants to search houses. See 267 U. S., at 150–157; Act of July 31, 1789, §24,1Stat.43. Here, Officer Rhodes did not invade a private wharf to undertake a search; he invaded the curtilage of a home.

585.US.2017_16-1348

Petitioner Michael Currier was indicted for burglary, grand larceny, and unlawful possession of a firearm by a convicted felon. Because the prosecution could introduce evidence of Mr. Currier’s prior burglary and larceny convictions to prove the felon-in-possession charge, and worried that evidence might prejudice the jury’s consideration of the other charges, Mr. Currier and the government agreed to a severance and asked the court to try the burglary and larceny charges first, followed by a second trial on the felon-in-possession charge. At the first trial, Mr. Currier was acquitted. He then sought to stop the second trial, arguing that it would amount to double jeopardy. Alternatively, he asked the court to prohibit the state from relitigating at the second trial any issue resolved in his favor at the first. The trial court denied his requests and allowed the second trial to proceed unfettered. The jury convicted him on the felon-in-possession charge. The Virginia Court of Appeals rejected his double jeopardy arguments, and the Virginia Supreme Court summarily affirmed. Held: The judgment is affirmed. 292 Va. 737, 798 S. E. 2d 164, affirmed. Justice Gorsuch delivered the opinion of the Court with respect to Parts I and II, concluding that, because Mr. Currier consented to a severance, his trial and conviction on the felon-in-possession charge did not violate the Double Jeopardy Clause, which provides that no person may be tried more than once “for the same offence.” Mr. Currier argues that Ashe v. Swenson, 397 U. S. 436, requires a ruling for him. There, the Court held that the Double Jeopardy Clause barred a defendant’s prosecution for robbing a poker player because the defendant’s acquittal in a previous trial for robbing a different poker player from the same game established that the defendant “was not one of the robbers,” id., at 446. Ashe’s suggestion that the relitigation of an issue may amount to the impermissible relitigation of an offense represented a significant innovation in this Court’s jurisprudence. But whatever else may be said about Ashe, the Court has emphasized that its test is a demanding one. Ashe forbids a second trial only if to secure a conviction the prosecution must prevail on an issue the jury necessarily resolved in the defendant’s favor in the first trial. A second trial is not precluded simply because it is unlikely—or even very unlikely—that the original jury acquitted without finding the fact in question. To say that the second trial is tantamount to a trial of the same offense as the first and thus forbidden by the Double Jeopardy Clause, the Court must be able to say that it would have been irrational for the jury in the first trial to acquit without finding in the defendant’s favor on a fact essential to a conviction in the second. Bearing all that in mind, a critical difference emerges between this case and Ashe: Even assuming that Mr. Currier’s second trial qualified as the retrial of the same offense under Ashe, he consented to the second trial. In Jeffers v. United States, 432 U. S. 137, where the issue was a trial on a greater offense after acquittal on a lesser- included offense, the Court held that the Double Jeopardy Clause is not violated when the defendant “elects to have the . . . offenses tried separately and persuades the trial court to honor his election.” Id., at 152. If consent can overcome a traditional double jeopardy complaint about a second trial for a greater offense, it must also suffice to overcome a double jeopardy complaint under Ashe’s more innovative approach. Holding otherwise would be inconsistent not only with Jeffers but with other cases too. See, e.g., United States v. Dinitz, 424 U. S. 600. And cases Mr. Currier cites for support, e.g., Harris v. Washington, 404 U. S. 55, merely applied Ashe’s test and concluded that a second trial was impermissible. They do not address the question whether the Double Jeopardy Clause prevents a second trial when the defendant consents to it. Mr. Currier contends that he had no choice but to seek two trials, because evidence of his prior convictions would have tainted the jury’s consideration of the burglary and larceny charges. This is not a case, however, where the defendant had to give up one constitutional right to secure another. Instead, Mr. Currier faced a lawful choice between two courses of action that each bore potential costs and rationally attractive benefits. Difficult strategic choices are “not the same as no choice,” United States v. Martinez-Salazar, 528 U. S. 304, 315, and the Constitution “does not . . . forbid requiring” a litigant to make them, McGautha v. California, 402 U. S. 183, 213. Pp. 3–8. Justice Gorsuch, joined by The Chief Justice, Justice Thomas, and Justice Alito, concluded in Part III that civil issue preclusion principles cannot be imported into the criminal law through the Double Jeopardy Clause to prevent parties from retrying any issue or introducing any evidence about a previously tried issue. Mr. Currier argues that, even if he consented to a second trial, that consent did not extend to the relitigation of any issues the first jury resolved in his favor. Even assuming for argument’s sake that Mr. Currier’s consent to holding a second trial didn’t more broadly imply consent to the manner it was conducted, his argument must be rejected on a narrower ground as refuted by the text and history of the Double Jeopardy Clause and by this Court’s contemporary double jeopardy cases, e.g., Blockburger v. United States, 284 U. S. 299; Dowling v. United States, 493 U. S. 342. Nor is it even clear that civil preclusion principles would help defendants like Mr. Currier. See, e.g., Bravo-Fernandez v. United States, 580 U. S. ___, ___. Grafting civil preclusion principles onto the criminal law could also invite ironies—e.g., making severances more costly might make them less freely available. Pp. 8–16. Justice Kennedy concluded that, because Parts I and II of the Court’s opinion resolve this case in a full and proper way, the extent of the Double Jeopardy Clause protections discussed and defined in Ashe need not be reexamined here. Pp. 1–2. Gorsuch, J., announced the judgment of the Court and delivered the opinion of the Court with respect to Parts I and II, in which Roberts, C. J., and Kennedy, Thomas, and Alito, JJ., joined, and an opinion with respect to Part III, in which Roberts, C. J., and Thomas and Alito, JJ., joined. Kennedy, J., filed an opinion concurring in part. Ginsburg, J., filed a dissenting opinion, in which Breyer, Sotomayor, and Kagan, JJ., joined.

and Turner v. Arkansas, 407 U. S. 366 (1972) (per curiam). But these cases merely applied Ashe’s test and concluded that a second trial was impermissible. They did not address the question whether double jeopardy protections apply if the defendant consents to a second trial. Meanwhile, as we’ve seen, Jeffers, Dinitz, and Scott focus on that question directly and make clear that a defendant’s consent dispels any specter of double jeopardy abuse that holding two trials might otherwise present. This Court’s teachings are consistent and plain: the “Clause, which guards against Government oppression, does not relieve a defendant from the consequences of his voluntary choice.” Scott, supra, at 99. Mr. Currier replies that he had no real choice but to seek two trials. Without a second trial, he says, evidence of his prior convictions would have tainted the jury’s consideration of the burglary and larceny charges. And, he notes, Virginia law guarantees a severance in cases like his unless the defendant and prosecution agree to a single trial. But no one disputes that the Constitution permitted Virginia to try all three charges at once with appropriate cautionary instructions. So this simply isn’t a case where the defendant had to give up one constitutional right to secure another. Instead, Mr. Currier faced a lawful choice between two courses of action that each bore potential costs and rationally attractive benefits. It might have been a hard choice. But litigants every day face difficult decisions. Whether it’s the defendant who finds himself in the shoes of Jeffers, Dinitz, and Scott and forced to choose between allowing an imperfect trial to proceed or seeking a second that promises its own risks. Or whether it’s the defendant who must decide between exercising his right to testify in his own defense or keeping impeachment evidence of past bad acts from the jury. See, e.g., Brown v. United States, 356 U. S. 148, 154–157 (1958). This Court has held repeatedly that difficult strategic choices like these are “not the same as no choice,” United States v. Martinez-Salazar, 528 U. S. 304, 315 (2000), and the Constitution “does not . . . forbid requiring” a litigant to make them, McGautha v. California, 402 U. S. 183, 213 (1971). III Even if he voluntarily consented to holding the second trial, Mr. Currier argues, that consent did not extend to the relitigation of any issues the first jury resolved in his favor. So, Mr. Currier says, the court should have excluded evidence suggesting he possessed the guns in Mr. Gar- rison’s home, leaving the prosecution to prove that he possessed them only later, maybe down by the river. To support this argument, Mr. Currier points to issue preclusion principles in civil cases and invites us to import them for the first time into the criminal law through the Double Jeopardy Clause. In his view, the Clause should do much more than bar the retrial of the same offense (or crimes tantamount to the same offense under Ashe); it should be read now to prevent the parties from retrying any issue or introducing any evidence about a previously tried issue. While the dissent today agrees with us that the trial court committed no double jeopardy violation in holding the second trial, on this alternative argument it sides with Mr. Currier. See post, at 11, 12, 14–15. We cannot. Even assuming for argument’s sake that Mr. Currier’s consent to holding a second trial didn’t more broadly imply consent to the manner it was conducted, we must reject his argument on a narrower ground. Just last Term this Court warned that issue preclusion principles should have only “guarded application . . . in criminal cases.” Bravo-Fernandez v. United States, 580 U. S. ___, ___ (2016) (slip op. at 4). We think that caution remains sound. Mr. Currier’s problems begin with the text of the Double Jeopardy Clause. As we’ve seen, the Clause speaks not about prohibiting the relitigation of issues or evidence but offenses. Contrast this with the language of the Reexamination Clause. There, the Seventh Amendment says that “[i]n Suits at common law . . . no fact tried by a jury, shall be otherwise re-examined in any Court of the United States, than according to the rules of the common law.” (Emphasis added.) Words in one provision are, of course, often understood “by comparing them with other words and sentences in the same instrument.” 1 J. Story, Commentaries on the Constitution of the United States §400, p. 384 (1833). So it’s difficult to ignore that only in the Seventh Amendment—and only for civil suits—can we find anything resembling contemporary issue preclusion doctrine. What problems the text suggests, the original public understanding of the Fifth Amendment confirms. The Double Jeopardy Clause took its cue from English common law pleas that prevented courts from retrying a criminal defendant previously acquitted or convicted of the crime in question. See Scott, 437 U. S., at 87; 4 W. Blackstone, Commentaries on the Laws of England 329–330 (1769). But those pleas barred only repeated “prosecution for the same identical act and crime,” not the retrial of particular issues or evidence. Id., at 330 (emphasis added). As Sir Matthew Hale explained: “If A. commit a burglary . . . and likewise at the same time steal goods out of the house, if he be indicted of larciny for the goods and acquitted, yet he may be indicted for the burglary notwithstanding the acquittal. And è converso, if indicted for the burglary and acquitted, yet he may be indicted of the larciny, for they are several offenses, tho committed at the same time.” 2 M. Hale, The History of the Pleas of the Crown, ch. 31, pp. 245–246 (1736 ed.). Both English and early American cases illustrate the point. In Turner’s Case, 30 Kel. J. 30, 84 Eng. Rep. 1068 (K. B. 1663), for example, a jury acquitted the defendant of breaking into a home and stealing money from the owner. Even so, the court held that the defendant could be tried later for the theft of money “stolen at the same time” from the owner’s servant. Ibid. In Commonwealth v. Roby, 12 Pickering 496 (Mass. 1832), the court, invoking Blackstone, held that “[i]n considering the identity of the of- fence, it must appear by the plea, that the offence charged in both cases was the same in law and in fact.” Id., at 509. The court explained that a second prosecution isn’t precluded “if the offences charged in the two indictments be perfectly distinct in point of law, however nearly they may be connected in fact.” Ibid. (emphasis added). Another court even ruled “that a man acquitted for stealing the horse hath yet been arraigned and convict for stealing the saddle, tho both were done at the same time.” 2 Hale, supra, at 246. These authorities and many more like them demonstrate that early courts regularly confronted cases just like ours and expressly rejected the notion that the Double Jeopardy Clause barred the relitigation of issues or facts. See also Grady v. Corbin, 495 U. S. 508, 533–535 (1990) (Scalia, J., dissenting) (collecting authorities); 2 W. Hawkins, Pleas of the Crown, ch. 35, p. 371 (1726 ed.); 1 J. Chitty, Criminal Law 452–457 (1816); M. Friedland, Double Jeopardy 179, and n. 2 (1969). Any suggestion that our case presents a new phenomenon, then, risks overlooking this long history. See post, at 4–5 (Ginsburg, J., dissenting). This Court’s contemporary double jeopardy cases confirm what the text and history suggest. Under Blockburger v. United States, 284 U. S. 299 (1932), the courts apply today much the same double jeopardy test they did at the founding. Id., at 304. To prevent a second trial on a new charge, the defendant must show an identity of statutory elements between the two charges against him; it’s not enough that “a substantial overlap [exists] in the proof offered to establish the crimes.” Iannelli v. United States, 420 U. S. 770, 785, n. 17 (1975) (emphasis added). Of course, Ashe later pressed Blockburger’s boundaries by suggesting that, in narrow circumstances, the retrial of an issue can be considered tantamount to the retrial of an offense. See Yeager, 557 U. S., at 119. But, as we’ve seen, even there a court’s ultimate focus remains on the practi- cal identity of offenses, and the only available remedy is the traditional double jeopardy bar against the retrial of the same offense—not a bar against the relitigation of issues or evidence. See id., at 119–120. Even at the outer reaches of our double jeopardy jurisprudence, then, this Court has never sought to regulate the retrial of issues or evidence in the name of the Double Jeopardy Clause. Nor in acknowledging this do we plow any new ground. In Dowling v. United States, 493 U. S. 342 (1990), the defendant faced charges of bank robbery. At trial, the prosecution introduced evidence of the defendant’s involvement in an earlier crime, even though the jury in that case had acquitted. Like Mr. Currier, the defendant in Dowling argued that the trial court should have barred relitigation of an issue resolved in his favor in an earlier case and therefore excluded evidence of the acquitted offense. But the Court refused the request and in doing so expressly “decline[d] to extend Ashe . . . to exclude in all circumstances, as [the defendant] would have it, relevant and probative evidence that is otherwise admissible under the Rules of Evidence simply because it relates to alleged criminal conduct for which a defendant has been acquitted.” Id., at 348. If a second trial is permissible, the admission of evidence at that trial is governed by normal evidentiary rules—not by the terms of the Double Jeopardy Clause. “So far as merely evidentiary . . . facts are concerned,” the Double Jeopardy Clause “is inoperative.” Yates v. United States, 354 U. S. 298, 338 (1957). On its own terms, too, any effort to transplant civil preclusion principles into the Double Jeopardy Clause would quickly meet trouble. While the Clause embodies a kind of “claim preclusion” rule, even this rule bears little in common with its civil counterpart. In civil cases, a claim generally may not be tried if it arises out of the same transaction or common nucleus of operative facts as another already tried. Restatement (Second) of Judg- ments §19 (1982); Moschzisker, Res Judicata, 38 Yale L. J. 299, 325 (1929). But in a criminal case, Blockburger precludes a trial on an offense only if a court has previously heard the same offense as measured by its statutory elements. 284 U. S., at 304. And this Court has emphatic- ally refused to import into criminal double jeopardy law the civil law’s more generous “same transaction” or same criminal “episode” test. See Garrett v. United States, 471 U.S. 773, 790 (1985); see also Ashe, 397 U. S., at 448 (Harlan, J., concurring). It isn’t even clear that civil preclusion principles would help defendants like Mr. Currier. Issue preclusion addresses the effect in a current case of a prior adjudication in another case. So it doesn’t often have much to say about the preclusive effects of rulings “within the framework of a continuing action.” 18A C. Wright & A. Miller, Federal Practice and Procedure §4434 (2d ed. 2002); see also id., §4478. Usually, only the more flexible law of the case doctrine governs the preclusive effect of an earlier decision “within a single action.” Ibid. And that doctrine might counsel against affording conclusive effect to a prior jury verdict on a particular issue when the parties agreed to hold a second trial covering much the same terrain at a later stage of the proceedings. Besides, even if issue preclusion is the right doctrine for cases like ours, its application usually depends “on ‘an underlying confidence that the result achieved in the initial litigation was substantially correct.’ ” Bravo-Fernandez, 580 U. S., at ___ (slip op., at 4) (quoting Standefer v. United States, 447 U. S. 10, 23, n. 18 (1980)). As a result, the doctrine does not often bar the relitigation of issues when “[t]he party against whom preclusion is sought could not, as a matter of law, have obtained review of the judgment in the initial action.” Restatement (Second) of Judgments §28. In criminal cases, of course, the government cannot obtain appellate review of acquittals. So a faithful application of civil preclusion principles in our case and others like it might actually militate against finding preclusion. See Bravo-Fernandez, supra, at ___ (slip op., at 4); Standefer, supra, at 22–23, and n. 18. Neither Mr. Currier nor the dissent offers a persuasive reply to these points. They cannot dispute that the text of the Double Jeopardy Clause, which bars a prosecution for the same offense, is inconsistent with an issue preclusion rule that purports to bar a “second prosecution involv[ing] . . . a different ‘offense.’ ” Post, at 4. They decline to “engage” with the Clause’s history, though the dissent appears to agree that the Clause was not originally understood to include an issue preclusion rule. See post, at 3–4, 13. Neither Mr. Currier nor the dissent seeks to show that, even taken on their own terms, civil issue preclusion principles would apply to cases like this one. Without text, history, or logic to stand on, the dissent leans heavily on a comparison to Dowling. In Dowling, the dissent emphasizes, the two trials involved different criminal episodes while the two trials here addressed the same set of facts. But Dowling did not rest its holding on this feature and the dissent does not explain its relevance. If issue preclusion really did exist in criminal law, why wouldn’t it preclude the retrial of any previously tried issue, regardless whether that issue stems from the same or a different “criminal episode”? In the end, Mr. Currier and the dissent must emphasize various policy reasons for adopting a new rule of issue preclusion into the criminal law. See post, at 4–5, 8–9. They contend that issue preclusion is “needed” to combat the “prosecutorial excesses” that could result from the proliferation of criminal offenses, post, at 4–5, though we aren’t sure what to make of this given the dissent’s later claim that “issue preclusion requires no showing of prosecutorial overreaching,” post, at 8. In any event, there are risks with the approach Mr. Currier and the dissent pro- pose. Consider, for example, the ironies that grafting civil preclusion principles onto the criminal law could invite. Issue preclusion is sometimes applied offensively against civil defendants who lost on an issue in an earlier case. Parklane Hosiery Co. v. Shore, 439 U. S. 322, 331–332 (1979). By parallel logic, could we expect the government to invoke the doctrine to bar criminal defendants from relitigating issues decided against them in a prior trial? It’s an outcome few defendants would welcome but one some have already promoted. See, e.g., Kennelly, Precluding the Accused: Offensive Collateral Estoppel in Criminal Cases, 80 Va. L. Rev. 1379, 1380–1381, 1416, 1426–1427 (1994); Vestal, Issue Preclusion and Criminal Prosecutions, 65 Iowa L. Rev. 281, 297, 320–321 (1980). Maybe worse yet, consider the possible effect on severances. Today, some state courts grant severance motions liberally to benefit defendants. But what would happen if this Court unilaterally increased the costs associated with severance in the form of allowing issue preclusion for defendants only? Granting a severance is no small thing. It means a court must expend resources for two trials where the Constitution would have permitted one. Witnesses and victims must endure a more protracted ordeal. States sometimes accept these costs to protect a defendant from potential prejudice. But 20 States appearing before us have warned that some jurisdictions might respond to any decision increasing the costs of severed trials by making them less freely available. See Brief for Indiana et al. as Amici Curiae 4, 16–20. Of course, that’s only a prediction. But it’s a hard if unwanted fact that “[t]oday’s elaborate body of procedural rules” can contribute to making “trials expensive [and] rare.” W. Stuntz, The Collapse of American Criminal Justice 39 (2011). And it would be a mistake to ignore the possibility that by making severances more costly we might wind up making them rarer too. The fact is, civil preclusion principles and double jeop- ardy are different doctrines, with different histories, serving different purposes. Historically, both claim and issue preclusion have sought to “promot[e] judicial economy by preventing needless litigation.” Parklane Hosiery, supra, at 326. That interest may make special sense in civil cases where often only money is at stake. But the Double Jeopardy Clause and the common law principles it built upon govern criminal cases and concern more than efficiency. They aim instead, as we’ve seen, to balance vital interests against abusive prosecutorial practices with consideration to the public’s safety. The Clause’s terms and history simply do not contain the rights Mr. Currier seeks. Nor are we at liberty to rewrite those terms or that history. While the growing number of criminal offenses in our statute books may be cause for concern, see post, at 4–5 (Ginsburg, J., dissenting), no one should expect (or want) judges to revise the Constitution to address every social problem they happen to perceive. The proper authorities, the States and Congress, are empowered to adopt new laws or rules experimenting with issue or claim preclusion in criminal cases if they wish. In fact, some States have already done so. On these matters, the Constitution dictates no answers but entrusts them to a self-governing people to resolve. * The judgment of the Virginia Supreme Court is Affirmed.

583.US.2017_15-1439

In the wake of the 1929 stock market crash, Congress enacted two laws, in successive years, to promote honest practices in the securities markets. The Securities Act of 1933 (1933 Act) creates private rights of action to aid the enforcement of obligations pertaining to securities offerings. The Act authorizes both federal and state courts to exercise jurisdiction over those private suits and, more unusually, bars the removal of such suits from state to federal court. The Securities Exchange Act of 1934 (1934 Act), which regulates not the original issuance of securities but all their subsequent trading, is also enforceable through private rights of action. But all suits brought under the 1934 Act fall within the exclusive jurisdiction of the federal courts. In 1995, the Private Securities Litigation Reform Act (Reform Act) amended both Acts, in order to stem perceived abuses of the class-action vehicle in securities litigation. The Reform Act included both substantive reforms, applicable in state and federal court alike, and procedural reforms, applicable only in federal court. Rather than face these new obstacles, plaintiffs began filing securities class actions under state law. To prevent this end run around the Reform Act, Congress passed the Securities Litigation Uniform Standards Act of 1998 (SLUSA), whose amendments to the 1933 Act are at issue in this case. As relevant here, those amendments include two operative provisions, two associated definitions, and two “conforming amendments.” First, 15 U. S. C. §77p(b) completely disallows (in both state and federal courts) “covered class actions” alleging dishonest practices “in connection with the purchase or sale of a covered security.” According to SLUSA’s definitions, the term “covered class action” means a class action in which “damages are sought on behalf of more than 50 persons.” §77p(f)(2). And the term “covered security” refers to a security listed on a national stock exchange. §77p(f)(3). Next, §77p(c) provides for the removal of certain class actions to federal court, where they are subject to dismissal. Finally, SLUSA’s “conforming amendments” add two new phrases to §77v(a), the 1933 Act’s jurisdictional provision. The first creates an exception to §77v(a)’s general removal bar through the language “[e]xcept as provided in section 77p(c).” The other—the key provision in this case—expresses a caveat to the general rule that state and federal courts have concurrent jurisdiction over all claims to enforce the 1933 Act. With this conforming amendment, §77v(a) now provides that state and federal courts shall have concurrent jurisdiction, “except as provided in section 77p . . . with respect to covered class actions.” The Court refers to this provision as the “except clause.” Respondents, three pension funds and an individual (Investors), purchased shares of stock in petitioner Cyan, Inc., in an initial public offering. After the stock declined in value, the Investors brought a damages class action against Cyan in state court, alleging 1933 Act violations. They did not assert any claims based on state law. Cyan moved to dismiss for lack of subject matter jurisdiction, arguing that SLUSA’s “except clause” stripped state courts of power to adjudicate 1933 Act claims in “covered class actions.” The Investors maintained that SLUSA left intact state courts’ jurisdiction over all suits—including “covered class actions”—alleging only 1933 Act claims. The state courts agreed with the Investors and denied Cyan’s motion to dismiss. This Court granted certiorari to decide whether SLUSA deprived state courts of jurisdiction over “covered class actions” asserting only 1933 Act claims. The Court also addresses a related question raised by the federal Government as amicus curiae and addressed by the parties in briefing and argument: whether SLUSA enabled defendants to remove 1933 Act class actions from state to federal court for adjudication. Held: 1. SLUSA did nothing to strip state courts of their longstanding jurisdiction to adjudicate class actions brought under the 1933 Act. Pp. 7–18. (a) SLUSA’s text, read most straightforwardly, leaves this jurisdiction intact. The background rule of §77v(a)—in place since the 1933 Act’s passage—gives state courts concurrent jurisdiction over all suits “brought to enforce any liability or duty created by” that statute. And the except clause—“except as provided in section 77p of this title with respect to covered class actions”—ensures that in any case in which §77v(a) and §77p conflict, §77p will control. The critical question for this case is therefore whether §77p limits state-court jurisdiction over class actions brought under the 1933 Act. It does not. Section 77p bars certain securities class actions based on state law but it says nothing, and so does nothing, to deprive state courts of jurisdiction over class actions based on federal law. That means §77v(a)’s background rule—under which a state court may hear the Investors’ 1933 Act suit—continues to govern. Cyan argues that the except clause’s reference to “covered class actions” points the reader to §77p(f)(2), which defines that term to mean a suit seeking damages on behalf of more than fifty persons—without mentioning anything about whether the suit is based on state or federal law. But that view cannot be squared with the except clause’s wording for two independent reasons. First, the except clause points to “section 77p” as a whole—not to paragraph 77p(f)(2). Had Congress intended to refer to §77p(f)(2)’s definition alone, it presumably would have done so. See NLRB v. SW General, Inc., 580 U. S. ___, ___. Second, a definition, like §77p(f)(2), does not “provide[]” an “except[ion],” but instead gives meaning to a term—and Congress well knows the difference between those two functions. Not one of the 30-plus provisions in the 1933 and 1934 Acts using the phrase “except as provided in . . .” cross-references a definition. Structure and context also support the Court’s reading of the except clause. Because Cyan treats the broad definition of “covered class action” as altering §77v(a)’s jurisdictional grant, its construction would prevent state courts from deciding any 1933 Act class suits seeking damages for more than fifty plaintiffs, thus stripping state courts of jurisdiction over suits about securities raising no particular national interest. That result is out of line with SLUSA’s overall scope. Moreover, it is highly unlikely that Congress upended the 65-year practice of state courts’ adjudicating all manner of 1933 Act cases (including class actions) by way of a mere conforming amendment. See Director of Revenue of Mo. v. CoBank ACB, 531 U. S. 316 . Pp. 8–12. (b) Cyan’s reliance on legislative purpose and history is unavailing. Pp. 12–18. (1) Cyan insists that the only way for SLUSA to serve the Reform Act’s objectives was by divesting state courts of jurisdiction over all sizable 1933 Act class actions. Specifically, it claims that its reading is necessary to prevent plaintiffs from circumventing the Reform Act’s procedural measures, which apply only in federal court, by bringing 1933 Act class actions in state court. But Cyan ignores a different way in which SLUSA served the Reform Act’s objectives—which the Court’s view of the statute fully effects. The Reform Act included substantive sections protecting defendants in suits brought under the federal securities laws. Plaintiffs circumvented those provisions by bringing their complaints of securities misconduct under state law instead. Hence emerged SLUSA’s bar on state-law class actions (and its removal provision to ensure their dismissal)—which guaranteed that the Reform Act’s heightened substantive standards would govern all future securities class litigation. SLUSA’s preamble states that the statute is designed “to limit the conduct of securities class actions under state law, and for other purposes,” 112Stat. 3227, and this Court has underscored, over and over, SLUSA’s “purpose to preclude certain vexing state-law class actions.” Kircher v. Putnam Funds Trust, 547 U. S. 633 . That object—which SLUSA’s text actually reflects—does not depend on stripping state courts of jurisdiction over 1933 Act class suits, as Cyan proposes. For wherever those suits go forward, the Reform Act’s substantive protections necessarily apply. SLUSA also went a good distance toward ensuring that federal courts would play the principal role in adjudicating securities class actions by means of its revisions to the 1934 Act. Because federal courts have exclusive jurisdiction over 1934 Act claims, forcing plaintiffs to bring class actions under the 1934 statute instead of state law also forced them to file in federal court. Pp. 12–15. (2) Cyan finally argues that the except clause would serve no purpose at all unless it works as Cyan says. But Congress could have envisioned the except clause as the ultimate fail-safe device, designed to safeguard §77p’s class-action bar come whatever might. Congress has been known to legislate in that hyper-vigilant way, to “remov[e] any doubt” as to things not particularly doubtful in the first instance. Marx v. General Revenue Corp., 568 U. S. 371 –384. If ever it had reason to legislate in that fashion, it was in SLUSA—whose very impetus lay in the success of class action attorneys in “bypass[ing] . . . the Reform Act.” Kircher, 547 U. S., at 636. And regardless of any uncertainty surrounding Congress’s reasons for drafting the except clause, there is no sound basis for giving that clause a broader reading than its language can bear, especially in light of the dramatic change such an interpretation would work in the 1933 Act’s jurisdictional framework. Pp. 15–18. 2. SLUSA does not permit defendants to remove class actions alleging only 1933 Act claims from state to federal court. The Government argues that §77p(c) allows defendants to remove 1933 Act class actions to federal court as long as they allege the kinds of misconduct listed in §77p(b). But most naturally read, §77p(c) refutes, not supports, the Government’s view. Section 77p(c) allows for removal of “[a]ny covered class action brought in any State court involving a covered security, as set forth in subsection (b).” The covered class actions “set forth” in §77p(b) are state-law class actions alleging securities misconduct. Federal-law suits are not “class action[s] . . . as set forth in subsection (b).” Thus, they remain subject to the 1933 Act’s removal ban. This Court has held as much, concluding that §§77p(b) and 77p(c) apply to the exact same universe of class actions. Kircher, 547 U. S., at 643–644. The “straightforward reading” of those two provisions is that removal under §77p(c) is “limited to those [actions] precluded by the terms of subsection (b).” Id., at 643. Pp. 18–24. Affirmed. Kagan, J., delivered the opinion for a unanimous Court.

This case presents two questions about the Securities Litigation Uniform Standards Act of 1998 (SLUSA), 112Stat. 3227. First, did SLUSA strip state courts of jurisdiction over class actions alleging violations of only the Securities Act of 1933 (1933 Act), 48Stat. 74, as amended, 15 U. S. C. §77a et seq.? And second, even if not, did SLUSA empower defendants to remove such actions from state to federal court? We answer both questions no. I A In the wake of the 1929 stock market crash, Congress enacted two laws, in successive years, to promote honest practices in the securities markets. The 1933 Act required companies offering securities to the public to make “full and fair disclosure” of relevant information. Pinter v. Dahl, 486 U. S. 622, 646 (1988) . And to aid enforcement of those obligations, the statute created private rights of action. Congress authorized both federal and state courts to exercise jurisdiction over those private suits. See §22(a), 48Stat. 86 (“The district courts of the United States . . . shall have jurisdiction[,] concurrent with State and Territorial courts, of all suits in equity and actions at law brought to enforce any liability or duty created by this title”). More unusually, Congress also barred the removal of such actions from state to federal court. Id., at 87 (“No case arising under this title and brought in any State court of competent jurisdiction shall be removed to any court of the United States”). So if a plaintiff chose to bring a 1933 Act suit in state court, the defendant could not change the forum. Congress’s next foray, the Securities Exchange Act of 1934 (1934 Act), operated differently. See 48Stat. 881, as amended, 15 U. S. C. §78a et seq. That statute regulated not the original issuance of securities but instead all their subsequent trading, most commonly on national stock exchanges. See Blue Chip Stamps v. Manor Drug Stores, 421 U. S. 723, 752 (1975) . The 1934 Act, this Court held, could also be enforced through private rights of action. See id., at 730, and n. 4. But Congress determined that all those suits should fall within the “exclusive jurisdiction” of the federal courts. §27, 48Stat. 902–903. So a plaintiff could never go to state court to litigate a 1934 Act claim. In 1995, the Private Securities Litigation Reform Act (Reform Act), 109Stat. 737, amended both the 1933 and the 1934 statutes in mostly identical ways. Congress passed the Reform Act principally to stem “perceived abuses of the class-action vehicle in litigation involving nationally traded securities.” Merrill Lynch, Pierce, Fenner & Smith Inc. v. Dabit, 547 U. S. 71, 81 (2006) . Some of the Reform Act’s provisions made substantive changes to the 1933 and 1934 laws, and applied even when a 1933 Act suit was brought in state court. For instance, the statute created a “safe harbor” from federal liability for certain “forward-looking statements” made by company officials. 15 U. S. C. §77z–2 (1933 Act); §78u–5 (1934 Act). Other Reform Act provisions modified the procedures used in litigating securities actions, and applied only when such a suit was brought in federal court. To take one example, the statute required a lead plaintiff in any class action brought under the Federal Rules of Civil Procedure to file a sworn certification stating, among other things, that he had not purchased the relevant security “at the direction of plaintiff’s counsel.” §77z–1(a)(2)(A)(ii) (1933 Act); §78u–4(a)(2)(A)(ii) (1934 Act). But the Reform Act fell prey to the law of “unintended consequence[s].” Dabit, 547 U. S., at 82. As this Court previously described the problem: “Rather than face the obstacles set in their path by the Reform Act, plaintiffs and their representatives began bringing class actions under state law.” Ibid. That “phenomenon was a novel one”—and an unwelcome one as well. Ibid. To prevent plaintiffs from circumventing the Reform Act, Congress again undertook to modify both securities laws. The result was SLUSA, whose amendments to the 1933 Act are at issue in this case. Those amendments include, as relevant here, two operative provisions, two associated definitions, and two “conforming amendments” to the 1933 law’s jurisdictional section. 112Stat. 3230. (SLUSA’s amendments to the 1934 Act include essentially the same operative provisions and definitions. See Dabit, 547 U. S., at 82, n. 6. But Congress decided that the 1934 law’s exclusive jurisdiction provision needed no conforming amendments.) The added material—now found in §§77p and 77v(a) and set out in full in this opinion’s appendix—goes as follows. First, §77p(b) altogether prohibits certain securities class actions based on state law. That provision—which we sometimes (and somewhat prosaically) refer to as the state-law class-action bar—reads: “No covered class action based upon the statutory or common law of any State . . . may be maintained in any State or Federal court by any private party alleging— “(1) an untrue statement or omission of a material fact in connection with the purchase or sale of a covered security; or “(2) that the defendant used or employed any manipulative or deceptive device or contrivance in connection with the purchase or sale of a covered security.” According to SLUSA’s definitions, the term “covered class action” means a class action in which “damages are sought on behalf of more than 50 persons.” §77p(f)(2). And the term “covered security” refers to a security listed on a national stock exchange. §77p(f)(3) (cross-referencing §77r(b)). So taken all in all, §77p(b) completely disallows (in both state and federal courts) sizable class actions that are founded on state law and allege dishonest practices respecting a nationally traded security’s purchase or sale. Next, §77p(c) provides for the removal of certain class actions to federal court, as well as for their subsequent disposition: “Any covered class action brought in any State court involving a covered security, as set forth in subsection (b) of this section, shall be removable to the Federal district court for the district in which the action is pending, and shall be subject to subsection (b) of this section.” The first chunk of that provision identifies the removable cases, partly by way of a cross-reference (“as set forth in subsection (b)”) to the just-described class-action bar. The final clause of the provision (“and shall be subject to subsection (b)”) indicates what should happen to a barred class suit after it has been removed: The “proper course is to dismiss” the action. Kircher v. Putnam Funds Trust, 547 U. S. 633, 644 (2006) . As this Court has explained, §77p(c) “avails a defendant of a federal forum in contemplation not of further litigation over the merits of a claim brought in state court, but of termination of the proceedings altogether.” Id., at 645, n. 12. The point of providing that option, everyone here agrees, was to ensure the dismissal of a prohibited state-law class action even when a state court “would not adequately enforce” §77p(b)’s bar. Brief for United States as Amicus Curiae 3; see Brief for Petitioners 7; Brief for Respondents 20. Finally, the 1933 Act’s jurisdictional provision, codified at §77v(a), now includes two new phrases framed as exemptions—SLUSA’s self-described “conforming amendments.” 112Stat. 3230; see supra, at 3. The less significant of the pair, for our purposes, reflects the allowance for removing certain class actions described above. Against the backdrop of the 1933 Act’s general removal bar, see supra, at 2, that added (italicized) material reads: “Except as provided in section 77p(c) of this title, no case arising under this subchapter and brought in any State court of competent jurisdiction shall be removed to any court of the United States.” The more important of the conforming amendments in this case expresses a caveat to the general rule, see supra, at 1–2, that state and federal courts have concurrent jurisdiction over all claims to enforce the 1933 Act. As amended (again, with the new material in italics), the relevant sentence now reads: “The district courts of the United States . . . shall have jurisdiction[,] concurrent with State and Territorial courts, except as provided in section 77p of this title with respect to covered class actions, of all suits in equity and actions at law brought to enforce any liability or duty created by this subchapter.” Throughout this opinion, we refer to the italicized words just above as the “except clause.” Its meaning is at the heart of the parties’ dispute in this Court. B The petitioners in this case are Cyan, a telecommunications company, and its officers and directors (together, Cyan). The respondents are three pension funds and an individual (together, Investors) who purchased shares of Cyan stock in an initial public offering. After the stock declined in value, the Investors brought a damages class action against Cyan in California Superior Court. Their complaint alleges that Cyan’s offering documents contained material misstatements, in violation of the 1933 Act. It does not assert any claims based on state law. Cyan moved to dismiss the Investors’ suit for lack of subject matter jurisdiction. It argued that what we have termed SLUSA’s “except clause”—i.e., the amendment made to §77v(a)’s concurrent-jurisdiction grant—stripped state courts of power to adjudicate 1933 Act claims in “covered class actions.” The Investors did not dispute that their suit qualifies as such an action under SLUSA’s definition, see §77p(f)(2). But they maintained that SLUSA left intact state courts’ jurisdiction over all suits—including “covered class actions”—alleging only 1933 Act claims. The California Superior Court agreed with the Investors and denied Cyan’s motion to dismiss. See App. to Pet. for Cert. 6a. The state appellate courts then denied review of that ruling. See id., at 15a–16a. We granted Cyan’s petition for certiorari, 581 U. S. ___ (2017), to resolve a split among state and federal courts about whether SLUSA deprived state courts of jurisdiction over “covered class actions” asserting only 1933 Act claims.[1] In opposing Cyan’s jurisdictional position here, the Federal Government as amicus curiae raised another question: whether SLUSA enabled defendants to remove 1933 Act class actions from state to federal court for adjudication. See Brief for United States as Amicus Curiae 23–31. That question is not directly presented because Cyan never attempted to remove the Investors’ suit. But the removal issue is related to the parties’ jurisdictional arguments, and both Cyan and the Investors addressed it in briefing and argument. See Brief for Petitioners 39–40; Brief for Respondents 31–35; Tr. of Arg. 31, 53–56, 74–76, 80. Accordingly, we consider as well the scope of §77p(c)’s removal authorization. II By its terms, §77v(a)’s “except clause” does nothing to deprive state courts of their jurisdiction to decide class actions brought under the 1933 Act. And Cyan’s various appeals to SLUSA’s purposes and legislative history fail to overcome the clear statutory language. The statute says what it says—or perhaps better put here, does not say what it does not say. State-court jurisdiction over 1933 Act claims thus continues undisturbed.[2] A SLUSA’s text, read most straightforwardly, leaves in place state courts’ jurisdiction over 1933 Act claims, including when brought in class actions. Recall that the background rule of §77v(a)—in place since the 1933 Act’s passage—gives state courts concurrent jurisdiction over all suits “brought to enforce any liability or duty created by” that statute. See supra, at 1–2. The except clause—once again, “except as provided in section 77p of this title with respect to covered class actions”—is drafted as a limitation on that rule: It ensures that in any case in which §77v(a) and §77p come into conflict, §77p will control. The critical question for this case is therefore whether §77p limits state-court jurisdiction over class actions brought under the 1933 Act. It does not. As earlier described, §77p bars certain securities class actions based on state law. See §77p(b); supra, at 3–4. And as a corollary of that prohibition, it authorizes removal of those suits so that a federal court can dismiss them. See §77p(c); supra, at 4–5. But the section says nothing, and so does nothing, to deprive state courts of jurisdiction over class actions based on federal law. That means the background rule of §77v(a)—under which a state court may hear the Investors’ 1933 Act suit—continues to govern. Cyan offers an alternative reading, in which one of SLUSA’s definitional provisions works to alter state-court jurisdiction. According to Cyan, the except clause’s reference to “covered class actions” points the reader to, and only to, §77p(f)(2)’s definition of that term. See Brief for Petitioners 16. And that definition states that a “covered class action” is a suit seeking damages on behalf of more than 50 persons—without mentioning anything about whether the suit is based on state or federal law. Cyan thus concludes that the except clause exempts all sizable class actions—including the Investors’ suit—from §77v(a)’s conferral of jurisdiction on state courts. But that view cannot be squared with the except clause’s wording for two independent reasons. To start with, the except clause points to “section 77p” as a whole—not to paragraph 77p(f)(2). Cyan wants to cherry pick from the material covered by the statutory cross-reference. But if Congress had intended to refer to the definition in §77p(f)(2) alone, it presumably would have done so—just by adding a letter, a number, and a few parentheticals. As this Court recently explained, “Congress often drafts statutes with hierarchical schemes—section, subsection, paragraph, and on down the line.” NLRB v. SW General, Inc., 580 U. S. ___, ___ (2017) (slip op., at 9). And “[w]hen Congress want[s] to refer only to a particular subsection or paragraph, it sa[ys] so.” Ibid. It said no such thing in the except clause. In any event, the definitional paragraph on which Cyan relies cannot be read to “provide[]” an “except[ion]” to the rule of concurrent jurisdiction, in the way SLUSA’s except clause requires. A definition does not provide an exception, but instead gives meaning to a term—and Congress well knows the difference between those two functions. Thousands of statutory provisions use the phrase “except as provided in . . .” followed by a cross-reference in order to indicate that one rule should prevail over another in any circumstance in which the two conflict; we count more than 30 such constructions in the 1933 and 1934 Acts alone.[3] Not one of those 30-plus provisions cross-references a definition; nor has Cyan pointed to a single such example from the whole rest of the U. S. Code. And the Congress enacting SLUSA had no reason to attempt that peculiar maneuver for the first time. If Congress had wanted to deprive state courts of jurisdiction over 1933 Act class actions, it had an easy way to do so: just insert into §77p an exclusive federal jurisdiction provision (like the 1934 Act’s) for such suits. That rule, when combined with the except clause, would have done the trick because it would have “provided” an “except[ion]” to §77v(a)’s grant of concurrent jurisdiction; by contrast, a mere definition of “covered class action” (as a damages suit on behalf of 50-plus people) does not so provide. SLUSA’s other conforming amendment illustrates the two ways in which Cyan’s construction of the except clause departs from its language. Recall that §77v(a) includes a general bar on removal. See supra, at 2. And recall that SLUSA appended to that prohibition the phrase “[e]xcept as provided in section 77p(c)” to reflect the statute’s new permission to remove certain class actions. See supra, at 5. In that “except as provided” phrase—just four sentences down from the except clause central to this case—Congress pinpointed a subsection of §77p, rather than citing the entire section for only one of its parts. Still more, that cross-referenced subsection contains an operative provision that could limit a rule, rather than a mere definition of a statutory term. In short, Congress wrote the removal bar’s except clause in just the way a reader of legislation would expect—and not in the wholly irregular way Cyan proposes for the except clause at issue here. Especially given the two provisions’ “interrelationship and close proximity,” Commissioner v. Lundy, 516 U. S. 235, 250 (1996) , the one conforming amendment highlights how far Cyan seeks to stretch the text of the other. Cyan’s interpretation also fits poorly with the remainder of the statutory scheme. Because Cyan treats the broad definition of “covered class action” as altering §77v(a)’s jurisdictional grant, its construction would prevent state courts from deciding any 1933 Act class suits seeking damages for more than 50 plaintiffs. That would include suits not involving a “covered security”—i.e., a security traded on a national stock exchange. §77p(f)(3); Brief for Petitioners 29 (conceding that point). But this Court has emphasized that SLUSA’s operative provisions (including its state-law class-action bar, see §77p(b)) apply to only “transactions in covered securities”: The statute “ex- presses no concern” with “transactions in uncovered securi- ties”—precisely because they are not traded on national markets. Chadbourne & Parke LLP v. Troice, 571 U. S. 377 , ___ (2014) (slip. op., at 9); see Brief for United States as Amicus Curiae 16–17 (SLUSA does not regard suits involving uncovered securities as “a matter of distinct federal concern”). Those securities, the Court explained, are “primarily of state concern,” and SLUSA “maintains state legal authority” to address them. Chadbourne, 571 U. S., at ___ (slip op., at 13). Except that under Cyan’s view, SLUSA would not. Instead, the law would strip state courts of jurisdiction over suits about securities raising no particular national interest. That result is out of line with SLUSA’s overall scope. And finally, Cyan’s take on the except clause reads too much into a mere “conforming amendment.” 112Stat. 3230. The change Cyan claims that clause made to state-court jurisdiction is the very opposite of a minor tweak. When Congress passed SLUSA, state courts had for 65 years adjudicated all manner of 1933 Act cases, including class actions. Indeed, defendants could not even remove those cases to federal court, as schemes of concurrent jurisdiction almost always allow. See supra, at 2. State courts thus had as much or more power over the 1933 Act’s enforcement as over any federal statute’s. To think Cyan right, we would have to believe that Congress upended that entrenched practice not by any direct means, but instead by way of a conforming amendment to §77v(a) (linked, in its view, with only a definition). But Congress does not make “radical—but entirely implicit—change[s]” through “technical and conforming amendments.” Director of Revenue of Mo. v. CoBank ACB, 531 U. S. 316, 324 (2001) (internal quotation marks omitted). Or to use the more general (and snappier) formulation of that rule, relevant to all “ancillary provisions,” Congress does not “hide elephants in mouseholes.” Whitman v. American Trucking Assns., Inc., 531 U. S. 457, 468 (2001) . That is yet one more reason to reject Cyan’s view of SLUSA’s text. B Faced with such recalcitrant statutory language, Cyan stakes much of its case on legislative purpose and history. See Brief for Petitioners 20–33, 36–37; Reply Brief 7–11, 17–21. Its claims come in two forms—one relating to the goals of SLUSA as a whole and the other relating to the aims of the except clause. Even assuming clear text can ever give way to purpose, Cyan would need some monster arguments on this score to create doubts about SLUSA’s meaning. The points Cyan raises come nowhere close to that level. 1 According to Cyan’s broad purposive argument, Congress could not “make good on the promise of the Reform Act”—which was its principal intention in enacting SLUSA—without divesting state courts of jurisdiction over all sizable 1933 Act class actions. Brief for Petitioners 20. Remember that the Reform Act contained a number of procedural measures (for example, a sworn-certification requirement for lead plaintiffs, see §77z–1(a)(2)(A)) that apply only in federal court. See supra, at 2–3. Plaintiffs bringing 1933 Act class actions could avoid those provisions simply by filing in state court; after all, those suits were not even removable by defendants. “So,” Cyan claims, “Congress enacted SLUSA to finish the job”—by shutting down the state forum and shifting all 1933 Act class actions to the federal one. Brief for Petitioners 21. In support of that view, Cyan cites several statements in SLUSA’s legislative reports—in particular, that SLUSA’s purpose was “to prevent plaintiffs from seeking to evade the protections that Federal law provides against abusive litigation by filing suit in State, rather than in Federal, court.” H. R. Conf. Rep. No. 105–803, p. 13 (1998); see H. R. Rep. No. 105–640, pp. 8–9 (1998); S. Rep. No. 105–182, p. 3 (1998). But to begin with, Cyan ignores a different way in which SLUSA “serve[d] the [Reform Act’s] objectives,” Brief for Petitioners 11—which our view of the statute fully effects. Recall that the Reform Act also included substantive sections protecting defendants (like a safe harbor for forward-looking statements) in suits brought under the federal securities laws. See §77z–2; supra, at 2. Plaintiffs could—and did—avoid those provisions by bringing their complaints of securities misconduct under state law instead. See supra, at 3. Hence emerged SLUSA’s bar on state-law class actions (and its removal provision to ensure their dismissal)—which guaranteed that the Reform Act’s heightened substantive standards would govern all future securities class litigation. SLUSA itself highlights that aim: Its preamble states that the statute is designed “to limit the conduct of securities class actions under State law, and for other purposes.” 112Stat. 3227. So too, this Court has underscored, over and over, SLUSA’s “purpose to preclude certain vexing state-law class actions.” Kircher, 547 U. S., at 645, n. 12; see Dabit, 547 U. S., at 82 (SLUSA stopped plaintiffs from “bringing class actions under state law”); Amgen Inc. v. Connecticut Retirement Plans and Trust Funds, 568 U. S. 455, 476 (2013) (SLUSA “curtailed plaintiffs’ ability to evade the [Reform Act] by bringing class-action suits under state rather than federal law”). That object—which SLUSA’s text actually reflects—does not depend on stripping state courts of jurisdiction over 1933 Act class suits, as Cyan proposes. For wherever those suits go forward, the Reform Act’s substantive protections necessarily apply. Still more, SLUSA ensured that federal courts would play the principal role in adjudicating securities class actions by means of its revisions to the 1934 Act. As explained earlier, SLUSA amended that statute in the same main way it did the 1933 Act—by adding a state-law class-action bar. See §78bb(f )(1); supra, at 3. But there, the change had a double effect: Because federal courts have exclusive jurisdiction over 1934 Act claims, forcing plaintiffs to bring class actions under the 1934 statute instead of state law also forced them to file in federal court. That meant the bulk of securities class actions would proceed in federal court—because the 1934 Act regulates all trading of securities whereas the 1933 Act addresses only securities offerings. See Blue Chip Stamps, 421 U. S., at 752 (characterizing the 1933 Act as “a far narrower statute”). So even without Cyan’s contrived reading of the except clause, SLUSA largely accomplished the purpose articulated in its Conference Report: moving securities class actions to federal court. To be sure, “largely” does not mean “entirely”—but then again, we do not generally expect statutes to fulfill 100% of all of their goals. See, e.g., Freeman v. Quicken Loans, Inc., 566 U. S. 624, 637 (2012) (“No legislation pursues its purposes at all costs” (internal quotation marks and alterations omitted)). Under our reading of SLUSA, all covered securities class actions must proceed under federal law; most (i.e., those alleging 1934 Act claims) must proceed in federal court; some (i.e., those alleging 1933 Act claims) may proceed in state court. We do not know why Congress declined to require as well that 1933 Act class actions be brought in federal court; perhaps it was because of the long and unusually pronounced tradition of according authority to state courts over 1933 Act litigation. See supra, at 11–12. But in any event, we will not revise that legislative choice, by reading a conforming amendment and a definition in a most improbable way, in an effort to make the world of securities litigation more consistent or pure. This Court has long rejected the notion that “whatever furthers the statute’s primary objective must be the law.” Rodriguez v. United States, 480 U. S. 522, 526 (1987) (per curiam). Even if Congress could or should have done more, still it “wrote the statute it wrote—meaning, a statute going so far and no further.” Michigan v. Bay Mills Indian Community, 572 U. S. ___, ___ (2014) (slip op., at 11) (internal quotation marks omitted). 2 Yet Cyan has a final argument—that the except clause would serve no purpose at all unless it works as Cyan says. See Brief for Petitioners 32–33; Reply Brief 8–11. Here, Cyan relies on an indubitable puzzle. Section 77v(a), as amended by SLUSA, gives state courts jurisdiction over 1933 Act suits “except as provided in §77p.” But §77p provides a bar on only certain state-law suits. So, Cyan contends, unless we take up its invitation to look to §77p(f)(2)’s definition of “covered class action,” the except clause excepts “exactly nothing.” Reply Brief 8. (To use an example of our own, it would be as if a parent told her child “you may have fruit after dinner, except for lollipops.”) What on earth, Cyan asks, would be the point of such a provision? The Investors answer that question with a theory about why Congress enacted the except clause. In their view, the clause was meant to deal with “mixed” securities class actions—containing both claims brought under the 1933 Act and claims arising under state law. See Brief for Respondents 12–13. If not for the except clause, the Investors posit, state courts would have been uncertain about how to handle those suits. Section 77p clearly instructs courts not to adjudicate the state-law claims; but (the Investors continue) §77v(a) gives state courts jurisdiction over entire “actions” brought to enforce the 1933 Act, even if they include additional state-law claims. What, then, to do? According to the Investors, the except clause’s purpose was to resolve that statutory conflict by making clear that §77p trumps §77v(a)—in other words, that a state court may not entertain state-law claims precluded by §77p(b) even when they are conjoined with 1933 Act claims falling within §77v(a)’s grant of jurisdiction. Truth be told, we are not sure whether Congress had that issue in mind. On the one hand (and contrary to what the Investors say), we doubt that the except clause was really necessary to address mixed class actions. Even without that clause, a competent state court faced with such a suit would understand that §77p requires dismissal of the state-law claims—and that §77v(a)’s jurisdictional grant over 1933 Act suits is not to the contrary. But on the other hand (and now supporting the Investors’ principal point), Congress may have thought that class-action lawyers would still try to circumvent SLUSA by tacking a 1933 Act claim onto a forbidden state-law class action, on the off chance of finding an error-prone judge. (After all, the worst that could happen was that the court would throw out the state-law claims, leaving the plaintiff with a permissible 1933 Act suit.) To prevent such gamesmanship—to make clear beyond peradventure that courts could not entertain the state-law half of mixed class actions—Congress might have added the except clause. But even if Congress never specifically considered mixed suits, it could well have added the except clause in a more general excess of caution—to safeguard §77p’s class-action bar come whatever might. This Court has encountered many examples of Congress legislating in that hyper-vigilant way, to “remov[e] any doubt” as to things not particularly doubtful in the first instance. Marx v. General Revenue Corp., 568 U. S. 371 –384 (2013) (citing Ali v. Federal Bureau of Prisons, 552 U. S. 214, 226 (2008) ; Fort Stewart Schools v. FLRA, 495 U. S. 641, 646 (1990) ). (The idea, to return to our prior example, is to make sure that even if the child thinks orange lollipops count as fruit, she will not act on that view.) And if ever Congress had reason to legislate in that fashion, it was in SLUSA—whose very impetus lay in the success of class-action attorneys in “bypass[ing] . . . the Reform Act.” Kircher, 547 U. S., at 636. Heedful of that history of machinations, Congress may have determined to eliminate any risk—even if unlikely or at the time unknown—that a pre-existing grant of power to state courts could be used to obstruct SLUSA’s new limitation on what they could decide. And so (this alternative explanation goes) Congress enacted the except clause—which, in insisting that the limitation prevailed, would function as the ultimate (though with any luck, unneeded) fail-safe device.[4] But the most important response to this purposive argument echoes what we have said before about the weaknesses of Cyan’s own construction of the except clause. In the end, the uncertainty surrounding Congress’s reasons for drafting that clause does not matter. Nor does the possibility that the risk Congress addressed (whether specific or inchoate) did not exist. Because irrespective of those points, we have no sound basis for giving the except clause a broader reading than its language can bear. And that is especially true in light of the dramatic change such an interpretation would work in the 1933 Act’s jurisdictional framework. Whatever questions remain as to the except clause’s precise purpose—and we do not gainsay there are some—they do not give us permission to devise a statute (and at that, a transformative one) of our own. III Our last task is to address the Federal Government’s proposed halfway-house position. The Government rejects Cyan’s view that SLUSA stripped state courts of jurisdiction over 1933 Act class actions, for roughly the same reasons we have given. See Brief for United States as Amicus Curiae 11–23. But like Cyan, the Government believes that “Congress would not have been content to leave” such suits “stuck in state court,” where the Reform Act’s procedural protections do not apply. Id., at 15 (internal quotation marks omitted). So the Government offers a reading of SLUSA—in particular, of §77p(c)—that would allow defendants to remove 1933 Act class actions to federal court, as long as they allege the kinds of misconduct listed in §77p(b) (e.g., false statements or deceptive devices in connection with a covered security’s purchase or sale). See id., at 24–25. But most naturally read, §77p(c)—SLUSA’s exception to the 1933 Act’s general bar on removal—refutes, not supports, the Government’s view. Once again, see supra, at 4, §77p(c) reads as follows: “Any covered class action brought in any State court involving a covered security, as set forth in subsection (b) of this section, shall be removable to the Federal district court for the district in which the action is pending, and shall be subject to subsection (b) of this section.” In other words, the covered class actions described in §77p(b) can be removed to federal court (and, once there, shall be subject to dismissal because precluded, see supra, at 5). And which are the covered class actions described in §77p(b)? By this point, no one should have to be reminded: They are state-law class actions alleging securities misconduct. See §77p(b) (prohibiting “class action[s] based upon the statutory or common law of any State”). So those state-law suits are removable. But conversely, federal-law suits like this one—alleging only 1933 Act claims—are not “class action[s] . . . as set forth in subsection (b).” So they remain subject to the 1933 Act’s removal ban. In fact, this Court already held as much, by concluding in Kircher that §§77p(b) and 77p(c) apply to the exact same universe of class actions. See 547 U. S., at 643–644. Kircher involved a securities suit that was unaffected by §77p(b)’s class-action bar—there, not because it was based on federal law but because it involved a form of conduct falling outside that subsection. The Court of Appeals decided that the suit could be removed under §77p(c) even though it was not precluded by §77p(b), thinking (as we later put it) that the removal issue and “the preclusion issue [were] distinct.” Id., at 638. We flatly rejected that understanding of the relationship between §77p(b) and §77p(c). The “straightforward reading” of those two provisions, we explained, is that removal is “limited to those [actions] precluded by the terms of subsection (b).” Id., at 643. And if that were not clear enough, we said it again: Removal under §77p(c) is “restricted to precluded actions defined by subsection (b).” Id., at 643–644. And just to pound the point home, we said it yet a third time: “A covered [class] action is removable if it is precluded.” Id., at 646. Kircher thus forecloses the Government’s argument. Section 77p(b) does not preclude federal-law class actions. So under our decision, §77p(c) does not authorize their removal.[5] The Government responds with a novel way of understanding §77p(c), which it thinks would allow us to disregard Kircher when a class action like this one is based on federal law. In the Government’s view, first presented at oral argument, see Tr. of Oral Arg. 32–33, the words “as set forth in subsection (b)” do not modify the entire preceding phrase (basically, any large class action involving a covered security). Instead, the Government claims, those words modify only the shorter phrase “involving a covered security.” To support that view, the Government invokes the “rule of the last antecedent”—under which “the limiting clause is most naturally applied to the thing that comes immediately before it.” Id., at 36. The Government then presents a theory of how subsection (b) “set[s] forth” the “involv[ement]” of a covered security. “[T]o figure out what that means,” the Government contends, “you look at [§77p](b)(1) and (b)(2), which talk about certain types of misconduct”—for example, false statements or deceptive devices in connection with a covered security’s sale. Id., at 33–34. As long as conduct of that kind is implicated in a suit, the Government concludes, it can be removed—even if it is based on federal law and thus does not fall within §77p(b) as a whole. That view is consistent with Kircher’s result because the action there did not involve the conduct described in §§77p(b)(1) and (2). And as to Kircher’s rationale . . . well, we should feel free to ignore it. But even putting aside respect for precedent, that argument is in many ways flawed. To start with, the Government provides no good reason to think that “as set forth in subsection (b)” modifies only the phrase “involving a covered security.” As stated above, the most natural way to view the modifier is as applying to the entire preceding clause—again, “[a]ny covered class action brought in any State court involving a covered security.” See supra, at 19. That is so because that clause hangs together as a unified whole, referring to a single thing (a type of class action). Consider the following, grammatically identical construction: “The woman dressed to the nines carrying an umbrella, as shown in the picture . . .” Would anyone think that “as shown in the picture” referred to anything less than the well-attired and rain-ready woman? No. And so too here, the modifier goes back to the beginning of the preceding clause. The rule of the last antecedent is not to the contrary. We have applied that rule when the alternative reading would “stretch[] the modifier too far” by asking it to qualify a remote or otherwise disconnected phrase. Jama v. Immigration and Customs Enforcement, 543 U. S. 335, 342 (2005) ; Lockhart v. United States, 577 U. S. ___, ___ (2016) (slip op., at 4) (using the rule “where it takes more than a little mental energy to process” a statute’s component parts, “making it a heavy lift to carry the modifier across them all”).[6] By contrast, we have not applied the rule when the modifier directly follows a concise and “integrated” clause. Jama, 543 U. S., at 344, n. 4. As it does here. But let us assume that the rule of the last antecedent governs: The Government then misapplies it by attaching the modifier to something more than the last thing before it. The rule, correctly used, would insist that “as set forth in subsection (b)” modifies only “a covered security”—because that is the closest “noun or noun phrase” that the modifier could reasonably reference. A. Scalia & B. Garner, Reading Law: The Interpretation of Legal Texts 144 (2012) (quoting R. Burchfield, Fowler’s Modern English Usage (3d ed. 1996)). But that standard way of applying the rule would not aid the Government’s construction, so it goes back yet another word: It attaches “as set forth in subsection (b)” to the longer phrase—and a verb phrase at that—“involving a covered security.” Tr. of Oral Arg. 35. That maneuver has no grammatical basis. (It is as if, in the example offered above, someone claimed that “as shown in the picture” modified not the woman, nor even the umbrella, but instead the in-between verb phrase “carrying an umbrella.”) The Government is choosing where to start in the sentence (that is, which words to qualify) based only on what best serves its argument. But let us even assume that “as set forth in subsection (b)” modifies “involving a covered security”: The language would still fail to explain the Government’s position. Remember that the Government reads the resulting phrase (again, “involving a covered security, as set forth in subsection (b)”) to point only to the forms of wrongful conduct listed in §§77p(b)(1) and (2)—for example, false statements or deceptive devices in securities sales. See supra, at 21. The problem is that no one would describe those misdeeds with that phrase. If Congress had meant to refer only to that behavior, rather than to everything in §77p(b), it would have done two things differently. First, Congress would have written “as set forth in paragraphs (b)(1) and (b)(2)” instead of “as set forth in subsection (b)” as a whole. See supra, at 9 (explaining that when Congress wants to refer only to a particular subsection or paragraph, it says so). And second, Congress would have written something like “involving allegations of misconduct,” rather than “involving a covered security”—because the latter phrase does not even passably describe §§77(b)(1) and (2)’s catalog of vices. We will not read “involving a covered security, as set forth in subsection (b)” to mean “involving allegations of misconduct, as set forth in paragraphs (b)(1) and (b)(2)” when Congress did not enact that formulation. See Lozano v. Montoya Alvarez, 572 U. S. 1 , ___ (2014) (slip op., at 14) (“Given that the drafters did not adopt that alternative, the natural implication is that they did not intend” to do so). And (finally, we promise) even if we could put out of mind all these difficulties, the Government’s position runs aground on §77p(c)’s last clause, which states that removed class actions “shall be subject to subsection (b).” That clause, properly understood, points toward dismissal of a removed action. As we earlier explained, and the Government concedes, Congress enacted §77p(c)’s removal provision out of “concern[] that state courts would not adequately enforce” §77p(b)’s state-law class-action prohibition. Brief for United States as Amicus Curiae 3; see supra, at 5. The idea was to allow removal so that a federal court could act as a backstop and order a class action’s dismissal—thereby “subject[ing]” it to §77p(b)’s bar. Kircher specifically said as much: Section 77p(c) “avails a defendant of a federal forum in contemplation not of further litigation over the merits of a claim brought in state court, but of termination of the proceedings altogether.” 547 U. S., at 645, n. 12; see supra, at 5. But of course, the Government contemplates “further litigation”—not “termination”—of a removed 1933 Act class action. See Brief for United States as Amicus Curiae 25. That decoupling of §77p(c)’s linkage between removal and dismissal provides the last reason to reject the Government’s argument. At bottom, the Government makes the same mistake as Cyan: It distorts SLUSA’s text because it thinks Congress simply must have wanted 1933 Act class actions to be litigated in federal court. But this Court has no license to “disregard clear language” based on an in- tuition that “Congress must have intended something broader.” Bay Mills, 572 U. S., at ___ (slip op., at 11) (internal quotation marks omitted). SLUSA did quite a bit to “make good on the promise of the Reform Act” (as Cyan puts it). Brief for Petitioners 20; see supra, at 12–13. If further steps are needed, they are up to Congress. IV SLUSA did nothing to strip state courts of their longstanding jurisdiction to adjudicate class actions alleging only 1933 Act violations. Neither did SLUSA authorize removing such suits from state to federal court. We accordingly affirm the judgment below. It is so ordered. APPENDIX “77p. Additional remedies; limitation on remedies . . . . . “(b) Class action limitations “No covered class action based upon the statutory or common law of any State or subdivision thereof may be maintained in any State or Federal court by any private party alleging— “(1) an untrue statement or omission of a material fact in connection with the purchase or sale of a covered security; or “(2) that the defendant used or employed any manipulative or deceptive device or contrivance in connection with the purchase or sale of a covered security. “(c) Removal of covered class actions “Any covered class action brought in any State court involving a covered security, as set forth in subsection (b) of this section, shall be removable to the Federal district court for the district in which the action is pending, and shall be subject to subsection (b) of this section. . . . . . “(f) Definitions “For purposes of this section, the following definitions shall apply: . . . . . “(2) Covered class action “(A) In general “The term “covered class action” means— “(i) any single lawsuit in which— “(I) damages are sought on behalf of more than 50 persons or prospective class members, and questions of law or fact common to those persons or members of the prospective class, without reference to issues of individualized reliance on an alleged misstatement or omission, predominate over any questions affecting only individual persons or members; or “(II) one or more named parties seek to recover damages on a representative basis on behalf of themselves and other unnamed parties similarly situated, and questions of law or fact common to those persons or members of the prospective class predominate over any questions affecting only individual persons or members; or “(ii) any group of lawsuits filed in or pending in the same court and involving common questions of law or fact, in which— “(I) damages are sought on behalf of more than 50 persons; and “(II) the lawsuits are joined, consolidated, or otherwise proceed as a single action for any purpose. . . . . . “(3) Covered security “The term “covered security” means a security that satisfies the standards for a covered security specified in paragraph (1) or (2) of section 77r(b) of this title at the time during which it is alleged that the misrepresentation, omission, or manipulative or deceptive conduct occurred, except that such term shall not include any debt security that is exempt from registration under this subchapter pursuant to rules issued by the Commission under section 77d(2) of this title.” “77v. Jurisdiction of offenses and suits “(a) Federal and State courts; venue; service of process; review; removal; costs “The district courts of the United States and the United States courts of any Territory shall have jurisdiction of offenses and violations under this subchapter and under the rules and regulations promulgated by the Commission in respect thereto, and, concurrent with State and Territorial courts, except as provided in section 77p of this title with respect to covered class actions, of all suits in equity and actions at law brought to enforce any liability or duty created by this subchapter. Any such suit or action may be brought in the district wherein the defendant is found or is an inhabitant or transacts business, or in the district where the offer or sale took place, if the defendant participated therein, and process in such cases may be served in any other district of which the defendant is an inhabitant or wherever the defendant may be found. In any action or proceeding instituted by the Commission under this subchapter in a United States district court for any judicial district, a subpoena issued to compel the attendance of a witness or the production of documents or tangible things (or both) at a hearing or trial may be served at any place within the United States. Rule 45(c)(3)(A)(ii) of the Federal Rules of Civil Procedure shall not apply to a subpoena issued under the preceding sentence. Judgments and decrees so rendered shall be subject to review as provided in sections 1254, 1291, 1292, and 1294 of title 28. Except as provided in section 77p(c) of this title, no case arising under this subchapter and brought in any State court of competent jurisdiction shall be removed to any court of the United States. No costs shall be assessed for or against the Commission in any proceeding under this subchapter brought by or against it in the Supreme Court or such other courts.” Notes 1 Compare, e.g., Luther v. Countrywide Financial Corp., 195 Cal. App. 4th 789, 797–798, 125 Cal. Rptr. 3d 716, 721 (2011) (holding that state courts have jurisdiction over covered class actions alleging only 1933 Act claims), with, e.g., Knox v. Agria Corp., 613 F. Supp. 2d 419, 425 (SDNY 2009) (holding that state courts lack jurisdiction over such actions). 2 This Court has often applied a “presumption in favor of concurrent state court jurisdiction” when interpreting federal statutes. Mims v. Arrow Financial Services, LLC, 565 U. S. 368, 378 (2012) (quoting Tafflin v. Levitt, 493 U. S. 455 –459 (1990)). Cyan argues that the presumption should not apply here because SLUSA included explicit “language addressing state-court jurisdiction” and “the only question is [its] scope.” Reply Brief 22. We need not address that contention because SLUSA’s text precludes Cyan’s position without aid from any presumption. 3 See, e.g., §77k(f)(1) (announcing a general rule of joint and several liability, “[e]xcept as provided in paragraph (2),” which sets out a different liability rule for outside directors); §77p(a) (announcing a general rule that “the rights and remedies provided” under the statute do not displace others, “[e]xcept as provided in subsection (b),” which restricts the right to bring class actions based on state law); §77z–2(c) (announcing a general safe harbor for certain forward-looking statements, “[e]xcept as provided in subsection (b),” which excludes protection for a subset of them). 4 In line with this precautionary function, the except clause could do some work to protect §77p’s state-law class-action bar in a set of suits beyond mixed cases. As this Court recently noted, some state-law securities claims “rise[] or fall[] on the plaintiff’s ability to prove the violation” of a federal securities statute. Merrill Lynch, Pierce, Fenner & Smith Inc. v. Manning, 578 U. S. ___, ___ (2016) (slip op., at 7) (addressing the 1934 Act). In such cases, we held, the state-law claims are “brought to enforce” a liability created by the federal statute, and thus fall within that law’s jurisdictional provision. See id., at ___ (slip op., at 8). In the absence of the except clause, then, class-action lawyers might well have argued that such state-law claims could be adjudicated under §77v(a) notwithstanding §77p’s bar. Once again, we think most courts would have rejected that claim and decided that §77p controls—but the except clause eliminates any chance of a contrary holding. 5 In light of SLUSA’s text and Kircher’s holding, it should come as no surprise that all seven Courts of Appeals to have considered the matter have concluded that §77p(c) allows removal of only class actions falling within §77p(b)’s prohibition. See Campbell v. American Int’l Group, Inc., 760 F.3d 62, 64 (CADC 2014); Hidalgo-Velez v. San Juan Asset Management, 758 F.3d 98, 103 (CA1 2014); Appert v. Morgan Stanley Dean Witter, Inc., 673 F.3d 609, 615–616 (CA7 2012); Madden v. Cowen & Co., 576 F.3d 957, 965 (CA9 2009); Sofonia v. Principal Life Ins. Co., 465 F.3d 873, 876 (CA8 2006); Dabit v. Merrill Lynch, Pierce, Fenner & Smith, Inc., 395 F.3d 25, 33 (CA2 2005), vacated on other grounds, 547 U. S. 71 (2006) ; Behlen v. Merrill Lynch, 311 F.3d 1087, 1092 (CA11 2002). 6 The classic example comes from Barnhart v. Thomas, 540 U. S. 20 (2003) . The statute at issue provided that a person is disabled if his impairment is so severe that “he is not only unable to do his previous work but cannot, considering his age, education, and work experience, engage in any other kind of substantial gainful work which exists in the national economy.” Id., at 23 (quoting 42 U. S. C. §423(d)(1)(A)) (emphasis altered). Invoking the rule of the last antecedent, we concluded that the italicized phrase “which exists in the national economy” modifies only “substantial gainful work,” and not the more distant “previous work.” See 540 U. S., at 26. Needless to say, that statutory provision is a far cry from the one at issue here.

584.US.2017_17-43

Under federal law, a judge normally may issue a wiretap order permitting the interception of communications only “within the territorial jurisdiction of the court in which the judge is sitting.” 18 U. S. C. §2518(3). Here, a judge for the District of Kansas authorized nine wiretap Orders as part of a Government investigation of a suspected drug distribution ring in Kansas. For the most part, the Government intercepted communications from a listening post within Kansas. But each Order also contained a sentence purporting to authorize interception outside of Kansas. Based on that authorization, the Government intercepted additional communications from a listening post in Missouri. Following the investigation, petitioners Los and Roosevelt Dahda were indicted for participating in an illegal drug distribution conspiracy. They moved to suppress the evidence derived from all the wiretaps under subparagraph (ii) of the wiretap statute’s suppression provision because the language authorizing interception beyond the District Court’s territorial jurisdiction rendered each Order “insufficient on its face.” §2518(10)(a)(ii). The Government agreed not to introduce any evidence arising from its Missouri listening post, and the District Court denied the Dahdas’ motion. On appeal, the Tenth Circuit rejected the Dahdas’ facial-insufficiency argument on the ground that the challenged language did not implicate Congress’ core statutory concerns in enacting the wiretap statute. Held: Because the Orders were not lacking any information that the statute required them to include and would have been sufficient absent the challenged language authorizing interception outside the court’s territorial jurisdiction, the Orders were not facially insufficient. Pp. 6–12. (a) The Tenth Circuit applied the “core concerns” test from United States v. Giordano, 416 U. S. 505, and held that subparagraph (ii) applies only where the insufficiency reflects an order’s failure to satisfy the “statutory requirements that directly and substantially implement the congressional intention to limit the use of” wiretapping, id., at 527. The court identified two such core concerns and concluded that neither applies to the statute’s territorial limitation. But Giordano involved a different suppression provision—subparagraph (i)—which applies only when a “communication was unlawfully intercepted.” §2518(10)(a)(i). The underlying point of Giordano’s limitation was to help distinguish subparagraph (i) of §2518(10)(a) from subparagraphs (ii) and (iii). It makes little sense to extend the “core concerns” test to subparagraph (ii) as well. Subparagraph (ii) therefore does not include a Giordano-like “core concerns” requirement. Pp. 6–8. (b) That said, this Court also cannot fully endorse the Dahdas’ interpretation of the statute. The Dahdas read subparagraph (ii) as applying to any legal defect that appears within the four corners of an order. Clearly, subparagraph (ii) covers at least an order’s failure to include information required by §§2518(4)(a)–(e). But that does not mean that every defect that may conceivably appear in an order results in an insufficiency. Here, the sentence authorizing interception outside Kansas is surplus. Its presence is not connected to any other relevant part of the Orders. Absent the challenged language, every wiretap that produced evidence introduced at the Dahdas’ trial was properly authorized under the statute. While the Orders do not specifically list the territorial area where they could lawfully take effect, they clearly set forth the authorizing judge’s territorial jurisdiction—the District of Kansas. And the statute itself presumptively limits every Order’s scope to the issuing court’s territorial jurisdiction. This interpretation of the term “insufficient” does not, as the Dahdas contend, produce bizarre results. Rather, it makes sense of the suppression provision as a whole. Pp. 8–12. 853 F. 3d 1101 (first judgment) and 852 F. 3d 1282 (second judgment), affirmed. Breyer, J., delivered the opinion of the Court, in which all other Members joined, except Gorsuch, J., who took no part in the consideration or decision of the cases. Notes 1 Together with Dahda v. United States (see this Court’s Rule 12.4), also on certiorari to the same court.

A federal statute allows judges to issue wiretap orders authorizing the interception of communications to help prevent, detect, or prosecute serious federal crimes. See Omnibus Crime Control and Safe Streets Act of 1968, 18 U. S. C. §2510 et seq. The statute requires the judge to find “probable cause” supporting issuance of the order, and it sets forth other detailed requirements governing both the application for a wiretap and the judicial order that authorizes it. See §2518. The statute provides for the suppression of “the contents of any wire or oral communication” that a wiretap “intercept[s]” along with any “evidence derived therefrom” if “(i) the communication was unlawfully intercepted; “(ii) the order of . . . approval under which it was intercepted is insufficient on its face; or “(iii) the interception was not made in conformity with the order of authorization or approval.” §2518(10)(a). This litigation concerns the second of these provisions—the provision that governs the “insufficien[cy]” of an order “on its face.” §2518(10)(a)(ii). Los and Roosevelt Dahda—defendants in the trial below and petitioners here—sought to suppress evidence derived from nine wiretap Orders used to obtain evidence of their participation in an unlawful drug distribution conspiracy. They argue that each Order is “insufficient on its face” because each contains a sentence authorizing interception “outside the territorial jurisdiction” of the authorizing judge, App. 97 (emphasis added), even though the statute normally allows a judge to authorize wiretaps only within his or her “territorial jurisdiction,” §2518(3). In deciding whether each Order was “insufficient on its face,” we assume that the Dahdas are right about the “territorial” requirement. That is to say, we assume the relevant sentence exceeded the judge’s statutory author- ity. But none of the communications unlawfully inter- cepted outside the judge’s territorial jurisdiction were intro- duced at trial, so the inclusion of the extra sentence had no significant adverse effect upon the Dahdas. Because the remainder of each Order was itself legally sufficient, we conclude that the Orders were not “insufficient” on their “face.” I A As we just said, the relevant statute permits a judge to issue an order authorizing the Government to intercept wire communications for an initial (but extendable) period of 30 days. §2518(5). To obtain that order, the Government must submit an application that describes the particular offense being investigated as well as the type of communications it seeks to intercept; that sets forth the basis for an appropriate finding of “probable cause”; that explains why other less intrusive methods are inadequate, have failed, or are too dangerous to try; and that meets other requirements, showing, for example, authorization by a specified governmental official. §2518(1). If the judge accepts the application, finds probable cause, and issues an authorizing order, that order must itself contain specified information, including, for example, the identity of the “person” whose “communications are to be intercepted”; the “nature and location of the [relevant] communications facilities”; a “particular description of the type of communication sought to be intercepted”; a statement of the “particular offense” to which the intercept “relates”; the “identity of the agency authorized to intercept”; the iden- tity of the “person authorizing the application”; and “the period of time during which” the “interception is authorized.” §§2518(4)(a)–(e). A judge’s authorizing authority normally extends only within statutorily defined bounds. The statute specifies that an order can permit the interception of communications “within the territorial jurisdiction of the court in which the judge is sitting.” §2518(3). (There is an exception allowing interception beyond the judge’s territorial jurisdiction if the judge authorizes a “mobile interception device,” ibid., but the parties now agree that exception does not apply to these Orders.) The Government here adds (without the Dahdas’ disagreement) that an intercept takes place either where the tapped telephone is located or where the Government’s “listening post” is located. See §2510(4) (defining “intercept” as “the aural or other acquisition of the contents of any wire, electronic, or oral communication through the use of any electronic, mechanical, or other device”); see also Brief for Petitioners 11; Brief for United States 6. As so interpreted, the statute generally requires that one or the other or both of these locations must be found within the authorizing judge’s “territorial jurisdiction.” B In 2011, the Government began investigating a sus- pected drug distribution ring based in Kansas. It submitted an application asking a federal judge for the District of Kansas to issue nine related wiretap Orders, and the judge issued them. For present purposes we assume, see infra, at 10–11, that all nine Orders met all statutory requirements with one exception. Each Order contained a sentence that read as follows: “Pursuant to Title 18, United States Code §2518(3), it is further Ordered that, in the event TARGET TELEPHONE #1, TARGET TELEPHONE #3 and TARGET TELEPHONE #4, are transported outside the territorial jurisdiction of the court, interception may take place in any other jurisdiction within the United States.” App. 105 (under seal) (emphasis added); see also id., at 97, 114, 123, 132, 140, 149, 158, 166, 174 (Orders containing identical language but targeting different telephones). Although they disputed it below, the parties now agree that this sentence could not lawfully allow a wiretap of a phone that was located outside Kansas in instances where the Government’s listening post was also located outside of Kansas. Pursuant to these Orders, the Government listened from a listening post within Kansas to conversations on mobile phones that were located within Kansas and conversations on mobile phones that were located outside of Kansas. But, in one instance, the Government listened from a listening post outside of Kansas (in Missouri) to conversations on a mobile phone that was also outside of Kansas (in California). That one instance concerned a mobile phone (Target Telephone #7) belonging to Philip Alarcon. In 2012, the Government indicted the Dahdas and several others, charging them with conspiracy to buy illegal drugs in California and sell them in Kansas. Prior to trial, the Dahdas moved to suppress all evidence derived from the wiretaps authorized by the nine Orders on the ground that the District Court could not authorize the interception of calls from the Missouri listening post to and from Alarcon’s mobile phone in California. In its response, the Government said it would not introduce any evidence arising from its Missouri listening post. A Magistrate Judge and subsequently the District Court denied the Dahdas’ suppression motion. App. to Pet. for Cert. 59a–76a. The Dahdas appealed. They argued that, even though the Government did not use any wiretap information from the Missouri listening post, the court should have suppressed all evidence derived from any of the Orders. That, they said, is because each Order was “insufficient on its face” given the extra sentence authorizing interception outside Kansas. Hence the second subparagraph of the statute’s suppression provision required the evidence to be suppressed. §2518(10)(a)(ii). The U. S. Court of Appeals for the Tenth Circuit rejected this argument on the ground that the claimed insuffi- ciency concerned the statute’s territorial requirement. 853 F. 3d 1101, 1114–1116 (2017). That requirement, in its view, did not “ ‘implemen[t]’ ” Congress’ core statutory concerns in enacting the wiretap statute. Id., at 1114 (quoting United States v. Giordano, 416 U. S. 505, 527 (1974)). And for that reason a violation of the territorial requirement did not warrant suppression. See also 852 F. 3d 1282, 1290 (2017). The Dahdas filed a petition for certiorari, seeking review of the Tenth Circuit’s determination. And, in light of different related holdings among the Circuits, we granted that petition. Compare 853 F. 3d, at 1114–1116 (suppression was not required for orders authorizing suppression beyond the District Court’s territorial jurisdiction), and Adams v. Lankford, 788 F. 2d 1493, 1500 (CA11 1986) (same), with United States v. Glover, 736 F.3d 509, 515 (CADC 2013) (suppression required for territorial defect). II A The question before us concerns the interpretation of the suppression provision’s second subparagraph, which requires suppression where a wiretap order is “insufficient on its face.” §2518(10)(a)(ii). The Dahdas ask us to read subparagraph (ii) as applying to any legal defect that appears within the four corners of the order. The Government replies that the Dahdas’ approach would require suppression of evidence of serious criminal behavior due to the most minor of technical failures, including those that have little or no relation to any statutory objective. The Tenth Circuit, agreeing with the Government, held that subparagraph (ii) applies only where the “insuffi- ciency” constitutes an order’s failure to satisfy a “ ‘statutory requiremen[t] that directly and substantially implement[s] the congressional intention to limit the use of intercept procedures to those situations clearly calling for the employment of this extraordinary investigative device.’ ” 853 F. 3d, at 1114 (quoting Giordano, supra, at 527; second alteration in original). The court identified two such core concerns—“ ‘(1) protecting the privacy of wire and oral communications, and (2) delineating on a uniform basis the circumstances and conditions under which the interception of wire and oral communications may be authorized’ ”—and concluded that neither applies to the statute’s territorial limitation. 853 F. 3d, at 1114 (quoting S. Rep. No. 90–1097, p. 66 (1986)). Like the Dahdas, we believe that the Tenth Circuit’s interpretation of this provision is too narrow. The Tenth Circuit took the test it applied from this Court’s decision in United States v. Giordano, supra, at 527. But Giordano involved a different provision. Keep in mind that the statute sets forth three grounds for suppression: “(i) the communication was unlawfully intercepted; “(ii) the order of . . . approval under which it was intercepted is insufficient on its face; or “(iii) the interception was not made in conformity with the order of authorization or approval.” §2518(10)(a). Giordano focused not, as here, on the second subparagraph but on the first subparagraph, which calls for the suppression of “unlawfully intercepted” communications. In Giordano, a criminal defendant sought suppression of wiretap-gathered information on the ground that the wiretap application was unlawfully authorized. 416 U. S., at 525. A provision of the wiretap statute that has since been amended required an application to be approved by either the Attorney General or a designated Assistant Attorney General. See 18 U. S. C. §2516(1) (1970 ed.). But, in Giordano’s case, an executive assistant to the Assistant Attorney General—not the Assistant Attorney General himself—had approved the application. 416 U. S., at 510. The Government argued that this statutory violation did not violate the first subparagraph, i.e., it did not lead to an “unlawfu[l] intercept[ion],” 18 U. S. C. §2518(10)(a)(i), because that subparagraph covers only violations of the Constitution, not statutes. Giordano, 416 U. S., at 525–526. Otherwise, the Government added, subparagraphs (ii) and (iii)—which clearly cover some statutory violations—would be superfluous. Id., at 526. But this Court held that the first subparagraph did cover certain statu- tory violations, namely, violations of those statutory provisions that “implemented” the wiretap-related congres- sional concerns the Tenth Circuit mentioned in its opinion. Id., at 527. So construed, the suppression provision left room for the second and third subparagraphs to have separate legal force. The Court went on to hold that a violation of the approval-by-the-Attorney-General provision implicated Congress’ core concerns. Subparagraph (i) thus covered that particular statutory provision. And, finding the provision violated, it ordered the wiretap evidence suppressed. Id., at 527–528. Here, by contrast, we focus upon subparagraph (ii), which requires suppression when an order is facially insufficient. And in respect to this subparagraph, we can find no good reason for applying Giordano’s test. The underlying point of Giordano’s limitation was to help give independent meaning to each of §2518(10)(a)’s subparagraphs. It thus makes little sense to extend the core concerns test to subparagraph (ii) as well. Doing so would “actually treat that subparagraph as ‘surplusage’—precisely what [this] Court tried to avoid in Giordano.” Glover, 736 F. 3d, at 514. We consequently conclude that subparagraph (ii) does not contain a Giordano-like “core concerns” requirement. The statute means what it says. That is to say, subparagraph (ii) applies where an order is “insufficient on its face.” §2518(10)(a)(ii). B Although we believe the Tenth Circuit erred in applying Giordano’s core concerns test to subparagraph (ii), we cannot fully endorse the Dahdas’ reading of the statute either. In our view, subparagraph (ii) does not cover each and every error that appears in an otherwise sufficient order. It is clear that subparagraph (ii) covers at least an order’s failure to include information that §2518(4) specifically requires the order to contain. See §§2518(4)(a)–(e) (requiring an order to specify, e.g., the “identity of the person, if known, whose communications are to be intercepted,” “a particular description of the type of communication sought to be intercepted, and a statement of the particular offense to which it relates”); Brief for United States 17. An order lacking that information would deviate from the uniform authorizing requirements that Congress explicitly set forth, while also falling literally within the phrase “insufficient on its face.” But the Dahdas would have us go further and conclude that any defect that may appear on an order’s face would render it insufficient. The lower courts in various contexts have debated just which kinds of defects subparagraph (ii) covers. See, e.g., United States v. Moore, 41 F. 3d 370, 375–376 (CA8 1994) (order missing judge’s signature); United States v. Joseph, 519 F. 2d 1068, 1070 (CA5 1975) (order identifying the wrong Government official as authorizing the application); United States v. Vigi, 515 F. 2d 290, 293 (CA6 1975) (same). We need not, however, resolve the questions that these many different cases raise. We need only determine whether the defects in the Orders before us render them “insufficient.” We conclude that they do not. We rest that conclusion upon an argument that the Government did not make below but which it did set forth in its response to the petition for certiorari and at the beginning of its brief on the merits. That argument is closely related to the arguments the Government did make below. It has been fully briefed by both sides. And as we may “affir[m]” a lower court judgment “on any ground permitted by the law and the record,” Murr v. Wis- consin, 582 U. S. ___, ___ (2017) (slip op., at 19), we see little to be gained by remanding this litigation for further consideration. The argument is simply this: Subparagraph (ii) refers to an order that is “insufficient on its face.” An order is “insufficient” insofar as it is “deficient” or “lacking in what is necessary or requisite.” 5 Oxford English Dictionary 359 (1933); accord, Webster’s New International Dictionary 1288 (2d ed. 1957). And, looking, as the Dahdas urge us to do, at “the four corners of the order itself,” Reply Brief 4, we cannot find any respect in which the Orders are deficient or lacking in anything necessary or requisite. The Orders do contain a defect, namely, the sentence authorizing interception outside Kansas, which we set forth above. See supra, at 4. But not every defect results in an insufficiency. In that sentence, the District Court “further” ordered that interception may take place “outside the territorial jurisdiction of the court.” App. 97. The sentence is without legal effect because, as the parties agree, the Orders could not legally authorize a wiretap outside the District Court’s “territorial jurisdiction.” But, more importantly, the sentence itself is surplus. Its presence is not connected to any other relevant part of the Orders. Were we to remove the sentence from the Orders, they would then properly authorize wiretaps within the authorizing court’s territorial jurisdiction. As we discussed above, a listening post within the court’s territorial jurisdiction could lawfully intercept communications made to or from telephones located within Kansas or outside Kansas. See supra, at 3. Consequently, every wiretap that produced evidence introduced at the Dahdas’ trial was properly authorized under the statute. The Dahdas argue that, without the offending sentence, the Orders are “insufficient” because they then do not specifically list the territorial area where they could lawfully take effect. Reply Brief 6. The Orders, however, clearly set forth the authorizing judge’s territorial jurisdiction: the “District of Kansas.” See App. 100. And the statute itself presumptively limits every Order’s scope to the issuing court’s territorial jurisdiction. See §2518(3). We consequently fail to see how the additional language here at issue could render the Orders facially insufficient. The Dahdas add that interpreting the term “insufficient” as we have just done will produce “bizarre results.” Reply Brief 5. They claim that, under the Government’s logic, an order authorizing interception for 180 days would not be facially insufficient even though the wiretap statute expressly limits the maximum duration of a wiretap order to 30 days. §2518(5). To be sure, a 180-day order may raise problems that the language at issue here does not. On the one hand, it may be argued that such an order would be facially insufficient because without the 180-day provision the order would not contain any time limit at all. See §2518(4)(e). On the other hand, one might argue that such an order merely would be overly broad—not facially insufficient—and that suppression would be warranted only for those communications unlawfully intercepted after 30 days. See §2518(10)(a)(i). Regardless, we need not now address the Dahdas’ 180-day hypothetical. It is enough to say that the problems that may be associated with such an order are not present in this litigation. Here, the Orders would have been sufficient even if they lacked the language authorizing interception outside Kansas. And the Dahdas cannot seek suppression under subparagraph (i) given that the unlawfully intercepted communications from the Missouri listening post were not introduced at trial. Our interpretation of subparagraph (ii) makes sense of the suppression provision as a whole. Where the Government’s use of a wiretap is unconstitutional or violates a statutory provision that reflects Congress’ core concerns, an aggrieved person may suppress improperly acquired evidence under subparagraph (i) (as “unlawfully intercepted,” see Giordano, 416 U. S., at 527). Where an order lacks information that the wiretap statute requires it to include, an aggrieved person may suppress the fruits of the order under subparagraph (ii) (as “insufficient on its face”). And where the Government fails to comply with conditions set forth in the authorizing order, an aggrieved person may suppress its fruits under subparagraph (iii) (as an “interception . . . not made in conformity with the order of authorization or approval”). For these reasons, the judgments of the Court of Appeals are affirmed. It is so ordered. Justice Gorsuch took no part in the consideration or decision of these cases.

583.US.2017_16-1276

Endeavoring to root out corporate fraud, Congress passed the Sarbanes-Oxley Act of 2002 (Sarbanes-Oxley) and the 2010 Dodd-Frank Wall Street Reform and Consumer Protection Act (Dodd-Frank). Both Acts shield whistleblowers from retaliation, but they differ in important respects. Sarbanes-Oxley applies to all “employees” who report misconduct to the Securities and Exchange Commission (SEC or Commission), any other federal agency, Congress, or an internal supervisor. 18 U. S. C. §1514A(a)(1). Dodd-Frank defines a “whistleblower” as “any individual who provides . . . information relating to a violation of the securities laws to the Commission, in a manner established, by rule or regulation, by the Commission.” 15 U. S. C. §78u–6(a)(6). A whistleblower so defined is eligible for an award if original information provided to the SEC leads to a successful enforcement action. §78u–6(b)–(g). And he or she is protected from retaliation in three situations, see §78u–6(h)(1)(A)(i)–(iii), including for “making disclosures that are required or protected under” Sarbanes-Oxley or other specified laws, §78u–6(h)(1)(A)(iii). Sarbanes-Oxley’s anti-retaliation provision contains an administrative-exhaustion requirement and a 180-day administrative complaint-filing deadline, see 18 U. S. C. §1514A(b)(1)(A), (2)(D), whereas Dodd-Frank permits a whistleblower to sue an employer directly in federal district court, with a default six-year limitation period, see §78u–6(h)(1)(B)(i), (iii)(I)(aa). The SEC’s regulations implementing the Dodd-Frank provision contain two discrete whistleblower definitions. For purposes of the award program, Rule 21F–2 requires a whistleblower to “provide the Commission with information” relating to possible securities-law violations. 17 CFR §240.21F–2(a)(1). For purposes of the anti-retaliation protections, however, the Rule does not require SEC reporting. See §240.21F–2(b)(1)(i)–(ii). Respondent Paul Somers alleges that petitioner Digital Realty Trust, Inc. (Digital Realty) terminated his employment shortly after he reported to senior management suspected securities-law violations by the company. Somers filed suit, alleging, inter alia, a claim of whistleblower retaliation under Dodd-Frank. Digital Realty moved to dismiss that claim on the ground that Somers was not a whistleblower under §78u–6(h) because he did not alert the SEC prior to his termination. The District Court denied the motion, and the Ninth Circuit affirmed. The Court of Appeals concluded that §78u–6(h) does not necessitate recourse to the SEC prior to gaining “whistleblower” status, and it accorded deference to the SEC’s regulation under Chevron U. S. A. Inc. v. Natural Resources Defense Council, Inc., 467 U. S. 837 . Held: Dodd-Frank’s anti-retaliation provision does not extend to an individual, like Somers, who has not reported a violation of the securities laws to the SEC. Pp. 9–19. (a) A statute’s explicit definition must be followed, even if it varies from a term’s ordinary meaning. Burgess v. United States, 553 U. S. 124 . Section 78u–6(a) instructs that the statute’s definition of “whistleblower” “shall apply” “[i]n this section,” that is, throughout §78u–6. The Court must therefore interpret the term “whistleblower” in §78u–6(h), the anti-retaliation provision, in accordance with that definition. The whistleblower definition operates in conjunction with the three clauses of §78u–6(h)(1)(A) to spell out the provision’s scope. The definition first describes who is eligible for protection—namely, a “whistleblower” who provides pertinent information “to the Commission.” §78u–6(a)(6). The three clauses then describe what conduct, when engaged in by a “whistleblower,” is shielded from employment discrimination. An individual who meets both measures may invoke Dodd-Frank’s protections. But an individual who falls outside the protected category of “whistleblowers” is ineligible to seek redress under the statute, regardless of the conduct in which that individual engages. This reading is reinforced by another whistleblower-protection provision in Dodd-Frank, see 12 U. S. C. §5567(b), which imposes no requirement that information be conveyed to a government agency. Pp. 9–11. (b) The Court’s understanding is corroborated by Dodd-Frank’s purpose and design. The core objective of Dodd-Frank’s whistleblower program is to aid the Commission’s enforcement efforts by “motivat[ing] people who know of securities law violations to tell the SEC.” S. Rep. No. 111–176, p. 38 (emphasis added). To that end, Congress provided monetary awards to whistleblowers who furnish actionable information to the Commission. Congress also complemented the financial incentives for SEC reporting by heightening protection against retaliation. Pp. 11–12. (c) Somers and the Solicitor General contend that Dodd-Frank’s “whistleblower” definition applies only to the statute’s award program and not, as the definition plainly states, to its anti-retaliation provision. Their concerns do not support a departure from the statutory text. Pp. 12–18. (1) They claim that the Court’s reading would vitiate the protections of clause (iii) for whistleblowers who make disclosures to persons and entities other than the SEC. See §78u–6(h)(1)(A)(iii). But the plain-text reading of the statute leaves the third clause with substantial meaning by protecting a whistleblower who reports misconduct both to the SEC and to another entity, but suffers retaliation because of the latter, non-SEC, disclosure. Pp. 13–15. (2) Nor would the Court’s reading jettison protections for auditors, attorneys, and other employees who are required to report information within the company before making external disclosures. Such employees would be shielded as soon as they also provide relevant information to the Commission. And Congress may well have considered adequate the safeguards already afforded to such employees by Sarbanes-Oxley. Pp. 15–16. (3) Applying the “whistleblower” definition as written, Somers and the Solicitor General further protest, will allow “identical misconduct” to “go punished or not based on the happenstance of a separate report” to the SEC. Brief for Respondent 37–38. But it is understandable that the statute’s retaliation protections, like its financial rewards, would be reserved for employees who have done what Dodd-Frank seeks to achieve by reporting information about unlawful activity to the SEC. P. 16. (4) The Solicitor General observes that the statute contains no apparent requirement of a “temporal or topical connection between the violation reported to the Commission and the internal disclosure for which the employee suffers retaliation.” Brief for United States as Amicus Curiae 25. The Court need not dwell on related hypotheticals, which veer far from the case at hand. Pp. 16–18. (5) Finally, the interpretation adopted here would not undermine clause (ii) of §78u–6(h)(1)(A), which prohibits retaliation against a whistleblower for “initiating, testifying in, or assisting in any investigation or . . . action of the Commission based upon” information conveyed to the SEC by a whistleblower in accordance with the statute. The statute delegates authority to the Commission to establish the “manner” in which a whistleblower may provide information to the SEC. §78u–6(a)(6). Nothing prevents the Commission from enumerating additional means of SEC reporting, including through testimony protected by clause (ii). P. 18. (d) Because “Congress has directly spoken to the precise question at issue,” Chevron, 467 U. S., at 842, deference is not accorded to the contrary view advanced by the SEC in Rule 21F–2. Pp. 18–19. 850 F. 3d 1045, reversed and remanded. Ginsburg, J., delivered the opinion of the Court, in which Roberts, C. J., and Kennedy, Breyer, Sotomayor, and Kagan, JJ., joined. Sotomayor, J., filed a concurring opinion, in which Breyer, J., joined. Thomas, J., filed an opinion concurring in part and concurring in the judgment, in which Alito and Gorsuch, JJ., joined.

Endeavoring to root out corporate fraud, Congress passed the Sarbanes-Oxley Act of 2002, 116Stat. 745 (Sarbanes-Oxley), and the 2010 Dodd-Frank Wall Street Reform and Consumer Protection Act, 124Stat. 1376 (Dodd-Frank). Both Acts shield whistleblowers from retaliation, but they differ in important respects. Most notably, Sarbanes-Oxley applies to all “employees” who report misconduct to the Securities and Exchange Commission (SEC or Commission), any other federal agency, Congress, or an internal supervisor. 18 U. S. C. §1514A(a)(1). Dodd-Frank delineates a more circumscribed class; it defines “whistleblower” to mean a person who provides “information relating to a violation of the securities laws to the Commission.” 15 U. S. C. §78u–6(a)(6). A whistleblower so defined is eligible for an award if original information he or she provides to the SEC leads to a successful enforcement action. §78u–6(b)–(g). And, most relevant here, a whistleblower is protected from retaliation for, inter alia, “making disclosures that are required or protected under” Sarbanes-Oxley, the Securities Exchange Act of 1934, the criminal anti-retaliation proscription at 18 U. S. C. §1513(e), or any other law subject to the SEC’s jurisdiction. 15 U. S. C. §78u–6(h)(1)(A)(iii). The question presented: Does the anti-retaliation provision of Dodd-Frank extend to an individual who has not reported a violation of the securities laws to the SEC and therefore falls outside the Act’s definition of “whistleblower”? Pet. for Cert. (I). We answer that question “No”: To sue under Dodd-Frank’s anti-retaliation provision, a person must first “provid[e] . . . information relating to a violation of the securities laws to the Commission.” §78u–6(a)(6). I A “To safeguard investors in public companies and restore trust in the financial markets following the collapse of Enron Corporation,” Congress enacted Sarbanes-Oxley in 2002. Lawson v. FMR LLC, 571 U. S. 429 , ___ (2014) (slip op., at 1). Most pertinent here, Sarbanes-Oxley created new protections for employees at risk of retaliation for reporting corporate misconduct. See 18 U. S. C. §1514A. Section 1514A prohibits certain companies from discharging or otherwise “discriminat[ing] against an employee in the terms and conditions of employment because” the employee “provid[es] information . . . or otherwise assist[s] in an investigation regarding any conduct which the employee reasonably believes constitutes a violation” of certain criminal fraud statutes, any SEC rule or regulation, or “any provision of Federal law relating to fraud against shareholders.” §1514A(a)(1). An employee qualifies for protection when he or she provides information or assistance either to a federal regulatory or law enforcement agency, Congress, or any “person with supervisory authority over the employee.” §1514A(a)(1)(A)–(C).[1] To recover under §1514A, an aggrieved employee must exhaust administrative remedies by “filing a complaint with the Secretary of Labor.” §1514A(b)(1)(A); see Lawson, 571 U. S., at ___–___ (slip op., at 5–6). Congress prescribed a 180-day limitation period for filing such a complaint. §1514A(b)(2)(D). If the agency “does not issue a final decision within 180 days of the filing of [a] complaint, and the [agency’s] delay is not due to bad faith on the claimant’s part, the claimant may proceed to federal district court for de novo review.” Id., at ___ (slip op., at 6) (citing §1514A(b)). An employee who prevails in a proceeding under §1514A is “entitled to all relief necessary to make the employee whole,” including reinstatement, backpay with interest, and any “special damages sustained as a result of the discrimination,” among such damages, litigation costs. §1514A(c). B 1 At issue in this case is the Dodd-Frank anti-retaliation provision enacted in 2010, eight years after the enactment of Sarbanes-Oxley. Passed in the wake of the 2008 financial crisis, Dodd-Frank aimed to “promote the financial stability of the United States by improving accountability and transparency in the financial system.” 124Stat. 1376. Dodd-Frank responded to numerous perceived shortcomings in financial regulation. Among them was the SEC’s need for additional “power, assistance and money at its disposal” to regulate securities markets. S. Rep. No. 111–176, pp. 36, 37 (2010). To assist the Commission “in identifying securities law violations,” the Act established “a new, robust whistleblower program designed to motivate people who know of securities law violations to tell the SEC.” Id., at 38. And recognizing that “whistleblowers often face the difficult choice between telling the truth and . . . committing ‘career suicide,’ ” Congress sought to protect whistleblowers from employment discrimination. Id., at 111, 112. Dodd-Frank implemented these goals by adding a new provision to the Securities Exchange Act of 1934: 15 U. S. C. §78u–6. Section 78u–6 begins by defining a “whistleblower” as “any individual who provides . . . information relating to a violation of the securities laws to the Commission, in a manner established, by rule or regulation, by the Commission.” §78u–6(a)(6) (emphasis added). That definition, the statute directs, “shall apply” “[i]n this section”—i.e., throughout §78u–6. §78u–6(a). Section 78u–6 affords covered whistleblowers both incentives and protection. First, the section creates an award program for “whistleblowers who voluntarily provid[e] original information to the Commission that le[ads] to the successful enforcement of [a] covered judicial or administrative action.” §78u–6(b)(1). A qualifying whistleblower is entitled to a cash award of 10 to 30 percent of the monetary sanctions collected in the enforcement action. See §78u–6(b)(1)(A)–(B). Second, §78u–6(h) prohibits an employer from discharging, harassing, or otherwise discriminating against a “whistleblower” “because of any lawful act done by the whistleblower” in three situations: first, “in providing information to the Commission in accordance with [§78u–6],” §78u–6(h)(1)(A)(i); second, “in initiating, testifying in, or assisting in any investigation or . . . action of the Commission based upon” information provided to the SEC in accordance with §78u–6, §78u–6(h)(1)(A)(ii); and third, “in making disclosures that are required or protected under” either Sarbanes-Oxley, the Securities Exchange Act of 1934, the criminal anti-retaliation prohibition at 18 U. S. C. §1513(e),[2] or “any other law, rule, or regulation subject to the jurisdiction of the Commission,” §78u–6(h)(1)(A)(iii). Clause (iii), by cross-referencing Sarbanes-Oxley and other laws, protects disclosures made to a variety of individuals and entities in addition to the SEC. For example, the clause shields an employee’s reports of wrongdoing to an internal supervisor if the reports are independently safeguarded from retaliation under Sarbanes-Oxley. See supra, at 2–3.[3] The recovery procedures under the anti-retaliation provisions of Dodd-Frank and Sarbanes-Oxley differ in critical respects. First, unlike Sarbanes-Oxley, which contains an administrative-exhaustion requirement and a 180-day administrative complaint-filing deadline, see 18 U. S. C. §1514A(b)(1)(A), (2)(D), Dodd-Frank permits a whistleblower to sue a current or former employer directly in federal district court, with a default limitation period of six years, see §78u–6(h)(1)(B)(i), (iii)(I)(aa). Second, Dodd-Frank instructs a court to award to a prevailing plaintiff double backpay with interest, see §78u–6(h)(1)(C)(ii), while Sarbanes-Oxley limits recovery to actual backpay with interest, see 18 U. S. C. §1514A(c)(2)(B). Like Sarbanes-Oxley, however, Dodd-Frank authorizes reinstatement and compensation for litigation costs, expert witness fees, and reasonable attorneys’ fees. Compare §78u–6(h)(1)(C)(i), (iii), with 18 U. S. C. §1514A(c)(2)(A), (C).[4] 2 Congress authorized the SEC “to issue such rules and regulations as may be necessary or appropriate to implement the provisions of [§78u–6] consistent with the purposes of this section.” §78u–6(j). Pursuant to this author- ity, the SEC published a notice of proposed rulemaking to “Implemen[t] the Whistleblower Provisions” of Dodd-Frank. 75 Fed. Reg. 70488 (2010). Proposed Rule 21F–2(a) defined a “whistleblower,” for purposes of both the award and anti-retaliation provisions of §78u–6, as one or more individuals who “provide the Commission with information relating to a potential violation of the securities laws.” Id., at 70519 (proposed 17 CFR §240.21F–2(a)). The proposed rule, the agency noted, “tracks the statutory definition of a ‘whistleblower’ ” by requiring information reporting to the SEC itself. 75 Fed. Reg. 70489. In promulgating the final Rule, however, the agency changed course. Rule 21F–2, in finished form, contains two discrete “whistleblower” definitions. See 17 CFR §240.21F–2(a)–(b) (2017). For purposes of the award program, the Rule states that “[y]ou are a whistleblower if . . . you provide the Commission with information . . . relat[ing] to a possible violation of the Federal securities laws.” §240.21F–2(a)(1) (emphasis added). The information must be provided to the SEC through its website or by mailing or faxing a specified form to the SEC Office of the Whistleblower. See ibid.; §240.21F–9(a)(1)–(2). “For purposes of the anti-retaliation protections,” however, the Rule states that “[y]ou are a whistleblower if . . . [y]ou possess a reasonable belief that the information you are providing relates to a possible securities law violation” and “[y]ou provide that information in a manner described in” clauses (i) through (iii) of §78u–6(h)(1)(A). 17 CFR §240.21F–2(b)(1)(i)–(ii). “The anti-retaliation protections apply,” the Rule emphasizes, “whether or not you satisfy the requirements, procedures and conditions to qualify for an award.” §240.21F–2(b)(1)(iii). An individual may therefore gain anti-retaliation protection as a “whistleblower” under Rule 21F–2 without providing information to the SEC, so long as he or she provides information in a manner shielded by one of the anti-retaliation provision’s three clauses. For example, a report to a company supervisor would qualify if the report garners protection under the Sarbanes-Oxley anti-retaliation provision.[5] C Petitioner Digital Realty Trust, Inc. (Digital Realty) is a real estate investment trust that owns, acquires, and develops data centers. See Brief for Petitioner 3. Digital Realty employed respondent Paul Somers as a Vice President from 2010 to 2014. See 119 F. Supp. 3d 1088, 1092 (ND Cal. 2015). Somers alleges that Digital Realty terminated him shortly after he reported to senior management suspected securities-law violations by the company. See ibid. Although nothing impeded him from alerting the SEC prior to his termination, he did not do so. See Tr. of Oral Arg. 45. Nor did he file an administrative complaint within 180 days of his termination, rendering him ineligible for relief under Sarbanes-Oxley. See ibid.; 18 U. S. C. §1514A(b)(2)(D). Somers brought suit in the United States District Court for the Northern District of California alleging, inter alia, a claim of whistleblower retaliation under Dodd-Frank. Digital Realty moved to dismiss that claim, arguing that “Somers does not qualify as a ‘whistleblower’ under [§78u–6(h)] because he did not report any alleged law violations to the SEC.” 119 F. Supp. 3d, at 1094. The District Court denied the motion. Rule 21F–2, the court observed, does not necessitate recourse to the SEC prior to gaining “whistleblower” status under Dodd-Frank. See id., at 1095–1096. Finding the statutory scheme ambiguous, the court accorded deference to the SEC’s Rule under Chevron U. S. A. Inc. v. Natural Resources Defense Council, Inc., 467 U. S. 837 (1984) . See 119 F. Supp. 3d, at 1096–1106. On interlocutory appeal, a divided panel of the Court of Appeals for the Ninth Circuit affirmed. 850 F. 3d 1045 (2017). The majority acknowledged that Dodd-Frank’s definitional provision describes a “whistleblower” as an individual who provides information to the SEC itself. Id., at 1049. But applying that definition to the anti-retaliation provision, the majority reasoned, would narrow the third clause of §78u–6(h)(1)(A) “to the point of absurd- ity”: The statute would protect employees only if they “reported possible securities violations both internally and to the SEC.” Ibid. Such dual reporting, the majority believed, was unlikely to occur. Ibid. Therefore, the majority concluded, the statute should be read to protect employees who make disclosures privileged by clause (iii) of §78u–6(h)(1)(A), whether or not those employees also provide information to the SEC. Id., at 1050. In any event, the majority held, the SEC’s resolution of any statutory ambiguity warranted deference. Ibid. Judge Owens dissented. In his view, the statutory definition of whistleblower was clear, left no room for interpretation, and plainly governed. Id., at 1051. Two other Courts of Appeals have weighed in on the question before us. The Court of Appeals for the Fifth Circuit has held that employees must provide information to the SEC to avail themselves of Dodd-Frank’s anti-retaliation safeguard. See Asadi v. G. E. Energy (USA), L. L. C., 720 F. 3d 620, 630 (2013). A divided panel of the Court of Appeals for the Second Circuit reached the opposite conclusion, over a dissent by Judge Jacobs. See Berman v. NEO@OGILVY LLC, 801 F. 3d 145, 155 (2013). We granted certiorari to resolve this conflict, 582 U. S. ___ (2017), and now reverse the Ninth Circuit’s judgment. II “When a statute includes an explicit definition, we must follow that definition,” even if it varies from a term’s ordinary meaning. Burgess v. United States, 553 U. S. 124, 130 (2008) (internal quotation marks omitted). This principle resolves the question before us. A Our charge in this review proceeding is to determine the meaning of “whistleblower” in §78u–6(h), Dodd-Frank’s anti-retaliation provision. The definition section of the statute supplies an unequivocal answer: A “whistleblower” is “any individual who provides . . . information relating to a violation of the securities laws to the Commission.” §78u–6(a)(6) (emphasis added). Leaving no doubt as to the definition’s reach, the statute instructs that the “definitio[n] shall apply” “[i]n this section,” that is, throughout §78u–6. §78u–6(a)(6). The whistleblower definition operates in conjunction with the three clauses of §78u–6(h)(1)(A) to spell out the provision’s scope. The definition first describes who is eligible for protection—namely, a whistleblower who provides pertinent information “to the Commission.” §78u–6(a)(6). The three clauses of §78u–6(h)(1)(A) then describe what conduct, when engaged in by a whistle- blower, is shielded from employment discrimination. See §78u–6(h)(1)(A)(i)–(iii). An individual who meets both measures may invoke Dodd-Frank’s protections. But an individual who falls outside the protected category of “whistleblowers” is ineligible to seek redress under the statute, regardless of the conduct in which that individual engages. Reinforcing our reading, another whistleblower-protection provision in Dodd-Frank imposes no requirement that information be conveyed to a government agency. Title 10 of the statute, which created the Consumer Financial Protection Bureau (CFPB), prohibits discrimination against a “covered employee” who, among other things, “provide[s] . . . information to [his or her] employer, the Bureau, or any other State, local, or Federal, government authority or law enforcement agency relating to” a violation of a law subject to the CFPB’s jurisdiction. 12 U. S. C. §5567(a)(1). To qualify as a “covered employee,” an individual need not provide information to the CFPB, or any other entity. See §5567(b) (“covered employee” means “any individual performing tasks related to the offering or provision of a consumer financial product or service”). “[W]hen Congress includes particular language in one section of a statute but omits it in another[,] . . . this Court presumes that Congress intended a difference in meaning.” Loughrin v. United States, 573 U. S. ___, ___ (2014) (slip op., at 6) (internal quotation marks and alteration omitted). Congress placed a government-reporting requirement in §78u–6(h), but not elsewhere in the same statute. Courts are not at liberty to dispense with the condition—tell the SEC—Congress imposed. B Dodd-Frank’s purpose and design corroborate our comprehension of §78u–6(h)’s reporting requirement. The “core objective” of Dodd-Frank’s robust whistleblower program, as Somers acknowledges, Tr. of Oral Arg. 45, is “to motivate people who know of securities law violations to tell the SEC,” S. Rep. No. 111–176, at 38 (emphasis added). By enlisting whistleblowers to “assist the Government [in] identify[ing] and prosecut[ing] persons who have violated securities laws,” Congress undertook to improve SEC enforcement and facilitate the Commission’s “recover[y] [of] money for victims of financial fraud.” Id., at 110. To that end, §78u–6 provides substantial monetary rewards to whistleblowers who furnish actionable information to the SEC. See §78u–6(b). Financial inducements alone, Congress recognized, may be insufficient to encourage certain employees, fearful of employer retaliation, to come forward with evidence of wrongdoing. Congress therefore complemented the Dodd-Frank monetary incentives for SEC reporting by heightening protection against retaliation. While Sarbanes-Oxley contains an administrative-exhaustion requirement, a 180-day administrative complaint-filing deadline, and a remedial scheme limited to actual damages, Dodd-Frank provides for immediate access to federal court, a generous statute of limitations (at least six years), and the opportunity to recover double backpay. See supra, at 5–6. Dodd-Frank’s award program and anti-retaliation provision thus work synchronously to motivate individuals with knowledge of illegal activity to “tell the SEC.” S. Rep. No. 111–176, at 38. When enacting Sarbanes-Oxley’s whistleblower regime, in comparison, Congress had a more far-reaching objective: It sought to disturb the “corporate code of silence” that “discourage[d] employees from reporting fraudulent behavior not only to the proper authorities, such as the FBI and the SEC, but even internally.” Lawson, 571 U. S., at ___ (slip op., at 4) (internal quotation marks omitted). Accordingly, the Sarbanes-Oxley anti-retaliation provision covers employees who report fraud not only to the SEC, but also to any other federal agency, Congress, or an internal supervisor. See 18 U. S. C. §1514A(a)(1). C In sum, Dodd-Frank’s text and purpose leave no doubt that the term “whistleblower” in §78u–6(h) carries the meaning set forth in the section’s definitional provision. The disposition of this case is therefore evident: Somers did not provide information “to the Commission” before his termination, §78u–6(a)(6), so he did not qualify as a “whistleblower” at the time of the alleged retaliation. He is therefore ineligible to seek relief under §78u–6(h). III Somers and the Solicitor General tender a different view of Dodd-Frank’s compass. The whistleblower definition, as they see it, applies only to the statute’s award program, not to its anti-retaliation provision, and thus not, as the definition plainly states, throughout “this section,” §78u–6(a). See Brief for Respondent 30; Brief for United States as Amicus Curiae 10–11. For purposes of the anti-retaliation provision alone, they urge us to construe the term “whistleblower” in its “ordinary sense,” i.e., without any SEC-reporting requirement. Brief for Respondent 18. Doing so, Somers and the Solicitor General contend, would align with our precedent, specifically Lawson v. Suwannee Fruit & S. S. Co., 336 U. S. 198 (1949) , and Utility Air Regulatory Group v. EPA, 573 U. S. ___ (2014). In those decisions, we declined to apply a statutory definition that ostensibly governed where doing so would have been “incompatible with . . . Congress’ regulatory scheme,” id., at ___ (slip op., at 18) (internal quotation marks omitted), or would have “destroy[ed] one of the [statute’s] major purposes,” Suwannee Fruit, 336 U. S., at 201. This case is of a piece, Somers and the Solicitor General maintain. Applying the statutory definition here, they variously charge, would “create obvious incongruities,” Brief for United States as Amicus Curiae 19 (internal quotation marks omitted), “produce anomalous results,” id., at 22, “vitiate much of the [statute’s] protection,” id., at 20 (internal quotation marks omitted), and, as the Court of Appeals put it, narrow clause (iii) of §78u–6(h)(1)(A) “to the point of absurdity,” Brief for Respondent 35 (quoting 850 F. 3d, at 1049). We next address these concerns and explain why they do not lead us to depart from the statutory text. A It would gut “much of the protection afforded by” the third clause of §78u–6(h)(1)(a), Somers and the Solicitor General urge most strenuously, to apply the whistleblower definition to the anti-retaliation provision. Brief for United States as Amicus Curiae 20 (internal quotation marks omitted); Brief for Respondent 28–29. As earlier noted, see supra, at 4–5, clause (iii) prohibits retaliation against a “whistleblower” for “making disclosures” to various persons and entities, including but not limited to the SEC, to the extent those disclosures are “required or protected under” various laws other than Dodd-Frank. §78u–6(h)(1)(A)(iii). Applying the statutory definition of whistleblower, however, would limit clause (iii)’s protection to “only those individuals who report to the Commission.” Brief for United States as Amicus Curiae 22. The plain-text reading of the statute undoubtedly shields fewer individuals from retaliation than the alternative proffered by Somers and the Solicitor General. But we do not agree that this consequence “vitiate[s]” clause (iii)’s protection, id., at 20 (internal quotation marks omitted), or ranks as “absur[d],” Brief for Respondent 35 (internal quotation marks omitted).[6] In fact, our reading leaves the third clause with “substantial meaning.” Brief for Petitioner 32. With the statutory definition incorporated, clause (iii) protects a whistleblower who reports misconduct both to the SEC and to another entity, but suffers retaliation because of the latter, non-SEC, disclosure. That would be so, for example, where the retaliating employer is un- aware that the employee has alerted the SEC. In such a case, without clause (iii), retaliation for internal reporting would not be reached by Dodd-Frank, for clause (i) applies only where the employer retaliates against the employee “because of” the SEC reporting. §78u–6(h)(1)(A). More- over, even where the employer knows of the SEC reporting, the third clause may operate to dispel a proof problem: The employee can recover under the statute without having to demonstrate whether the retaliation was motivated by the internal report (thus yielding protection under clause (iii)) or by the SEC disclosure (thus gaining protection under clause (i)). While the Solicitor General asserts that limiting the protections of clause (iii) to dual reporters would “shrink to insignificance the [clause’s] ban on retaliation,” Brief for United States as Amicus Curiae 22 (internal quotation marks omitted), he offers scant evidence to support that assertion. Tugging in the opposite direction, he reports that approximately 80 percent of the whistleblowers who received awards in 2016 “reported internally before reporting to the Commission.” Id., at 23. And Digital Realty cites real-world examples of dual reporters seeking Dodd-Frank or Sarbanes-Oxley recovery for alleged retaliation. See Brief for Petitioner 33, and n. 4 (collecting cases). Overlooked by Somers and the Solicitor General, in dual-reporting cases, retaliation not prompted by SEC disclosures (and thus unaddressed by clause (i)) is likely commonplace: The SEC is required to protect the identity of whistleblowers, see §78u–6(h)(2)(A), so employers will often be unaware that an employee has reported to the Commission. In any event, even if the number of individuals qualifying for protection under clause (iii) is relatively limited, “[i]t is our function to give the statute the effect its language suggests, however modest that may be.” Morrison v. National Australia Bank Ltd., 561 U. S. 247, 270 (2010) . B Somers and the Solicitor General express concern that our reading would jettison protection for auditors, attorneys, and other employees subject to internal-reporting requirements. See Brief for Respondent 35; Brief for United States as Amicus Curiae 21. Sarbanes-Oxley, for example, requires auditors and attorneys to report certain information within the company before making disclosures externally. See 15 U. S. C. §§78j–1(b), 7245; 17 CFR §205.3. If the whistleblower definition applies, Somers and the Solicitor General fear, these professionals will be “le[ft] . . . vulnerable to discharge or other retaliatory action for complying with” their internal-reporting obligations. Brief for United States as Amicus Curiae 22 (internal quotation marks omitted). Our reading shields employees in these circumstances, however, as soon as they also provide relevant information to the Commission. True, such employees will remain ineligible for Dodd-Frank’s protection until they tell the SEC, but this result is consistent with Congress’ aim to encourage SEC disclosures. See S. Rep. No. 111–176, at 38; supra, at 3–4, 11. Somers worries that lawyers and auditors will face retaliation quickly, before they have a chance to report to the SEC. Brief for Respondent 35–36. But he offers nothing to show that Congress had this concern in mind when it enacted §78u–6(h). Indeed, Congress may well have considered adequate the safeguards already afforded by Sarbanes-Oxley, protections specifi- cally designed to shield lawyers, accountants, and similar professionals. See Lawson, 571 U. S., at ___ (slip op., at 17). C Applying the whistleblower definition as written, Somers and the Solicitor General further protest, will create “an incredibly unusual statutory scheme”: “[I]dentical misconduct”—i.e., retaliating against an employee for internal reporting—will “go punished or not based on the happenstance of a separate report” to the SEC, of which the wrongdoer may “not even be aware.” Brief for Respondent 37–38. See also Brief for United States as Amicus Curiae 24. The upshot, the Solicitor General warns, “would [be] substantially diminish[ed] Dodd-Fran[k] deterrent effect.” Ibid. Overlooked in this protest is Dodd-Frank’s core objective: to prompt reporting to the SEC. Supra, at 3–4, 11. In view of that precise aim, it is understandable that the statute’s retaliation protections, like its financial rewards, would be reserved for employees who have done what Dodd-Frank seeks to achieve, i.e., they have placed information about unlawful activity before the Commission to aid its enforcement efforts. D Pointing to another purported anomaly attending the reading we adopt today, the Solicitor General observes that neither the whistleblower definition nor §78u–6(h) contains any requirement of a “temporal or topical connection between the violation reported to the Commission and the internal disclosure for which the employee suffers retaliation.” Brief for United States as Amicus Curiae 25. It is therefore possible, the Solicitor General posits, that “an employee who was fired for reporting accounting fraud to his supervisor in 2017 would have a cause of action under [§78u–6(h)] if he had reported an insider-trading violation by his previous employer to the Commission in 2012.” Ibid. For its part, Digital Realty agrees that this scenario could arise, but does not see it as a cause for concern: “Congress,” it states, “could reasonably have made the policy judgment that individuals who report securities-law violations to the SEC should receive broad protection over time against retaliation for a variety of disclosures.” Reply Brief 11. We need not dwell on the situation hypothesized by the Solicitor General, for it veers far from the case before us. We note, however, that the interpretation offered by Somers and the Solicitor General—i.e., ignoring the statutory definition when construing the anti-retaliation provision—raises an even thornier question about the law’s scope. Their view, which would not require an employee to provide information relating to a securities-law violation to the SEC, could afford Dodd-Frank protection to an employee who reports information bearing no relationship whatever to the securities laws. That prospect could be imagined based on the broad array of federal statutes and regulations cross-referenced by clause (iii) of the anti-retaliation provision. E.g., 18 U. S. C. §1513(e) (criminalizing retaliation for “providing to a law enforcement officer any truthful information relating to the commission . . . of any Federal offense” (emphasis added)); see supra, at 5, and n. 2. For example, an employee fired for reporting a coworker’s drug dealing to the Federal Bureau of Investigation might be protected. Brief for Petitioner 38. It would make scant sense, however, to rank an FBI drug-trafficking informant a whistleblower under Dodd-Frank, a law concerned only with encouraging the reporting of “securities law violations.” S. Rep. No. 111–176, at 38 (emphasis added). E Finally, the interpretation we adopt, the Solicitor General adds, would undermine not just clause (iii) of §78u–6(h)(1)(A), but clause (ii) as well. Clause (ii) prohibits retaliation against a whistleblower for “initiating, testifying in, or assisting in any investigation or . . . action of the Commission based upon” information conveyed to the SEC by a whistleblower in accordance with the statute. §78u–6(h)(1)(A)(ii). If the whistleblower definition is applied to §78u–6(h), the Solicitor General states, “an employer could fire an employee for giving . . . testimony [to the SEC] if the employee had not previously reported to the Commission online or through the specified written form”—i.e., the methods currently prescribed by Rule 21F–9 for a whistleblower to provide information to the Commission. Brief for United States as Amicus Curiae 20–21 (citing 17 CFR §240.21F–9(a)(1)–(2)). But the statute expressly delegates authority to the SEC to establish the “manner” in which information may be provided to the Commission by a whistleblower. See §78u–6(a)(6). Nothing in today’s opinion prevents the agency from enumerating additional means of SEC reporting—including through testimony protected by clause (ii). IV For the foregoing reasons, we find the statute’s definition of “whistleblower” clear and conclusive. Because “Congress has directly spoken to the precise question at issue,” Chevron, 467 U. S., at 842, we do not accord deference to the contrary view advanced by the SEC in Rule 21F–2. See 17 CFR §240.21F–2(b)(1); supra, at 6–7. The statute’s unambiguous whistleblower definition, in short, precludes the Commission from more expansively interpreting that term. See Burgess, 553 U. S., at 130. * * * The judgment of the Court of Appeals for the Ninth Circuit is reversed, and the case is remanded for further proceedings consistent with this opinion. It is so ordered. Notes 1 Sarbanes-Oxley also prohibits retaliation against an “employee” who “file[s], . . . testif[ies], participate[s] in, or otherwise assist[s] in a proceeding filed or about to be filed . . . relating to an alleged violation of” the same provisions of federal law addressed in 18 U. S. C. §1514A(a)(1). See §1514A(a)(2). 2 Section 1513(e) provides: “Whoever knowingly, with the intent to retaliate, takes any action harmful to any person, including interference with the lawful employment or livelihood of any person, for providing to a law enforcement officer any truthful information relating to the commission or possible commission of any Federal offense, shall be fined under this title or imprisoned not more than 10 years, or both.” 3 Section 78u–6(h)(1)(A) reads in full: 4 Unlike Dodd-Frank, Sarbanes-Oxley explicitly entitles a prevailing employee to “all relief necessary to make the employee whole,” including “compensation for any special damages sustained as a result of the discrimination.” 18 U. S. C. §1514A(c)(1), (2)(C). 5 In 2015, the SEC issued an interpretive rule reiterating that anti-retaliation protection is not contingent on a whistleblower’s provision of information to the Commission. See 80 Fed. Reg. 47829 (2015). 6 The Solicitor General, unlike Somers, acknowledges that it would not be absurd to apply the “whistleblower” definition to the anti-retaliation provision. Tr. of Oral Arg. 52.

584.US.2017_16-1362

Respondents, current and former service advisors for petitioner Encino Motorcars, LLC, sued petitioner for backpay, alleging that petitioner violated the Fair Labor Standards Act (FLSA) by failing to pay them overtime. Petitioner moved to dismiss, arguing that service advisors are exempt from the FLSA’s overtime-pay requirement under 29 U. S. C. §213(b)(10)(A), which applies to “any salesman, partsman, or mechanic primarily engaged in selling or servicing automobiles, trucks, or farm implements.” The District Court agreed and dismissed the suit. The Court of Appeals for the Ninth Circuit reversed. It found the statute ambiguous and the legislative history inconclusive, and it deferred to a 2011 Department of Labor rule that interpreted “salesman” to exclude service advisors. This Court vacated the Ninth Circuit’s judgment, holding that courts could not defer to the procedurally defective 2011 rule, Encino Motorcars, LLC v. Navarro, 579 U. S. ___, ___–___ (Encino I), but not deciding whether the exemption covers service advisors, id., at ___. On remand, the Ninth Circuit again held that the exemption does not include service advisors. Held: Because service advisors are “salesm[e]n . . . primarily engaged in . . . servicing automobiles,” they are exempt from the FLSA’s overtime-pay requirement. Pp. 5–11. (a) A service advisor is obviously a “salesman.” The ordinary meaning of “salesman” is someone who sells goods or services, and service advisors “sell [customers] services for their vehicles,” Encino I, supra, at ___. P. 6. (b) Service advisors are also “primarily engaged in . . . servicing automobiles.” “Servicing” can mean either “the action of maintaining or repairing a motor vehicle” or “[t]he action of providing a service.” 15 Oxford English Dictionary 39. Service advisors satisfy both definitions because they are integral to the servicing process. They “mee[t] customers; liste[n] to their concerns about their cars; sugges[t] repair and maintenance services; sel[l] new accessories or replacement parts; recor[d] service orders; follo[w] up with customers as the services are performed (for instance, if new problems are discovered); and explai[n] the repair and maintenance work when customers return for their vehicles.” Encino I, supra, at ___. While service advisors do not spend most of their time physically repairing automobiles, neither do partsmen, who the parties agree are “primarily engaged in . . . servicing automobiles.” Pp. 6–7. (c) The Ninth Circuit invoked the distributive canon—matching “salesman” with “selling” and “partsman [and] mechanic” with “[servicing]”—to conclude that the exemption simply does not apply to “salesm[e]n . . . primarily engaged in . . . servicing automobiles.” But the word “or,” which connects all of the exemption’s nouns and gerunds, is “almost always disjunctive.” United States v. Woods, 571 U. S. 31 . Using “or” to join “selling” and “servicing” thus suggests that the exemption covers a salesman primarily engaged in either activity. Statutory context supports this reading. First, the distributive canon has the most force when one-to-one matching is present, but here, the statute would require matching some of three nouns with one of two gerunds. Second, the distributive canon has the most force when an ordinary, disjunctive reading is linguistically impossible. But here, “salesman . . . primarily engaged in . . . servicing automobiles” is an apt description of a service advisor. Third, a narrow distributive phrasing is an unnatural fit here because the entire exemption bespeaks breadth, starting with “any” and using the disjunctive “or” three times. Pp. 7–9. (d) The Ninth Circuit also invoked the principle that exemptions to the FLSA should be construed narrowly. But the Court rejects this principle as a guide to interpreting the FLSA. Because the FLSA gives no textual indication that its exemptions should be construed narrowly, they should be given a fair reading. P. 9. (e) Finally, the Ninth Circuit’s reliance on two extraneous sources to support its interpretation—the 1966–1967 Occupational Outlook Handbook and the FLSA’s legislative history—is unavailing. Pp. 9–11. 845 F. 3d 925, reversed and remanded. Thomas, J., delivered the opinion of the Court, in which Roberts, C. J., and Kennedy, Alito, and Gorsuch, JJ., joined. Ginsburg, J., filed a dissenting opinion, in which Breyer, Sotomayor, and Kagan, JJ., joined.

The Fair Labor Standards Act (FLSA), 52Stat. 1060, as amended, 29 U. S. C. §201 et seq., requires employers to pay overtime compensation to covered employees. The FLSA exempts from the overtime-pay requirement “any salesman, partsman, or mechanic primarily engaged in selling or servicing automobiles” at a covered dealership. §213(b)(10)(A). We granted certiorari to decide whether this exemption applies to service advisors—employees at car dealerships who consult with customers about their servicing needs and sell them servicing solutions. We conclude that service advisors are exempt. I A Enacted in 1938, the FLSA requires employers to pay overtime to covered employees who work more than 40 hours in a week. 29 U. S. C. §207(a). But the FLSA exempts many categories of employees from this requirement. See §213. Employees at car dealerships have long been among those exempted. Congress initially exempted all employees at car dealerships from the overtime-pay requirement. See Fair Labor Standards Amendments of 1961, §9, 75Stat. 73. Congress then narrowed that exemption to cover “any salesman, partsman, or mechanic primarily engaged in selling or servicing automobiles, trailers, trucks, farm implements, or aircraft.” Fair Labor Standards Amendments of 1966, §209, 80Stat. 836. In 1974, Congress enacted the version of the exemption at issue here. It provides that the FLSA’s overtime-pay requirement does not apply to “any salesman, partsman, or mechanic primarily engaged in selling or servicing automobiles, trucks, or farm implements, if he is employed by a nonmanufacturing establishment primarily engaged in the business of selling such vehicles or implements to ultimate purchasers.” §213(b)(10)(A). This language has long been understood to cover service advisors. Although the Department of Labor initially interpreted it to exclude them, 35 Fed. Reg. 5896 (1970) (codified at 29 CFR §779.372(c)(4) (1971)), the federal courts rejected that view, see Brennan v. Deel Motors, Inc., 475 F. 2d 1095 (CA5 1973); Brennan v. North Bros. Ford, Inc., 76 CCH LC ¶33, 247 (ED Mich. 1975), aff’d sub nom. Dunlop v. North Bros. Ford, Inc., 529 F. 2d 524 (CA6 1976) (table). After these decisions, the Department issued an opinion letter in 1978, explaining that service advisors are exempt in most cases. See Dept. of Labor, Wage & Hour Div., Opinion Letter No. 1520 (WH–467) (1978), [1978–1981 Transfer Binder] CCH Wages–Hours Administrative Rulings ¶31,207. From 1978 to 2011, Congress made no changes to the exemption, despite amending §213 nearly a dozen times. The Department also continued to acquiesce in the view that service advisors are exempt. See Dept. of Labor, Wage & Hour Div., Field Operations Handbook, Insert No. 1757, 24L04(k) (Oct. 20, 1987), online at https://perma.cc/5GHD-KCJJ (as last visited Mar. 28, 2018). In 2011, however, the Department reversed course. It issued a rule that interpreted “salesman” to exclude service advisors. 76 Fed. Reg. 18832, 18859 (2011) (codified at 29 CFR §779.372(c)). That regulation prompted this litigation. B Petitioner Encino Motorcars, LLC, is a Mercedes-Benz dealership in California. Respondents are current and former service advisors for petitioner. Service advisors “interact with customers and sell them services for their vehicles.” Encino Motorcars, LLC v. Navarro, 579 U. S. ___, ___ (2016) (Encino I) (slip op., at 2). They “mee[t] customers; liste[n] to their concerns about their cars; sugges[t] repair and maintenance services; sel[l] new accessories or replacement parts; recor[d] service orders; follo[w] up with customers as the services are performed (for instance, if new problems are discovered); and explai[n] the repair and maintenance work when customers return for their vehicles.” Ibid. In 2012, respondents sued petitioner for backpay. Relying on the Department’s 2011 regulation, respondents alleged that petitioner had violated the FLSA by failing to pay them overtime. Petitioner moved to dismiss, arguing that service advisors are exempt under §213(b)(10)(A). The District Court agreed with petitioner and dismissed the complaint, but the Court of Appeals for the Ninth Circuit reversed. Finding the text ambiguous and the legislative history “inconclusive,” the Ninth Circuit deferred to the Department’s 2011 rule under Chevron U. S. A. Inc. v. Natural Resources Defense Council, Inc., 467 U. S. 837 (1984) . Encino, 780 F. 3d 1267, 1275 (2015). We granted certiorari and vacated the Ninth Circuit’s judgment. We explained that courts cannot defer to the 2011 rule because it is procedurally defective. See Encino I, 579 U. S., at ___–___ (slip op., at 8–12). Specifically, the regulation undermined significant reliance interests in the automobile industry by changing the treatment of service advisors without a sufficiently reasoned explanation. Id., at ___ (slip op., at 10). But we did not decide whether, without administrative deference, the exemption covers service advisors. Id., at ___ (slip op., at 12). We remanded that issue for the Ninth Circuit to address in the first instance. Ibid. C On remand, the Ninth Circuit again held that the exemption does not include service advisors. The Court of Appeals agreed that a service advisor is a “ ‘salesman’ ” in a “generic sense,” 845 F. 3d 925, 930 (2017), and is “ ‘primarily engaged in . . . servicing automobiles’ ” in a “general sense,” id., at 931. Nonetheless, it concluded that “Congress did not intend to exempt service advisors.” Id., at 929. The Ninth Circuit began by noting that the Department’s 1966–1967 Occupational Outlook Handbook listed 12 job titles in the table of contents that could be found at a car dealership, including “automobile mechanics,” “automobile parts countermen,” “automobile salesmen,” and “automobile service advisors.” Id., at 930. Because the FLSA exemption listed three of these positions, but not service advisors, the Ninth Circuit concluded that service advisors are not exempt. Ibid. The Ninth Circuit also determined that service advisors are not primarily engaged in “servicing” automobiles, which it defined to mean “only those who are actually occupied in the repair and maintenance of cars.” Id., at 931. And the Ninth Circuit further concluded that the exemption does not cover salesmen who are primarily engaged in servicing. Id., at 933. In reaching this conclusion, the Ninth Circuit invoked the distributive canon. See A. Scalia & B. Garner, Reading Law 214 (2012) (“Distributive phrasing applies each expression to its appropriate referent”). It reasoned that “Congress intended the gerunds—selling and servicing—to be distributed to their appropriate subjects—salesman, partsman, and mechanic. A salesman sells; a partsman services; and a mechanic services.” Id., at 934. Finally, the Court of Appeals noted that its interpretation was supported by the principle that exemptions to the FLSA should be construed narrowly, id., at 935, and the lack of any “mention of service advisors” in the legislative history, id., at 939. We granted certiorari, 582 U. S. ___ (2017), and now reverse. II The FLSA exempts from its overtime-pay requirement “any salesman, partsman, or mechanic primarily engaged in selling or servicing automobiles, trucks, or farm implements, if he is employed by a nonmanufacturing establishment primarily engaged in the business of selling such vehicles or implements to ultimate purchasers.” §213(b)(10)(A). The parties agree that petitioner is a “nonmanufacturing establishment primarily engaged in the business of selling [automobiles] to ultimate purchasers.” The parties also agree that a service advisor is not a “partsman” or “mechanic,” and that a service advisor is not “primarily engaged . . . in selling automobiles.” The question, then, is whether service advisors are “salesm[e]n . . . primarily engaged in . . . servicing automobiles.” We conclude that they are. Under the best reading of the text, service advisors are “salesm[e]n,” and they are “primarily engaged in . . . servicing automobiles.” The distributive canon, the practice of construing FLSA exemptions narrowly, and the legislative history do not persuade us otherwise. A A service advisor is obviously a “salesman.” The term “salesman” is not defined in the statute, so “we give the term its ordinary meaning.” Taniguchi v. Kan Pacific Saipan, Ltd., 566 U. S. 560, 566 (2012) . The ordinary meaning of “salesman” is someone who sells goods or services. See 14 Oxford English Dictionary 391 (2d ed. 1989) (“[a] man whose business it is to sell goods or conduct sales”); Random House Dictionary of the English Language 1262 (1966) (“a man who sells goods, services, etc.”). Service advisors do precisely that. As this Court previously explained, service advisors “sell [customers] services for their vehicles.” Encino I, 579 U. S., at ___ (slip op., at 2). B Service advisors are also “primarily engaged in . . . servicing automobiles.” §213(b)(10)(A). The word “servicing” in this context can mean either “the action of maintaining or repairing a motor vehicle” or “[t]he action of providing a service.” 15 Oxford English Dictionary, at 39; see also Random House Dictionary of the English Language, at 1304 (“to make fit for use; repair; restore to condition for service”). Service advisors satisfy both definitions. Service advisors are integral to the servicing process. They “mee[t] customers; liste[n] to their concerns about their cars; sugges[t] repair and maintenance services; sel[l] new accessories or replacement parts; recor[d] service orders; follo[w] up with customers as the services are performed (for instance, if new problems are discovered); and explai[n] the repair and maintenance work when customers return for their vehicles.” Encino I, supra, at ___ (slip op., at 2). If you ask the average customer who services his car, the primary, and perhaps only, person he is likely to identify is his service advisor. True, service advisors do not spend most of their time physically repairing automobiles. But the statutory language is not so constrained. All agree that partsmen, for example, are “primarily engaged in . . . servicing automobiles.” Brief for Petitioner 40; Brief for Respondents 41–44. But partsmen, like service advisors, do not spend most of their time under the hood. Instead, they “obtain the vehicle parts . . . and provide those parts to the mechanics.” Encino I, supra, at ___ (slip op., at 2); see also 1 Dept. of Labor, Dictionary of Occupational Titles 33 (3d ed. 1965) (defining “partsman” as someone who “[p]urchases, stores, and issues spare parts for automotive and industrial equipment”). In other words, the phrase “primarily engaged in . . . servicing automobiles” must include some individuals who do not physically repair automobiles themselves but who are integrally involved in the servicing process. That description applies to partsmen and service advisors alike. C The Ninth Circuit concluded that service advisors are not covered because the exemption simply does not apply to “salesm[e]n . . . primarily engaged in . . . servicing automobiles.” The Ninth Circuit invoked the distributive canon to reach this conclusion. Using that canon, it matched “salesman” with “selling” and “partsma[n] [and] mechanic” with “servicing.” We reject this reasoning. The text of the exemption covers “any salesman, partsman, or mechanic primarily engaged in selling or servicing automobiles, trucks, or farm implements.” §213(b)(10)(A). The exemption uses the word “or” to connect all of its nouns and gerunds, and “or” is “almost always disjunctive.” United States v. Woods, 571 U. S. 31, 45 (2013) . Thus, the use of “or” to join “selling” and “servicing” suggests that the exemption covers a salesman primarily engaged in either activity. Unsurprisingly, statutory context can overcome the ordinary, disjunctive meaning of “or.” The distributive canon, for example, recognizes that sometimes “[w]here a sentence contains several antecedents and several consequents,” courts should “read them distributively and apply the words to the subjects which, by context, they seem most properly to relate.” 2A N. Singer & S. Singer, Sutherland Statutes and Statutory Construction §47:26, p. 448 (rev. 7th ed. 2014). But here, context favors the ordinary disjunctive meaning of “or” for at least three reasons. First, the distributive canon has the most force when the statute allows for one-to-one matching. But here, the distributive canon would mix and match some of three nouns—“salesman, partsman, or mechanic”—with one of two gerunds—“selling or servicing.” §213(b)(10)(A). We doubt that a legislative drafter would leave it to the reader to figure out the precise combinations. Second, the distributive canon has the most force when an ordinary, disjunctive reading is linguistically impossible. Cf., e.g., Huidekoper’s Lessee v. Douglass, 3 Cranch 1, 67 (1805) (Marshall, C. J.) (applying the distributive canon when a purely disjunctive reading “would involve a contradiction in terms”). But as explained above, the phrase “salesman . . . primarily engaged in . . . servicing automobiles” not only makes sense; it is an apt description of a service advisor. Third, a narrow distributive phrasing is an unnatural fit here because the entire exemption bespeaks breadth. It begins with the word “any.” See Ali v. Federal Bureau of Prisons, 552 U. S. 214, 219 (2008) (noting the “expansive meaning” of “any”). And it uses the disjunctive word “or” three times. In fact, all agree that the third list in the exemption—“automobiles, trucks, or farm implements”—modifies every other noun and gerund. But it would be odd to read the exemption as starting with a distributive phrasing and then, halfway through and without warning, switching to a disjunctive phrasing—all the while using the same word (“or”) to signal both meanings. See Brown v. Gardner, 513 U. S. 115, 118 (1994) (noting the “vigorous” presumption that, “when a term is repeated within a given sentence,” it “is used to mean the same thing”). The more natural reading is that the exemption covers any combination of its nouns, gerunds, and objects. D The Ninth Circuit also invoked the principle that exemptions to the FLSA should be construed narrowly. 845 F. 3d, at 935–936. We reject this principle as a useful guidepost for interpreting the FLSA. Because the FLSA gives no “textual indication” that its exemptions should be construed narrowly, “there is no reason to give [them] anything other than a fair (rather than a ‘narrow’) interpretation.” Scalia, Reading Law, at 363. The narrow-construction principle relies on the flawed premise that the FLSA “ ‘pursues’ ” its remedial purpose “ ‘at all costs.’ ” American Express Co. v. Italian Colors Restaurant, 570 U. S. 228, 234 (2013) (quoting Rodriguez v. United States, 480 U. S. 522 –526 (1987) (per curiam)); see also Henson v. Santander Consumer USA Inc., 582 U. S. ___, ___ (2017) (slip op., at 9) (“[I]t is quite mistaken to assume . . . that whatever might appear to further the statute’s primary objective must be the law” (internal quotation marks and alterations omitted)). But the FLSA has over two dozen exemptions in §213(b) alone, including the one at issue here. Those exemptions are as much a part of the FLSA’s purpose as the overtime-pay requirement. See id., at ___ (slip op., at 9) (“Legislation is, after all, the art of compromise, the limitations expressed in statutory terms often the price of passage”). We thus have no license to give the exemption anything but a fair reading. E Finally, the Ninth Circuit relied on two extraneous sources to support its interpretation: the Department’s 1966–1967 Occupational Outlook Handbook and the FLSA’s legislative history. We find neither persuasive. 1 The Ninth Circuit first relied on the Department’s 1966–1967 Occupational Outlook Handbook. It identified 12 jobs from the Handbook’s table of contents that it thought could be found at automobile dealerships. See 845 F. 3d, at 930. The Ninth Circuit then stressed that the exemption aligns with three of those job titles—“[a]utomobile mechanics,” “[a]utomobile parts countermen,” and “[a]utomobile salesmen”—but not “[a]utomobile service advisors.” Ibid. The Ninth Circuit cited nothing, however, suggesting that the exemption was meant to align with the job titles listed in the Handbook. To the contrary, the exemption applies to “any salesman . . . primarily engaged in selling or servicing automobiles.” It is not limited, like the term in the Handbook, to “automobile salesmen.” And the ordinary meaning of “salesman” plainly includes service advisors. 2 The Ninth Circuit also relied on legislative history to support its interpretation. See id., at 936–939. Specifi- cally, it noted that the legislative history discusses “automobile salesmen, partsmen, and mechanics” but never discusses service advisors. Id., at 939. Although the Ninth Circuit had previously found that same legislative history “inconclusive,” Encino, 780 F. 3d, at 1275, on remand it was “firmly persuaded” that the legislative history demonstrated Congress’ desire to exclude service advisors, 845 F. 3d, at 939. The Ninth Circuit was right the first time. As we have explained, the best reading of the statute is that service advisors are exempt. Even for those Members of this Court who consider legislative history, silence in the legislative history, “no matter how ‘clanging,’ ” cannot defeat the better reading of the text and statutory context. Sedima, S. P. R. L. v. Imrex Co., 473 U. S. 479 , n. 13 (1985). If the text is clear, it needs no repetition in the legislative history; and if the text is ambiguous, silence in the legislative history cannot lend any clarity. See Avco Corp. v. Department of Justice, 884 F. 2d 621, 625 (CADC 1989). Even if Congress did not foresee all of the applications of the statute, that is no reason not to give the statutory text a fair reading. See Union Bank v. Wolas, 502 U. S. 151, 158 (1991) . * * * In sum, we conclude that service advisors are exempt from the overtime-pay requirement of the FLSA because they are “salesm[e]n . . . primarily engaged in . . . servicing automobiles.” §213(b)(10)(A). Accordingly, we reverse the judgment of the Court of Appeals and remand the case for further proceedings consistent with this opinion. It is so ordered.

584.US.2017_16-285

In each of these cases, an employer and employee entered into a contract providing for individualized arbitration proceedings to resolve employment disputes between the parties. Each employee nonetheless sought to litigate Fair Labor Standards Act and related state law claims through class or collective actions in federal court. Although the Federal Arbitration Act generally requires courts to enforce arbitration agreements as written, the employees argued that its “saving clause” removes this obligation if an arbitration agreement violates some other federal law and that, by requiring individualized proceedings, the agreements here violated the National Labor Relations Act. The employers countered that the Arbitration Act protects agreements requiring arbitration from judicial interference and that neither the saving clause nor the NLRA demands a different conclusion. Until recently, courts as well as the National Labor Relations Board’s general counsel agreed that such arbitration agreements are enforceable. In 2012, however, the Board ruled that the NLRA effectively nullifies the Arbitration Act in cases like these, and since then other courts have either agreed with or deferred to the Board’s position. Held: Congress has instructed in the Arbitration Act that arbitration agreements providing for individualized proceedings must be enforced, and neither the Arbitration Act’s saving clause nor the NLRA suggests otherwise. Pp. 5–25. (a) The Arbitration Act requires courts to enforce agreements to arbitrate, including the terms of arbitration the parties select. See 9 U. S. C. §§2, 3, 4. These emphatic directions would seem to resolve any argument here. The Act’s saving clause—which allows courts to refuse to enforce arbitration agreements “upon such grounds as exist at law or in equity for the revocation of any contract,” §2—recognizes only “ ‘generally applicable contract defenses, such as fraud, duress, or unconscionability,’ ” AT&T Mobility LLC v. Concepcion, 563 U. S. 333, 339, not defenses targeting arbitration either by name or by more subtle methods, such as by “interfer[ing] with fundamental attributes of arbitration,” id., at 344. By challenging the agreements precisely because they require individualized arbitration instead of class or collective proceedings, the employees seek to interfere with one of these fundamental attributes. Pp. 5–9. (b) The employees also mistakenly claim that, even if the Arbitration Act normally requires enforcement of arbitration agreements like theirs, the NLRA overrides that guidance and renders their agreements unlawful yet. When confronted with two Acts allegedly touching on the same topic, this Court must strive “to give effect to both.” Morton v. Mancari, 417 U. S. 535, 551. To prevail, the employees must show a “ ‘clear and manifest’ ” congressional intention to displace one Act with another. Ibid. There is a “stron[g] presum[ption]” that disfavors repeals by implication and that “Congress will specifically address” preexisting law before suspending the law’s normal operations in a later statute. United States v. Fausto, 484 U. S. 439, 452, 453. The employees ask the Court to infer that class and collective actions are “concerted activities” protected by §7 of the NLRA, which guarantees employees “the right to self-organization, to form, join, or assist labor organizations, to bargain collectively . . . , and to engage in other concerted activities for the purpose of collective bargaining or other mutual aid or protection,” 29 U. S. C. §157. But §7 focuses on the right to organize unions and bargain collectively. It does not mention class or collective action procedures or even hint at a clear and manifest wish to displace the Arbitration Act. It is unlikely that Congress wished to confer a right to class or collective actions in §7, since those procedures were hardly known when the NLRA was adopted in 1935. Because the catchall term “other concerted activities for the purpose of . . . other mutual aid or protection” appears at the end of a detailed list of activities, it should be understood to protect the same kind of things, i.e., things employees do for themselves in the course of exercising their right to free association in the workplace. The NLRA’s structure points to the same conclusion. After speaking of various “concerted activities” in §7, the statute establishes a detailed regulatory regime applicable to each item on the list, but gives no hint about what rules should govern the adjudication of class or collective actions in court or arbitration. Nor is it at all obvious what rules should govern on such essential issues as opt-out and opt-in procedures, notice to class members, and class certification standards. Telling too is the fact that Congress has shown that it knows exactly how to specify certain dispute resolution procedures, cf., e.g., 29 U. S. C. §§216(b), 626, or to override the Arbitration Act, see, e.g., 15 U. S. C. §1226(a)(2), but Congress has done nothing like that in the NLRA. The employees suggest that the NLRA does not discuss class and collective action procedures because it means to confer a right to use existing procedures provided by statute or rule, but the NLRA does not say even that much. And if employees do take existing rules as they find them, they must take them subject to those rules’ inherent limitations, including the principle that parties may depart from them in favor of individualized arbitration. In another contextual clue, the employees’ underlying causes of action arise not under the NLRA but under the Fair Labor Standards Act, which permits the sort of collective action the employees wish to pursue here. Yet they do not suggest that the FLSA displaces the Arbitration Act, presumably because the Court has held that an identical collective action scheme does not prohibit individualized arbitration proceedings, see Gilmer v. Interstate/Johnson Lane Corp., 500 U. S. 20, 32. The employees’ theory also runs afoul of the rule that Congress “does not alter the fundamental details of a regulatory scheme in vague terms or ancillary provisions,” Whitman v. American Trucking Assns., Inc., 531 U. S. 457, 468, as it would allow a catchall term in the NLRA to dictate the particulars of dispute resolution procedures in Article III courts or arbitration proceedings—matters that are usually left to, e.g., the Federal Rules of Civil Procedure, the Arbitration Act, and the FLSA. Nor does the employees’ invocation of the Norris-LaGuardia Act, a predecessor of the NLRA, help their argument. That statute declares unenforceable contracts in conflict with its policy of protecting workers’ “concerted activities for the purpose of collective bargaining or other mutual aid or protection,” 29 U. S. C. §102, and just as under the NLRA, that policy does not conflict with Congress’s directions favoring arbitration. Precedent confirms the Court’s reading. The Court has rejected many efforts to manufacture conflicts between the Arbitration Act and other federal statutes, see, e.g. American Express Co. v. Italian Colors Restaurant, 570 U. S. 228; and its §7 cases have generally involved efforts related to organizing and collective bargaining in the workplace, not the treatment of class or collective action procedures in court or arbitration, see, e.g., NLRB v. Washington Aluminum Co., 370 U. S. 9. Finally, the employees cannot expect deference under Chevron U. S. A. Inc. v. Natural Resources Defense Council, Inc., 467 U. S. 837, because Chevron’s essential premises are missing. The Board sought not to interpret just the NLRA, “which it administers,” id., at 842, but to interpret that statute in a way that limits the work of the Arbitration Act, which the agency does not administer. The Board and the Solicitor General also dispute the NLRA’s meaning, articulating no single position on which the Executive Branch might be held “accountable to the people.” Id., at 865. And after “employing traditional tools of statutory construction,” id., at 843, n. 9, including the canon against reading conflicts into statutes, there is no unresolved ambiguity for the Board to address. Pp. 9–21. No. 16–285, 823 F. 3d 1147, and No. 16–300, 834 F. 3d 975, reversed and remanded; No. 16–307, 808 F. 3d 1013, affirmed. Gorsuch, J., delivered the opinion of the Court, in which Roberts, C. J., and Kennedy, Thomas, and Alito, JJ., joined. Thomas, J., filed a concurring opinion. Ginsburg, J., filed a dissenting opinion, in which Breyer, Sotomayor, and Kagan, JJ., joined. Notes 1 Together with No. 16–300, Ernst & Young LLP et al. v. Morris et al., on certiorari to the United States Court of Appeals for the Ninth Circuit, and No. 16–307, National Labor Relations Board v. Murphy Oil USA, Inc., et al., on certiorari to the United States Court of Appeals for the Fifth Circuit.

Should employees and employers be allowed to agree that any disputes between them will be resolved through one-on-one arbitration? Or should employees always be permitted to bring their claims in class or collective actions, no matter what they agreed with their employers? As a matter of policy these questions are surely debat- able. But as a matter of law the answer is clear. In the Federal Arbitration Act, Congress has instructed federal courts to enforce arbitration agreements according to their terms—including terms providing for individualized proceedings. Nor can we agree with the employees’ suggestion that the National Labor Relations Act (NLRA) offers a conflicting command. It is this Court’s duty to interpret Congress’s statutes as a harmonious whole rather than at war with one another. And abiding that duty here leads to an unmistakable conclusion. The NLRA secures to employees rights to organize unions and bargain collectively, but it says nothing about how judges and arbitrators must try legal disputes that leave the workplace and enter the courtroom or arbitral forum. This Court has never read a right to class actions into the NLRA—and for three quarters of a century neither did the National Labor Relations Board. Far from conflicting, the Arbitration Act and the NLRA have long enjoyed separate spheres of influence and neither permits this Court to declare the parties’ agreements unlawful. I The three cases before us differ in detail but not in substance. Take Ernst & Young LLP v. Morris. There Ernst & Young and one of its junior accountants, Stephen Morris, entered into an agreement providing that they would arbitrate any disputes that might arise between them. The agreement stated that the employee could choose the arbitration provider and that the arbitrator could “grant any relief that could be granted by . . . a court” in the relevant jurisdiction. App. in No. 16–300, p. 43. The agreement also specified individualized arbitration, with claims “pertaining to different [e]mployees [to] be heard in separate proceedings.” Id., at 44. After his employment ended, and despite having agreed to arbitrate claims against the firm, Mr. Morris sued Ernst & Young in federal court. He alleged that the firm had misclassified its junior accountants as professional employees and violated the federal Fair Labor Standards Act (FLSA) and California law by paying them salaries without overtime pay. Although the arbitration agreement provided for individualized proceedings, Mr. Morris sought to litigate the federal claim on behalf of a nationwide class under the FLSA’s collective action provision, 29 U. S. C. §216(b). He sought to pursue the state law claim as a class action under Federal Rule of Civil Procedure 23. Ernst & Young replied with a motion to compel arbitration. The district court granted the request, but the Ninth Circuit reversed this judgment. 834 F. 3d 975 (2016). The Ninth Circuit recognized that the Arbitration Act gener- ally requires courts to enforce arbitration agreements as written. But the court reasoned that the statute’s “saving clause,” see 9 U. S. C. §2, removes this obligation if an arbitration agreement violates some other federal law. And the court concluded that an agreement requiring individualized arbitration proceedings violates the NLRA by barring employees from engaging in the “concerted activit[y],” 29 U. S. C. §157, of pursuing claims as a class or collective action. Judge Ikuta dissented. In her view, the Arbitration Act protected the arbitration agreement from judicial interference and nothing in the Act’s saving clause suggested otherwise. Neither, she concluded, did the NLRA demand a different result. Rather, that statute focuses on protecting unionization and collective bargaining in the workplace, not on guaranteeing class or collective action procedures in disputes before judges or arbitrators. Although the Arbitration Act and the NLRA have long coexisted—they date from 1925 and 1935, respectively—the suggestion they might conflict is something quite new. Until a couple of years ago, courts more or less agreed that arbitration agreements like those before us must be enforced according to their terms. See, e.g., Owen v. Bristol Care, Inc., 702 F. 3d 1050 (CA8 2013); Sutherland v. Ernst & Young LLP, 726 F. 3d 290 (CA2 2013); D. R. Horton, Inc. v. NLRB, 737 F. 3d 344 (CA5 2013); Iskanian v. CLS Transp. Los Angeles, LLC, 59 Cal. 4th 348, 327 P. 3d 129 (2014); Tallman v. Eighth Jud. Dist. Court, 131 Nev. 71, 359 P. 3d 113 (2015); 808 F. 3d 1013 (CA5 2015) (case below in No. 16–307). The National Labor Relations Board’s general counsel expressed much the same view in 2010. Remarking that employees and employers “can benefit from the relative simplicity and informality of resolving claims before arbitrators,” the general counsel opined that the validity of such agreements “does not involve consideration of the policies of the National Labor Relations Act.” Memorandum GC 10–06, pp. 2, 5 (June 16, 2010). But recently things have shifted. In 2012, the Board—for the first time in the 77 years since the NLRA’s adoption—asserted that the NLRA effectively nullifies the Arbitration Act in cases like ours. D. R. Horton, Inc., 357 N. L. R. B. 2277. Initially, this agency decision received a cool reception in court. See D. R. Horton, 737 F. 3d, at 355–362. In the last two years, though, some circuits have either agreed with the Board’s conclusion or thought themselves obliged to defer to it under Chevron U. S. A. Inc. v. Natural Resources Defense Council, Inc., 467 U. S. 837 (1984). See 823 F. 3d 1147 (CA7 2016) (case below in No. 16–285); 834 F. 3d 975 (case below in No. 16–300); NLRB v. Alternative Entertainment, Inc., 858 F. 3d 393 (CA6 2017). More recently still, the disagreement has grown as the Executive has disavowed the Board’s (most recent) position, and the Solicitor General and the Board have offered us battling briefs about the law’s meaning. We granted certiorari to clear the confusion. 580 U. S. ___ (2017). II We begin with the Arbitration Act and the question of its saving clause. Congress adopted the Arbitration Act in 1925 in response to a perception that courts were unduly hostile to arbitration. No doubt there was much to that perception. Before 1925, English and American common law courts routinely refused to enforce agreements to arbitrate disputes. Scherk v. Alberto-Culver Co., 417 U. S. 506, 510, n. 4 (1974). But in Congress’s judgment arbitration had more to offer than courts recognized—not least the promise of quicker, more informal, and often cheaper resolutions for everyone involved. Id., at 511. So Congress directed courts to abandon their hostility and instead treat arbitration agreements as “valid, irrevocable, and enforceable.” 9 U. S. C. §2. The Act, this Court has said, establishes “a liberal federal policy favoring arbitration agreements.” Moses H. Cone Memorial Hospital v. Mercury Constr. Corp., 460 U. S. 1, 24 (1983) (citing Prima Paint Corp. v. Flood & Conklin Mfg. Co., 388 U. S. 395 (1967)); see id., at 404 (discussing “the plain meaning of the statute” and “the unmistakably clear congressional purpose that the arbitration procedure, when selected by the parties to a contract, be speedy and not subject to delay and obstruction in the courts”). Not only did Congress require courts to respect and enforce agreements to arbitrate; it also specifically directed them to respect and enforce the parties’ chosen arbitration procedures. See §3 (providing for a stay of litigation pending arbitration “in accordance with the terms of the agreement”); §4 (providing for “an order directing that . . . arbitration proceed in the manner provided for in such agreement”). Indeed, we have often observed that the Arbitration Act requires courts “rigorously” to “enforce arbitration agreements according to their terms, including terms that specify with whom the parties choose to arbitrate their disputes and the rules under which that arbitration will be conducted.” American Express Co. v. Italian Colors Restaurant, 570 U. S. 228, 233 (2013) (some emphasis added; citations, internal quotation marks, and brackets omitted). On first blush, these emphatic directions would seem to resolve any argument under the Arbitration Act. The parties before us contracted for arbitration. They pro- ceeded to specify the rules that would govern their arbitrations, indicating their intention to use individualized rather than class or collective action procedures. And this much the Arbitration Act seems to protect pretty absolutely. See AT&T Mobility LLC v. Concepcion, 563 U. S. 333 (2011); Italian Colors, supra; DIRECTV, Inc. v. Imburgia, 577 U. S. ___ (2015). You might wonder if the balance Congress struck in 1925 between arbitration and litigation should be revisited in light of more contemporary developments. You might even ask if the Act was good policy when enacted. But all the same you might find it difficult to see how to avoid the statute’s application. Still, the employees suggest the Arbitration Act’s saving clause creates an exception for cases like theirs. By its terms, the saving clause allows courts to refuse to enforce arbitration agreements “upon such grounds as exist at law or in equity for the revocation of any contract.” §2. That provision applies here, the employees tell us, because the NLRA renders their particular class and collective action waivers illegal. In their view, illegality under the NLRA is a “ground” that “exists at law . . . for the revocation” of their arbitration agreements, at least to the extent those agreements prohibit class or collective action proceedings. The problem with this line of argument is fundamental. Put to the side the question whether the saving clause was designed to save not only state law defenses but also defenses allegedly arising from federal statutes. See 834 F. 3d, at 991–992, 997 (Ikuta, J., dissenting). Put to the side the question of what it takes to qualify as a ground for “revocation” of a contract. See Concepcion, supra, at 352–355 (Thomas, J., concurring); post, at 1–2 (Thomas, J., concurring). Put to the side for the moment, too, even the question whether the NLRA actually renders class and collective action waivers illegal. Assuming (but not granting) the employees could satisfactorily answer all those questions, the saving clause still can’t save their cause. It can’t because the saving clause recognizes only defenses that apply to “any” contract. In this way the clause establishes a sort of “equal-treatment” rule for arbitration contracts. Kindred Nursing Centers L. P. v. Clark, 581 U. S. ___, ___ (2017) (slip op., at 4). The clause “permits agreements to arbitrate to be invalidated by ‘generally applicable contract defenses, such as fraud, duress, or unconscionability.’ ” Concepcion, 563 U. S., at 339. At the same time, the clause offers no refuge for “defenses that apply only to arbitration or that derive their meaning from the fact that an agreement to arbitrate is at issue.” Ibid. Under our precedent, this means the saving clause does not save defenses that target arbitration either by name or by more subtle methods, such as by “interfer[ing] with fundamental attributes of arbitration.” Id., at 344; see Kindred Nursing, supra, at ___ (slip op., at 5). This is where the employees’ argument stumbles. They don’t suggest that their arbitration agreements were extracted, say, by an act of fraud or duress or in some other unconscionable way that would render any contract unenforceable. Instead, they object to their agreements precisely because they require individualized arbitration proceedings instead of class or collective ones. And by attacking (only) the individualized nature of the arbitration proceedings, the employees’ argument seeks to interfere with one of arbitration’s fundamental attributes. We know this much because of Concepcion. There this Court faced a state law defense that prohibited as unconscionable class action waivers in consumer contracts. The Court readily acknowledged that the defense formally applied in both the litigation and the arbitration context. 563 U. S., at 338, 341. But, the Court held, the defense failed to qualify for protection under the saving clause because it interfered with a fundamental attribute of arbitration all the same. It did so by effectively permitting any party in arbitration to demand classwide proceedings despite the traditionally individualized and informal nature of arbitration. This “fundamental” change to the traditional arbitration process, the Court said, would “sacrific[e] the principal advantage of arbitration—its informality—and mak[e] the process slower, more costly, and more likely to generate procedural morass than final judgment.” Id., at 347, 348. Not least, Concepcion noted, arbitrators would have to decide whether the named class representatives are sufficiently representative and typical of the class; what kind of notice, opportunity to be heard, and right to opt out absent class members should enjoy; and how discovery should be altered in light of the classwide nature of the proceedings. Ibid. All of which would take much time and effort, and introduce new risks and costs for both sides. Ibid. In the Court’s judgment, the virtues Congress originally saw in arbitration, its speed and simplicity and inexpensiveness, would be shorn away and arbitration would wind up looking like the litigation it was meant to displace. Of course, Concepcion has its limits. The Court recognized that parties remain free to alter arbitration procedures to suit their tastes, and in recent years some parties have sometimes chosen to arbitrate on a classwide basis. Id., at 351. But Concepcion’s essential insight remains: courts may not allow a contract defense to reshape traditional individualized arbitration by mandating classwide arbitration procedures without the parties’ consent. Id., at 344–351; see also Stolt-Nielsen S. A. v. AnimalFeeds Int’l Corp., 559 U. S. 662, 684–687 (2010). Just as judicial antagonism toward arbitration before the Arbitration Act’s enactment “manifested itself in a great variety of devices and formulas declaring arbitration against public policy,” Concepcion teaches that we must be alert to new devices and formulas that would achieve much the same result today. 563 U. S., at 342 (internal quotation marks omitted). And a rule seeking to declare individualized arbitration proceedings off limits is, the Court held, just such a device. The employees’ efforts to distinguish Concepcion fall short. They note that their putative NLRA defense would render an agreement “illegal” as a matter of federal statutory law rather than “unconscionable” as a matter of state common law. But we don’t see how that distinction makes any difference in light of Concepion’s rationale and rule. Illegality, like unconscionability, may be a traditional, generally applicable contract defense in many cases, including arbitration cases. But an argument that a contract is unenforceable just because it requires bilateral arbitration is a different creature. A defense of that kind, Concepcion tells us, is one that impermissibly disfavors arbitration whether it sounds in illegality or unconscion- ability. The law of precedent teaches that like cases should generally be treated alike, and appropriate respect for that principle means the Arbitration Act’s saving clause can no more save the defense at issue in these cases than it did the defense at issue in Concepcion. At the end of our encounter with the Arbitration Act, then, it appears just as it did at the beginning: a congressional command requiring us to enforce, not override, the terms of the arbitration agreements before us. III But that’s not the end of it. Even if the Arbitration Act normally requires us to enforce arbitration agreements like theirs, the employees reply that the NLRA overrides that guidance in these cases and commands us to hold their agreements unlawful yet. This argument faces a stout uphill climb. When confronted with two Acts of Congress allegedly touching on the same topic, this Court is not at “liberty to pick and choose among congressional enactments” and must instead strive “ ‘to give effect to both.’ ” Morton v. Mancari, 417 U. S. 535, 551 (1974). A party seeking to suggest that two statutes cannot be harmonized, and that one displaces the other, bears the heavy burden of showing “ ‘a clearly expressed congressional intention’ ” that such a result should follow. Vimar Seguros y Reaseguros, S. A. v. M/V Sky Reefer, 515 U. S. 528, 533 (1995). The intention must be “ ‘clear and manifest.’ ” Morton, supra, at 551. And in approaching a claimed conflict, we come armed with the “stron[g] presum[ption]” that repeals by implication are “disfavored” and that “Congress will specifically address” preexisting law when it wishes to suspend its normal operations in a later statute. United States v. Fausto, 484 U. S. 439, 452, 453 (1988). These rules exist for good reasons. Respect for Congress as drafter counsels against too easily finding irreconcilable conflicts in its work. More than that, respect for the separation of powers counsels restraint. Allowing judges to pick and choose between statutes risks transforming them from expounders of what the law is into policymakers choosing what the law should be. Our rules aiming for harmony over conflict in statutory interpretation grow from an appreciation that it’s the job of Congress by legislation, not this Court by supposition, both to write the laws and to repeal them. Seeking to demonstrate an irreconcilable statutory conflict even in light of these demanding standards, the employees point to Section 7 of the NLRA. That provision guarantees workers “the right to self-organization, to form, join, or assist labor organizations, to bargain collectively through representatives of their own choosing, and to engage in other concerted activities for the purpose of collective bargaining or other mutual aid or protection.” 29 U. S. C. §157. From this language, the employees ask us to infer a clear and manifest congressional command to displace the Arbitration Act and outlaw agreements like theirs. But that much inference is more than this Court may make. Section 7 focuses on the right to organize unions and bargain collectively. It may permit unions to bargain to prohibit arbitration. Cf. 14 Penn Plaza LLC v. Pyett, 556 U. S. 247, 256–260 (2009). But it does not express approval or disapproval of arbitration. It does not mention class or collective action procedures. It does not even hint at a wish to displace the Arbitration Act—let alone accomplish that much clearly and manifestly, as our precedents demand. Neither should any of this come as a surprise. The notion that Section 7 confers a right to class or collective actions seems pretty unlikely when you recall that procedures like that were hardly known when the NLRA was adopted in 1935. Federal Rule of Civil Procedure 23 didn’t create the modern class action until 1966; class arbitration didn’t emerge until later still; and even the Fair Labor Standards Act’s collective action provision postdated Section 7 by years. See Rule 23–Class Actions, 28 U. S. C. App., p. 1258 (1964 ed., Supp. II); 52Stat. 1069; Concepcion, 563 U. S., at 349; see also Califano v. Yamasaki, 442 U. S. 682, 700–701 (1979) (noting that the “usual rule” then was litigation “conducted by and on behalf of individual named parties only”). And while some forms of group litigation existed even in 1935, see 823 F. 3d, at 1154, Section 7’s failure to mention them only reinforces that the statute doesn’t speak to such procedures. A close look at the employees’ best evidence of a potential conflict turns out to reveal no conflict at all. The employees direct our attention to the term “other con- certed activities for the purpose of . . . other mutual aid or protection.” This catchall term, they say, can be read to include class and collective legal actions. But the term appears at the end of a detailed list of activities speaking of “self-organization,” “form[ing], join[ing], or assist[ing] labor organizations,” and “bargain[ing] collectively.” 29 U. S. C. §157. And where, as here, a more general term follows more specific terms in a list, the general term is usually understood to “ ‘embrace only objects similar in nature to those objects enumerated by the preceding specific words.’ ” Circuit City Stores, Inc. v. Adams, 532 U. S. 105, 115 (2001) (discussing ejusdem generis canon); National Assn. of Mfrs. v. Department of Defense, 583 U. S. ___, ___ (2018) (slip op., at 10). All of which suggests that the term “other concerted activities” should, like the terms that precede it, serve to protect things employees “just do” for themselves in the course of exercising their right to free association in the workplace, rather than “the highly regulated, courtroom-bound ‘activities’ of class and joint litigation.” Alternative Entertainment, 858 F. 3d, at 414–415 (Sutton, J., concurring in part and dissenting in part) (emphasis deleted). None of the preceding and more specific terms speaks to the procedures judges or arbitrators must apply in disputes that leave the workplace and enter the courtroom or arbitral forum, and there is no textually sound reason to suppose the final catchall term should bear such a radically different object than all its predecessors. The NLRA’s broader structure underscores the point. After speaking of various “concerted activities” in Section 7, Congress proceeded to establish a regulatory regime applicable to each of them. The NLRA provides rules for the recognition of exclusive bargaining representatives, 29 U. S. C. §159, explains employees’ and employers’ obligation to bargain collectively, §158(d), and conscribes certain labor organization practices, §§158(a)(3), (b). The NLRA also touches on other concerted activities closely related to organization and collective bargaining, such as picketing, §158(b)(7), and strikes, §163. It even sets rules for adjudicatory proceedings under the NLRA itself. §§160, 161. Many of these provisions were part of the original NLRA in 1935, see 49Stat. 449, while others were added later. But missing entirely from this careful regime is any hint about what rules should govern the adjudication of class or collective actions in court or arbitration. Without some comparably specific guidance, it’s not at all obvious what procedures Section 7 might protect. Would opt-out class action procedures suffice? Or would opt-in procedures be necessary? What notice might be owed to absent class members? What standards would govern class certification? Should the same rules always apply or should they vary based on the nature of the suit? Nothing in the NLRA even whispers to us on any of these essential questions. And it is hard to fathom why Congress would take such care to regulate all the other matters mentioned in Section 7 yet remain mute about this matter alone—unless, of course, Section 7 doesn’t speak to class and collective action procedures in the first place. Telling, too, is the fact that when Congress wants to mandate particular dispute resolution procedures it knows exactly how to do so. Congress has spoken often and clearly to the procedures for resolving “actions,” “claims,” “charges,” and “cases” in statute after statute. E.g., 29 U. S. C. §§216(b), 626; 42 U. S. C. §§2000e–5(b), (f)(3)–(5). Congress has likewise shown that it knows how to override the Arbitration Act when it wishes—by explaining, for example, that, “[n]otwithstanding any other provision of law, . . . arbitration may be used . . . only if” certain conditions are met, 15 U. S. C. §1226(a)(2); or that “[n]o predispute arbitration agreement shall be valid or enforceable” in other circumstances, 7 U. S. C. §26(n)(2); 12 U. S. C. §5567(d)(2); or that requiring a party to arbitrate is “unlawful” in other circumstances yet, 10 U. S. C. §987(e)(3). The fact that we have nothing like that here is further evidence that Section 7 does nothing to address the question of class and collective actions. In response, the employees offer this slight reply. They suggest that the NLRA doesn’t discuss any particular class and collective action procedures because it merely confers a right to use existing procedures provided by statute or rule, “on the same terms as [they are] made available to everyone else.” Brief for Respondent in No. 16–285, p. 53, n. 10. But of course the NLRA doesn’t say even that much. And, besides, if the parties really take existing class and collective action rules as they find them, they surely take them subject to the limitations inherent in those rules—including the principle that parties may (as here) contract to depart from them in favor of individualized arbitration procedures of their own design. Still another contextual clue yields the same message. The employees’ underlying causes of action involve their wages and arise not under the NLRA but under an en- tirely different statute, the Fair Labor Standards Act. The FLSA allows employees to sue on behalf of “themselves and other employees similarly situated,” 29 U. S. C. §216(b), and it’s precisely this sort of collective action the employees before us wish to pursue. Yet they do not offer the seemingly more natural suggestion that the FLSA overcomes the Arbitration Act to permit their class and collective actions. Why not? Presumably because this Court held decades ago that an identical collective action scheme (in fact, one borrowed from the FLSA) does not displace the Arbitration Act or prohibit individualized arbitration proceedings. Gilmer v. Interstate/Johnson Lane Corp., 500 U. S. 20, 32 (1991) (discussing Age Discrimination in Employment Act). In fact, it turns out that “[e]very circuit to consider the question” has held that the FLSA allows agreements for individualized arbitration. Alternative Entertainment, 858 F. 3d, at 413 (opinion of Sutton, J.) (collecting cases). Faced with that obstacle, the employees are left to cast about elsewhere for help. And so they have cast in this direction, suggesting that one statute (the NLRA) steps in to dictate the procedures for claims under a different statute (the FLSA), and thereby overrides the commands of yet a third statute (the Arbitration Act). It’s a sort of interpretive triple bank shot, and just stating the theory is enough to raise a judicial eyebrow. Perhaps worse still, the employees’ theory runs afoul of the usual rule that Congress “does not alter the fundamental details of a regulatory scheme in vague terms or ancillary provisions—it does not, one might say, hide elephants in mouseholes.” Whitman v. American Trucking Assns., Inc., 531 U. S. 457, 468 (2001). Union organization and collective bargaining in the workplace are the bread and butter of the NLRA, while the particulars of dispute resolution procedures in Article III courts or arbitration proceedings are usually left to other statutes and rules—not least the Federal Rules of Civil Procedure, the Arbitration Act, and the FLSA. It’s more than a little doubtful that Congress would have tucked into the mousehole of Section 7’s catchall term an elephant that tramples the work done by these other laws; flattens the parties’ contracted-for dispute resolution procedures; and seats the Board as supreme superintendent of claims arising under a statute it doesn’t even administer. Nor does it help to fold yet another statute into the mix. At points, the employees suggest that the Norris-LaGuardia Act, a precursor of the NLRA, also renders their arbitration agreements unenforceable. But the Norris-LaGuardia Act adds nothing here. It declares “[un]enforceable” contracts that conflict with its policy of protecting workers’ “concerted activities for the purpose of collective bargaining or other mutual aid or protection.” 29 U. S. C. §§102, 103. That is the same policy the NLRA advances and, as we’ve seen, it does not conflict with Congress’s statutory directions favoring arbitration. See also Boys Markets, Inc. v. Retail Clerks, 398 U. S. 235 (1970) (holding that the Norris-LaGuardia Act’s anti-injunction provisions do not bar enforcement of arbitration agreements). What all these textual and contextual clues indicate, our precedents confirm. In many cases over many years, this Court has heard and rejected efforts to conjure conflicts between the Arbitration Act and other federal statutes. In fact, this Court has rejected every such effort to date (save one temporary exception since overruled), with statutes ranging from the Sherman and Clayton Acts to the Age Discrimination in Employment Act, the Credit Repair Organizations Act, the Securities Act of 1933, the Securities Exchange Act of 1934, and the Racketeer Influenced and Corrupt Organizations Act. Italian Colors, 570 U. S. 228; Gilmer, 500 U. S. 20; CompuCredit Corp. v. Greenwood, 565 U. S. 95 (2012); Rodriguez de Quijas v. Shearson/American Express, Inc., 490 U. S. 477 (1989) (over- ruling Wilko v. Swan, 346 U. S. 427 (1953)); Shearson/American Express Inc. v. McMahon, 482 U. S. 220 (1987). Throughout, we have made clear that even a statute’s express provision for collective legal actions does not necessarily mean that it precludes “ ‘individual attempts at conciliation’ ” through arbitration. Gilmer, supra, at 32. And we’ve stressed that the absence of any specific statutory discussion of arbitration or class actions is an important and telling clue that Congress has not displaced the Arbitration Act. CompuCredit, supra, at 103–104; McMahon, supra, at 227; Italian Colors, supra, at 234. Given so much precedent pointing so strongly in one direction, we do not see how we might faithfully turn the other way here. Consider a few examples. In Italian Colors, this Court refused to find a conflict between the Arbitration Act and the Sherman Act because the Sherman Act (just like the NLRA) made “no mention of class actions” and was adopted before Rule 23 introduced its exception to the “usual rule” of “individual” dispute resolution. 570 U. S., at 234 (internal quotation marks omitted). In Gilmer, this Court “had no qualms in enforcing a class waiver in an arbitration agreement even though” the Age Discrimination in Employment Act “expressly permitted collective legal actions.” Italian Colors, supra, at 237 (citing Gilmer, supra, at 32). And in CompuCredit, this Court refused to find a conflict even though the Credit Repair Organizations Act expressly provided a “right to sue,” “repeated[ly]” used the words “action ” and “court” and “class action,” and even declared “[a]ny waiver” of the rights it provided to be “void.” 565 U. S., at 99–100 (internal quotation marks omitted). If all the statutes in all those cases did not provide a congressional command sufficient to displace the Arbitration Act, we cannot imagine how we might hold that the NLRA alone and for the first time does so today. The employees rejoin that our precedential story is complicated by some of this Court’s cases interpreting Section 7 itself. But, as it turns out, this Court’s Section 7 cases have usually involved just what you would expect from the statute’s plain language: efforts by employees related to organizing and collective bargaining in the workplace, not the treatment of class or collective actions in court or arbitration proceedings. See, e.g., NLRB v. Washington Aluminum Co., 370 U. S. 9 (1962) (walkout to protest workplace conditions); NLRB v. Textile Workers, 409 U. S. 213 (1972) (resignation from union and refusal to strike); NLRB v. J. Weingarten, Inc., 420 U. S. 251 (1975) (request for union representation at disciplinary interview). Neither do the two cases the employees cite prove otherwise. In Eastex, Inc. v. NLRB, 437 U. S. 556, 558 (1978), we simply addressed the question whether a union’s distribution of a newsletter in the workplace qualified as a protected concerted activity. We held it did, noting that it was “undisputed that the union undertook the distribution in order to boost its support and improve its bargaining position in upcoming contract negotiations,” all part of the union’s “ ‘continuing organizational efforts.’ ” Id., at 575, and n. 24. In NLRB v. City Disposal Systems, Inc., 465 U. S. 822, 831–832 (1984), we held only that an employee’s assertion of a right under a collective bargaining agreement was protected, reasoning that the collective bargaining “process—beginning with the organization of the union, continuing into the negotiation of a collective-bargaining agreement, and extending through the enforcement of the agreement—is a single, collective activ- ity.” Nothing in our cases indicates that the NLRA guarantees class and collective action procedures, let alone for claims arising under different statutes and despite the express (and entirely unmentioned) teachings of the Arbitration Act. That leaves the employees to try to make something of our dicta. The employees point to a line in Eastex observing that “it has been held” by other courts and the Board “that the ‘mutual aid or protection’ clause protects employees from retaliation by their employers when they seek to improve working conditions through resort to administrative and judicial forums.” 437 U. S., at 565–566; see also Brief for National Labor Relations Board in No. 16–307, p. 15 (citing similar Board decisions). But even on its own terms, this dicta about the holdings of other bodies does not purport to discuss what procedures an employee might be entitled to in litigation or arbitration. Instead this passage at most suggests only that “resort to administrative and judicial forums” isn’t “entirely unprotected.” Id., at 566. Indeed, the Court proceeded to explain that it did not intend to “address . . . the question of what may constitute ‘concerted’ activities in this [litigation] context.” Ibid., n. 15. So even the employees’ dicta, when viewed fairly and fully, doesn’t suggest that individualized dispute resolution procedures might be insufficient and collective procedures might be mandatory. Neither should this come as a surprise given that not a single one of the lower court or Board decisions Eastex discussed went so far as to hold that Section 7 guarantees a right to class or collective action procedures. As we’ve seen, the Board did not purport to discover that right until 2012, and no federal appellate court accepted it until 2016. See D. R. Horton, 357 N. L. R. B. 2277; 823 F. 3d 1147 (case below in No. 16–285). With so much against them in the statute and our precedent, the employees end by seeking shelter in Chevron. Even if this Court doesn’t see what they see in Section 7, the employees say we must rule for them anyway because of the deference this Court owes to an administrative agency’s interpretation of the law. To be sure, the employees do not wish us to defer to the general counsel’s judgment in 2010 that the NLRA and the Arbitration Act coexist peaceably; they wish us to defer instead to the Board’s 2012 opinion suggesting the NLRA displaces the Arbitration Act. No party to these cases has asked us to reconsider Chevron deference. Cf. SAS Institute Inc. v. Iancu, ante, at 11. But even under Chevron’s terms, no deference is due. To show why, it suffices to outline just a few of the most obvious reasons. The Chevron Court justified deference on the premise that a statutory ambiguity represents an “implicit” delegation to an agency to interpret a “statute which it administers.” 467 U. S., at 841, 844. Here, though, the Board hasn’t just sought to interpret its statute, the NLRA, in isolation; it has sought to interpret this statute in a way that limits the work of a second statute, the Arbitration Act. And on no account might we agree that Congress implicitly delegated to an agency authority to address the meaning of a second statute it does not administer. One of Chevron’s essential premises is simply missing here. It’s easy, too, to see why the “reconciliation” of distinct statutory regimes “is a matter for the courts,” not agencies. Gordon v. New York Stock Exchange, Inc., 422 U. S. 659, 685–686 (1975). An agency eager to advance its statutory mission, but without any particular interest in or expertise with a second statute, might (as here) seek to diminish the second statute’s scope in favor of a more expansive interpretation of its own—effectively “ ‘bootstrap[ping] itself into an area in which it has no jurisdiction.’ ” Adams Fruit Co. v. Barrett, 494 U. S. 638, 650 (1990). All of which threatens to undo rather than honor legislative intentions. To preserve the balance Congress struck in its statutes, courts must exercise independent interpretive judgment. See Hoffman Plastic Compounds, Inc. v. NLRB, 535 U. S. 137, 144 (2002) (noting that this Court has “never deferred to the Board’s remedial preferences where such preferences potentially trench upon federal statutes and policies unrelated to the NLRA”). Another justification the Chevron Court offered for deference is that “policy choices” should be left to Executive Branch officials “directly accountable to the people.” 467 U. S., at 865. But here the Executive seems of two minds, for we have received competing briefs from the Board and from the United States (through the Solicitor General) disputing the meaning of the NLRA. And whatever argument might be mustered for deferring to the Executive on grounds of political accountability, surely it becomes a garble when the Executive speaks from both sides of its mouth, articulating no single position on which it might be held accountable. See Hemel & Nielson, Chevron Step One-and-a-Half, 84 U. Chi. L. Rev. 757, 808 (2017) (“If the theory undergirding Chevron is that voters should be the judges of the executive branch’s policy choices, then presumably the executive branch should have to take ownership of those policy choices so that voters know whom to blame (and to credit)”). In these circumstances, we will not defer. Finally, the Chevron Court explained that deference is not due unless a “court, employing traditional tools of statutory construction,” is left with an unresolved ambiguity. 467 U. S., at 843, n. 9. And that too is missing: the canon against reading conflicts into statutes is a traditional tool of statutory construction and it, along with the other traditional canons we have discussed, is more than up to the job of solving today’s interpretive puzzle. Where, as here, the canons supply an answer, “Chevron leaves the stage.” Alternative Entertainment, 858 F. 3d, at 417 (opinion of Sutton, J.). IV The dissent sees things a little bit differently. In its view, today’s decision ushers us back to the Lochner era when this Court regularly overrode legislative policy judgments. The dissent even suggests we have resur- rected the long-dead “yellow dog” contract. Post, at 3–17, 30 (opinion of Ginsburg, J.). But like most apocalyptic warnings, this one proves a false alarm. Cf. L. Tribe, American Constitutional Law 435 (1978) (“ ‘Lochnerizing’ has become so much an epithet that the very use of the label may obscure attempts at understanding”). Our decision does nothing to override Congress’s policy judgments. As the dissent recognizes, the legislative policy embodied in the NLRA is aimed at “safeguard[ing], first and foremost, workers’ rights to join unions and to engage in collective bargaining.” Post, at 8. Those rights stand every bit as strong today as they did yesterday. And rather than revive “yellow dog” contracts against union organizing that the NLRA outlawed back in 1935, today’s decision merely declines to read into the NLRA a novel right to class action procedures that the Board’s own general counsel disclaimed as recently as 2010. Instead of overriding Congress’s policy judgments, today’s decision seeks to honor them. This much the dissent surely knows. Shortly after invoking the specter of Lochner, it turns around and criticizes the Court for trying too hard to abide the Arbitration Act’s “ ‘liberal federal policy favoring arbitration agreements,’ ” Howsam v. Dean Witter Reynolds, Inc., 537 U. S. 79, 83 (2002), saying we “ ‘ski’ ” too far down the “ ‘slippery slope’ ” of this Court’s arbitration precedent, post, at 23. But the dissent’s real complaint lies with the mountain of precedent itself. The dissent spends page after page relitigating our Arbitration Act precedents, rehashing arguments this Court has heard and rejected many times in many cases that no party has asked us to revisit. Compare post, at 18–23, 26 (criticizing Mitsubishi Motors Corp. v. Soler Chrysler-Plymouth, Inc., 473 U. S. 614 (1985), Gilmer, 500 U. S. 20, Circuit City, 532 U. S. 105, Concepcion, 563 U. S. 333, Italian Colors, 570 U. S. 228, and CompuCredit, 565 U. S. 95), with Mitsubishi, supra, at 645–650 (Stevens, J., dissenting), Gilmer, supra, at 36, 39–43 (Stevens, J., dissenting), Circuit City, supra, at 124–129 (Stevens, J., dissenting), Concepcion, supra, at 357–367 (Breyer, J., dissenting), Italian Colors, supra, at 240–253 (Kagan, J., dissenting), and CompuCredit, supra, at 116–117 (Ginsburg, J., dissenting). When at last it reaches the question of applying our precedent, the dissent offers little, and understandably so. Our precedent clearly teaches that a contract defense “conditioning the enforceability of certain arbitration agreements on the availability of classwide arbitration procedures” is inconsistent with the Arbitration Act and its saving clause. Concepcion, supra, at 336 (opinion of the Court). And that, of course, is exactly what the employees’ proffered defense seeks to do. Nor is the dissent’s reading of the NLRA any more available to us than its reading of the Arbitration Act. The dissent imposes a vast construction on Section 7’s language. Post, at 9. But a statute’s meaning does not always “turn solely” on the broadest imaginable “definitions of its component words.” Yates v. United States, 574 U. S. ___, ___ (2015) (plurality opinion) (slip op., at 7). Linguistic and statutory context also matter. We have offered an extensive explanation why those clues support our reading today. By contrast, the dissent rests its interpretation on legislative history. Post, at 3–5; see also post, at 19–21. But legislative history is not the law. “It is the business of Congress to sum up its own debates in its legislation,” and once it enacts a statute “ ‘[w]e do not inquire what the legislature meant; we ask only what the statute means.’ ” Schwegmann Brothers v. Calvert Distillers Corp., 341 U. S. 384, 396, 397 (1951) (Jackson, J., concurring) (quoting Justice Holmes). Besides, when it comes to the legislative history here, it seems Congress “did not discuss the right to file class or consolidated claims against employers.” D. R. Horton, 737 F. 3d, at 361. So the dissent seeks instead to divine messages from congressional commentary directed to different questions altogether—a project that threatens to “substitute [the Court] for the Congress.” Schwegmann, supra, at 396. Nor do the problems end there. The dissent proceeds to argue that its expansive reading of the NLRA conflicts with and should prevail over the Arbitration Act. The NLRA leaves the Arbitration Act without force, the dissent says, because it provides the more “pinpointed” direction. Post, at 25. Even taken on its own terms, though, this argument quickly faces trouble. The dissent says the NLRA is the more specific provision because it supposedly “speaks directly to group action by employees,” while the Arbitration Act doesn’t speak to such actions. Ibid. But the question before us is whether courts must enforce particular arbitration agreements according to their terms. And it’s the Arbitration Act that speaks directly to the enforceability of arbitration agreements, while the NLRA doesn’t mention arbitration at all. So if forced to choose between the two, we might well say the Arbitration Act offers the more on-point instruction. Of course, there is no need to make that call because, as our precedents demand, we have sought and found a persuasive interpretation that gives effect to all of Congress’s work, not just the parts we might prefer. Ultimately, the dissent retreats to policy arguments. It argues that we should read a class and collective action right into the NLRA to promote the enforcement of wage and hour laws. Post, at 26–30. But it’s altogether unclear why the dissent expects to find such a right in the NLRA rather than in statutes like the FLSA that actually regulate wages and hours. Or why we should read the NLRA as mandating the availability of class or collective actions when the FLSA expressly authorizes them yet allows parties to contract for bilateral arbitration instead. 29 U. S. C. §216(b); Gilmer, supra, at 32. While the dissent is no doubt right that class actions can enhance enforcement by “spread[ing] the costs of litigation,” post, at 9, it’s also well known that they can unfairly “plac[e] pressure on the defendant to settle even unmeritorious claims,” Shady Grove Orthopedic Associates, P. A. v. Allstate Ins. Co., 559 U. S. 393, 445, n. 3 (2010) (Ginsburg, J., dissenting). The respective merits of class actions and private arbitration as means of enforcing the law are questions constitution- ally entrusted not to the courts to decide but to the policy- makers in the political branches where those questions remain hotly contested. Just recently, for example, one federal agency banned individualized arbitration agreements it blamed for underenforcement of certain laws, only to see Congress respond by immediately repealing that rule. See 82 Fed. Reg. 33210 (2017) (cited post, at 28, n. 15); Pub. L. 115–74, 131Stat. 1243. This Court is not free to substitute its preferred economic policies for those chosen by the people’s representatives. That, we had always understood, was Lochner’s sin. * The policy may be debatable but the law is clear: Congress has instructed that arbitration agreements like those before us must be enforced as written. While Congress is of course always free to amend this judgment, we see nothing suggesting it did so in the NLRA—much less that it manifested a clear intention to displace the Arbitration Act. Because we can easily read Congress’s statutes to work in harmony, that is where our duty lies. The judgments in Epic, No. 16–285, and Ernst & Young, No. 16–300, are reversed, and the cases are remanded for further proceedings consistent with this opinion. The judgment in Murphy Oil, No. 16–307, is affirmed. So ordered.

585.US.2017_142-orig

This original action concerns the proper apportionment of water from an interstate river basin. Three rivers form the heart of the Basin. The Chattahoochee and Flint Rivers begin near Atlanta, flow south through Georgia, and ultimately converge at Lake Seminole, just north of Florida, where the Apalachicola River begins and flows 106 miles south into the Gulf of Mexico. In 2013, Florida, the downstream State, sued Georgia, the upstream State, asking the Court to issue a decree equitably apportioning the Basin’s waters. The Court agreed to exercise its original jurisdiction and appointed a Special Master. The United States declined to waive its sovereign immunity from suit in the case. After conducting lengthy evidentiary proceedings, the Master submitted a Report recommending that the Court dismiss Florida’s complaint. That recommendation, the parties agree, turns on a single issue—namely, whether Florida met its initial burden in respect to redressability. The Master concluded that Florida failed to make the requisite showing because it did not present clear and convincing evidence that its injuries could be redressed by a decree capping Georgia’s upstream water consumption if the decree does not also bind the Corps. Florida has filed exceptions to the Master’s Report. Held: 1. The Special Master applied too strict a standard in concluding that Florida failed to meet its initial burden of demonstrating that the Court can eventually fashion an effective equitable decree. Pp. 10–18. (a) Where, as here, the Court is asked to resolve an interstate water dispute raising questions beyond the interpretation of specific language of an interstate compact, the doctrine of equitable apportionment applies. In this realm, several related but more specific sets of principles guide the Court’s review. First, both Georgia and Florida possess “an equal right to make a reasonable use of the waters of” the Flint River. United States v. Willow River Power Co., 324 U. S. 499, 505. Second, when confronted with competing claims to interstate water, the Court’s “effort always is to secure an equitable apportionment without quibbling over formulas.” New Jersey v. New York, 283 U. S. 336, 343. Third, in light of the sovereign status and “equal dignity” of States, a complaining State’s burden is “much greater” than the burden ordinarily shouldered by a private party seeking an injunction. Connecticut v. Massachusetts, 282 U. S. 660, 669. Among other things, it must demonstrate, by “ ‘clear and convincing evidence,’ ” that it has suffered a “ ‘threatened invasion of rights’ ” that is “ ‘of serious magnitude.’ ” Washington v. Oregon, 297 U. S. 517, 522. And to the extent the Court has addressed the “initial burden” a State bears in respect to redressability, the Court has said that “it should be clear that [the complaining] State has not merely some technical right, but also a right with a corresponding benefit” as a precondition to any equitable apportionment. Kansas v. Colorado, 206 U. S. 46, 102, 109. An effort to shape a decree cannot be “a vain thing.” Foster v. Mansfield, C. & L. M. R. Co., 146 U. S. 88, 101. Finally, because equitable apportionment is “ ‘flexible,’ ” not “formulaic,” this Court will seek to “arrive at a ‘ “just and equitable” apportionment’ of an interstate stream” by “consider[ing] ‘all relevant factors,’ ” South Carolina v. North Carolina, 558 U. S. 256, 271, including, inter alia, “ ‘physical and climatic conditions, the consumptive use of water in the several sections of the river, the character and rate of return flows, the extent of established uses, the availability of storage water, the practical effect of wasteful uses on downstream areas, [and] the damage to upstream areas as compared to the benefits to downstream areas if a limitation is imposed on the former.’ ” Colorado v. New Mexico, 459 U. S. 176, 183. Because all relevant factors must be weighed, extensive and specific factual findings are essential for the Court to properly apply the doctrine of equitable apportionment. See Nebraska v. Wyoming, 325 U. S. 589, 618. Pp. 10–15. (b) The Special Master applied too strict a standard when he determined that the Court would not be able to fashion an appropriate equitable decree. The Master referred to this as a “threshold” showing. But it is “threshold” only in the sense that the Master has not yet determined key remedy-related matters, including the approximate amount of water that must flow into the Apalachicola River in order for Florida to receive a significant benefit from a cap on Georgia’s use of Flint River waters. Unless and until the Special Master makes the findings of fact necessary to determine the nature and scope of likely harm caused by the absence of water and the amount of additional water necessary to ameliorate that harm significantly, the complaining State should not have to prove with specificity the details of an eventually workable decree by “clear and convincing” evidence. Rather, the complaining State should have to show that, applying the principles of “flexibility” and “approximation,” it is likely to prove possible to fashion such a decree. To require “clear and convincing evidence” about the workability of a decree before the Court or a Special Master has a view about likely harms and likely amelioration is, at least in this case, to put the cart before the horse. Pp. 15–18. 2. The Court reserves judgment as to the ultimate disposition of this case, addressing here only the narrow “threshold” question the Master addressed below—namely, whether Florida has shown that its “injur[ies can] effectively be redressed by limiting Georgia’s consumptive use of water from the Basin without a decree binding the Corps.” Report 30–31. Florida has made a legally sufficient showing as to the possibility of fashioning an effective remedial decree. Pp. 18–37. (a) The Report makes several key assumptions. First, the Master assumed Florida has suffered harm as a result of decreased water flow into the Apalachicola River. Second, the Master further assumed that Florida has shown that Georgia, contrary to equitable principles, has taken too much water from the Flint River. Third, the Master assumed that Georgia’s inequitable use of the water injured Florida. At this stage of the proceeding and in light of these assumptions, Florida made a sufficient showing that the extra water that would result from its proposed consumption cap would both lead to increased streamflow in Florida’s Apalachicola River and significantly redress the economic and ecological harm that Florida has alleged. In addition, the United States has made clear that the Corps will cooperate in helping to implement any determinations and obligations the Court sets forth in a final decree in this case. While the Corps must take account of a variety of circumstances and statutory obligations when it allocates water, it cannot now be said that an effort to shape a decree here will prove “a vain thing,” Foster, supra, at 101, since the record indicates that, if necessary and with the help of the United States, the Special Master, and the parties, the Court should be able to fashion a decree. Pp. 20–35. (b) Further findings, however, are needed on all of these evidentiary issues. Florida will be entitled to a decree only if it is shown that “the benefits of the [apportionment] substantially outweigh the harm that might result.” Colorado, 459 U. S., at 187. On remand, before fashioning a remedy, the Special Master must address several evidentiary questions that are assumed or found plausible here. Pp. 35–37. Case remanded. Breyer, J., delivered the opinion of the Court, in which Roberts, C. J., and Kennedy, Ginsburg, and Sotomayor, JJ., joined. Thomas, J., filed a dissenting opinion, in which Alito, Kagan, and Gorsuch, JJ., joined.

This case concerns the proper apportionment of the water of an interstate river basin. Florida, a downstream State, brought this lawsuit against Georgia, an upstream State, claiming that Georgia has denied it an equitable share of the basin’s waters. We found that the dispute lies within our original jurisdiction, and we appointed a Special Master to take evidence and make recommendations. After lengthy evidentiary proceedings, the Special Master submitted a report in which he recommends that the Court deny Florida’s request for relief on the ground that “Florida has not proven by clear and convincing evidence that its injury can be redressed by an order equitably apportioning the waters of the Basin.” Report of Special Master 3. The case is before us on Florida’s exceptions to the Special Master’s Report. In light of our examination of the Report and relevant portions of the record, we remand the case to the Master for further findings and such further proceedings as the Master believes helpful. I A This original action arises out of a dispute over the division of water from an interstate river basin known as the Apalachicola-Chattahoochee-Flint River Basin. The Basin drains an area of more than 20,000 square miles across the southeastern United States. Three interstate rivers form the heart of the Basin and are central to this case. They are the Chattahoochee River, the Flint River, and the Apalachicola River. It is easiest to think of these three rivers as forming the capital letter “Y,” with each branch starting at a different point in northeastern Georgia near Atlanta and the stem running through the Flor- ida panhandle and emptying into Apalachicola Bay in the Gulf of Mexico. See Appendix, infra. The Chattahoochee River is the western branch of this Y-shaped river system. It runs from the foothills of Georgia’s Blue Ridge Mountains, through most of Georgia, down to Lake Seminole, just north of Florida. The United States Army Corps of Engineers operates several dams and reservoirs along the Chattahoochee where it both stores water and controls the amount of water that flows downstream to Florida in accordance with the terms of its recently revised Master Water Control Manual (Master Manual). As we shall discuss in more detail, Part IV, infra, the Corps’ operations are important to the resolution of this case. The Flint River, the eastern branch of the “Y,” runs from just south of Atlanta down to the same lake, namely, Lake Seminole. Unlike the Chattahoochee, there are no dams along the Flint River; it flows unimpeded through southern Georgia’s farmland, where the greatest share of the Basin’s water is consumed by agricultural irrigation. After water from the Flint and Chattahoochee Rivers mixes at Lake Seminole, the mixed water (now forming the stem of the Y) continues its southward journey. At the southern end of Lake Seminole, it flows through the Woodruff Dam—a dam also controlled by the Corps. The mixed waters then change their name. They are called the Apalachicola River, and under that name they flow 106 miles through the Florida Panhandle and finally empty into the Gulf of Mexico. There, the fresh water of the Apalachicola River mixes with the Gulf’s saltwater, forming Apalachicola Bay, which the United Nations, the United States, and the State of Florida have all recognized as one of the Northern Hemisphere’s most productive estuaries. In total, the Apalachicola River accounts for 35% of the fresh water that flows along Florida’s western coast. See Joint Exh. 168, p. 39. B Florida and Georgia have long disputed the apportionment of the Basin’s waters. Florida contends that Georgia is consuming more than its equitable share of Flint River water. It adds that, were Georgia to consume less water from the Flint River, more water would flow into Lake Seminole, pass through the Woodruff Dam and subsequently flow down the Apalachicola River (the Y’s stem) and into Apalachicola Bay. The additional water that would result from a cap on Georgia’s consumption would, Florida argues, help (among other things) to recover and maintain its oyster industry, which collapsed following a drought in 2012. Georgia believes that it should not have to cut back on its Flint River water consumption because, in its view, it consumes no more than its equitable share. “This Court has recognized for more than a century its inherent authority, as part of the Constitution’s grant of original jurisdiction, to equitably apportion interstate streams between States.” Kansas v. Nebraska, 574 U. S. ___, ___ (2015) (slip op., at 7). But we have long noted our “preference” that States “settle their controversies by ‘mutual accommodation and agreement.’ ” Arizona v. California, 373 U.S. 546, 564 (1963) (quoting Colorado v. Kansas, 320 U.S. 383, 392 (1943) (Kansas II )); see also id., at 392 (“[Interstate] controversies may appropriately be composed by negotiation and agreement, pursuant to the compact clause of the federal Constitution”); Kansas v. Nebraska, supra, at ___ (slip op., at 2–3) (describing codification of Republican River Compact); Montana v. Wyoming, 563 U.S. 368, 372 (2011) (interpreting Yellowstone River Compact); Kansas v. Colorado, 543 U.S. 86 (2004) (resolving dispute over Arkansas River Compact). We recognize that Florida and Georgia (sometimes with the help of the Federal Government) have long tried to do so. But so far they have failed. In 1992, for example, the States signed a memorandum of agreement in which they “committed to a process for cooperative management and development” of the three-river Basin and agreed to “participate fully as equal partners” in a “comprehensive, basin-wide study” of its waters. Joint Exh. 004, at 1. Five years later, the States signed—and Congress approved—a compact, the Apalachicola-Chattahoochee-Flint River Basin Compact, in which they agreed: “to develop an allocation formula for equitably apportioning the surface waters of the ACF Basin among the states while protecting the water quality, ecology and biodiversity of the ACF.” 111Stat. 2222–2223. But five years of negotiations under the Compact proved fruitless, and in 2003, the Compact expired. More than a decade later, in 2014, Congress again recognized the need for an equitable apportionment of Basin waters. See Water Resources Reform and Development Act of 2014, Pub. L. 113–121, §1051(a), 128Stat. 1259. But once again, despite drought, expanding city populations, and a dramatic increase in acreage devoted to agricultural irrigation, no agreement has been reached. The “last effort to reach an amicable resolution of this complex equitable apportionment proceeding” in 2017 was “unsuccessful.” Report 24. The States instead have come to this Court. II A In 2013, Florida, the downstream State, sought to sue Georgia, the upstream State, asking us to exercise our “original and exclusive jurisdiction” and issue a decree equitably apportioning the waters of the Basin. 28 U. S. C. §1251(a); see U. S. Const. Art. III, §2; see also this Court’s Rule 17. In its complaint, Florida alleged that Georgia’s consumption of Flint River water “reduce[s] the amount of water flowing to the Apalachicola River at all times,” and noted that “the effects are especially apparent during the low flow summer and fall periods.” Complaint 9, ¶21; see also id., at 17, ¶49 (complaining that the impact of Georgia’s water consumption “is significant, particularly during dry periods”). In addition, Florida alleged that “[a]s Georgia’s upstream storage and consumption grows over time, low flow events will become more frequent and increase in severity, diminishing the likelihood that key species will survive and precluding any chance of recovery over the long term.” Id., at 20, ¶59. To remedy these harms, Florida seeks a cap on Georgia’s consumption of water from the Flint River. Id., at 21. Georgia filed a brief in opposition, arguing that Florida failed to allege an injury sufficient to warrant this Court’s exercise of original jurisdiction. See State of Georgia’s Opposition to Florida’s Motion for Leave to File a Complaint 31 (“Florida has not pleaded facts plausibly suggesting that it will be able to establish clear and convincing evidence that it suffers substantial injury as a result of Georgia’s consumption of water”). At our request, the United States filed a brief in which it told us that “Florida has pleaded an interstate water dispute of sufficient importance to warrant this Court’s exercise of its original jurisdiction, and no other judicial forum is suitable for resolving the overall controversy.” Brief for United States as Amicus Curiae 12 (Sept. 18, 2014). But, the United States also warned that “[p]ractical considerations . . . weigh against the Court’s resolution of Florida’s claims before the Corps has completed its process of updating the Master Manual for the federal projects in the ACF Basin.” Ibid. It suggested that the Court could “grant Florida leave to file, but stay or provide for tailoring of any further proceedings until the Corps has issued the revised Master Manual” in March 2017, id., at 13 (which Florida has now done, see Brief for United States as Amicus Curiae 3, n. 1, 10–12). We subsequently agreed to exercise our original jurisdiction and appointed a Special Master “with authority to . . . direct subsequent proceedings,” “take such evidence as may be introduced and such as he may deem it necessary to call for,” and “submit Reports as he may deem appropriate.” 574 U. S. ___ (2014). At the outset, the United States declined to waive its sovereign immunity from suit in this case. And shortly thereafter, Georgia asked the Special Master to dismiss the case on the grounds that the United States was a necessary party but could not be forced to intervene. See Fed. Rule Civ. Proc. 19(b). The Master concluded that the motion to dismiss Florida’s complaint should be denied. The Master reasoned that a decree binding the Corps might not prove necessary. Order on State of Georgia’s Motion To Dismiss 14–15 (June 19, 2015). Rather, the Master concluded that “the few facts before me at this stage of the proceeding support the conclusion that” a cap on Georgia’s Flint River water consumption could, at least in principle, redress Florida’s injuries either by increasing the amount of water that flows into Florida’s Apalachicola River or by “render[ing] periods of reduced flow releases [into the Apalachicola River] fewer and further between because of the increased reservoir levels that would result from Georgia’s reduced consumption.” Id., at 14, and n. 5. The Special Master pointed out that Florida would have to show that “a consumption cap is justified and will afford adequate relief.” Id., at 13. B The Master then held lengthy discovery and evidentiary proceedings. See Brief for Georgia 11; post, at 23 (opinion of Thomas, J.) (“During their 18 months of discovery, the parties produced 7.2 million pages of documents”). Ultimately, the Master submitted a 70-page Report to this Court in February 2017. He recommended that the Court dismiss Florida’s complaint. In particular, despite the very large factual record amassed and “the extensive testimony bearing on numerous issues,” the Special Master stated: “I have concluded that there is a single, discrete issue that resolves this case: even assuming that Florida has sustained injury as a result of unreasonable upstream water use by Georgia, can Florida’s injury effectively be redressed by limiting Georgia’s consumptive use of water from the Basin without a decree binding the [Army] Corps [of Engineers]? I conclude that Florida has not proven that its injury can be remedied without such a decree. The evidence does not provide sufficient certainty that an effective remedy is available without the presence of the Corps as a party in this case.” Report 30–31 (emphasis added). For present purposes, we note that Florida and Georgia agree that the Master’s recommendation “turned on a ‘single, discrete issue’—whether Florida had shown that a cap on Georgia’s consumption would redress its injury if the decree did not bind the Corps as well.” Florida Brief in Support of Exceptions 23–24; see also Georgia’s Reply to Florida’s Exceptions 23 (“The Special Master reserved ruling on any issue other than effective redress”); Brief for United States as Amicus Curiae 19–20 (Aug. 7, 2017) (same). In reviewing this determination, we do not agree with the dissent’s view that the Master applied the “ordinary balance-of-harms test” that our equitable apportionment cases require. Post, at 14 (opinion of Thomas, J.); see also Part III–A, infra, (describing equitable apportionment doctrine). As we shall explain, the dissent’s assertion that “the balance of harms cannot tip in Florida’s favor” is, at best, premature. Post, at 34–35. That judgment may eventually prove right or it may prove wrong. Here, as we just said, we consider only the “single” and “threshold” question of “redressability” upon which the Master rested his conclusion and which the parties have now argued here. In determining precisely what we now review, we rely upon (and do not go beyond) the Report’s specific and key statements, which include the following: “As a threshold matter, equitable apportionment is only available to a state that has suffered ‘real and substantial injury’ as a result of proposed or actual upstream water use” and “the injury must be redressable by the Court.” Report 24 (emphasis added). “Florida points to real harm and, at the very least, likely misuse of resources by Georgia. There is little question that Florida has suffered harm from decreased flows in the [Apalachicola] River,” including “an unprecedented collapse of its oyster fisheries in 2012.” Id., at 31. “Much more could be said and would need to be said on these [and other] issues . . . .” Id., at 34. “I need only address the narrow question of which party bears the burden of proving injury and redressability.” Id., at 28–29 (emphasis added). “Florida bears the burden to prove that the proposed remedy will provide redress for Florida’s injury.” Id., at 30. “Florida has not proven by clear and convincing evidence that any additional streamflow in the Flint River or in the Chattahoochee River would be released from Jim Woodruff Dam into the Apalachicola River at a time that would provide a material benefit to Florida (i.e., during dry periods), thereby alleviating Florida’s injury.” Id., at 47 (emphasis added). “Florida has provided no evidence that a decree in this case could provide an effective remedy during normal (i.e., non-drought) periods.” Id., at 68. “[T]he Corps can likely offset increased streamflow in the Flint River by storing additional water in its reservoirs along the Chattahoochee River during dry periods [and so] . . . [t]here is no guarantee that the Corps will exercise its discretion to release or hold back water at any particular time.” Id., at 69 (emphasis added). “[W]ithout the Corps as a party, the Court cannot order the Corps to take any particular action.” Id., at 69–70. C Florida has filed exceptions to the Special Master’s Report. Florida first challenges the legal standard the Master applied in resolving what the Master called the “threshold” question whether Florida had “proven. . . that its injury can be redressed by an order equitably apportioning the waters of the Basin.” Id., at 24, 3. The Master wrote that Florida must meet a “clear and convincing evidence” evidentiary burden. Id., at 3. Second, Florida argues that, in any event, its showing in respect to redressability was sufficient. We consider each of these exceptions in turn. III A We note at the outset that our role in resolving disputes between sovereign States under our original jurisdiction “significantly differs from the one the Court undertakes ‘in suits between private parties.” Kansas v. Nebraska, 574 U. S., at ___ (slip op., at 6) (internal quotation marks and alterations omitted). “In this singular sphere,” we have observed, “ ‘the court may regulate and mould the process it uses in such a manner as in its judgment will best promote the purposes of justice.’ ” Id., at ___ (slip op., at 6–7) (quoting Kentucky v. Dennison, 24 How. 66, 98 (1861)). We must approach interstate disputes “in the untechnical spirit proper for dealing with a quasi-international controversy, remembering that there is no municipal code governing the matter, and that this court may be called on to adjust differences that cannot be dealt with by Congress or disposed of by the legislature of either State alone.” Virginia v. West Virginia, 220 U.S. 1, 27 (1911) (Holmes, J.). Where, as here, the Court is asked to resolve an interstate water dispute raising questions beyond the interpretation of specific language of an interstate compact, the doctrine of equitable apportionment governs our inquiry. See Colorado v. New Mexico, 459 U.S. 176, 183 (1982) (Colorado I); Virginia v. Maryland, 540 U.S. 56, 74, n. 9 (2003) (“Federal common law governs interstate bodies of water, ensuring that the water is equitably apportioned between the States and that neither State harms the other’s interest in the river”). In this realm, we have kept in mind several related but more specific sets of principles. First, as the Special Master pointed out, “the relevant guiding principle in this case” is a simple one. Report 26–27. Given the laws of the States, both Georgia and Florida possess “ ‘an equal right to make a reasonable use of the waters of the stream’ ”—which, in this case, is the Flint River. Id., at 26 (quoting United States v. Willow River Power Co., 324 U.S. 499, 505 (1945)); see also Colorado I supra, at 184 (“Our prior cases clearly establish that equitable apportionment will only protect those rights to water that are ‘reasonably required and applied.’ . . . [W]asteful or inefficient uses will not be protected (quoting Wyoming v. Colorado, 259 U.S. 419, 484 (1922))); Idaho ex rel. Evans v. Oregon, 462 U.S. 1017, 1025 (1983) (Idaho II ) (“States have an affirmative duty under the doctrine of equitable apportionment to take reasonable steps to conserve and even to augment the natural resources within their borders for the benefit of other States”); Nebraska v. Wyoming, 325 U.S. 589, 618 (1945); Kansas II, 320 U. S., at 394; Washington v. Oregon, 297 U.S. 517, 522, 527–528 (1936); New Jersey v. New York, 283 U.S. 336, 342–343 (1931); North Dakota v. Minnesota, 263 U.S. 365, 372 (1923) (reaffirming that an upstream State may not “burden his lower neighbor with more than is reasonable”); Kansas v. Colorado, 206 U.S. 46, 102 (1907) (Kansas I); Tyler v. Wilkinson, 24 F. Cas. 472, 474 (No. 14,312) (CC RI 1827) (Story, J.) (setting forth the principle of “reasonable use”). Second, our prior decisions emphasize that, when we are confronted with competing claims to interstate water, the Court’s “effort always is to secure an equitable apportionment without quibbling over formulas.” New Jersey v. New York, 283 U. S., at 342–343 (Holmes, J.). Where “[b]oth States have real and substantial interests in the River,” those interests “must be reconciled as best they may be.” Id., at 342–343. We have added that “[u]ncertainties about the future . . . do not provide a basis for declining to fashion a decree.” Idaho II, 462 U. S., at 1026; see also ibid. (“Reliance on reasonable predictions of future conditions is necessary”); Colorado v. New Mexico, 467 U.S. 310, 322 (1984) (Colorado II ) (requiring “absolute precision in forecasts . . . would be unrealistic”); North Dakota v. Minnesota, supra, at 386 (emphasizing the need to “draw inferences as to the probabilities”); Kansas I, supra, at 97–98. Third, in light of the sovereign status and “equal dignity” of States, a complaining State must bear a burden that is “much greater” than the burden ordinarily shouldered by a private party seeking an injunction. Connecticut v. Massachusetts, 282 U.S. 660, 669 (1931); see Kansas II, supra, at 392 (“The reason for judicial caution in adjudicating the relative rights of States in such cases is that, while we have jurisdiction of such disputes, they involve the interests of quasi-sovereigns, present complicated and delicate questions, and, due to the possibility of future change of conditions, necessitate expert administration rather than judicial imposition of a hard and fast rule” (footnote omitted)). In particular, “ ‘[b]efore this court can be moved to exercise its extraordinary power under the Constitution to control the conduct of one State at the suit of another,’ ” the complaining State must demonstrate that it has suffered a “ ‘threatened invasion of rights’ ” that is “ ‘of serious magnitude.’ ” Washington v. Oregon, supra, at 524 (quoting New York v. New Jersey, 256 U.S. 296, 309 (1921)). The State must make that showing by “ ‘clear and convincing evidence.’ ” Washington v. Oregon, supra, at 522 (quoting New York v. New Jersey, supra, at 309); see also Idaho II, supra, at 1027 (“A State seeking equitable apportionment under our original jurisdiction must prove by clear and convincing evidence some real and substantial injury or damage”); Colorado I, supra, at 187–188, n. 13 (“[A] state seeking to prevent or enjoin [an upstream] diversion by another State” must “bear the initial burden of showing that a diversion by [the upstream State] will cause substantial injury to [the downstream State’s] interests” (emphasis added)). In addition, to the extent the Court has addressed the “initial burden” a State bears in respect to redressability, our prior decisions make clear that, as a general matter, “[t]o constitute a justiciable controversy, it must appear that the complaining State has suffered a wrong through the action of the other State, furnishing a ground for judicial redress, or is asserting a right against the other State which is susceptible of judicial enforcement according to the accepted principles of the common law or equity systems of jurisprudence.” Massachusetts v. Missouri, 308 U.S. 1, 15 (1939)); see also Wyoming v. Oklahoma, 502 U.S. 437, 447, 452 (1992) (same); Maryland v. Louisiana, 451 U.S. 725, 735–736 (1981). More specifically, we have said that “it should be clear that [the complaining] State has not merely some technical right, but also a right with a corresponding benefit” as a precondition to any equitable apportionment. Kansas I, supra, at 109. An effort to shape a decree cannot be “a vain thing.” Foster v. Mansfield, C. & L. M. R. Co., 146 U.S. 88, 101 (1892). A State “will not be granted [relief] against something merely feared as liable to occur at some indefinite time in the future,” Connecticut v. Massachusetts, supra, at 674, or when there is “no other or better purpose [at stake] than to vindicate a barren right.” Washington v. Oregon, supra, at 523; cf. Idaho II, supra, at 1026 (assessing whether “the formulation of a workable decree is impossible”). Fourth, in an interstate water matter, where a complaining State meets its “initial burden of showing ‘real or substantial injury,’ ” Colorado II, supra, at 317 (quoting Colorado I, 459 U. S., at 188, n. 13), this Court, recalling that equitable apportionment is “ ‘flexible,’ ” not “formu- laic,” will seek to “arrive at a ‘ “just and equitable” apportionment’ of an interstate stream” by “consider[ing] ‘all relevant factors.’ ” South Carolina v. North Carolina, 558 U.S. 256, 271 (2010) (quoting Colorado I, 459 U. S., at 183); see also id., at 190 (“Whether [relief] should be permitted will turn on an examination of all factors relevant to a just apportionment”); Kansas II, 320 U. S., at 393–394 (“[I]n determining whether one State is using, or threatening to use, more than its equitable share of the benefits of a stream, all the factors which create equities in favor of one State or the other must be weighed”) (emphasis added). These factors include (but are not limited to): “physical and climatic conditions, the consumptive use of water in the several sections of the river, the character and rate of return flows, the extent of established uses, the availability of storage water, the practical effect of wasteful uses on downstream areas, [and] the damage to upstream areas as compared to the benefits to downstream areas if a limitation is imposed on the former.” Nebraska v. Wyoming, 325 U. S., at 618. Because “all the factors which create equities in favor of one State or the other must be weighed,” Kansas II, supra, at 394 (emphasis added), extensive and “specific factual findings” are essential for the Court to properly apply the doctrine of equitable apportionment. Colorado I, supra, at 189–190 (emphasis added). And given the complexity of many water-division cases, the need to secure equitable solutions, the need to respect the sovereign status of the States, and the importance of finding flexible solutions to multi-factor problems, we typically appoint a Special Master and benefit from detailed factual findings. Without the full range of factual findings, we have said, the Court may lack an adequate basis on which to make “the delicate adjustment of interests” that the law requires. Nebraska v. Wyoming, supra, at 618; Washington v. Oregon, 297 U. S., at 519, 523–524 (emphasizing that “the Master’s Report finds the facts fully”); see also Colorado I, supra, at 183, 189–190 (remanding “with instructions to the Special Master to make further findings of fact”); Colorado II, 467 U. S., at 312–315 (explaining that because “the Master’s report [was] unclear,” the Court remanded to the Special Master “for additional factual findings on five specific issues” even after “a lengthy trial at which both States presented extensive evidence” in order “to assist this Court in balancing the benefit and harm”); Texas v. New Mexico, 462 U.S. 554, 575–576, and n. 21 (1983) (“[W]e return this case to the Special Master for determination of the unresolved issues framed in his pretrial order”); 3 A. Kelley, Water and Water Rights §45.02(c), p. 45–14 (3d ed. 2018) (“If the factual findings in the report are insufficient for the Court to decide whether the master correctly applied the doctrine of equitable apportionment, the Court may refer the case back to the master for additional findings”). B Applying the principles just described, we conclude that the Special Master applied too strict a standard when he determined that the Court would not be able to fashion an appropriate equitable decree. See Report 3 (“Florida has not proven by clear and convincing evidence that its injury can be redressed by an order equitably apportioning the waters of the Basin”); see also id., at 31 (“The evidence does not provide sufficient certainty that an effective remedy is available without the presence of the Corps as a party in this case”). The Special Master referred to the relevant showing that Florida must make in this respect as a “threshold” showing. Report 24. We agree that the matter is “threshold” in one particular sense—namely, the sense that the Master has not yet determined several key remedy-related matters, including the approximate amount of water that must flow into the Apalachicola River in order for Florida to receive a significant benefit from a cap on Georgia’s use of Flint River waters. See infra, at 28. The Master also wrote that Florida had failed to show “with sufficient certainty that the Corps must (or will choose to) operate its projects so as to permit all additional flows in the Flint River” or “the entire marginal increase in streamflow” to reach Florida “without any substantial delay.” Id., at 48 (emphasis added); see also id., at 24, 70 (similar). He added that there “is no guarantee” that the Corps will exercise its relevant discretion. Id., at 69 (emphasis added). And he said that Florida must show the existence of a workable remedy by “clear and convincing evidence.” Id., at 3; see also, e.g., id., at 28–29, 47, 51, 69–70. We believe the Master’s standard, as indicated by these statements, is too strict. In our view, unless and until the Special Master makes the findings of fact necessary to determine the nature and scope of likely harm caused by the absence of water and the amount of additional water necessary to ameliorate that harm significantly, the complaining State should not have to prove with specificity the details of an eventually workable decree by “clear and convincing” evidence. Rather, the complaining State should have to show that, applying the principles of “flexibility” and “approximation” we discussed above, it is likely to prove possible to fashion such a decree. See supra, at 12. To require more definite proof at the outset may well (at least on some occasions) make little sense. Suppose, for example, downstream State A claims that upstream State B wastes at least 10,000 cubic feet per second (cfs) of water. And suppose further that no decree could enforce a 10,000 cfs consumption cap but that it may well prove possible to enforce a lesser requirement. If so, we would have to know at least approximately how much water will significantly ameliorate State A’s water problem before we could know whether it is possible to shape a workable decree. And the workability of decrees themselves, approximate as they may be, may depend upon more precise findings in respect to the nature and scope of the range of likely harms and likely benefits that a Special Master finds are actually likely to exist. To require “clear and convincing evidence” about the workability of a decree before the Court or a Special Master has a view about likely harms and likely amelioration is, at least in this case, to put the cart before the horse. And that, we fear, is what the Master’s statements, with their apparent references to a “clear and convincing” evidence standard in respect to “redressability” (where that refers to the availability of an eventual decree) have done here. Cf. post, at 17–19. That is also why our cases, while referring to the use of a “clear and convincing” evidentiary standard in respect to an initial showing of “invasion of rights” and “substantial injury,” have never referred to that standard in respect to a showing of “remedy” or “redressability.” See Nebraska v. Wyoming, 515 U.S. 1, 8 (1995) (repeating that as a threshold matter, a “ ‘threatened invasion of rights must be of a serious magnitude and it must be established by clear and convincing evidence’ ” without addressing the required initial burden in respect to remedy (quoting New York v. New Jersey, 256 U. S., at 309)); Colorado II, supra, at 317 (describing the “initial burden” a State bears to show “ ‘real or substantial injury’ ” (quoting Colorado I, 459 U. S., at 187–188, n. 13)); Idaho II, 462 U. S., at 1027; Colorado I, supra, at 187–188, and n. 13 (“[A] State seeking to prevent or enjoin [an upstream] diversion by another State” must “bear the initial burden of showing that a diversion by [the upstream State] will cause substantial injury to [the downstream State’s] interests” (emphasis added)); Washington v. Oregon, 297 U. S., at 522; Connecticut v. Massachusetts, 282 U. S., at 672; New Jersey v. New York, 283 U. S., at 344–345; Kansas II, 320 U. S., at 393–394. The dissent does not dispute this. See post, at 12. As discussed, supra, at 12–13, our prior decisions have said that the “right” a complaining State asserts must be more than “merely some technical right” and must be “a right with a corresponding benefit,” Kansas I, 206 U. S., at 109 (emphasis added)—an effort to shape an equitable apportionment decree cannot be “a vain thing.” Foster, 146 U. S., at 101; see also Idaho II, supra, at 1026 (assessing whether “the formulation of a workable decree is impossible”); Washington v. Oregon, supra, at 523. But these statements apply to the general availability of judicial relief—not to the details of a final decree or to the workability of a decree that will depend on those details. Cf. Idaho ex rel. Evans v. Oregon 444 U.S. 380, 392 (1980) (Idaho I ) (explaining that the question whether a State’s proposed remedy will have an “appreciable effect” is a question that “goes to the merits” of the equitable apportionment inquiry). And, of course, to insist upon the use of such a strict standard, in respect to an eventual decree, runs directly contrary to the statements in, and holdings of, cases to which we have referred when discussing the need for “approximation” and “flexibility.” See supra, at 13–14. IV We next address Florida’s exceptions to the Master’s evidentiary determinations. In doing so, we recognize that the record in this case is long. It addresses a number of highly technical matters on a range of subjects—from biology to hydrology to the workings of the Corps’ newly revised Master Manual governing the organization’s complex operations in the Basin. Insofar as the Special Master made findings of fact, those findings “deserve respect and a tacit presumption of correctness.” Colorado II, 467 U. S., at 317. But at the end of the day, “the ultimate responsibility for deciding what are correct findings of fact remains with us.” Ibid. We have therefore read those portions of the record to which the parties, amici, or the Master refer, along with several other portions that we have found potentially relevant. Our “independent examination of the record,” Kansas v. Missouri, 322 U.S. 213, 232 (1944), leads us to conclude that, at this stage, Florida has met its “initial burden” in respect to remedy. But, we also believe that a remand is necessary to conduct the equitable-balancing inquiry. Cf. Colorado I, supra, at 183–190. We reserve judgment as to the ultimate disposition of this case, addressing here only the narrow “threshold” question the Master addressed below—namely, whether Florida has shown that its “injur[ies can] effectively be redressed by limiting Georgia’s consumptive use of water from the Basin without a decree binding the Corps.” Report 30–31. This dispositive threshold question leads us, in turn, to focus upon five subsidiary questions: First, has Florida suffered harm as a result of decreased water flow into the Apalachicola River? (The Special Master assumed “yes.”) Second, has Florida shown that Georgia, contrary to equitable principles, has taken too much water from the Flint River (the eastern branch of the Y-shaped river system)? (Again, the Special Master assumed “yes.”) Third, if so, has Georgia’s inequitable use of Basin waters injured Florida? (The Special Master assumed “yes.”) Fourth, if so, would an equity-based cap on Georgia’s use of the Flint River lead to a significant increase in streamflow from the Flint River into Florida’s Apalachicola River (the stem of the Y)? (This is the basic question before us.) Fifth, if so, would the amount of extra water that reaches the Apalachicola River significantly redress the economic and ecological harm that Florida has suffered? (This question is mostly for remand.) As our parentheticals suggest, the Special Master assumed that the answer to the first three questions was “yes.” The fourth question is the question before us now. And the fifth question is partly for us now and partly for the Master to answer on remand. A The Report indicates that the Special Master assumed the answer to the first question is “yes.” The Report says that the Special Master reached his conclusion on the “single, discrete issue that resolves this case” by “assuming that Florida has sustained injury.” Id., at 30 (emphasis added); see also id., at 2 (repeating Georgia’s argument that “without an order binding the Corps, Florida will not be assured any relief—assuming it has suffered any injury at all—by a decree entered in this proceeding because the Corps has the ability to impound water in various reservoirs that it maintains in the Basin” (emphasis added)); id., at 65 (“Even if there were evidence of harm from other than low-flow conditions . . . ”). At the same time, the Report states that “Florida points to real harm.” Id., at 31. And the Master specified that there is “little question that Florida has suffered harm from decreased flows in the [Apalachicola] River.” Id., at 31 (emphasis added). That harm—caused (at least in part) by increased salinity—includes “an unprecedented collapse of [Florida’s] oyster fisheries in 2012.” Ibid.; see id., at 32 (stating that “the evidence presented tends to show that increased salinity . . . led to the collapse” of Apalachicola Bay’s oysters and “greatly harmed the oystermen of the Apalachicola Region, threatening their longterm sustainability”). Cf. New Jersey v. New York, 283 U. S., at 343, 345 (finding redressable harm to oysters caused by diminished water flow and increased salinity). The harms of reduced streamflow may extend to other species in the Apalachicola Region, including in the river and its floodplain, which, as the Master noted, “is home to the highest species density of amphibians and reptiles in all of North America, and supports hundreds of endangered or threatened animal and plant species,” including three “endangered” or “threatened” mussel species, the “[t]hreatened Gulf sturgeon,” and the largest stand of Tupelo trees—of Tupelo Honey fame—in the world. Report 7–8; see also Joint Exh. 168, at 193, 195–196. B The Master also appears to have assumed the answer to the second question is “yes.” The Report reached its key conclusion that Florida’s (assumed) injuries cannot “effectively be redressed” by “assuming that Florida has sustained injury as a result of unreasonable upstream water use by Georgia.” Report 30 (emphasis added). But, at the same time, the Master acknowledged that “Florida points to real harm and, at the very least, likely misuse of resources by Georgia.” Id., at 31 (emphasis added). And the Report “provide[s] the Court a brief descriptive background regarding . . . the unreasonableness of Georgia’s consumptive water use.” Ibid.; see, e.g., id., at 32 (“Georgia’s upstream agricultural water use has been—and continues to be—largely unrestrained”); id., at 33 (“Despite early warnings of oncoming drought, Georgi[a] . . . chose not to declare a drought in 2011—apparently hoping for the best, and clearly not wishing to incur the cost of preventative action”); id., at 34 (“Georgia’s position—practically, politically, and legally—can be summarized as follows: Georgia’s agricultural water use should be subject to no limitations, regardless of the long-term consequences for the Basin”). C In respect to the third question, the Master again assumed the answer “yes.” In particular, the Report “assume[s]” that “Florida has sustained injury as a result of unreasonable upstream water use by Georgia.” Id., at 30 (emphasis added). And as relevant to each of the first three questions, the Master added that “[m]uch more could be said and would need to be said about” Florida’s injuries, the reasonableness of Georgia’s water consumption, and “other issues, such as causation,” if the case proceeds. Id., at 34. As we have explained, our prior equitable apportionment decisions make clear that “all factors which create equities in favor of one State or the other must be weighed.” Kansas II, 320 U. S., at 393–394 (emphasis added). Thus, a remand is necessary to consider each of the relevant factors, including those upon which the dissent focuses. See infra, at 27; Nebraska v. Wyoming, 325 U. S., at 618; cf. Colorado II, 467 U. S., at 323–324. D We now turn to the fourth question, the basic question before us. Would an equity-based cap on Georgia’s use of the Flint River lead to a significant increase in streamflow from the Flint River into Florida’s Apalachicola River (the stem of the Y)? The answer depends upon (1) the amount of extra water that would flow into Lake Seminole as a result of a cap on Georgia’s Flint River water consumption; and (2) the amount of water that could actually flow through the Corps-controlled Woodruff Dam at Lake Seminole’s southern end and into Florida’s Apalachicola River. 1 The record shows that Florida’s proposed cap on Georgia’s water consumption could result in the release of considerable extra water into Lake Seminole. Florida’s expert, Dr. David Sunding, testified that the cap would limit the average amount of water that Georgia could use annually and also reduce the amount of water that Georgia could use during drought years, which could “materially reduce [Georgia’s] depletions of river flows . . . by 1,500 to over 2,000 cubic feet per second (cfs) in peak summer months of drought years.” Updated Pre-Filed Direct Testimony (PFDT) of Sunding ¶8; see also id., ¶¶88–90. Dr. Sunding added that it would cost Georgia roughly $35 million annually (less than 0.2% of Georgia’s annual bud- get) to reduce streamflow depletions by 2,000 cfs. Id., ¶113, Table 4. Georgia’s expert, Dr. Robert Stavins, disputed these conclusions. See Direct Testimony of Stavins ¶¶4, 90, 136; see also Brief for Georgia 18. The Master did not make specific findings of fact regarding this aspect of Florida’s proposed remedy. Rather than expressly making any findings, the Master apparently “accept[ed] Florida’s estimates of the increased streamflow that would result from a consumption cap.” Report 67, n. 43. At this stage, we shall do the same. And as we shall later discuss, the record suggests that an increase in streamflow of 1,500 to 2,000 cfs is reason- ably likely to benefit Florida significantly. See infra, at 39–40 (citing record evidence of benefits); see also Updated PFDT of J. David Allan ¶¶3d, 26, 67 (Allan) (discussing ecological benefits of increasing streamflow by 300 to 500 cfs); 10 Tr. 2629:7–15 (Kondolf) (detailing benefits of increasing streamflow into the Apalachicola River from 5,000 to 7,000 cfs); 3 id., at 591:6–593:4, 596:17–598:1 (Allan). 2 The key question, however, is whether the 1,500 to 2,000 cfs of extra water that will flow into Lake Seminole from the Flint River as a result of a cap on Georgia’s water consumption will flow beyond Lake Seminole, through the Woodruff Dam, and into the Apalachicola River at the relevant times. That is where the Army Corps of Engineers enters the picture. And it is where Florida disagrees with the Special Master and with Georgia. The Special Master and Georgia believe that—at any relevant time—the Corps might “offset” any extra Flint River water that flows into Lake Seminole by simultaneously reducing the amount of water that flows into that lake from the Chattahoochee River. See Report 48–53. Thus, if the 1,500 to 2,000 cfs of extra water that would reach Lake Seminole from the Flint as a result of Florida’s proposed consumption cap, the question is whether and to what extent the Corps will “offset” that extra streamflow by releasing 1,500 to 2,000 cfs less water into Lake Seminole from its upstream Chattahoochee reservoirs. Of course, the Corps might, under certain circumstances, be authorized to “offset” extra streamflow from the Flint River. As the Special Master wrote, “[t]here is no guarantee that the Corps will exercise its discretion to release or hold back water at a particular time.” Id., at 69. But as the United States has explained, increased streamflow into Lake Seminole (that is, increased Basin Inflow) “would generally benefit the ACF system by delaying the onset of drought operations, by allowing the Corps to meet the 5000 cfs minimum flow longer during extended drought, and by quickening the resumption of normal operations after drought.” Brief for United States as Amicus Curiae 28 (Aug. 7, 2017). And our reading of the record convinces us it is highly unlikely that the Corps will always reduce the flow in this way; it leads us to believe that, acting in accordance with the its own revised Master Manual, the Corps is likely to permit, and in some cases may be required to ensure that, material amounts of additional Flint water to flow through the Woodruff Dam and into the Apalachicola River. At the very least, we believe that more proceedings are necessary to reach a definitive determination. As an initial matter, the Master Manual makes clear that the amount of water the Corps will release turns in part on the amount of water stored in the Corps’ Chattahoochee reservoirs. See U. S. Army Corps of Engineers, Master Manual, Apalachicola-Chattahoochee-Flint River Basin, Florida and Georgia, App. A, pp. 7–4 to 7–5, 7–7. More specifically, the amount of water storage in those reservoirs dictates whether the Corps is conducting one of two possible types of “operations”—namely, “drought operations” or “nondrought operations.” These are technical terms. See id., at 7–14 to 7–16. The term “drought operations” need not correspond to dry periods, nor need the term “nondrought operations” refer to wet periods. Rather their applicability depends in part upon the amount of water that is stored behind the Corps’ Chattahoochee dams. As the United States explained, “[t]he term ‘drought operations’ refers to more conservative operations that [the Corps conducts, which] are intended to enable the Corps to preserve water and operate its reservoir projects more effectively as drought conditions arise.” Brief for United States as Amicus Curiae 9 (Aug. 7, 2017). We therefore must clearly distinguish what the record tells us about the amount of extra water that could flow into Florida as a result of a consumption cap during each of these two distinct types of Corps operations. a Nondrought Operations When the Corps is conducting “nondrought operations,” the Master Manual requires the Corps to release into Florida all or some of any extra water that flows from the Flint River into Lake Seminole, where it will then flow through the Woodruff Dam. See App. to Brief for United States as Amicus Curiae 2a (Aug. 7, 2017) (detailing Corps operational protocol). As the United States has explained, when the total streamflow into Lake Seminole is between 5,000 and 10,000 cfs during “nondrought operations,” the following facts are true: “[A]ny additional basin inflow . . . would generally be passed straight through to Florida. If, for example, the conservation measures advocated by Florida as part of a consumption cap actually resulted in an increased flow in the Flint River of 2,000 cfs, see Pre-Filed Direct Testimony of David Sunding, Ph. D. at 44, Table 4, then flows into Florida would also increase by roughly that amount.” United States Post-Trial Brief 12–13 (Dec. 15, 2016); see also Brief for United States as Amicus Curiae 18 (Aug. 7, 2017) (reaffirming that under these circumstances “flows in the Apalachicola would increase by the amount of increased Flint River flows” including during summer months). As far as we can tell, under the Corps’ current operational protocol, the Corps may remain in “nondrought operations” even during the driest summer months of the driest years. For example, in 2007 the Corps conducted “nondrought operations” not only during late autumn, winter, and spring months, but also during the hottest summer and early autumn months “when streamflow is at its lowest.” See Direct Testimony of Phillip Bedient ¶¶48–53 (stating that “[i]f 2007’s Basin Inflow were repeated today and Drought Operations were not triggered,” the Corps would have had 92 days of “nondrought operations,” including 19 days “during summer and fall months, when streamflow was at its lowest” on which 100% of extra water resulting from a consumption cap would reach Florida). We note that these 19 days fell during a period of severe drought in which no extra water (let alone 2,000 cfs of extra water) was flowing into Lake Seminole. And, unsurprisingly, the same trend appears to be true in dry summer months of other years: all or some of the extra water that would result from a consumption cap would also pass through to Florida. See, e.g., Ga. Exh. 949 (reporting streamflow data indicating several days in 2009 on which extra Flint River water would have passed through to Florida); Joint Exh. 128 (providing link to U. S. Geological Survey data indicating a similar trend based on streamflow into the Apalachicola River, including in 2016 and 2017). b Drought Operations The Corps’ “drought operations” are different. Again, whether the Corps must initiate drought operations is not a matter of discretion; it depends, as we have said, upon the total amount of water the Corps has stored behind the dams it controls along the Chattahoochee River. The Master Manual requires that, when the total amount of water stored in pools behind the Corps’ Chattahoochee dams drops below a certain level, the Corps must reduce the amount of water it releases from the Woodruff Dam to 5,000 cfs, or, in instances of extreme low water levels in the storage pools, to 4,500 cfs. Master Manual App. A, at 7–14 to 7–16. Accordingly, if additional water were to flow into Lake Seminole from the Flint River while the Corps is in drought operations, the Corps, pursuant to its Master Manual, must reduce the flow of its controlled upstream Chattahoochee water in order to maintain a defined water level in the pools behind its Chattahoochee dams, and no more than 4,500 cfs or 5,000 cfs can flow beyond the Woodruff Dam regardless. Brief for United States as Amicus Curiae 7. But even then, as we just said, the Corps must make certain that at least 4,500 cfs and more often 5,000 cfs flows though the Woodruff Dam. And, if more water flows from the Flint into Lake Seminole, and if the Corps uses that water to keep the water level high in its Chattahoochee reservoirs, then there will be fewer days in which the Corps is conducting either “drought operations” or “extreme drought operations.” Instead, there will be more “nondrought operations” days where the Corps must pass most or all additional streamflow that exceeds 5,000 cfs through the Woodruff (because there will be more days, given the added Flint water, when its upstream Chattahoochee reservoirs are sufficiently high). The United States adds that “a cap on Georgia’s consumption” could, among other things, generate increased streamflow that “would provide a cushion during low-flow periods, so that it would be possible to maintain a flow rate of greater than 5,000 cfs for a longer period of time without any alteration of the Corps’ operations.” United States Post-Trial Brief 18–19 (Dec. 15, 2016) (emphasis added); see also Brief for United States as Amicus Curiae 18 (Aug. 7, 2017) (same). We repeat this point with an example for purposes of clarity. Assume the following: (1) that it is August 13 and the Corps is conducting “drought operations”; (2) that as a result of a cap on Georgia’s consumption, 2,000 cfs more water flows down the Flint and into Lake Seminole; and (3) that, consistent with the Master Manual, 5,000 cfs will flow from Lake Seminole, through the Woodruff Dam, and into Florida’s Apalachicola River. On these three assumptions in all likelihood, as the dissent points out, no extra water will flow into Florida. But (and this “but” is key), the extra 2,000 cfs of water that flows into Lake Seminole on August 13 as a result of a cap on Georgia’s from the Flint River water consumption will allow the Corps to store more water behind its upstream Chattahoochee dams (while still complying with the Master Manual’s minimum release requirements). And that fact means that the Corps is likely to remain in “drought operations” for fewer days because whether the Corps remains in “drought operations” depends upon the water level behind the Chattahoochee dams. And the fewer days the Corps conducts “drought operations,” the more days the Corps, consistent with its Master Manual, will allow all (or some) of the 2,000 cfs extra water that would result from a consumption cap to flow through the Woodruff Dam and into Florida’s Apalachicola River. Again, record evidence makes clear that this is not a fanciful possibility. For example, Florida points to record evidence that suggests a consumption cap could have prevented the Corps from entering drought operations in 2011–2012 without departing from the terms of its Master Manual. See, e.g., Florida Brief in Support of Exceptions 48–49, and n. 12 (citing record evidence, including Ga. Exh. 924 and Fla. Exh. 811, that the Special Master did not address suggesting that Florida’s proposed consumption cap could have helped the Corps to “avoi[d] drought operations entirely” in 2011–2012 without departing from the Master Manual’s requirements). The upshot is that, even when the Corps conducts its operations in accordance with the Master Manual, Florida’s proposed consumption cap would likely mean more water in the Apalachicola—as much as 2,000 cfs more water when the Corps is conducting normal or “nondrought operations,” which could take place in dry periods, including the driest days of summer, and 500 cfs more on days when the Corps is conducting “drought operations.” And a cap would likely allow the Corps to conduct “nondrought operations” (i.e., reservoirs-sufficiently-full operations) more often as well. 3 We cannot agree with the dissent’s efforts to deny these conclusions. To begin with, the dissent says that our conclusion “depends on the premise that, during droughts, the natural streamflow into Florida is between ‘5,000 and 10,000 cubic feet per second.’ ” Post, at 29. If the dissent means by “droughts” simply dry days, or summer days, then it is obviously wrong, for pursuant to the Corps’ Master Manual, the Corps must allow all or some of the 2,000 cfs extra water that would flow into Lake Seminole to continue through the Woodruff Dam into Florida during dry summer days when the Corps is not conducting “drought operations.” This was true, as the dissent concedes, even during 19 summer days in 2007, which was among the driest years in the Basin’s history. Or, does the dissent mean by “droughts” days on which the Corps is conducting “drought operations”? If so, then we agree that on such days, the Corps will normally allow no more than 5,000 cfs to flow into Florida. But, for the reasons just stated in the last few paragraphs, Florida’s proposed consumption cap—which could result in as much as 2,000 extra cubic feet of water per second flowing from the Flint into Lake Seminole—will mean (consistent with the testimony of the very Georgia expert that the dissent so frequently quotes) that there will be significantly fewer such days. Is there a mistake then in the “concrete example” the dissent offers to support its point? See post, at 29–30. Invoking a hypothetical posed by Georgia’s expert, the dissent says: “[I]f the natural flows in the Apalachicola River were 2,600 cubic feet per second, then the Corps would release 2,400 cubic feet per second from its [Chattahoochee] reservoirs . . . . And if a cap on Georgia[ ’s Flint River consumption] increased the River’s natural flow to 4,100 cubic feet per second, the Corps would release 900 cubic feet per second. . . . In either case, the total flow on the Apalachicola River would remain the same: 5,000 cubic feet per second. Thus, so long as the natural flows remain significantly less than 5,000 cubic feet per second, a cap on Georgia would only decrease the amount of water that the Corps releases from storage; it would not increase the overall amount of water flowing into the Apalachicola River.” Id., at 29–30 (citing Bedient ¶¶45–47). If, however, a consumption cap causes 1,500 cfs extra water (from the Flint) to flow into Lake Seminole (as we assume Florida’s proposed cap would), under the dissent’s example, the Corps will reduce (or “offset”) the amount of water it releases from its upstream Chattahoochee dams from 2,400 cfs to 900 cfs. That is because 2,400 cfs minus 900 cfs is 1,500 cfs. What happens to that 1,500 cfs extra water? When the Corps is in drought operations, the answer according to the Master Manual is that the Corps must store that water in its upstream Chattahoochee reservoirs. And with that 1,500 cfs extra water each day, the water levels in those reservoirs will rise (or, at a minimum, deplete less rapidly) and allow the Corps to resume “nondrought operations” more quickly. The United States repeats precisely this point—namely, when more water flows into Lake Seminole, it benefits Florida by “quickening the [Corps’] resumption of normal [i.e., “nondrought”] operations.” Brief for United States as Amicus Curiae 28 (Aug. 7, 2017). (That extra water also means that there will be more days when 5,000 cfs, rather than 4,500 cfs, flows from Lake Seminole into the Apalachicola River). And it means, as no one denies, that on days when the Corps conducts “nondrought operations” (which, as Georgia’s own expert report shows, occur even during dry summer months), more water will reach Florida when Florida needs it. What about the dissent’s point that Georgia’s expert, Dr. Bedient, said that the extra 2,000 cfs would mean more water for Florida “only 19 days ‘during the summer and fall months when streamflow was at its lowest’ ”? Post, at 30. Dr. Bediant’s exact words, as the dissent points out, were that in “ ‘dry years (e.g., 2007 and 2011), . . . even significant changes in Georgia’s consumptive use would lead to virtually no change in state-line flows during the low-flow months (e.g., June, July, August, September).’ ” Bedient ¶¶48–53. At this point, in our view, the dissent has pointed to record evidence with which other record evidence conflicts. It seems from record evidence, from the statements of the United States, from geological data, and from laws of mechanics, that 2,000 cfs extra water flowing into Lake Seminole when, in the dissent’s words, “drought operations were not in effect” would have to mean more water in Florida. Post, at 30. And the dissent does not dispute that some of these days are in the summer. Ibid. Our own check of the record reinforces the point. In particular, data from the U. S. Geological Survey’s website, which the parties entered into the record at Joint Exh. 128, indicates that between May 2016 and August 2016, streamflow into the Apalachicola River was above 6,000 cfs each day with the exception of two days: August 30, 2016 and August 31, 2016. Nothing in the record suggests that the Corps was in drought operations during these days, and so it appears that under these conditions, any additional streamflow resulting from a cap on Georgia’s Flint River consumption would pass through into Florida. However, without explicit findings, it is neither possible nor prudent for us in the first instance to read through this voluminous record and discover who is right on this matter of how much extra water there will be, when, and how much Florida would benefit from the extra water that there might be. That is why we are sending this case back for more findings. Finally, while the dissent suggests that “[i]t is incredibly odd to conclude that a Special Master’s merits determination is ‘premature’ after a full trial,” post, at 17, this Court has repeatedly concluded that remand is “appropriate” to resolve certain issues in an equitable apportionment case even where, as here, there has already been a “lengthy trial at which both States presented extensive evidence.” Colorado II, 467 U. S., at 313; see also Wyoming v. Colorado, 259 U. S., at 456–457 (explaining that “the evidence was taken” over the course of two years and presented to the Court two years later and that “[t]he case has been argued at bar three times” including because of the “importance of some of the questions involved”). Moreover, we note that adequate factfinding is especially important where, as here, no interstate compact guides our inquiry or sets forth a congressionally ratified water allocation formula. When such a compact exists, as it often does, our effort is relatively simple and focuses upon “declar[ing] rights under the Compact and enforc[ing] its terms.” Kansas v. Nebraska, 574 U. S., at ___ (slip op., at 8) (citing Texas v. New Mexico, 462 U. S., at 567); id., at 567–568 (“If there is a compact, it is a law of the United States, and our first and last order of business is interpreting the compact”). Here, no compact guides our inquiry and it would appear to be important that we approach this complex controversy with the care and thoroughness that our precedent requires. E Our final question is this: Would the amount of extra water that reaches the Apalachicola significantly redress the economic and ecological harm that Florida has suffered? There is evidence indicating that the answer to the question is in the affirmative. See, e.g., Allan ¶3d, 26, 67 (“Even relatively modest increases in flows—on the order of 300 to 500 cfs during key periods of the year—could reduce harm to the [Apalachicola Region’s] ecosystem and halt the cycle that is leading to irreversible harm” while “[g]reater increases could make even more dramatic improvements”); Updated PFDT of Patricia Glibert ¶¶5, 28–32, 58–60, and Table 1, Figs. 10, 19b; supra, at 21–22 (citing record evidence of benefits); see also 10 Tr. 2629:7–15 (Kondolf) (detailing benefits of increasing streamflow from 5,000 to 7,000 cfs); 3 id., at 591:6–593:4, 596:17–598:1 (Allan). But the Master’s Report does not explicitly answer this question. We consequently must remand the case to find the answer to this question (and others). * * * In sum, in respect to the evidentiary questions at issue, the Master assumed that: (1) Florida has likely suffered harm as a result of decreased water flow into the Apalachicola River; (2) Florida has made some showing that Georgia, contrary to equitable principles, has taken too much water from the Flint River; and (3) Georgia’s inequitable use of the water may have injured Florida, but more findings are needed. And in light of the Master’s assumptions, we conclude that: (4) an equity-based cap on Georgia’s use of the Flint River would likely lead to a material increase in streamflow from the Flint River into Florida’s Apalachicola River; and (5) the amount of extra water that reaches the Apalachicola may significantly redress the economic and ecological harm that Florida has suffered. Further findings, however, are needed on all of these evidentiary issues on remand. We add the following: The United States has made clear that the Corps will work to accommodate any determinations or obligations the Court sets forth if a final decree equitably apportioning the Basin’s waters proves justified in this case. It states in its brief here that if a decree results “in more water flowing to Florida . . . under existing Corps protocols, then the Corps would likely not need to change its operations.” Brief for United States as Amicus Curiae 28 (Aug. 7, 2017). It has added that, in any event, a decree “would necessarily form part of the constellation of laws to be considered by the Corps when deciding how best to operate the federal projects.” Id., at 32. And in issuing its revised Master Manual, the Corps stated that it would “review any final decision from the U. S. Supreme Court and consider any operational adjustments that are appropriate in light of that decision, including modifications to the then-existing [Master Manual], if applicable.” Record of Decision 18. The United States has “continually asserted its preparedness to implement, in accordance with federal law, any [agreed-upon] comprehensive water allocation formula.” Id., at 4; see also Joint Exh. 124, at 6–35. And, of course, the Administrative Procedure Act requires the Corps to make decisions that are reasonable, i.e., not “arbitrary, capricious, an abuse of dis- cretion” or “in excess of [the Corps’] statutory jurisdiction.” 5 U. S. C. §706(2). We recognize that the Corps must take account of a variety of circumstances and statutory obligations when it allocates water. New circumstances may require the Corps to revise its Master Manual or devote more water from the Chattahoochee River to other uses. But, given the considerations we have set forth, we cannot agree with the Special Master that the Corps’ “inheren[t] discretio[n]” renders effective relief impermissibly “uncertain” or that meaningful relief is otherwise precluded. Report 56, n. 38. We cannot now say that Florida has “merely some technical right” without “a corresponding benefit,” Kansas I, 206 U. S., at 109, or that an effort to shape a decree will prove “a vain thing.” Foster, 146 U. S., at 101. Ordinarily “[u]ncertainties about the future” do not “provide a basis for declining to fashion a decree.” See Idaho II, 462 U. S., at 1026. And in this case, the record leads us to believe that, if necessary and with the help of the United States, the Special Master, and the parties, we should be able to fashion one. V We keep in mind what our prior decisions make clear: “ ‘The difficulties of drafting and enforcing a decree’ ” do not necessarily provide a convincing “ ‘justification for us to refuse to perform the important function entrusted to us by the Constitution.’ ” Idaho I, 444 U. S., at 390, n. 7 (quoting Nebraska v. Wyoming, 325 U. S., at 616); see also Idaho II, supra, at 1027 (“Although the computation is complicated and somewhat technical, that fact does not prevent the issuance of an equitable decree”). For this reason and the others we have discussed, we agree with Florida that it has made a legally sufficient showing as to the possibility of fashioning an effective remedial decree. We repeat, however, that Florida will be entitled to a decree only if it is shown that “the benefits of the [apportionment] substantially outweigh the harm that might result.” Colorado I, 459 U. S., at 187. In assessing whether that showing has been made, the Master may find it necessary to address in the first instance many of the evidentiary and legal questions the answers to which we have here assumed or found plausible enough to allow us to resolve the threshold remedial question. In order to determine whether Florida can eventually prove its right to cap Georgia’s use of Flint River waters, it may find it necessary for the Special Master to make more specific factual findings and definitive recommendations regarding such questions as: To what extent does Georgia take too much water from the Flint River? To what extent has Florida sustained injuries as a result? To what extent would a cap on Georgia’s water consumption increase the amount of water that flows from the Flint River into Lake Seminole? To what extent (under the Corps’ revised Master Manual or under reasonable modifications that could be made to that Manual) would additional water resulting from a cap on Georgia’s water consumption result in additional streamflow in the Apalachicola River? To what extent would that additional streamflow into the Apalachicola River ameliorate Florida’s injuries? The Special Master may make other factual findings he believes necessary and hold hearings (or take additional evidence) as he believes necessary. Cf. Colorado I, 459 U. S., at 190, n. 14. Consistent with the principles that guide our inquiry in this context, answers need not be “mathematically precise or based on definite present and future conditions.” Id., at 1026. Approximation and reasonable estimates may prove “necessary to protect the equitable rights of a State.” Ibid. And the answers may change over time. Cf. New Jersey v. New York, 347 U.S. 995, 996–1005 (1954); New Jersey v. New York, 283 U. S., at 344–346. Flexibility and approximation are often the keys to success in our efforts to resolve water disputes between sovereign States that neither Congress “nor the legislature of either State” has been able to resolve. Virginia v. West Virginia, 220 U. S., at 27. We consequently do not dismiss this case. Rather, we remand the case to the Special Master for further proceedings consistent with this opinion. It is so ordered. APPENDIX

585.US.2017_16-1161

Members of the Wisconsin Legislature are elected from single-member legislative districts. Under the Wisconsin Constitution, the legislature must redraw the boundaries of those districts following each census. After the 2010 census, the legislature passed a new districting plan known as Act 43. Twelve Democratic voters, the plaintiffs in this case, alleged that Act 43 harms the Democratic Party’s ability to convert Democratic votes into Democratic seats in the legislature. They asserted that Act 43 does this by “cracking” certain Democratic voters among different districts in which those voters fail to achieve electoral majorities and “packing” other Democratic voters in a few districts in which Democratic candidates win by large margins. The plaintiffs argued that the degree to which packing and cracking has favored one political party over another can be measured by an “efficiency gap” that compares each party’s respective “wasted” votes—i.e., votes cast for a losing candidate or for a winning candidate in excess of what that candidate needs to win—across all legislative districts. The plaintiffs claimed that the statewide enforcement of Act 43 generated an excess of wasted Democratic votes, thereby violating the plaintiffs’ First Amendment right of association and their Fourteenth Amendment right to equal protection. The defendants, several members of the state election commission, moved to dismiss the plaintiffs’ claims. They argued that the plaintiffs lacked standing to challenge the constitutionality of Act 43 as a whole because, as individual voters, their legally protected interests extend only to the makeup of the legislative district in which they vote. The three-judge District Court denied the defendants’ motion and, following a trial, concluded that Act 43 was an unconstitutional partisan gerrymander. Regarding standing, the court held that the plaintiffs had suffered a particularized injury to their equal protection rights. Held: The plaintiffs have failed to demonstrate Article III standing. Pp. 8–22. (a) Over the past five decades this Court has repeatedly been asked to decide what judicially enforceable limits, if any, the Constitution sets on partisan gerrymandering. Previous attempts at an answer have left few clear landmarks for addressing the question and have generated conflicting views both of how to conceive of the injury arising from partisan gerrymandering and of the appropriate role for the Federal Judiciary in remedying that injury. See Gaffney v. Cummings, 412 U. S. 735, Davis v. Bandemer, 478 U. S. 109, Vieth v. Jubelirer, 541 U. S. 267, and League of United Latin American Citizens v. Perry, 548 U. S. 399. Pp. 8–12. (b) A plaintiff may not invoke federal-court jurisdiction unless he can show “a personal stake in the outcome of the controversy,” Baker v. Carr, 369 U. S. 186, 204. That requirement ensures that federal courts “exercise power that is judicial in nature,” Lance v. Coffman, 549 U. S. 437, 439, 441. To meet that requirement, a plaintiff must show an injury in fact—his pleading and proof that he has suffered the “invasion of a legally protected interest” that is “concrete and particularized,” i.e., which “affect[s] the plaintiff in a personal and individual way.” Lujan v. Defenders of Wildlife, 504 U. S. 555, 560, and n. 1. The right to vote is “individual and personal in nature,” Reynolds v. Sims, 377 U. S. 533, 561, and “voters who allege facts showing disadvantage to themselves as individuals have standing to sue” to remedy that disadvantage, Baker, 369 U. S., at 206. The plaintiffs here alleged that they suffered such injury from partisan gerrymandering, which works through the “cracking” and “packing” of voters. To the extent that the plaintiffs’ alleged harm is the dilution of their votes, that injury is district specific. An individual voter in Wisconsin is placed in a single district. He votes for a single representative. The boundaries of the district, and the composition of its voters, determine whether and to what extent a particular voter is packed or cracked. A plaintiff who complains of gerrymandering, but who does not live in a gerrymandered district, “assert[s] only a generalized grievance against governmental conduct of which he or she does not approve.” United States v. Hays, 515 U. S. 737, 745. The plaintiffs argue that their claim, like the claims presented in Baker and Reynolds, is statewide in nature. But the holdings in those cases were expressly premised on the understanding that the injuries giving rise to those claims were “individual and personal in nature,” Reynolds, 377 U. S., at 561, because the claims were brought by voters who alleged “facts showing disadvantage to themselves as individuals,” Baker, 369 U. S., at 206. The plaintiffs’ mistaken insistence that the claims in Baker and Reynolds were “statewide in nature” rests on a failure to distinguish injury from remedy. In those malapportionment cases, the only way to vindicate an individual plaintiff’s right to an equally weighted vote was through a wholesale “restructuring of the geographical distribution of seats in a state legislature.” Reynolds, 377 U. S., at 561. Here, the plaintiffs’ claims turn on allegations that their votes have been diluted. Because that harm arises from the particular composition of the voter’s own district, remedying the harm does not necessarily require restructuring all of the State’s legislative districts. It requires revising only such districts as are necessary to reshape the voter’s district. This fits the rule that a “remedy must of course be limited to the inadequacy that produced the injury in fact that the plaintiff has established.” Lewis v. Casey, 518 U. S. 343, 357. The plaintiffs argue that their legal injury also extends to the statewide harm to their interest “in their collective representation in the legislature,” and in influencing the legislature’s overall “composition and policymaking.” Brief for Appellees 31. To date, however, the Court has not found that this presents an individual and personal injury of the kind required for Article III standing. A citizen’s interest in the overall composition of the legislature is embodied in his right to vote for his representative. The harm asserted by the plaintiffs in this case is best understood as arising from a burden on their own votes. Pp. 12–17. (c) Four of the plaintiffs in this case pleaded such a particularized burden. But as their case progressed to trial, they failed to pursue their allegations of individual harm. They instead rested their case on their theory of statewide injury to Wisconsin Democrats, in support of which they offered three kinds of evidence. First, they presented testimony pointing to the lead plaintiff’s hope of achieving a Democratic majority in the legislature. Under the Court’s cases to date, that is a collective political interest, not an individual legal interest. Second, they produced evidence regarding the mapmakers’ deliberations as they drew district lines. The District Court relied on this evidence in concluding that those mapmakers sought to understand the partisan effect of the maps they were drawing. But the plaintiffs’ establishment of injury in fact turns on effect, not intent, and requires a showing of a burden on the plaintiffs’ votes that is “actual or imminent, not ‘conjectural’ or ‘hypothetical.’ ” Defenders of Wildlife, 504 U. S., at 560. Third, the plaintiffs presented partisan-asymmetry studies showing that Act 43 had skewed Wisconsin’s statewide map in favor of Republicans. Those studies do not address the effect that a gerrymander has on the votes of particular citizens. They measure instead the effect that a gerrymander has on the fortunes of political parties. That shortcoming confirms the fundamental problem with the plaintiffs’ case as presented on this record. It is a case about group political interests, not individual legal rights. Pp. 17–21. (d) Where a plaintiff has failed to demonstrate standing, this Court usually directs dismissal. See, e.g., DaimlerChrysler Corp. v. Cuno, 547 U. S. 332, 354. Here, however, where the case concerns an unsettled kind of claim that the Court has not agreed upon, the contours and justiciability of which are unresolved, the case is remanded to the District Court to give the plaintiffs an opportunity to prove concrete and particularized injuries using evidence that would tend to demonstrate a burden on their individual votes. Cf. Alabama Legislative Black Caucus v. Alabama, 575 U. S. ___, ___. Pp. 21–22. 218 F. Supp. 3d 837, vacated and remanded. Roberts, C. J., delivered the opinion of the Court, in which Kennedy, Ginsburg, Breyer, Alito, Sotomayor, and Kagan, JJ., joined, and in which Thomas and Gorsuch, JJ., joined except as to Part III. Kagan, J., filed a concurring opinion, in which Ginsburg, Breyer, and Sotomayor, JJ., joined. Thomas, J., filed an opinion concurring in part and concurring in the judgment, in which Gorsuch, J., joined.

The State of Wisconsin, like most other States, entrusts to its legislature the periodic task of redrawing the boundaries of the State’s legislative districts. A group of Wisconsin Democratic voters filed a complaint in the District Court, alleging that the legislature carried out this task with an eye to diminishing the ability of Wisconsin Democrats to convert Democratic votes into Democratic seats in the legislature. The plaintiffs asserted that, in so doing, the legislature had infringed their rights under the First and Fourteenth Amendments. But a plaintiff seeking relief in federal court must first demonstrate that he has standing to do so, including that he has “a personal stake in the outcome,” Baker v. Carr, 369 U. S. 186, 204 (1962), distinct from a “generally available grievance about government,” Lance v. Coffman, 549 U. S. 437, 439 (2007) (per curiam). That threshold requirement “ensures that we act as judges, and do not engage in policymaking properly left to elected representatives.” Hollingsworth v. Perry, 570 U. S. 693, 700 (2013). Certain of the plaintiffs before us alleged that they had such a personal stake in this case, but never followed up with the requisite proof. The District Court and this Court therefore lack the power to resolve their claims. We vacate the judgment and remand the case for further proceedings, in the course of which those plaintiffs may attempt to demonstrate standing in accord with the analysis in this opinion. I Wisconsin’s Legislature consists of a State Assembly and a State Senate. Wis. Const., Art. IV, §1. The 99 members of the Assembly are chosen from single districts that must “consist of contiguous territory and be in as compact form as practicable.” §4. State senators are likewise chosen from single-member districts, which are laid on top of the State Assembly districts so that three Assembly districts form one Senate district. See §5; Wis. Stat. §4.001 (2011). The Wisconsin Constitution gives the legislature the responsibility to “apportion and district anew the members of the senate and assembly” at the first session following each census. Art. IV, §3. In recent decades, however, that responsibility has just as often been taken up by federal courts. Following the census in 1980, 1990, and 2000, federal courts drew the State’s legislative districts when the Legislature and the Governor—split on party lines—were unable to agree on new districting plans. The Legislature has broken the logjam just twice in the last 40 years. In 1983, a Democratic Legislature passed, and a Democratic Governor signed, a new districting plan that remained in effect until the 1990 census. See 1983 Wis. Laws ch. 4. In 2011, a Republican Legislature passed, and a Republican Governor signed, the districting plan at issue here, known as Act 43. See Wis. Stat. §§ 4.009, 4.01–4.99; 2011 Wis. Laws ch. 4. Following the passage of Act 43, Republicans won majorities in the State Assembly in the 2012 and 2014 elections. In 2012, Republicans won 60 Assembly seats with 48.6% of the two-party statewide vote for Assembly candidates. In 2014, Republicans won 63 Assembly seats with 52% of the statewide vote. 218 F. Supp. 3d 837, 853 (WD Wis. 2016). In July 2015, twelve Wisconsin voters filed a complaint in the Western District of Wisconsin challenging Act 43. The plaintiffs identified themselves as “supporters of the public policies espoused by the Democratic Party and of Democratic Party candidates.” 1 App. 32, Complaint ¶15. They alleged that Act 43 is a partisan gerrymander that “unfairly favor[s] Republican voters and candidates,” and that it does so by “cracking” and “packing” Democratic voters around Wisconsin. Id., at 28–30, ¶¶5–7. As they explained: “Cracking means dividing a party’s supporters among multiple districts so that they fall short of a majority in each one. Packing means concentrating one party’s backers in a few districts that they win by overwhelming margins.” Id., at 29, ¶5. Four of the plaintiffs—Mary Lynne Donohue, Wendy Sue Johnson, Janet Mitchell, and Jerome Wallace—alleged that they lived in State Assembly districts where Democrats have been cracked or packed. Id., at 34–36, ¶¶20, 23, 24, 26; see id., at 50–53, ¶¶60–70 (describing packing and cracking in Assembly Districts 22, 26, 66, and 91). All of the plaintiffs also alleged that, regardless of “whether they themselves reside in a district that has been packed or cracked,” they have been “harmed by the manipulation of district boundaries” because Democrats statewide “do not have the same opportunity provided to Republicans to elect representatives of their choice to the Assembly.” Id., at 33, ¶16. The plaintiffs argued that, on a statewide level, the degree to which packing and cracking has favored one party over another can be measured by a single calculation: an “efficiency gap” that compares each party’s respective “wasted” votes across all legislative districts. “Wasted” votes are those cast for a losing candidate or for a win- ning candidate in excess of what that candidate needs to win. Id., at 28–29, ¶5. The plaintiffs alleged that Act 43 resulted in an unusually large efficiency gap that favored Republicans. Id., at 30, ¶7. They also submitted a “Demonstration Plan” that, they asserted, met all of the legal criteria for apportionment, but was at the same time “almost perfectly balanced in its partisan consequences.” Id., at 31, ¶10. They argued that because Act 43 gener- ated a large and unnecessary efficiency gap in favor of Re- publicans, it violated the First Amendment right of association of Wisconsin Democratic voters and their Fourteenth Amendment right to equal protection. The plaintiffs named several members of the state election commission as defendants in the action. Id., at 36, ¶¶28–30. The election officials moved to dismiss the complaint. They argued, among other things, that the plaintiffs lacked standing to challenge the constitutionality of Act 43 as a whole because, as individual voters, their legally protected interests extend only to the makeup of the legislative districts in which they vote. A three-judge panel of the District Court, see 28 U. S. C. §2284(a), denied the defendants’ motion. In the District Court’s view, the plaintiffs “identif[ied] their injury as not simply their inability to elect a representative in their own districts, but also their reduced opportunity to be represented by Democratic legislators across the state.” Whitford v. Nichol, 151 F. Supp. 3d 918, 924 (WD Wis. 2015). It therefore followed, in the District Court’s opinion, that “[b]ecause plaintiffs’ alleged injury in this case relates to their statewide representation, . . . they should be permitted to bring a statewide claim.” Id., at 926. The case proceeded to trial, where the plaintiffs presented testimony from four fact witnesses. The first was lead plaintiff William Whitford, a retired law professor at the University of Wisconsin in Madison. Whitford testified that he lives in Madison in the 76th Assembly District, and acknowledged on cross-examination that this is, under any plausible circumstances, a heavily Democratic district. Under Act 43, the Democratic share of the Assembly vote in Whitford’s district is 81.9%; under the plaintiffs’ ideal map—their Demonstration Plan—the projected Democratic share of the Assembly vote in Whitford’s district would be 82%. 147 Record 35–36. Whitford therefore conceded that Act 43 had not “affected [his] ability to vote for and elect a Democrat in [his] district.” Id., at 37. Whitford testified that he had nevertheless suffered a harm “relate[d] to [his] ability to engage in campaign activity to achieve a majority in the Assembly and the Senate.” Ibid. As he explained, “[t]he only practical way to accomplish my policy objectives is to get a majority of the Democrats in the Assembly and the Senate ideally in order to get the legislative product I prefer.” Id., at 33. The plaintiffs also presented the testimony of legislative aides Adam Foltz and Tad Ottman, as well as that of Professor Ronald Gaddie, a political scientist who helped design the Act 43 districting map, regarding how that map was designed and adopted. In particular, Professor Gaddie testified about his creation of what he and the District Court called “S curves”: color-coded tables of the estimated partisan skew of different draft redistricting maps. See 218 F. Supp. 3d, at 850, 858. The colors corresponded with assessments regarding whether different districts tilted Republican or Democratic under various statewide political scenarios. The S curve for the map that was eventually adopted projected that “Republicans would maintain a majority under any likely voting scenario,” with Democrats needing 54% of the statewide vote to secure a majority in the legislature. Id., at 852. Finally, the parties presented testimony from four expert witnesses. The plaintiffs’ experts, Professor Kenneth Mayer and Professor Simon Jackman, opined that—according to their efficiency-gap analyses—the Act 43 map would systematically favor Republicans for the duration of the decade. See id., at 859–861. The defendants’ experts, Professor Nicholas Goedert and Sean Trende, opined that efficiency gaps alone are unreliable measures of durable partisan advantage, and that the political geography of Wisconsin currently favors Republicans because Democrats—who tend to be clustered in large cities—are inefficiently distributed in many parts of Wisconsin for purposes of winning elections. See id., at 861–862. At the close of evidence, the District Court concluded—over the dissent of Judge Griesbach—that the plaintiffs had proved a violation of the First and Fourteenth Amendments. The court set out a three-part test for identifying unconstitutional gerrymanders: A redistricting map violates the First Amendment and the Equal Protection Clause of the Fourteenth Amendment if it “(1) is intended to place a severe impediment on the effectiveness of the votes of individual citizens on the basis of their political affiliation, (2) has that effect, and (3) cannot be justified on other, legitimate legislative grounds.” Id., at 884. The court went on to find, based on evidence concerning the manner in which Act 43 had been adopted, that “one of the purposes of Act 43 was to secure Republican control of the Assembly under any likely future electoral scenario for the remainder of the decade.” Id., at 896. It also found that the “more efficient distribution of Republican voters has allowed the Republican Party to translate its votes into seats with significantly greater ease and to achieve—and preserve—control of the Wisconsin legislature.” Id., at 905. As to the third prong of its test, the District Court concluded that the burdens the Act 43 map imposed on Democrats could not be explained by “legitimate state prerogatives [or] neutral factors.” Id., at 911. The court recognized that “Wisconsin’s political geography, particularly the high concentration of Democratic voters in urban centers like Milwaukee and Madison, affords the Republican Party a natural, but modest, advantage in the districting process,” but found that this inherent geographic disparity did not account for the magnitude of the Republican advantage. Id., at 921, 924. Regarding standing, the court held that the plaintiffs had a “cognizable equal protection right against state-imposed barriers on [their] ability to vote effectively for the party of [their] choice.” Id., at 928. It concluded that Act 43 “prevent[ed] Wisconsin Democrats from being able to translate their votes into seats as effectively as Wisconsin Republicans,” and that “Wisconsin Democrats, therefore, have suffered a personal injury to their Equal Protection rights.” Ibid. The court turned away the defendants’ argument that the plaintiffs’ injury was not sufficiently particularized by finding that “[t]he harm that the plaintiffs have experienced . . . is one shared by Democratic voters in the State of Wisconsin. The dilution of their votes is both personal and acute.” Id., at 930. Judge Griesbach dissented. He wrote that, under this Court’s existing precedents, “partisan intent” to benefit one party rather than the other in districting “is not illegal, but is simply the consequence of assigning the task of redistricting to the political branches.” Id., at 939. He observed that the plaintiffs had not attempted to prove that “specific districts . . . had been gerrymandered,” but rather had “relied on statewide data and calculations.” Ibid. And he argued that the plaintiffs’ proof, resting as it did on statewide data, had “no relevance to any gerrymandering injury alleged by a voter in a single district.” Id., at 952. On that basis, Judge Griesbach would have entered judgment for the defendants. The District Court enjoined the defendants from using the Act 43 map in future elections and ordered them to have a remedial districting plan in place no later than November 1, 2017. The defendants appealed directly to this Court, as provided under 28 U. S. C. §1253. We stayed the District Court’s judgment and postponed consideration of our jurisdiction. 582 U. S. ___ (2017). II A Over the past five decades this Court has been repeat- edly asked to decide what judicially enforceable limits, if any, the Constitution sets on the gerrymandering of voters along partisan lines. Our previous attempts at an answer have left few clear landmarks for addressing the question. What our precedents have to say on the topic is, however, instructive as to the myriad competing considerations that partisan gerrymandering claims involve. Our efforts to sort through those considerations have generated conflicting views both of how to conceive of the injury arising from partisan gerrymandering and of the appropriate role for the Federal Judiciary in remedying that injury. Our first consideration of a partisan gerrymandering claim came in Gaffney v. Cummings, 412 U. S. 735 (1973). There a group of plaintiffs challenged the constitutionality of a Connecticut redistricting plan that “consciously and overtly adopted and followed a policy of ‘political fairness,’ which aimed at a rough scheme of proportional representation of the two major political parties.” Id., at 738. To that end, the redistricting plan broke up numerous towns, “wiggl[ing] and joggl[ing]” district boundary lines in order to “ferret out pockets of each party’s strength.” Id., at 738, and n. 3, 752, n. 18. The plaintiffs argued that, notwithstanding the rough population equality of the districts, the plan was unconstitutional because its consciously political design was “nothing less than a gigantic political gerrymander.” Id., at 752. This Court rejected that claim. We reasoned that it would be “idle” to hold that “any political consideration taken into account in fashioning a reapportionment plan is sufficient to invalidate it,” because districting “inevitably has and is intended to have substantial political consequences.” Id., at 752–753. Thirteen years later came Davis v. Bandemer, 478 U. S. 109 (1986). Unlike the bipartisan gerrymander at issue in Gaffney, the allegation in Bandemer was that Indiana Republicans had gerrymandered Indiana’s legislative districts “to favor Republican incumbents and candidates and to disadvantage Democratic voters” through what the plaintiffs called the “stacking” (packing) and “splitting” (cracking) of Democrats. 478 U. S., at 116–117 (plurality opinion). A majority of the Court agreed that the case before it was justiciable. Id., at 125, 127. The Court could not, however, settle on a standard for what constitutes an unconstitutional partisan gerrymander. Four Justices would have required the Bandemer plaintiffs to “prove both intentional discrimination against an identifiable political group and an actual discriminatory effect on that group.” Id., at 127. In that plurality’s view, the plaintiffs had failed to make a sufficient showing on the latter point because their evidence of unfavorable election results for Democrats was limited to a single election cycle. See id., at 135. Three Justices, concurring in the judgment, would have held that the “Equal Protection Clause does not supply judicially manageable standards for resolving purely political gerrymandering claims.” Id., at 147 (opinion of O’Connor, J.). Justice O’Connor took issue, in particular, with the plurality’s focus on factual questions concerning “statewide electoral success.” Id., at 158. She warned that allowing district courts to “strike down apportionment plans on the basis of their prognostications as to the outcome of future elections or future apportionments invites ‘findings’ on matters as to which neither judges nor anyone else can have any confidence.” Id., at 160. Justice Powell, joined by Justice Stevens, concurred in part and dissented in part. In his view, the plaintiffs’ claim was not simply that their “voting strength was diluted statewide,” but rather that “certain key districts were grotesquely gerrymandered to enhance the election prospects of Republican candidates.” Id., at 162, 169. Thus, he would have focused on the question “whether the boundaries of the voting districts have been distorted deliberately and arbitrarily to achieve illegitimate ends.” Id., at 165. Eighteen years later, we revisited the issue in Vieth v. Jubelirer, 541 U. S. 267 (2004). In that case the plaintiffs argued that Pennsylvania’s Legislature had created “meandering and irregular” congressional districts that “ignored all traditional redistricting criteria, including the preservation of local government boundaries,” in order to provide an advantage to Republican candidates for Congress. Id., at 272–273 (plurality opinion) (brackets omitted). The Vieth Court broke down on numerous lines. Writing for a four-Justice plurality, Justice Scalia would have held that the plaintiffs’ claims were nonjusticiable because there was no “judicially discernible and manageable standard” by which to decide them. Id., at 306. On those grounds, the plurality affirmed the dismissal of the claims. Ibid. Justice Kennedy concurred in the judgment. He noted that “there are yet no agreed upon substantive principles of fairness in districting,” and that, consequently, “we have no basis on which to define clear, manageable, and politically neutral standards for measuring the particular burden” on constitutional rights. Id., at 307–308. He rejected the principle advanced by the plaintiffs—that “a majority of voters in [Pennsylvania] should be able to elect a majority of [Pennsylvania’s] congressional delegation”—as a “precept” for which there is “no authority.” Id., at 308. Yet Justice Kennedy recognized the possibility that “in another case a standard might emerge that suit- ably demonstrates how an apportionment’s de facto incorporation of partisan classifications burdens” representational rights. Id., at 312. Four Justices dissented in three different opinions. Justice Stevens would have permitted the plaintiffs’ claims to proceed on a district-by-district basis, using a legal standard similar to the standard for racial gerrymandering set forth in Shaw v. Hunt, 517 U. S. 899 (1996). See 541 U. S., at 335–336, 339. Under this standard, any district with a “bizarre shape” for which the only possible explanation was “a naked desire to increase partisan strength” would be found unconstitutional under the Equal Protection Clause. Id., at 339. Justice Souter, joined by Justice Ginsburg, agreed that a plaintiff alleging unconstitutional partisan gerrymandering should proceed on a district-by-district basis, as “we would be able to call more readily on some existing law when we defined what is suspect at the district level.” See id., at 346–347. Justice Breyer dissented on still other grounds. In his view, the drawing of single-member legislative districts—even according to traditional criteria—is “rarely . . . politically neutral.” Id., at 359. He therefore would have distinguished between gerrymandering for passing political advantage and gerrymandering leading to the “unjustified entrenchment” of a political party. Id., at 360–361. The Court last took up this question in League of United Latin American Citizens v. Perry, 548 U. S. 399 (2006) (LULAC). The plaintiffs there challenged a mid-decade redistricting map passed by the Texas Legislature. As in Vieth, a majority of the Court could find no justiciable standard by which to resolve the plaintiffs’ partisan gerrymandering claims. Relevant to this case, an amicus brief in support of the LULAC plaintiffs proposed a “symmetry standard” to “measure partisan bias” by comparing how the two major political parties “would fare hypothetically if they each . . . received a given percentage of the vote.” 548 U. S., at 419 (opinion of Kennedy, J.). Justice Kennedy noted some wariness at the prospect of “adopting a constitutional standard that invalidates a map based on unfair results that would occur in a hypothetical state of affairs.” Id., at 420. Aside from that problem, he wrote, the partisan bias standard shed no light on “how much partisan dominance is too much.” Ibid. Justice Kennedy therefore concluded that “asymmetry alone is not a reliable measure of unconstitutional partisanship.” Ibid. Justice Stevens would have found that the Texas map was a partisan gerrymander based in part on the asymmetric advantage it conferred on Republicans in converting votes to seats. Id., at 466–467, 471–473 (opinion concurring in part and dissenting in part). Justice Souter, writing for himself and Justice Ginsburg, noted that he would not “rule out the utility of a criterion of symmetry,” and that “further attention could be devoted to the administrability of such a criterion at all levels of redistricting and its review.” Id., at 483–484 (opinion concurring in part and dissenting in part). B At argument on appeal in this case, counsel for the plaintiffs argued that this Court can address the problem of partisan gerrymandering because it must: The Court should exercise its power here because it is the “only institution in the United States” capable of “solv[ing] this problem.” Tr. of Oral Arg. 62. Such invitations must be answered with care. “Failure of political will does not justify unconstitutional remedies.” Clinton v. City of New York, 524 U. S. 417, 449 (1998) (Kennedy, J., concurring). Our power as judges to “say what the law is,” Marbury v. Madison, 1 Cranch 137, 177 (1803), rests not on the default of politically accountable officers, but is instead grounded in and limited by the necessity of resolving, according to legal principles, a plaintiff’s particular claim of legal right. Our considerable efforts in Gaffney, Bandemer, Vieth, and LULAC leave unresolved whether such claims may be brought in cases involving allegations of partisan gerrymandering. In particular, two threshold questions remain: what is necessary to show standing in a case of this sort, and whether those claims are justiciable. Here we do not decide the latter question because the plaintiffs in this case have not shown standing under the theory upon which they based their claims for relief. To ensure that the Federal Judiciary respects “the proper—and properly limited—role of the courts in a democratic society,” Allen v. Wright, 468 U. S. 737, 750 (1984), a plaintiff may not invoke federal-court jurisdiction unless he can show “a personal stake in the outcome of the controversy.” Baker, 369 U. S., at 204. A federal court is not “a forum for generalized grievances,” and the requirement of such a personal stake “ensures that courts exercise power that is judicial in nature.” Lance, 549 U. S., at 439, 441. We enforce that requirement by insisting that a plaintiff satisfy the familiar three-part test for Article III standing: that he “(1) suffered an injury in fact, (2) that is fairly traceable to the challenged conduct of the defendant, and (3) that is likely to be redressed by a favorable judicial decision.” Spokeo, Inc. v. Robins, 578 U. S. ___, ___ (2016) (slip op., at 6). Foremost among these requirements is injury in fact—a plaintiff’s pleading and proof that he has suffered the “invasion of a legally protected interest” that is “concrete and particularized,” i.e., which “affect[s] the plaintiff in a personal and individual way.” Lujan v. Defenders of Wildlife, 504 U. S. 555, 560, and n. 1 (1992). We have long recognized that a person’s right to vote is “individual and personal in nature.” Reynolds v. Sims, 377 U. S. 533, 561 (1964). Thus, “voters who allege facts showing disadvantage to themselves as individuals have standing to sue” to remedy that disadvantage. Baker, 369 U. S., at 206. The plaintiffs in this case alleged that they suffered such injury from partisan gerrymandering, which works through “packing” and “cracking” voters of one party to disadvantage those voters. 1 App. 28–29, 32–33, Complaint ¶¶5, 15. That is, the plaintiffs claim a constitutional right not to be placed in legislative districts deliberately designed to “waste” their votes in elections where their chosen candidates will win in landslides (packing) or are destined to lose by closer margins (cracking). Id., at 32–33, ¶15. To the extent the plaintiffs’ alleged harm is the dilution of their votes, that injury is district specific. An individual voter in Wisconsin is placed in a single district. He votes for a single representative. The boundaries of the district, and the composition of its voters, determine whether and to what extent a particular voter is packed or cracked. This “disadvantage to [the voter] as [an] individual[ ],” Baker, 369 U. S., at 206, therefore results from the boundaries of the particular district in which he resides. And a plaintiff’s remedy must be “limited to the inadequacy that produced [his] injury in fact.” Lewis v. Casey, 518 U. S. 343, 357 (1996). In this case the remedy that is proper and sufficient lies in the revision of the boundaries of the individual’s own district. For similar reasons, we have held that a plaintiff who alleges that he is the object of a racial gerrymander—a drawing of district lines on the basis of race—has standing to assert only that his own district has been so gerrymandered. See United States v. Hays, 515 U. S. 737, 744–745 (1995). A plaintiff who complains of gerrymandering, but who does not live in a gerrymandered district, “assert[s] only a generalized grievance against governmental conduct of which he or she does not approve.” Id., at 745. Plaintiffs who complain of racial gerrymandering in their State cannot sue to invalidate the whole State’s legislative districting map; such complaints must proceed “district-by-district.” Alabama Legislative Black Caucus v. Alabama, 575 U. S. ___, ___ (2015) (slip op., at 6). The plaintiffs argue that their claim of statewide injury is analogous to the claims presented in Baker and Reynolds, which they assert were “statewide in nature” because they rested on allegations that “districts throughout a state [had] been malapportioned.” Brief for Appellees 29. But, as we have already noted, the holdings in Baker and Reynolds were expressly premised on the understanding that the injuries giving rise to those claims were “individual and personal in nature,” Reynolds, 377 U. S., at 561, because the claims were brought by voters who alleged “facts showing disadvantage to themselves as individuals,” Baker, 369 U. S., at 206. The plaintiffs’ mistaken insistence that the claims in Baker and Reynolds were “statewide in nature” rests on a failure to distinguish injury from remedy. In those malapportionment cases, the only way to vindicate an individual plaintiff’s right to an equally weighted vote was through a wholesale “restructuring of the geographical distribution of seats in a state legislature.” Reynolds, 377 U. S., at 561; see, e.g., Moss v. Burkhart, 220 F. Supp. 149, 156–160 (WD Okla. 1963) (directing the county-by-county reapportionment of the Oklahoma Legislature), aff’d sub nom. Williams v. Moss, 378 U. S. 558 (1964) ( per curiam). Here, the plaintiffs’ partisan gerrymandering claims turn on allegations that their votes have been diluted. That harm arises from the particular composition of the voter’s own district, which causes his vote—having been packed or cracked—to carry less weight than it would carry in another, hypothetical district. Remedying the individual voter’s harm, therefore, does not necessarily require restructuring all of the State’s legislative districts. It requires revising only such districts as are necessary to reshape the voter’s district—so that the voter may be unpacked or uncracked, as the case may be. Cf. Alabama Legislative Black Caucus, 575 U. S., at ___ (slip op., at 7). This fits the rule that a “remedy must of course be limited to the inadequacy that produced the injury in fact that the plaintiff has established.” Lewis, 518 U. S., at 357. The plaintiffs argue that their legal injury is not limited to the injury that they have suffered as individual voters, but extends also to the statewide harm to their interest “in their collective representation in the legislature,” and in influencing the legislature’s overall “composition and policymaking.” Brief for Appellees 31. But our cases to date have not found that this presents an individual and personal injury of the kind required for Article III standing. On the facts of this case, the plaintiffs may not rely on “the kind of undifferentiated, generalized grievance about the conduct of government that we have refused to countenance in the past.” Lance, 549 U. S., at 442. A citizen’s interest in the overall composition of the legislature is embodied in his right to vote for his representative. And the citizen’s abstract interest in policies adopted by the legislature on the facts here is a nonjusticiable “general interest common to all members of the public.” Ex parte Lévitt, 302 U. S. 633, 634 (1937) (per curiam). We leave for another day consideration of other possible theories of harm not presented here and whether those theories might present justiciable claims giving rise to statewide remedies. Justice Kagan’s concurring opinion endeavors to address “other kinds of constitutional harm,” see post, at 8, perhaps involving different kinds of plaintiffs, see post, at 9, and differently alleged burdens, see ibid. But the opinion of the Court rests on the understanding that we lack jurisdiction to decide this case, much less to draw speculative and advisory conclusions regarding others. See Public Workers v. Mitchell, 330 U. S. 75, 90 (1947) (noting that courts must “respect the limits of [their] unique authority” and engage in “[j]udicial exposition . . . only when necessary to decide definite issues between litigants”). The reasoning of this Court with respect to the disposition of this case is set forth in this opinion and none other. And the sum of the standing principles articulated here, as applied to this case, is that the harm asserted by the plaintiffs is best understood as arising from a burden on those plaintiffs’ own votes. In this gerrymandering context that burden arises through a voter’s placement in a “cracked” or “packed” district. C Four of the plaintiffs in this case—Mary Lynne Donohue, Wendy Sue Johnson, Janet Mitchell, and Jerome Wallace—pleaded a particularized burden along such lines. They alleged that Act 43 had “dilut[ed] the influence” of their votes as a result of packing or cracking in their legislative districts. See 1 App. 34–36, Complaint ¶¶20, 23, 24, 26. The facts necessary to establish standing, however, must not only be alleged at the pleading stage, but also proved at trial. See Defenders of Wildlife, 504 U. S., at 561. As the proceedings in the District Court progressed to trial, the plaintiffs failed to meaningfully pursue their allegations of individual harm. The plaintiffs did not seek to show such requisite harm since, on this record, it appears that not a single plaintiff sought to prove that he or she lives in a cracked or packed district. They instead rested their case at trial—and their arguments before this Court—on their theory of statewide injury to Wisconsin Democrats, in support of which they offered three kinds of evidence. First, the plaintiffs presented the testimony of the lead plaintiff, Professor Whitford. But Whitford’s testimony does not support any claim of packing or cracking of himself as a voter. Indeed, Whitford expressly acknowledged that Act 43 did not affect the weight of his vote. 147 Record 37. His testimony points merely to his hope of achieving a Democratic majority in the legislature—what the plaintiffs describe here as their shared interest in the composition of “the legislature as a whole.” Brief for Appellees 32. Under our cases to date, that is a collective political interest, not an individual legal interest, and the Court must be cautious that it does not become “a forum for generalized grievances.” Lance, 549 U. S., at 439, 441. Second, the plaintiffs provided evidence regarding the mapmakers’ deliberations as they drew district lines. As the District Court recounted, the plaintiffs’ evidence showed that the mapmakers “test[ed] the partisan makeup and performance of districts as they might be configured in different ways.” 218 F. Supp. 3d, at 891. Each of the mapmakers’ alternative configurations came with a table that listed the number of “Safe” and “Lean” seats for each party, as well as “Swing” seats. Ibid. The mapmakers also labeled certain districts as ones in which “GOP seats [would be] strengthened a lot,” id., at 893; 2 App. 344, or which would result in “Statistical Pick Ups” for Republicans. 218 F. Supp. 3d, at 893 (alterations omitted). And they identified still other districts in which “GOP seats [would be] strengthened a little,” “weakened a little,” or were “likely lost.” Ibid. The District Court relied upon this evidence in concluding that, “from the outset of the redistricting process, the drafters sought to understand the partisan effect of the maps they were drawing.” Id., at 895. That evidence may well be pertinent with respect to any ultimate determination whether the plaintiffs may prevail in their claims against the defendants, assuming such claims present a justiciable controversy. But the question at this point is whether the plaintiffs have established injury in fact. That turns on effect, not intent, and requires a showing of a burden on the plaintiffs’ votes that is “actual or imminent, not ‘conjectural’ or ‘hypothetical.’ ” Defenders of Wildlife, 504 U. S., at 560. Third, the plaintiffs offered evidence concerning the impact that Act 43 had in skewing Wisconsin’s statewide political map in favor of Republicans. This evidence, which made up the heart of the plaintiffs’ case, was derived from partisan-asymmetry studies similar to those discussed in LULAC. The plaintiffs contend that these studies measure deviations from “partisan symmetry,” which they describe as the “social scientific tenet that [districting] maps should treat parties symmetrically.” Brief for Appellees 37. In the District Court, the plaintiffs’ case rested largely on a particular measure of partisan asymmetry—the “efficiency gap” of wasted votes. See supra, at 3–4. That measure was first developed in two academic articles published shortly before the initiation of this lawsuit. See Stephanopoulos & McGhee, Partisan Gerrymandering and the Efficiency Gap, 82 U. Chi. L. Rev. 831 (2015); McGhee, Measuring Partisan Bias in Single-Member District Electoral Systems, 39 Leg. Studies Q. 55 (2014). The plaintiffs asserted in their complaint that the “efficiency gap captures in a single number all of a district plan’s cracking and packing.” 1 App. 28–29, Complaint ¶5 (emphasis deleted). That number is calculated by subtracting the statewide sum of one party’s wasted votes from the statewide sum of the other party’s wasted votes and dividing the result by the statewide sum of all votes cast, where “wasted votes” are defined as all votes cast for a losing candidate and all votes cast for a winning candidate beyond the 50% plus one that ensures victory. See Brief for Eric McGhee as Amicus Curiae 6, and n. 3. The larger the number produced by that calculation, the greater the asymmetry between the parties in their efficiency in converting votes into legislative seats. Though they take no firm position on the matter, the plaintiffs have suggested that an efficiency gap in the range of 7% to 10% should trigger constitutional scrutiny. See Brief for Appellees 52–53, and n. 17. The plaintiffs and their amici curiae promise us that the efficiency gap and similar measures of partisan asymmetry will allow the federal courts—armed with just “a pencil and paper or a hand calculator”—to finally solve the problem of partisan gerrymandering that has confounded the Court for decades. Brief for Heather K. Gerken et al. as Amici Curiae 27 (citing Wang, Let Math Save Our Democracy, N. Y. Times, Dec. 5, 2015). We need not doubt the plaintiffs’ math. The difficulty for standing purposes is that these calculations are an average measure. They do not address the effect that a gerrymander has on the votes of particular citizens. Partisan-asymmetry metrics such as the efficiency gap measure something else en- tirely: the effect that a gerrymander has on the fortunes of political parties. Consider the situation of Professor Whitford, who lives in District 76, where, defendants contend, Democrats are “naturally” packed due to their geographic concentration, with that of plaintiff Mary Lynne Donohue, who lives in Assembly District 26 in Sheboygan, where Democrats like her have allegedly been deliberately cracked. By all accounts, Act 43 has not affected Whitford’s individual vote for his Assembly representative—even plaintiffs’ own demonstration map resulted in a virtually identical district for him. Donohue, on the other hand, alleges that Act 43 burdened her individual vote. Yet neither the effi- ciency gap nor the other measures of partisan asymmetry offered by the plaintiffs are capable of telling the difference between what Act 43 did to Whitford and what it did to Donohue. The single statewide measure of partisan advantage delivered by the efficiency gap treats Whitford and Donohue as indistinguishable, even though their individual situations are quite different. That shortcoming confirms the fundamental problem with the plaintiffs’ case as presented on this record. It is a case about group political interests, not individual legal rights. But this Court is not responsible for vindicating generalized partisan preferences. The Court’s constitutionally prescribed role is to vindicate the individual rights of the people appearing before it. III In cases where a plaintiff fails to demonstrate Article III standing, we usually direct the dismissal of the plaintiff’s claims. See, e.g., DaimlerChrysler Corp. v. Cuno, 547 U. S. 332, 354 (2006). This is not the usual case. It concerns an unsettled kind of claim this Court has not agreed upon, the contours and justiciability of which are unresolved. Under the circumstances, and in light of the plaintiffs’ allegations that Donohue, Johnson, Mitchell, and Wallace live in districts where Democrats like them have been packed or cracked, we decline to direct dismissal. We therefore remand the case to the District Court so that the plaintiffs may have an opportunity to prove concrete and particularized injuries using evidence—unlike the bulk of the evidence presented thus far—that would tend to demonstrate a burden on their individual votes. Cf. Alabama Legislative Black Caucus, 575 U. S., at ___ (slip op., at 8) (remanding for further consideration of the plaintiffs’ gerrymandering claims on a district-by-district basis). We express no view on the merits of the plaintiffs’ case. We caution, however, that “standing is not dispensed in gross”: A plaintiff’s remedy must be tailored to redress the plaintiff’s particular injury. Cuno, 547 U. S., at 353. The judgment of the District Court is vacated, and the case is remanded for further proceedings consistent with this opinion. It is so ordered.

584.US.2017_16-1150

Respondent Samuel Hall served as caretaker and legal advisor to his mother Ethlyn Hall, a property owner in the United States Virgin Islands. After falling out with Samuel, Ethlyn transferred her property into a trust and designated her daughter, petitioner Elsa Hall, as her successor trustee. Ethlyn sued Samuel and his law firm over the handling of her affairs (the “trust case”). When Ethlyn died, Elsa took Ethlyn’s place as trustee and as plaintiff. Samuel later filed a separate complaint against Elsa in her individual capacity (the “individual case”). On Samuel’s motion, the District Court consolidated the trust and individual cases under Federal Rule of Civil Procedure 42(a). The District Court held a single trial of the consolidated cases. In the individual case, the jury returned a verdict for Samuel, but the District Court granted Elsa a new trial. In the trust case, the jury returned a verdict against Elsa, and she filed a notice of appeal from the judgment in that case. Samuel moved to dismiss the appeal on jurisdictional grounds, arguing that the judgment in the trust case was not final and appealable because his claims against Elsa remained unresolved in the individual case. The Court of Appeals for the Third Circuit agreed and dismissed the appeal. Held: When one of several cases consolidated under Rule 42(a) is finally decided, that decision confers upon the losing party the immediate right to appeal, regardless of whether any of the other consolidated cases remain pending. Pp. 4–18. (a) Title 28 U. S. C. §1291 vests the courts of appeals with jurisdiction over “appeals from all final decisions of the district courts,” except those directly appealable to this Court. Under §1291, “any litigant armed with a final judgment from a lower federal court is entitled to take an appeal.” Arizona v. Manypenny, 451 U. S. 232 . Here an appeal would normally lie from the judgment in the trust case. But Samuel argues that because the trust and individual cases were consolidated under Rule 42(a)(2), they merged and should be regarded as one case, such that the judgment in the trust case was merely interlocutory and not appealable before the consolidated cases in the aggregate are finally resolved. Pp. 4–5. (b) Rule 42(a)(2) provides that if “actions before the court involve a common question of law or fact, the court may . . . consolidate the actions.” The meaning of the term “consolidate” in this context is ambiguous. But the term has a legal lineage stretching back at least to the first federal consolidation statute, enacted by Congress in 1813. Act of July 22, 1813, §3, 3Stat. 21 (later codified as Rev. Stat. §921 and 28 U. S. C. §734 (1934 ed.)). That history makes clear that one of multiple cases consolidated under the Rule retains its independent character, at least to the extent it is appealable when finally resolved, regardless of any ongoing proceedings in the other cases. Pp. 5–6. (c) Under the consolidation statute—which was in force for 125 years, until its replacement by Rule 42(a)—consolidation was understood not as completely merging the constituent cases into one, but as enabling more efficient case management while preserving the distinct identities of the cases and rights of the separate parties in them. See, e.g., Rich v. Lambert, 12 How. 347; Mutual Life Ins. Co. v. Hillmon, 145 U. S. 285 ; Stone v. United States, 167 U. S. 178 . Just five years before Rule 42(a) became law, the Court reiterated that, under the consolidation statute, consolidation did not result in the merger of constituent cases. Johnson v. Manhattan R. Co., 289 U. S. 479 –497. This body of law supports the inference that, prior to Rule 42(a), a judgment completely resolving one of several consolidated cases was an immediately appealable final decision. Pp. 6–12. (d) Rule 42(a) was expressly modeled on the consolidation statute. Because the Rule contained no definition of “consolidate,” the term presumably carried forward the same meaning ascribed to it under the statute and reaffirmed in Johnson. Samuel nonetheless asserts that “consolidate” took on a different meaning under Rule 42(a). He describes the Rule as permitting two forms of consolidation: consolidation for limited purposes and consolidation for all purposes. He locates textual authority for the former in a new provision, subsection (a)(1), which permits courts to “join for hearing or trial any or all matters at issue in the actions.” And he contends that subsection (a)(2), so as not to be superfluous, must permit the merger of cases that have been consolidated for all purposes into a single, undifferentiated case. But the narrow grant of authority in subsection (a)(1) cannot fairly be read as the exclusive source of a district court’s power to consolidate cases for limited purposes, because there is much more to litigation than hearings or trials. Instead, that undisputed power must stem from subsection (a)(2). That defeats Samuel’s argument that interpreting subsection (a)(2) to adopt the traditional understanding of consolidation would render it duplicative of subsection (a)(1), and that subsection (a)(2) therefore must permit courts to merge the actions into a single unit. Moreover, a Federal Rules Advisory Committee would not take a term that had long meant that separate actions do not merge into one, and silently and abruptly reimagine the same term to mean that they do. Nothing in the pertinent Committee proceedings supports the notion that Rule 42(a) was meant to overturn the settled understanding of consolidation; the Committee simply commented that Rule 42(a) “is based upon” its statutory predecessor, “but insofar as the statute differs from this rule, it is modified.” Advisory Committee’s Notes on 1937 Adoption of Fed. Rule Civ. Proc. 42(a), 28 U. S. C. App., p. 887. The limited extent to which this Court has addressed consolidation since adoption of Rule 42(a) confirms that the traditional understanding remains in place. See, e.g., Bank Markazi v. Peterson, 578 U. S. ___, ___–___; Butler v. Dexter, 425 U. S. 262 –267. This decision does not mean that district courts may not consolidate cases for all purposes in appropriate circumstances. But constituent cases retain their separate identities at least to the extent that a final decision in one is immediately appealable by the losing party. Pp. 12–17. 679 Fed. Appx. 142, reversed and remanded. Roberts, C. J., delivered the opinion for a unanimous Court.

Three Terms ago, we held that one of multiple cases consolidated for multidistrict litigation under 28 U. S. C. §1407 is immediately appealable upon an order disposing of that case, regardless of whether any of the others remain pending. Gelboim v. Bank of America Corp., 574 U. S. ___ (2015). We left open, however, the question whether the same is true with respect to cases consoli- dated under Rule 42(a) of the Federal Rules of Civil Procedure. Id., at ___, n. 4 (slip op., at 7, n. 4). This case presents that question. I Petitioner Elsa Hall and respondent Samuel Hall are siblings enmeshed in a long-running family feud. Their mother, Ethlyn Hall, lived and owned property in the United States Virgin Islands. Samuel, a lawyer in the Virgin Islands, served as Ethlyn’s caretaker and provided her with legal assistance. But trouble eventually came to paradise, and Samuel and Ethlyn fell out over Samuel’s management of Ethlyn’s real estate holdings. During a visit from Elsa, Ethlyn established an inter vivos trust, transferred all of her property into the trust, and designated Elsa as her successor trustee. Ethlyn then moved to Miami—under circumstances disputed by the parties—to live with her daughter. The family squabble made its way to court in May 2011. Ethlyn, acting in her individual capacity and as trustee of her inter vivos trust, sued Samuel and his law firm in Federal District Court (the “trust case”). Ethlyn’s claims—for breach of fiduciary duty, legal malpractice, conversion, fraud, and unjust enrichment—concerned the handling of her affairs by Samuel and his law firm before she left for Florida. Then Ethlyn died, and Elsa stepped into her shoes as trustee and accordingly as plaintiff in the trust case. Samuel promptly filed counterclaims in that case against Elsa—in both her individual and representative capacities—for intentional infliction of emotional distress, fraud, breach of fiduciary duty, conversion, and tortious inter- ference. Samuel contended that Elsa had turned their mother against him by taking advantage of Ethlyn’s alleged mental frailty. But Samuel ran into an obstacle: Elsa was not a party to the trust case in her individual capacity (only Ethlyn had been). So Samuel filed a new complaint against Elsa in her individual capacity in the same District Court (the “individual case”), raising the same claims that he had asserted as counterclaims in the trust case. The trust and individual cases initially proceeded along separate tracks. Eventually, on Samuel’s motion, the District Court consolidated the cases under Rule 42(a) of the Federal Rules of Civil Procedure, ordering that “[a]ll submissions in the consolidated case shall be filed in” the docket assigned to the trust case. App. to Pet. for Cert. A–15. Just before the trial commenced, the District Court dismissed from the trust case Samuel’s counterclaims against Elsa. Those claims remained in the individual case. The parties then tried the consolidated cases to- gether before a jury. In the individual case, the jury returned a verdict for Samuel on his intentional infliction of emotional distress claim against Elsa, awarding him $500,000 in compensatory damages and $1.5 million in punitive damages. The clerk entered judgment in that case, but the District Court granted Elsa a new trial, which had the effect of reopening the judgment. The individual case remains pending before the District Court. In the trust case, the jury returned a verdict against Elsa, in her representative capacity, on her claims against Samuel and his law firm. The clerk entered judgment in that case directing that Elsa “recover nothing” and that “the action be dismissed on the merits.” Id., at A–12. Elsa filed a notice of appeal from the District Court’s judgment in the trust case. Samuel and his law firm moved to dismiss the appeal on jurisdictional grounds, arguing that the judgment was not final and appealable because his claims against Elsa remained unresolved in the individual case. The Court of Appeals for the Third Circuit agreed. When two cases have been consolidated for all purposes, the court reasoned, a final decision on one set of claims is generally not appealable while the second set remains pending. The court explained that it considers “whether a less-than-complete judgment is appealable” on a “case-by-case basis.” 679 Fed. Appx. 142, 145 (2017). Here, the fact that the claims in the trust and individual cases had been “scheduled together and tried before a single jury” “counsel[ed] in favor of keeping the claims together on appeal.” Ibid. The court dismissed Elsa’s appeal for lack of jurisdiction. We granted certiorari, 582 U. S. ___ (2017), and now reverse. II A Had the District Court never consolidated the trust and individual cases, there would be no question that Elsa could immediately appeal from the judgment in the trust case. Title 28 U. S. C. §1291 vests the courts of appeals with jurisdiction over “appeals from all final decisions of the district courts,” except those directly appealable to this Court. A final decision “ends the litigation on the merits and leaves nothing for the court to do but execute the judgment.” Ray Haluch Gravel Co. v. Central Pension Fund of Operating Engineers and Participating Employers, 571 U. S. 177, 183 (2014) . The archetypal final decision is “one[ ] that trigger[s] the entry of judgment.” Mohawk Industries, Inc. v. Carpenter, 558 U. S. 100, 103 (2009) . Appeal from such a final decision is a “matter of right.” Gelboim, 574 U. S., at ___ (slip op., at 1). Under §1291, “any litigant armed with a final judgment from a lower federal court is entitled to take an appeal,” Arizona v. Manypenny, 451 U. S. 232, 244 (1981) , which generally must be filed within 30 days, 28 U. S. C. §2107(a). Here the jury’s verdict against Elsa resolved all of the claims in the trust case, and the clerk accordingly entered judgment in that case providing that “the action be dismissed on the merits.” App. to Pet. for Cert. A–12. With the entry of judgment, the District Court “completed its adjudication of [Elsa’s] complaint and terminated [her] action.” Gelboim, 574 U. S., at ___ (slip op., at 7). An appeal would normally lie from that judgment. But, Samuel contends, there is more to the litigation than the suit Elsa pursued against him in her representative capacity. There is also his suit against her in her individual capacity, which has not yet been decided. Because the District Court consolidated the trust and individual cases under Rule 42(a)(2), he argues, they merged and should be regarded as one case. Viewed that way, the judgment in the trust case was merely interlocutory, and more remains to be done in the individual case before the consolidated cases in the aggregate are finally resolved and subject to appeal. B Rule 42(a)—entitled “[c]onsolidation”—provides that if “actions before the court involve a common question of law or fact, the court may” take one of three measures. First, the court may “join for hearing or trial any or all matters at issue in the actions.” Fed. Rule Civ. Proc. 42(a)(1). Second, the court may “consolidate the actions.” Rule 42(a)(2). Third, the court may “issue any other orders to avoid unnecessary cost or delay.” Rule 42(a)(3). Whether the judgment entered in the trust case is an immediately appealable final decision turns on the effect of consolidation under Rule 42(a). Samuel, looking to dictionary definitions, asserts that the “plain meaning of the phrase ‘consolidate the actions’ is . . . to unite two or more actions into one whole—that is, to join them into a single case.” Brief for Respondents 23 (citing Black’s Law Dictionary (10th ed. 2014); some internal quotation marks and alterations omitted). But the meaning of “consolidate” in the present context is am- biguous. When Rule 42(a) was adopted, the term was generally defined, as it is now, as meaning to “unite, as various particulars, into one mass or body; to bring together in close union; to combine.” Webster’s New International Dictionary 570 (2d ed. 1942). Consolidation can thus sometimes signify the complete merger of discrete units: “The company consolidated two branches.” But the term can also mean joining together discrete units without causing them to lose their independent character. The United States, for example, is composed of States “unite[d], as various particulars, into one mass or body,” “br[ought] together in close union,” or “combine[d].” Yet all agree that entry into our Union “by no means implies the loss of distinct and individual existence . . . by the States.” Texas v. White, 7 Wall. 700, 725 (1869). “She consolidated her books” hardly suggests that the “books” became “book.” The very metaphor Samuel offers—that consolidation “make[s] two one, like marriage”—highlights this point. Tr. of Oral Arg. 56. However dear to each other, spouses would be surprised to hear that their union extends beyond the metaphysical. This is not a plain meaning case. It is instead about a term—consolidate—with a legal lineage stretching back at least to the first federal consolidation statute, enacted by Congress in 1813. Act of July 22, 1813, §3, 3Stat. 21 (later codified as Rev. Stat. §921 and 28 U. S. C. §734 (1934 ed.)). Over 125 years, this Court, along with the courts of appeals and leading treatises, interpreted that term to mean the joining together—but not the complete merger—of constituent cases. Those authorities particularly emphasized that constituent cases remained independent when it came to judgments and appeals. Rule 42(a), promulgated in 1938, was expressly based on the 1813 statute. The history against which Rule 42(a) was adopted resolves any ambiguity regarding the meaning of “consolidate” in subsection (a)(2). It makes clear that one of multiple cases consolidated under the Rule retains its independent character, at least to the extent it is appealable when finally resolved, regardless of any ongoing proceedings in the other cases. C Lord Mansfield pioneered the consolidation of related cases in England, and the practice quickly took root in American courts. See Mutual Life Ins. Co. v. Hillmon, 145 U. S. 285, 292 (1892) . In 1813, Congress authorized the newly formed federal courts, when confronted with “causes of like nature, or relative to the same question,” to “make such orders and rules concerning proceedings therein as may be conformable to the principles and usages belonging to courts for avoiding unnecessary costs or delay in the administration of justice” and to “consolidate[ ]” the causes when it “shall appear reasonable.” §3, 3Stat. 21. This consolidation statute applied at law, equity, and admi- ralty, see 1 W. Rose, A Code of Federal Procedure §823(a) (1907) (Rose), and remained in force for 125 years, until its replacement by Rule 42(a). From the outset, we understood consolidation not as completely merging the constituent cases into one, but instead as enabling more efficient case management while preserving the distinct identities of the cases and the rights of the separate parties in them. In Rich v. Lambert, 12 How. 347 (1852), for example, we considered an appeal from several consolidated cases in admiralty. The appellees, the owners of cargo damaged during shipment, raised a challenge to our jurisdiction that turned on the nature of the consolidation. At the time, we could exercise appellate jurisdiction only over cases involving at least $2,000 in controversy. The damages awarded to the cargo owners in the consolidated cases surpassed $2,000 in the aggregate, but most of the constituent cases did not individually clear that jurisdictional hurdle. Id., at 352–353. We declined to view the consolidated cases as one for purposes of appeal, concluding that we had jurisdiction only over those constituent cases that individually involved damages exceeding $2,000. Ibid. As we explained, “although [a consolidated] proceeding assumes the form of a joint suit, it is in reality a mere joinder of distinct causes of action by distinct parties, arising out of a common injury, and which are heard and determined, so far as the merits are concerned, the same as in the case of separate libels for each cause of action.” Id., at 353. Consolidation was “allowed by the practice of the court for its convenience, and the saving of time and expense to the parties.” Ibid. The trial court’s decree, we noted, had the effect of individually resolving each constituent case. Ibid. (“The same decree . . . is entered as in the case of separate suits.”); see Black’s Law Dictionary 532 (3d ed. 1933) (“decree” is a “judgment of a court of equity or admiralty, answering for most purposes to the judgment of a court of common law”). Accordingly, we did “not perceive . . . any ground for a distinction as to the right of appeal from a decree as entered in these cases from that which exists where the proceedings have been distinct and separate throughout.” Rich, 12 How., at 353; see Hanover Fire Ins. Co. v. Kinneard, 129 U. S. 176, 177 (1889) (evaluating appellate jurisdiction over a writ of error in one of several consolidated cases without reference to the others). We elaborated on the principles underlying consolidation in Mutual Life Insurance Co. v. Hillmon, 145 U. S. 285 . Hillmon, a staple of law school courses on evidence, involved three separate actions instituted against different life insurance companies by one Sallie Hillmon, the beneficiary on policies purchased by her husband John. Sallie claimed she was entitled to the sizable proceeds of the policies because John had died while journeying through southern Kansas with two companions in search of a site for a cattle ranch. The three companies countered that John was in fact still alive, having conspired with one of the companions to murder the other and pass his corpse off as John’s, all as part of an insurance fraud scheme. The trial court consolidated the cases and tried them together. Id., at 285–287. The court, for purposes of determining the number of peremptory juror challenges to which each defendant was entitled, treated the three cases as though they had merged into one. Ibid. On appeal we disagreed, holding that each defendant should receive the full complement of peremptory challenges. Id., at 293. That was because, “although the defendants might lawfully be compelled, at the discretion of the court, to try the cases together, the causes of action remained distinct, and required separate verdicts and judgments; and no defendant could be deprived, without its consent, of any right material to its defence . . . to which it would have been entitled if the cases had been tried separately.” Ibid. On remand, one case settled, and a consolidated trial of the others “result[ed] in separate judgments” for Sallie. Connecticut Mut. Life Ins. Co. v. Hillmon, 188 U. S. 208, 209 (1903) . In Stone v. United States, 167 U. S. 178, 189 (1897) , we held that a party appealing from the judgment in one of two cases consolidated for trial could not also raise claims with respect to the other case. John Stone was the sole defendant in one case and one of three defendants in the other. Id., at 179–181. After a consolidated trial, the jury returned a verdict in the case against Stone alone; its verdict in the multidefendant case was set aside. Id., at 181. Stone appealed from the judgment in his case, arguing that the failure to grant a peremptory challenge in the multidefendant case affected the jury’s verdict in his. Id., at 189. We rejected that claim, punctiliously respecting the distinction between the constituent cases. There was “no merit in the objection,” we said, because in the case before us Stone had “had the benefit of the three peremptory challenges” to which he was entitled in that case. Ibid.; see Stone v. United States, 64 F. 667, 672 (CA9 1894) (“The two cases, although consolidated, were separate and distinct. Defendant had exercised all the rights and privileges he was entitled to in this case.”). And just five years before Rule 42(a) became law, we reiterated that, under the consolidation statute, consolidation did not result in the merger of constituent cases. Johnson v. Manhattan R. Co., 289 U. S. 479 –497 (1933). A major case of its day, Johnson arose from the “financial embarrassment” during the Great Depression of two companies involved in operating the New York subway system. Johnson v. Manhattan R. Co., 61 F. 2d 934, 936 (CA2 1932). In the resulting litigation, the District Court consolidated two suits, apparently with the intent to “effect an intervention of the parties to the [first suit] in the [second] suit”—in other words, to make the two suits one. Id., at 940. Judge Learned Hand, writing for the Second Circuit on appeal, would have none of it: “consolidation does not merge the suits; it is a mere matter of convenience in administration, to keep them in step. They remain as independent as before.” Ibid. We affirmed, relying on Hillmon and several lower court cases reflecting the same understanding of consolidation. Johnson, 289 U. S., at 497, n. 8. We explained once more that “consolidation is permitted as a matter of convenience and econ- omy in administration, but does not merge the suits into a single cause, or change the rights of the parties, or make those who are parties in one suit parties in another.” Id., at 496–497. Decisions by the Courts of Appeals, with isolated departures,[1]* reflected the same understanding in cases involving all manners of consolidation. See, e.g., Baltimore S. S. Co., Inc. v. Koppel Indus. Car & Equip. Co., 299 F. 158, 160 (CA4 1924) (“the consolidation for convenience of trial did not merge the two causes of action” or “deprive either party of any right or relieve it of any burden incident to the libel or cross-libel as a separate proceeding”); Taylor v. Logan Trust Co., 289 F. 51, 53 (CA8 1923) (parties to one constituent case could not appeal orders in the other because “consolidation did not make the parties to one suit parties to the other”; cited in Johnson); Toledo, St. L. & K. C. R. Co. v. Continental Trust Co., 95 F. 497, 506 (CA6 1899) (consolidation “operates as a mere carrying on together of two separate suits supposed to involve identical issues” and “does not avoid the necessity of separate decrees in each case”; cited in Johnson). One frequently cited case illustrates the point. In Adler v. Seaman, 266 F. 828, 831 (CA8 1920), the District Court “sought to employ consolidation as a medium of getting the two independent suits united,” but the Court of Appeals made clear that the consolidation statute did not authorize such action. The court explained that constituent cases sometimes “assume certain natural attitudes toward each other, such as ‘in the nature of’ a cross-bill or intervention.” Id., at 838. Be that as it may, the court continued, “this is purely a rule of convenience, and does not result in actually making such parties defendants or interveners in the other suit.” Ibid. The court described “the result of consolidation” as instead “merely to try cases together, necessitating separate verdicts and judgments or separate decrees,” and to “leave” the constituent cases “separate, independent action[s].” Id., at 838, 840. Treatises summarizing federal precedent applying the consolidation statute also concluded that consolidated cases “remain distinct.” 1 Rose §823(c), at 758. They recognized that consolidated cases should “remain separate as to parties, pleadings, and judgment,” W. Simkins, Federal Practice 63 (rev. ed. 1923), and that “[t]here must be separate verdicts, judgments or decrees, even although the consolidating party wished for one verdict,” 1 Rose §823(c), at 758; see also G. Virden, Consolidation Under Rule 42 of the Federal Rules of Civil Procedure, in 141 F. R. D. 169, 173–174 (1992) (Virden) (“as of 1933 and the Johnson case of that year, it was well settled that consolidation in the federal courts did not merge the separate cases into a single action”). Several aspects of this body of law support the inference that, prior to Rule 42(a), a judgment completely resolving one of several consolidated cases was an immediately appealable final decision. We made clear, for example, that each constituent case must be analyzed individually on appeal to ascertain jurisdiction and to decide its disposition—a compartmentalized analysis that would be gratuitous if the cases had merged into a single case subject to a single appeal. We emphasized that constituent cases should end in separate decrees or judgments—the traditional trigger for the right to appeal, for which there would be no need if an appeal could arise only from the resolution of the consolidated cases as a whole. We explained that the parties to one case did not become parties to the other by virtue of consolidation—indicating that the right of each to pursue his individual case on appeal should not be compromised by the litigation conduct of the other. And, finally, we held that consolidation could not prejudice rights to which the parties would have been due had consolidation never occurred. Forcing an aggrieved party to wait for other cases to conclude would substantially impair his ability to appeal from a final decision fully resolving his own case—a “matter of right,” Gelboim, 574 U. S., at ___ (slip op., at 1), to which he was “entitled,” Manypenny, 451 U. S., at 244. D Against this background, two years after Johnson, the Rules Advisory Committee began discussion of what was to become Rule 42(a). The Rule, which became effective in 1938, was expressly modeled on its statutory predecessor, the Act of July 22, 1813. See Advisory Committee’s Notes on 1937 Adoption of Fed. Rule Civ. Proc. 42(a), 28 U. S. C. App., p. 887. The Rule contained no definition of “consolidate,” so the term presumably carried forward the same meaning we had ascribed to it under the consolidation statute for 125 years, and had just recently reaffirmed in Johnson. See Frankfurter, Some Reflections on the Reading of Statutes, 47 Colum. L. Rev. 527, 537 (1947) (“if a word is obviously transplanted from another legal source, whether the common law or other legislation, it brings the old soil with it”); cf. Class v. United States, 583 U. S. ___, ___ (2018) (slip op., at 10) (Federal Rule of Criminal Procedure 11(a)(2) did not silently alter existing doctrine established by this Court’s past decisions). Samuel nonetheless asserts that there is a significant distinction between the original consolidation statute and Rule 42(a). The statute authorized district courts to “consolidate” related “causes when it appears reasonable to do so” or to “make such orders and rules . . . as may be conformable to the usages of courts for avoiding unnecessary costs or delay in the administration of justice.” 28 U. S. C. §734 (1934 ed.). Rule 42(a) permits district courts not only to “consolidate the actions” (subsection (a)(2)) and “issue any other orders to avoid unnecessary cost or delay” (subsection (a)(3)), but also to “join for hearing or trial any or all matters at issue in the actions” (subsection (a)(1)). Whatever “consolidate” meant under the statute, Sam- uel posits, it took on a different meaning under Rule 42(a) with the addition of subsection (a)(1). Samuel describes the Rule as “permit[ting] two forms of consolidation”: consolidation that “extend[s] only to certain proceedings,” such as discovery, and consolidation “for all purposes.” Brief for Respondents 4–5. He locates textual authority for the former in subsection (a)(1), which he says empowers courts to “join[ ] multiple actions for procedural purposes.” Id., at 23. In light of this broad grant of authority, he contends, subsection (a)(2) must provide for something more if it is not to be superfluous. And Samuel sees that something more as the ability to merge cases that have been consolidated for “all purposes” into a single, undifferentiated case—one appealable only when all issues in each formerly distinct case have been decided. See id., at 22–24 (to “consolidate” separate actions is “to join them into a single case” or “meld [them] into a single unit” (alterations omitted)). We disagree. It is only by substantially overreading subsection (a)(1) that Samuel can argue that its addition compels a radical reinterpretation of the familiar term “consolidate” in subsection (a)(2). The text of subsection (a)(1) permits the joining of cases only for “hearing or trial.” That narrow grant of authority cannot fairly be read as the exclusive source of a district court’s power to “join[ ] multiple actions for procedural purposes.” Brief for Respondents 23. There is, after all, much more to litigation than hearings or trials—such as motions practice or discovery. A district court’s undisputed ability to consolidate cases for such limited purposes must therefore stem from subsection (a)(2). That defeats Samuel’s argument that interpreting subsection (a)(2) to adopt the traditional understanding of consolidation would render it “wholly duplicative of [subsection] (a)(1),” and that subsection (a)(2) “therefore must permit courts . . . to ‘consolidate’ the actions themselves into a single unit.” Id., at 23–24. Samuel’s reinterpretation of “consolidate” is, in other words, a solution in search of a problem. We think, moreover, that if Rule 42(a) were meant to transform consolidation into something sharply contrary to what it had been, we would have heard about it. Congress, we have held, “does not alter the fundamental details” of an existing scheme with “vague terms” and “subtle device[s].” Whitman v. American Trucking Assns., Inc., 531 U. S. 457, 468 (2001) ; cf. Class, 583 U. S., at ___ (slip op., at 10). That is true in spades when it comes to the work of the Federal Rules Advisory Committees. Their laborious drafting process requires years of effort and many layers of careful review before a proposed Rule is presented to this Court for possible submission to Congress. See Report of Advisory Committee on Rules for Civil Procedure (Apr. 1937) (describing the exhaustive process undertaken to draft the first Federal Rules of Civil Procedure). No sensible draftsman, let alone a Federal Rules Advisory Committee, would take a term that had meant, for more than a century, that separate actions do not merge into one, and silently and abruptly reimagine the same term to mean that they do. Similarly, nothing in the pertinent proceedings of the Rules Advisory Committee supports the notion that Rule 42(a) was meant to overturn the settled understanding of consolidation. See United States v. Vonn, 535 U. S. 55 , n. 6 (2002) (Advisory Committee Notes are “a reliable source of insight into the meaning of a rule”). In this instance, the Committee simply commented that Rule 42(a) “is based upon” its statutory predecessor, “but insofar as the statute differs from this rule, it is modified.” Advisory Committee’s Notes on 1937 Adoption of Fed. Rule Civ. Proc. 42(a), 28 U. S. C. App., at 887. The Committee did not identify any specific instance in which Rule 42(a) changed the statute, let alone the dramatic transformation Samuel would have us recognize. See Virden 174–181 (evaluating the history of the development of Rule 42(a) and finding no evidence that the Committee intended a shift in meaning along the lines proposed by Samuel). This is significant because when the Committee intended a new rule to change existing federal practice, it typically explained the departure. See, e.g., Advisory Committee’s Notes on 1937 Adoption of Fed. Rule Civ. Proc. 4, 28 U. S. C. App., p. 747 (a predecessor statute “is substantially continued insofar as it applies to a summons, but its requirements as to teste of process are superseded”); Advisory Committee’s Notes on 1937 Adoption of Fed. Rule Civ. Proc. 18, 28 U. S. C. App., p. 802 (“In respect to fraudulent conveyances the rule changes the former rule requiring a prior judgment against the owner . . . to conform to the provisions of the Uniform Fraudulent Conveyance Act, §§ 9 and 10.”). As a leading treatise explained at the time, through consolidation under Rule 42(a) “one or many or all of the phases of the several actions may be merged. But merger is never so complete in consolidation as to deprive any party of any substantial rights which he may have possessed had the actions proceeded separately.” 3 J. Moore & J. Friedman, Moore’s Federal Practice §42.01, pp. 3050–3051 (1938). Thus, “separate verdicts and judgments are normally necessary.” Id., at 3051, n. 12. The limited extent to which this Court has addressed consolidation since adoption of Rule 42(a) confirms the traditional understanding. Just recently in Bank Markazi v. Peterson, 578 U. S. ___, ___–___ (2016) (slip op., at 19–20), for example, the Court determined that cases “consolidated for administrative purposes at the execution stage . . . were not independent of the original actions for damages and each claim retained its separate character.” The Court quoted as authority a treatise explaining that “actions do not lose their separate identity because of consolidation.” Id., at ___ (slip op., at 20) (quoting 9A C. Wright & A. Miller, Federal Practice and Procedure §2382, p. 10 (3d ed. 2008) (Wright & Miller)). In Butler v. Dexter, 425 U. S. 262 –267 (1976) (per curiam), we dismissed an appeal because the constitutional question that supplied our jurisdiction had been raised not in the case before us, but instead only in other cases with which it had been consolidated. We explained that “[e]ach case . . . must be considered separately to determine whether or not this Court has jurisdiction to consider its merits”. Id., at 267, n. 12; see Rich, 12 How., at 352–353. And in Alfred Dunhill of London, Inc. v. Republic of Cuba, 425 U.S. 682 , and n. 22 (1976) (Marshall, J., dissenting), four dissenting Justices—reaching an issue not addressed by the majority—cited Johnson for the proposition that actions are “not merged” and do “not lose their separate identities because of . . . consolidation” under Rule 42(a). In the face of all the foregoing, we cannot accept Sam- uel’s contention that “consolidate” in Rule 42(a) carried a very different meaning—with very different consequences—than it had in Johnson, just five years before the Rule was adopted. None of this means that district courts may not consolidate cases for “all purposes” in appropriate circumstances. District courts enjoy substantial discretion in deciding whether and to what extent to consolidate cases. See 9A Wright & Miller §2383 (collecting cases). What our decision does mean is that constituent cases retain their separate identities at least to the extent that a final decision in one is immediately appealable by the losing party. That is, after all, the point at which, by definition, a “district court disassociates itself from a case.” Swint v. Chambers County Comm’n, 514 U. S. 35, 42 (1995) . We thus express no view on any issue arising prior to that time. * * * The normal rule is that a “final decision” confers upon the losing party the immediate right to appeal. That rule provides clear guidance to litigants. Creating exceptions to such a critical step in litigation should not be under- taken lightly. Congress has granted us the authority to prescribe rules “defin[ing] when a ruling of a district court is final for the purposes of appeal under” §1291, 28 U. S. C. §2072(c), and we have explained that changes with respect to the meaning of final decision “are to come from rulemaking, . . . not judicial decisions in particular controversies,” Microsoft Corp. v. Baker, 582 U. S. ___, ___ (2017) (slip op., at 15). If, as Samuel fears, our holding in this case were to give rise to practical problems for district courts and litigants, the appropriate Federal Rules Advisory Committees would certainly remain free to take the matter up and recommend revisions accordingly. Rule 42(a) did not purport to alter the settled understanding of the consequences of consolidation. That understanding makes clear that when one of several consolidated cases is finally decided, a disappointed litigant is free to seek review of that decision in the court of appeals. We reverse the judgment of the Court of Appeals for the Third Circuit and remand the case for further proceedings consistent with this opinion. It is so ordered. Notes 1 * See, e.g., Edward P. Allis Co. v. Columbia Mill Co., 65 F. 52, 54 (CA8 1894) (involving two suits “consolidated, and tried as one action,” with the “complaint in the first suit . . . treated as a counterclaim interposed in the second suit”). State practice was varied. Compare, e.g., East Bay Municipal Util. Dist. v. Kieffer, 99 Cal. App. 240, 263 (1929) (denial of rehearing) (“By such consolidation the three proceedings became one proceeding and should have been determined by a single verdict, ‘a single set of findings and a single judgment.’ ”), with Missouri Pac. R. Co. v. Helmert, 196 Ark. 1073, 121 S. W. 2d 103 (1938) (consolidated cases resulted in separate judgments).

583.US.2017_16-658

An appeal filing deadline prescribed by statute is considered “jurisdictional,” meaning that late filing of the appeal notice necessitates dismissal of the appeal. See Bowles v. Russell, 551 U. S. 205 –213. In contrast, a time limit prescribed only in a court-made rule is not jurisdictional. It is a mandatory claim-processing rule that may be waived or forfeited. Ibid. This Court and other forums have sometimes overlooked this critical distinction. See Reed Elsevier, Inc. v. Muchnick, 559 U. S. 154 . Petitioner Charmaine Hamer filed an employment discrimination suit against respondents. The District Court granted respondents’ motion for summary judgment, entering final judgment on September 14, 2015. Before October 14, the date Hamer’s notice of appeal was due, her attorneys filed a motion to withdraw as counsel and a motion for an extension of the appeal filing deadline to give Hamer time to secure new counsel. The District Court granted both motions, extending the deadline to December 14, a two-month extension, even though the governing Federal Rule of Appellate Procedure, Rule 4(a)(5)(C), confines such extensions to 30 days. Concluding that Rule 4(a)(5)(C)’s time prescription is jurisdictional, the Court of Appeals dismissed Hamer’s appeal. Held: The Court of Appeals erred in treating as jurisdictional Rule 4(a)(5)(C)’s limitation on extensions of time to file a notice of appeal. Pp. 5–10. (a) The 1948 version of 28 U. S. C. §2107 allowed extensions of time to file a notice of appeal, not exceeding 30 days, “upon a showing of excusable neglect based on failure of a party to learn of the entry of the judgment,” but the statute said nothing about extensions when the judgment loser did receive notice of the entry of judgment. In 1991, the statute was amended, broadening the class of prospective appellants who could gain extensions to include all who showed “excusable neglect or good cause” and reducing the time prescription for appellants who lacked notice of the entry of judgment from 30 to 14 days. §2107(c). For other cases, the statute does not say how long an extension may run. Rule 4(a)(5)(C), however, does prescribe a limit: “No extension [of time for filing a notice of appeal] may exceed 30 days after the prescribed time [for filing a notice of appeal] or 14 days after the date [of] the order granting the [extension] motion . . . , whichever is later.” Pp. 5–6. (b) This Court’s precedent shapes a rule of decision that is both clear and easy to apply: If a time prescription governing the transfer of adjudicatory authority from one Article III court to another appears in a statute, the limitation is jurisdictional; otherwise, the time specification fits within the claim-processing category. In concluding otherwise, the Court of Appeals relied on Bowles. There, Bowles filed a notice of appeal outside a limitation set by Congress in §2107(c). This Court held that, as a result, the Court of Appeals lacked jurisdiction over his tardy appeal. 551 U. S., at 213. In conflating Rule 4(a)(5)(C) with §2107(c) here, the Seventh Circuit failed to grasp the distinction between jurisdictional appeal filing deadlines and deadlines stated only in mandatory claim-processing rules. It therefore misapplied Bowles. Bowles’s statement that “the taking of an appeal within the prescribed time is ‘mandatory and jurisdictional,’ ” id., at 209, is a characterization left over from days when the Court was “less than meticulous” in using the term “jurisdictional,” Kontrick v. Ryan, 540 U. S. 443 . The statement was correct in Bowles, where the time prescription was imposed by Congress, but it would be incorrect here, where only Rule 4(a)(5)(C) limits the length of the extension. Pp. 7–10. 835 F. 3d 761, vacated and remanded. Ginsburg, J., delivered the opinion for a unanimous Court.

This case presents a question of time, specifically, time to file a notice of appeal from a district court’s judgment. In Bowles v. Russell, 551 U. S. 205 –213 (2007), this Court clarified that an appeal filing deadline prescribed by statute will be regarded as “jurisdictional,” meaning that late filing of the appeal notice necessitates dismissal of the appeal. But a time limit prescribed only in a court-made rule, Bowles acknowledged, is not jurisdictional; it is, instead, a mandatory claim-processing rule subject to forfeiture if not properly raised by the appellee. Ibid.; Kontrick v. Ryan, 540 U. S. 443, 456 (2004) . Because the Court of Appeals held jurisdictional a time limit specified in a rule, not in a statute, 835 F. 3d 761, 763 (CA7 2016), we vacate that court’s judgment dismissing the appeal. I A “Only Congress may determine a lower federal court’s subject-matter jurisdiction.” Kontrick, 540 U. S., at 452 (citing U. S. Const., Art. III, §1); Owen Equipment & Erection Co. v. Kroger, 437 U. S. 365, 370 (1978) (“[I]t is axiomatic that the Federal Rules of Civil Procedure do not create or withdraw federal jurisdiction.”). Accordingly, a provision governing the time to appeal in a civil action qualifies as jurisdictional only if Congress sets the time. See Bowles, 551 U. S., at 211–212 (noting “the jurisdictional distinction between court-promulgated rules and limits enacted by Congress”); Sibbach v. Wilson & Co., 312 U. S. 1, 10 (1941) (noting “the inability of a court, by rule, to extend or restrict the jurisdiction conferred by a statute”). A time limit not prescribed by Congress ranks as a mandatory claim-processing rule, serving “to promote the orderly progress of litigation by requiring that the parties take certain procedural steps at certain specified times.” Henderson v. Shinseki, 562 U. S. 428, 435 (2011) . This Court and other forums have sometimes overlooked this distinction, “mischaracteriz[ing] claim-processing rules or elements of a cause of action as jurisdictional limitations, particularly when that characterization was not central to the case, and thus did not require close analysis.” Reed Elsevier, Inc. v. Muchnick, 559 U. S. 154, 161 (2010) . But prevailing precedent makes the distinction critical. Failure to comply with a jurisdictional time prescription, we have maintained, deprives a court of adjudicatory authority over the case, necessitating dismissal—a “drastic” result. Shinseki, 562 U. S., at 435; Bowles, 551 U. S., at 213 (“[W]hen an ‘appeal has not been prosecuted . . . within the time limited by the acts of Congress, it must be dismissed for want of jurisdiction.’ ” (quoting United States v. Curry, 6 How. 106, 113 (1848))). The jurisdictional defect is not subject to waiver or forfeiture[1] and may be raised at any time in the court of first instance and on direct appeal. Kontrick, 540 U. S., at 455.[2] In contrast to the ordinary operation of our adversarial system, courts are obliged to notice jurisdictional issues and raise them on their own initiative. Shinseki, 562 U. S., at 434. Mandatory claim-processing rules are less stern. If properly invoked, mandatory claim-processing rules must be enforced, but they may be waived or forfeited. Manrique v. United States, 581 U. S. ___, ___ (2017) (slip op., at 4). “[C]laim-processing rules . . . [ensure] relief to a party properly raising them, but do not compel the same result if the party forfeits them.” Eberhart v. United States, 546 U. S. 12, 19 (2005) (per curiam).[3] B Petitioner Charmaine Hamer filed a complaint against respondents Neighborhood Housing Services of Chicago and Fannie Mae alleging employment discrimination in violation of Title VII of the Civil Rights Act of 1964, 42 U. S. C. §2000e et seq., and the Age Discrimination in Employment Act of 1967, 29 U. S. C. §621 et seq. The District Court granted respondents’ motion for summary judgment on September 10, 2015, and entered final judgment on September 14, 2015. In the absence of a time extension, Hamer’s notice of appeal would have been due by October 14, 2015. Fed. Rule App. Proc. 4(a)(1)(A). On October 8, 2015, before the October 14 deadline for filing Hamer’s notice of appeal, her attorneys made two motions.[4] First, they sought to withdraw as counsel because of their disagreement with Hamer on pursuit of an appeal. Second, they sought a two-month extension of the notice of appeal filing date, so that Hamer would have adequate time to engage new counsel for her appeal. App. to Pet. for Cert. 57–59. The District Court granted both motions on the same day and ordered extension of the deadline for Hamer’s notice of appeal from October 14 to December 14, 2015. Id., at 60. Respondents did not move for reconsideration or otherwise raise any objection to the length of the extension. In the docketing statement respondents filed in the Court of Appeals, they stated: “The United States Court of Appeals for the Seventh Circuit has jurisdiction over this appeal under 28 U. S. C. §1291, in that on December 11, 2015, [Hamer] filed a timely Notice of Appeal from a final judgment of the United States District Court for the Northern District of Illinois that disposed of all of [Hamer’s] claims against [respondents].” Id., at 63. Respondents’ statement later reiterated: “On December 11, 2015, [Hamer] timely filed a Notice of Appeal . . . .” Id., at 64. Nevertheless, the Court of Appeals, on its own ini-tiative, questioned the timeliness of the appeal and instructed respondents to brief the issue. 835 F. 3d, at 762. Respondents did so and, for the first time, asserted that the appeal was untimely, citing the relevant Rule confining extensions to 30 days. Id., at 762–763 (citing Fed. Rule App. Proc. 4(a)(5)(C)). Concluding that it lacked jurisdiction to reach the merits, the Court of Appeals dismissed Hamer’s appeal. 835 F. 3d, at 763.[5] We granted certio-rari. 580 U. S. ___ (2017). II A Section 2107 of Title 28 of the U. S. Code, as enacted in 1948, allowed extensions of the time to file a notice of appeal, not exceeding 30 days, “upon a showing of excus-able neglect based on failure of a party to learn of the entry of the judgment.” Act of June 25, 1948, §2107, 62Stat. 963.[6] Nothing in the statute provided for extension of the time to file a notice of appeal when, as in this case, the judgment loser did receive notice of the entry of judgment. In 1991, Congress broadened the class of persons who could gain extensions to include all prospective appellants who showed “excusable neglect or good cause.” §12, 105Stat. 1627. In addition, Congress retained a time prescription covering appellants who lacked notice of the entry of judgment: “[A] party entitled to notice of the entry of a judgment . . . [who] did not receive such notice from the clerk or any party within 21 days of [the judgment’s] entry” qualifies for a 14-day extension,[7] if “no party would be prejudiced [thereby].” §2107(c). In full, §2107(c) now provides: “(c) The district court may, upon motion filed not later than 30 days after the expiration of the time otherwise set for bringing appeal, extend the time for appeal upon a showing of excusable neglect or good cause. In addition, if the district court finds— “(1) that a party entitled to notice of the entry of a judgment or order did not receive such notice from the clerk or any party within 21 days of its entry, and “(2) that no party would be prejudiced, “the district court may, upon motion filed within 180 days after entry of the judgment or order or within 14 days after receipt of such notice, whichever is earlier, reopen the time for appeal for a period of 14 days from the date of entry of the order reopening the time for appeal.” In short, current §2107(c), like the provision as initially enacted, specifies the length of an extension for cases in which the appellant lacked notice of the entry of judgment.[8] For other cases, the statute does not say how long an extension may run. But Federal Rule of Appellate Procedure 4(a)(5)(C) does prescribe a limit: “No extension [of time for filing a notice of appeal] may exceed 30 days after the prescribed time [for filing a notice of appeal] or 14 days after the date [of] the order granting the [extension] motion . . . , whichever is later.” Unlike §2107(c), we note, Rule 4(a)(5)(C) limits extensions of time to file a notice of appeal in all circumstances, not just in cases in which the prospective appellant lacked notice of the entry of judgment. B Although Rule 4(a)(5)(C)’s limit on extensions of time appears nowhere in the text of §2107(c), respondents now contend that Rule 4(a)(5)(C) has a “statutory basis” because §2107(c) once limited extensions (to the extent it did authorize them) to 30 days. Brief for Respondents 17. No matter, respondents submit, that Congress struck the 30-day limit in 1991 and replaced it with a 14-day limit governing, as the 30-day limit did, only lack-of-notice cases; deleting the 30-day prescription, respondents conjecture, was “probably inadverten[t].” Id., at 1. In support of their argument that Congress accidentally failed to impose an all-purpose limit on extensions, respondents observe that the 1991 statute identifies Congress’ aim as the enactment of “certain technical corrections in . . . provisions of law relating to the courts.” 105Stat. 1623. They also note the caption of the relevant section of the amending statute: “Conformity with Rules of Appellate Procedure.” Id., at 1627. Because striking the 30-day limit from §2107 made the statute less like Rule 4(a)(5)(C), respondents reason, Congress likely erased the relevant paragraph absentmindedly. Hence, respondents conclude, “there is no reason to interpret the 1991 amendment as stripping Rule 4(a)(5)(C) of its jurisdictional significance.” Brief for Respondents 2. Overlooked by respondents, pre-1991 §2107 never spoke to extensions for reasons other than lack of notice. In any event, we resist speculating whether Congress acted inadvertently. See Henson v. Santander Consumer USA Inc., 582 U. S. ___, ___–___ (2017) (slip op., at 9–10) (“[W]e will not presume with [respondents] that any result consistent with their account of the statute’s overarching goal must be the law but will presume more modestly instead ‘that [the] legislature says . . . what it means and means . . . what it says.’ ” (quoting Dodd v. United States, 545 U. S. 353, 357 (2005) )); Magwood v. Patterson, 561 U. S. 320 , 334 (2010) (“We cannot replace the actual text with speculation as to Congress’ intent.”). The rule of decision our precedent shapes is both clear and easy to apply: If a time prescription governing the transfer of adjudicatory authority from one Article III court to another appears in a statute, the limitation is jurisdictional, supra, at 2; otherwise, the time specification fits within the claim-processing category, ibid.[9] In dismissing Hamer’s appeal for want of jurisdiction, the Court of Appeals relied heavily on our decision in Bowles. We therefore reiterate what that precedent conveys. There, petitioner Keith Bowles did not receive timely notice of the entry of a postjudgment order and consequently failed to file a timely notice of appeal. Bowles v. Russell, 432 F. 3d 668, 670 (CA6 2005). When Bowles learned of the postjudgment order, he moved for an extension under Federal Rule of Appellate Procedure 4(a)(6), which implements §2107(c)’s authorization of extensions in lack-of-notice cases. Ibid. The District Court granted Bowles’s motion, but inexplicably provided a 17-day extension, rather than the 14-day extension authorized by §2107(c). Bowles, 551 U. S., at 207. Bowles filed his notice of appeal within the 17 days allowed by the District Court but outside the 14 days allowed by §2107(c). Ibid. “Because Congress specifically limited the amount of time by which district courts can extend the notice-of-appeal period in §2107(c),” we explained, the Court of Appeals lacked jurisdiction over Bowles’s tardy appeal. Id., at 213. Quoting Bowles at length, the Court of Appeals in this case reasoned that “[l]ike Rule 4(a)(6), Rule 4(a)(5)(C) is the vehicle by which §2107(c) is employed and it limits a district court’s authority to extend the notice of appeal filing deadline to no more than an additional 30 days.” 835 F. 3d, at 763. In conflating Rule 4(a)(5)(C) with §2107(c), the Court of Appeals failed to grasp the distinction our decisions delineate between jurisdictional appeal filing deadlines and mandatory claim-processing rules, and therefore misapplied Bowles. Several Courts of Appeals,[10] including the Court of Appeals in Hamer’s case, have tripped over our statement in Bowles that “the taking of an appeal within the prescribed time is ‘mandatory and jurisdictional.’ ” 551 U. S., at 209 (quoting Griggs v. Provident Consumer Discount Co., 459 U. S. 56, 61 (1982) (per curiam)). The “mandatory and jurisdictional” formulation is a characterization left over from days when we were “less than meticulous” in our use of the term “jurisdictional.” Kontrick, 540 U. S., at 454.[11] The statement was correct as applied in Bowles because, as the Court there explained, the time prescription at issue in Bowles was imposed by Congress. 551 U. S., at 209–213. But “mandatory and jurisdictional” is erroneous and confounding terminology where, as here, the relevant time prescription is absent from the U. S. Code. Because Rule 4(a)(5)(C), not §2107, limits the length of the extension granted here, the time prescription is not jurisdictional. See Youkelsone v. FDIC, 660 F. 3d 473, 475 (CADC 2011) (“Rule 4(a)(5)(C)’s thirty-day limit on the length of any extension ultimately granted appears nowhere in the U. S. Code.”). * * * For the reasons stated, the Court of Appeals erroneously treated as jurisdictional Rule 4(a)(5)(C)’s 30-day limitation on extensions of time to file a notice of appeal. We therefore vacate that court’s judgment and remand the case for further proceedings consistent with this opinion. We note, in this regard, that our decision does not reach issues raised by Hamer, but left unaddressed by the Court of Appeals, including: (1) whether respondents’ failure to raise any objection in the District Court to the overlong time extension, by itself, effected a forfeiture, see Brief for Petitioner 21–22; (2) whether respondents could gain review of the District Court’s time extension only by filing their own appeal notice, see id., at 23–27; and (3) whether equitable considerations may occasion an exception to Rule 4(a)(5)(C)’s time constraint, see id., at 29–43. It is so ordered.Notes 1 The terms waiver and forfeiture—though often used interchange-ably by jurists and litigants—are not synonymous. “[F]orfeiture is the failure to make the timely assertion of a right[;] waiver is the ‘intentional relinquishment or abandonment of a known right.’ ” United States v. Olano, 507 U. S. 725, 733 (1993) (quoting Johnson v. Zerbst, 304 U. S. 458, 464 (1938) ). 2 Subject-matter jurisdiction cannot be attacked collaterally, however. Kontrick v. Ryan, 540 U. S. 443, 455, n. 9 (2004) (citing Des Moines Nav. & R. Co. v. Iowa Homestead Co., 123 U. S. 552 –559 (1887)). 3 We have reserved whether mandatory claim-processing rules may be subject to equitable exceptions. See Kontrick, 540 U. S., at 457. 4 Movants were the attorney appointed by the court to represent Hamer and two other attorneys who entered appearances as co-counsel. App. to Pet. for Cert. 57–59. 5 The Court of Appeals incorrectly stated that respondents, answering the Seventh Circuit’s inquiry, asserted that the appeals court “lack[ed] jurisdiction over [Hamer’s] appeal.” 835 F. 3d, at 763. In fact, respondents maintained that “the timeliness of Hamer’s appeal d[id] not appear to be jurisdictional according to [Circuit] law.” App. to Pet. for Cert. 71 (capitalization and footnote omitted). That was so, respondents explained, because “the time limits found [in] Fed. R[ule] App. P[roc.] 4(a)(5)(C) . . . lack a statutory basis.” Id., at 77. Even if not jurisdictional, respondents continued, the Rule is mandatory and must be observed unless forfeited or waived. Ibid. 6 As enacted, the pertinent paragraph of §2107 provided in full: “The district court, in any such action, suit or proceeding, may extend the time for appeal not exceeding thirty days from the expiration of the original time herein prescribed, upon a showing of excusable neglect based on failure of a party to learn of the entry of the judgment, order or decree.” Act of June 25, 1948, §2107, 62Stat. 963. 7 The 14-day prescription cuts back the original limit of 30 days. 8 The statute describes the 14-day extension permitted in lack-of-notice cases as a “reopening [of] the time for appeal.” §2107(c). The “reopening” period is the functional equivalent of an extension. See Brief for American Academy of Appellate Lawyers as Amicus Curiae5–6. 9 In cases not involving the timebound transfer of adjudicatoryauthority from one Article III court to another, we have additionally applied a clear-statement rule: “A rule is jurisdictional ‘[i]f the Legislature clearly states that a threshold limitation on a statute’s scope shall count as jurisdictional.’ ” Gonzalez v. Thaler, 565 U. S. 134, 141 (2012) (quoting Arbaugh v. Y & H Corp., 546 U. S. 500, 515 (2006) ). See also, e.g., Henderson v. Shinseki, 562 U. S. 428, 431 (2011) (statutory deadline for filing notice of appeal with Article I tribunal held not jurisdictional). “This is not to say that Congress must incant magic words in order to speak clearly,” however. Sebelius v. Auburn Regional Medical Center, 568 U. S. 145, 153 (2013) . In determining whether Congress intended a particular provision to be jurisdictional, “[w]e consider ‘context, including this Court’s interpretations of similar provisions in many years past,’ as probative of [Congress’ intent].” Id., at 153–154 (quoting Reed Elsevier, Inc. v. Muchnick, 559 U. S. 154, 168 (2010) ). Even so, “in applying th[e] clear statement rule, we have made plain that most [statutory] time bars are nonjurisdictional.” United States v. Kwai Fun Wong, 575 U. S. ___, ___ (2015) (slip op., at 6). 10 See Freidzon v. OAO LUKOIL, 644 Fed. Appx. 52, 53 (CA2 2016); Peters v. Williams, 353 Fed. Appx. 136, 137 (CA10 2009); United States v. Hawkins, 298 Fed. Appx. 275 (CA4 2008). 11 Indeed, the formulation took flight from a case in which we mistakenly suggested that a claim-processing rule was “mandatory and jurisdictional.” See United States v. Robinson, 361 U. S. 220, 224 (1960) . We have since clarified that “Robinson is correct not because the District Court lacked subject-matter jurisdiction, but because district courts must observe the clear limits of the Rules of Criminal Procedure when they are properly invoked.” Eberhart v. United States, 546 U. S. 12, 17 (2005) (per curiam).

584.US.2017_17-155

In Freeman v. United States, 564 U. S. 522, this Court considered whether a prisoner who had been sentenced under a plea agreement authorized by the Federal Rules of Criminal Procedure could have his sentence reduced under 18 U. S. C. §3582(c)(2) when his Federal Guidelines sentencing range was lowered retroactively. No single interpretation or rationale commanded a majority, however. Some Courts of Appeals, turning to Marks v. United States, 430 U. S. 188, for guidance, adopted the reasoning of Justice Sotomayor’s opinion concurring in the judgment. Others interpreted Marks differently and adopted the plurality’s reasoning. Because this Court can now resolve the substantive, sentencing issue discussed in Freeman, it is unnecessary to reach questions regarding the proper application of Marks. The Sentencing Reform Act of 1984 authorizes the United States Sentencing Commission to establish, and retroactively amend, Sentencing Guidelines. Though the Guidelines are only advisory, see United States v. Booker, 543 U. S. 220, a district court must consult them during sentencing, id., at 264, along with other factors specified in 18 U. S. C. §3553(a), including “the need to avoid unwarranted sentence disparities,” §3553(a)(6). When an amendment applies retroactively, district courts may reduce the sentences of prisoners whose sentences were “based on a sentencing range that has subsequently been lowered by the Sentencing Commission.” §3582(c)(2). This case concerns the issue whether a defendant may seek relief under §3582(c)(2) if he entered a plea agreement under Federal Rule of Criminal Procedure 11(c)(1)(C) (Type-C agreement), which permits the defendant and the Government to “agree that a specific sentence or sentencing range is the appropriate disposition of the case,” and “binds the court [to the agreed-upon sentence] once [it] accepts the plea agreement.” In making its decision, the district court must consider the Sentencing Guidelines. And it may not accept the agreement unless the sentence is within the applicable Guidelines range, or it is outside that range for justifiable reasons specifically set out. After petitioner Erik Hughes was indicted on drug and gun charges, he and the Government negotiated a Type-C plea agreement, which stipulated that Hughes would receive a sentence of 180 months but did not refer to a particular Guidelines range. Hughes pleaded guilty. At his sentencing hearing, the District Court accepted the agreement and sentenced him to 180 months. In so doing, it calculated Hughes’ Guidelines range as 188 to 235 months and determined that the sentence was in accordance with the Guidelines and other factors the court was required to consider. Less than two months later, the Sentencing Commission adopted, and made retroactive, an amendment that had the effect of reducing Hughes’ sentencing range to 151 to 188 months. The District Court denied Hughes’ motion for a reduced sentence under §3582(c)(2), and the Eleventh Circuit affirmed. Both courts concluded that, under the Freeman concurrence, Hughes was ineligible for a reduced sentence because his plea agreement did not expressly rely on a Guidelines range. Held: 1. A sentence imposed pursuant to a Type-C agreement is “based on” the defendant’s Guidelines range so long as that range was part of the framework the district court relied on in imposing the sentence or accepting the agreement. Pp. 7–14. (a) A principal purpose of the Sentencing Guidelines is to promote sentencing uniformity. But in the aftermath of Freeman, a defendant’s eligibility for a reduced sentence under §3582(c)(2) turns on the Circuit in which the case arises. Even within Circuits that follow the Freeman concurrence, unwarranted disparities have resulted depending on whether a defendant’s Type-C agreement has a specific-enough reference to a Guidelines range. This Court’s precedents since Freeman have confirmed that the Guidelines remain the foundation of federal sentencing decisions. See, e.g., Peugh v. United States, 569 U. S. 530; Molina-Martinez v. United States, 578 U. S. ___. Pp. 7–9. (b) A district court imposes a sentence that is “based on” a Guidelines range for purposes of §3582(c)(2) if the range was a basis for the court’s exercise of discretion in imposing a sentence. Given the standard legal definition of “base,” there will be no question in the typical case that the defendant’s Guidelines range was a basis for his sentence. A district court is required to calculate and consider a defendant’s Guidelines range in every case. §3553(a). Indeed, the Guidelines are “the starting point for every sentencing calculation in the federal system.” Peugh, supra, at 542. Thus, in general, §3582(c)(2) allows district courts to reconsider a prisoner’s sentence based on a new starting point—that is, a lower Guidelines range—and determine whether a reduction is appropriate. A sentence imposed pursuant to a Type-C agreement is no exception to the general rule that a defendant’s Guidelines range is the starting point and a basis for his ultimate sentence. The Government and the defendant may agree to a specific sentence, but the Sentencing Guidelines prohibit district courts from accepting Type-C agreements without first evaluating the recommended sentence in light of the defendant’s Guidelines range. So in the usual case the court’s acceptance of a Type-C agreement and the sentence to be imposed pursuant to that agreement are “based on” the defendant’s Guidelines range. Since the Guidelines are a district court’s starting point, when the Commission lowers the range, the defendant will be eligible for relief under §3582(c)(2) absent clear demonstration, based on the record as a whole, that the court would have imposed the same sentence regardless of the Guidelines. This interpretation furthers §3582(c)(2)’s purpose, as well as the broader purposes of the Sentencing Reform Act. It is also reinforced by Molina-Martinez and Peugh, which both confirm that the Guidelines remain a basis for almost all federal sentences. Experience has shown that, although the interpretation proffered by Justice Sotomayor’s concurring opinion in Freeman could be one permissible reading of §3582(c)(2), as a systemic, structural matter the system Congress put in place is best implemented by the interpretation confirmed in this case. Pp. 9–12. (c) The Government’s counterarguments—that allowing defendants with Type-C agreements to seek reduced sentences under §3582(c)(2) would deprive the Government of a benefit of its bargain, namely, the defendant’s agreement to a particular sentence; and that allowing courts to reduce the sentences of defendants like Hughes would be inconsistent with one of the Commission’s policy statements—are unpersuasive. Pp. 12–14. 2. Hughes is eligible for relief under §3582(c)(2). The District Court accepted his Type-C agreement after concluding that a 180-month sentence was consistent with the Guidelines, and then calculated Hughes’ sentencing range and imposed a sentence it deemed “compatible” with the Guidelines. The sentencing range was thus a basis for the sentence imposed. And that range has since been lowered by the Commission. The District Court has discretion to decide whether to reduce Hughes’ sentence after considering the §3553(a) factors and the Commission’s relevant policy statements. P. 14. 849 F. 3d 1008, reversed and remanded. Kennedy, J., delivered the opinion of the Court, in which Ginsburg, Breyer, Sotomayor, Kagan, and Gorsuch, JJ., joined. Sotomayor, J., filed a concurring opinion. Roberts, C. J., filed a dissenting opinion, in which Thomas and Alito, JJ., joined.

The proper construction of federal sentencing statutes and the Federal Rules of Criminal Procedure can present close questions of statutory and textual interpretation when implementing the Federal Sentencing Guidelines. Seven Terms ago the Court considered one of these issues in a case involving a prisoner’s motion to reduce his sentence, where the prisoner had been sentenced under a plea agreement authorized by a specific Rule of criminal procedure. Freeman v. United States, 564 U. S. 522 (2011). The prisoner maintained that his sentence should be reduced under 18 U. S. C. §3582(c)(2) when his Guidelines sentencing range was lowered retroactively. 564 U. S., at 527–528 (plurality opinion). No single interpretation or rationale in Freeman commanded a majority of the Court. The courts of appeals then confronted the question of what principle or principles considered in Freeman controlled when an opinion by four Justices and a concurring opinion by a single Justice had allowed a majority of this Court to agree on the judgment in Freeman but not on one interpretation or rule that courts could follow in later cases when similar questions arose under the same statute and Rule. For guidance courts turned to this Court’s opinion in Marks v. United States, 430 U. S. 188 (1977). Some courts interpreted Marks as directing them to follow the “narrowest” opinion in Freeman that was necessary for the judgment in that case; and, accordingly, they adopted the reasoning of the opinion concurring in the judgment by Justice Sotomayor. See United States v. Rivera-Martinez, 665 F. 3d 344, 348 (CA1 2011); United States v. Thompson, 682 F. 3d 285, 290 (CA3 2012); United States v. Brown, 653 F. 3d 337, 340, n. 1 (CA4 2011); United States v. Benitez, 822 F. 3d 807, 811 (CA5 2016); United States v. Smith, 658 F. 3d 608, 611 (CA6 2011); United States v. Dixon, 687 F. 3d 356, 359 (CA7 2012); United States v. Browne, 698 F. 3d 1042, 1045 (CA8 2012); United States v. Graham, 704 F. 3d 1275, 1277–1278 (CA10 2013). In contrast, the Courts of Appeals for the District of Columbia and Ninth Circuits held that no opinion in Freeman provided a controlling rule because the reasoning in the concurrence was not a “logical subset” of the reasoning in the plurality. United States v. Davis, 825 F. 3d 1014, 1021–1022 (CA9 2016) (en banc); United States v. Epps, 707 F. 3d 337, 350 (CADC 2013). Those courts have adopted the plurality’s opinion as the most persuasive interpretation of §3582(c)(2). Davis, supra, at 1026; Epps, supra, at 351. To resolve these differences over the proper application of Marks and the proper interpretation of §3582(c)(2), the Court granted certiorari in the present case. 583 U. S. ___ (2017). The first two questions, relating to Marks, are as follows: (1) “Whether this Court’s decision in Marks means that the concurring opinion in a 4–1–4 decision represents the holding of the Court where neither the plurality’s reasoning nor the concurrence’s reasoning is a logical subset of the other”; and (2) “Whether, under Marks, the lower courts are bound by the four-Justice plurality opinion in Freeman, or, instead, by Justice Sotomayor’s separate concurring opinion with which all eight other Justices disagreed.” Pet. for Cert. i. The third question is directed to the underlying statu- tory issue in this case, the substantive, sentencing issue the Court discussed in the three opinions issued in Freeman. That question is: “Whether, as the four-Justice plurality in Freeman concluded, a defendant who enters into a Fed. R. Crim. P. 11(c)(1)(C) plea agreement is generally eligible for a sentence reduction if there is a later, retroactive amendment to the relevant Sentencing Guidelines range.” Pet. for Cert. ii. Taking instruction from the cases decided in the wake of Freeman and the systemic concerns that have arisen in some Circuits, and considering as well the arguments of the parties as to question three, a majority of the Court in the instant case now can resolve the sentencing issue on its merits. So it will be unnecessary to consider questions one and two despite the extensive briefing and careful argument the parties presented to the Court concerning the proper application of Marks. The opinion that follows resolves the sentencing issue in this case; and, as well, it should give the necessary guidance to federal district courts and to the courts of appeals with respect to plea agreements of the kind presented here and in Freeman. With that explanation, the Court now turns to the circumstances of this case and the sentencing issue it presents. I A Under the Sentencing Reform Act of 1984, the United States Sentencing Commission establishes Sentencing Guidelines based on the seriousness of a defendant’s offense and his criminal history. Dillon v. United States, 560 U. S. 817, 820 (2010). In combination, these two factors yield a range of potential sentences for a district court to choose from in sentencing a particular defendant. “The Sentencing Guidelines provide the framework for the tens of thousands of federal sentencing proceedings that occur each year.” Molina-Martinez v. United States, 578 U. S. ___, ___ (2016) (slip op., at 2). After this Court’s decision in United States v. Booker, 543 U. S. 220 (2005), the Guidelines are advisory only. But a district court still “must consult those Guidelines and take them into account when sentencing.” Id., at 264; see also 18 U. S. C. §3553(a)(4). Courts must also consider various other sentencing factors listed in §3553(a), including “the need to avoid unwarranted sentence disparities among defendants with similar records who have been found guilty of similar conduct.” §3553(a)(6). The Act requires the Commission to review and revise the Guidelines from time to time. 28 U. S. C. §994(o). When the Commission amends the Guidelines in a way that reduces the Guidelines range for “a particular offense or category of offenses,” the Commission must “specify in what circumstances and by what amount the sentences of prisoners serving terms of imprisonment for the offense may be reduced.” §994(u). In this way the Act requires the Commission to decide whether amendments to the Guidelines should have retroactive effect. If an amendment applies retroactively, the Act authorizes district courts to reduce the sentences of prisoners who were sentenced based on a Guidelines range that would have been lower had the amendment been in place when they were sentenced. 18 U. S. C. §3582(c)(2). Specifically, §3582(c)(2) provides: “[I]n the case of a defendant who has been sentenced to a term of imprisonment based on a sentencing range that has subsequently been lowered by the Sentencing Commission pursuant to 28 U. S. C. §994(o), . . . the court may reduce the term of imprisonment, after considering the factors set forth in section 3553(a) to the extent that they are applicable, if such a reduction is consistent with applicable policy statements issued by the Sentencing Commission.” B The controlling issue here is whether a defendant may seek relief under §3582(c)(2) if he entered a plea agreement specifying a particular sentence under Federal Rule of Criminal Procedure 11(c)(1)(C). This kind of plea agreement is sometimes referred to as a “Type-C agreement.” In a Type-C agreement the Government and a defendant “agree that a specific sentence or sentencing range is the appropriate disposition of the case, or that a particular provision of the Sentencing Guidelines, or policy statement, or sentencing factor does or does not apply,” and “such a recommendation or request binds the court once the court accepts the plea agreement.” Rule 11(c)(1)(C). When the Government and a defendant enter a Type-C agreement, the district court has three choices: It “may accept the agreement, reject it, or defer a decision until the court has reviewed the presentence report.” Rule 11(c)(3)(A). If the court rejects the agreement, the defendant may withdraw his guilty plea. Rule 11(c)(5)(B). In deciding whether to accept an agreement that includes a specific sentence, the district court must consider the Sentencing Guidelines. The court may not accept the agreement unless the court is satisfied that “(1) the agreed sentence is within the applicable guideline range; or (2)(A) the agreed sentence is outside the applicable guideline range for justifiable reasons; and (B) those reasons are set forth with specificity.” United States Sentencing Commission, Guidelines Manual §6B1.2(c) (Nov. 2016) (USSG). “[T]he decision whether to accept the agreement will often be deferred until the sentencing hearing,” which means that “the decision whether to accept the plea agreement will often be made at the same time that the defendant is sentenced.” United States v. Hyde, 520 U. S. 670, 678 (1997). C 1 In 2013 petitioner Erik Hughes was indicted on drug and gun charges for his participation in a conspiracy to distribute methamphetamine. About four months later, the Government and Hughes negotiated a Type-C plea agreement. Hughes agreed to plead guilty to two of the four charges (conspiracy to distribute methamphetamine and being a felon in possession of a gun); and in exchange the Government agreed to dismiss the other two charges and to refrain from filing an information giving formal notification to the District Court of his prior drug felonies. If the Government had filed the information, Hughes would have been subject to a mandatory sentence of life in prison. See 21 U. S. C. §§841(b)(1)(A), 851(a). The agreement stipulated that Hughes would receive a sentence of 180 months, but it did not refer to any particular Guidelines range. Hughes entered his guilty plea in December 2013. The District Court accepted the plea at that time, but it deferred consideration of the plea agreement (and hence the stipulated 180-month sentence) until sentencing. Three months later, at the sentencing hearing, the District Court accepted the agreement and sentenced Hughes to 180 months in prison. The court stated that it had “considered the plea agreement [and] the sentencing guidelines, particularly the provisions of [§3553(a)],” and that it would “accept and approve the binding plea agreement.” App. to Pet. for Cert. 32a–33a. The court calcu- lated Hughes’ Guidelines range as 188 to 235 months in prison and heard statements from Hughes’ daughter, mother, and Hughes himself. Id., at 37a–43a. When it imposed the agreed 180-month sentence the court reiterated that it was “a reasonable sentence in this case compatible with the advisory United States Sentencing Guidelines but in accordance with the mandatory matters the Court is required to consider in ultimately determining a sentence.” Id., at 44a, 47a. 2 Less than two months after the District Court sentenced Hughes, the Sentencing Commission adopted amendment 782 to the Guidelines. USSG App. C, Amdt. 782 (Supp. Nov. 2012–Nov. 2016). The amendment reduced the base offense level by two levels for most drug offenses. The Commission later made amendment 782 retroactive for defendants who, like Hughes, already had been sentenced under the higher offense levels. Amdt. 788. Under the revised Guidelines, Hughes’ sentencing range is 151 to 188 months—about three to four years lower than the range in effect when he was sentenced. Hughes filed a motion for a reduced sentence under §3582(c)(2). The District Court denied the motion, concluding that Hughes is ineligible for relief; and the Court of Appeals for the Eleventh Circuit affirmed. 849 F. 3d 1008, 1016 (2017); App. to Pet. for Cert. 28a. Both courts concluded that the Freeman concurrence stated the holding of this Court under Marks, and that under the concurrence’s interpretation Hughes was ineligible for a reduced sentence because his plea agreement did not expressly rely on a Guidelines range. 849 F. 3d, at 1015; App. to Pet. for Cert. 25a. This Court granted certiorari. 583 U. S. ___. II A principal purpose of the Sentencing Guidelines is to promote “uniformity in sentencing imposed by different federal courts for similar criminal conduct.” Molina-Martinez, 578 U. S., at ___ (slip op., at 2) (internal quotation marks and alteration omitted; emphasis deleted). Yet in the aftermath of Freeman, a defendant’s eligibility for a reduced sentence under §3582(c)(2) turns on the Circuit in which the case arises. Further, even within Circuits that follow the Freeman concurrence, unwarranted disparities have resulted depending on the fortuity of whether a defendant’s Type-C agreement includes a specific-enough reference to a Guidelines range. See Brief for National Association of Criminal Defense Lawyers et al. as Amici Curiae 13–20. In some cases defendants have been held ineligible for relief even where the sentencing hearing makes it crystal clear that the Government and the defendant agreed to a Guidelines sentence and the district court imposed one. See, e.g., United States v. McNeese, 819 F. 3d 922, 929 (CA6 2016). In addition this Court’s precedents since Freeman have further confirmed that the Guidelines remain the foundation of federal sentencing decisions. In Peugh v. United States, 569 U. S. 530 (2013), for example, the Court held that the Ex Post Facto Clause prohibits retroactive application of amended Guidelines that increase a defendant’s sentencing range. Id., at 544. The Court reasoned that, Booker notwithstanding, the Guidelines remain “the lodestone of sentencing.” 569 U. S., at 544. And in Molina-Martinez, the Court held that in the ordinary case a defendant suffers prejudice from a Guidelines error because of “the systemic function of the selected Guidelines range.” 578 U. S., at ___ (slip op., at 10). “The post-Booker federal sentencing scheme aims to achieve uniformity by ensuring that sentencing decisions are anchored by the Guidelines.” Peugh, supra, at 541. In this context clarity and consistency are essential. To resolve the uncertainty that resulted from this Court’s divided decision in Freeman, the Court now holds that a sentence imposed pursuant to a Type-C agreement is “based on” the defendant’s Guidelines range so long as that range was part of the framework the district court relied on in imposing the sentence or accepting the agreement. A As already mentioned, §3582(c)(2) authorizes a district court to reduce a defendant’s sentence if the defendant “has been sentenced to a term of imprisonment based on a sentencing range that has subsequently been lowered by the Sentencing Commission.” A district court imposes a sentence that is “based on” a Guidelines range if the range was a basis for the court’s exercise of discretion in imposing a sentence. To “base” means “[t]o make, form, or serve as a foundation for,” or “[t]o use (something) as the thing from which something else is developed.” Black’s Law Dictionary 180 (10th ed. 2014). Likewise, a “base” is “[t]he starting point or foundational part of something,” or “[a] point, part, line, or quantity from which a reckoning or conclusion proceeds.” Ibid.; see also ibid. (similarly defining “basis”). In the typical sentencing case there will be no question that the defendant’s Guidelines range was a basis for his sentence. The Sentencing Reform Act requires a district court to calculate and consider a defendant’s Guidelines range in every case. 18 U. S. C. §3553(a). Indeed, the Guidelines are “the starting point for every sentencing calculation in the federal system.” Peugh, supra, at 542; see also Molina-Martinez, 578 U. S., at ___ (slip op., at 9) (“The Court has made clear that the Guidelines are to be the sentencing court’s starting point and initial benchmark” (internal quotation marks and alteration omitted)). “Even if the sentencing judge sees a reason to vary from the Guidelines, if the judge uses the sentencing range as the beginning point to explain the decision to deviate from it, then the Guidelines are in a real sense the basis for the sentence.” Ibid. (internal quotation marks omitted; emphasis deleted). In general, §3582(c)(2) allows district courts to reconsider a prisoner’s sentence based on a new starting point—that is, a lower Guidelines range—and determine whether a reduction in the prisoner’s sentence is appropriate. A sentence imposed pursuant to a Type-C agreement is no exception to the general rule that a defendant’s Guidelines range is both the starting point and a basis for his ultimate sentence. Although in a Type-C agreement the Government and the defendant may agree to a specific sentence, that bargain is contingent on the district court accepting the agreement and its stipulated sentence. Freeman, 564 U. S., at 529–530. The Sentencing Guidelines prohibit district courts from accepting Type-C agreements without first evaluating the recommended sentence in light of the defendant’s Guidelines range. USSG §6B1.2(c). So in the usual case the court’s acceptance of a Type-C agreement and the sentence to be imposed pursuant to that agreement are “based on” the defendant’s Guidelines range. To be sure, the Guidelines are advisory only, and so not every sentence will be consistent with the relevant Guidelines range. See Koons v. United States, post, at 5 (defendants’ Guidelines ranges “clearly did not” form a basis of the ultimate sentences). For example, in Koons the Court today holds that five defendants’ sentences were not “based on” subsequently lowered Guidelines ranges because in that case the Guidelines and the record make clear that the sentencing judge “discarded” their sen- tencing ranges “in favor of mandatory minimums and substantial-assistance factors.” Post, at 5–6; see also Molina-Martinez, supra, at ___ (slip op., at 11) (“The record in a case may show, for example, that the district court thought the sentence it chose was appropriate irrespective of the Guidelines range”). If the Guidelines range was not “a relevant part of the analytic framework the judge used to determine the sentence or to approve the agreement,” Freeman, supra, at 530, then the defendant’s sentence was not based on that sentencing range, and relief under §3582(c)(2) is unavail- able. And that is so regardless of whether a defendant pleaded guilty pursuant to a Type-C agreement or whether the agreement itself referred to a Guidelines range. The statutory language points to the reasons for the sentence that the district court imposed, not the reasons for the parties’ plea agreement. Still, cases like Koons are a narrow exception to the general rule that, in most cases, a defendant’s sentence will be “based on” his Guidelines range. In federal sentencing the Guidelines are a district court’s starting point, so when the Commission lowers a defendant’s Guidelines range the defendant will be eligible for relief under §3582(c)(2) absent clear demonstration, based on the record as a whole, that the court would have imposed the same sentence regardless of the Guidelines. See Koons, post, at 4–6. This interpretation furthers §3582(c)(2)’s purpose, as well as the broader purposes of the Sentencing Reform Act. “The Act aims to create a comprehensive sentencing scheme in which those who commit crimes of similar severity under similar conditions receive similar sentences.” Freeman, 564 U. S., at 533. “Section 3582(c)(2) contri- butes to that goal by ensuring that district courts may adjust sentences imposed pursuant to a range that the Commission concludes [is] too severe, out of step with the seriousness of the crime and the sentencing ranges of analogous offenses, and inconsistent with the Act’s purposes.” Ibid. And there is no reason a defendant’s eligibility for relief should turn on the form of his plea agreement. Two cases decided after Freeman now reinforce this proposition. See Molina-Martinez, 578 U. S., at ___–___ (slip op., at 9–11); Peugh, 569 U. S., at 541–544. These cases confirm that the Guidelines remain a basis for almost all federal sentences. In Peugh, the Court recognized that “[e]ven after Booker rendered the Sentencing Guidelines advisory, district courts have in the vast majority of cases imposed either within-Guidelines sentences or sentences that depart downward from the Guidelines on the Government’s motion.” Id., at 543. And in Molina-Martinez, the Court explained that “[t]he Commission’s statistics demonstrate the real and pervasive effect the Guidelines have on sentencing.” 578 U. S., at ___ (slip op., at 10). In short, experience has shown that, although the interpretation proffered by Justice Sotomayor’s concurring opinion in Freeman could be one permissible reading of §3582(c)(2), the system Congress put in place is best implemented, as a systemic, structural matter, by the interpretation confirmed in the instant case. B In response, the Government largely recycles arguments that a majority of this Court rejected in Freeman. For example, the Government contends that allowing defendants who enter Type-C agreements to seek reduced sentences under §3582(c)(2) would deprive the Government of one of the benefits of its bargain—namely, the defendant’s agreement to a particular sentence. But that has nothing to do with whether a defendant’s sentence was based on the Sentencing Guidelines under §3582(c)(2). Freeman, 564 U. S., at 531; see also id., at 540 (opinion of Sotomayor, J.). And in any event, “[w]hat is at stake in this case is a defendant’s eligibility for relief, not the extent of that relief.” Id., at 532 (plurality opinion). Even if a defendant is eligible for relief, before a district court grants a reduction it must consider “the factors set forth in section 3553(a) to the extent that they are applicable” and the Commission’s “applicable policy statements.” §3582(c)(2). The district court can consider the benefits the defendant gained by entering a Type-C agreement when it decides whether a reduction is appropriate (or when it determines the extent of any reduction), “for the statute permits but does not require the court to reduce a sentence.” Freeman, supra, at 532. The Government also contends that allowing courts to reduce the sentences of defendants like Hughes would be inconsistent with the Commission’s policy statement in USSG §1B1.10, which provides that when a district court modifies a sentence under §3582(c)(2) it “shall substitute only the [retroactive] amendments listed in subsection (d) for the corresponding guidelines provisions that were applied when the defendant was sentenced and shall leave all other guideline application decisions unaffected.” USSG §1B1.10(b)(1). According to the Government, no “guidelines provisions” are “applied” when a defendant enters a Type-C agreement because at the moment of sentencing—that is, after the court has already accepted the agreement—Rule 11 prohibits the court from imposing any sentence other than the one the parties bargained for. This argument fails for at least two reasons. First, the Government’s interpretation of §1B1.10 depends on an artificial distinction between a court’s decision to accept a Type-C agreement and its decision to impose the agreed-upon sentence. As explained above, a district court must consider the defendant’s “applicable Guidelines range” when it decides whether to accept or reject the agreement, USSG §6B1.2(c)—often, as here, at the sentencing hearing, after the court has reviewed the presentence report. And as the Government itself points out, once the district court accepts the agreement, the agreed-upon sentence is the only sentence the court may impose. Thus, there is no meaningful difference between a court’s decision to accept a Type-C agreement that includes a particular sentence and the court’s decision (sometimes, as here, just minutes later) to impose that sentence. Second, the Commission’s policy statement “seeks to isolate whatever marginal effect the since-rejected Guideline had on the defendant’s sentence.” Freeman, 564 U. S., at 530. Accordingly, relief under §3582(c)(2) should be available to permit the district court to reconsider a prior sentence to the extent the prisoner’s Guidelines range was a relevant part of the framework the judge used to accept the agreement or determine the sentence. Ibid. If the district court concludes that it would have imposed the same sentence even if the defendant had been subject to the lower range, then the court retains discretion to deny relief. C In this case the District Court accepted Hughes’ Type-C agreement after concluding that a 180-month sentence was consistent with the Sentencing Guidelines. App. to Pet. for Cert. 33a. The court then calculated Hughes’ sentencing range and imposed a sentence that the court deemed “compatible” with the Guidelines. Id., at 36a, 47a. Thus, the sentencing range was a basis for the sentence that the District Court imposed. That range has “subsequently been lowered by the Sentencing Commission,” so Hughes is eligible for relief under §3582(c)(2). The Court expresses no view as to whether the District Court should exercise its discretion to reduce Hughes’ sentence after considering the §3553(a) factors and the Commission’s relevant policy statements. See 18 U. S. C. §3582(c)(2). * * * For these reasons, the judgment of the Court of Appeals is reversed, and the case is remanded for further proceedings consistent with this opinion. It is so ordered.

584.US.2017_16-980

The National Voter Registration Act (NVRA) addresses the removal of ineligible voters from state voting rolls, 52 U. S. C. §20501(b), including those who are ineligible “by reason of” a change in residence, §20507(a)(4). The Act prescribes requirements that a State must meet in order to remove a name on change-of-residence grounds, §§20507(b), (c), (d). The most relevant of these are found in subsection (d), which provides that a State may not remove a name on change-of-residence grounds unless the registrant either (A) confirms in writing that he or she has moved or (B) fails to return a preaddressed, postage prepaid “return card” containing statutorily prescribed content and then fails to vote in any election during the period covering the next two general federal elections. In addition to these specific change-of-residence requirements, the NVRA also contains a general “Failure-to-Vote Clause,” §20507(b)(2), consisting of two parts. It first provides that a state removal program “shall not result in the removal of the name of any person . . . by reason of the person’s failure to vote.” Second, as added by the Help America Vote Act of 2002 (HAVA), it specifies that “nothing in [this prohibition] may be construed to prohibit a State from using the procedures” described above—sending a return card and removing registrants who fail to return the card and fail to vote for the requisite time. Since one of the requirements for removal under subsection (d) is the failure to vote, the explanation added by HAVA makes clear that the Failure-to-Vote Clause’s prohibition on removal “by reason of the person’s failure to vote” does not categorically preclude using nonvoting as part of a test for removal. Another provision makes this point even more clearly by providing that “no registrant may be removed solely by reason of a failure to vote.” §21083(a)(4)(A) (emphasis added). Respondents contend that Ohio’s process for removing voters on change-of-residence grounds violates this federal law. The Ohio process at issue relies on the failure to vote for two years as a rough way of identifying voters who may have moved. It sends these nonvoters a preaddressed, postage prepaid return card, asking them to verify that they still reside at the same address. Voters who do not return the card and fail to vote in any election for four more years are presumed to have moved and are removed from the rolls. Held:The process that Ohio uses to remove voters on change-of-residence grounds does not violate the Failure-to-Vote Clause or any other part of the NVRA. Pp. 8–21. (a) Ohio’s law does not violate the Failure-to-Vote Clause. Pp. 8–16. (1) Ohio’s removal process follows subsection (d) to the letter: It does not remove a registrant on change-of-residence grounds unless the registrant is sent and fails to mail back a return card and then fails to vote for an additional four years. See §20507(d)(1)(B). Pp. 8–9. (2) Nonetheless, respondents argue that Ohio’s process violates subsection (b)’s Failure-to-Vote Clause by using a person’s failure to vote twice over: once as the trigger for sending return cards and again as one of the two requirements for removal. But Congress could not have meant for the Failure-to-Vote Clause to cannibalize subsection (d) in that way. Instead, the Failure-to-Vote Clause, both as originally enacted in the NVRA and as amended by HAVA, simply forbids the use of nonvoting as the sole criterion for removing a registrant, and Ohio does not use it that way. The phrase “by reason of” in the Failure-to-Vote Clause denotes some form of causation, see Gross v. FBL Financial Services, Inc., 557 U. S. 167, 176, and in context sole causation is the only type of causation that harmonizes the Failure-to-Vote Clause and subsection (d). Any other reading would mean that a State that follows subsection (d) nevertheless can violate the Failure-to-Vote Clause. When Congress enacted HAVA, it made this point explicit by adding to the Failure-to-Vote Clause an explanation of how the clause is to be read, i.e., in a way that does not contradict subsection (d). Pp. 9–12. (3) Respondents’ and the dissent’s alternative reading is inconsistent with both the text of the Failure-to-Vote Clause and the clarification of its meaning in §21083(a)(4). Among other things, their reading would make HAVA’s new language worse than redundant, since no sensible person would read the Failure-to-Vote Clause as prohibiting what subsections (c) and (d) expressly allow. Nor does the Court’s interpretation render the Failure-to-Vote Clause superfluous; the clause retains meaning because it prohibits States from using nonvoting both as the ground for removal and as the sole evidence for another ground for removal (e.g., as the sole evidence that someone has died). Pp. 12–15. (4) Respondents’ additional argument—that so many registered voters discard return cards upon receipt that the failure to send cards back is worthless as evidence that an addressee has moved—is based on a dubious empirical conclusion that conflicts with the congressional judgment found in subsection (d). Congress clearly did not think that the failure to send back a return card was of no evidentiary value, having made that conduct one of the two requirements for removal under subsection (d). Pp. 15–16. (b) Nor has Ohio violated other NVRA provisions. Pp. 16–21. (1) Ohio removes the registrants at issue on a permissible ground: change of residence. The failure to return a notice and the failure to vote simply serve as evidence that a registrant has moved, not as the ground itself for removal. Pp. 16–17. (2) The NVRA contains no “reliable indicator” prerequisite to sending notices, requiring States to have good information that someone has moved before sending them a return card. So long as the trigger for sending such notices is “uniform, nondiscriminatory, and in compliance with the Voting Rights Act,” §20507(b)(1), States may use whatever trigger they think best, including the failure to vote. Pp. 17–19. (3) Ohio has not violated the NVRA’s “reasonable effort” provision, §20507(a)(4). Even assuming that this provision authorizes federal courts to go beyond the restrictions set out in subsections (b), (c), and (d) and strike down a state law that does not meet some standard of “reasonableness,” Ohio’s process cannot be unreasonable because it uses the change-of-residence evidence that Congress said it could: the failure to send back a notice coupled with the failure to vote for the requisite period. Ohio’s process is accordingly lawful. Pp. 19–21. 838 F. 3d 699, reversed. Alito, J., delivered the opinion of the Court, in which Roberts, C. J., and Kennedy, Thomas, and Gorsuch, JJ., joined. Thomas, J., filed a concurring opinion. Breyer, J., filed a dissenting opinion, in which Ginsburg, Sotomayor, and Kagan, JJ., joined. Sotomayor, J., filed a dissenting opinion.

It has been estimated that 24 million voter registrations in the United States—about one in eight—are either invalid or significantly inaccurate. Pew Center on the States, Election Initiatives Issue Brief (Feb. 2012). And about 2.75 million people are said to be registered to vote in more than one State. Ibid. At issue in today’s case is an Ohio law that aims to keep the State’s voting lists up to date by removing the names of those who have moved out of the district where they are registered. Ohio uses the failure to vote for two years as a rough way of identifying voters who may have moved, and it then sends a preaddressed, postage prepaid card to these individuals asking them to verify that they still reside at the same address. Voters who do not return this card and fail to vote in any election for four more years are presumed to have moved and are removed from the rolls. We are asked to decide whether this program complies with federal law. I A Like other States, Ohio requires voters to reside in the district in which they vote. Ohio Rev. Code Ann. §3503.01(A) (West Supp. 2017); see National Conference of State Legislatures, Voting by Nonresidents and Non- citizens (Feb. 27, 2015). When voters move out of that district, they become ineligible to vote there. See §3503.01(A). And since more than 10% of Americans move every year,[1] deleting the names of those who have moved away is no small undertaking. For many years, Congress left it up to the States to maintain accurate lists of those eligible to vote in federal elections, but in 1993, with the enactment of the National Voter Registration Act (NVRA), Congress intervened. The NVRA “erect[s] a complex superstructure of federal regulation atop state voter-registration systems.” Arizona v. Inter Tribal Council of Ariz., Inc., 570 U. S. 1, 5 (2013). The Act has two main objectives: increasing voter registration and removing ineligible persons from the States’ voter registration rolls. See §2, 107Stat. 77, 52 U. S. C. §20501(b). To achieve the latter goal, the NVRA requires States to “conduct a general program that makes a reasonable effort to remove the names” of voters who are ineligible “by reason of” death or change in residence. §20507(a)(4). The Act also prescribes requirements that a State must meet in order to remove a name on change-of-residence grounds. §§20507(b), (c), (d). The most important of these requirements is a prior notice obligation. Before the NVRA, some States removed registrants without giving any notice. See J. Harris, Nat. Munic. League, Model Voter Registration System 45 (rev. 4th ed. 1957). The NVRA changed that by providing in §20507(d)(1) that a State may not remove a registrant’s name on change-of-residence grounds unless either (A) the registrant confirms in writing that he or she has moved or (B) the registrant fails to return a preaddressed, postage prepaid “return card” containing statutorily prescribed content. This card must explain what a registrant who has not moved needs to do in order to stay on the rolls, i.e., either return the card or vote during the period covering the next two general federal elections. §20507(d)(2)(A). And for the benefit of those who have moved, the card must contain “information concerning how the registrant can continue to be eligible to vote.” §20507(d)(2)(B). If the State does not send such a card or otherwise get written notice that the person has moved, it may not remove the registrant on change-of-residence grounds. See §20507(d)(1).[2] While the NVRA is clear about the need to send a “return card” (or obtain written confirmation of a move) before pruning a registrant’s name, no provision of federal law specifies the circumstances under which a return card may be sent. Accordingly, States take a variety of approaches. See Nat. Assn. of Secretaries of State (NASS) Report: Maintenance of State Voter Registration Lists 5–6 (Dec. 2017). The NVRA itself sets out one option. A State may send these cards to those who have submitted “change-of-address information” to the United States Postal Service. §20507(c)(1). Thirty-six States do at least that. See NASS Report, supra, at 5, and n. v (listing States). Other States send notices to every registered voter at specified intervals (say, once a year). See, e.g., Iowa Code §48 A. 28.3 (2012); S. C. Code Ann. §§7–5–330(F), 7–5–340(2)–(3) (2017 Cum. Supp.); see also S. Rep. No. 103–6, p. 46 (1993). Still other States, including Ohio, take an intermediate approach, see NASS Report, supra, at 5–6, such as sending notices to those who have turned in their driver’s licenses, e.g., Ind. Code §§3–7–38.2–2(b)(2), (c)(4) (2004), or sending notices to those who have not voted for some period of time, see, e.g., Ga. Code Ann. §21–2–234 (Supp. 2017); Ohio Rev. Code Ann. §3503.21(B)(2); Okla. Admin. Code §230:15–11–19(a)(3) (2016); Pa. Stat. Ann., Tit. 25, §1901(b)(3) (Purdon 2007); Wis. Stat. Ann. §6.50(1) (2017 West Cum. Supp.). When a State receives a return card confirming that a registrant has left the district, the State must remove the voter’s name from the rolls. §§20507(d)(1)(A), (3). And if the State receives a card stating that the registrant has not moved, the registrant’s name must be kept on the list. See §20507(d)(2)(A). What if no return card is mailed back? Congress obviously anticipated that some voters who received cards would fail to return them for any number of reasons, and it addressed this contingency in §20507(d), which, for convenience, we will simply call “subsection (d).” Subsection (d) treats the failure to return a card as some evidence—but by no means conclusive proof—that the voter has moved. Instead, the voter’s name is kept on the list for a period covering two general elections for federal office (usually about four years). Only if the registrant fails to vote during that period and does not otherwise confirm that he or she still lives in the district (e.g., by updating address information online) may the registrant’s name be removed. §20507(d)(2)(A); see §§20507(d)(1)(B), (3). In addition to these specific change-of-residence requirements, the NVRA also imposes two general limitations that are applicable to state removal programs. First, all such programs must be “uniform, nondiscriminatory, and in compliance with the Voting Rights Act of 1965.” §20507(b)(1). Second, the NVRA contains what we will call the “Failure-to-Vote Clause.” See §20507(b)(2). At present, this clause contains two parts. The first is a prohibition that was included in the NVRA when it was originally enacted in 1993. It provides that a state program “shall not result in the removal of the name of any person . . . by reason of the person’s failure to vote.” Ibid. The second part, added by the Help America Vote Act of 2002 (HAVA), 116Stat. 1666, explains the meaning of that prohibition. This explanation says that “nothing in [the prohibition] may be construed to prohibit a State from using the procedures described in [§§20507](c) and (d) to remove an individual from the official list of eligible voters.” §20507(b)(2). These referenced subsections, §§20507(c) and (d), are the provisions allowing the removal of registrants who either submitted change-of-address information to the Postal Service (subsection (c)) or did not mail back a return card and did not vote during a period covering two general federal elections (subsection (d)). And since one of the requirements for removal under subsection (d) is the failure to vote during this period, the explanation added by HAVA in 2002 makes it clear that the statutory phrase “by reason of the person’s failure to vote” in the Failure-to-Vote Clause does not categorically preclude the use of nonvoting as part of a test for removal. Another provision of HAVA makes this point more directly. After directing that “registrants who have not responded to a notice and . . . have not voted in 2 consecutive general elections for Federal office shall be removed,” it adds that “no registrant may be removed solely by reason of a failure to vote.” §21083(a)(4)(A) (emphasis added). B Since 1994, Ohio has used two procedures to identify and remove voters who have lost their residency qualification. First, the State utilizes the Postal Service option set out in the NVRA. The State sends notices to registrants whom the Postal Service’s “national change of address service” identifies as having moved. Ohio Rev. Code Ann. §3503.21(B)(1). This procedure is undisputedly lawful. See 52 U. S. C. §20507(c)(1). But because according to the Postal Service “[a]s many as 40 percent of people who move do not inform the Postal Service,”[3] Ohio does not rely on this information alone. In its so-called Supplemental Process, Ohio “identif[ies] electors whose lack of voter activity indicates they may have moved.” Record 401 (emphasis deleted). Under this process, Ohio sends notices to registrants who have “not engage[d] in any voter activity for a period of two consecutive years.” Id., at 1509. “Voter activity” includes “casting a ballot” in any election—whether general, primary, or special and whether federal, state, or local. See id., at 1507. (And Ohio regularly holds elections on both even and odd years.) Moreover, the term “voter activity” is broader than simply voting. It also includes such things as “sign[ing] a petition,” “filing a voter registration form, and updating a voting address with a variety of [state] entities.” Id., at 295, 357. After sending these notices, Ohio removes registrants from the rolls only if they “fai[l] to respond” and “continu[e] to be inactive for an additional period of four consecutive years, including two federal general elections.” Id., at 1509; see Ohio Rev. Code Ann. §3503.21(B)(2). Federal law specifies that a registration may be canceled if the registrant does not vote “in an election during the period” covering two general federal elections after notice, §20507(d)(1)(B)(ii), but Ohio rounds up to “four consecutive years” of nonvoting after notice, Record 1509. Thus, a person remains on the rolls if he or she votes in any election during that period—which in Ohio typically means voting in any of the at least four elections after notice. Combined with the two years of nonvoting before notice is sent, that makes a total of six years of nonvoting before removal. Ibid. C A pair of advocacy groups and an Ohio resident (respondents here) think that Ohio’s Supplemental Process violates the NVRA and HAVA. They sued petitioner, Ohio’s Secretary of State, seeking to enjoin this process. Respondents alleged, first, that Ohio removes voters who have not actually moved, thus purging the rolls of eligible voters. They also contended that Ohio violates the NVRA’s Failure-to-Vote Clause because the failure to vote plays a prominent part in the Ohio removal scheme: Failure to vote for two years triggers the sending of a return card, and if the card is not returned, failure to vote for four more years results in removal. The District Court rejected both of these arguments and entered judgment for the Secretary. It held that Ohio’s Supplemental Process “mirror[s] the procedures established by the NVRA” for removing people on change-of-residence grounds and does not violate the Failure-to-Vote Clause because it does not remove anyone “solely for [their] failure to vote.” App. to Pet. for Cert. 43a, 57a, 69a–70a. A divided panel of the Court of Appeals for the Sixth Circuit reversed. 838 F. 3d 699 (2016). It focused on respondents’ second argument, holding that Ohio violates the Failure-to-Vote Clause because it sends change-of-residence notices “based ‘solely’ on a person’s failure to vote.” Id., at 711. In dissent, Judge Siler explained why he saw the case as a simple one: “The State cannot remove the registrant’s name from the rolls for a failure to vote only, and Ohio does not do [that].” Id., at 716. We granted certiorari, 581 U. S. ___ (2017), and now reverse. II A As noted, subsection (d), the provision of the NVRA that directly addresses the procedures that a State must follow before removing a registrant from the rolls on change-of-residence grounds, provides that a State may remove a registrant who “(i) has failed to respond to a notice” and “(ii) has not voted or appeared to vote . . . during the pe- riod beginning on the date of the notice and ending on the day after the date of the second general election for Fed- eral office that occurs after the date of the notice” (about four years). 52 U. S. C. §20507(d)(1)(B). Not only are States allowed to remove registrants who satisfy these requirements, but federal law makes this removal mandatory. §20507(d)(3); see also §21083(a)(4)(A). Ohio’s Supplemental Process follows subsection (d) to the letter. It is undisputed that Ohio does not remove a registrant on change-of-residence grounds unless the registrant is sent and fails to mail back a return card and then fails to vote for an additional four years. B Respondents argue (and the Sixth Circuit held) that, even if Ohio’s process complies with subsection (d), it nevertheless violates the Failure-to-Vote Clause—the clause that generally prohibits States from removing people from the rolls “by reason of [a] person’s failure to vote.” §20507(b)(2); see also §21083(a)(4)(A). Respondents point out that Ohio’s Supplemental Process uses a person’s failure to vote twice: once as the trigger for sending return cards and again as one of the requirements for removal. Respondents conclude that this use of nonvoting is illegal. We reject this argument because the Failure-to-Vote Clause, both as originally enacted in the NVRA and as amended by HAVA, simply forbids the use of nonvoting as the sole criterion for removing a registrant, and Ohio does not use it that way. Instead, as permitted by subsection (d), Ohio removes registrants only if they have failed to vote and have failed to respond to a notice. When Congress clarified the meaning of the NVRA’s Failure-to-Vote Clause in HAVA, here is what it said: “[C]onsistent with the [NVRA], . . . no registrant may be removed solely by reason of a failure to vote.” §21083(a)(4)(A) (emphasis added). The meaning of these words is straightforward. “Solely” means “alone.” Webster’s Third New International Dictionary 2168 (2002); American Heritage Dictionary 1654 (4th ed. 2000). And “by reason of” is a “quite formal” way of saying “[b]ecause of.” C. Ammer, American Heritage Dictionary of Idioms 67 (2d ed. 2013). Thus, a State violates the Failure-to-Vote Clause only if it removes registrants for no reason other than their failure to vote. This explanation of the meaning of the Failure-to-Vote Clause merely makes explicit what was implicit in the clause as originally enacted. At that time, the clause simply said that a state program “shall not result in the removal of the name of any person from the [rolls for federal elections] by reason of the person’s failure to vote.” 107Stat. 83. But that prohibition had to be read together with subsection (d), which authorized removal if a registrant did not send back a return card and also failed to vote during a period covering two successive general elections for federal office. If possible, “[w]e must interpret the statute to give effect to both provisions,” Ricci v. DeStefano, 557 U. S. 557, 580 (2009), and here, that is quite easy. The phrase “by reason of” denotes some form of causation. See Gross v. FBL Financial Services, Inc., 557 U. S. 167, 176 (2009). Thus, the Failure-to-Vote Clause applies when nonvoting, in some sense, causes a registrant’s name to be removed, but the law recognizes several types of causation. When a statutory provision includes an undefined causation requirement, we look to context to decide whether the statute demands only but-for cause as opposed to proximate cause or sole cause. See Holmes v. Securities Investor Protection Corporation, 503 U. S. 258, 265–268 (1992). Cf. CSX Transp., Inc. v. McBride, 564 U. S. 685, 692–693 (2011). Which form of causation is required by the Failure-to-Vote Clause? We can readily rule out but-for causation. If “by reason of” in the Failure-to-Vote Clause meant but-for causation, a State would violate the clause if the failure to vote played a necessary part in the removal of a name from the list. Burrage v. United States, 571 U. S. 204, 211 (2014). But the removal process expressly authorized by subsection (d) allows a State to remove a registrant if the registrant, in addition to failing to send back a return card, fails to vote during a period covering two general federal elections. So if the Failure-to-Vote Clause were read in this way, it would cannibalize subsection (d). Interpreting the Failure-to-Vote Clause as incorporating a proximate cause requirement would lead to a similar problem. Proximate cause is an elusive concept, see McBride, supra, at 692–693, but no matter how the term is understood, it is hard to escape the conclusion that the failure to vote is a proximate cause of removal under subsection (d). If a registrant, having failed to send back a return card, also fails to vote during the period covering the next two general federal elections, removal is the direct, foreseeable, and closely connected consequence. See Paroline v. United States, 572 U. S. 434, 444–445 (2014); Bridge v. Phoenix Bond & Indemnity Co., 553 U. S. 639, 654 (2008). By process of elimination, we are left with sole causation. This reading harmonizes the Failure-to-Vote Clause and subsection (d) because the latter provision does not authorize removal solely by reason of a person’s failure to vote. Instead, subsection (d) authorizes removal only if a registrant also fails to mail back a return card. For these reasons, we conclude that the Failure-to-Vote Clause, as originally enacted, referred to sole causation. And when Congress enacted HAVA, it made this point explicit. It added to the Failure-to-Vote Clause itself an explanation of how it is to be read, i.e., in a way that does not contradict subsection (d). And in language that cannot be misunderstood, it reiterated what the clause means: “[R]egistrants who have not responded to a notice and who have not voted in 2 consecutive general elections for Federal office shall be removed from the official list of eligible voters, except that no registrant may be removed solely by reason of a failure to vote.” §21083(a)(4)(A) (emphasis added). In this way, HAVA dispelled any doubt that a state removal program may use the failure to vote as a factor (but not the sole factor) in removing names from the list of registered voters. That is exactly what Ohio’s Supplemental Process does. It does not strike any registrant solely by reason of the failure to vote. Instead, as expressly permitted by federal law, it removes registrants only when they have failed to vote and have failed to respond to a change-of-residence notice. C Respondents and the dissent advance an alternative interpretation of the Failure-to-Vote Clause, but that reading is inconsistent with both the text of the clause and the clarification of its meaning in §21083(a)(4)(A). Respondents argue that the clause allows States to consider nonvoting only to the extent that subsection (d) requires—that is, only after a registrant has failed to mail back a notice. Any other use of the failure to vote, including as the trigger for mailing a notice, they claim, is proscribed. In essence, respondents read the language added to the clause by HAVA—“except that nothing in this paragraph may be construed to prohibit a State from using the procedures described in subsections (c) and (d)”—as an exception to the general rule forbidding the use of nonvoting. See Brief for Respondents 37. And the Sixth Circuit seemed to find this point dispositive, reasoning that “ ‘exceptions in statutes must be strictly construed.’ ” 838 F. 3d, at 708 (quoting Detroit Edison Co. v. SEC, 119 F. 2d 730, 739 (CA6 1941)). We reject this argument for three reasons. First, it distorts what the new language added by HAVA actually says. The new language does not create an exception to a general rule against the use of nonvoting. It does not say that the failure to vote may not be used “except that this paragraph does not prohibit a State from using the procedures described in subsections (c) and (d).” Instead, it says that “nothing in this paragraph may be construed” to have that effect. §20507(b)(2) (emphasis added). Thus, it sets out not an exception, but a rule of interpretation. It does not narrow the language that precedes it; it clarifies what that language means. That is precisely what Congress said when it enacted HAVA: It added the “may not be construed” provision to “[c]larif[y],” not to alter, the prohibition’s scope. §903, 116Stat. 1728. Second, under respondents’ reading, HAVA’s new language is worse than superfluous. Even without the added language, no sensible person would read the Failure-to-Vote Clause as prohibiting what subsections (c) and (d) expressly allow. Yet according to respondents, that is all that the new language accomplishes. So at a minimum, it would be redundant. But the implications of this reading are actually worse than that. There is no reason to create an exception to a prohibition unless the prohibition would otherwise forbid what the exception allows. So if the new language were an exception, it would seem to follow that prior to HAVA, the Failure-to-Vote Clause did outlaw what subsections (c) and (d) specifically authorize. And that, of course, would be nonsensical. Third, respondents’ reading of the language that HAVA added to the Failure-to-Vote Clause makes it hard to understand why Congress prescribed in another section of the same Act, i.e., §21083(a)(4)(A), that “no registrant may be removed solely by reason of a failure to vote.” As interpreted by respondents, the amended Failure-to-Vote Clause prohibits any use of nonvoting with just two narrow exceptions—the uses allowed by subsections (c) and (d). So, according to respondents, the amended Failure-to-Vote Clause prohibits much more than §21083(a)(4)(A). That provision, in addition to allowing the use of nonvoting in accordance with subsections (c) and (d), also permits the use of nonvoting in any other way that does not treat nonvoting as the sole basis for removal. There is no plausible reason why Congress would enact the provision that respondents envision. As interpreted by respondents, HAVA would be like a law that contains one provision making it illegal to drive with a blood alcohol level of 0.08 or higher and another provision making it illegal to drive with a blood alcohol level of 0.10 or higher. The second provision would not only be redundant; it would be confusing and downright silly. Our reading, on the other hand, gives the new language added to the Failure-to-Vote Clause “real and substantial effect.” Husky Int’l Electronics, Inc. v. Ritz, 578 U. S. ___, ___ (2016) (slip op., at 4) (internal quotation marks omitted). It clarifies the meaning of the prohibition against removal by reason of nonvoting, a matter that troubled some States prior to HAVA’s enactment. See, e.g., FEC Report on the NVRA to the 106th Congress 19 (1999). Respondents and the dissent separately claim that the Failure-to-Vote Clause must be read to bar the use of nonvoting as a trigger for sending return cards because otherwise it would be “superfluous.” Post, at 17 (opinion of Breyer, J.); see Brief for Respondents 29. After all, subsection (d) already prohibits States from removing registrants because of a failure to vote alone. See §20507(d)(1). To have meaning independent of subsection (d), respondents reason, the Failure-to-Vote Clause must prohibit other uses of the failure to vote, including its use as a trigger for sending out notices. This argument is flawed because the Failure-to-Vote Clause has plenty of work to do under our reading. Most important, it prohibits the once-common state practice of removing registered voters simply because they failed to vote for some period of time. Not too long ago, “[c]ancellation for failure to vote [was] the principal means used . . . to purge the [voter] lists.” Harris, Model Voter Registration System, at 44. States did not use a person’s failure to vote as evidence that the person had died or moved but as an independent ground for removal. See ibid.[4] Ohio was one such State. Its Constitution provided that “[a]ny elector who fails to vote in at least one election during any period of four consecutive years shall cease to be an elector unless he again registers to vote.” Art. V, §1 (1977). In addition, our reading prohibits States from using the failure to vote as the sole cause for removal on any ground, not just because of a change of residence. Recall that subsection (d)’s removal process applies only to change-of-residence removals but that the Failure-to-Vote Clause applies to all removals. Without the Failure-to-Vote Clause, therefore, States could use the failure to vote as conclusive evidence of ineligibility for some reason other than change of residence, such as death, mental incapac- ity, or a criminal conviction resulting in prolonged imprisonment. D Respondents put forth one additional argument regarding the Failure-to-Vote Clause. In essence, it boils down to this. So many properly registered voters simply discard return cards upon receipt that the failure to send them back is worthless as evidence that the addressee has moved. As respondents’ counsel put it at argument, “a notice that doesn’t get returned” tells the State “absolutely nothing about whether the person has moved.” Tr. of Oral Arg. 41, 58. According to respondents, when Ohio removes registrants for failing to respond to a notice and failing to vote, it functionally “removes people solely for non-voting” unless the State has additional “reliable evidence” that a registrant has moved. Id., at 49, 71. This argument is based on a dubious empirical conclusion that the NVRA and HAVA do not allow us to indulge. Congress clearly did not think that the failure to send back a return card was of no evidentiary value because Congress made that conduct one of the two requirements for removal under subsection (d). Requiring additional evidence not only second-guesses the congressional judgment embodied in subsection (d)’s removal process, but it also second-guesses the judgment of the Ohio Legislature as expressed in the State’s Supplemental Process. The Constitution gives States the authority to set the qualifications for voting in congressional elections, Art. I, §2, cl. 1; Amdt. 17, as well as the authority to set the “Times, Places and Manner” to conduct such elections in the absence of contrary congressional direction, Art. I, §4, cl. 1. We have no authority to dismiss the considered judgment of Congress and the Ohio Legislature regarding the probative value of a registrant’s failure to send back a return card. See Inter Tribal, 570 U. S., at 16–19; see also id., at 36–37 (Thomas, J., dissenting); id., at 42–43, 46 (Alito, J., dissenting). For all these reasons, we hold that Ohio law does not violate the Failure-to-Vote Clause. III We similarly reject respondents’ argument that Ohio violates other provisions of the NVRA and HAVA. A Respondents contend that Ohio removes registered voters on a ground not permitted by the NVRA. They claim that the NVRA permits the removal of a name for only a few specified reasons—a person’s request, criminal conviction, mental incapacity, death, change of residence, and initial ineligibility. Brief for Respondents 25–26; see 52 U. S. C. §§20507(a)(3), (4).[5] And they argue that Ohio removes registrants for other reasons, namely, for failing to respond to a notice and failing to vote. This argument plainly fails. Ohio simply treats the failure to return a notice and the failure to vote as evidence that a registrant has moved, not as a ground for removal. And in doing this, Ohio simply follows federal law. Subsection (d), which governs removals “on the ground that the registrant has changed residence,” treats the failure to return a notice and the failure to vote as evidence that this ground is satisfied. §20507(d)(1). If respondents’ argument were correct, then it would also be illegal to remove a name under §20507(c) because that would constitute removal for submitting change-of-address information to the Postal Service. Likewise, if a State removed a name after receiving a death certificate or a judgment of criminal conviction, that would be illegal because receipt of such documents is not listed as a permitted ground for removal under §20507(a)(3) or §20507(a)(4). About this argument no more need be said. B Respondents maintain, finally, that Ohio’s procedure is illegal because the State sends out notices without having any “reliable indicator” that the addressee has moved. Brief for Respondents 31. The “[f]ailure to vote for a mere two-year period,” they argue, does not reliably “indicate that a registrant has moved out of the jurisdiction.” Id., at 30; see also, e.g., Brief for State of New York et al. as Amici Curiae 13–28. This argument also fails. The degree of correlation between the failure to vote for two years and a change of residence is debatable, but we know from subsection (d) that Congress thought that the failure to vote for a period of two consecutive general elections was a good indicator of change of residence, since it made nonvoting for that period an element of subsection (d)’s requirements for removal. In a similar vein, the Ohio Legislature appar- ently thought that nonvoting for two years was sufficiently correlated with a change of residence to justify sending a return card. What matters for present purposes is not whether the Ohio Legislature overestimated the correlation between nonvoting and moving or whether it reached a wise policy judgment about when return cards should be sent. For us, all that matters is that no provision of the NVRA prohibits the legislature from implementing that judgment. Neither subsection (d) nor any other provision of the NVRA demands that a State have some particular quantum of evidence of a change of residence before sending a registrant a return card. So long as the trigger for sending such notices is “uniform, nondiscriminatory, and in compliance with the Voting Rights Act,” §20507(b)(1), States can use whatever plan they think best. That may be why not even the Sixth Circuit relied on this rationale. Respondents attempt to find support for their argument in subsection (c), which allows States to send notices based on Postal Service change-of-address information. This provision, they argue, implicitly sets a minimum reliability requirement. Thus, they claim, a State may not send out a return card unless its evidence of change of residence is at least as probative as the information obtained from the Postal Service. See Tr. of Oral Arg. 56. Nothing in subsection (c) suggests that it is designed to play this role. Subsection (c) says that “[a] State may meet” its obligation “to remove the names” of ineligible voters on change-of-residence grounds by sending notices to voters who are shown by the Postal Service information to have moved, but subsection (c) does not even hint that it imposes any sort of minimum reliability requirement for sending such notices. §§20507(a)(4), (c). By its terms, subsection (c) simply provides one way—the minimal way—in which a State “may meet the [NVRA’s] requirement[s]” for change-of-residence removals. §20507(c) (emphasis added). As respondents agreed at argument, it is not the only way. Tr. of Oral Arg. 53. C Nothing in the two dissents changes our analysis of the statutory language. 1 Despite its length and complexity, the principal dissent sets out only two arguments. See post, at 7–8 (opinion of Breyer, J.). The first is one that we have already discussed at length, namely, that the Failure-to-Vote Clause prohibits any use of the failure to vote except as permitted by subsections (c) and (d). We have explained why this argument is insupportable, supra, at 12–16, and the dissent has no answer to any of the problems we identify. The dissent’s only other argument is that Ohio’s process violates §20507(a)(4), which requires States to make a “reasonable effort” to remove the names of ineligible voters from the rolls. The dissent thinks that this provision authorizes the federal courts to go beyond the restrictions set out in subsections (b), (c), and (d) and to strike down any state law that does not meet their own standard of “reasonableness.” But see Brief for United States as Amicus Curiae 28–29. The dissent contends that Ohio’s system violates this supposed “reasonableness” requirement primarily because it relies on the failure to mail back the postcard sent to those who have not engaged in voter activity for two years. Based on its own cobbled-together statistics, post, at 12–13, and a feature of human nature of which the dissent has apparently taken judicial notice (i.e., “the human tendency not to send back cards received in the mail,” post, at 13), the dissent argues that the failure to send back the card in question “has no tendency to reveal accurately whether the registered voter has changed residences”; it is an “irrelevant factor” that “shows nothing at all that is statutorily significant.” Post, at 13–14, 17. Whatever the meaning of §20507(a)(4)’s reference to reasonableness, the principal dissent’s argument fails since it is the federal NVRA, not Ohio law, that attaches importance to the failure to send back the card. See §§20507(d)(1)(B)(i), (d)(2)(A). The dissenters may not think that the failure to send back the card means anything, but that was not Congress’s view. The NVRA plainly reflects Congress’s judgment that the failure to send back the card, coupled with the failure to vote during the period covering the next two general federal elections, is significant evidence that the addressee has moved. It is not our prerogative to judge the reasonableness of that congressional judgment, but we note that, whatever the general “human tendency” may be with respect to mailing back cards received in the mail, the notice sent under subsection (d) is nothing like the solicitations for commercial products or contributions that recipients may routinely discard. The notice in question here warns recipients that unless they take the simple and easy step of mailing back the preaddressed, postage prepaid card—or take the equally easy step of updating their information online—their names may be removed from the voting rolls if they do not vote during the next four years. See Record 295–296, 357. It was Congress’s judgment that a reasonable person with an interest in voting is not likely to ignore notice of this sort. 2 Justice Sotomayor’s dissent says nothing about what is relevant in this case—namely, the language of the NVRA—but instead accuses us of “ignor[ing] the history of voter suppression” in this country and of “uphold[ing] a program that appears to further the . . . disenfranchisement of minority and low-income voters.” Post, at 5. Those charges are misconceived. The NVRA prohibits state programs that are discriminatory, see §20507(b)(1), but respondents did not assert a claim under that provision. And Justice Sotomayor has not pointed to any evidence in the record that Ohio instituted or has carried out its program with discriminatory intent. * * * The dissents have a policy disagreement, not just with Ohio, but with Congress. But this case presents a question of statutory interpretation, not a question of policy. We have no authority to second-guess Congress or to decide whether Ohio’s Supplemental Process is the ideal method for keeping its voting rolls up to date. The only question before us is whether it violates federal law. It does not. The judgment of the Sixth Circuit is reversed. It is so ordered. Notes 1 United States Census Bureau, CB16–189, Americans Moving at Historically Low Rates (Nov. 16, 2016), available at https://www.census.gov/newsroom/press-releases/2016/cb16-189.html (all Internet materials as last visited June 8, 2018). States must update the addresses of even those voters who move within their county of residence, for (among other reasons) counties may contain multiple voting districts. Cf. post, at 12 (Breyer, J., dissenting). For example, Cuyahoga County contains 11 State House districts. See House District Map, Ohio House Districts 2012–2022, online at http://www.ohiohouse.gov/members/district-map. 2 The principal dissent attaches a misleading label to this return card, calling it a “ ‘last chance’ notice.” Post, at 6–7, 9–12 (opinion of Breyer, J.). It is actually no such thing. Sending back the notice does not represent a voter’s “last chance” to avoid having his or her name stricken from the rolls. Instead, such a voter has many more chances over a period of four years to avoid that result. All that the voter must do is vote in any election during that time. See 52 U. S. C. §20507(d)(1)(B). 3 U. S. Postal Service, Office of Inspector Gen., MS–MA–15–006, Strategies for Reducing Undeliverable as Addressed Mail 15 (2015); see also Brief for Buckeye Institute as Amicus Curiae 10. Respondents and one of their amici dispute this statistic. See Tr. of Oral Arg. 46; Brief for Asian Americans Advancing Justice et al. as Amici Curiae 27–28. 4 See, e.g., Haw. Rev. Stat. §11–17(a) (1993); Idaho Code Ann. §34–435 (1981); Minn. Stat. §201.171 (1992); Mont. Code Ann. §13–2–401(1) (1993); N. J. Stat. Ann. §19:31–5 (West Supp. 1989); Okla. Stat., Tit. 26, §4–120.2 (1991); Utah Code §20–2–24(1)(b) (1991). 5 We assume for the sake of argument that Congress has the constitutional authority to limit voting eligibility requirements in the way respondents suggest.

585.US.2017_16-1466

Illinois law permits public employees to unionize. If a majority of the employees in a bargaining unit vote to be represented by a union, that union is designated as the exclusive representative of all the employees, even those who do not join. Only the union may engage in collective bargaining; individual employees may not be represented by another agent or negotiate directly with their employer. Nonmembers are required to pay what is generally called an “agency fee,” i.e., a percentage of the full union dues. Under Abood v. Detroit Bd. of Ed., 431 U. S. 209, 235–236, this fee may cover union expenditures attributable to those activities “germane” to the union’s collective-bargaining activities (chargeable expenditures), but may not cover the union’s political and ideological projects (nonchargeable expenditures). The union sets the agency fee annually and then sends nonmembers a notice explaining the basis for the fee and the breakdown of expenditures. Here it was 78.06% of full union dues. Petitioner Mark Janus is a state employee whose unit is represented by a public-sector union (Union), one of the respondents. He refused to join the Union because he opposes many of its positions, including those taken in collective bargaining. Illinois’ Governor, similarly opposed to many of these positions, filed suit challenging the constitutionality of the state law authorizing agency fees. The state attorney general, another respondent, intervened to defend the law, while Janus moved to intervene on the Governor’s side. The District Court dismissed the Governor’s challenge for lack of standing, but it simultaneously allowed Janus to file his own complaint challenging the constitutionality of agency fees. The District Court granted respondents’ motion to dismiss on the ground that the claim was foreclosed by Abood. The Seventh Circuit affirmed. Held: 1. The District Court had jurisdiction over petitioner’s suit. Petitioner was undisputedly injured in fact by Illinois’ agency-fee scheme and his injuries can be redressed by a favorable court decision. For jurisdictional purposes, the court permissibly treated his amended complaint in intervention as the operative complaint in a new lawsuit. United States ex rel. Texas Portland Cement Co. v. McCord, 233 U. S. 157, distinguished. Pp. 6–7. 2. The State’s extraction of agency fees from nonconsenting public-sector employees violates the First Amendment. Abood erred in concluding otherwise, and stare decisis cannot support it. Abood is therefore overruled. Pp. 7–47. (a) Abood’s holding is inconsistent with standard First Amendment principles. Pp. 7–18. (1) Forcing free and independent individuals to endorse ideas they find objectionable raises serious First Amendment concerns. E.g., West Virginia Bd. of Ed. v. Barnette, 319 U. S. 624, 633. That includes compelling a person to subsidize the speech of other private speakers. E.g., Knox v. Service Employees, 567 U. S. 298, 309. In Knox and Harris v. Quinn, 573 U. S. ___, the Court applied an “exacting” scrutiny standard in judging the constitutionality of agency fees rather than the more traditional strict scrutiny. Even under the more permissive standard, Illinois’ scheme cannot survive. Pp. 7–11. (2) Neither of Abood’s two justifications for agency fees passes muster under this standard. First, agency fees cannot be upheld on the ground that they promote an interest in “labor peace.” The Abood Court’s fears of conflict and disruption if employees were represented by more than one union have proved to be unfounded: Exclusive representation of all the employees in a unit and the exaction of agency fees are not inextricably linked. To the contrary, in the Federal Government and the 28 States with laws prohibiting agency fees, millions of public employees are represented by unions that effectively serve as the exclusive representatives of all the employees. Whatever may have been the case 41 years ago when Abood was decided, it is thus now undeniable that “labor peace” can readily be achieved through less restrictive means than the assessment of agency fees. Second, avoiding “the risk of ‘free riders,’ ” Abood, supra, at 224, is not a compelling state interest. Free-rider “arguments . . . are generally insufficient to overcome First Amendment objections,” Knox, supra, at 311, and the statutory requirement that unions represent members and nonmembers alike does not justify different treatment. As is evident in non-agency-fee jurisdictions, unions are quite willing to represent nonmembers in the absence of agency fees. And their duty of fair representation is a necessary concomitant of the authority that a union seeks when it chooses to be the exclusive representative. In any event, States can avoid free riders through less restrictive means than the imposition of agency fees. Pp. 11–18. (b) Respondents’ alternative justifications for Abood are similarly unavailing. Pp. 18–26. (1) The Union claims that Abood is supported by the First Amendment’s original meaning. But neither founding-era evidence nor dictum in Connick v. Myers, 461 U. S. 138, 143, supports the view that the First Amendment was originally understood to allow States to force public employees to subsidize a private third party. If anything, the opposite is true. Pp. 18–22. (2) Nor does Pickering v. Board of Ed. of Township High School Dist. 205, Will Cty., 391 U. S. 563, provide a basis for Abood. Abood was not based on Pickering, and for good reasons. First, Pickering’s framework was developed for use in cases involving “one employee’s speech and its impact on that employee’s public responsibilities,” United States v. Treasury Employees, 513 U. S. 454, 467, while Abood and other agency-fee cases involve a blanket requirement that all employees subsidize private speech with which they may not agree. Second, Pickering’s framework was designed to determine whether a public employee’s speech interferes with the effective operation of a government office, not what happens when the government compels speech or speech subsidies in support of third parties. Third, the categorization schemes of Pickering and Abood do not line up. For example, under Abood, nonmembers cannot be charged for speech that concerns political or ideological issues; but under Pickering, an employee’s free speech interests on such issues could be overcome if outweighed by the employer’s interests. Pp. 22–26. (c) Even under some form of Pickering, Illinois’ agency-fee arrangement would not survive. Pp. 26–33. (1) Respondents compare union speech in collective bargaining and grievance proceedings to speech “pursuant to [an employee’s] official duties,” Garcetti v. Ceballos, 547 U. S. 410, 421, which the State may require of its employees. But in those situations, the employee’s words are really the words of the employer, whereas here the union is speaking on behalf of the employees. Garcetti therefore does not apply. Pp. 26–27. (2) Nor does the union speech at issue cover only matters of private concern, which the State may also generally regulate under Pickering. To the contrary, union speech covers critically important and public matters such as the State’s budget crisis, taxes, and collective bargaining issues related to education, child welfare, healthcare, and minority rights. Pp. 27–31. (3) The government’s proffered interests must therefore justify the heavy burden of agency fees on nonmembers’ First Amendment interests. They do not. The state interests asserted in Abood—promoting “labor peace” and avoiding free riders—clearly do not, as explained earlier. And the new interests asserted in Harris and here—bargaining with an adequately funded agent and improving the efficiency of the work force—do not suffice either. Experience shows that unions can be effective even without agency fees. Pp. 31–33. (d) Stare decisis does not require retention of Abood. An analysis of several important factors that should be taken into account in deciding whether to overrule a past decision supports this conclusion. Pp. 33–47. (1) Abood was poorly reasoned, and those arguing for retaining it have recast its reasoning, which further undermines its stare decisis effect, e.g., Citizens United v. Federal Election Comm’n, 558 U. S. 310, 363. Abood relied on Railway Employes v. Hanson, 351 U. S. 225, and Machinists v. Street, 367 U. S. 740, both of which involved private-sector collective-bargaining agreements where the government merely authorized agency fees. Abood did not appreciate the very different First Amendment question that arises when a State requires its employees to pay agency fees. Abood also judged the constitutionality of public-sector agency fees using Hanson’s deferential standard, which is inappropriate in deciding free speech issues. Nor did Abood take into account the difference between the effects of agency fees in public- and private-sector collective bargaining, anticipate administrative problems with classifying union expenses as chargeable or nonchargeable, foresee practical problems faced by nonmembers wishing to challenge those decisions, or understand the inherently political nature of public-sector bargaining. Pp. 35–38. (2) Abood’s lack of workability also weighs against it. Its line between chargeable and nonchargeable expenditures has proved to be impossible to draw with precision, as even respondents recognize. See, e.g., Lehnert v. Ferris Faculty Assn., 500 U. S. 507, 519. What is more, a nonmember objecting to union chargeability determinations will have much trouble determining the accuracy of the union’s reported expenditures, which are often expressed in extremely broad and vague terms. Pp. 38–41. (3) Developments since Abood, both factual and legal, have “eroded” the decision’s “underpinnings” and left it an outlier among the Court’s First Amendment cases. United States v. Gaudin, 515 U. S. 506, 521. Abood relied on an assumption that “the principle of exclusive representation in the public sector is dependent on a union or agency shop,” Harris, 573 U. S., at ___–___, but experience has shown otherwise. It was also decided when public-sector unionism was a relatively new phenomenon. Today, however, public-sector union membership has surpassed that in the private sector, and that ascendency corresponds with a parallel increase in public spending. Abood is also an anomaly in the Court’s First Amendment jurisprudence, where exacting scrutiny, if not a more demanding standard, generally applies. Overruling Abood will also end the oddity of allowing public employers to compel union support (which is not supported by any tradition) but not to compel party support (which is supported by tradition), see, e.g., Elrod v. Burns, 427 U. S. 347. Pp. 42–44. (4) Reliance on Abood does not carry decisive weight. The uncertain status of Abood, known to unions for years; the lack of clarity it provides; the short-term nature of collective-bargaining agreements; and the ability of unions to protect themselves if an agency-fee provision was crucial to its bargain undermine the force of reliance. Pp. 44–47. 3. For these reasons, States and public-sector unions may no longer extract agency fees from nonconsenting employees. The First Amendment is violated when money is taken from nonconsenting employees for a public-sector union; employees must choose to support the union before anything is taken from them. Accordingly, neither an agency fee nor any other form of payment to a public-sector union may be deducted from an employee, nor may any other attempt be made to collect such a payment, unless the employee affirmatively consents to pay. Pp. 48–49. 851 F. 3d 746, reversed and remanded. Alito, J., delivered the opinion of the Court, in which Roberts, C. J., and Kennedy, Thomas, and, Gorsuch, JJ., joined. Sotomayor, J., filed a dissenting opinion. Kagan, J., filed a dissenting opinion, in which Ginsburg, Breyer, and Sotomayor, JJ., joined.

Under Illinois law, public employees are forced to subsidize a union, even if they choose not to join and strongly object to the positions the union takes in collective bargaining and related activities. We conclude that this arrangement violates the free speech rights of nonmembers by compelling them to subsidize private speech on matters of substantial public concern. We upheld a similar law in Abood v. Detroit Bd. of Ed., 431 U. S. 209 (1977), and we recognize the importance of following precedent unless there are strong reasons for not doing so. But there are very strong reasons in this case. Fundamental free speech rights are at stake. Abood was poorly reasoned. It has led to practical problems and abuse. It is inconsistent with other First Amendment cases and has been undermined by more recent decisions. Developments since Abood was handed down have shed new light on the issue of agency fees, and no reliance interests on the part of public-sector unions are sufficient to justify the perpetuation of the free speech violations that Abood has countenanced for the past 41 years. Abood is therefore overruled. I A Under the Illinois Public Labor Relations Act (IPLRA), employees of the State and its political subdivisions are permitted to unionize. See Ill. Comp. Stat., ch. 5, §315/6(a) (West 2016). If a majority of the employees in a bargaining unit vote to be represented by a union, that union is designated as the exclusive representative of all the employees. §§315/3(s)(1), 315/6(c), 315/9. Employees in the unit are not obligated to join the union selected by their co-workers, but whether they join or not, that union is deemed to be their sole permitted representative. See §§315/6(a), (c). Once a union is so designated, it is vested with broad authority. Only the union may negotiate with the employer on matters relating to “pay, wages, hours[,] and other conditions of employment.” §315/6(c). And this authority extends to the negotiation of what the IPLRA calls “policy matters,” such as merit pay, the size of the work force, layoffs, privatization, promotion methods, and non-discrimination policies. §315/4; see §315/6(c); see gener- ally, e.g., Illinois Dept. of Central Management Servs. v. AFSCME, Council 31, No. S–CB–16–17 etc., 33 PERI ¶67 (ILRB Dec. 13, 2016) (Board Decision). Designating a union as the employees’ exclusive representative substantially restricts the rights of individual employees. Among other things, this designation means that individual employees may not be represented by any agent other than the designated union; nor may individual employees negotiate directly with their employer. §§315/6(c)–(d), 315/10(a)(4); see Matthews v. Chicago Transit Authority, 2016 IL 117638, 51 N. E. 3d 753, 782; accord, Medo Photo Supply Corp. v. NLRB, 321 U. S. 678, 683–684 (1944). Protection of the employees’ interests is placed in the hands of the union, and therefore the union is required by law to provide fair representation for all employees in the unit, members and nonmembers alike. §315/6(d). Employees who decline to join the union are not assessed full union dues but must instead pay what is generally called an “agency fee,” which amounts to a percentage of the union dues. Under Abood, nonmembers may be charged for the portion of union dues attributable to activities that are “germane to [the union’s] duties as collective-bargaining representative,” but nonmembers may not be required to fund the union’s political and ideological projects. 431 U. S., at 235; see id., at 235–236. In labor-law parlance, the outlays in the first category are known as “chargeable” expenditures, while those in the latter are labeled “nonchargeable.” Illinois law does not specify in detail which expenditures are chargeable and which are not. The IPLRA provides that an agency fee may compensate a union for the costs incurred in “the collective bargaining process, contract administration[,] and pursuing matters affecting wages, hours[,] and conditions of employment.” §315/6(e); see also §315/3(g). Excluded from the agency-fee calculation are union expenditures “related to the election or support of any candidate for political office.” §315/3(g); see §315/6(e). Applying this standard, a union categorizes its expenditures as chargeable or nonchargeable and thus determines a nonmember’s “proportionate share,” §315/6(e); this determination is then audited; the amount of the “proportionate share” is certified to the employer; and the employer automatically deducts that amount from the nonmembers’ wages. See ibid.; App. to Pet. for Cert. 37a; see also Harris v. Quinn, 573 U. S. ___, ___–___ (2014) (slip op., at 19–20) (describing this process). Nonmembers need not be asked, and they are not required to consent before the fees are deducted. After the amount of the agency fee is fixed each year, the union must send nonmembers what is known as a Hudson notice. See Teachers v. Hudson, 475 U. S. 292 (1986). This notice is supposed to provide nonmembers with “an adequate explanation of the basis for the [agency] fee.” Id., at 310. If nonmembers “suspect that a union has improperly put certain expenses in the [chargeable] category,” they may challenge that determination. Harris, supra, at ___ (slip op., at 19). As illustrated by the record in this case, unions charge nonmembers, not just for the cost of collective bargaining per se, but also for many other supposedly connected activities. See App. to Pet. for Cert. 28a–39a. Here, the nonmembers were told that they had to pay for “[l]obbying,” “[s]ocial and recreational activities,” “advertising,” “[m]embership meetings and conventions,” and “litigation,” as well as other unspecified “[s]ervices” that “may ultimately inure to the benefit of the members of the local bargaining unit.” Id., at 28a–32a. The total chargeable amount for nonmembers was 78.06% of full union dues. Id., at 34a. B Petitioner Mark Janus is employed by the Illinois Department of Healthcare and Family Services as a child support specialist. Id., at 10a. The employees in his unit are among the 35,000 public employees in Illinois who are represented by respondent American Federation of State, County, and Municipal Employees, Council 31 (Union). Ibid. Janus refused to join the Union because he opposes “many of the public policy positions that [it] advocates,” including the positions it takes in collective bargaining. Id., at 10a, 18a. Janus believes that the Union’s “behavior in bargaining does not appreciate the current fiscal crises in Illinois and does not reflect his best interests or the interests of Illinois citizens.” Id., at 18a. Therefore, if he had the choice, he “would not pay any fees or otherwise subsidize [the Union].” Ibid. Under his unit’s collective-bargaining agreement, however, he was required to pay an agency fee of $44.58 per month, id., at 14a—which would amount to about $535 per year. Janus’s concern about Illinois’ current financial situation is shared by the Governor of the State, and it was the Governor who initially challenged the statute authorizing the imposition of agency fees. The Governor commenced an action in federal court, asking that the law be declared unconstitutional, and the Illinois attorney general (a respondent here) intervened to defend the law. App. 41. Janus and two other state employees also moved to intervene—but on the Governor’s side. Id., at 60. Respondents moved to dismiss the Governor’s challenge for lack of standing, contending that the agency fees did not cause him any personal injury. E.g., id., at 48–49. The District Court agreed that the Governor could not maintain the lawsuit, but it held that petitioner and the other individuals who had moved to intervene had standing because the agency fees unquestionably injured them. Accordingly, “in the interest of judicial economy,” the court dismissed the Governor as a plaintiff, while simultane- ously allowing petitioner and the other employees to file their own complaint. Id., at 112. They did so, and the case proceeded on the basis of this new complaint. The amended complaint claims that all “nonmember fee deductions are coerced political speech” and that “the First Amendment forbids coercing any money from the nonmembers.” App. to Pet. for Cert. 23a. Respondents moved to dismiss the amended complaint, correctly recognizing that the claim it asserted was foreclosed by Abood. The District Court granted the motion, id., at 7a, and the Court of Appeals for the Seventh Circuit affirmed, 851 F. 3d 746 (2017). Janus then sought review in this Court, asking us to overrule Abood and hold that public-sector agency-fee arrangements are unconstitutional. We granted certiorari to consider this important question. 582 U. S. ___ (2017). II Before reaching this question, however, we must con- sider a threshold issue. Respondents contend that the Dis- trict Court lacked jurisdiction under Article III of the Constitution because petitioner “moved to intervene in [the Governor’s] jurisdictionally defective lawsuit.” Union Brief in Opposition 11; see also id., at 13–17; State Brief in Opposition 6; Brief for Union Respondent i, 16–17; Brief for State Respondents 14, n. 1. This argument is clearly wrong. It rests on the faulty premise that petitioner intervened in the action brought by the Governor, but that is not what happened. The District Court did not grant petitioner’s motion to intervene in that lawsuit. Instead, the court essentially treated petitioner’s amended complaint as the operative complaint in a new lawsuit. App. 110–112. And when the case is viewed in that way, any Article III issue vanishes. As the District Court recognized—and as respondents concede—petitioner was injured in fact by Illinois’ agency-fee scheme, and his injuries can be redressed by a favorable court decision. Ibid.; see Record 2312–2313, 2322–2323. Therefore, he clearly has Article III standing. Lujan v. Defenders of Wildlife, 504 U. S. 555, 560–561 (1992). It is true that the District Court docketed petitioner’s complaint under the number originally assigned to the Governor’s complaint, instead of giving it a new number of its own. But Article III jurisdiction does not turn on such trivialities. The sole decision on which respondents rely, United States ex rel. Texas Portland Cement Co. v. McCord, 233 U. S. 157 (1914), actually works against them. That case concerned a statute permitting creditors of a government contractor to bring suit on a bond between 6 and 12 months after the completion of the work. Id., at 162. One creditor filed suit before the 6-month starting date, but another intervened within the 6-to-12-month window. The Court held that the “[t]he intervention [did] not cure th[e] vice in the original [prematurely filed] suit,” but the Court also contemplated treating “intervention . . . as an original suit” in a case in which the intervenor met the requirements that a plaintiff must satisfy—e.g., filing a separate complaint and properly serving the defendants. Id., at 163–164. Because that is what petitioner did here, we may reach the merits of the question presented. III In Abood, the Court upheld the constitutionality of an agency-shop arrangement like the one now before us, 431 U. S., at 232, but in more recent cases we have recognized that this holding is “something of an anomaly,” Knox v. Service Employees, 567 U. S. 298, 311 (2012), and that Abood’s “analysis is questionable on several grounds,” Harris, 573 U. S., at ___ (slip op., at 17); see id., at ___–___ (slip op., at 17–20) (discussing flaws in Abood’s reasoning). We have therefore refused to extend Abood to situations where it does not squarely control, see Harris, supra, at ___–___ (slip op., at 27–29), while leaving for another day the question whether Abood should be overruled, Harris, supra, at ___, n. 19 (slip op., at 27, n. 19); see Knox, supra, at 310–311. We now address that question. We first consider whether Abood’s holding is consistent with standard First Amendment principles. A The First Amendment, made applicable to the States by the Fourteenth Amendment, forbids abridgment of the freedom of speech. We have held time and again that freedom of speech “includes both the right to speak freely and the right to refrain from speaking at all.” Wooley v. Maynard, 430 U. S. 705, 714 (1977); see Riley v. National Federation of Blind of N. C., Inc., 487 U. S. 781, 796–797 (1988); Harper & Row, Publishers, Inc. v. Nation Enterprises, 471 U. S. 539, 559 (1985); Miami Herald Publishing Co. v. Tornillo, 418 U. S. 241, 256–257 (1974); accord, Pacific Gas & Elec. Co. v. Public Util. Comm’n of Cal., 475 U. S. 1, 9 (1986) (plurality opinion). The right to eschew association for expressive purposes is likewise protected. Roberts v. United States Jaycees, 468 U. S. 609, 623 (1984) (“Freedom of association . . . plainly presupposes a freedom not to associate”); see Pacific Gas & Elec., supra, at 12 (“[F]orced associations that burden protected speech are impermissible”). As Justice Jackson memorably put it: “If there is any fixed star in our constitutional constellation, it is that no official, high or petty, can prescribe what shall be orthodox in politics, nationalism, religion, or other matters of opinion or force citizens to confess by word or act their faith therein.” West Virginia Bd. of Ed. v. Barnette, 319 U. S. 624, 642 (1943) (emphasis added). Compelling individuals to mouth support for views they find objectionable violates that cardinal constitutional command, and in most contexts, any such effort would be universally condemned. Suppose, for example, that the State of Illinois required all residents to sign a document expressing support for a particular set of positions on controversial public issues—say, the platform of one of the major political parties. No one, we trust, would seriously argue that the First Amendment permits this. Perhaps because such compulsion so plainly violates the Constitution, most of our free speech cases have involved restrictions on what can be said, rather than laws compelling speech. But measures compelling speech are at least as threatening. Free speech serves many ends. It is essential to our democratic form of government, see, e.g., Garrison v. Louisiana, 379 U. S. 64, 74–75 (1964), and it furthers the search for truth, see, e.g., Thornhill v. Alabama, 310 U. S. 88, 95 (1940). Whenever the Federal Government or a State prevents individuals from saying what they think on important matters or compels them to voice ideas with which they disagree, it undermines these ends. When speech is compelled, however, additional damage is done. In that situation, individuals are coerced into betraying their convictions. Forcing free and independent individuals to endorse ideas they find objectionable is always demeaning, and for this reason, one of our landmark free speech cases said that a law commanding “involuntary affirmation” of objected-to beliefs would require “even more immediate and urgent grounds” than a law demanding silence. Barnette, supra, at 633; see also Riley, supra, at 796–797 (rejecting “deferential test” for compelled speech claims). Compelling a person to subsidize the speech of other private speakers raises similar First Amendment concerns. Knox, supra, at 309; United States v. United Foods, Inc., 533 U. S. 405, 410 (2001); Abood, supra, at 222, 234–235. As Jefferson famously put it, “to compel a man to furnish contributions of money for the propagation of opinions which he disbelieves and abhor[s] is sinful and tyrannical.” A Bill for Establishing Religious Freedom, in 2 Papers of Thomas Jefferson 545 (J. Boyd ed. 1950) (emphasis deleted and footnote omitted); see also Hudson, 475 U. S., at 305, n. 15. We have therefore recognized that a “ ‘significant impingement on First Amendment rights’ ” occurs when public employees are required to provide financial support for a union that “takes many positions during collective bargaining that have powerful political and civic consequences.” Knox, supra, at 310–311 (quoting Ellis v. Railway Clerks, 466 U. S. 435, 455 (1984)). Because the compelled subsidization of private speech seriously impinges on First Amendment rights, it cannot be casually allowed. Our free speech cases have identified “levels of scrutiny” to be applied in different contexts, and in three recent cases, we have considered the standard that should be used in judging the constitutionality of agency fees. See Knox, supra; Harris, supra; Friedrichs v. California Teachers Assn., 578 U. S. ___ (2016) (per cu- riam) (affirming decision below by equally divided Court). In Knox, the first of these cases, we found it sufficient to hold that the conduct in question was unconstitutional under even the test used for the compulsory subsidization of commercial speech. 567 U. S., at 309–310, 321–322. Even though commercial speech has been thought to enjoy a lesser degree of protection, see, e.g., Central Hudson Gas & Elec. Corp. v. Public Serv. Comm’n of N. Y., 447 U. S. 557, 562–563 (1980), prior precedent in that area, specifically United Foods, supra, had applied what we characterized as “exacting” scrutiny, Knox, 567 U. S., at 310, a less demanding test than the “strict” scrutiny that might be thought to apply outside the commercial sphere. Under “exacting” scrutiny, we noted, a compelled subsidy must “serve a compelling state interest that cannot be achieved through means significantly less restrictive of associa- tional freedoms.” Ibid. (internal quotation marks and altera- tions omitted). In Harris, the second of these cases, we again found that an agency-fee requirement failed “exacting scrutiny.” 573 U. S., at ___ (slip op., at 33). But we questioned whether that test provides sufficient protection for free speech rights, since “it is apparent that the speech compelled” in agency-fee cases “is not commercial speech.” Id., at ___ (slip op., at 30). Picking up that cue, petitioner in the present case contends that the Illinois law at issue should be subjected to “strict scrutiny.” Brief for Petitioner 36. The dissent, on the other hand, proposes that we apply what amounts to rational-basis review, that is, that we ask only whether a government employer could reasonably believe that the exaction of agency fees serves its interests. See post, at 4 (Kagan, J., dissenting) (“A government entity could reasonably conclude that such a clause was needed”). This form of minimal scrutiny is foreign to our free-speech jurisprudence, and we reject it here. At the same time, we again find it unnecessary to decide the issue of strict scrutiny because the Illinois scheme cannot survive under even the more permissive standard applied in Knox and Harris. In the remainder of this part of our opinion (Parts III–B and III–C), we will apply this standard to the justifications for agency fees adopted by the Court in Abood. Then, in Parts IV and V, we will turn to alternative rationales proffered by respondents and their amici. B In Abood, the main defense of the agency-fee arrangement was that it served the State’s interest in “labor peace,” 431 U. S., at 224. By “labor peace,” the Abood Court meant avoidance of the conflict and disruption that it envisioned would occur if the employees in a unit were represented by more than one union. In such a situation, the Court predicted, “inter-union rivalries” would foster “dissension within the work force,” and the employer could face “conflicting demands from different unions.” Id., at 220–221. Confusion would ensue if the employer entered into and attempted to “enforce two or more agreements specifying different terms and conditions of employment.” Id., at 220. And a settlement with one union would be “subject to attack from [a] rival labor organizatio[n].” Id., at 221. We assume that “labor peace,” in this sense of the term, is a compelling state interest, but Abood cited no evidence that the pandemonium it imagined would result if agency fees were not allowed, and it is now clear that Abood’s fears were unfounded. The Abood Court assumed that designation of a union as the exclusive representative of all the employees in a unit and the exaction of agency fees are inextricably linked, but that is simply not true. Harris, supra, at ___ (slip op., at 31). The federal employment experience is illustrative. Under federal law, a union chosen by majority vote is designated as the exclusive representative of all the employees, but federal law does not permit agency fees. See 5 U. S. C. §§7102, 7111(a), 7114(a). Nevertheless, nearly a million federal employees—about 27% of the federal work force—are union members.[1] The situation in the Postal Service is similar. Although permitted to choose an exclusive representative, Postal Service employees are not required to pay an agency fee, 39 U. S. C. §§1203(a), 1209(c), and about 400,000 are union members.[2] Likewise, millions of public employees in the 28 States that have laws generally prohibiting agency fees are represented by unions that serve as the exclusive representatives of all the employees.[3] Whatever may have been the case 41 years ago when Abood was handed down, it is now undeniable that “labor peace” can readily be achieved “through means significantly less restrictive of associational freedoms” than the assessment of agency fees. Harris, supra, at ___ (slip op., at 30) (internal quotation marks omitted). C In addition to the promotion of “labor peace,” Abood cited “the risk of ‘free riders’ ” as justification for agency fees, 431 U. S., at 224. Respondents and some of their amici endorse this reasoning, contending that agency fees are needed to prevent nonmembers from enjoying the benefits of union representation without shouldering the costs. Brief for Union Respondent 34–36; Brief for State Respondents 41–45; see, e.g., Brief for International Brotherhood of Teamsters as Amicus Curiae 3–5. Petitioner strenuously objects to this free-rider label. He argues that he is not a free rider on a bus headed for a destination that he wishes to reach but is more like a person shanghaied for an unwanted voyage. Whichever description fits the majority of public employees who would not subsidize a union if given the option, avoiding free riders is not a compelling interest. As we have noted, “free-rider arguments . . . are generally insufficient to overcome First Amendment objections.” Knox, 567 U. S., at 311. To hold otherwise across the board would have startling consequences. Many private groups speak out with the objective of obtaining government action that will have the effect of benefiting nonmembers. May all those who are thought to benefit from such efforts be compelled to subsidize this speech? Suppose that a particular group lobbies or speaks out on behalf of what it thinks are the needs of senior citizens or veterans or physicians, to take just a few examples. Could the government require that all seniors, veterans, or doctors pay for that service even if they object? It has never been thought that this is permissible. “[P]rivate speech often furthers the interests of nonspeakers,” but “that does not alone empower the state to compel the speech to be paid for.” Lehnert v. Ferris Faculty Assn., 500 U. S. 507, 556 (1991) (Scalia, J., concurring in judgment in part and dissenting in part). In simple terms, the First Amendment does not permit the government to compel a person to pay for another party’s speech just because the government thinks that the speech furthers the interests of the person who does not want to pay.[4] Those supporting agency fees contend that the situation here is different because unions are statutorily required to “represen[t] the interests of all public employees in the unit,” whether or not they are union members. §315/6(d); see, e.g., Brief for State Respondents 40–41, 45; post, at 7 (Kagan, J., dissenting). Why might this matter? We can think of two possible arguments. It might be argued that a State has a compelling interest in requiring the payment of agency fees because (1) unions would otherwise be unwilling to represent nonmembers or (2) it would be fundamentally unfair to require unions to provide fair representation for nonmembers if nonmembers were not required to pay. Neither of these arguments is sound. First, it is simply not true that unions will refuse to serve as the exclusive representative of all employees in the unit if they are not given agency fees. As noted, unions represent millions of public employees in jurisdictions that do not permit agency fees. No union is ever compelled to seek that designation. On the contrary, designation as exclusive representative is avidly sought.[5] Why is this so? Even without agency fees, designation as the exclusive representative confers many benefits. As noted, that status gives the union a privileged place in negotiations over wages, benefits, and working conditions. See §315/6(c). Not only is the union given the exclusive right to speak for all the employees in collective bargaining, but the employer is required by state law to listen to and to bargain in good faith with only that union. §315/7. Designation as exclusive representative thus “results in a tremendous increase in the power” of the union. American Communications Assn. v. Douds, 339 U. S. 382, 401 (1950). In addition, a union designated as exclusive representative is often granted special privileges, such as obtaining information about employees, see §315/6(c), and having dues and fees deducted directly from employee wages, §§315/6(e)–(f ). The collective-bargaining agreement in this case guarantees a long list of additional privileges. See App. 138–143. These benefits greatly outweigh any extra burden imposed by the duty of providing fair representation for nonmembers. What this duty entails, in simple terms, is an obligation not to “act solely in the interests of [the union’s] own members.” Brief for State Respondents 41; see Cintron v. AFSCME, Council 31, No. S–CB–16–032, p. 1, 34 PERI ¶105 (ILRB Dec. 13, 2017) (union may not intentionally direct “animosity” toward nonmembers based on their “dissident union practices”); accord, 14 Penn Plaza LLC v. Pyett, 556 U. S. 247, 271 (2009); Vaca v. Sipes, 386 U. S. 171, 177 (1967). What does this mean when it comes to the negotiation of a contract? The union may not negotiate a collective-bargaining agreement that discriminates against nonmembers, see Steele v. Louisville & Nashville R. Co., 323 U. S. 192, 202–203 (1944), but the union’s bargaining latitude would be little different if state law simply prohibited public employers from entering into agreements that discriminate in that way. And for that matter, it is questionable whether the Constitution would permit a public-sector employer to adopt a collective-bargaining agreement that discriminates against nonmembers. See id., at 198–199, 202 (analogizing a private-sector union’s fair-representation duty to the duty “the Constitution imposes upon a legislature to give equal protection to the interests of those for whom it legislates”); cf. Rumsfeld v. Forum for Academic and Institutional Rights, Inc., 547 U. S. 47, 69 (2006) (recognizing that government may not “impose penalties or withhold benefits based on membership in a disfavored group” where doing so “ma[kes] group membership less attractive”). To the extent that an employer would be barred from acceding to a discriminatory agreement anyway, the union’s duty not to ask for one is superfluous. It is noteworthy that neither respondents nor any of the 39 amicus briefs supporting them—nor the dissent—has explained why the duty of fair representation causes public-sector unions to incur significantly greater expenses than they would otherwise bear in negotiating collective-bargaining agreements. What about the representation of nonmembers in grievance proceedings? Unions do not undertake this activity solely for the benefit of nonmembers—which is why Illinois law gives a public-sector union the right to send a representative to such proceedings even if the employee declines union representation. §315/6(b). Representation of nonmembers furthers the union’s interest in keeping control of the administration of the collective-bargaining agreement, since the resolution of one employee’s grievance can affect others. And when a union controls the grievance process, it may, as a practical matter, effectively subordinate “the interests of [an] individual em- ployee . . . to the collective interests of all employees in the bargaining unit.” Alexander v. Gardner-Denver Co., 415 U.S. 36, 58, n. 19 (1974); see Stahulak v. Chicago, 184 Ill. 2d 176, 180–181, 703 N. E. 2d 44, 46–47 (1998); Mahoney v. Chicago, 293 Ill. App. 3d 69, 73–74, 687 N.E.2d 132, 135–137 (1997) (union has “ ‘discretion to refuse to process’ ” a grievance, provided it does not act “arbitrar[ily]” or “in bad faith” (emphasis deleted)). In any event, whatever unwanted burden is imposed by the representation of nonmembers in disciplinary matters can be eliminated “through means significantly less restrictive of associational freedoms” than the imposition of agency fees. Harris, 573 U. S., at ___ (slip op., at 30) (internal quotation marks omitted). Individual nonmembers could be required to pay for that service or could be denied union representation altogether.[6] Thus, agency fees cannot be sustained on the ground that unions would otherwise be unwilling to represent nonmembers. Nor can such fees be justified on the ground that it would otherwise be unfair to require a union to bear the duty of fair representation. That duty is a necessary concomitant of the authority that a union seeks when it chooses to serve as the exclusive representative of all the employees in a unit. As explained, designating a union as the exclusive representative of nonmembers substantially restricts the nonmembers’ rights. Supra, at 2–3. Protection of their interests is placed in the hands of the union, and if the union were free to disregard or even work against those interests, these employees would be wholly unprotected. That is why we said many years ago that serious “constitutional questions [would] arise” if the union were not subject to the duty to represent all employees fairly. Steele, supra, at 198. In sum, we do not see any reason to treat the free-rider interest any differently in the agency-fee context than in any other First Amendment context. See Knox, 567 U. S., at 311, 321. We therefore hold that agency fees cannot be upheld on free-rider grounds. IV Implicitly acknowledging the weakness of Abood’s own reasoning, proponents of agency fees have come forward with alternative justifications for the decision, and we now address these arguments. A The most surprising of these new arguments is the Union respondent’s originalist defense of Abood. According to this argument, Abood was correctly decided because the First Amendment was not originally understood to provide any protection for the free speech rights of public employees. Brief for Union Respondent 2–3, 17–20. As an initial matter, we doubt that the Union—or its members—actually want us to hold that public employees have “no [free speech] rights.” Id., at 1. Cf., e.g., Brief for National Treasury Employees Union as Amicus Curiae in Garcetti v. Ceballos, O. T. 2005, No. 04–473, p. 7 (arguing for “broa[d]” public-employee First Amendment rights); Brief for AFL–CIO as Amicus Curiae in No. 04–473 (similar). It is particularly discordant to find this argument in a brief that trumpets the importance of stare decisis. See Brief for Union Respondent 47–57. Taking away free speech protection for public employees would mean overturning decades of landmark precedent. Under the Union’s theory, Pickering v. Board of Ed. of Township High School Dist. 205, Will Cty., 391 U. S. 563 (1968), and its progeny would fall. Yet Pickering, as we will discuss, is now the foundation for respondents’ chief defense of Abood. And indeed, Abood itself would have to go if public employees have no free speech rights, since Abood holds that the First Amendment prohibits the exaction of agency fees for political or ideological purposes. 431 U. S., at 234–235 (finding it “clear” that “a government may not require an individual to relinquish rights guaranteed him by the First Amendment as a condition of public employment”). Our political patronage cases would be doomed. See, e.g., Rutan v. Republican Party of Ill., 497 U. S. 62 (1990); Branti v. Finkel, 445 U. S. 507 (1980); Elrod v. Burns, 427 U. S. 347 (1976). Also imperiled would be older precedents like Wieman v. Updegraff, 344 U. S. 183 (1952) (loyalty oaths), Shelton v. Tucker, 364 U. S. 479 (1960) (disclosure of memberships and contributions), and Keyishian v. Board of Regents of Univ. of State of N. Y., 385 U. S. 589 (1967) (subversive speech). Respondents presumably want none of this, desiring instead that we apply the Constitution’s supposed original meaning only when it suits them—to retain the part of Abood that they like. See Tr. of Oral Arg. 56–57. We will not engage in this halfway originalism. Nor, in any event, does the First Amendment’s original meaning support the Union’s claim. The Union offers no persuasive founding-era evidence that public employees were understood to lack free speech protections. While it observes that restrictions on federal employees’ activities have existed since the First Congress, most of its historical examples involved limitations on public officials’ outside business dealings, not on their speech. See Ex parte Curtis, 106 U. S. 371, 372–373 (1882). The only early speech restrictions the Union identifies are an 1806 statute prohibiting military personnel from using “ ‘contemptuous or disrespectful words against the President’ ” and other officials, and an 1801 directive limiting electioneering by top government employees. Brief for Union Respondent 3. But those examples at most show that the government was understood to have power to limit employee speech that threatened important governmental interests (such as maintaining military discipline and preventing corruption)—not that public employees’ speech was entirely unprotected. Indeed, more recently this Court has upheld similar restrictions even while recognizing that government employees possess First Amendment rights. See, e.g., Brown v. Glines, 444 U. S. 348, 353 (1980) (upholding military restriction on speech that threatened troop readiness); Civil Service Comm’n v. Letter Carriers, 413 U. S. 548, 556–557 (1973) (upholding limits on public employees’ political activities). Ultimately, the Union relies, not on founding-era evidence, but on dictum from a 1983 opinion of this Court stating that, “[f]or most of th[e 20th] century, the unchallenged dogma was that a public employee had no right to object to conditions placed upon the terms of employment—including those which restricted the exercise of constitutional rights.” Connick v. Myers, 461 U. S. 138, 143; see Brief for Union Respondent 2, 17. Even on its own terms, this dictum about 20th-century views does not purport to describe how the First Amendment was understood in 1791. And a careful examination of the decisions by this Court that Connick cited to support its dictum, see 461 U. S., at 144, reveals that none of them rested on the facile premise that public employees are unprotected by the First Amendment. Instead, they considered (much as we do today) whether particular speech restrictions were “necessary to protect” fundamental government interests. Curtis, supra, at 374. The Union has also failed to show that, even if public employees enjoyed free speech rights, the First Amendment was nonetheless originally understood to allow forced subsidies like those at issue here. We can safely say that, at the time of the adoption of the First Amendment, no one gave any thought to whether public-sector unions could charge nonmembers agency fees. Entities resembling labor unions did not exist at the founding, and public-sector unions did not emerge until the mid-20th century. The idea of public-sector unionization and agency fees would astound those who framed and ratified the Bill of Rights.[7] Thus, the Union cannot point to any accepted founding-era practice that even remotely resembles the compulsory assessment of agency fees from public-sector employees. We do know, however, that prominent members of the founding generation condemned laws requiring public employees to affirm or support beliefs with which they disagreed. As noted, Jefferson denounced compelled support for such beliefs as “ ‘sinful and tyrannical,’ ” supra, at 9, and others expressed similar views.[8] In short, the Union has offered no basis for concluding that Abood is supported by the original understanding of the First Amendment. B The principal defense of Abood advanced by respondents and the dissent is based on our decision in Pickering, 391 U. S. 563, which held that a school district violated the First Amendment by firing a teacher for writing a letter critical of the school administration. Under Pickering and later cases in the same line, employee speech is largely unprotected if it is part of what the employee is paid to do, see Garcetti v. Ceballos, 547 U. S. 410, 421–422 (2006), or if it involved a matter of only private concern, see Connick, supra, at 146–149. On the other hand, when a public employee speaks as a citizen on a matter of public concern, the employee’s speech is protected unless “ ‘the interest of the state, as an employer, in promoting the efficiency of the public services it performs through its employees’ outweighs ‘the interests of the [employee], as a citizen, in commenting upon matters of public concern.’ ” Harris, 573 U. S., at ___ (slip op., at 35) (quoting Pickering, supra, at 568). Pickering was the centerpiece of the defense of Abood in Harris, see 573 U. S., at ___–___ (slip op., at 17–21) (Kagan, J., dissenting), and we found the argument unpersuasive, see id., at ___–___ (slip op., at 34–37). The intervening years have not improved its appeal. 1 As we pointed out in Harris, Abood was not based on Pickering. 573 U. S., at ___, and n. 26 (slip op., at 34, and n. 26). The Abood majority cited the case exactly once—in a footnote—and then merely to acknowledge that “there may be limits on the extent to which an employee in a sensitive or policymaking position may freely criticize his superiors and the policies they espouse.” 431 U. S., at 230, n. 27. That aside has no bearing on the agency-fee issue here.[9] Respondents’ reliance on Pickering is thus “an effort to find a new justification for the decision in Abood.” Harris, supra, at ___ (slip op., at 34). And we have previously taken a dim view of similar attempts to recast problematic First Amendment decisions. See, e.g., Citizens United v. Federal Election Comm’n, 558 U. S. 310, 348–349, 363 (2010) (rejecting efforts to recast Austin v. Michigan Chamber of Commerce, 494 U. S. 652 (1990)); see also Citizens United, supra, at 382–385 (Roberts, C. J., concurring). We see no good reason, at this late date, to try to shoehorn Abood into the Pickering framework. 2 Even if that were attempted, the shoe would be a painful fit for at least three reasons. First, the Pickering framework was developed for use in a very different context—in cases that involve “one employee’s speech and its impact on that employee’s public responsibilities.” United States v. Treasury Employees, 513 U. S. 454, 467 (1995). This case, by contrast, involves a blanket requirement that all employees subsidize speech with which they may not agree. While we have sometimes looked to Pickering in considering general rules that affect broad categories of employees, we have acknowledged that the standard Pickering analysis requires modification in that situation. See 513 U. S., at 466–468, and n. 11. A speech-restrictive law with “widespread impact,” we have said, “gives rise to far more serious concerns than could any single supervisory decision.” Id., at 468. Therefore, when such a law is at issue, the government must shoulder a correspondingly “heav[ier]” burden, id., at 466, and is entitled to considerably less deference in its assessment that a predicted harm justifies a particular impingement on First Amendment rights, see id., at 475–476, n. 21; accord, id., at 482–483 (O’Connor, J., concurring in judgment in part and dissenting in part). The end product of those adjustments is a test that more closely resembles exacting scrutiny than the traditional Pickering analysis. The core collective-bargaining issue of wages and benefits illustrates this point. Suppose that a single employee complains that he or she should have received a 5% raise. This individual complaint would likely constitute a matter of only private concern and would therefore be unprotected under Pickering. But a public-sector union’s demand for a 5% raise for the many thousands of employees it represents would be another matter entirely. Granting such a raise could have a serious impact on the budget of the government unit in question, and by the same token, denying a raise might have a significant effect on the performance of government services. When a large number of employees speak through their union, the category of speech that is of public concern is greatly enlarged, and the category of speech that is of only private concern is substantially shrunk. By disputing this, post, at 13–14, the dissent denies the obvious. Second, the Pickering framework fits much less well where the government compels speech or speech subsidies in support of third parties. Pickering is based on the insight that the speech of a public-sector employee may interfere with the effective operation of a government office. When a public employer does not simply restrict potentially disruptive speech but commands that its employees mouth a message on its own behalf, the calculus is very different. Of course, if the speech in question is part of an employee’s official duties, the employer may insist that the employee deliver any lawful message. See Garcetti, 547 U. S., at 421–422, 425–426. Otherwise, however, it is not easy to imagine a situation in which a public employer has a legitimate need to demand that its employees recite words with which they disagree. And we have never applied Pickering in such a case. Consider our decision in Connick. In that case, we held that an assistant district attorney’s complaints about the supervisors in her office were, for the most part, matters of only private concern. 461 U. S., at 148. As a result, we held, the district attorney could fire her for making those comments. Id., at 154. Now, suppose that the assistant had not made any critical comments about the supervisors but that the district attorney, out of the blue, demanded that she circulate a memo praising the supervisors. Would her refusal to go along still be a matter of purely private concern? And if not, would the order be justified on the ground that the effective operation of the office demanded that the assistant voice complimentary sentiments with which she disagreed? If Pickering applies at all to compelled speech—a question that we do not decide—it would certainly require adjustment in that context. Third, although both Pickering and Abood divided speech into two categories, the cases’ categorization schemes do not line up. Superimposing the Pickering scheme on Abood would significantly change the Abood regime. Let us first look at speech that is not germane to collective bargaining but instead concerns political or ideological issues. Under Abood, a public employer is flatly prohibited from permitting nonmembers to be charged for this speech, but under Pickering, the employees’ free speech interests could be overcome if a court found that the employer’s interests outweighed the employees’. A similar problem arises with respect to speech that is germane to collective bargaining. The parties dispute how much of this speech is of public concern, but respondents concede that much of it falls squarely into that category. See Tr. of Oral Arg. 47, 65. Under Abood, nonmembers may be required to pay for all this speech, but Pickering would permit that practice only if the employer’s interests outweighed those of the employees. Thus, recasting Abood as an application of Pickering would substantially alter the Abood scheme. For all these reasons, Pickering is a poor fit indeed. V Even if we were to apply some form of Pickering, Illinois’ agency-fee arrangement would not survive. A Respondents begin by suggesting that union speech in collective-bargaining and grievance proceedings should be treated like the employee speech in Garcetti, i.e., as speech “pursuant to [an employee’s] official duties,” 547 U. S., at 421. Many employees, in both the public and private sectors, are paid to write or speak for the purpose of furthering the interests of their employers. There are laws that protect public employees from being compelled to say things that they reasonably believe to be untrue or improper, see id., at 425–426, but in general when public employees are performing their job duties, their speech may be controlled by their employer. Trying to fit union speech into this framework, respondents now suggest that the union speech funded by agency fees forms part of the official duties of the union officers who engage in the speech. Brief for Union Respondent 22–23; see Brief for State Respondents 23–24. This argument distorts collective bargaining and grievance adjustment beyond recognition. When an employee engages in speech that is part of the employee’s job duties, the employee’s words are really the words of the employer. The employee is effectively the employer’s spokesperson. But when a union negotiates with the employer or represents employees in disciplinary proceedings, the union speaks for the employees, not the employer. Otherwise, the employer would be negotiating with itself and disputing its own actions. That is not what anybody understands to be happening. What is more, if the union’s speech is really the employer’s speech, then the employer could dictate what the union says. Unions, we trust, would be appalled by such a suggestion. For these reasons, Garcetti is totally inapposite here. B Since the union speech paid for by agency fees is not controlled by Garcetti, we move on to the next step of the Pickering framework and ask whether the speech is on a matter of public or only private concern. In Harris, the dissent’s central argument in defense of Abood was that union speech in collective bargaining, including speech about wages and benefits, is basically a matter of only private interest. See 573 U. S., at ___–___ (slip op., at 19–20) (Kagan, J., dissenting). We squarely rejected that argument, see id., at ___–___ (slip op., at 35–36), and the facts of the present case substantiate what we said at that time: “[I]t is impossible to argue that the level of . . . state spending for employee benefits . . . is not a matter of great public concern,” id., at ___ (slip op., at 36). Illinois, like some other States and a number of counties and cities around the country, suffers from severe budget problems.[10] As of 2013, Illinois had nearly $160 billion in unfunded pension and retiree healthcare liabilities.[11] By 2017, that number had only grown, and the State was grappling with $15 billion in unpaid bills.[12] We are told that a “quarter of the budget is now devoted to paying down” those liabilities.[13] These problems and others led Moody’s and S&P to downgrade Illinois’ credit rating to “one step above junk”—the “lowest ranking on record for a U. S. state.”[14] The Governor, on one side, and public-sector unions, on the other, disagree sharply about what to do about these problems. The State claims that its employment-related debt is “ ‘squeezing core programs in education, public safety, and human services, in addition to limiting [the State’s] ability to pay [its] bills.’ ” Securities Act of 1933 Release No. 9389, 105 S. E. C. Docket 3381 (2013). It therefore “told the Union that it would attempt to address th[e financial] crisis, at least in part, through collective bargaining.” Board Decision 12–13. And “the State’s desire for savings” in fact “dr[o]ve [its] bargaining” positions on matters such as health-insurance benefits and holiday, overtime, and promotion policies. Id., at 13; Illinois Dept. of Central Management Servs. v. AFSCME, Council 31, No. S–CB–16–17 etc., 33 PERI ¶67 (ILRB Dec. 13, 2016) (ALJ Decision), pp. 26–28, 63–66, 224. But when the State offered cost-saving proposals on these issues, the Union countered with very different suggestions. Among other things, it advocated wage and tax increases, cutting spending “to Wall Street financial institutions,” and reforms to Illinois’ pension and tax systems (such as closing “corporate tax loopholes,” “[e]xpanding the base of the state sales tax,” and “allowing an income tax that is adjusted in accordance with ability to pay”). Id., at 27–28. To suggest that speech on such matters is not of great public concern—or that it is not directed at the “public square,” post, at 16 (Kagan, J., dissenting)—is to deny reality. In addition to affecting how public money is spent, union speech in collective bargaining addresses many other important matters. As the examples offered by respondents’ own amici show, unions express views on a wide range of subjects—education, child welfare, healthcare, and minority rights, to name a few. See, e.g., Brief for American Federation of Teachers as Amicus Curiae 15–27; Brief for Child Protective Service Workers et al. as Amici Curiae 5–13; Brief for Human Rights Campaign et al. as Amici Curiae 10–17; Brief for National Women’s Law Center et al. as Amici Curiae 14–30. What unions have to say on these matters in the context of collective bargaining is of great public importance. Take the example of education, which was the focus of briefing and argument in Friedrichs. The public importance of subsidized union speech is especially apparent in this field, since educators make up by far the largest category of state and local government employees, and education is typically the largest component of state and local government expenditures.[15] Speech in this area also touches on fundamental questions of education policy. Should teacher pay be based on seniority, the better to retain experienced teachers? Or should schools adopt merit-pay systems to encourage teachers to get the best results out of their students?[16] Should districts transfer more experienced teachers to the lower performing schools that may have the greatest need for their skills, or should those teachers be allowed to stay where they have put down roots?[17] Should teachers be given tenure protection and, if so, under what conditions? On what grounds and pursuant to what procedures should teachers be subject to discipline or dismissal? How should teacher performance and student progress be measured—by standardized tests or other means? Unions can also speak out in collective bargaining on controversial subjects such as climate change,[18] the Confederacy,[19] sexual orientation and gender identity,[20] evolution,[21] and minority religions.[22] These are sensitive political topics, and they are undoubtedly matters of profound “ ‘value and concern to the public.’ ” Snyder v. Phelps, 562 U. S. 443, 453 (2011). We have often recognized that such speech “ ‘occupies the highest rung of the hierarchy of First Amendment values’ ” and merits “ ‘special protection.’ ” Id., at 452. What does the dissent say about the prevalence of such issues? The most that it is willing to admit is that “some” issues that arise in collective bargaining “raise important non-budgetary disputes.” Post, at 17. Here again, the dissent refuses to recognize what actually occurs in public-sector collective bargaining. Even union speech in the handling of grievances may be of substantial public importance and may be directed at the “public square.” Post, at 16. For instance, the Union respondent in this case recently filed a grievance seeking to compel Illinois to appropriate $75 million to fund a 2% wage increase. State v. AFSCME Council 31, 2016 IL 118422, 51 N. E. 3d 738, 740–742, and n. 4. In short, the union speech at issue in this case is overwhelmingly of substantial public concern. C The only remaining question under Pickering is whether the State’s proffered interests justify the heavy burden that agency fees inflict on nonmembers’ First Amendment interests. We have already addressed the state interests asserted in Abood—promoting “labor peace” and avoiding free riders, see supra, at 11–18—and we will not repeat that analysis. In Harris and this case, defenders of Abood have as- serted a different state interest—in the words of the Harris dissent, the State’s “interest in bargaining with an adequately funded exclusive bargaining agent.” 573 U. S., at ___ (Kagan, J., dissenting) (slip op., at 7); see also post, at 6–7 (Kagan, J., dissenting). This was not “the interest Abood recognized and protected,” Harris, supra, at ___ (slip op., at 7) (Kagan, J., dissenting), and, in any event, it is insufficient. Although the dissent would accept without any serious independent evaluation the State’s assertion that the absence of agency fees would cripple public-sector unions and thus impair the efficiency of government operations, see post, at 8–9, 11, ample experience, as we have noted, supra, at 12, shows that this is questionable. Especially in light of the more rigorous form of Pickering analysis that would apply in this context, see supra, at 23–25, the balance tips decisively in favor of the employees’ free speech rights.[23] We readily acknowledge, as Pickering did, that “the State has interests as an employer in regulating the speech of its employees that differ significantly from those it possesses in connection with regulation of the speech of the citizenry in general.” 391 U. S., at 568. Our analysis is consistent with that principle. The exacting scrutiny standard we apply in this case was developed in the context of commercial speech, another area where the government has traditionally enjoyed greater-than-usual power to regulate speech. See supra, at 10. It is also not disputed that the State may require that a union serve as exclusive bargaining agent for its employees—itself a significant impingement on associational freedoms that would not be tolerated in other contexts. We simply draw the line at allowing the government to go further still and require all employees to support the union irrespective of whether they share its views. Nothing in the Pickering line of cases requires us to uphold every speech restriction the government imposes as an employer. See Pickering, supra, at 564–566 (holding teacher’s dismissal for criticizing school board unconstitutional); Rankin v. McPherson, 483 U. S. 378, 392 (1987) (holding clerical employ- ee’s dismissal for supporting assassination attempt on President unconstitutional); Treasury Employees, 513 U. S., at 477 (holding federal-employee honoraria ban unconstitutional). VI For the reasons given above, we conclude that public-sector agency-shop arrangements violate the First Amendment, and Abood erred in concluding otherwise. There remains the question whether stare decisis nonetheless counsels against overruling Abood. It does not. “Stare decisis is the preferred course because it promotes the evenhanded, predictable, and consistent development of legal principles, fosters reliance on judicial decisions, and contributes to the actual and perceived integrity of the judicial process.” Payne v. Tennessee, 501 U. S. 808, 827 (1991). We will not overturn a past decision unless there are strong grounds for doing so. United States v. International Business Machines Corp., 517 U. S. 843, 855–856 (1996); Citizens United, 558 U. S., at 377 (Roberts, C. J., concurring). But as we have often recognized, stare decisis is “ ‘not an inexorable command.’ ” Pearson v. Callahan, 555 U. S. 223, 233 (2009); see also Lawrence v. Texas, 539 U. S. 558, 577 (2003); State Oil Co. v. Khan, 522 U. S. 3, 20 (1997); Agostini v. Felton, 521 U. S. 203, 235 (1997); Seminole Tribe of Fla. v. Florida, 517 U. S. 44, 63 (1996); Payne, supra, at 828. The doctrine “is at its weakest when we interpret the Constitution because our interpretation can be altered only by constitutional amendment or by overruling our prior decisions.” Agostini, supra, at 235. And stare decisis applies with perhaps least force of all to decisions that wrongly denied First Amendment rights: “This Court has not hesitated to overrule decisions offensive to the First Amendment (a fixed star in our constitutional constellation, if there is one).” Federal Election Comm’n v. Wisconsin Right to Life, Inc., 551 U. S. 449, 500 (2007) (Scalia, J., concurring in part and concurring in judgment) (internal quotation marks omitted); see also Citizens United, supra, at 362–365 (overruling Austin, 494 U. S. 652); Barnette, 319 U. S., at 642 (overruling Minersville School Dist. v. Gobitis, 310 U. S. 586 (1940)). Our cases identify factors that should be taken into account in deciding whether to overrule a past decision. Five of these are most important here: the quality of Abood’s reasoning, the workability of the rule it established, its consistency with other related decisions, developments since the decision was handed down, and reliance on the decision. After analyzing these factors, we conclude that stare decisis does not require us to retain Abood. A An important factor in determining whether a precedent should be overruled is the quality of its reasoning, see Citizens United, 558 U. S., at 363–364; id., at 382–385 (Roberts, C. J., concurring); Lawrence, 539 U. S., at 577–578, and as we explained in Harris, Abood was poorly reasoned, see 573 U. S., at ___–___ (slip op., at 17–20). We will summarize, but not repeat, Harris’s lengthy discussion of the issue. Abood went wrong at the start when it concluded that two prior decisions, Railway Employes v. Hanson, 351 U. S. 225 (1956), and Machinists v. Street, 367 U. S. 740 (1961), “appear[ed] to require validation of the agency-shop agreement before [the Court].” 431 U. S., at 226. Properly understood, those decisions did no such thing. Both cases involved Congress’s “bare authorization” of private-sector union shops under the Railway Labor Act. Street, supra, at 749 (emphasis added).[24] Abood failed to appreciate that a very different First Amendment question arises when a State requires its employees to pay agency fees. See Harris, supra, at ___ (slip op., at 17). Moreover, neither Hanson nor Street gave careful consideration to the First Amendment. In Hanson, the primary questions were whether Congress exceeded its power under the Commerce Clause or violated substantive due process by authorizing private union-shop arrangements under the Commerce and Due Process Clauses. 351 U. S., at 233–235. After deciding those questions, the Court summarily dismissed what was essentially a facial First Amendment challenge, noting that the record did not substantiate the challengers’ claim. Id., at 238; see Harris, supra, at ___ (slip op., at 17). For its part, Street was decided as a matter of statutory construction, and so did not reach any constitutional issue. 367 U. S., at 749–750, 768–769. Abood nevertheless took the view that Hanson and Street “all but decided” the important free speech issue that was before the Court. Harris, 573 U. S., at ___ (slip op., at 17). As we said in Harris, “[s]urely a First Amendment issue of this importance deserved better treatment.” Ibid. Abood’s unwarranted reliance on Hanson and Street appears to have contributed to another mistake: Abood judged the constitutionality of public-sector agency fees under a deferential standard that finds no support in our free speech cases. (As noted, supra, at 10–11, today’s dissent makes the same fundamental mistake.) Abood did not independently evaluate the strength of the government interests that were said to support the challenged agency-fee provision; nor did it ask how well that provision actually promoted those interests or whether they could have been adequately served without impinging so heavily on the free speech rights of nonmembers. Rather, Abood followed Hanson and Street, which it interpreted as having deferred to “the legislative assessment of the important contribution of the union shop to the system of labor relations established by Congress.” 431 U. S., at 222 (emphasis added). But Hanson deferred to that judgment in deciding the Commerce Clause and substantive due process questions that were the focus of the case. Such deference to legislative judgments is inappropriate in deciding free speech issues. If Abood had considered whether agency fees were actually needed to serve the asserted state interests, it might not have made the serious mistake of assuming that one of those interests—“labor peace”—demanded, not only that a single union be designated as the exclusive representative of all the employees in the relevant unit, but also that nonmembers be required to pay agency fees. Deferring to a perceived legislative judgment, Abood failed to see that the designation of a union as exclusive representative and the imposition of agency fees are not inextricably linked. See supra, at 11–12; Harris, supra, at ___ (slip op., at 31). Abood also did not sufficiently take into account the difference between the effects of agency fees in public- and private-sector collective bargaining. The challengers in Abood argued that collective bargaining with a government employer, unlike collective bargaining in the private sector, involves “inherently ‘political’ ” speech. 431 U. S., at 226. The Court did not dispute that characterization, and in fact conceded that “decisionmaking by a public employer is above all a political process” driven more by policy concerns than economic ones. Id., at 228; see id., at 228–231. But (again invoking Hanson), the Abood Court asserted that public employees do not have “weightier First Amendment interest[s]” against compelled speech than do private employees. Id., at 229. That missed the point. Assuming for the sake of argument that the First Amendment applies at all to private-sector agency-shop arrangements, the individual interests at stake still differ. “In the public sector, core issues such as wages, pensions, and benefits are important political issues, but that is generally not so in the private sector.” Harris, 573 U. S., at ___ (slip op., at 17). Overlooking the importance of this distinction, “Abood failed to appreciate the conceptual difficulty of distinguishing in public-sector cases between union expenditures that are made for collective-bargaining purposes and those that are made to achieve political ends.” Id., at ___ (slip op., at 18). Likewise, “Abood does not seem to have anticipated the magnitude of the practical administrative problems that would result in attempting to classify public-sector union expenditures as either ‘chargeable’ . . . or nonchargeable.” Ibid. Nor did Abood “foresee the practical problems that would face objecting nonmembers.” Id., at ___ (slip op., at 19). In sum, as detailed in Harris, Abood was not well reasoned.[25] B Another relevant consideration in the stare decisis calculus is the workability of the precedent in question, Montejo v. Louisiana, 556 U. S. 778, 792 (2009), and that factor also weighs against Abood. 1 Abood’s line between chargeable and nonchargeable union expenditures has proved to be impossible to draw with precision. We tried to give the line some definition in Lehnert. There, a majority of the Court adopted a three-part test requiring that chargeable expenses (1) be “ ‘germane’ ” to collective bargaining, (2) be “justified” by the government’s labor-peace and free-rider interests, and (3) not add “significantly” to the burden on free speech, 500 U. S., at 519, but the Court splintered over the application of this test, see id., at 519–522 (plurality opinion); id., at 533–534 (Marshall, J., concurring in part and dissenting in part). That division was not surprising. As the Lehnert dissenters aptly observed, each part of the majority’s test “involves a substantial judgment call,” id., at 551 (opinion of Scalia, J.), rendering the test “altogether malleable” and “no[t] principled,” id., at 563 (Kennedy, J., concurring in judgment in part and dissenting in part). Justice Scalia presciently warned that Lehnert’s amorphous standard would invite “perpetua[l] give-it-a-try litigation,” id., at 551, and the Court’s experience with union lobbying expenses illustrates the point. The Lehnert plurality held that money spent on lobbying for increased education funding was not chargeable. Id., at 519–522. But Justice Marshall—applying the same three-prong test—reached precisely the opposite conclusion. Id., at 533–542. And Lehnert failed to settle the matter; States and unions have continued to “give it a try” ever since. In Knox, for example, we confronted a union’s claim that the costs of lobbying the legislature and the electorate about a ballot measure were chargeable expenses under Lehnert. See Brief for Respondent in Knox v. Service Employees, O. T. 2011, No. 10–1121, pp. 48–53. The Court rejected this claim out of hand, 567 U. S., at 320–321, but the dissent refused to do so, id., at 336 (opinion of Breyer, J.). And in the present case, nonmembers are required to pay for unspecified “[l]obbying” expenses and for “[s]ervices” that “may ultimately inure to the benefit of the members of the local bargaining unit.” App. to Pet. for Cert. 31a–32a. That formulation is broad enough to encompass just about anything that the union might choose to do. Respondents agree that Abood’s chargeable-nonchargeable line suffers from “a vagueness problem,” that it sometimes “allows what it shouldn’t allow,” and that “a firm[er] line c[ould] be drawn.” Tr. of Oral Arg. 47–48. They therefore argue that we should “consider revisiting” this part of Abood. Tr. of Oral Arg. 66; see Brief for Union Respondent 46–47; Brief for State Respondents 30. This concession only underscores the real- ity that Abood has proved unworkable: Not even the parties defending agency fees support the line that it has taken this Court over 40 years to draw. 2 Objecting employees also face a daunting and expensive task if they wish to challenge union chargeability determinations. While Hudson requires a union to provide nonmembers with “sufficient information to gauge the propriety of the union’s fee,” 475 U. S., at 306, the Hudson notice in the present case and in others that have come before us do not begin to permit a nonmember to make such a determination. In this case, the notice lists categories of expenses and sets out the amount in each category that is said to be attributable to chargeable and nonchargeable expenses. Here are some examples regarding the Union respondent’s expenditures: See App. to Pet. for Cert. 35a–36a. How could any nonmember determine whether these numbers are even close to the mark without launching a legal challenge and retaining the services of attorneys and accountants? Indeed, even with such services, it would be a laborious and difficult task to check these figures.[26] The Union respondent argues that challenging its chargeability determinations is not burdensome because the Union pays for the costs of arbitration, see Brief for Union Respondent 10–11, but objectors must still pay for the attorneys and experts needed to mount a serious challenge. And the attorney’s fees incurred in such a proceeding can be substantial. See, e.g., Knox v. Chiang, 2013 WL 2434606, *15 (ED Cal., June 5, 2013) (attorney’s fees in Knox exceeded $1 million). The Union respondent’s suggestion that an objector could obtain adequate review without even showing up at an arbitration, see App. to Pet. for Cert. 40a–41a, is therefore farfetched. C Developments since Abood, both factual and legal, have also “eroded” the decision’s “underpinnings” and left it an outlier among our First Amendment cases. United States v. Gaudin, 515 U. S. 506, 521 (1995). 1 Abood pinned its result on the “unsupported empirical assumption” that “the principle of exclusive representation in the public sector is dependent on a union or agency shop.” Harris, 573 U. S., at ___ (slip op., at 20); Abood, 431 U. S., at 220–222. But, as already noted, experience has shown otherwise. See supra, at 11–12. It is also significant that the Court decided Abood against a very different legal and economic backdrop. Public-sector unionism was a relatively new phenomenon in 1977. The first State to permit collective bargaining by government employees was Wisconsin in 1959, R. Kearney & P. Mareschal, Labor Relations in the Public Sector 64 (5th ed. 2014), and public-sector union membership remained relatively low until a “spurt” in the late 1960’s and early 1970’s, shortly before Abood was decided, Freeman, Unionism Comes to the Public Sector, 24 J. Econ. Lit. 41, 45 (1986). Since then, public-sector union membership has come to surpass private-sector union membership, even though there are nearly four times as many total private-sector employees as public-sector employees. B. Hirsch & D. Macpherson, Union Membership and Earnings Data Book 9–10, 12, 16 (2013 ed.). This ascendance of public-sector unions has been marked by a parallel increase in public spending. In 1970, total state and local government expenditures amounted to $646 per capita in nominal terms, or about $4,000 per capita in 2014 dollars. See Dept. of Commerce, Statistical Abstract of the United States: 1972, p. 419; CPI Inflation Calculator, BLS, http://data.bls.gov/cgi-bin/cpicalc.pl. By 2014, that figure had ballooned to approximately $10,238 per capita. ProQuest, Statistical Abstract of the United States: 2018, pp. 17, Table 14, 300, Table 469. Not all that increase can be attributed to public-sector unions, of course, but the mounting costs of public-employee wages, benefits, and pensions undoubtedly played a substantial role. We are told, for example, that Illinois’ pension funds are underfunded by $129 billion as a result of generous public-employee retirement packages. Brief for Jason R. Barclay et al. as Amici Curiae 9, 14. Unsustainable collective-bargaining agreements have also been blamed for multiple municipal bankruptcies. See Brief for State of Michigan et al. as Amici Curiae 10–19. These developments, and the political debate over public spending and debt they have spurred, have given collective-bargaining issues a political valence that Abood did not fully appreciate. 2 Abood is also an “anomaly” in our First Amendment jurisprudence, as we recognized in Harris and Knox. Harris, supra, at ___ (slip op., at 8); Knox, 567 U. S., at 311. This is not an altogether new observation. In Abood itself, Justice Powell faulted the Court for failing to perform the “ ‘exacting scrutiny’ ” applied in other cases involving significant impingements on First Amendment rights. 431 U. S., at 259; see id., at 259–260, and n. 14. Our later cases involving compelled speech and association have also employed exacting scrutiny, if not a more demanding standard. See, e.g., Roberts, 468 U. S., at 623; United Foods, 533 U. S., at 414. And we have more recently refused, even in agency-fee cases, to extend Abood beyond circumstances where it directly controls. See Knox, supra, at 314; Harris, supra, at ___–___ (slip op., at 28–29). Abood particularly sticks out when viewed against our cases holding that public employees generally may not be required to support a political party. See Elrod, 427 U. S. 347; Branti, 445 U. S. 507; Rutan, 497 U. S. 62; O’Hare Truck Service, Inc. v. City of Northlake, 518 U. S. 712 (1996). The Court reached that conclusion despite a “long tradition” of political patronage in government. Rutan, supra, at 95 (Scalia, J., dissenting); see also Elrod, 427 U. S., at 353 (plurality opinion); id., at 377–378 (Powell, J., dissenting). It is an odd feature of our First Amendment cases that political patronage has been deemed largely unconstitutional, while forced subsidization of union speech (which has no such pedigree) has been largely permitted. As Justice Powell observed: “I am at a loss to understand why the State’s decision to adopt the agency shop in the public sector should be worthy of greater deference, when challenged on First Amendment grounds, than its decision to adhere to the tradition of political patronage.” Abood, supra, at 260, n. 14 (opinion concurring in judgment) (citing Elrod, supra, at 376–380, 382–387 (Powell, J., dissenting); emphasis added). We have no occasion here to reconsider our political patronage decisions, but Justice Powell’s observation is sound as far as it goes. By overruling Abood, we end the oddity of privileging compelled union support over compelled party support and bring a measure of greater coherence to our First Amendment law. D In some cases, reliance provides a strong reason for adhering to established law, see, e.g., Hilton v. South Carolina Public Railways Comm’n, 502 U. S. 197, 202–203 (1991), and this is the factor that is stressed most strongly by respondents, their amici, and the dissent. They contend that collective-bargaining agreements now in effect were negotiated with agency fees in mind and that unions may have given up other benefits in exchange for provisions granting them such fees. Tr. of Oral Arg. 67–68; see Brief for State Respondents 54; Brief for Union Respondent 50; post, at 22–26 (Kagan, J., dissenting). In this case, however, reliance does not carry decisive weight. For one thing, it would be unconscionable to permit free speech rights to be abridged in perpetuity in order to preserve contract provisions that will expire on their own in a few years’ time. “The fact that [public-sector unions] may view [agency fees] as an entitlement does not establish the sort of reliance interest that could outweigh the countervailing interest that [nonmembers] share in having their constitutional rights fully protected.” Arizona v. Gant, 556 U. S. 332, 349 (2009). For another, Abood does not provide “a clear or easily applicable standard, so arguments for reliance based on its clarity are misplaced.” South Dakota v. Wayfair, Inc., ante, at 20; see supra, at 38–41. This is especially so because public-sector unions have been on notice for years regarding this Court’s misgivings about Abood. In Knox, decided in 2012, we described Abood as a First Amendment “anomaly.” 567 U. S., at 311. Two years later in Harris, we were asked to overrule Abood, and while we found it unnecessary to take that step, we cataloged Abood’s many weaknesses. In 2015, we granted a petition for certiorari asking us to review a decision that sustained an agency-fee arrangement under Abood. Friedrichs v. California Teachers Assn., 576 U. S. ___. After exhaustive briefing and argument on the question whether Abood should be overruled, we affirmed the decision below by an equally divided vote. 578 U. S. ___ (2016) (per curiam). During this period of time, any public-sector union seeking an agency-fee provision in a collective-bargaining agreement must have understood that the constitutionality of such a provision was uncertain. That is certainly true with respect to the collective-bargaining agreement in the present case. That agreement initially ran from July 1, 2012, until June 30, 2015. App. 331. Since then, the agreement has been extended pursuant to a provision providing for automatic renewal for an additional year unless either party gives timely notice that it desires to amend or terminate the contract. Ibid. Thus, for the past three years, the Union could not have been confident about the continuation of the agency-fee arrangement for more than a year at a time. Because public-sector collective-bargaining agreements are generally of rather short duration, a great many of those now in effect probably began or were renewed since Knox (2012) or Harris (2014). But even if an agreement antedates those decisions, the union was able to protect itself if an agency-fee provision was essential to the overall bargain. A union’s attorneys undoubtedly understand that if one provision of a collective-bargaining agreement is found to be unlawful, the remaining provisions are likely to remain in effect. See NLRB v. Rockaway News Supply Co., 345 U. S. 71, 76–79 (1953); see also 8 R. Lord, Williston on Contracts §19:70 (4th ed. 2010). Any union believing that an agency-fee provision was essential to its bargain could have insisted on a provision giving it greater protection. The agreement in the present case, by contrast, provides expressly that the invalidation of any part of the agreement “shall not invalidate the remaining portions,” which “shall remain in full force and effect.” App. 328. Such severability clauses ensure that “entire contracts” are not “br[ought] down” by today’s ruling. Post, at 23, n. 5 (Kagan, J., dissenting). In short, the uncertain status of Abood, the lack of clarity it provides, the short-term nature of collective-bargaining agreements, and the ability of unions to protect themselves if an agency-fee provision was crucial to its bargain all work to undermine the force of reliance as a factor supporting Abood.[27] * * * We recognize that the loss of payments from nonmembers may cause unions to experience unpleasant transition costs in the short term, and may require unions to make adjustments in order to attract and retain members. But we must weigh these disadvantages against the considerable windfall that unions have received under Abood for the past 41 years. It is hard to estimate how many billions of dollars have been taken from nonmembers and transferred to public-sector unions in violation of the First Amendment. Those unconstitutional exactions cannot be allowed to continue indefinitely. All these reasons—that Abood’s proponents have abandoned its reasoning, that the precedent has proved unworkable, that it conflicts with other First Amendment decisions, and that subsequent developments have eroded its underpinnings—provide the “ ‘special justification[s]’ ” for overruling Abood. Post, at 19 (Kagan, J., dissenting) (quoting Kimble v. Marvel Entertainment, LLC, 576 U. S. ___, ___ (2015) (slip op., at 8)).[28] VII For these reasons, States and public-sector unions may no longer extract agency fees from nonconsenting employees. Under Illinois law, if a public-sector collective-bargaining agreement includes an agency-fee provision and the union certifies to the employer the amount of the fee, that amount is automatically deducted from the nonmember’s wages. §315/6(e). No form of employee consent is required. This procedure violates the First Amendment and cannot continue. Neither an agency fee nor any other payment to the union may be deducted from a nonmember’s wages, nor may any other attempt be made to collect such a payment, unless the employee affirmatively consents to pay. By agreeing to pay, nonmembers are waiving their First Amendment rights, and such a waiver cannot be presumed. Johnson v. Zerbst, 304 U. S. 458, 464 (1938); see also Knox, 567 U. S., at 312–313. Rather, to be effective, the waiver must be freely given and shown by “clear and compelling” evidence. Curtis Publishing Co. v. Butts, 388 U. S. 130, 145 (1967) (plurality opinion); see also College Savings Bank v. Florida Prepaid Postsecondary Ed. Expense Bd., 527 U. S. 666, 680–682 (1999). Unless employees clearly and affirmatively consent before any money is taken from them, this standard cannot be met. * * * Abood was wrongly decided and is now overruled. The judgment of the United States Court of Appeals for the Seventh Circuit is reversed, and the case is remanded for further proceedings consistent with this opinion. It is so ordered. Notes 1 See Bureau of Labor Statistics (BLS), Labor Force Statistics From the Current Population Survey (Table 42) (2017), https://www.bls.gov/cps/tables.htm (all Internet materials as visited June 26, 2018). 2 See Union Membership and Coverage Database From the Current Population Survey (Jan. 21, 2018), unionstats.com. 3 See National Conference of State Legislatures, Right-to-Work States (2018), http://www.ncsl.org/research/labor-and-employment/right-to-work-laws-and-bills.aspx#chart; see also, e.g., Brief for Mackinac Center for Public Policy as Amicus Curiae 27–28, 34–36. 4 The collective-action problem cited by the dissent, post, at 6, is not specific to the agency-fee context. And contrary to the dissent’s suggestion, it is often not practical for an entity that lobbies or advocates on behalf of the members of a group to tailor its message so that only its members benefit from its efforts. Consider how effective it would be for a group that advocates on behalf of, say, seniors, to argue that a new measure should apply only to its dues-paying members. 5 In order to obtain that status, a union must petition to be recognized and campaign to win majority approval. Ill. Comp. Stat., ch. 5, §315/9(a) (2016); see, e.g., County of Du Page v. Illinois Labor Relations Bd., 231 Ill. 2d 593, 597–600, 900 N. E. 2d 1095, 1098–1099 (2008). And unions eagerly seek this support. See, e.g., Brief for Employees of the State of Minnesota Court System as Amici Curiae 9–17. 6 There is precedent for such arrangements. Some States have laws providing that, if an employee with a religious objection to paying an agency fee “requests the [union] to use the grievance procedure or arbitration procedure on the employee’s behalf, the [union] is authorized to charge the employee for the reasonable cost of using such procedure.” E.g., Cal. Govt. Code Ann. §3546.3 (West 2010); cf. Ill. Comp. Stat., ch. 5, §315/6(g) (2016). This more tailored alternative, if applied to other objectors, would prevent free ridership while imposing a lesser burden on First Amendment rights. 7 Indeed, under common law, “collective bargaining was unlawful,” Teamsters v. Terry, 494 U. S. 558, 565–566 (1990) (plurality opinion); see N. Citrine, Trade Union Law 4–7, 9–10 (2d ed. 1960); Notes, Legal-ity of Trade Unions at Common Law, 25 Harv. L. Rev. 465, 466 (1912), and into the 20th century, every individual employee had the “liberty of contract” to “sell his labor upon such terms as he deem[ed] proper,” Adair v. United States, 208 U. S. 161, 174–175 (1908); see R. Morris, Government and Labor in Early America 208, 529 (1946). So even the concept of a private third-party entity with the power to bind employees on the terms of their employment likely would have been foreign to the Founders. We note this only to show the problems inherent in the Union respondent’s argument; we are not in any way questioning the foundations of modern labor law. 8 See, e.g., Ellsworth, The Landholder, VII (1787), in Essays on the Constitution of the United States 167–171 (P. Ford ed. 1892); Webster, On Test Laws, Oaths of Allegiance and Abjuration, and Partial Exclusions from Office, in A Collection of Essays and Fugitiv[e] Writings 151–153 (1790). 9 Justice Powell’s separate opinion did invoke Pickering in a relevant sense, but he did so only to acknowledge the State’s relatively greater interest in regulating speech when it acts as employer than when it acts as sovereign. Abood v. Detroit Bd. of Ed., 431 U. S. 209, 259 (1977) (concurring in judgment). In the very next sentence, he explained that “even in public employment, a significant impairment of First Amendment rights must survive exacting scrutiny.” Ibid. (internal quotation marks omitted). That is the test we apply today. 10 See Brief for State of Michigan et al. as Amici Curiae 9–24. Nationwide, the cost of state and local employees’ wages and benefits, for example, is nearly $1.5 trillion—more than half of those jurisdictions’ total expenditures. See Dept. of Commerce, Bureau of Economic Analysis, National Data, GDP & Personal Income, Table 6.2D, line 92 (Aug. 3, 2017), and Table 3.3, line 37 (May 30, 2018), https://www.bea.gov/iTable/iTable.cfm?reqid=19&step=2#reqid=19&step=2&isuri=1&1921=survey. And many States and cities struggle with unfunded pension and retiree healthcare liabilities and other budget issues. 11 PEW Charitable Trusts, Fiscal 50: State Trends and Analysis (updated May 17, 2016), http://www.pewtrusts.org/en/research-and-analysis/data-visualizations/2014/fiscal-50#ind4. 12 See Brief for Jason R. Barclay et al. as Amici Curiae 9; M. Egan, How Illinois Became America’s Most Messed-Up State, CNN Money (July 1, 2017), https://cnnmon.ie/2tp9NX5. 13 Brief for Jason R. Barclay et al. as Amici Curiae 9. 14 E. Campbell, S&P, Moody’s Downgrade Illinois to Near Junk, Lowest Ever for a U. S. State, Bloomberg (June 1, 2017), https://bloom.bg/2roEJUc. 15 See National Association of State Budget Officers, Summary: Spring 2018 Fiscal Survey of States 2 (June 14, 2018), http://www.nasbo.org; ProQuest Statistical Abstract of the United States: 2018, pp. 306, Table 476, 321, Table 489. 16 See Rogers, School Districts ‘Race to the Top’ Despite Teacher Dispute, Marin Independent J., June 19, 2010. 17 See Sawchuk, Transferring Top Teachers Has Benefits: Study Probes Moving Talent to Low-Performing Schools, Education Week, Nov. 13, 2013, pp. 1, 13. 18 See Tucker, Textbooks Equivocate on Global Warming: Stanford Study Finds Portrayal ‘Dishonest,’ San Francisco Chronicle, Nov. 24, 2015, p. C1. 19 See Reagan, Anti-Confederacy Movement Rekindles Texas Textbook Controversy, San Antonio Current, Aug. 4, 2015. 20 See Watanabe, How To Teach Gay Issues in 1st Grade? A New Law Requiring California Schools To Have Lessons About LGBT Americans Raises Tough Questions, L. A. Times, Oct. 16, 2011, p. A1. 21 See Goodstein, A Web of Faith, Law and Science in Evolution Suit, N. Y. Times, Sept. 26, 2005, p. A1. 22 See Golden, Defending the Faith: New Battleground in Textbook Wars: Religion in History, Wall St. J., Jan. 25, 2006, p. A1. 23 Claiming that our decision will hobble government operations, the dissent asserts that it would prevent a government employer from taking action against disruptive non-unionized employees in two carefully constructed hypothetical situations. See post, at 17–18. Both hypotheticals are short on potentially important details, but in any event, neither would be affected by our decision in this case. Rather, both would simply call for the application of the standard Pickering test. In one of the hypotheticals, teachers “protest merit pay in the school cafeteria.” Post, at 17. If such a case actually arose, it would be important to know, among other things, whether the teachers involved were supposed to be teaching in their classrooms at the time in question and whether the protest occurred in the presence of students during the student lunch period. If both those conditions were met, the teachers would presumably be violating content-neutral rules regarding their duty to teach at specified times and places, and their conduct might well have a disruptive effect on the educational process. Thus, in the dissent’s hypothetical, the school’s interests might well outweigh those of the teachers, but in this hypothetical case, as in all Pickering cases, the particular facts would be very important. In the other hypothetical, employees agitate for a better health plan “at various inopportune times and places.” Post, at 17. Here, the lack of factual detail makes it impossible to evaluate how the Pickering balance would come out. The term “agitat[ion]” can encompass a wide range of conduct, as well as speech. Post, at 17. And the time and place of the agitation would also be important. 24 No First Amendment issue could have properly arisen in those cases unless Congress’s enactment of a provision allowing, but not requiring, private parties to enter into union-shop arrangements was sufficient to establish governmental action. That proposition was debatable when Abood was decided, and is even more questionable today. See American Mfrs. Mut. Ins. Co. v. Sullivan, 526 U. S. 40, 53 (1999); Jackson v. Metropolitan Edison Co., 419 U. S. 345, 357 (1974). Compare, e.g., White v. Communications Workers of Am., AFL–CIO, Local 13000, 370 F. 3d 346, 350 (CA3 2004) (no state action), and Kolinske v. Lubbers, 712 F. 2d 471, 477–478 (CADC 1983) (same), with Beck v. Communications Workers of Am., 776 F. 2d 1187, 1207 (CA4 1985) (state action), and Linscott v. Millers Falls Co., 440 F. 2d 14, 16, and n. 2 (CA1 1971) (same). We reserved decision on this question in Communications Workers v. Beck, 487 U. S. 735, 761 (1988), and do not resolve it here. 25 Contrary to the dissent’s claim, see post, at 19, and n. 4, the fact that “[t]he rationale of [Abood] does not withstand careful analysis” is a reason to overrule it, e.g., Lawrence v. Texas, 539 U. S. 558, 577 (2003). And that is even truer when, as here, the defenders of the precedent do not attempt to “defend [its actual] reasoning.” Citizens United v. Federal Election Comm’n, 558 U. S. 310, 363 (2010); id., at 382–385 (Roberts, C. J., concurring). 26 For this reason, it is hardly surprising that chargeability issues have not arisen in many Court of Appeals cases. See post, at 22 (Kagan, J., dissenting). 27 The dissent emphasizes another type of reliance, namely, that “[o]ver 20 States have by now enacted statutes authorizing [agency-fee] provisions.” Post, at 23. But as we explained in Citizens United, “[t]his is not a compelling interest for stare decisis. If it were, legislative acts could prevent us from overruling our own precedents, thereby interfering with our duty ‘to say what the law is.’ ” 558 U. S., at 365 (quoting Marbury v. Madison, 1 Cranch 137, 177 (1803)). Nor does our decision “ ‘require an extensive legislative response.’ ” Post, at 23. States can keep their labor-relations systems exactly as they are—only they cannot force nonmembers to subsidize public-sector unions. In this way, these States can follow the model of the federal government and 28 other States. 28 Unfortunately, the dissent sees the need to resort to accusations that we are acting like “black-robed rulers” who have shut down an “energetic policy debate.” Post, at 27–28. We certainly agree that judges should not “overrid[e] citizens’ choices” or “pick the winning side,” ibid.—unless the Constitution commands that they do so. But when a federal or state law violates the Constitution, the American doctrine of judicial review requires us to enforce the Constitution. Here, States with agency-fee laws have abridged fundamental free speech rights. In holding that these laws violate the Constitution, we are simply enforcing the First Amendment as properly understood, “[t]he very purpose of [which] was to withdraw certain subjects from the vicissitudes of political controversy, to place them beyond the reach of majorities and officials and to establish them as legal principles to be applied by the courts.” West Virginia Bd. of Ed. v. Barnette, 319 U. S. 624, 638 (1943).

583.US.2017_15-1204

Immigration officials are authorized to detain certain aliens in the course of immigration proceedings while they determine whether those aliens may be lawfully present in the country. For example, §1225(b) of Title 8 of the U. S. Code authorizes the detention of certain aliens seeking to enter the country. Section 1225(b)(1) applies to aliens initially determined to be inadmissible due to fraud, misrepresentation, or lack of valid documentation, and to certain other aliens designated by the Attorney General in his discretion. Section 1225(b)(2) is a catchall provision that applies to most other applicants for admission not covered by §1225(b)(1). Under §1225(b)(1), aliens are normally ordered removed “without further hearing or review,” §1225(b)(1)(A)(i), but an alien indicating either an intention to apply for asylum or a credible fear of persecution, §1225(b)(1)(A)(ii), “shall be detained” while that alien’s asylum application is pending, §1225(b)(1)(B)(ii). Aliens covered by §1225(b)(2) in turn “shall be detained for a [removal] proceeding” if an immigration officer “determines that [they are] not clearly and beyond a doubt entitled” to admission. §1225(b)(2)(A). The Government is also authorized to detain certain aliens already in the country. Section 1226(a)’s default rule permits the Attorney General to issue warrants for the arrest and detention of these aliens pending the outcome of their removal proceedings. The Attorney General “may release” these aliens on bond, “[e]xcept as provided in subsection (c) of this section.” Section 1226(c) in turn states that the Attorney General “shall take into custody any alien” who falls into one of the enumerated categories involving criminal offenses and terrorist activities, §1226(c)(1), and specifies that the Attorney General “may release” one of those aliens “only if the Attorney General decides” both that doing so is necessary for witness-protection purposes and that the alien will not pose a danger or flight risk, §1226(c)(2). After a 2004 conviction, respondent Alejandro Rodriguez, a Mexican citizen and a lawful permanent resident of the United States, was detained pursuant to §1226 while the Government sought to remove him. In May 2007, while still litigating his removal, Rodriguez filed a habeas petition, claiming that he was entitled to a bond hearing to determine whether his continued detention was justified. As relevant here, he and the class of aliens he represents allege that §§1225(b), 1226(a), and 1226(c) do not authorize “prolonged” detention in the absence of an individualized bond hearing at which the Government proves by clear and convincing evidence that detention remains justified. The District Court entered a permanent injunction, and the Ninth Circuit affirmed. Relying on the canon of constitutional avoidance, the Ninth Circuit construed §§1225(b) and 1226(c) as imposing an implicit 6-month time limit on an alien’s detention under those sections. After that point, the court held, the Government may continue to detain the alien only under the authority of §1226(a). The court then construed §1226(a) to mean that an alien must be given a bond hearing every six months and that detention beyond the initial 6-month period is permitted only if the Government proves by clear and convincing evidence that further detention is justified. Held: The judgment is reversed, and the case is remanded. 804 F. 3d 1060, reversed and remanded. Justice Alito delivered the opinion of the Court, except as to Part II, concluding that §§1225(b), 1226(a), and 1226(c) do not give detained aliens the right to periodic bond hearings during the course of their detention. The Ninth Circuit misapplied the canon of constitutional avoidance in holding otherwise. Pp. 12–31. (a) The canon of constitutional avoidance “comes into play only when, after the application of ordinary textual analysis, the statute is found to be susceptible of more than one [plausible] construction.” Clark v. Martinez, 543 U. S. 371 . The Ninth Circuit’s interpretations of the provisions at issue, however, are implausible. Pp. 12–13. (b) Read most naturally, §§1225(b)(1) and (b)(2) mandate detention of applicants for admission until certain proceedings have concluded. Until that point, nothing in the statutory text imposes a limit on the length of detention, and neither provision says anything about bond hearings. Pp. 13–19. (1) Nothing in the text of §1225(b)(1) or §1225(b)(2) hints that those provisions have an implicit 6-month time limit on the length of detention. Respondents must show that this is a plausible reading in order to prevail under the canon of constitutional avoidance, but they simply invoke the canon without making any attempt to defend their reading. The Ninth Circuit also all but ignored the statutory text, relying instead on Zadvydas v. Davis, 533 U. S. 678 , as authority for grafting a time limit onto §1225(b)’s text. There, this Court invoked the constitutional-avoidance canon, construing §1231(a)(6)—which provides than an alien subject to a removal order “may be detained” beyond the section’s 90-day removal period—to mean that the alien may not be detained beyond “a period reasonably necessary to secure removal,” id., at 699, presumptively six months, id., at 701. The Court detected ambiguity in the statutory phrase “may be detained” and noted the absence of any explicit statutory limit on the length of permissible detention following the entry of an order of removal. Several material differences distinguish the provisions at issue in this case from Zadvydas’s interpretation of §1231(a)(6). To start, the provisions here, unlike §1231(a)(6), mandate detention for a specified period of time: until immigration officers have finished “consider[ing]” the asylum application, §1225(b)(1)(B)(ii), or until removal proceedings have concluded, §1225(b)(2)(A). Section 1231(a)(6) also uses the ambiguous “may,” while §§1225(b)(1) and (b)(2) use the unequivocal mandate “shall be detained.” There is also a specific provision authorizing temporary parole from §1225(b) detention “for urgent humanitarian reasons or significant public benefit,” §1182(d)(5)(A), but no similar release provision applies to §1231(a)(6). That express exception implies that there are no other circumstances under which aliens detained under §1225(b) may be released. Pp. 14–17. (2) Respondents also claim that the term “for” in §§1225(b)(1) and (b)(2) mandates detention only until the start of applicable proceedings. That is inconsistent with the meanings of “for”—“[d]uring [or] throughout,” 6 Oxford English Dictionary 26, and “with the object or purpose of,” id., at 23—that make sense in the context of the statutory scheme as a whole. Nor does respondents’ reading align with the historical use of “for” in §1225. Pp. 17–19. (c) Section 1226(c)’s language is even clearer. By allowing aliens to be released “only if” the Attorney General decides that certain conditions are met, that provision reinforces the conclusion that aliens detained under its authority are not entitled to be released under any circumstances other than those expressly recognized by the statute. Together with §1226(a), §1226(c) makes clear that detention of aliens within its scope must continue “pending a decision” on removal. Section 1226(c) is thus not silent as to the length of detention. See Demore v. Kim, 538 U. S. 510 . The provision, by expressly stating that covered aliens may be released “only if” certain conditions are met, also unequivocally imposes an affirmative prohibition on releasing them under any other conditions. Finally, because §1226(c) and the PATRIOT Act apply to different categories of aliens in different ways, adopting §1226(c)’s plain meaning will not make any part of the PATRIOT Act, see §1226a(a)(3), superfluous. Pp. 19–22. (d) Nothing in §1226(a), which authorizes the Attorney General to arrest and detain an alien “pending a decision” on removal and which permits the Attorney General to release the alien on bond, supports the imposition of periodic bond hearings every six months in which the Attorney General must prove by clear and convincing evidence that continued detention is necessary. Nor does it hint that the length of detention prior to the bond hearing must be considered in determining whether an alien should be released. Pp. 22–23. (e) The Ninth Circuit should consider the merits of respondents’ constitutional arguments in the first instance. But before doing so, it should also reexamine whether respondents can continue litigating their claims as a class. Pp. 29–31. Alito, J., delivered the opinion of the Court, except as to Part II. Roberts, C. J., and Kennedy, J., joined that opinion in full; Thomas and Gorsuch, JJ., joined as to all but Part II; and Sotomayor, J., joined as to Part III–C. Thomas, J., filed an opinion concurring in part and concurring in the judgment, in which Gorsuch, J., joined except for footnote 6. Breyer, J., filed a dissenting opinion, in which Ginsburg and Sotomayor, JJ., joined. Kagan, J., took no part in the decision of the case.

except as to Part II.[1]*Every day, immigration officials must determine whether to admit or remove the many aliens who have arrived at an official “port of entry” (e.g., an international airport or border crossing) or who have been apprehended trying to enter the country at an unauthorized location. Immigration officials must also determine on a daily basis whether there are grounds for removing any of the aliens who are already present inside the country. The vast majority of these determinations are quickly made, but in some cases deciding whether an alien should be admitted or removed is not as easy. As a result, Congress has authorized immigration officials to detain some classes of aliens during the course of certain immigration proceedings. Detention during those proceedings gives immigration officials time to determine an alien’s status without running the risk of the alien’s either absconding or engaging in criminal activity before a final decision can be made.In this case we are asked to interpret three provisions of U. S. immigration law that authorize the Government to detain aliens in the course of immigration proceedings. All parties appear to agree that the text of these provisions, when read most naturally, does not give detained aliens the right to periodic bond hearings during the course of their detention. But by relying on the constitutional-avoidance canon of statutory interpretation, the Court of Appeals for the Ninth Circuit held that detained aliens have a statutory right to periodic bond hearings under the provisions at issue.Under the constitutional-avoidance canon, when statutory language is susceptible of multiple interpretations, a court may shun an interpretation that raises serious constitutional doubts and instead may adopt an alternative that avoids those problems. But a court relying on that canon still must interpret the statute, not rewrite it. Because the Court of Appeals in this case adopted implausible constructions of the three immigration provisions at issue, we reverse its judgment and remand for further proceedings.IATo implement its immigration policy, the Government must be able to decide (1) who may enter the country and (2) who may stay here after entering.1That process of decision generally begins at the Nation’s borders and ports of entry, where the Government must determine whether an alien seeking to enter the country is admissible. Under122Stat.867,8 U. S. C. §1225, an alien who “arrives in the United States,” or “is present” in this country but “has not been admitted,” is treated as “an applicant for admission.” §1225(a)(1). Applicants for admission must “be inspected by immigration officers” to ensure that they may be admitted into the country consistent with U. S. immigration law. §1225(a)(3).As relevant here, applicants for admission fall into one of two categories, those covered by §1225(b)(1) and those covered by §1225(b)(2). Section 1225(b)(1) applies to aliens initially determined to be inadmissible due to fraud, misrepresentation, or lack of valid documentation. See §1225(b)(1)(A)(i) (citing §§1182(a)(6)(C), (a)(7)). Section 1225(b)(1) also applies to certain other aliens designated by the Attorney General in his discretion. See §1225(b)(1)(A)(iii). Section 1225(b)(2) is broader. It serves as a catchall provision that applies to all applicants for admission not covered by §1225(b)(1) (with specific exceptions not relevant here). See §§1225(b)(2)(A), (B).Both §1225(b)(1) and §1225(b)(2) authorize the detention of certain aliens. Aliens covered by §1225(b)(1) are normally ordered removed “without further hearing or review” pursuant to an expedited removal process. §1225(b)(1)(A)(i). But if a §1225(b)(1) alien “indicates either an intention to apply for asylum . . . or a fear of persecution,” then that alien is referred for an asylum interview. §1225(b)(1)(A)(ii). If an immigration officer determines after that interview that the alien has a credible fear of persecution, “the alien shall be detained for further consideration of the application for asylum.” §1225(b)(1)(B)(ii). Aliens who are instead covered by §1225(b)(2) are detained pursuant to a different process. Those aliens “shall be detained for a [removal] proceeding” if an immigration officer “determines that [they are] not clearly and beyond a doubt entitled to be admitted” into the country. §1225(b)(2)(A).Regardless of which of those two sections authorizes their detention, applicants for admission may be temporarily released on parole “for urgent humanitarian reasons or significant public benefit.” §1182(d)(5)(A); see also 8 CFR §§212.5(b), 235.3 (2017). Such parole, however, “shall not be regarded as an admission of the alien.”8 U. S. C. §1182(d)(5)(A). Instead, when the purpose of the parole has been served, “the alien shall forthwith return or be returned to the custody from which he was paroled and thereafter his case shall continue to be dealt with in the same manner as that of any other applicant for admission to the United States.” Ibid.2Even once inside the United States, aliens do not have an absolute right to remain here. For example, an alien present in the country may still be removed if he or she falls “within one or more . . . classes of deportable aliens.” §1227(a). That includes aliens who were inadmissible at the time of entry or who have been convicted of certain criminal offenses since admission. See §§1227(a)(1), (2).Section 1226 generally governs the process of arresting and detaining that group of aliens pending their removal. As relevant here, §1226 distinguishes between two different categories of aliens. Section 1226(a) sets out the default rule: The Attorney General may issue a warrant for the arrest and detention of an alien “pending a decision on whether the alien is to be removed from the United States.” §1226(a). “Except as provided in subsection (c) of this section,” the Attorney General “may release” an alien detained under §1226(a) “on bond . . . or conditional parole.” Ibid.Section 1226(c), however, carves out a statutory category of aliens who may not be released under §1226(a). Under §1226(c), the “Attorney General shall take into custody any alien” who falls into one of several enumerated categories involving criminal offenses and terrorist activities. §1226(c)(1). The Attorney General may release aliens in those categories “only if the Attorney General decides . . . that release of the alien from custody is necessary” for witness-protection purposes and “the alien satisfies the Attorney General that the alien will not pose a danger to the safety of other persons or of property and is likely to appear for any scheduled proceeding.” §1226(c)(2). Any release under those narrow conditions “shall take place in accordance with a procedure that considers the severity of the offense committed by the alien.” Ibid.[2]In sum, U. S. immigration law authorizes the Government to detain certain aliens seeking admission into the country under §§1225(b)(1) and (b)(2). It also authorizes the Government to detain certain aliens already in the country pending the outcome of removal proceedings under §§1226(a) and (c). The primary issue is the proper interpretation of §§1225(b), 1226(a), and 1226(c).BRespondent Alejandro Rodriguez is a Mexican citizen. Since 1987, he has also been a lawful permanent resident of the United States. In April 2004, after Rodriguez was convicted of a drug offense and theft of a vehicle, the Government detained him under §1226 and sought to remove him from the country. At his removal hearing, Rodriguez argued both that he was not removable and, in the alternative, that he was eligible for relief from removal. In July 2004, an Immigration Judge ordered Rodriguez deported to Mexico. Rodriguez chose to appeal that decision to the Board of Immigration Appeals, but five months later the Board agreed that Rodriguez was subject to mandatory removal. Once again, Rodriguez chose to seek further review, this time petitioning the Court of Appeals for the Ninth Circuit for review of the Board’s decision.In May 2007, while Rodriguez was still litigating his removal in the Court of Appeals, he filed a habeas petition in the District Court for the Central District of California, alleging that he was entitled to a bond hearing to determine whether his continued detention was justified. Rodriguez’s case was consolidated with another, similar case brought by Alejandro Garcia, and together they moved for class certification. The District Court denied their motion, but the Court of Appeals for the Ninth Circuit reversed. See Rodriguez v. Hayes, 591 F. 3d 1105, 1111 (2010). It concluded that the proposed class met the certification requirements of Rule 23 of the Federal Rules of Civil Procedure, and it remanded the case to the District Court. Id., at 1111, 1126.On remand, the District Court certified the following class:“[A]ll non-citizens within the Central District of California who: (1) are or were detained for longer than six months pursuant to one of the general immigration detention statutes pending completion of removal proceedings, including judicial review, (2) are not and have not been detained pursuant to a national security detention statute, and (3) have not been afforded a hearing to determine whether their detention is justified.” Class Certification Order in Rodriguez v. Hayes, CV 07–03239 (CD Cal., Apr. 5, 2010).The District Court named Rodriguez as class representative of the newly certified class, ibid., and then organized the class into four subclasses based on the four “general immigration detention statutes” under which it understood the class members to be detained: Sections 1225(b), 1226(a), 1226(c), and 1231(a). See Order Granting Plaintiff’s Motion for Class Certification in Rodriguez v. Holder, CV 07–03239 (CD Cal., Mar. 8, 2011) (2011 Order); Rodriguez v. Robbins, 715 F. 3d 1127, 1130–1131 (CA9 2013). Each of the four subclasses was certified to pursue declaratory and injunctive relief. 2011 Order. On appeal, the Court of Appeals held that the §1231(a) subclass had been improperly certified, but it affirmed the certification of the other three subclasses. See Rodriguez v. Robbins, 804 F. 3d 1060, 1074, 1085–1086 (CA9 2015).In their complaint, Rodriguez and the other respondents argued that the relevant statutory provisions—§§1225(b), 1226(a), and 1226(c)—do not authorize “prolonged” detention in the absence of an individualized bond hearing at which the Government proves by clear and convincing evidence that the class member’s detention remains justified. Absent such a bond-hearing requirement, respondents continued, those three provisions would violate the Due Process Clause of the Fifth Amendment. In their prayer for relief, respondents thus asked the District Court to require the Government “to provide, after giving notice, individual hearings before an immigration judge for . . . each member of the class, at which [the Government] will bear the burden to prove by clear and convincing evidence that no reasonable conditions will ensure the detainee’s presence in the event of removal and protect the community from serious danger, despite the prolonged length of detention at issue.” Third Amended Complaint in Rodriguez v. Holder, CV 07–03239, p. 31 (CD Cal., Oct. 20, 2010). Respondents also sought declaratory relief. Ibid.As relevant here, the District Court entered a permanent injunction in line with the relief sought by respondents, and the Court of Appeals affirmed. See 804 F. 3d, at 1065. Relying heavily on the canon of constitutional avoidance, the Court of Appeals construed §§1225(b) and 1226(c) as imposing an implicit 6-month time limit on an alien’s detention under these sections. Id., at 1079, 1082. After that point, the Court of Appeals held, the Government may continue to detain the alien only under the authority of §1226(a). Ibid. The Court of Appeals then construed §1226(a) to mean that an alien must be given a bond hearing every six months and that detention beyond the initial 6-month period is permitted only if the Government proves by clear and convincing evidence that further detention is justified. Id., at 1085, 1087.The Government petitioned this Court for review of that decision, and we granted certiorari. 579 U. S. ___ (2016).IIBefore reaching the merits of the lower court’s interpretation, we briefly address whether we have jurisdiction to entertain respondents’ claims. We discuss two potential obstacles, 8 U. S. C. §§1252(b)(9) and 1226(e).AUnder §1252(b)(9):“Judicial review of all questions of law and fact, including interpretation and application of constitutional and statutory provisions, arising from any action taken or proceeding brought to remove an alien from the United States under this subchapter [including §§1225 and 1226] shall be available only in judicial review of a final order under this section.”This provision does not deprive us of jurisdiction. We are required in this case to decide “questions of law,” specifically, whether, contrary to the decision of the Court of Appeals, certain statutory provisions require detention without a bond hearing. We assume for the sake of argument that the actions taken with respect to all the aliens in the certified class constitute “action[s] taken . . . to remove [them] from the United States.”[3] On that assumption, the applicability of §1252(b)(9) turns on whether the legal questions that we must decide “aris[e] from” the actions taken to remove these aliens.It may be argued that this is so in the sense that if those actions had never been taken, the aliens would not be in custody at all. But this expansive interpretation of §1252(b)(9) would lead to staggering results. Suppose, for example, that a detained alien wishes to assert a claim under Bivens v. Six Unknown Fed. Narcotics Agents,403 U. S. 388 (1971), based on allegedly inhumane conditions of confinement. See, e.g., Ziglar v. Abbasi, 582 U. S. ___, ___–___ (2017) (slip op., at 23–29). Or suppose that a detained alien brings a state-law claim for assault against a guard or fellow detainee. Or suppose that an alien is injured when a truck hits the bus transporting aliens to a detention facility, and the alien sues the driver or owner of the truck. The “questions of law and fact” in all those cases could be said to “aris[e] from” actions taken to remove the aliens in the sense that the aliens’ injuries would never have occurred if they had not been placed in detention. But cramming judicial review of those questions into the review of final removal orders would be absurd.Interpreting “arising from” in this extreme way would also make claims of prolonged detention effectively unreviewable. By the time a final order of removal was eventually entered, the allegedly excessive detention would have already taken place. And of course, it is possible that no such order would ever be entered in a particular case, depriving that detainee of any meaningful chance for judicial review.In past cases, when confronted with capacious phrases like “ ‘arising from,’ ” we have eschewed “ ‘uncritical literalism’ ” leading to results that “ ‘no sensible person could have intended.’ ” Gobeille v. Liberty Mut. Ins. Co., 577 U. S. ___, ___ (2016) (slip op., at 6) (interpreting phrase “relate to” in the Employee Retirement Income Security Act of 1974’s pre-emption provision). See also, e.g., FERC v. Electric Power Supply Assn., 577 U. S. ___, ___–___ (2016) (slip op., at 15–16) (interpreting term “affecting” in Federal Power Act); Maracich v. Spears,570 U. S. 48–61 (2013) (interpreting phrase “in connection with” in Driver’s Privacy Protection Act); Dan’s City Used Cars, Inc. v. Pelkey,569 U. S. 251–261 (2013) (interpreting phrase “related to” in Federal Aviation Administration Authorization Act); Celotex Corp. v. Edwards,514 U. S. 300,308 (1995) (interpreting phrase “related to” in Bankruptcy Act). In Reno v. American-Arab Anti-Discrimination Comm.,525 U. S. 471,482 (1999), we took this approach in construing the very phrase that appears in §1252(b)(9). A neighboring provision of the Immigration and Nationality Act refers to “any cause or claim by or on behalf of any alien arising from the decision or action by the Attorney General to commence proceedings, adjudicate cases, or execute removal orders against any alien under this chapter.”8 U. S. C. §1252(g) (emphasis added). We did not interpret this language to sweep in any claim that can technically be said to “arise from” the three listed actions of the Attorney General. Instead, we read the language to refer to just those three specific actions themselves. American-Arab Anti-Discrimination Comm., supra, at 482–483.The parties in this case have not addressed the scope of §1252(b)(9), and it is not necessary for us to attempt to provide a comprehensive interpretation. For present purposes, it is enough to note that respondents are not asking for review of an order of removal; they are not challenging the decision to detain them in the first place or to seek removal; and they are not even challenging any part of the process by which their removability will be determined. Under these circumstances, §1252(b)(9) does not present a jurisdictional bar.[4]BWe likewise hold that §1226(e) does not bar us from considering respondents’ claims.That provision states:“The Attorney General’s discretionary judgment regarding the application of [§1226] shall not be subject to review. No court may set aside any action or decision by the Attorney General under this section regarding the detention or release of any alien or the grant, revocation, or denial of bond or parole.” §1226(e).As we have previously explained, §1226(e) precludes an alien from “challeng[ing] a ‘discretionary judgment’ by the Attorney General or a ‘decision’ that the Attorney General has made regarding his detention or release.” Demore v. Kim,538 U. S. 510,516 (2003). But §1226(e) does not preclude “challenges [to] the statutory framework that permits [the alien’s] detention without bail.” Id., at 517.Respondents mount that second type of challenge here. First and foremost, they are challenging the extent of the Government’s detention authority under the “statutory framework” as a whole. If that challenge fails, they are then contesting the constitutionality of the entire statutory scheme under the Fifth Amendment. Because the extent of the Government’s detention authority is not a matter of “discretionary judgment,” “action,” or “decision,” respondents’ challenge to “the statutory framework that permits [their] detention without bail,” ibid., falls outside of the scope of §1226(e). We may therefore consider the merits of their claims.IIIWhen “a serious doubt” is raised about the constitutionality of an act of Congress, “it is a cardinal principle that this Court will first ascertain whether a construction of the statute is fairly possible by which the question may be avoided.” Crowell v. Benson,285 U. S. 22,62 (1932). Relying on this canon of constitutional avoidance, the Court of Appeals construed §§1225(b), 1226(a), and 1226(c) to limit the permissible length of an alien’s detention without a bond hearing. Without such a construction, the Court of Appeals believed, the “ ‘prolonged detention without adequate procedural protections’ ” authorized by the provisions “ ‘would raise serious constitutional concerns.’ ” 804 F. 3d, at 1077 (quoting Casas-Castrillon v. DHS, 535 F. 3d 942, 950 (CA9 2008)).The canon of constitutional avoidance “comes into play only when, after the application of ordinary textual analysis, the statute is found to be susceptible of more than one construction.” Clark v. Martinez,543 U. S. 371,385 (2005). In the absence of more than one plausible construction, the canon simply “ ‘has no application.’ ” Warger v. Shauers, 574 U. S. ___, ___ (2014) (slip op., at 10) (quoting United States v. Oakland Cannabis Buyers’ Cooperative,532 U. S. 483,494 (2001)).The Court of Appeals misapplied the canon in this case because its interpretations of the three provisions at issue here are implausible. In Parts III–A and III–B, we hold that, subject only to express exceptions, §§1225(b) and 1226(c) authorize detention until the end of applicable proceedings. And in Part III–C, we hold that there is no justification for any of the procedural requirements that the Court of Appeals layered onto §1226(a) without any arguable statutory foundation.AAs noted, §1225(b) applies primarily to aliens seeking entry into the United States (“applicants for admission” in the language of the statute). Section 1225(b) divides these applicants into two categories. First, certain aliens claiming a credible fear of persecution under §1225(b)(1) “shall be detained for further consideration of the application for asylum.” §1225(b)(1)(B)(ii). Second, aliens falling within the scope of §1225(b)(2) “shall be detained for a [removal] proceeding.” §1225(b)(2)(A).Read most naturally, §§1225(b)(1) and (b)(2) thus mandate detention of applicants for admission until certain proceedings have concluded. Section 1225(b)(1) aliens are detained for “further consideration of the application for asylum,” and §1225(b)(2) aliens are in turn detained for “[removal] proceeding[s].” Once those proceedings end, detention under §1225(b) must end as well. Until that point, however, nothing in the statutory text imposes any limit on the length of detention. And neither §1225(b)(1) nor §1225(b)(2) says anything whatsoever about bond hearings.Despite the clear language of §§1225(b)(1) and (b)(2), respondents argue—and the Court of Appeals held—that those provisions nevertheless can be construed to contain implicit limitations on the length of detention. But neither of the two limiting interpretations offered by respondents is plausible.1First, respondents argue that §§1225(b)(1) and (b)(2) contain an implicit 6-month limit on the length of detention. Once that 6-month period elapses, respondents contend, aliens previously detained under those provisions must instead be detained under the authority of §1226(a), which allows for bond hearings in certain circumstances.There are many problems with this interpretation. Nothing in the text of §1225(b)(1) or §1225(b)(2) even hints that those provisions restrict detention after six months, but respondents do not engage in any analysis of the text. Instead, they simply cite the canon of constitutional avoidance and urge this Court to use that canon to read a “six-month reasonableness limitation” into §1225(b). Brief for Respondents 48.That is not how the canon of constitutional avoidance works. Spotting a constitutional issue does not give a court the authority to rewrite a statute as it pleases. Instead, the canon permits a court to “choos[e] between competing plausible interpretations of a statutory text.” Clark, supra, at 381 (emphasis added). To prevail, respondents must thus show that §1225(b)’s detention provisions may plausibly be read to contain an implicit 6-month limit. And they do not even attempt to defend that reading of the text.In much the same manner, the Court of Appeals all but ignored the statutory text. Instead, it read Zadvydas v. Davis,533 U. S. 678 (2001), as essentially granting a license to graft a time limit onto the text of §1225(b). Zadvydas, however, provides no such authority.Zadvydas concerned §1231(a)(6), which authorizes the detention of aliens who have already been ordered removed from the country. Under this section, when an alien is ordered removed, the Attorney General is directed to complete removal within a period of 90 days,8 U. S. C. §1231(a)(1)(A), and the alien must be detained during that period, §1231(a)(2). After that time elapses, however, §1231(a)(6) provides only that certain aliens “may be detained” while efforts to complete removal continue. (Emphasis added.)In Zadvydas, the Court construed §1231(a)(6) to mean that an alien who has been ordered removed may not be detained beyond “a period reasonably necessary to secure removal,” 533 U. S., at 699, and it further held that six months is a presumptively reasonable period, id., at 701. After that, the Court concluded, if the alien “provides good reason to believe that there is no significant likelihood of removal in the reasonably foreseeable future,” the Government must either rebut that showing or release the alien. Ibid.The Zadvydas Court justified this interpretation by invoking the constitutional-avoidance canon, and the Court defended its resort to that canon on the ground that §1231(a)(6) is ambiguous. Specifically, the Court detected ambiguity in the statutory phrase “may be detained.” “ ‘[M]ay,’ ” the Court said, “suggests discretion” but not necessarily “unlimited discretion. In that respect the word ‘may’ is ambiguous.” Id., at 697. The Court also pointed to the absence of any explicit statutory limit on the length of permissible detention following the entry of an order of removal. Ibid.Zadvydas represents a notably generous application of the constitutional-avoidance canon, but the Court of Appeals in this case went much further. It failed to address whether Zadvydas’s reasoning may fairly be applied in this case despite the many ways in which the provision in question in Zadvydas, §1231(a)(6), differs materially from those at issue here, §§1225(b)(1) and (b)(2). Those dif- ferences preclude the reading adopted by the Court of Appeals.To start, §§1225(b)(1) and (b)(2), unlike §1231(a)(6), provide for detention for a specified period of time. Section 1225(b)(1) mandates detention “for further consideration of the application for asylum,” §1225(b)(1)(B)(ii), and §1225(b)(2) requires detention “for a [removal] proceeding,” §1225(b)(2)(A). The plain meaning of those phrases is that detention must continue until immigration officers have finished “consider[ing]” the application for asylum, §1225(b)(1)(B)(ii), or until removal proceedings have concluded, §1225(b)(2)(A). By contrast, Congress left the permissible length of detention under §1231(a)(6) unclear.Moreover, in Zadvydas, the Court saw ambiguity in §1231(a)(6)’s use of the word “may.” Here, by contrast, §§1225(b)(1) and (b)(2) do not use the word “may.” Instead, they unequivocally mandate that aliens falling within their scope “shall” be detained. “Unlike the word ‘may,’ which implies discretion, the word ‘shall’ usually connotes a requirement.” Kingdomware Technologies, Inc. v. United States, 579 U. S. ___, ___ (2016) (slip op., at 9). That requirement of detention precludes a court from finding ambiguity here in the way that Zadvydas found ambiguity in §1231(a)(6).Zadvydas’s reasoning is particularly inapt here because there is a specific provision authorizing release from §1225(b) detention whereas no similar release provision applies to §1231(a)(6). With a few exceptions not relevant here, the Attorney General may “for urgent humanitarian reasons or significant public benefit” temporarily parole aliens detained under §§1225(b)(1) and (b)(2).8 U. S. C. §1182(d)(5)(A). That express exception to detention implies that there are no other circumstances under which aliens detained under §1225(b) may be released. See A. Scalia & B. Garner, Reading Law 107 (2012) (“Negative-Implication Canon[:] The expression of one thing implies the exclusion of others (expressio unius est exclusio al- terius)”). That negative implication precludes the sort of implicit time limit on detention that we found in Zadvydas.[5]In short, a series of textual signals distinguishes the provisions at issue in this case from Zadvydas’s interpretation of §1231(a)(6). While Zadvydas found §1231(a)(6) to be ambiguous, the same cannot be said of §§1225(b)(1) and (b)(2): Both provisions mandate detention until a certain point and authorize release prior to that point only under limited circumstances. As a result, neither provision can reasonably be read to limit detention to six months.2In this Court, respondents advance an interpretation of the language of §§1225(b)(1) and (b)(2) that was never made below, namely, that the term “for,” which appears in both provisions, mandates detention only until the start of applicable proceedings rather than all the way through to their conclusion. Respondents contrast the language of §§1225(b)(1) and (b)(2) authorizing detention “for” further proceedings with another provision’s authorization of detention “pending” further proceedings. See8 U. S. C. §1225(b)(1)(B)(iii)(IV) (“Any alien . . . shall be detained pending a final determination of credible fear of persecution and, if found not to have such a fear, until removed”). According to respondents, that distinction between “for” and “pending” makes an enormous difference. As they see things, the word “pending” authorizes detention throughout subsequent proceedings, but the term “for” means that detention authority ends once subsequent proceedings begin. As a result, respondents argue, once the applicable proceedings commence, §§1225(b)(1) and (b)(2) no longer authorize detention, and the Government must instead look to §1226(a) for continued detention authority.That interpretation is inconsistent with ordinary English usage and is incompatible with the rest of the statute. To be sure, “for” can sometimes mean “in preparation for or anticipation of.” 6 Oxford English Dictionary 24 (2d ed. 1989). But “for” can also mean “[d]uring [or] throughout,” id., at 26, as well as “with the object or purpose of,” id., at 23; see also American Heritage Dictionary 709 (3d ed. 1992) (“Used to indicate the object, aim, or purpose of an action or activity”; “Used to indicate amount, extent, or duration”); Random House Dictionary of the English Language 747 (2d ed. 1987) (“with the object or purpose of”; “during the continuance of”); Webster’s Third New International Dictionary 886 (1993) (“with the purpose or object of”; “to the . . . duration of”). And here, only that second set of definitions makes sense in the context of the statutory scheme as a whole.For example, respondents argue that, once detention authority ends under §§1225(b)(1) and (b)(2), aliens can be detained only under §1226(a). But that section authorizes detention only “[o]n a warrant issued” by the Attorney General leading to the alien’s arrest. §1226(a). If respondents’ interpretation of §1225(b) were correct, then the Government could detain an alien without a warrant at the border, but once removal proceedings began, the Attorney General would have to issue an arrest warrant in order to continue detaining the alien. To put it lightly, that makes little sense.Nor does respondents’ interpretation of the word “for” align with the way Congress has historically used that word in §1225. Consider that section’s text prior to the enactment of the Illegal Immigration Reform and Immigrant Responsibility Act of 1996,110Stat.3009–546. Under the older version of §1225(b), “[e]very alien” within its scope “who may not appear . . . to be clearly and beyond a doubt entitled to [entry] shall be detained for further inquiry to be conducted by a special inquiry officer.”8 U. S. C. §1225(b) (1994 ed.). It would make no sense to read “for further inquiry” as authorizing detention of the alien only until the start of the inquiry; Congress obviously did not mean to allow aliens to feel free to leave once immigration officers asked their first question.In sum, §§1225(b)(1) and (b)(2) mandate detention of aliens throughout the completion of applicable proceedings and not just until the moment those proceedings begin. Of course, other provisions of the immigration statutes do authorize detention “pending” other proceedings or “until” a certain point. See post, at 22–23 (Breyer, J., dissenting) (quoting §1225(b)(1)(B)(iii)(IV)). But there is no “canon of interpretation that forbids interpreting different words used in different parts of the same statute to mean roughly the same thing.” Kirtsaeng v. John Wiley & Sons, Inc.,568 U. S. 519,540 (2013). We decline to invent and apply such a canon here.BWhile the language of §§1225(b)(1) and (b)(2) is quite clear, §1226(c) is even clearer. As noted, §1226 applies to aliens already present in the United States. Section 1226(a) creates a default rule for those aliens by permitting—but not requiring—the Attorney General to issue warrants for their arrest and detention pending removal proceedings. Section 1226(a) also permits the Attorney General to release those aliens on bond, “[e]xcept as provided in subsection (c) of this section.” Section 1226(c) in turn states that the Attorney General “shall take into custody any alien” who falls into one of the enumerated categories involving criminal offenses and terrorist activities.8 U. S. C. §1226(c)(1). Section 1226(c) then goes on to specify that the Attorney General “may release” one of those aliens “only if the Attorney General decides” both that doing so is necessary for witness-protection purposes and that the alien will not pose a danger or flight risk. §1226(c)(2) (emphasis added).Like §1225(b), §1226(c) does not on its face limit the length of the detention it authorizes. In fact, by allowing aliens to be released “only if” the Attorney General decides that certain conditions are met, §1226(c) reinforces the conclusion that aliens detained under its authority are not entitled to be released under any circumstances other than those expressly recognized by the statute. And together with §1226(a), §1226(c) makes clear that detention of aliens within its scope must continue “pending a decision on whether the alien is to be removed from the United States.” §1226(a).In a reprise of their interpretation of §1225(b), respondents argue, and the Court of Appeals held, that §1226(c) should be interpreted to include an implicit 6-month time limit on the length of mandatory detention. Once again, that interpretation falls far short of a “plausible statutory construction.”In defense of their statutory reading, respondents first argue that §1226(c)’s “silence” as to the length of detention “cannot be construed to authorize prolonged mandatory detention, because Congress must use ‘clearer terms’ to authorize ‘long-term detention.’ ” Brief for Respondents 34 (quoting Zadvydas, 533 U. S., at 697). But §1226(c) is not “silent” as to the length of detention. It mandates detention “pending a decision on whether the alien is to be removed from the United States,” §1226(a), and it expressly prohibits release from that detention except for narrow, witness-protection purposes. Even if courts were permitted to fashion 6-month time limits out of statutory silence, they certainly may not transmute existing statutory language into its polar opposite. The constitutional-avoidance canon does not countenance such textual alchemy.Indeed, we have held as much in connection with §1226(c) itself. In Demore v. Kim, 538 U. S., at 529, we distinguished §1226(c) from the statutory provision in Zadvydas by pointing out that detention under §1226(c) has “a definite termination point”: the conclusion of removal proceedings. As we made clear there, that “definite termination point”—and not some arbitrary time limit devised by courts—marks the end of the Government’s detention authority under §1226(c).Respondents next contend that §1226(c)’s limited authorization for release for witness-protection purposes does not imply that other forms of release are forbidden, but this argument defies the statutory text. By expressly stating that the covered aliens may be released “only if” certain conditions are met,8 U. S. C. §1226(c)(2), the statute expressly and unequivocally imposes an affirmative prohibition on releasing detained aliens under any other conditions.Finally, respondents point to a provision enacted as part of the PATRIOT Act[6] and contend that their reading of §1226(c) is needed to prevent that provision from being superfluous. That argument, however, misreads both statutory provisions. Although the two provisions overlap in part, they are by no means congruent.Two differences stand out. First, §1226(c) and the PATRIOT Act cover different categories of aliens. Both apply to certain terrorist suspects, but only §1226(c) reaches aliens convicted of other more common criminal offenses. See §§1226(c)(1)(A)–(C) (aliens inadmissible or deportable under §1182(a)(2); §§1227(a)(2)(A)(ii), (A)(iii), (B), (C), and (D); and §1227(a)(2)(A)(i) under certain conditions). For its part, the PATRIOT Act casts a wider net than §1226(c) insofar as it encompasses certain threats to national security not covered by §1226(c). See §1226a(a)(3) (aliens described in §§1182(a)(3)(A)(i), (iii), and 1227(a)(4)(A)(i), (iii), as well as aliens “engaged in any other activity that endangers the national security of the United States”). In addition, the Government’s detention authority under §1226(c) and the PATRIOT Act is not the same. Under §1226(c), the Government must detain an alien until “a decision on whether the alien is to be removed” is made. §1226(a) (emphasis added). But, subject to exceptions not relevant here, the PATRIOT Act authorizes the Government to detain an alien “until the alien is removed.” §1226a(a)(2) (emphasis added).Far from being redundant, then, §1226(c) and the PATRIOT Act apply to different categories of aliens in different ways. There is thus no reason to depart from the plain meaning of §1226(c) in order to avoid making the provision superfluous.We hold that §1226(c) mandates detention of any alien falling within its scope and that detention may end prior to the conclusion of removal proceedings “only if” the alien is released for witness-protection purposes.CFinally, as noted, §1226(a) authorizes the Attorney General to arrest and detain an alien “pending a decision on whether the alien is to be removed from the United States.” §1226(a). As long as the detained alien is not covered by §1226(c), the Attorney General “may release” the alien on “bond . . . or conditional parole.” §1226(a). Federal regulations provide that aliens detained under §1226(a) receive bond hearings at the outset of detention. See 8 CFR §§236.1(d)(1), 1236.1(d)(1).The Court of Appeals ordered the Government to provide procedural protections that go well beyond the initial bond hearing established by existing regulations—namely, periodic bond hearings every six months in which the Attorney General must prove by clear and convincing evidence that the alien’s continued detention is necessary. Nothing in §1226(a)’s text—which says only that the Attorney General “may release” the alien “on . . . bond”—even remotely supports the imposition of either of those requirements. Nor does §1226(a)’s text even hint that the length of detention prior to a bond hearing must specifically be considered in determining whether the alien should be released.IVFor these reasons, the meaning of the relevant statutory provisions is clear—and clearly contrary to the decision of the Court of Appeals. But the dissent is undeterred. It begins by ignoring the statutory language for as long as possible, devoting the first two-thirds of its opinion to a disquisition on the Constitution. Only after a 19-page prologue does the dissent acknowledge the relevant statutory provisions.The dissent frames the question of interpretation as follows: Can §§1225(b), 1226(c), and 1226(a) be read to require bond hearings every six months “without doing violence to the statutory language,” post, at 20 (opinion of Breyer, J.)? According to the dissent, the answer is “yes,” but the dissent evidently has a strong stomach when it comes to inflicting linguistic trauma. Thus, when Congress mandated that an “alien shall be detained,” §1225(b)(1)(B)(ii), what Congress really meant, the dissent insists, is that the alien may be released from custody provided only that his freedom of movement is restricted in some way, such as by “the imposition of a curfew,” post, at 21. And when Congress stressed that “[t]he Attorney General may release an alien . . . only if . . . release . . . from custody is necessary” to protect the safety of a witness, §1226(c)(2) (emphasis added), what Congress meant, the dissent tells us, is that the Attorney General must release an alien even when no witness is in need of protection—so long as the alien is neither a flight risk nor a danger to the community, see post, at 25–27. The contortions needed to reach these remarkable conclusions are a sight to behold.Let us start with the simple term “detain.” According to the dissent, “detain” means the absence of “unrestrained freedom.” Post, at 21. An alien who is subject to any one of “numerous restraints”—including “a requirement to obtain medical treatment,” “to report at regular intervals,” or even simply to comply with “a curfew”—is “detained” in the dissent’s eyes, even if that alien is otherwise free to roam the streets. Ibid.This interpretation defies ordinary English usage. The dictionary cited by the dissent, the Oxford English Dictionary (OED), defines “detain” as follows: “[t]o keep in confinement or under restraint; to keep prisoner.” 4 OED 543 (2d ed. 1989) (emphasis added); see also OED (3d ed. 2012), http://www.oed.com/view/Entry/51176 (same). Other general-purpose dictionaries provide similar definitions. See, e.g., Webster’s Third New International Dictionary 616 (1961) (“to hold or keep in or as if in custody ”); Webster’s New International Dictionary 710 (2d ed. 1934) (“[t]o hold or keep as in custody”); American Heritage Dictionary 508 (def. 2) (3d ed. 1992) (“To keep in custody or temporary confinement”); Webster’s New World College Dictionary 375 (3d ed. 1997) (“to keep in custody; confine”). And legal dictionaries define “detain” the same way. See, e.g., Ballentine’s Law Dictionary 343 (3d ed. 1969) (“To hold; to keep in custody; to keep”); Black’s Law Dictionary 459 (7th ed. 1999) (“The act or fact of holding a person in custody; confinement or compulsory delay”).How does the dissent attempt to evade the clear meaning of “detain”? It resorts to the legal equivalent of a sleight-of-hand trick. First, the dissent cites a passage in Blackstone stating that arrestees could always seek release on bail. Post, at 8–9. Then, having established the obvious point that a person who is initially detained may later be released from detention, the dissent reasons that this means that a person may still be regarded as detained even after he is released from custody. Post, at 21. That, of course, is a nonsequitur. Just because a person who is initially detained may later be released, it does not follow that the person is still “detained” after his period of detention comes to an end.If there were any doubt about the meaning of the term “detain” in the relevant statutory provisions, the context in which they appear would put that doubt to rest. Title 8 of the United States Code, the title dealing with immigration, is replete with references that distinguish between “detained” aliens and aliens who are free to walk the streets in the way the dissent imagines. Section 1226(a), for instance, distinguishes between the power to “continue to detain the arrested alien” and the power to “release the alien on . . . bond.” But if the dissent were right, that distinction would make no sense: An “alien released on bond” would also be a “detained alien.” Here is another example: In §1226(b), Congress gave the Attorney General the power to “revoke” at any time “a bond or parole authorized under subsection (a) of this section, rearrest the alien under the original warrant, and detain the alien.” It beggars belief that Congress would have given the Attorney General the power to detain a class of aliens who, under the dissent’s reading, are already “detained” because they are free on bond. But that is what the dissent would have us believe. Consider, finally, the example of §1226(c). As noted, that provision obligates the Attorney General to “take into custody” certain aliens whenever they are “released, without regard to whether the alien is released on parole, supervised release, or probation.” On the dissent’s view, however, even aliens “released on parole, supervised release, or probation” are “in custody”—and so there would be no need for the Attorney General to take them into custody again.[7]Struggling to prop up its implausible interpretation, the dissent looks to our prior decisions for aid, but that too fails. The best case it can find is Tod v. Waldman,266 U. S. 547 (1925), a grant of a petition for rehearing in which the Court clarified that “[n]othing in [its original] order . . . shall prejudice an application for release on bail of the respondents pending compliance with the mandate of this Court.” Id., at 548. According to the dissent, that two-page decision from almost a century ago supports its reading because the underlying immigration statute in that case—like some of the provisions at issue here—mandated that the relevant class of aliens “ ‘shall be detained’ ” pending the outcome of an inspection process. See post, at 21–22 (quoting Act of Feb. 5, 1917, §16,39Stat.886).That reads far too much into Waldman. To start, the Court did not state that the aliens at issue were entitled to bail or even that bail was available to them. Instead, the Court merely noted that its decision should not “prejudice” any application the aliens might choose to file. That is notable, for in their petition for rehearing the aliens had asked the Court to affirmatively “authorize [them] to give bail.” Petition for Rehearing in Tod v. Waldman, O. T. 1924, No. 95, p. 17 (emphasis added). By refusing to do so, the Court may have been signaling its skepticism about their request. But it is impossible to tell. That is precisely why we, unlike the dissent, choose not to go beyond what the sentence actually says. And Waldman says nothing about how the word “detain” should be read in the context of §§1225(b), 1226(c), and 1226(a).[8]Neither does Zadvydas. It is true, as the dissent points out, that Zadvydas found “that the words ‘ “may be detained” ’ [are] consistent with requiring release from long-term detention,” post, at 23 (quoting 533 U. S., at 682), but that is not because there is any ambiguity in the term “detain.” As we have explained, the key statutory provision in Zadvydas said that the aliens in question “may,” not “shall,” be detained, and that provision also failed to specify how long detention was to last. Here, the statutory provisions at issue state either that the covered aliens “shall” be detained until specified events take place, see8 U. S. C. §1225(b)(1)(B)(ii) (“further consideration of the application for asylum”); §1225(b)(2)(A) (“a [removal] proceeding”), or provide that the covered aliens may be released “only if” specified conditions are met, §1226(c)(2). The term that the Zadvydas Court found to be ambiguous was “may,” not “detain.” See 533 U. S., at 697. And the opinion in that case consistently used the words “detain” and “custody” to refer exclusively to physical confinement and restraint. See id., at 690 (referring to “[f]reedom from imprisonment—from government custody, detention, or other forms of physical restraint” (emphasis added)); id., at 683 (contrasting aliens “released on bond” with those “held in custody”).[9]The dissent offers no plausible interpretation of §§1225(b), 1226(c), and 1226(a). But even if we were to accept the dissent’s interpretation and hold that “detained” aliens in the “custody” of the Government include aliens released on bond, that would still not justify the dissent’s proposed resolution of this case. The Court of Appeals held that aliens detained under the provisions at issue must be given periodic bond hearings, and the dissent agrees. See post, at 2 (“I would interpret the statute as requiring bail hearings, presumptively after six months of confinement”). But the dissent draws that 6-month limitation out of thin air. However broad its interpretation of the words “detain” and “custody,” nothing in any of the relevant provisions imposes a 6-month time limit on detention without the possibility of bail. So if the dissent’s interpretation is right, then aliens detained under §§1225(b), 1226(c), and 1226(a) are entitled to bail hearings as soon as their detention begins rather than six months later. “Detained” does not mean “released on bond,” and it certainly does not mean “released on bond but only after six months of mandatory physical confinement.”The dissent’s utterly implausible interpretation of the statutory language cannot support the decision of the court below.VBecause the Court of Appeals erroneously concluded that periodic bond hearings are required under the immigration provisions at issue here, it had no occasion to consider respondents’ constitutional arguments on their merits. Consistent with our role as “a court of review, not of first view,” Cutter v. Wilkinson,544 U. S. 709, n. 7 (2005), we do not reach those arguments. Instead, we remand the case to the Court of Appeals to consider them in the first instance.Before the Court of Appeals addresses those claims, however, it should reexamine whether respondents can continue litigating their claims as a class. When the District Court certified the class under Rule 23(b)(2) of the Federal Rules of Civil Procedure, it had their statutory challenge primarily in mind. Now that we have resolved that challenge, however, new questions emerge.Specifically, the Court of Appeals should first decide whether it continues to have jurisdiction despite8 U. S. C. §1252(f )(1). Under that provision, “no court (other than the Supreme Court) shall have jurisdiction or authority to enjoin or restrain the operation of [§§1221–1232] other than with respect to the application of such provisions to an individual alien against whom proceedings under such part have been initiated.” Section 1252(f )(1) thus “prohibits federal courts from granting classwide injunctive relief against the operation of §§1221–123[2].” American-Arab Anti-Discrimination Comm., 525 U. S., at 481. The Court of Appeals held that this provision did not affect its jurisdiction over respondents’ statutory claims because those claims did not “seek to enjoin the operation of the immigration detention statutes, but to enjoin conduct . . . not authorized by the statutes.” 591 F. 3d, at 1120. This reasoning does not seem to apply to an order granting relief on constitutional grounds, and therefore the Court of Appeals should consider on remand whether it may issue classwide injunctive relief based on respondents’ constitutional claims. If not, and if the Court of Appeals concludes that it may issue only declaratory relief, then the Court of Appeals should decide whether that remedy can sustain the class on its own. See, e. g., Rule 23(b)(2) (requiring “that final injunctive relief or corresponding declaratory relief [be] appropriate respecting the class as a whole” (emphasis added)).The Court of Appeals should also consider whether a Rule 23(b)(2) class action continues to be the appropriate vehicle for respondents’ claims in light of Wal-Mart Stores, Inc. v. Dukes,564 U. S. 338 (2011). We held in Dukes that “Rule 23(b)(2) applies only when a single injunction or declaratory judgment would provide relief to each member of the class.” Id., at 360. That holding may be relevant on remand because the Court of Appeals has already acknowledged that some members of the certified class may not be entitled to bond hearings as a constitutional matter. See, e. g., 804 F. 3d, at 1082; 715 F. 3d, at 1139–1141 (citing, e. g., Shaughnessy v. United States ex rel. Mezei,345 U. S. 206 (1953)). Assuming that is correct, then it may no longer be true that the complained-of “ ‘conduct is such that it can be enjoined or declared unlawful only as to all of the class members or as to none of them.’ ” Dukes, supra, at 360 (quoting Nagareda, Class Certification in the Age of Aggregate Proof, 84 N. Y. U. L. Rev. 97, 132 (2009)).Similarly, the Court of Appeals should also consider on remand whether a Rule 23(b)(2) class action litigated on common facts is an appropriate way to resolve respondents’ Due Process Clause claims. “[D]ue process is flexible,” we have stressed repeatedly, and it “calls for such procedural protections as the particular situation demands.” Morrissey v. Brewer,408 U. S. 471,481 (1972); see also Landon v. Plasencia,459 U. S. 21,34 (1982).VIWe reverse the judgment of the United States Court of Appeals for the Ninth Circuit and remand the case for further proceedings.It is so ordered.Justice Kagan took no part in the decision of this case. Notes 1 * Justice Sotomayor joins only Part III–C of this opinion. 2 Anyone who believes that he is not covered by §1226(c) may also ask for what is known as a “Joseph hearing.” See Matter of Joseph, 22 I. & N. Dec. 799 (BIA 1999). At a Joseph hearing, that person “may avoid mandatory detention by demonstrating that he is not an alien, was not convicted of the predicate crime, or that the [Government] is otherwise substantially unlikely to establish that he is in fact subject to mandatory detention.” Demore v. Kim,538 U. S. 510, n. 3 (2003). Whether respondents are entitled to Joseph hearings is not before this Court. 3 It is questionable whether this is true for aliens who are detained under8 U. S. C. §1225(b)(1)(B)(ii) for consideration of their asylum applications. 4 The concurrence contends that “detention is an ‘action taken . . . to remove’ an alien” and that therefore “even the narrowest reading of ‘arising from’ must cover” the claims raised by respondents. Post, at 6. We do not follow this logic. We will assume for the sake of argument that detention is an action taken “to remove an alien,” i.e., for the purpose of removing an alien, rather than simply an action aimed at ensuring that the alien does not flee or commit a crime while his proceedings are pending. But even if we proceed on the basis of that assumption, we do not see what it proves. The question is not whether detention is an action taken to remove an alien but whether the legal questions in this case arise from such an action. And for the reasons explained above, those legal questions are too remote from the actions taken to fall within the scope of §1252(b)(9). 5 According to the dissent, we could have applied the expressio unius canon in Zadvydas as well because there was also an “alternative avenue for relief, namely, bail,” available for aliens detained under §1231(a)(6). Post, at 25 (opinion of Breyer, J.). But the dissent overlooks the fact that the provision granting bail was precisely the same provision that the Court purported to be interpreting, so the canon was not applicable. See 533 U. S., at 683. 6 See Uniting and Strengthening America by Providing Appropriate Tools Required to Intercept and Obstruct Terrorism Act of 2001 (PATRIOT Act),115Stat.272. 7 As the dissent notes, §1158(d)(2) regulates employment authorization for certain “applicant[s] for asylum.” Were all asylum applicants detained, the dissent says, that provision would make no sense, because detained aliens do not need work authorizations. Post, at 23–24. But §1158(d)(2) applies not only to aliens seeking asylum status “in accordance with . . . section 1225(b)” (and thus aliens who are detained), but also to all aliens already “physically present in the United States.” §1158(a)(1). Many of those aliens will be in the country lawfully, and thus they will not be detained and will be able to work pending the outcome of their asylum application. For example, an alien may apply for asylum after being admitted into the country on a short-term visa. While the application is pending, §1158 may offer a way for that alien to find employment. In response, the dissent accuses us of “apply[ing] this provision to some asylum applicants but not the ones before us.” Post, at 23–24. That is not remotely what we are doing. We do not doubt that §1158(d)(2) “applies” to all “applicant[s] for asylum” as it says, even if some of those applicants are not as likely to receive an employment authorization (for instance, because they are detained) as others. 8 It should not be surprising by this point that even the aliens in Waldman understood “detention” in contradistinction to “bail.” See Petition for Rehearing in Tod v. Waldman, O. T. 1924, No. 95, pp. 17–18 (“[T]he Court’s mandate should authorize relators to give bail, instead of having [them] go to Ellis Island and remain there in custody pending an appeal . . . which may involve very long detention pending hearing of the appeal . . .” (capitalization omitted and emphasis added)). 9 The dissent argues that because “the question at issue [in Zadvydas] was release from detention,” “the key word was consequently ‘may.’ ” Post, at 23. We agree but fail to see the point. If, as the dissent admits, Zadvydas was about “release from detention” and not about what qualifies as “detention,” then it is unclear why the dissent thinks that decision supports its unorthodox interpretation of the word “detention.”

584.US.2017_16-499

Petitioners filed suits under the Alien Tort Statute (ATS), alleging that they, or the persons on whose behalf they assert claims, were injured or killed by terrorist acts committed abroad, and that those acts were in part caused or facilitated by respondent Arab Bank, PLC, a Jordanian financial institution with a branch in New York. They seek to impose liability on the bank for the conduct of its human agents, including high-ranking bank officials. They claim that the bank used its New York branch to clear dollar-denominated transactions that benefited terrorists through the Clearing House Interbank Payments System (CHIPS) and to launder money for a Texas-based charity allegedly affiliated with Hamas. While the litigation was pending, this Court held, in Kiobel v. Royal Dutch Petroleum Co., 569 U. S. 108 , that the ATS does not extend to suits against foreign corporations when “all the relevant conduct took place outside the United States,” id., at 124, but it left unresolved the Second Circuit’s broader holding in its Kiobel decision: that foreign corporations may not be sued under the ATS. Deeming that broader holding binding precedent, the District Court dismissed petitioners’ ATS claims and the Second Circuit affirmed. Held: The judgment is affirmed. 808 F. 3d 144, affirmed. Justice Kennedy delivered the opinion of the Court with respect to Parts I, II–B–1, and II–C, concluding that foreign corporations may not be defendants in suits brought under the ATS. Pp. 6–11, 18–19, and 25–27. (a) The Judiciary Act of 1789 included what is now known as the ATS, which provides: “The district courts shall have original jurisdiction of any civil action by an alien for a tort only, committed in violation of the law of nations or a treaty of the United States.” 28 U. S. C. §1350. The ATS is “strictly jurisdictional” and does not by its own terms provide or delineate the definition of a cause of action for international-law violations. Sosa v. Alvarez-Machain, 542 U. S. 692, 713 –714. It was enacted against the backdrop of the general common law, which in 1789 recognized a limited category of “torts in violation of the law of nations,” id., at 714; and one of its principal objectives was to avoid foreign entanglements by ensuring the availability of a federal forum where the failure to have one might cause another nation to hold the United States responsible for an injury to a foreign citizen, see id., at 715–719. The ATS was invoked but a few times over its first 190 years, but with the evolving recognition—e.g., in the Nuremberg trials—that certain crimes against humanity violate basic precepts of international law, courts began to give some redress for violations of clear and unambiguous international human-rights protections. After the Second Circuit first permitted plaintiffs to bring ATS actions based on modern human-rights laws, Congress enacted the Torture Victim Protection Act of 1991 (TVPA), creating an express cause of action for victims of torture and extrajudicial killing in violation of international law. ATS suits became more frequent; and modern ATS litigation has the potential to involve groups of foreign plaintiffs suing foreign corporations in the United States for alleged human-rights violations in other nations. In Sosa, the Court held that in certain narrow circumstances courts may recognize a common-law cause of action for claims based on the present-day law of nations, 542 U. S., at 732, but it explicitly held that ATS litigation implicates serious separation-of-powers and foreign-relations concerns, id., at 727–728. The Court subsequently held in Kiobel that “the presumption against extraterritoriality applies to [ATS] claims,” 569 U. S., at 124, and that even claims that “touch and concern the territory of the United States . . . must do so with sufficient force to displace” that presumption, id., at 124–125. Pp. 6–11. (b) Sosa is consistent with this Court’s general reluctance to extend judicially created private rights of action. Recent precedents cast doubt on courts’ authority to extend or create private causes of action, even in the realm of domestic law, rather than leaving such decisions to the Legislature, which is better positioned “to consider if the public interest would be served by imposing a new substantive legal liability,” Ziglar v. Abbasi, 582 U. S. ___, ___ (internal quotation marks omitted). This caution extends to the question whether the courts should exercise the judicial authority to mandate a rule imposing liability upon artificial entities like corporations. Thus, in Correctional Services Corp. v. Malesko, 534 U. S. 61, 72 , the Court concluded that Congress, not the courts, should decide whether corporate defendants could be held liable in actions under Bivens v. Six Unknown Fed. Narcotics Agents, 403 U. S. 388 . Neither the language of the ATS nor precedent supports an exception to these general principles in this context. Separation-of-powers concerns that counsel against courts creating private rights of action apply with particular force in the context of the ATS, which implicates foreign-policy concerns that are the province of the political branches. And courts must exercise “great caution” before recognizing new forms of liability under the ATS. Sosa, supra, at 728. The question whether a proper application of Sosa would preclude courts from ever recognizing new ATS causes of action need not be decided here, for either way it would be inappropriate for courts to extend ATS liability to foreign corporations absent further action from Congress. Pp. 18–19. (c) The ATS was intended to promote harmony in international relations by ensuring foreign plaintiffs a remedy for international-law violations when the absence of such a remedy might provoke foreign nations to hold the United States accountable. But here, and in similar cases, the opposite is occurring. Petitioners are foreign nationals seeking millions of dollars in damages from a major Jordanian financial institution for injuries suffered in attacks by foreign terrorists in the Middle East. The only alleged connections to the United States are the CHIPS transactions in Arab Bank’s New York branch and a brief allegation about a charity in Texas. At a minimum, the relatively minor connection between the terrorist attacks and the alleged conduct in the United States illustrates the perils of extending the scope of ATS liability to foreign multinational corporations like Arab Bank. For 13 years, this litigation has caused considerable diplomatic tensions with Jordan, a critical ally that considers the litigation an affront to its sovereignty. And this is not the first time that a foreign sovereign has raised objections to ATS litigation in this Court. See Sosa, supra, at 733, n. 21. These are the very foreign-relations tensions the First Congress sought to avoid. Nor are the courts well suited to make the required policy judgments implicated by foreign corporate liability. Like the presumption against extraterritoriality, judicial caution under Sosa “guards against our courts triggering . . . serious foreign policy consequences, and instead defers such decisions, quite appropriately, to the political branches.” Kiobel, supra, at 124. Accordingly, the Court holds that foreign corporations may not be defendants in suits brought under the ATS. Pp. 25–27. Justice Kennedy, joined by The Chief Justice and Justice Thomas, concluded in Parts II–A, II–B–2, II–B–3, and III: (a) Before recognizing an ATS common-law action, federal courts must apply the two-part test announced in Sosa. The threshold question is whether a plaintiff can demonstrate that the alleged violation is “ ‘of a norm that is specific, universal, and obligatory.’ ” 542 U. S., at 732. Assuming that such a norm can control, it must be determined whether allowing the case to proceed under the ATS is a proper exercise of judicial discretion or whether caution requires the political branches to grant specific authority before corporate liability can be imposed. Id., at 732–733, and nn. 20–21. With regard to the first Sosa question, the Court need not resolve whether corporate liability is a question governed by international law or whether that law imposes liability on corporations, because, as shown by the parties’ opposing arguments, there is at least sufficient doubt on the point to turn to Sosa’s second question: whether the Judiciary must defer to Congress to determine in the first instance whether that universal norm has been recognized and, if so, whether it should be enforced in ATS suits. Pp. 11–18. (b) Especially here, in the realm of international law, it is important to look to analogous statutes for guidance on the appropriate boundaries of judge-made causes of action. The logical statutory analogy for an ATS common-law action is the TVPA—the only ATS cause of action created by Congress rather than the courts. Drafted as “an unambiguous and modern basis for [an ATS] cause of action,” H. R. Rep. No. 102–367, p. 3, the TVPA reflects Congress’ considered judgment of the proper structure for such an action. Absent a compelling justification, courts should not deviate from that model. Relevant here, the TVPA limits liability to “individuals,” a term which unambiguously limits liability to natural persons, Mohamad v. Palestinian Authority, 566 U. S. 449, 453 –456. Congress’ decision to exclude liability for corporations in TVPA actions is all but dispositive in this case. Pp. 19–23. (c) Other considerations relevant to the exercise of judicial discretion also counsel against allowing liability under the ATS for foreign corporations, absent congressional instructions. Corporate liability under the ATS has not been shown to be essential to serving that statute’s goals, the ATS will seldom be the only way for plaintiffs to hold the perpetrators liable, and plaintiffs still can sue the individual corporate employees responsible for a violation of international law under the ATS. That the corporate form can be an instrument for inflicting grave harm and suffering poses serious and complex questions for the international community and for Congress. And this complexity makes it all the more important that Congress determine whether victims of human-rights abuses may sue foreign corporations in federal court. Pp. 23–25. (d) In making its determination, Congress might decide that violations of international law do, or should, impose that liability to ensure that corporations make every effort to deter human-rights violations, and so that compensation for injured persons will be a cost of doing business. Or Congress could conclude that neutral judicial safeguards may not be ensured in every country and that, as a reciprocal matter, ATS liability for foreign corporations should be subject to some limitations or preconditions. Finally, Congress might find that corporate liability should be limited to cases where a corporation’s management was actively complicit in the crime. Pp. 27–29. Justice Alito concluded that the outcome in this case is justified not only by “judicial caution” but also by the separation of powers. Assuming that Sosa v. Alvarez-Machain, 542 U. S. 692 , correctly held that federal courts, exercising their authority in limited circumstances to make federal common law, may create causes of action under the ATS, this Court should not create such causes of action against foreign corporate defendants. The objective for courts in any case requiring the creation of federal common law must be “to find the rule that will best effectuate the federal policy.” Textile Workers v. Lincoln Mills of Ala., 353 U. S. 448, 457 . The First Congress enacted the ATS to help the United States avoid diplomatic friction. Putting that objective together with the rules governing federal common law generally, the following principle emerges: Federal courts should decline to create federal common law causes of action whenever doing so would not materially advance the ATS’s objective of avoiding diplomatic strife. Applying that principle here, it is clear that courts should not create causes of action under the ATS against foreign corporate defendants. Customary international law does not generally require corporate liability, so declining to create it under the ATS cannot give other nations just cause for complaint against the United States. To the contrary, creating causes of action against foreign corporations under the ATS may instead provoke exactly the sort of diplomatic strife inimical to the statute’s fundamental purpose. Pp. 1–7. Justice Gorsuch concluded that there are two more fundamental reasons why this lawsuit should be dismissed. Pp. 1–14. (a) This Court has suggested that Congress originally enacted the ATS to afford federal courts jurisdiction to hear tort claims related to three violations of international law that were already embodied in English common law: violations of safe conducts extended to aliens, interference with ambassadors, and piracy. Sosa v. Alvarez-Machain, 542 U. S. 692, 715 . Here, the plaintiffs seek much more. They want the federal courts to recognize a new cause of action, one that did not exist at the time of the statute’s adoption, one that Congress has never authorized. They find support in a passage suggesting that the ATS may afford federal judges “discretion [to] conside[r] [creating] new cause[s] of action” if they “rest on a norm of international character accepted by the civilized world and defined with a specificity comparable to the features of the [three specified] 18th-century” torts. Id., at 725. This is doubtful, for the people’s elected representatives, not judges, make the laws that govern them. But even accepting Sosa’s framework, a proper application of that framework would preclude courts from recognizing any new causes of action under the ATS. When courts are confronted with a request to fashion a new cause of action, “separation-of-powers principles are or should be central to the analysis.” Ziglar v. Abbasi, 582 U. S. ___, ___. The first and most important question is whether Congress or the courts should decide, and the right answer “most often will be Congress.” Ibid. There is no reason to make a special exception for the ATS, which was designed as “a jurisdictional statute creating no new causes of action.” Sosa, 542 U. S., at 724. The context in which any Sosa discretion would be exercised confirms the wisdom of restraint. The “practical consequences” that might follow a decision to create a new ATS cause of action, see id., at 732–733, would likely involve questions of foreign affairs and national security—matters implicating the expertise and authority not of the Judiciary but of the political branches. Pp. 2–5. (b) Another independent problem is that this suit is by foreigners against a foreigner over the meaning of international norms. The original understanding of the ATS, which was but one clause in one section of the Judiciary Act of 1789, likely would have required a domestic defendant in order to comply with the requirements of the Diversity-of-Citizenship Clause of Article III. Precedent interpreting a neighboring provision of the Judiciary Act confirms that conclusion. See Mossman v. Higginson, 4 Dall. 12, 14. In any event, separation-of-powers limits on the judicial function and deference to the political branches should lead federal courts to require a domestic defendant before agreeing to exercise any Sosa-generated discretion to entertain an ATS suit. Pp. 5–14. Kennedy, J., announced the judgment of the Court and delivered the opinion of the Court with respect to Parts I, II–B–1, and II–C, in which Roberts, C. J., and Thomas, Alito, and Gorsuch, JJ., joined, and an opinion with respect to Parts II–A, II–B–2, II–B–3, and III, in which Roberts, C. J., and Thomas, J., joined. Thomas, J., filed a concurring opinion. Alito, J., and Gorsuch, J., filed opinions concurring in part and concurring in the judgment. Sotomayor, J., filed a dissenting opinion, in which Ginsburg, Breyer, and Kagan, JJ., joined.

of the Court with respect to Parts I, II–B–1, and II–C, and an opinion with respect to Parts II–A, II–B–2, II–B–3, and III, in which The Chief Justice and Justice Thomas join. Petitioners in this case, or the persons on whose behalf petitioners now assert claims, allegedly were injured or killed by terrorist acts committed abroad. Those terrorist acts, it is contended, were in part caused or facilitated by a foreign corporation. Petitioners now seek to impose liability on the foreign corporation for the conduct of its human agents, including its then-chairman and other high-ranking management officials. The suits were filed in a United States District Court under the Alien Tort Statute, commonly referred to as the ATS. See 28 U. S. C. §1350. The foreign corporation charged with liability in these ATS suits is Arab Bank, PLC; and it is respondent here. Some of Arab Bank’s officials, it is alleged, allowed the Bank to be used to transfer funds to terrorist groups in the Middle East, which in turn enabled or facilitated criminal acts of terrorism, causing the deaths or injuries for which petitioners now seek compensation. Petitioners seek to prove Arab Bank helped the terrorists receive the moneys in part by means of currency clearances and bank transactions passing through its New York City offices, all by means of electronic transfers. It is assumed here that those individuals who inflicted death or injury by terrorism committed crimes in violation of well-settled, fundamental precepts of international law, precepts essential for basic human-rights protections. It is assumed as well that individuals who knowingly and purposefully facilitated banking transactions to aid, en- able, or facilitate the terrorist acts would themselves be committing crimes under the same international-law prohibitions. Petitioners contend that international and domestic laws impose responsibility and liability on a corporation if its human agents use the corporation to commit crimes in violation of international laws that protect human rights. The question here is whether the Judiciary has the authority, in an ATS action, to make that determination and then to enforce that liability in ATS suits, all without any explicit authorization from Congress to do so. The answer turns upon the proper interpretation and implementation of the ATS. The statute provides: “The district courts shall have original jurisdiction of any civil action by an alien for a tort only, committed in violation of the law of nations or a treaty of the United States.” §1350. The Court must first ask whether the law of nations imposes liability on corporations for human-rights violations committed by its employees. The Court must also ask whether it has authority and discretion in an ATS suit to impose liability on a corporation without a specific direction from Congress to do so. I A Petitioners are plaintiffs in five ATS lawsuits filed against Arab Bank in the United States District Court for the Eastern District of New York. The suits were filed between 2004 and 2010. A significant majority of the plaintiffs in these lawsuits—about 6,000 of them—are foreign nationals whose claims arise under the ATS. These foreign nationals are petitioners here. They allege that they or their family members were injured by terrorist attacks in the Middle East over a 10-year period. Two of the five lawsuits also included claims brought by American nationals under the Anti-Terrorism Act, 18 U. S. C. §2333(a), but those claims are not at issue. Arab Bank is a major Jordanian financial institution with branches throughout the world, including in New York. According to the Kingdom of Jordan, Arab Bank “accounts for between one-fifth and one-third of the total market capitalization of the Amman Stock Exchange.” Brief for Hashemite Kingdom of Jordan as Amicus Curiae 2. Petitioners allege that Arab Bank helped finance attacks by Hamas and other terrorist groups. Among other claims, petitioners allege that Arab Bank maintained bank accounts for terrorists and their front groups and allowed the accounts to be used to pay the families of suicide bombers. Most of petitioners’ allegations involve conduct that occurred in the Middle East. Yet petitioners allege as well that Arab Bank used its New York branch to clear dollar-denominated transactions through the Clearing House Interbank Payments System. That elaborate system is commonly referred to as CHIPS. It is alleged that some of these CHIPS transactions benefited terrorists. Foreign banks often use dollar-clearing transactions to facilitate currency exchanges or to make payments in dollars from one foreign bank account to another. Arab Bank and certain amici point out that CHIPS transactions are enormous both in volume and in dollar amounts. The transactions occur predominantly in the United States but are used by major banks both in the United States and abroad. The CHIPS system is used for dollar-denominated transactions and for transactions where the dollar is used as an intermediate currency to facilitate a currency exchange. Brief for Institute of International Bankers as Amicus Curiae 12–13, and n. 8. In New York each day, on average, about 440,000 of these transfers occur, in dollar amounts totaling about $1.5 trillion. Id., at 14. The “clearance activity is an entirely mechanical function; it occurs without human intervention in the proverbial ‘blink of an eye.’ ” Ibid. There seems to be no dispute that the speed and volume of these transactions are such that individual supervision is simply not a systemic reality. As noted below, substantial regulations govern these transactions, both in the United States and in Jordan. In addition to the dollar-clearing transactions, petitioners allege that Arab Bank’s New York branch was used to launder money for the Holy Land Foundation for Relief and Development (HLF), a Texas-based charity that petitioners say is affiliated with Hamas. According to petitioners, Arab Bank used its New York branch to facilitate the transfer of funds from HLF to the bank accounts of terrorist-affiliated charities in the Middle East. During the pendency of this litigation, there was an unrelated case that also implicated the issue whether the ATS is applicable to suits in this country against foreign corporations. See Kiobel v. Royal Dutch Petroleum Co., 621 F. 3d 111 (CA2 2010). That suit worked its way through the trial court and the Court of Appeals for the Second Circuit. The Kiobel litigation did not involve banking transactions. Its allegations were that holding companies incorporated in the Netherlands and the United Kingdom had, through a Nigerian subsidiary, aided and abetted the Nigerian Government in human-rights abuses. Id., at 123. In Kiobel, the Court of Appeals held that the ATS does not extend to suits against corporations. Id., at 120. This Court granted certiorari in Kiobel. 565 U. S. 961 (2011) . After additional briefing and reargument in Kiobel, this Court held that, given all the circumstances, the suit could not be maintained under the ATS. Kiobel v. Royal Dutch Petroleum Co., 569 U. S. 108, 114, 124 –125 (2013). The rationale of the holding, however, was not that the ATS does not extend to suits against foreign corporations. That question was left unresolved. The Court ruled, instead, that “all the relevant conduct took place outside the United States.” Id., at 124. Dismissal of the action was required based on the presumption against extraterritorial application of statutes. So while this Court in Kiobel affirmed the ruling that the action there could not be maintained, it did not address the broader holding of the Court of Appeals that dismissal was required because corporations may not be sued under the ATS. Still, the courts of the Second Circuit deemed that broader holding to be binding precedent. As a consequence, in the instant case the District Court dismissed petitioners’ ATS claims based on the earlier Kiobel holding in the Court of Appeals; and on review of the dismissal order the Court of Appeals, also adhering to its earlier holding, affirmed. In re Arab Bank, PLC Alien Tort Statute Litigation, 808 F. 3d 144 (2015). This Court granted certiorari in the instant case. 581 U. S. ___ (2017). Since the Court of Appeals relied on its Kiobel holding in the instant case, it is instructive to begin with an analysis of that decision. The majority opinion in Kiobel, written by Judge Cabranes, held that the ATS does not apply to alleged international-law violations by a corporation. 621 F. 3d, at 120. Judge Cabranes relied in large part on the fact that international criminal tribunals have consistently limited their jurisdiction to natural persons. Id., at 132–137. Judge Leval filed a separate opinion. He concurred in the judgment on other grounds but disagreed with the proposition that the foreign corporation was not subject to suit under the ATS. Id., at 196. Judge Leval conceded that “international law, of its own force, imposes no liabilities on corporations or other private juridical entities.” Id., at 186. But he reasoned that corporate liability for violations of international law is an issue of “civil compensatory liability” that international law leaves to individual nations. Ibid. Later decisions in the Courts of Appeals for the Seventh, Ninth, and District of Columbia Circuits agreed with Judge Leval and held that corporations can be subject to suit under the ATS. See Flomo v. Firestone Nat. Rubber Co., 643 F. 3d 1013, 1017–1021 (CA7 2011); Doe I v. Nestle USA, Inc., 766 F. 3d 1013, 1020–1022 (CA9 2014); Doe VIII v. Exxon Mobil Corp., 654 F. 3d 11, 40–55 (CADC 2011), vacated on other grounds, 527 Fed. Appx. 7 (CADC 2013). The respective opinions by Judges Cabranes and Leval are scholarly and extensive, providing significant guidance for this Court in the case now before it. With this background, it is now proper to turn to the history of the ATS and the decisions interpreting it. B Under the Articles of Confederation, the Continental Congress lacked authority to “ ‘cause infractions of treaties, or of the law of nations to be punished.’ ” Sosa v. Alvarez-Machain, 542 U. S. 692, 716 (2004) (quoting J. Madison, Journal of the Constitutional Convention 60 (E. Scott ed. 1893)). The Continental Congress urged the States to authorize suits for damages sustained by foreign citizens as a result of violations of international law; but the state courts’ vindication of the law of nations remained unsatisfactory. Concerns with the consequent international-relations tensions “persisted through the time of the Constitutional Convention.” 542 U. S., at 717. Under the Articles of Confederation, the inability of the central government to ensure adequate remedies for foreign citizens caused substantial foreign-relations problems. In 1784, the French Minister lodged a protest with the Continental Congress after a French adventurer, the Chevalier de Longchamps, assaulted the Secretary of the French Legation in Philadelphia. See Kiobel, 569 U. S., at 120. A few years later, a New York constable caused an international incident when he entered the house of the Dutch Ambassador and arrested one of his servants. Ibid. Under the Articles of Confederation, there was no national forum available to resolve disputes like these under any binding laws that were or could be enacted or enforced by a central government. The Framers addressed these matters at the 1787 Philadelphia Convention; and, as a result, Article III of the Constitution extends the federal judicial power to “all cases affecting ambassadors, other public ministers and consuls,” and “to controversies . . . between a state, or the citizens thereof, and foreign states, citizens, or subjects.” §2. The First Congress passed a statute to implement these provisions: The Judiciary Act of 1789 authorized federal jurisdiction over suits involving disputes between aliens and United States citizens and suits involving diplomats. §§9, 11, 1Stat. 76–79. The Judiciary Act also included what is now the statute known as the ATS. §9, id., at 76. As noted, the ATS is central to this case and its brief text bears repeating. Its full text is: “The district courts shall have original jurisdiction of any civil action by an alien for a tort only, committed in violation of the law of nations or a treaty of the United States.” 28 U. S. C. §1350. The ATS is “strictly jurisdictional” and does not by its own terms provide or delineate the definition of a cause of action for violations of international law. Sosa, 542 U. S., at 713–714. But the statute was not enacted to sit on a shelf awaiting further legislation. Id., at 714. Rather, Congress enacted it against the backdrop of the general common law, which in 1789 recognized a limited category of “torts in violation of the law of nations.” Ibid. In the 18th century, international law primarily governed relationships between and among nation-states, but in a few instances it governed individual conduct occurring outside national borders (for example, “disputes relating to prizes, to shipwrecks, to hostages, and ransom bills”). Id., at 714–715 (internal quotation marks omitted). There was, furthermore, a narrow domain in which “rules binding individuals for the benefit of other individuals overlapped with” the rules governing the relationships between nation-states. Id., at 715. As understood by Blackstone, this domain included “three specific offenses against the law of nations addressed by the criminal law of England: violation of safe conducts, infringement of the rights of ambassadors, and piracy.” Ibid. (citing 4 W. Blackstone, Commentaries on the Laws of England 68 (1769)). “It was this narrow set of violations of the law of nations, admitting of a judicial remedy and at the same time threatening serious consequences in international affairs, that was probably on the minds of the men who drafted the ATS.” 542 U. S., at 715. This history teaches that Congress drafted the ATS “to furnish jurisdiction for a relatively modest set of actions alleging violations of the law of nations.” Id., at 720. The principal objective of the statute, when first enacted, was to avoid foreign entanglements by ensuring the availability of a federal forum where the failure to provide one might cause another nation to hold the United States responsible for an injury to a foreign citizen. See id., at 715–719; Kiobel, 569 U. S., at 123–124. Over the first 190 years or so after its enactment, the ATS was invoked but a few times. Yet with the evolving recognition—for instance, in the Nuremberg trials after World War II—that certain acts constituting crimes against humanity are in violation of basic precepts of international law, courts began to give some redress for violations of international human-rights protections that are clear and unambiguous. In the modern era this began with the decision of the Court of Appeals for the Second Circuit in Filartiga v. Pena-Irala, 630 F. 2d 876 (1980). In Filartiga, it was alleged that a young man had been tortured and murdered by Peruvian police officers, and that an officer named Pena-Irala was one of the supervisors and perpetrators. Some members of the victim’s family were in the United States on visas. When they discovered that Pena-Irala himself was living in New York, they filed suit against him. The action, seeking damages for the suffering and death he allegedly had caused, was filed in the United States District Court for the Eastern District of New York. The Court of Appeals found that there was jurisdiction under the ATS. For this holding it relied upon the universal acknowledgment that acts of official torture are contrary to the law of nations. Id., at 890. This Court did not review that decision. In the midst of debates in the courts of appeals over whether the court in Filartiga was correct in holding that plaintiffs could bring ATS actions based on modern human- rights laws absent an express cause of action created by an additional statute, Congress enacted the Torture Victim Protection Act of 1991 (TVPA), 106Stat. 73, note following 28 U. S. C. §1350. H. R. Rep. No. 102–367, pp. 3–4 (1991) (H. R. Rep.) (citing Tel-Oren v. Libyan Arab Republic, 726 F. 2d 774 (CADC 1984)); S. Rep. No. 102–249, pp. 3–5 (1991) (S. Rep.) (same). The TVPA—which is codified as a note following the ATS—creates an express cause of action for victims of torture and extrajudicial killing in violation of international law. After Filartiga and the TVPA, ATS lawsuits became more frequent. Modern ATS litigation has the potential to involve large groups of foreign plaintiffs suing foreign corporations in the United States for alleged human-rights violations in other nations. For example, in Kiobel the plaintiffs were Nigerian nationals who sued Dutch, British, and Nigerian corporations for alleged crimes in Nigeria. 569 U. S., at 111–112. The extent and scope of this litigation in United States courts have resulted in criticism here and abroad. See id., at 124 (noting objections to ATS litigation by Canada, Germany, Indonesia, Papua New Guinea, South Africa, Switzerland, and the United Kingdom). In Sosa, the Court considered the question whether courts may recognize new, enforceable international norms in ATS lawsuits. 542 U. S., at 730–731. The Sosa Court acknowledged the decisions made in Filartiga and similar cases; and it held that in certain narrow circumstances courts may recognize a common-law cause of action for claims based on the present-day law of nations, in addition to the “historical paradigms familiar when §1350 was enacted.” 542 U. S., at 732. The Court was quite explicit, however, in holding that ATS litigation implicates serious separation-of-powers and foreign-relations concerns. Id., at 727–728. Thus, ATS claims must be “subject to vigilant doorkeeping.” Id., at 729. This Court next addressed the ATS in Kiobel, the case already noted. There, this Court held that “the presumption against extraterritoriality applies to claims under the ATS, and that nothing in the statute rebuts the presumption.” 569 U. S., at 124. The Court added that “even where the claims touch and concern the territory of the United States, they must do so with sufficient force to displace the presumption against extraterritorial application.” Id., at 124–125. II With these principles in mind, this Court now must decide whether common-law liability under the ATS extends to a foreign corporate defendant. It could be argued, under the Court’s holding in Kiobel, that even if, under accepted principles of international law and federal common law, corporations are subject to ATS liability for human-rights crimes committed by their human agents, in this case the activities of the defendant corporation and the alleged actions of its employees have insufficient connections to the United States to subject it to jurisdiction under the ATS. Various amici urge this as a rationale to affirm here, while the Government argues that the Court should remand this case so the Court of Appeals can address the issue in the first instance. There are substantial arguments on both sides of that question; but it is not the question on which this Court granted certiorari, nor is it the question that has divided the Courts of Appeals. The question whether foreign corporations are subject to liability under the ATS should be addressed; for, if there is no liability for Arab Bank, the lengthy and costly litigation concerning whether corporate contacts like those alleged here suffice to impose liability would be pointless. In addition, a remand to the Court of Appeals would require prolonging litigation that already has caused significant diplomatic tensions with Jordan for more than a decade. So it is proper for this Court to decide whether corporations, or at least foreign corporations, are subject to liability in an ATS suit filed in a United States district court. Before recognizing a common-law action under the ATS, federal courts must apply the test announced in Sosa. An initial, threshold question is whether a plaintiff can demonstrate that the alleged violation is “of a norm that is specific, universal, and obligatory.” 542 U. S., at 732 (internal quotation marks omitted). And even assuming that, under international law, there is a specific norm that can be controlling, it must be determined further whether allowing this case to proceed under the ATS is a proper exercise of judicial discretion, or instead whether caution requires the political branches to grant specific authority before corporate liability can be imposed. See id., at 732–733, and nn. 20–21. “[T]he potential implications for the foreign relations of the United States of recognizing such causes should make courts particularly wary of impinging on the discretion of the Legislative and Executive Branches in managing foreign affairs.” Id., at 727. It must be said that some of the considerations that pertain to determining whether there is a specific, universal, and obligatory norm that is established under international law are applicable as well in determining whether deference must be given to the political branches. For instance, the fact that the charters of some international tribunals and the provisions of some congressional statutes addressing international human-rights violations are specifically limited to individual wrongdoers, and thus foreclose corporate liability, has significant bearing both on the content of the norm being asserted and the question whether courts should defer to Congress. The two inquiries inform each other and are, to that extent, not altogether discrete. With that introduction, it is proper now to turn first to the question whether there is an international-law norm imposing liability on corporations for acts of their employees that contravene fundamental human rights. A Petitioners and Arab Bank disagree as to whether corporate liability is a question of international law or only a question of judicial authority and discretion under domes- tic law. The dispute centers on a footnote in Sosa. In the course of holding that international norms must be “sufficiently definite to support a cause of action,” the Court in Sosa noted that a “related consideration is whether international law extends the scope of liability for a violation of a given norm to the perpetrator being sued, if the defendant is a private actor such as a corporation or individual.” Id., at 732, and n. 20. In the Court of Appeals’ decision in Kiobel, the majority opinion by Judge Cabranes interpreted footnote 20 to mean that corporate defendants may be held liable under the ATS only if there is a specific, universal, and obligatory norm that corporations are liable for violations of international law. 621 F. 3d, at 127. In Judge Cabranes’ view, “[i]nternational law is not silent on the question of the subjects of international law—that is, those that, to varying extents, have legal status, personality, rights, and duties under international law,” “[n]or does international law leave to individual States the responsibility of defining those subjects.” Id., at 126 (internal quotation marks omitted). There is considerable force and weight to the position articulated by Judge Cabranes. And, assuming the Court of Appeals was correct that under Sosa corporate liability is a question of international law, there is an equally strong argument that petitioners cannot satisfy the high bar of demonstrating a specific, universal, and obligatory norm of liability for corporations. Indeed, Judge Leval agreed with the conclusion that international law does “not provide for any form of liability of corporations.” Kiobel, 621 F. 3d, at 186. 1 In modern times, there is no doubt, of course, that “the international community has come to recognize the common danger posed by the flagrant disregard of basic human rights,” leading “the nations of the world to recognize that respect for fundamental human rights is in their individual and collective interest.” Filartiga, 630 F. 2d, at 890. That principle and commitment support the conclusion that human-rights norms must bind the individual men and women responsible for committing humanity’s most terrible crimes, not just nation-states in their interactions with one another. “The singular achievement of international law since the Second World War has come in the area of human rights,” where international law now imposes duties on individuals as well as nation-states. Kiobel, 621 F. 3d, at 118. It does not follow, however, that current principles of international law extend liability—civil or criminal—for human-rights violations to corporations or other artificial entities. This is confirmed by the fact that the charters of respective international criminal tribunals often exclude corporations from their jurisdictional reach. The Charter for the Nuremberg Tribunal, created by the Allies after World War II, provided that the Tribunal had jurisdiction over natural persons only. See Agreement for Prosecution and Punishment of Major War Criminals of the European Axis, Art. 6, Aug. 8, 1945, 59Stat. 1547, E. A. S. 472. Later, a United States Military Tribunal prosecuted 24 executives of the German corporation IG Farben. 7 Trials of War Criminals Before the Nuernberg Military Tribunals Under Control Council Law No. 10, pp. 11–60 (1952) (The Farben Case). Among other crimes, Farben’s employees had operated a slave-labor camp at Auschwitz and “knowingly and intentionally manufactured and provided” the poison gas used in the Nazi death chambers. Kiobel, 621 F. 3d, at 135. Although the Military Tribunal “used the term ‘Farben’ as descriptive of the instrumentality of cohesion in the name of which” the crimes were committed, the Tribunal noted that “corporations act through individuals.” 8 The Farben Case, at 1153. Farben itself was not held liable. See ibid. The jurisdictional reach of more recent international tribunals also has been limited to “natural persons.” See Statute of the International Criminal Tribunal for the Former Yugoslavia, S. C. Res. 827 (May 25, 1993), adopting U. N. Secretary-General Rep. Pursuant to Paragraph 2 of Security Council Resolution 808, Art. 6, U. N. Doc. S/25704 (May 3, 1993); Statute of the International Tribunal for Rwanda, Art. 5, S. C. Res. 955, Art. 5 (Nov. 8, 1994). The Rome Statute of the International Criminal Court, for example, limits that tribunal’s jurisdiction to “natural persons.” See Rome Statute of the International Criminal Court, Art. 25(1), July 17, 1998, 2187 U. N. T. S. 90. The drafters of the Rome Statute considered, but rejected, a proposal to give the International Criminal Court jurisdiction over corporations. Eser, Individual Criminal Responsibility, in 1 Rome Statute of the International Criminal Court 767, 778–779 (A. Cassese et al. eds. 2002). The international community’s conscious decision to limit the authority of these international tribunals to natural persons counsels against a broad holding that there is a specific, universal, and obligatory norm of corporate liability under currently prevailing international law. 2 In light of the sources just discussed, the sources petitioners rely on to support their contention that liability for corporations is well established as a matter of international law lend weak support to their position. Petitioners first point to the International Convention for the Suppression of the Financing of Terrorism. This Convention imposes an obligation on “Each State Party” “to enable a legal entity located in its territory or organized under its laws to be held liable when a person responsible for the management or control of that legal entity has, in that capacity,” violated the Convention. International Convention for the Suppression of the Financing of Terrorism, Dec. 9, 1999, S. Treaty Doc. No. 106–49, 2178 U. N. T. S. 232. But by its terms the Convention imposes its obligations only on nation-states “to enable” corporations to be held liable in certain circumstances under domestic law. The United States and other nations, including Jordan, may fulfill their obligations under the Convention by adopting detailed regulatory regimes governing financial institutions. See, e.g., 18 U. S. C. §2333(a) (private right of action under the Anti-Terrorism Act); 31 U. S. C. §5311 et seq. (Bank Secrecy Act); 31 CFR pt. 595 (2017) (Terrorism Sanctions Regulations); Brief for Central Bank of Jordan as Amicus Curiae 5 (describing Jordan’s “comprehensive approach to preventing money laundering and terrorist financing”). The Convention neither requires nor authorizes courts, without congressional authorization, to displace those detailed regulatory regimes by allowing common-law actions under the ATS. And nothing in the Convention’s text requires signatories to hold corporations liable in common-law tort actions raising claims under international law. In addition, petitioners and their amici cite a few cases from other nations and the Special Tribunal for Lebanon that, according to petitioners, are examples of corporations being held liable for violations of international law. E.g., Brief for Petitioners 50–51. Yet even assuming that these cases are relevant examples, at most they demonstrate that corporate liability might be permissible under international law in some circumstances. That falls far short of establishing a specific, universal, and obligatory norm of corporate liability. It must be remembered that international law is distinct from domestic law in its domain as well as its objectives. International human-rights norms prohibit acts repugnant to all civilized peoples—crimes like genocide, torture, and slavery, that make their perpetrators “enem[ies] of all mankind.” Sosa, 542 U. S., at 732 (internal quotation marks omitted). In the American legal system, of course, corporations are often subject to liability for the conduct of their human employees, and so it may seem necessary and natural that corporate entities are liable for violations of international law under the ATS. It is true, furthermore, that the enormity of the offenses that can be committed against persons in violation of international human-rights protections can be cited to show that corporations should be subject to liability for the crimes of their human agents. But the international community has not yet taken that step, at least in the specific, universal, and obligatory manner required by Sosa. Indeed, there is precedent to the contrary in the statement during the Nuremberg proceedings that “[c]rimes against international law are committed by men, not by abstract entities, and only by punishing individuals who commit such crimes can the provisions of international law be enforced.” The Nurnberg Trial, 6 F. R. D. 69, 110 (1946). Petitioners also contend that international law leaves questions of remedies open for determination under domestic law. As they see it, corporate liability is a remedial consideration, not a substantive principle that must be supported by a universal and obligatory norm if it is to be implemented under the ATS. According to petitioners, footnote 20 in Sosa does no more than recognize the distinction in international law between state and private actors. But, as just explained, there is a similar distinction in international law between corporations and natural persons. And it is far from obvious why the question whether corporations may be held liable for the international crimes of their employees is a mere question of remedy. In any event, the Court need not resolve the questions whether corporate liability is a question that is governed by international law, or, if so, whether international law imposes liability on corporations. There is at least sufficient doubt on the point to turn to Sosa’s second question—whether the Judiciary must defer to Congress, allowing it to determine in the first instance whether that universal norm has been recognized and, if so, whether it is prudent and necessary to direct its enforcement in suits under the ATS. B 1 Sosa is consistent with this Court’s general reluctance to extend judicially created private rights of action. The Court’s recent precedents cast doubt on the authority of courts to extend or create private causes of action even in the realm of domestic law, where this Court has “recently and repeatedly said that a decision to create a private right of action is one better left to legislative judgment in the great majority of cases.” 542 U. S., at 727 (citing Correctional Services Corp. v. Malesko, 534 U. S. 61, 68 (2001) ; Alexander v. Sandoval, 532 U. S. 275, 286 –287 (2001)). That is because “the Legislature is in the better position to consider if the public interest would be served by imposing a new substantive legal liability.” Ziglar v. Abbasi, 582 U. S. ___, ___ (2017) (slip op., at 12) (internal quotation marks omitted). Thus, “if there are sound reasons to think Congress might doubt the efficacy or necessity of a damages remedy, . . . courts must refrain from creating the remedy in order to respect the role of Congress.” Id., at ___ (slip op., at 13). This caution extends to the question whether the courts should exercise the judicial authority to mandate a rule that imposes liability upon artificial entities like corporations. Thus, in Malesko the Court held that corporate defendants may not be held liable in Bivens actions. See Bivens v. Six Unknown Fed. Narcotics Agents, 403 U. S. 388 (1971) . Allowing corporate liability would have been a “marked extension” of Bivens that was unnecessary to advance its purpose of holding individual officers responsible for “engaging in unconstitutional wrongdoing.” Malesko, 534 U. S., at 74. Whether corporate defendants should be subject to suit was “a question for Congress, not us, to decide.” Id., at 72. Neither the language of the ATS nor the precedents interpreting it support an exception to these general principles in this context. In fact, the separation-of-powers concerns that counsel against courts creating private rights of action apply with particular force in the context of the ATS. See infra, at 25–26. The political branches, not the Judiciary, have the responsibility and institutional capacity to weigh foreign-policy concerns. See Kiobel, 569 U. S., at 116–117. That the ATS implicates foreign relations “is itself a reason for a high bar to new private causes of action for violating international law.” Sosa, supra, at 727. In Sosa, the Court emphasized that federal courts must exercise “great caution” before recognizing new forms of liability under the ATS. 542 U. S., at 728. In light of the foreign-policy and separation-of-powers concerns inherent in ATS litigation, there is an argument that a proper application of Sosa would preclude courts from ever recognizing any new causes of action under the ATS. But the Court need not resolve that question in this case. Either way, absent further action from Congress it would be inappropriate for courts to extend ATS liability to foreign corporations. 2 Even in areas less fraught with foreign-policy consequences, the Court looks to analogous statutes for guidance on the appropriate boundaries of judge-made causes of action. See, e.g., Miles v. Apex Marine Corp., 498 U. S. 19, 24 (1990) ; Blue Chip Stamps v. Manor Drug Stores, 421 U. S. 723, 736 (1975) . Doing so is even more important in the realm of international law, where “the general practice has been to look for legislative guidance before exercising innovative authority over substantive law.” Sosa, supra, at 726. Here, the logical place to look for a statutory analogy to an ATS common-law action is the TVPA—the only cause of action under the ATS created by Congress rather than the courts. As explained above, Congress drafted the TVPA to “establish an unambiguous and modern basis for a cause of action” under the ATS. H. R. Rep., at 3; S. Rep., at 4–5. Congress took care to delineate the TVPA’s boundaries. In doing so, it could weigh the foreign-policy implications of its rule. Among other things, Congress specified who may be liable, created an exhaustion requirement, and established a limitations period. Kiobel, 569 U. S., at 117. In Kiobel, the Court recognized that “[e]ach of these decisions carries with it significant foreign policy implications.” Ibid. The TVPA reflects Congress’ considered judgment of the proper structure for a right of action under the ATS. Absent a compelling justification, courts should not deviate from that model. The key feature of the TVPA for this case is that it limits liability to “individuals,” which, the Court has held, unambiguously limits liability to natural persons. Mohamad v. Palestinian Authority, 566 U. S. 449, 453 –456 (2012). Congress’ decision to exclude liability for corporations in actions brought under the TVPA is all but dispositive of the present case. That decision illustrates that significant foreign-policy implications require the courts to draw a careful balance in defining the scope of actions under the ATS. It would be inconsistent with that balance to create a remedy broader than the one created by Congress. Indeed, it “would be remarkable to take a more aggressive role in exercising a jurisdiction that remained largely in shadow for much of the prior two centuries.” Sosa, supra, at 726. According to petitioners, the TVPA is not a useful guidepost because Congress limited liability under that statute to “individuals” out of concern for the sovereign immunity of foreign governmental entities, not out of general hesitation about corporate liability under the ATS. The argument seems to run as follows: The TVPA provides a right of action to victims of torture and extrajudicial killing, and under international law those human-rights violations require state action. For a corporation’s employees to violate these norms therefore would require the corporation to be an instrumentality of a foreign state or other sovereign entity. That concern is absent, petitioners insist, for crimes that lack a state-action requirement—for example, genocide, slavery, or, in the present case, the financing of terrorists. At least two flaws inhere in this argument. First, in Mohamad the Court unanimously rejected petitioners’ account of the TVPA’s legislative history. 566 U. S., at 453, 458–460. The Court instead read that history to demonstrate that Congress acted to exclude all corporate entities, not just the sovereign ones. Id., at 459–460 (citing Hearing and Markup on H. R. 1417 before the House Committee on Foreign Affairs and Its Subcommittee on Human Rights and International Organizations, 100th Cong., 2d Sess., 87–88 (1988)); see also 566 U. S., at 461–462 (Breyer, J., concurring). Second, even for international-law norms that do not require state action, plaintiffs can still use corporations as surrogate defendants to challenge the conduct of foreign governments. In Kiobel, for example, the plaintiffs sought to hold a corporate defendant liable for “aiding and abetting the Nigerian Government in committing,” among other things, “crimes against humanity.” 569 U. S., at 114; see also, e.g., Sarei v. Rio Tinto, PLC, 671 F. 3d 736, 761–763 (CA9 2011) (en banc) (corporate defendant allegedly used Papua New Guinea’s military to commit genocide), vacated and remanded, 569 U. S. 945 (2013) . Petitioners contend that, instead of the TVPA, the most analogous statute here is the Anti-Terrorism Act. That Act does permit suits against corporate entities. See 18 U. S. C. §§ 2331(3), 2333(d)(2). In fact, in these suits some of the foreign plaintiffs joined their claims to those of United States nationals suing Arab Bank under the Anti-Terrorism Act. But the Anti-Terrorism Act provides a cause of action only to “national[s] of the United States,” and their “estate, survivors, or heirs.” §2333(a). In contrast, the ATS is available only for claims brought by “an alien.” 28 U. S. C. §1350. A statute that excludes foreign nationals (with the possible exception of foreign survivors or heirs) is an inapt analogy for a common-law cause of action that provides a remedy for foreign nationals only. To the extent, furthermore, that the Anti-Terrorism Act is relevant it suggests that there should be no common-law action under the ATS for allegations like petitioners’. Otherwise, foreign plaintiffs could bypass Congress’ express limitations on liability under the Anti-Terrorism Act simply by bringing an ATS lawsuit. The Anti-Terrorism Act, as mentioned above, is part of a comprehensive statutory and regulatory regime that prohibits terrorism and terrorism financing. The detailed regulatory structures prescribed by Congress and the federal agencies charged with oversight of financial institutions reflect the careful deliberation of the political branches on when, and how, banks should be held liable for the financing of terrorism. It would be inappropriate for courts to displace this considered statutory and regulatory structure by holding banks subject to common-law liability in actions filed under the ATS. In any event, even if the Anti-Terrorism Act were a suitable model for an ATS suit, Congress’ decision in the TVPA to limit liability to individuals still demonstrates that there are two reasonable choices. In this area, that is dispositive—Congress, not the Judiciary, must decide whether to expand the scope of liability under the ATS to include foreign corporations. 3 Other considerations relevant to the exercise of judicial discretion also counsel against allowing liability under the ATS for foreign corporations, absent instructions from Congress to do so. It has not been shown that corporate liability under the ATS is essential to serve the goals of the statute. As to the question of adequate remedies, the ATS will seldom be the only way for plaintiffs to hold the perpetrators liable. See, e.g., 18 U. S. C. §1091 (criminal prohibition on genocide); §1595 (civil remedy for victims of slavery). And plaintiffs still can sue the individual corporate employees responsible for a violation of international law under the ATS. If the Court were to hold that foreign corporations have liability for international-law violations, then plaintiffs may well ignore the human perpetrators and concentrate instead on multinational corporate entities. As explained above, in the context of criminal tribunals international law itself generally limits liability to natural persons. Although the Court need not decide whether the seeming absence of a specific, universal, and obligatory norm of corporate liability under international law by itself forecloses petitioners’ claims against Arab Bank, or whether this is an issue governed by international law, the lack of a clear and well-established international-law rule is of critical relevance in determining whether courts should extend ATS liability to foreign corporations without specific congressional authorization to do so. That is especially so in light of the TVPA’s limitation of liability to natural persons, which parallels the distinction between corporations and individuals in international law. If, moreover, the Court were to hold that foreign corporations may be held liable under the ATS, that precedent-setting principle “would imply that other nations, also applying the law of nations, could hale our [corporations] into their courts for alleged violations of the law of nations.” Kiobel, 569 U. S., at 124. This judicially mandated doctrine, in turn, could subject American corporations to an immediate, constant risk of claims seeking to impose massive liability for the alleged conduct of their employees and subsidiaries around the world, all as determined in foreign courts, thereby “hinder[ing] global investment in developing economies, where it is most needed.” Brief for United States as Amicus Curiae in American Isuzu Motors, Inc. v. Ntsebeza, O. T. 2007, No. 07–919, p. 20 (internal quotation marks omitted). In other words, allowing plaintiffs to sue foreign corporations under the ATS could establish a precedent that discourages American corporations from investing abroad, including in developing economies where the host government might have a history of alleged human-rights violations, or where judicial systems might lack the safeguards of United States courts. And, in consequence, that often might deter the active corporate investment that contributes to the economic development that so often is an essential foundation for human rights. It is also true, of course, that natural persons can and do use corporations for sinister purposes, including conduct that violates international law. That the corporate form can be an instrument for inflicting grave harm and suffering poses serious and complex questions both for the international community and for Congress. So there are strong arguments for permitting the victims to seek relief from corporations themselves. Yet the urgency and complexity of this problem make it all the more important that Congress determine whether victims of human-rights abuses may sue foreign corporations in federal courts in the United States. Congress, not the Judiciary, is the branch with “the facilities necessary to make fairly such an important policy decision where the possibilities of international discord are so evident and retaliative action so certain.” Kiobel, 569 U. S., at 116 (internal quotation marks omitted). As noted further below, there are many delicate and important considerations that Congress is in a better position to examine in determining whether and how best to impose corporate liability. And, as the TVPA illustrates, Congress is well aware of the necessity of clarifying the proper scope of liability under the ATS in a timely way. C The ATS was intended to promote harmony in international relations by ensuring foreign plaintiffs a remedy for international-law violations in circumstances where the absence of such a remedy might provoke foreign nations to hold the United States accountable. Brief for United States as Amicus Curiae 7. But here, and in similar cases, the opposite is occurring. Petitioners are foreign nationals seeking hundreds of millions of dollars in damages from a major Jordanian financial institution for injuries suffered in attacks by foreign terrorists in the Middle East. The only alleged connections to the United States are the CHIPS transactions in Arab Bank’s New York branch and a brief allegation regarding a charity in Texas. The Court of Appeals did not address, and the Court need not now decide, whether these allegations are sufficient to “touch and concern” the United States under Kiobel. See 569 U. S., at 124–125. At a minimum, the relatively minor connection between the terrorist attacks at issue in this case and the alleged conduct in the United States well illustrates the perils of extending the scope of ATS liability to foreign multinational corporations like Arab Bank. For 13 years, this litigation has “caused significant diplomatic tensions” with Jordan, a critical ally in one of the world’s most sensitive regions. Brief for United States as Amicus Curiae 30. “Jordan is a key counterterrorism partner, especially in the global campaign to defeat the Islamic State in Iraq and Syria.” Id., at 31. The United States explains that Arab Bank itself is “a constructive partner with the United States in working to prevent terrorist financing.” Id., at 32 (internal quotation marks omitted). Jordan considers the instant litigation to be a “grave affront” to its sovereignty. See Brief for Hashemite Kingdom of Jordan as Amicus Curiae 3; see ibid. (“By exposing Arab Bank to massive liability, this suit thus threatens to destabilize Jordan’s economy and undermine its cooperation with the United States”). This is not the first time, furthermore, that a foreign sovereign has appeared in this Court to note its objections to ATS litigation. Sosa, 542 U. S., at 733, n. 21 (noting objections by the European Commission and South Africa); Brief for the Federal Republic of Germany as Amicus Curiae in Kiobel v. Royal Dutch Petroleum Co., O. T. 2012, No. 10–1491, p. 1; Brief for the Government of the United Kingdom of Great Britain and Northern Ireland and the Kingdom of the Netherlands as Amici Curiae in No. 10–1491, p. 3. These are the very foreign-relations tensions the First Congress sought to avoid. Petitioners insist that whatever the faults of this litigation—for example, its tenuous connections to the United States and the prolonged diplomatic disruptions it has caused—the fact that Arab Bank is a foreign corporate entity, as distinct from a natural person, is not one of them. That misses the point. As demonstrated by this litigation, foreign corporate defendants create unique problems. And courts are not well suited to make the required policy judgments that are implicated by corporate liability in cases like this one. Like the presumption against extraterritoriality, judicial caution under Sosa “guards against our courts triggering . . . serious foreign policy consequences, and instead defers such decisions, quite appropriately, to the political branches.” Kiobel, 569 U. S., at 124. If, in light of all the concerns that must be weighed before imposing liability on foreign corporations via ATS suits, the Court were to hold that it has the discretion to make that determination, then the cautionary language of Sosa would be little more than empty rhetoric. Accordingly, the Court holds that foreign corporations may not be defendants in suits brought under the ATS. III With the ATS, the First Congress provided a federal remedy for a narrow category of international-law violations committed by individuals. Whether, more than two centuries on, a similar remedy should be available against foreign corporations is similarly a decision that Congress must make. The political branches can determine, referring to international law to the extent they deem proper, whether to impose liability for human-rights violations upon foreign corporations in this Nation’s courts, and, conversely, that courts in other countries should be able to hold United States corporations liable. Congress might determine that violations of international law do, or should, impose that liability to ensure that corporations make every effort to deter human-rights violations, and so that, even when those efforts cannot be faulted, compensation for injured persons will be a cost of doing business. If Congress and the Executive were to determine that corporations should be liable for violations of international law, that decision would have special power and force because it would be made by the branches most immediately responsive to, and accountable to, the electorate. It is still another possibility that, in the careful exercise of its expertise in the field of foreign affairs, Congress might conclude that neutral judicial safeguards may not be ensured in every country; and so, as a reciprocal matter, it could determine that liability of foreign corporations under the ATS should be subject to some limitations or preconditions. Congress might deem this more careful course to be the best way to encourage American corporations to undertake the extensive investments and foreign operations that can be an important beginning point for creating the infrastructures that allow human rights, as well as judicial safeguards, to emerge. These delicate judgments, involving a balance that it is the prerogative of the political branches to make, especially in the field of foreign affairs, would, once again, also be entitled to special respect, especially because those careful distinctions might themselves advance the Rule of Law. All this underscores the important separation-of-powers concerns that require the Judiciary to refrain from making these kinds of decisions under the ATS. The political branches, moreover, surely are better positioned than the Judiciary to determine if corporate liability would, or would not, create special risks of disrupting good relations with foreign governments. Finally, Congress might find that corporate liability should be limited to cases where a corporation’s management was actively complicit in the crime. Cf. ALI, Model Penal Code §2.07(1)(c) (1985) (a corporation may be held criminally liable where “the commission of the offense was authorized, requested, commanded, performed or recklessly tolerated by the board of directors or by a high managerial agent acting on behalf of the corporation within the scope of his office or employment”). Again, the political branches are better equipped to make the preliminary findings and consequent conclusions that should inform this determination. These and other considerations that must shape and instruct the formulation of principles of international and domestic law are matters that the political branches are in the better position to define and articulate. For these reasons, judicial deference requires that any imposition of corporate liability on foreign corporations for violations of international law must be determined in the first instance by the political branches of the Government. The judgment of the Court of Appeals is affirmed. It is so ordered.

584.US.2017_17-5716

The five petitioners pleaded guilty to drug conspiracy charges that subjected them to mandatory minimum sentences under 21 U. S. C. §841(b)(1). Before imposing their sentences, the District Court calculated their advisory Guidelines ranges. But because the top end of the Guidelines ranges fell below the mandatory minimums, the court concluded that the mandatory minimums superseded the Guidelines ranges. After discarding these ranges, the court departed downward from the mandatory minimums under 18 U. S. C. §3553(e) to reflect petitioners’ substantial assistance to the Government in prosecuting other drug offenders. In settling on the final sentences, the court considered the relevant “substantial assistance factors” set out in the Guidelines, but it did not consider the original Guidelines ranges that it had earlier discarded. After petitioners were sentenced, the Sentencing Commission amended the Guidelines and reduced the base offense levels for certain drug offenses, including those for which petitioners were convicted. Petitioners sought sentence reductions under §3582(c)(2), which makes defendants eligible if they were sentenced “based on a sentencing range” that was later lowered by the Sentencing Commission. The courts below held that petitioners were not eligible because they could not show that their sentences were “based on” the now-lowered Guidelines ranges. Held: Petitioners do not qualify for sentence reductions under §3582(c)(2) because their sentences were not “based on” their lowered Guidelines ranges but, instead, were “based on” their mandatory minimums and on their substantial assistance to the Government. Pp. 3–7. (a) For a sentence to be “based on” a lowered Guidelines range, the range must have at least played “a relevant part [in] the framework the [sentencing] judge used” in imposing the sentence. Hughes v. United States, ante, at ___. Petitioners’ sentences do not fall into this category because the District Court did not consider the Guidelines ranges in imposing its ultimate sentences. On the contrary, the court scrapped the ranges in favor of the mandatory minimums and never considered the ranges again. Thus, petitioners may not receive §3582(c)(2) sentence reductions. Pp. 3–5. (b) Petitioners’ four counterarguments are unavailing. First, they insist that because this Court has said that the Guidelines ranges serve as “the starting point for every sentencing calculation in the federal system,” Peugh v. United States, 569 U. S. 530, 542, all sentences are “based on” Guidelines ranges. But that does not follow. Just because district courts routinely calculate defendants’ Guidelines ranges does not mean that any sentence subsequently imposed must be regarded as “based on” a Guidelines range. What matters instead is the role that the Guidelines range played in the selection of the sentence eventually imposed. And here the ranges played no relevant role. Second, petitioners argue that even if their sentences were not actually based on the Guidelines ranges, they are eligible under §3582(c)(2) because their sentences should have been based on those ranges. But even assuming that this is the correct interpretation of “based on,” petitioners are not eligible because the District Court made no mistake in sentencing them. The court properly discarded their Guidelines ranges and permissibly considered only factors related to substantial assistance when departing downward. Third, petitioners stress that the Sentencing Commission’s policy statement shows that defendants in their shoes should be eligible for sentence reductions. Policy statements, however, cannot make defendants eligible when §3582(c)(2) makes them ineligible. Fourth, petitioners contend that the Court’s rule creates unjustifiable sentencing disparities, but, in fact, the rule avoids such disparities. Identically situated defendants sentenced today may receive the same sentences petitioners received, and those defendants, like petitioners, are not eligible for sentence reductions under §3582(c)(2). Pp. 5–7. 850 F. 3d 973, affirmed. Alito, J., delivered the opinion for a unanimous Court.

Under 18 U. S. C. §3582(c)(2), a defendant is eligible for a sentence reduction if he was initially sentenced “based on a sentencing range” that was later lowered by the United States Sentencing Commission. The five petitioners in today’s case claim to be eligible under this provision. They were convicted of drug offenses that carried statutory mandatory minimum sentences, but they received sentences below these mandatory minimums, as another statute allows, because they substantially assisted the Government in prosecuting other drug offenders. We hold that petitioners’ sentences were “based on” their mandatory minimums and on their substantial assistance to the Government, not on sentencing ranges that the Commission later lowered. Petitioners are therefore ineligible for §3582(c)(2) sentence reductions. I All five petitioners pleaded guilty before the same sentencing judge to methamphetamine conspiracy offenses that subjected them to mandatory minimum sentences under 21 U. S. C. §841(b)(1). Before the District Court imposed those sentences, however, it first calculated petitioners’ advisory Guidelines ranges, as district courts do in sentencing proceedings all around the country. These ranges take into account the seriousness of a defendant’s offense and his criminal history in order to produce a set of months as a recommended sentence (e.g., 151 to 188 months for petitioner Koons). But not only are these ranges advisory, they are also tentative: They can be overridden by other considerations, such as a congressionally mandated minimum sentence. Indeed, the Guidelines themselves instruct that “[w]here a statutorily required minimum sentence is greater than the maximum of the applicable guideline range, the statutorily required minimum sentence shall be the [final] guideline sentence.” United States Sentencing Commission, Guidelines Manual §5G1.1(b) (Nov. 2016) (USSG); see also §1B1.1(a)(8). That is what happened here. In each of petitioners’ cases, the top end of the Guidelines range fell below the applicable mandatory minimum sentence, and so the court concluded that the mandatory minimum superseded the Guidelines range. E.g., App. 197; see also id., at 70. Thus, in all five cases, the court discarded the advisory ranges in favor of the mandatory minimum sentences. See id., at 114–115, 148, 174, 197, 216. When a statute sets out a mandatory minimum sentence, a defendant convicted under that statute will generally receive a sentence at or above the mandatory minimum—but not always. If the defendant has substantially assisted the Government “in the investigation or prosecution of another person,” the Government may move under 18 U. S. C. §3553(e) to allow the district court to “impose a sentence below” the mandatory minimum “so as to reflect [the] defendant’s substantial assistance.” The Government filed such motions in each of petitioners’ cases, and in each case, the District Court departed downward from the mandatory minimum because of petitioners’ substantial assistance. In settling on the final sentences, the court considered the so-called “substantial-assistance factors” found in §5K1.1(a) of the Guidelines, all of which relate to the assistance defendants supply the Government. App. 80, 197; see, e.g., USSG §§5K1.1(a)(1)–(3), (5) (the “extent,” “timeliness,” “significance[,] and usefulness” of the defendant’s assistance and the “truthfulness, completeness, and reliability of [the] information” provided). In no case did the court consider the original drug Guidelines ranges that it had earlier discarded. See App. 115–116, 148–154, 174–177, 197–198, 216–218. The sentences ultimately imposed in these cases represented downward departures from the mandatory minimums of between 25 and 45 percent. See Brief for United States 3. Years after petitioners’ sentences became final, the Sentencing Commission issued amendment 782, which reduced the Guidelines’ base offense levels for certain drug offenses, including those for which petitioners were convicted. See USSG App. C, Amdt. 782 (Supp. Nov. 2012–Nov. 2016); see also Hughes v. United States, ante, at 7. And because the amendment applied retroactively, ibid., it made defendants previously convicted of those offenses potentially eligible for a sentence reduction under §3582(c)(2). Petitioners sought such reductions, but in order to qualify, they had to show that their sentences were “based on” the now-lowered drug Guidelines ranges. §3582(c)(2). The courts below held that petitioners could not make that showing, App. 93–97; 850 F. 3d 973, 977 (CA8 2017), and we granted certiorari to review the question, 583 U. S. ___ (2017). II We hold that petitioners do not qualify for sentence reductions under §3582(c)(2) because their sentences were not “based on” their lowered Guidelines ranges. Instead, their sentences were “based on” their mandatory minimums and on their substantial assistance to the Government.[1] A For a sentence to be “based on” a lowered Guidelines range, the range must have at least played “a relevant part [in] the framework the [sentencing] judge used” in imposing the sentence. Hughes, ante, at 14; see ante, at 10–11. The Guidelines range will often play that part, for district judges must calculate the defendant’s advisory range and then will frequently tie the sentence they impose to that range. See ante, at 9–10; see also §3553(a)(4). But that is not always the case. After all, the Guidelines are advisory, and in some instances they even explicitly call for the ranges to be tossed aside. When that happens—when the ranges play no relevant part in the judge’s determination of the defendant’s ultimate sentence—the resulting sentence is not “based on” a Guidelines range. Petitioners’ sentences fall into this latter category of cases. Their sentences were not “based on” the lowered Guidelines ranges because the District Court did not consider those ranges in imposing its ultimate sentences. On the contrary, the court scrapped the ranges in favor of the mandatory minimums, and never considered the ranges again; as the court explained, the ranges dropped out of the case. App. 114–115, 148, 174, 197, 216. And once out of the case, the ranges could not come close to forming the “basis for the sentence that the District Court imposed,” Hughes, ante, at 14, and petitioners thus could not receive §3582(c)(2) sentence reductions. B Petitioners’ four counterarguments do not change our conclusion. First, petitioners insist that because the Guidelines ranges serve as “the starting point for every sentencing calculation in the federal system,” Peugh v. United States, 569 U. S. 530, 542 (2013), all sentences are “based on” Guidelines ranges. See Brief for Petitioners 21–22; Reply Brief 16–17. It is true that our cases require sentencing judges to calculate the now-advisory Guidelines range in every sentencing proceeding. And it is true that many judges use those ranges as “the foundation of [their] sentencing decisions.” Hughes, ante, at 8. But it does not follow that any sentence subsequently imposed must be regarded as “based on” a Guidelines range. What matters, instead, is the role that the Guidelines range played in the selection of the sentence eventually imposed—not the role that the range played in the initial calculation. And here, while consideration of the ranges may have served as the “starting point” in the sense that the court began by calculating those ranges, the ranges clearly did not form the “foundation” of the sentences ultimately selected. See Hughes, ante, at 9–11. In constructing a house, a builder may begin by considering one design but may ultimately decide to use entirely different plans. While the first design would represent the starting point in the builder’s decisionmaking process, the house finally built would not be “based on” that design. The same is true here. Petitioners’ sentences were not “based on” Guidelines ranges that the sentencing judge dis- carded in favor of mandatory minimums and substantial-assistance factors. Second, petitioners argue that even if their sentences were not actually based on their Guidelines ranges, they are eligible under §3582(c)(2) because their sentences should have been based on those ranges. See Brief for Petitioners 25–34.[2] But even under that reading of “based on,” petitioners are not eligible because the District Court made no mistake at sentencing. Petitioners emphasize that when a court departs downward because of a defendant’s substantial assistance, §3553(e) requires it to impose a sentence “in accordance with the guidelines.” Id., at 28 (emphasis deleted). But that does not mean “in accordance with the guidelines range.” Instead, a court imposes a sentence “in accordance with the guidelines” when it follows the Guidelines—including the parts of the Guidelines that instruct it to disregard the advisory ranges, see USSG §§1B1.1(a)(8), 5G1.1(b)—in settling on a sentence. And that is precisely what the court did here. It properly discarded the advisory ranges, ibid., and permissibly considered only factors related to petitioners’ substantial assistance, rather than factors related to the advisory ranges, as a guide in determining how far to depart downward, USSG §5K1.1. See §3553(e).[3] Third, petitioners stress that the Sentencing Commission’s policy statement makes clear that the Commission wanted defendants in their shoes to be eligible for sentence reductions. Brief for Petitioners 35–38; see USSG §1B1.10(c) (policy statement). But the Commission’s policy statement cannot alter §3582(c)(2), which applies only when a sentence was “based on” a subsequently lowered range. The Sentencing Commission may limit the application of its retroactive Guidelines amendments through its “ ‘applicable policy statements.’ ” Dillon v. United States, 560 U. S. 817, 824–826 (2010). But policy statements cannot make a defendant eligible when §3582(c)(2) makes him ineligible. See id., at 824–825. In short, because petitioners do not satisfy §3582(c)(2)’s threshold “based on” requirement, the Commission had no power to enable their sentence reductions. Fourth and finally, far from creating “unjustifiable sentencing disparities,” Brief for Petitioners 38–42, our rule avoids such disparities. Identically situated defendants sentenced today may receive the same sentences as petitioners received. See App. 89–90. Now, as then, district courts calculate the advisory Guidelines ranges, see USSG §1B1.1(a)(7); discard them in favor of the mandatory minimum sentences, §§1B1.1(a)(8), 5G1.1(b); and then may use the substantial-assistance factors to determine how far to depart downward, §§1B1.1(b), 5K1.1(a). See §3553(e). Those resulting sentences, like the sentences here, are not “based on” a lowered Guidelines range—they are “based on” the defendants’ mandatory minimums and substantial assistance to the Government. And those defendants, like petitioners, are not eligible for sentence reductions under §3582(c)(2). * * * For these reasons, we affirm. It is so ordered. Notes 1 The Government argues that defendants subject to mandatory minimum sentences can never be sentenced “based on a sentencing range” that the Commission has lowered, 18 U. S. C. §3582(c)(2), because such defendants’ “sentencing range[s]” are the mandatory minimums, which the Commission has no power to lower. See Brief for United States 19–28. We need not resolve the meaning of “sentencing range” today. Even if it referred to the discarded Guidelines range rather than the mandatory minimum—as petitioners contend, see Brief for Petitioners 20–21—petitioners still would not be eligible for sentence reductions: As explained in the text that follows, their sentences were not “based on” even that range. 2 We assume for argument’s sake that what should have happened at the initial sentencing proceedings, rather than what actually happened, matters for purposes of §3582(c)(2). But cf. Dillon v. United States, 560 U. S. 817, 825–826, 831 (2010). 3 Many courts have held that §3553(e) prohibits consideration of the advisory Guidelines ranges in determining how far to depart downward. See, e.g., United States v. Spinks, 770 F. 3d 285, 287–288, and n. 1 (CA4 2014) (collecting cases). We take no view on that issue. All we must decide today is that, at the least, neither §3553(e) nor the Guidelines required the District Court to use the advisory ranges in determining how far to depart downward.

584.US.2017_16-1519

Petitioner Sergio Fernando Lagos was convicted of using a company he controlled to defraud a lender of tens of millions of dollars. After the fraudulent scheme came to light and Lagos’ company went bankrupt, the lender conducted a private investigation of Lagos’ fraud and participated as a party in the company’s bankruptcy proceedings. Between the private investigation and the bankruptcy proceedings, the lender spent nearly $5 million in legal, accounting, and consulting fees related to the fraud. After Lagos pleaded guilty to federal wire fraud charges, the District Court ordered him to pay restitution to the lender for those fees. The Fifth Circuit affirmed, holding that such restitution was required by the Mandatory Victims Restitution Act of 1996, which requires defendants convicted of certain federal offenses, including wire fraud, to, among other things, “reimburse the victim for lost income and necessary child care, transportation, and other expenses incurred during participation in the investigation or prosecution of the offense or attendance at proceedings related to the offense,” 18 U. S. C. §3663A(b)(4). Held: 1. The words “investigation” and “proceedings” in subsection (b)(4) of the Mandatory Victims Restitution Act are limited to government investigations and criminal proceedings and do not include private investigations and civil or bankruptcy proceedings. The word “investigation” appears in the phrase “the investigation or prosecution.” Because the word “prosecution” must refer to a government’s criminal prosecution, this suggests that the word “investigation” refers to a government’s criminal investigation. Similar reasoning suggests that the immediately following reference to “proceedings” refers to criminal proceedings. Furthermore, the statute refers to the victim’s “participation” in the “investigation,” and “attendance” at “proceedings,” which would be odd ways to describe a victim’s role in its own private investigation and as a party in noncriminal court proceedings, but which are natural ways to describe a victim’s role in a government’s investigation and in the criminal proceedings that a government conducts. Moreover, the statute lists three specific items that must be reimbursed: lost income, child care expenses, and transportation expenses. These are precisely the kind of expenses that a victim is likely to incur when missing work and traveling to participate in a government investigation or to attend criminal proceedings. In contrast, the statute says nothing about the kinds of expenses a victim would often incur during private investigations or noncriminal proceedings, namely, the costs of hiring private investigators, attorneys, or accountants. This supports the Court’s more limited reading of the statute. A broad reading would also require district courts to resolve difficult, fact-intensive disputes about whether particular expenses “incurred during” participation in a private investigation were in fact “necessary,” and about whether proceedings such as a licensing proceeding or a Consumer Products Safety Commission hearing were sufficiently “related to the offense.” The Court’s narrower interpretation avoids such controversies, which are often irrelevant to the victim because over 90% of criminal restitution is never collected. The Court’s interpretation means that some victims will not receive restitution for all of their losses from a crime, but that is consistent with the Mandatory Victims Restitution Act’s enumeration of limited categories of covered expenses, in contrast with the broader language that other federal restitution statutes use, see, e.g., 18 U. S. C. §§2248(b), 2259(b), 2264(b), 2327(b). Pp. 3–7. 2. That the victim shared the results of its private investigation with the Government does not make the costs of conducting the private investigation “necessary . . . other expenses incurred during participation in the investigation . . . of the offense.” §3663A(b)(4). That language does not cover the costs of a private investigation that the victim chooses on its own to conduct, which are not “incurred during” participation in a government’s investigation. Pp. 7–8. 864 F. 3d 320, reversed and remanded. Breyer, J., delivered the opinion for a unanimous Court.

The Mandatory Victims Restitution Act of 1996 requires defendants convicted of a listed range of offenses to “reimburse the victim for lost income and necessary child care, transportation, and other expenses incurred during participation in the investigation or prosecution of the offense or attendance at proceedings related to the offense.” 18 U. S. C. §3663A(b)(4) (emphasis added). We must decide whether the words “investigation” and “proceedings” are limited to government investigations and criminal proceedings, or whether they include private investigations and civil proceedings. In our view, they are limited to government investigations and criminal proceedings. I The petitioner, Sergio Fernando Lagos, was convicted of using a company that he controlled (Dry Van Logistics) to defraud a lender (General Electric Capital Corporation, or GE) of tens of millions of dollars. The fraud involved generating false invoices for services that Dry Van Logistics had not actually performed and then borrowing money from GE using the false invoices as collateral. Eventually, the scheme came to light. Dry Van Logistics went bankrupt. GE investigated. The Government indicted Lagos. Lagos pleaded guilty to wire fraud. And the judge, among other things, ordered him to pay GE restitution. The issue here concerns the part of the restitution order that requires Lagos to reimburse GE for expenses GE incurred during its own investigation of the fraud and during its participation in Dry Van Logistics’ bankruptcy proceedings. The amounts are substantial (about $5 million), and primarily consist of professional fees for attorneys, accountants, and consultants. The Government argued that the District Court must order restitution of these amounts under the Mandatory Victims Restitution Act because these sums were “necessary . . . other expenses incurred during participation in the investigation . . . of the offense or attendance at proceedings related to the offense.” §3663A(b)(4). The District Court agreed, as did the U. S. Court of Appeals for the Fifth Circuit. 864 F. 3d 320, 323 (2017). Lagos filed a petition for certiorari. And in light of a division of opinion on the matter, we granted the petition. Compare United States v. Papagno, 639 F. 3d 1093, 1100 (CADC 2011) (subsection (b)(4) of the Mandatory Victims Restitution Act does not cover private investigation costs), with United States v. Elson, 577 F. 3d 713, 726–729 (CA6 2009) (statute not so limited); United States v. Hosking, 567 F. 3d 329, 331–332 (CA7 2009) (same); United States v. Stennis-Williams, 557 F. 3d 927, 930 (CA8 2009) (same); United States v. Amato, 540 F. 3d 153, 159–163 (CA2 2008) (same); United States v. Gordon, 393 F. 3d 1044, 1056–1057 (CA9 2004) (same). II The Mandatory Victims Restitution Act is one of several federal statutes that govern federal court orders requiring defendants convicted of certain crimes to pay their victims restitution. It concerns “crime[s] of violence,” “offense[s] against property . . . , including any offense committed by fraud or deceit,” and two specific offenses, one concern- ing tampering with a consumer product and the other concerning theft of medical products. 18 U. S. C. §3663A(c)(1)(A). It requires, in the case of property offenses, return of the property taken or its value, §3663A(b)(1); in the case of bodily injury, the payment of medical expenses and lost income, §3663A(b)(2); in the case of death, the payment of funeral expenses, §3663A(b)(3); and, as we have said, supra, at 1, in all cases, “reimburse[ment]” to “the victim for lost income and necessary child care, transportation, and other expenses incurred during participation in the investigation or prosecution of the offense or attendance at proceedings related to the offense.” §3663A(b)(4) (emphasis added). We here consider the meaning of that italicized phrase. Specifically, we ask whether the scope of the words “investigation” and “proceedings” is limited to government investigations and criminal proceedings, or whether it includes private investigations and civil or bankruptcy litigation. We conclude that those words are limited to government investigations and criminal proceedings. Our conclusion rests in large part upon the statute’s wording, both its individual words and the text taken as a whole. The individual words suggest (though they do not demand) our limited interpretation. The word “investigation” is directly linked by the word “or” to the word “prosecution,” with which it shares the article “the.” This suggests that the “investigation[s]” and “prosecution[s]” that the statute refers to are of the same general type. And the word “prosecution” must refer to a government’s criminal prosecution, which suggests that the word “investigation” may refer to a government’s criminal investigation. A similar line of reasoning suggests that the immediately following reference to “proceedings” also refers to criminal proceedings in particular, rather than to “proceedings” of any sort. Furthermore, there would be an awkwardness about the statute’s use of the word “participation” to refer to a victim’s role in its own private investigation, and the word “attendance” to refer to a victim’s role as a party in noncriminal court proceedings. A victim opting to pursue a private investigation of an offense would be more naturally said to “provide for” or “conduct” the private investigation (in which he may, or may not, actively “participate”). And a victim who pursues civil or bankruptcy litigation does not merely “atten[d]” such other “proceedings related to the offense” but instead “participates” in them as a party. In contrast, there is no awkwardness, indeed it seems perfectly natural, to say that a victim “participat[es] in the investigation” or “attend[s] . . . proceedings related to the offense” if the investigation at issue is a government’s criminal investigation, and if the proceedings at issue are criminal proceedings conducted by a government. Moreover, to consider the statutory phrase as a whole strengthens these linguistic points considerably. The phrase lists three specific items that must be reimbursed, namely, lost income, child care, and transportation; and it then adds the words, “and other expenses.” §3663A(b)(4). Lost income, child care expenses, and transportation expenses are precisely the kind of expenses that a victim would be likely to incur when he or she (or, for a corporate victim like GE, its employees) misses work and travels to talk to government investigators, to participate in a government criminal investigation, or to testify before a grand jury or attend a criminal trial. At the same time, the statute says nothing about the kinds of expenses a victim would often incur when private investigations, or, say, bankruptcy proceedings are at issue, namely, the costs of hiring private investigators, attorneys, or accountants. Thus, if we look to noscitur a sociis, the well-worn Latin phrase that tells us that statutory words are often known by the company they keep, we find here both the presence of company that suggests limitation and the absence of company that suggests breadth. See, e.g., Yates v. United States, 574 U. S. ___, ___ (2015) (slip op., at 14). We add a practical fact: A broad reading would create significant administrative burdens. The statute provides for mandatory restitution, and the portion we construe is limited to “necessary . . . other expenses.” §3663A(b)(4) (emphasis added). The word “necessary” would, if the statute is broadly interpreted, invite disputes as to whether particular expenses “incurred during” participation in a private investigation or attendance at, say, a bankruptcy proceeding, were in fact “necessary.” Such disputes may become burdensome in cases involving multimillion dollar investigation expenses for teams of lawyers and accountants. A district court might, for example, need to decide whether each witness interview and each set of documents reviewed was really “necessary” to the investigation. Similarly, the statute also limits restitution to expenses incurred only during “attendance at proceedings related to the offense,” ibid. (emphasis added), inviting disputes as to whether, say, a licensing proceeding, a human resources review, an in-house disciplinary proceeding, a job interview, a Consumer Product Safety Commission hearing, or a neighborhood watch meeting qualified as “proceedings” sufficiently “related to the offense” so as to be eligible for restitution. To interpret the statute broadly is to invite controversy on those and other matters; our narrower construction avoids it. And one begins to doubt whether Congress intended, in making this restitution mandatory, to require courts to resolve these potentially time-consuming controversies as part of criminal sentencing—particularly once one realizes that few victims are likely to benefit because more than 90% of criminal restitution is never collected. See GAO, Federal Criminal Restitution: Most Debt Is Outstanding and Oversight of Collections Could Be Improved 25 (GAO–18–203, 2018) (explaining that the Justice Department considers 91% of outstanding criminal restitution to be “uncollectible”). There are, of course, contrary arguments—arguments favoring a broad interpretation. The Government points out, in particular, that our narrow interpretation will sometimes leave a victim without a restitution remedy sufficient to cover some expenses (say, those related to his private investigation) which he undoubtedly incurred as a result of the offense. Leaving the victim without that restitution remedy, the Government adds, runs contrary to the broad purpose of the Mandatory Victims Restitution Act, namely, “to ensure that victims of a crime receive full restitution.” Dolan v. United States, 560 U. S. 605, 612 (2010). But a broad general purpose of this kind does not always require us to interpret a restitution statute in a way that favors an award. After all, Congress has enacted many different restitution statutes with differing language, governing different circumstances. Some of those statutes specifically require restitution for the “full amount of the victim’s losses,” defined to include “any . . . losses suffered by the victim as a proximate result of the offense.” See 18 U. S. C. §§2248(b), 2259(b), 2264(b), 2327(b). The Mandatory Victims Restitution Act, however, contains no such language; it specifically lists the kinds of losses and expenses that it covers. Moreover, in at least one other statute Congress has expressly provided for restitution of “the value of the time reasonably spent by the victim in an attempt to remediate the intended or actual harm incurred by the victim from the offense.” §3663(b)(6). Again the Mandatory Victims Restitution Act has no similar provision. And given those differences between the Mandatory Victims Restitution Act and other restitution statutes, we conclude that the considerations we have mentioned, particularly those based on a reading of the statute as a whole, tip the balance in favor of our more limited interpretation. We add that this interpretation does not leave a victim such as GE totally without a remedy for additional losses not covered by the Mandatory Victims Restitution Act. GE also brought a civil lawsuit against Lagos for the full extent of its losses, and obtained an over-$30 million judgment against him. The Government says that GE has largely been unable to collect on that judgment, but there is no reason to think that collection efforts related to a criminal restitution award would prove any more successful. The Government makes one additional argument. It points out that GE shared with the Government the information that its private investigation uncovered. And that fact, the Government says, should bring the expenses of that investigation within the terms of the statute even if the “investigation” referred to by the statute is a government’s criminal investigation. The short, conclusive answer to that claim, however, lies in the fact that the statute refers to “necessary child care, transportation, and other expenses incurred during participation in the investigation or prosecution of the offense.” §3663A(b)(4) (emphasis added). It does not refer to expenses incurred before the victim’s participation in a government’s investigation began. And the Government does not deny that it is those preparticipation expenses—the expenses of conducting GE’s investigation, not those of sharing the results from it—that are at issue here. We therefore need not address in this case whether this part of the Mandatory Victims Restitution Act would cover similar expenses incurred during a private investigation that was pursued at a government’s invitation or request. It is enough to hold that it does not cover the costs of a private investigation that the victim chooses on its own to conduct. * * * For the reasons stated, we conclude that the words “investigation” and “proceedings” in the Mandatory Victims Restitution Act refer to government investigations and criminal proceedings. Consequently Lagos is not obliged to pay the portion of the restitution award that he here challenges. We reverse the Court of Appeals’ judgment to the contrary, and we remand the case for further proceedings consistent with this opinion. It is so ordered.

584.US.2017_16-1215

Respondent R. Scott Appling fell behind on his bills owed to petitioner law firm Lamar, Archer & Cofrin, LLP, which threatened to withdraw representation and place a lien on its work product if Appling did not pay. Appling told Lamar that he could cover owed and future legal expenses with an expected tax refund, so Lamar agreed to continue representation. However, Appling used the refund, which was for much less than he had stated, for business expenses. When he met with Lamar again, he told the firm he was still waiting on the refund, so Lamar agreed to complete pending litigation. Appling never paid the final invoice, so Lamar sued him and obtained a judgment. Shortly thereafter, Appling and his wife filed for Chapter 7 bankruptcy. Lamar initiated an adversary proceeding against Appling in Bankruptcy Court, arguing that his debt to Lamar was nondischargeable pursuant to 11 U. S. C. §523(a)(2)(A), which bars discharge of specified debts arising from “false pretenses, a false representation, or actual fraud, other than a statement respecting the debtor’s . . . financial condition.” Appling moved to dismiss on the ground that his alleged misrepresentations were “statement[s] respecting the debtor’s . . . financial condition,” which §523(a)(2)(B) requires to be “in writing.” The Bankruptcy Court disagreed and denied Appling’s motion. Finding that Appling knowingly made two false representations on which Lamar justifiably relied and that Lamar incurred damages as a result, the court concluded that Appling’s debt to Lamar was nondischargeable under §523(a)(2)(A). The District Court affirmed, but the Eleventh Circuit reversed, holding that a “statement respecting the debtor’s financial condition” may include a statement about a single asset. Because Appling’s statements were not in writing, the court held, §523(a)(2)(B) did not bar him from discharging his debt to Lamar. Held: A statement about a single asset can be a “statement respecting the debtor’s financial condition” under §523(a)(2). Pp. 4–15. (a) The key word in the relevant statutory phrase here is the preposition “respecting.” In ordinary usage, “respecting” means “concerning; about; regarding; in regard to; relating to.” Lamar contends that the definitions “about,” “concerning,” “with reference to,” and “as regards” denote a more limited scope than “related to.” And under that more limited meaning, Lamar asserts, a formal financial statement providing a detailed accounting of one’s assets and liabilities would qualify as “a statement respecting the debtor’s financial condition,” but a statement about a single asset would not. But the overlapping and circular definitions of these words belie the clear distinction Lamar attempts to impose. And the firm gives no example of a phrase in a legal context similar to the one at issue here in which toggling between “related to” and “about” has any pertinent significance. Use of the word “respecting” in a legal context generally has a broadening effect, ensuring that a provision’s scope covers not only its subject but also matters relating to that subject. Cf. Kleppe v. New Mexico, 426 U. S. 529, 539. Indeed, this Court has typically read the phrase “relating to”—one of respecting’s meanings—expansively. See, e.g., Coventry Health Care of Mo., Inc. v. Nevils, 581 U. S. ___, ___. Appling and the United States, as amicus curiae, accordingly advance an expansive interpretation here. This Court agrees with them that, given the ordinary meaning of “respecting,” Lamar’s statutory construction must be rejected, for it reads “respecting” out of the statute. See TRW Inc. v. Andrews, 534 U. S. 19, 31. Had Congress intended §523(a)(2)(B) to encompass only statements expressing the balance of a debtor’s assets and liabilities, it could have so specified—e.g., “statement of the debtor’s financial condition.” The Court also agrees that a statement is “respecting” a debtor’s financial condition if it has a direct relation to or impact on the debtor’s overall financial status. A single asset has a direct relation to and impact on aggregate financial condition, so a statement about that asset bears on a debtor’s overall financial condition and can help indicate whether a debtor is solvent or insolvent. A statement about a single asset, thus, can be a “statement respecting the debtor’s financial condition.” Pp. 5–9. (b) Lamar’s interpretation would yield incoherent results. For instance, on Lamar’s view, a misrepresentation about a single asset made in the context of a formal financial statement or balance sheet would constitute a “statement respecting the debtor’s financial condition” and trigger §523(a)(2)(B)’s heightened nondischargeability requirements, but the same misrepresentation made on its own, or in the context of a list of some but not all of the debtor’s assets and liabilities, would not. Lamar does not explain why Congress would draw such seemingly arbitrary distinctions. Pp. 9–10. (c) The statutory history of the phrase “statement respecting the debtor’s financial condition” corroborates this Court’s reading. Between 1926, when the phrase was introduced, and 1978, when Congress enacted the Bankruptcy Code, Courts of Appeals consistently construed the phrase to encompass statements addressing just one or some of a debtor’s assets or liabilities. When Congress used the materially same language in §523(a)(2), it presumptively was aware of this longstanding judicial interpretation and intended for the phrase to retain its established meaning. Pp. 10–11. (d) Lamar’s additional arguments are unpersuasive. First, Lamar contends that Appling’s construction gives §523(a)(2)(B) an implausibly broad reach, such that little would be covered by §523(a)(2)(A)’s general rule rendering nondischargeable debts arising from “false pretenses, a false representation, or actual fraud.” But §523(a)(2)(A) still retains significant function when the phrase “statement respecting the debtor’s financial condition” is interpreted to encompass a statement about a single asset. See, e.g., Husky Int’l Electronics, Inc. v. Ritz, 578 U. S. ___, ___. Second, Lamar asserts that Appling’s interpretation is inconsistent with the overall principle that the Bankruptcy Code exists to afford relief only to the “ ‘honest but unfortunate debtor.’ ” Cohen v. de la Cruz, 523 U. S. 213, 217. The text of §523(a)(2), however, plainly heightens the bar to discharge when the fraud at issue was effectuated via a “statement respecting the debtor’s financial condition.” The heightened requirements, moreover, are not a shield for dishonest debtors. Rather, they reflect Congress’ effort to balance the potential misuse of such statements by both debtors and creditors. See Field v. Mans, 516 U. S. 59, 76–77. Pp. 12–15. 848 F. 3d 953, affirmed. Sotomayor, J., delivered the opinion of the Court, in which Roberts, C. J., and Kennedy, Ginsburg, Breyer, and Kagan, JJ., joined, and in which Thomas, Alito, and Gorsuch, JJ., joined as to all but Part III–B.

if she returns to work is not a statement that respects the debtor’s financial condition”). 6 See, e.g., In re Bocchino, 794 F. 3d 376, 380–383 (CA3 2015) (holding nondischargeable under §523(a)(2)(A) civil judgment debts against a debtor-stockbroker who made misrepresentations about investments); In re Cohen, 106 F. 3d 52, 54–55 (CA3 1997) (holding that a landlord’s misrepresentations about the rent that legally could be charged for an apartment constituted fraud under §523(a)(2)(A)); United States v. Spicer, 57 F. 3d 1152, 1154, 1161 (CADC 1995) (holding nondischargeable under §523(a)(2)(A) a settlement agreement owed to the Government based on an investor’s misrepresentations of downpayment amounts in mortgage applications). 7 In addition to the writing requirement, §523(a)(2)(B) requires a creditor to show reasonable reliance. 11 U. S. C. §523(a)(2)(B)(iii). Section 523(a)(2)(A), by contrast, requires only the lesser showing of “justifiable reliance.” Field v. Mans, 516 U. S. 59, 61, 70–75 (1995).

585.US.2017_17-21

After petitioner Lozman towed his floating home into a slip in a marina owned by the city of Riviera Beach, he became an outspoken critic of the City’s plan to use its eminent domain power to seize waterfront homes for private development and often made critical comments about officials during the public-comment period of city council meetings. He also filed a lawsuit alleging that the City Council’s approval of an agreement with developers violated Florida’s open-meetings laws. In June 2006 the Council held a closed-door session, in part to discuss Lozman’s lawsuit. He alleges that the meeting’s transcript shows that councilmembers devised an official plan to intimidate him, and that many of his subsequent disputes with city officials and employees were part of the City’s retaliation plan. Five months after the closed-door meeting, the Council held a public meeting. During the public-comment session, Lozman began to speak about the arrests of officials from other jurisdictions. When he refused a councilmember’s request to stop making his remarks, the councilmember told the police officer in attendance to “carry him out.” The officer handcuffed Lozman and ushered him out of the meeting. The City contends that he was arrested for violating the City Council’s rules of procedure by discussing issues unrelated to the City and then refusing to leave the podium. Lozman claims that his arrest was to retaliate for his lawsuit and his prior public criticisms of city officials. The State’s attorney determined that there was probable cause for his arrest, but decided to dismiss the charges. Lozman then filed suit under 42 U. S. C. §1983, alleging a number of incidents that, under his theory, showed the City’s purpose was to harass him, including by initiating an admiralty lawsuit against his floating home, see Lozman v. Riviera Beach, 568 U. S. 115. The jury returned a verdict for the City on all of the claims. The District Court instructed the jury that, for Lozman to prevail on his claim of a retaliatory arrest at the city council meeting, he had to prove that the arresting officer was motivated by impermissible animus against Lozman’s protected speech and that the officer lacked probable cause to make the arrest. The Eleventh Circuit affirmed, concluding that any error the District Court made when it instructed the jury to consider the officer’s retaliatory animus was harmless because the jury necessarily determined that the arrest was supported by probable cause when it found for the City on Lozman’s other claims. The existence of probable cause, the court ruled, defeated a First Amendment claim for retaliatory arrest. Held: The existence of probable cause does not bar Lozman’s First Amendment retaliation claim under the circumstances of this case. Pp. 5–13. (a) The issue here is narrow. Lozman concedes that there was probable cause for his arrest. Nonetheless, he claims, the arrest violated the First Amendment because it was ordered in retaliation for his earlier, protected speech: his open-meetings lawsuit and his prior public criticisms of city officials. Pp. 5–6. (b) In a §1983 case, a city or other local governmental entity cannot be subject to liability unless the harm was caused in the implementation of “official municipal policy.” Monell v. New York City Dept. of Social Servs., 436 U. S. 658, 691. The Court assumes that Lozman’s arrest was taken pursuant to an official city policy. Two major precedents bear on the issue whether the conceded existence of probable cause for the arrest bars recovery regardless of any intent or purpose to retaliate for past speech. Lozman argues that the controlling rule is found in Mt. Healthy City Bd. of Ed. v. Doyle, 429 U. S. 274, a civil case in which a city board of education decided not to rehire an untenured teacher after a series of incidents, including a telephone call to a local radio station. The phone call was protected speech, but, the Court held, there was no liability unless the alleged constitutional violation was a but-for cause of the employment termination. Id., at 285–287. The City counters that the applicable precedent is Hartman v. Moore, 547 U. S. 250, where the Court held that a plaintiff alleging a retaliatory prosecution must show the absence of probable cause for the underlying criminal charge, id., at 265–266. If there was probable cause, the case ends. If the plaintiff proves the absence of probable cause, then the Mt. Healthy test governs. Pp. 6–10. (c) Whether Hartman or Mt. Healthy governs here is a determination that must await a different case. For Lozman’s claim is far afield from the typical retaliatory arrest claim, and the difficulties that might arise if Mt. Healthy is applied to the mine run of arrests made by police officers are not present here. Lozman alleges that the City itself retaliated against him pursuant to an “official municipal policy” of intimidation. Monell, supra, at 691. The fact that he must prove the existence and enforcement of an official policy motivated by retaliation separates his claim from the typical retaliatory arrest claim. An official retaliatory policy can be long term and pervasive, unlike an ad hoc, on-the-spot decision by an individual officer. And it can be difficult to dislodge. A citizen can seek to have an individual officer disciplined or removed from service, but there may be little practical recourse when the government itself orchestrates the retaliation. Lozman’s allegations, if proved, also alleviate the problems that the City says will result from applying Mt. Healthy in retaliatory arrest cases, for it is unlikely that the connection between the alleged animus and injury in a case like this will be “weakened . . . by [an official’s] legitimate consideration of speech,” Reichle v. Howards, 566 U. S. 658, 668, and there is little risk of a flood of retaliatory arrest suits against high-level policymakers. Because Lozman alleges that the City deprived him of the right to petition, “ ‘one of the most precious of the liberties safeguarded by the Bill of Rights,’ ” BE&K Constr. Co. v. NLRB, 536 U. S. 516, 524, his speech is high in the hierarchy of First Amendment values. On these facts, Mt. Healthy provides the correct standard for assessing a retaliatory arrest claim. On remand, the Eleventh Circuit may consider any arguments in support of the District Court’s judgment that have been preserved by the City, including whether a reasonable juror could find that the City formed a retaliatory policy to intimidate Lozman during its closed-door session, whether a reasonable juror could find that the arrest constituted an official act by the City, and whether, under Mt. Healthy, the City has proved that it would have arrested Lozman regardless of any retaliatory animus. Pp. 10–13. 681 Fed. Appx. 746, vacated and remanded. Kennedy, J., delivered the opinion of the Court, in which Roberts, C. J., and Ginsburg, Breyer, Alito, Sotomayor, Kagan, and Gorsuch, JJ., joined. Thomas, J., filed a dissenting opinion.

This case requires the Court to address the intersection of principles that define when arrests are lawful and principles that prohibit the government from retaliating against a person for having exercised the right to free speech. An arrest deprives a person of essential liberties, but if there is probable cause to believe the person has committed a criminal offense there is often no recourse for the deprivation. See, e.g., Devenpeck v. Alford, 543 U. S. 146, 153 (2004). At the same time, the First Amendment prohibits government officials from retaliating against individuals for engaging in protected speech. Crawford-El v. Britton, 523 U. S. 574, 592 (1998). The petitioner in this case alleges that high-level city policymakers adopted a plan to retaliate against him for protected speech and then ordered his arrest when he attempted to make remarks during the public-comment portion of a city council meeting. The petitioner now concedes there was probable cause for the arrest. The question is whether the presence of probable cause bars petitioner’s retaliatory arrest claim under these circumstances. I The city of Riviera Beach is on the Atlantic coast of Florida, about 75 miles north of Miami. The petitioner here is Fane Lozman. In 2006 Lozman towed his floating home into a slip in the City-owned marina, where he became a resident. Thus began his contentious relationship with the City’s elected officials. Soon after his arrival Lozman became an outspoken critic of the City’s plan to use its eminent domain power to seize homes along the waterfront for private development. Lozman often spoke during the public-comment period at city council meetings and criticized councilmembers, the mayor, and other public employees. He also filed a lawsuit alleging that the Council’s approval of an agreement with developers violated Florida’s open-meetings laws. In June 2006 the Council held a closed-door session, in part to discuss the open-meetings lawsuit that Lozman recently had filed. According to the transcript of the meeting, Councilmember Elizabeth Wade suggested that the City use its resources to “intimidate” Lozman and others who had filed lawsuits against the City. App. 176. Later in the meeting a different councilmember asked whether there was “a consensus of what Ms. Wade is saying,” and others responded in the affirmative. Id., at 181–182. Lozman alleges that these remarks formed an official plan to intimidate him. The City, on the other hand, maintains that the only consensus reached during the meeting was to invest the money and resources necessary to prevail in the litigation against it. In all events, Lozman became embroiled in a number of disputes with city officials and employees over the ensuing years, many of which Lozman says were part of the City’s plan of retaliation. The dispute that led to this litigation took place in 2006. In November of that year, five months after the closed-door meeting where the “intimidate” com- ment was made, the City Council held a public meeting. The agenda included a public-comment session in which citizens could address the Council for a few minutes. As he had done on earlier occasions and would do more than 200 times over the coming years, see Tr. in No. 9:08–cv–80134 (SD Fla.), Doc. 785, p. 61, Lozman stepped up to the podium to give remarks. He began to discuss the recent arrest of a former county official. Councilmember Wade interrupted Lozman, directing him to stop making those remarks. Lozman continued speaking, this time about the arrest of a former official from the city of West Palm Beach. Wade then called for the assistance of the police officer in attendance. The officer approached Lozman and asked him to leave the podium. Lozman refused. So Wade told the officer to “carry him out.” The officer handcuffed Lozman and ushered him out of the meeting. The incident was recorded on video. See Record, Def. Exh. 505, Doc. 687, available at https://www.supremecourt.gov/media/ video/mp4files/Lozman_v_RivieraBeach.mp4. According to the City, Lozman was arrested because he violated the City Council’s rules of procedure by discussing issues unrelated to the City and then refused to leave the po- dium. According to Lozman, the arrest was to retaliate for his open-meetings lawsuit against the City and his prior public criticisms of city officials. Under arrest, Lozman was escorted to police headquarters. He was charged with disorderly conduct and resisting arrest without violence and then released. Later, the State’s attorney determined there was probable cause to arrest Lozman for those offenses but decided to dismiss the charges. Lozman filed this lawsuit under Rev. Stat. §1979, 42 U. S. C. §1983. The complaint described a number of alleged incidents that, under Lozman’s theory, showed the City’s purpose to harass him in different ways. These ranged from a city employee telling Lozman that his dog needed a muzzle to the City’s initiation of an admiralty lawsuit against Lozman’s floating home—the latter resulting in an earlier decision by this Court. See Lozman v. Riviera Beach, 568 U. S. 115 (2013). The evidence and arguments presented by both parties with respect to all the matters alleged in Lozman’s suit consumed 19 days of trial before a jury. The jury returned a verdict for the City on all of the claims. Before this Court, Lozman seeks a reversal only as to the City’s alleged retaliatory arrest at the November 2006 city council meeting. The District Court instructed the jury that, for Lozman to prevail on this claim, he had to prove that the arresting officer was himself motivated by impermissible animus against Lozman’s protected speech and that the officer lacked probable cause to make the arrest. The District Court determined that the evidence was insufficient as a matter of law to support probable cause for the offenses charged at the time of the arrest (disorderly conduct and resisting arrest without violence). But the District Court concluded that there may have been probable cause to arrest Lozman for violating a Florida statute that prohibits interruptions or disturbances in schools, churches, or other public assemblies. Fla. Stat. §871.01 (2017). (The City had brought this statute to the District Court’s attention during the course of the litigation.) The District Court allowed the jury to decide whether there was probable cause to arrest for the public-disturbance offense. Judgment having been entered for the City after the jury’s verdict, Lozman appealed. The Court of Appeals for the Eleventh Circuit affirmed. 681 Fed. Appx. 746 (2017). As relevant here, the Court of Appeals assumed that the District Court erred when it instructed the jury that the officer, rather than the City, must have harbored the retaliatory animus. But the Court of Appeals held that any error was harmless because the jury necessarily determined that the arrest was supported by probable cause when it found for the City on some of Lozman’s other claims—specifically, his claims that the arrest violated the Fourth Amendment and state law. Id., at 751–752. And, under precedents which the Court of Appeals deemed controlling, the existence of probable cause defeated a First Amendment claim for retaliatory arrest. See id., at 752 (citing Dahl v. Holley, 312 F. 3d 1228, 1236 (CA11 2002)). This Court granted certiorari, 583 U. S. ___ (2017), on the issue whether the existence of probable cause defeats a First Amendment claim for retaliatory arrest under §1983. The Court considered this issue once before, see Reichle v. Howards, 566 U. S. 658, 663 (2012), but resolved the case on different grounds. II The issue before the Court is a narrow one. In this Court Lozman does not challenge the constitutionality of Florida’s statute criminalizing disturbances at public assemblies. He does not argue that the statute is overly broad, e.g., Terminiello v. Chicago, 337 U. S. 1 (1949); Watchtower Bible & Tract Soc. of N. Y., Inc. v. Village of Stratton, 536 U. S. 150 (2002); or that it impermissibly targets speech based on its content or viewpoint, e.g., Texas v. Johnson, 491 U. S. 397 (1989); Cohen v. California, 403 U. S. 15 (1971); or that it was enforced in a way that curtailed Lozman’s right to peaceful assembly, e.g., Brown v. Louisiana, 383 U. S. 131 (1966). Lozman, furthermore, does not challenge the validity of the City Council’s asserted limitations on the subjects speakers may discuss during the public-comment portion of city council meetings (although he continues to dispute whether those limitations in fact existed). Instead Lozman challenges only the lawfulness of his arrest, and even that challenge is a limited one. There is no contention that the City ordered Lozman’s arrest to discriminate against him based on protected classifications, or that the City denied Lozman his equal protection rights by placing him in a “class of one.” See Village of Willowbrook v. Olech, 528 U. S. 562 (2000) (per curiam). Lozman, moreover, now concedes that there was probable cause for the arrest. Although Lozman does not indicate what facts he believes support this concession, it appears that the existence of probable cause must be based on the assumption that Lozman failed to depart the podium after receiving a lawful order to leave. Lozman’s claim is that, notwithstanding the presence of probable cause, his arrest at the city council meeting violated the First Amendment because the arrest was ordered in retaliation for his earlier, protected speech: his open-meetings lawsuit and his prior public criticisms of city officials. The question this Court is asked to consider is whether the existence of probable cause bars that First Amendment retaliation claim. III It is well established that in a §1983 case a city or other local governmental entity cannot be subject to liability at all unless the harm was caused in the implementation of “official municipal policy.” Monell v. New York City Dept. of Social Servs., 436 U. S. 658, 691 (1978); see Los Angeles County v. Humphries, 562 U. S. 29, 36 (2010). Lozman’s §1983 damages claim is against only the City itself, based on the acts of its officers and employees—here, the members of the City Council. Lozman says that the City, through its city councilmembers, formed an official policy to retaliate against him and ordered his arrest. The Court assumes in the discussion to follow that the arrest was taken pursuant to an official city policy, but whether there was such a policy and what its content may have been are issues not decided here. This brings the discussion to the issue the parties deem central to the case: whether the conceded existence of probable cause for the arrest bars recovery regardless of any intent or purpose to retaliate for past speech. Two major precedents could bear on this point, and the parties disagree on which should be applicable here. The first is this Court’s decision in Mt. Healthy City Bd. of Ed. v. Doyle, 429 U. S. 274 (1977). See also Board of Comm’rs, Wabaunsee Cty. v. Umbehr, 518 U. S. 668 (1996). Lozman urges that the rule of Mt. Healthy should control and that under it he is entitled to recover. The second is this Court’s decision in Hartman v. Moore, 547 U. S. 250 (2006), which the City cites for the proposition that once there is probable cause there can be no further claim that the arrest was retaliation for protected speech. Mt. Healthy arose in a civil, not criminal, context. A city board of education decided not to rehire an untenured school teacher after a series of incidents indicating unprofessional demeanor. 429 U. S., at 281–283. One of the incidents was a telephone call the teacher made to a local radio station to report on a new school policy. Id., at 282. Because the board of education did not suggest that the teacher violated any established policy in making the call, this Court accepted a finding by the District Court that the call was protected speech. Id., at 284. The Court went on to hold, however, that since the other incidents, standing alone, would have justified the dismissal, relief could not be granted if the board could show that the discharge would have been ordered even without reference to the protected speech. Id., at 285–287. In terms of precepts in the law of torts, the Court held that even if retaliation might have been a substantial motive for the board’s action, still there was no liability unless the alleged constitutional violation was a but-for cause of the employment termination. Ibid.; see also Umbehr, supra, at 675. The City resists the applicability of the Mt. Healthy test as the sole determinant here. It contends that, where there was probable cause for the arrest, the applicable precedent is Hartman—a case that was in the criminal sphere and that turned on the existence of probable cause. The background in Hartman was that a company and its chief executive, William Moore, had engaged in an extensive lobbying and governmental relations campaign opposing a particular postal service policy. 547 U. S., at 252–253. Moore and the company were later prosecuted for violating federal statutes in the course of that lobbying. Id., at 253–254. After being acquitted, Moore filed suit against five postal inspectors, alleging that they had violated his First Amendment rights when they instigated his prosecution in retaliation for his criticisms of the Postal Service. Id., at 254. This Court held that a plaintiff alleging a retaliatory prosecution must show the absence of probable cause for the underlying criminal charge. Id., at 265–266. If there was probable cause, the case ends. If the plaintiff proves the absence of probable cause, then the Mt. Healthy test governs: The plaintiff must show that the retaliation was a substantial or motivating factor behind the prosecution, and, if that showing is made, the defendant can prevail only by showing that the prosecution would have been initiated without respect to retaliation. See 547 U. S., at 265–266. The Court in Hartman deemed it necessary to inquire as to the existence of probable cause because proving the link between the defendant’s retaliatory animus and the plaintiff’s injury in retaliatory prosecution cases “is usually more complex than it is in other retaliation cases.” Id., at 261. An action for retaliatory prosecution “will not be brought against the prosecutor, who is absolutely immune from liability for the decision to prosecute.” Id., at 261–262. Instead, the plaintiff must sue some other government official and prove that the official “induced the prosecutor to bring charges that would not have been initiated without his urging.” Id., at 262. Noting that inquiries with respect to probable cause are commonplace in criminal cases, the Court determined that requiring plaintiffs in retaliatory prosecution cases to prove the lack of probable cause would help “bridge the gap between the nonprosecuting government agent’s motive and the prosecutor’s action.” Id., at 263. The City’s argument here is that, just as probable cause is a bar in retaliatory prosecution cases, so too should it be a bar in this case, involving a retaliatory arrest. There is undoubted force in the City’s position. Reichle, 566 U. S., at 667–668. There are on average about 29,000 arrests per day in this country. Dept. of Justice–FBI, Uniform Crime Report, Crime in the United States, 2016 (Fall 2017). In deciding whether to arrest, police officers often make split-second judgments. The content of the suspect’s speech might be a consideration in circumstances where the officer must decide whether the suspect is ready to cooperate, or, on the other hand, whether he may present a continuing threat to interests that the law must protect. See, e.g., District of Columbia v. Wesby, 583 U. S. ___, ___ (2018) (slip op., at 10) (“suspect’s untruthful and evasive answers to police questioning could support probable cause” (internal quotation marks omitted)). For these reasons retaliatory arrest claims, much like retaliatory prosecution claims, can “present a tenuous causal connection between the defendant’s alleged animus and the plaintiff’s injury.” Reichle, 566 U. S., at 668. That means it can be difficult to discern whether an arrest was caused by the officer’s legitimate or illegitimate consideration of speech. Ibid. And the complexity of proving (or disproving) causation in these cases creates a risk that the courts will be flooded with dubious retaliatory arrest suits. See Brief for District of Columbia et al. as Amici Curiae 5–11. At the same time, there are substantial arguments that Hartman’s framework is inapt in retaliatory arrest cases, and that Mt. Healthy should apply without a threshold inquiry into probable cause. For one thing, the causation problem in retaliatory arrest cases is not the same as the problem identified in Hartman. Hartman relied in part on the fact that, in retaliatory prosecution cases, the causal connection between the defendant’s animus and the prosecutor’s decision to prosecute is weakened by the “presumption of regularity accorded to prosecutorial decisionmaking.” 547 U. S., at 263. That presumption does not apply in this context. See Reichle, supra, at 669. In addition, there is a risk that some police officers may exploit the arrest power as a means of suppressing speech. See Brief for Institute for Free Speech as Amicus Curiae. IV The parties’ arguments raise difficult questions about the scope of First Amendment protections when speech is made in connection with, or contemporaneously to, criminal activity. But whether in a retaliatory arrest case the Hartman approach should apply, thus barring a suit where probable cause exists, or, on the other hand, the inquiry should be governed only by Mt. Healthy is a determination that must await a different case. For Lozman’s claim is far afield from the typical retaliatory arrest claim, and the difficulties that might arise if Mt. Healthy is applied to the mine run of arrests made by police offi- cers are not present here. Here Lozman does not sue the officer who made the arrest. Indeed, Lozman likely could not have maintained a retaliation claim against the arresting officer in these circumstances, because the officer appears to have acted in good faith, and there is no showing that the officer had any knowledge of Lozman’s prior speech or any motive to arrest him for his earlier expressive activities. Instead Lozman alleges more governmental action than simply an arrest. His claim is that the City itself retali- ated against him pursuant to an “official municipal policy” of intimidation. Monell, 436 U. S., at 691. In particular, he alleges that the City, through its legislators, formed a premeditated plan to intimidate him in retaliation for his criticisms of city officials and his open-meetings lawsuit. And he asserts that the City itself, through the same high officers, executed that plan by ordering his arrest at the November 2006 city council meeting. The fact that Lozman must prove the existence and enforcement of an official policy motivated by retaliation separates Lozman’s claim from the typical retaliatory arrest claim. An official retaliatory policy is a particularly troubling and potent form of retaliation, for a policy can be long term and pervasive, unlike an ad hoc, on-the-spot decision by an individual officer. An official policy also can be difficult to dislodge. A citizen who suffers retaliation by an individual officer can seek to have the officer disciplined or removed from service, but there may be little practical recourse when the government itself orchestrates the retaliation. For these reasons, when retaliation against protected speech is elevated to the level of official policy, there is a compelling need for adequate avenues of redress. In addition, Lozman’s allegations, if proved, alleviate the problems that the City says will result from applying Mt. Healthy in retaliatory arrest cases. The causation problem in arrest cases is not of the same difficulty where, as is alleged here, the official policy is retaliation for prior, protected speech bearing little relation to the criminal offense for which the arrest is made. In determining whether there was probable cause to arrest Lozman for disrupting a public assembly, it is difficult to see why a city official could have legitimately considered that Lozman had, months earlier, criticized city officials or filed a lawsuit against the City. So in a case like this one it is unlikely that the connection between the alleged animus and injury will be “weakened . . . by [an official’s] legitimate consideration of speech.” Reichle, 566 U. S., at 668. This unique class of retaliatory arrest claims, moreover, will require objective evidence of a policy motivated by retaliation to survive summary judgment. Lozman, for instance, cites a transcript of a closed-door city council meeting and a video recording of his arrest. There is thus little risk of a flood of retaliatory arrest suits against high-level policymakers. As a final matter, it must be underscored that this Court has recognized the “right to petition as one of the most precious of the liberties safeguarded by the Bill of Rights.” BE&K Constr. Co. v. NLRB, 536 U. S. 516, 524 (2002) (internal quotation marks omitted). Lozman alleges the City deprived him of this liberty by retaliating against him for his lawsuit against the City and his criticisms of public officials. Thus, Lozman’s speech is high in the hierarchy of First Amendment values. See Connick v. Myers, 461 U. S. 138, 145 (1983). For these reasons, Lozman need not prove the absence of probable cause to maintain a claim of retaliatory arrest against the City. On facts like these, Mt. Healthy provides the correct standard for assessing a retaliatory arrest claim. The Court need not, and does not, address the elements required to prove a retaliatory arrest claim in other contexts. This is not to say, of course, that Lozman is ultimately entitled to relief or even a new trial. On remand, the Court of Appeals, applying Mt. Healthy and other relevant precedents, may consider any arguments in support of the District Court’s judgment that have been preserved by the City. Among other matters, the Court of Appeals may wish to consider (1) whether any reasonable juror could find that the City actually formed a retaliatory policy to intimidate Lozman during its June 2006 closed-door session; (2) whether any reasonable juror could find that the November 2006 arrest constituted an official act by the City; and (3) whether, under Mt. Healthy, the City has proved that it would have arrested Lozman regardless of any retaliatory animus—for example, if Lozman’s conduct during prior city council meetings had also violated valid rules as to proper subjects of discussion, thus explaining his arrest here. For these reasons, the judgment of the Court of Appeals is vacated, and the case is remanded for further proceedings consistent with this opinion. It is so ordered.

585.US.2017_17-130

The Securities and Exchange Commission (SEC or Commission) has statutory authority to enforce the nation’s securities laws. One way it can do so is by instituting an administrative proceeding against an alleged wrongdoer. Typically, the Commission delegates the task of presiding over such a proceeding to an administrative law judge (ALJ). The SEC currently has five ALJs. Other staff members, rather than the Commission proper, selected them all. An ALJ assigned to hear an SEC enforcement action has the “authority to do all things necessary and appropriate” to ensure a “fair and orderly” adversarial proceeding. 17 CFR §§201.111, 200.14(a). After a hearing ends, the ALJ issues an initial decision. The Commission can review that decision, but if it opts against review, it issues an order that the initial decision has become final. See §201.360(d). The initial decision is then “deemed the action of the Commission.” 15 U. S. C. §78d–1(c). The SEC charged petitioner Raymond Lucia with violating certain securities laws and assigned ALJ Cameron Elliot to adjudicate the case. Following a hearing, Judge Elliot issued an initial decision concluding that Lucia had violated the law and imposing sanctions. On appeal to the SEC, Lucia argued that the administrative proceeding was invalid because Judge Elliot had not been constitutionally appointed. According to Lucia, SEC ALJs are “Officers of the United States” and thus subject to the Appointments Clause. Under that Clause, only the President, “Courts of Law,” or “Heads of Departments” can appoint such “Officers.” But none of those actors had made Judge Elliot an ALJ. The SEC and the Court of Appeals for the D. C. Circuit rejected Lucia’s argument, holding that SEC ALJs are not “Officers of the United States,” but are instead mere employees—officials with lesser responsibilities who are not subject to the Appointments Clause. Held: The Commission’s ALJs are “Officers of the United States,” subject to the Appointments Clause. Pp. 5–13. (a) This Court’s decisions in United States v. Germaine, 99 U. S. 508, and Buckley v. Valeo, 424 U. S. 1, set out the basic framework for distinguishing between officers and employees. To qualify as an officer, rather than an employee, an individual must occupy a “continuing” position established by law, Germaine, 99 U. S., at 511, and must “exercis[e] significant authority pursuant to the laws of the United States,” Buckley, 424 U. S., at 126. In Freytag v. Commissioner, 501 U. S. 868, the Court applied this framework to “special trial judges” (STJs) of the United States Tax Court. STJs could issue the final decision of the Tax Court in “comparatively narrow and minor matters.” Id., at 873. In more major matters, they could preside over the hearing but could not issue a final decision. Instead, they were to “prepare proposed findings and an opinion” for a regular Tax Court judge to consider. Ibid. The proceeding challenged in Freytag was a major one. The losing parties argued on appeal that the STJ who presided over their hearing was not constitutionally appointed. This Court held that STJs are officers. Citing Germaine, the Freytag Court first found that STJs hold a continuing office established by law. See 501 U. S., at 881. The Court then considered, as Buckley demands, the “significance” of the “authority” STJs wield. 501 U. S., at 881. The Government had argued that STJs are employees in all cases in which they could not enter a final decision. But the Court thought that the Government’s focus on finality “ignore[d] the significance of the duties and discretion that [STJs] possess.” Ibid. Describing the responsibilities involved in presiding over adversarial hearings, the Court said: STJs “take testimony, conduct trials, rule on the admissibility of evidence, and have the power to enforce compliance with discovery orders.” Id., at 881–882. And the Court observed that “[i]n the course of carrying out these important functions,” STJs “exercise significant discretion.” Id., at 882. Freytag’s analysis decides this case. The Commission’s ALJs, like the Tax Court’s STJs, hold a continuing office established by law. SEC ALJs “receive[ ] a career appointment,” 5 CFR §930.204(a), to a position created by statute, see 5 U. S. C. §§556–557, 5372, 3105. And they exercise the same “significant discretion” when carrying out the same “important functions” as STJs do. Freytag, 501 U. S., at 882. Both sets of officials have all the authority needed to ensure fair and orderly adversarial hearings—indeed, nearly all the tools of federal trial judges. The Commission’s ALJs, like the Tax Court’s STJs, “take testimony,” “conduct trials,” “rule on the admissibility of evidence,” and “have the power to enforce compliance with discovery orders.” Id., at 881–882. So point for point from Freytag’s list, SEC ALJs have equivalent duties and powers as STJs in conducting adversarial inquiries. Moreover, at the close of those proceedings, SEC ALJs issue decisions much like that in Freytag. STJs prepare proposed findings and an opinion adjudicating charges and assessing tax liabilities. Similarly, the Commission’s ALJs issue initial decisions containing factual findings, legal conclusions, and appropriate remedies. And what happens next reveals that the ALJ can play the more autonomous role. In a major Tax Court case, a regular Tax Court judge must always review an STJ’s opinion, and that opinion comes to nothing unless the regular judge adopts it. By contrast, the SEC can decide against reviewing an ALJ’s decision, and when it does so the ALJ’s decision itself “becomes final” and is “deemed the action of the Commission.” 17 CFR §201.360(d)(2); 15 U. S. C. §78d–1(c). Pp. 5–11. (b) Judge Elliot heard and decided Lucia’s case without a constitutional appointment. “[O]ne who makes a timely challenge to the constitutional validity of the appointment of an officer who adjudicates his case” is entitled to relief. Ryder v. United States, 515 U. S. 177, 182. Lucia made just such a timely challenge. And the “appropriate” remedy for an adjudication tainted with an appointments violation is a new “hearing before a properly appointed” official. Id., at 183, 188. In this case, that official cannot be Judge Elliot, even if he has by now received a constitutional appointment. Having already both heard Lucia’s case and issued an initial decision on the merits, he cannot be expected to consider the matter as though he had not adjudicated it before. To cure the constitutional error, another ALJ (or the Commission itself) must hold the new hearing. Pp. 12–13. 868 F. 3d 1021, reversed and remanded. Kagan, J., delivered the opinion of the Court, in which Roberts, C. J., and Kennedy, Thomas, Alito, and Gorsuch, JJ., joined. Thomas, J., filed a concurring opinion, in which Gorsuch, J., joined. Breyer, J., filed an opinion concurring in the judgment in part and dissenting in part, in which Ginsburg and Sotomayor, JJ., joined as to Part III. Sotomayor, J., filed a dissenting opinion, in which Ginsburg, J., joined.

The Appointments Clause of the Constitution lays out the permissible methods of appointing “Officers of the United States,” a class of government officials distinct from mere employees. Art. II, §2, cl. 2. This case requires us to decide whether administrative law judges (ALJs) of the Securities and Exchange Commission (SEC or Commission) qualify as such “Officers.” In keeping with Freytag v. Commissioner, 501 U. S. 868 (1991), we hold that they do. I The SEC has statutory authority to enforce the nation’s securities laws. One way it can do so is by instituting an administrative proceeding against an alleged wrongdoer. By law, the Commission may itself preside over such a proceeding. See 17 CFR §201.110 (2017). But the Commission also may, and typically does, delegate that task to an ALJ. See ibid.; 15 U. S. C. §78d–1(a). The SEC currently has five ALJs. Other staff members, rather than the Commission proper, selected them all. See App. to Pet. for Cert. 295a–297a. An ALJ assigned to hear an SEC enforcement action has extensive powers—the “authority to do all things necessary and appropriate to discharge his or her duties” and ensure a “fair and orderly” adversarial proceeding. §§201.111, 200.14(a). Those powers “include, but are not limited to,” supervising discovery; issuing, revoking, or modifying subpoenas; deciding motions; ruling on the admissibility of evidence; administering oaths; hearing and examining witnesses; generally “[r]egulating the course of” the proceeding and the “conduct of the parties and their counsel”; and imposing sanctions for “[c]ontemptuous conduct” or violations of procedural requirements. §§201.111, 201.180; see §§200.14(a), 201.230. As that list suggests, an SEC ALJ exercises authority “comparable to” that of a federal district judge conducting a bench trial. Butz v. Economou, 438 U. S. 478, 513 (1978). After a hearing ends, the ALJ issues an “initial decision.” §201.360(a)(1). That decision must set out “findings and conclusions” about all “material issues of fact [and] law”; it also must include the “appropriate order, sanction, relief, or denial thereof.” §201.360(b). The Commission can then review the ALJ’s decision, either upon request or sua sponte. See §201.360(d)(1). But if it opts against review, the Commission “issue[s] an order that the [ALJ’s] decision has become final.” §201.360(d)(2). At that point, the initial decision is “deemed the action of the Commission.” §78d–1(c). This case began when the SEC instituted an administrative proceeding against petitioner Raymond Lucia and his investment company. Lucia marketed a retirement savings strategy called “Buckets of Money.” In the SEC’s view, Lucia used misleading slideshow presentations to deceive prospective clients. The SEC charged Lucia under the Investment Advisers Act, §80b–1 et seq., and assigned ALJ Cameron Elliot to adjudicate the case. After nine days of testimony and argument, Judge Elliot issued an initial decision concluding that Lucia had violated the Act and imposing sanctions, including civil penalties of $300,000 and a lifetime bar from the investment industry. In his decision, Judge Elliot made factual findings about only one of the four ways the SEC thought Lucia’s slideshow misled investors. The Commission thus remanded for factfinding on the other three claims, explaining that an ALJ’s “personal experience with the witnesses” places him “in the best position to make findings of fact” and “resolve any conflicts in the evidence.” App. to Pet. for Cert. 241a. Judge Elliot then made additional findings of deception and issued a revised initial decision, with the same sanctions. See id., at 118a. On appeal to the SEC, Lucia argued that the administrative proceeding was invalid because Judge Elliot had not been constitutionally appointed. According to Lucia, the Commission’s ALJs are “Officers of the United States” and thus subject to the Appointments Clause. Under that Clause, Lucia noted, only the President, “Courts of Law,” or “Heads of Departments” can appoint “Officers.” See Art. II, §2, cl. 2. And none of those actors had made Judge Elliot an ALJ. To be sure, the Commission itself counts as a “Head[ ] of Department[ ].” Ibid.; see Free Enterprise Fund v. Public Company Accounting Oversight Bd., 561 U. S. 477, 511–513 (2010). But the Commission had left the task of appointing ALJs, including Judge Elliot, to SEC staff members. See supra, at 1. As a result, Lucia contended, Judge Elliot lacked constitutional authority to do his job. The Commission rejected Lucia’s argument. It held that the SEC’s ALJs are not “Officers of the United States.” Instead, they are “mere employees”—officials with lesser responsibilities who fall outside the Appointments Clause’s ambit. App. to Pet. for Cert. 87a. The Commission reasoned that its ALJs do not “exercise significant authority independent of [its own] supervision.” Id., at 88a. Because that is so (said the SEC), they need no special, high-level appointment. See id., at 86a. Lucia’s claim fared no better in the Court of Appeals for the D. C. Circuit. A panel of that court seconded the Commission’s view that SEC ALJs are employees rather than officers, and so are not subject to the Appointments Clause. See 832 F.3d 277, 283–289 (2016). Lucia then petitioned for rehearing en banc. The Court of Appeals granted that request and heard argument in the case. But the ten members of the en banc court divided evenly, resulting in a per curiam order denying Lucia’s claim. See 868 F.3d 1021 (2017). That decision conflicted with one from the Court of Appeals for the Tenth Circuit. See Bandimere v. SEC, 844 F.3d 1168, 1179 (2016). Lucia asked us to resolve the split by deciding whether the Commission’s ALJs are “Officers of the United States within the meaning of the Appointments Clause.” Pet. for Cert. i. Up to that point, the Federal Government (as represented by the Department of Justice) had defended the Commission’s position that SEC ALJs are employees, not officers. But in responding to Lucia’s petition, the Government switched sides.[1] So when we granted the petition, 583 U. S. ___ (2018), we also appointed an amicus curiae to defend the judgment below.[2] We now reverse. II The sole question here is whether the Commission’s ALJs are “Officers of the United States” or simply employees of the Federal Government. The Appointments Clause prescribes the exclusive means of appointing “Officers.” Only the President, a court of law, or a head of department can do so. See Art. II, §2, cl. 2.[3] And as all parties agree, none of those actors appointed Judge Elliot before he heard Lucia’s case; instead, SEC staff members gave him an ALJ slot. See Brief for Petitioners 15; Brief for United States 38; Brief for Court-Appointed Amicus Cu- riae 21. So if the Commission’s ALJs are constitutional officers, Lucia raises a valid Appointments Clause claim. The only way to defeat his position is to show that those ALJs are not officers at all, but instead non-officer employees—part of the broad swath of “lesser functionaries” in the Government’s workforce. Buckley v. Valeo, 424 U. S. 1, 126, n. 162 (1976) (per curiam). For if that is true, the Appointments Clause cares not a whit about who named them. See United States v. Germaine, 99 U. S. 508, 510 (1879). Two decisions set out this Court’s basic framework for distinguishing between officers and employees. Germaine held that “civil surgeons” (doctors hired to perform various physical exams) were mere employees because their duties were “occasional or temporary” rather than “continuing and permanent.” Id., at 511–512. Stressing “ideas of tenure [and] duration,” the Court there made clear that an individual must occupy a “continuing” position established by law to qualify as an officer. Id., at 511. Buckley then set out another requirement, central to this case. It determined that members of a federal commission were officers only after finding that they “exercis[ed] significant authority pursuant to the laws of the United States.” 424 U. S., at 126. The inquiry thus focused on the extent of power an individual wields in carrying out his assigned functions. Both the amicus and the Government urge us to elaborate on Buckley’s “significant authority” test, but another of our precedents makes that project unnecessary. The standard is no doubt framed in general terms, tempting advocates to add whatever glosses best suit their arguments. See Brief for Amicus Curiae 14 (contending that an individual wields “significant authority” when he has “(i) the power to bind the government or private parties (ii) in her own name rather than in the name of a superior officer”); Reply Brief for United States 2 (countering that an individual wields that authority when he has “the power to bind the government or third parties on significant matters” or to undertake other “important and distinctively sovereign functions”). And maybe one day we will see a need to refine or enhance the test Buckley set out so concisely. But that day is not this one, because in Freytag v. Commissioner, 501 U. S. 868 (1991), we applied the unadorned “significant authority” test to adjudicative officials who are near-carbon copies of the Commission’s ALJs. As we now explain, our analysis there (sans any more detailed legal criteria) necessarily decides this case. The officials at issue in Freytag were the “special trial judges” (STJs) of the United States Tax Court. The authority of those judges depended on the significance of the tax dispute before them. In “comparatively narrow and minor matters,” they could both hear and definitively resolve a case for the Tax Court. Id., at 873. In more major matters, they could preside over the hearing, but could not issue the final decision; instead, they were to “prepare proposed findings and an opinion” for a regular Tax Court judge to consider. Ibid. The proceeding challenged in Freytag was a major one, involving $1.5 billion in alleged tax deficiencies. See id., at 871, n. 1. After conducting a 14-week trial, the STJ drafted a proposed decision in favor of the Government. A regular judge then adopted the STJ’s work as the opinion of the Tax Court. See id., at 872. The losing parties argued on appeal that the STJ was not constitutionally appointed. This Court held that the Tax Court’s STJs are officers, not mere employees. Citing Germaine, the Court first found that STJs hold a continuing office established by law. See 501 U. S., at 881. They serve on an ongoing, rather than a “temporary [or] episodic[,] basis”; and their “duties, salary, and means of appointment” are all specified in the Tax Code. Ibid. The Court then considered, as Buckley demands, the “significance” of the “authority” STJs wield. 501 U. S., at 881. In addressing that issue, the Government had argued that STJs are employees, rather than officers, in all cases (like the one at issue) in which they could not “enter a final decision.” Ibid. But the Court thought the Government’s focus on finality “ignore[d] the significance of the duties and discretion that [STJs] possess.” Ibid. Describing the responsibilities involved in presiding over adversarial hearings, the Court said: STJs “take testimony, conduct trials, rule on the admissibility of evidence, and have the power to enforce compliance with discovery orders.” Id., at 881–882. And the Court observed that “[i]n the course of carrying out these important functions, the [STJs] exercise significant discretion.” Id., at 882. That fact meant they were offi- cers, even when their decisions were not final.[4] Freytag says everything necessary to decide this case. To begin, the Commission’s ALJs, like the Tax Court’s STJs, hold a continuing office established by law. See id., at 881. Indeed, everyone here—Lucia, the Government, and the amicus—agrees on that point. See Brief for Petitioners 21; Brief for United States 17–18, n. 3; Brief for Amicus Curiae 22, n. 7. Far from serving temporarily or episodically, SEC ALJs “receive[ ] a career appointment.” 5 CFR §930.204(a) (2018). And that appointment is to a position created by statute, down to its “duties, salary, and means of appointment.” Freytag, 501 U. S., at 881; see 5 U. S. C. §§556–557, 5372, 3105. Still more, the Commission’s ALJs exercise the same “significant discretion” when carrying out the same “important functions” as STJs do. Freytag, 501 U. S., at 882. Both sets of officials have all the authority needed to ensure fair and orderly adversarial hearings—indeed, nearly all the tools of federal trial judges. See Butz, 438 U. S., at 513; supra, at 2. Consider in order the four specific (if overlapping) powers Freytag mentioned. First, the Commission’s ALJs (like the Tax Court’s STJs) “take testimony.” 501 U. S., at 881. More precisely, they “[r]eceiv[e] evidence” and “[e]xamine witnesses” at hearings, and may also take pre-hearing depositions. 17 CFR §§201.111(c), 200.14(a)(4); see 5 U. S. C. §556(c)(4). Second, the ALJs (like STJs) “conduct trials.” 501 U. S., at 882. As detailed earlier, they administer oaths, rule on motions, and generally “regulat[e] the course of” a hearing, as well as the conduct of parties and counsel. §201.111; see §§200.14(a)(1), (a)(7); supra, at 2. Third, the ALJs (like STJs) “rule on the admissibility of evidence.” 501 U. S., at 882; see §201.111(c). They thus critically shape the administrative record (as they also do when issuing document subpoenas). See §201.111(b). And fourth, the ALJs (like STJs) “have the power to enforce compliance with discovery orders.” 501 U. S., at 882. In particular, they may punish all “[c]ontemptuous conduct,” including violations of those orders, by means as severe as excluding the offender from the hearing. See §201.180(a)(1). So point for point—straight from Freytag’s list—the Commission’s ALJs have equivalent duties and powers as STJs in conducting adversarial inquiries. And at the close of those proceedings, ALJs issue decisions much like that in Freytag—except with potentially more independent effect. As the Freytag Court recounted, STJs “prepare proposed findings and an opinion” adjudicating charges and assessing tax liabilities. 501 U. S., at 873; see supra, at 7. Similarly, the Commission’s ALJs issue decisions containing factual findings, legal conclusions, and appropriate remedies. See §201.360(b); supra, at 2. And what happens next reveals that the ALJ can play the more autonomous role. In a major case like Freytag, a regular Tax Court judge must always review an STJ’s opinion. And that opinion counts for nothing unless the regular judge adopts it as his own. See 501 U. S., at 873. By contrast, the SEC can decide against reviewing an ALJ decision at all. And when the SEC declines review (and issues an order saying so), the ALJ’s decision itself “becomes final” and is “deemed the action of the Commission.” §201.360(d)(2); 15 U. S. C. §78d–1(c); see supra, at 2. That last-word capacity makes this an a fortiori case: If the Tax Court’s STJs are officers, as Freytag held, then the Commission’s ALJs must be too. The amicus offers up two distinctions to support the opposite conclusion. His main argument relates to “the power to enforce compliance with discovery orders”—the fourth of Freytag’s listed functions. 501 U. S., at 882. The Tax Court’s STJs, he states, had that power “because they had authority to punish contempt” (including discovery violations) through fines or imprisonment. Brief for Amicus Curiae 37; see id., at 37, n. 10 (citing 26 U. S. C. §7456(c)). By contrast, he observes, the Commission’s ALJs have less capacious power to sanction misconduct. The amicus’s secondary distinction involves how the Tax Court and Commission, respectively, review the factfinding of STJs and ALJs. The Tax Court’s rules state that an STJ’s findings of fact “shall be presumed” correct. Tax Court Rule 183(d). In comparison, the amicus notes, the SEC’s regulations include no such deferential standard. See Brief for Amicus Curiae 10, 38, n. 11. But those distinctions make no difference for officer status. To start with the amicus’s primary point, Freytag referenced only the general “power to enforce compliance with discovery orders,” not any particular method of doing so. 501 U. S., at 882. True enough, the power to toss malefactors in jail is an especially muscular means of enforcement—the nuclear option of compliance tools. But just as armies can often enforce their will through conventional weapons, so too can administrative judges. As noted earlier, the Commission’s ALJs can respond to discovery violations and other contemptuous conduct by excluding the wrongdoer (whether party or lawyer) from the proceedings—a powerful disincentive to resist a court order. See §201.180(a)(1)(i); supra, at 9. Similarly, if the offender is an attorney, the ALJ can “[s]ummarily suspend” him from representing his client—not something the typical lawyer wants to invite. §201.180(a)(1)(ii). And finally, a judge who will, in the end, issue an opinion complete with factual findings, legal conclusions, and sanctions has substantial informal power to ensure the parties stay in line. Contrary to the amicus’s view, all that is enough to satisfy Freytag’s fourth item (even supposing, which we do not decide, that each of those items is necessary for someone conducting adversarial hearings to count as an officer). And the amicus’s standard-of-review distinction fares just as badly. The Freytag Court never suggested that the deference given to STJs’ factual findings mattered to its Appointments Clause analysis. Indeed, the relevant part of Freytag did not so much as mention the subject (even though it came up at oral argument, see Tr. of Oral Arg. 33–41). And anyway, the Commission often accords a similar deference to its ALJs, even if not by regulation. The Commission has repeatedly stated, as it did below, that its ALJs are in the “best position to make findings of fact” and “resolve any conflicts in the evidence.” App. to Pet. for Cert. 241a (quoting In re Nasdaq Stock Market, LLC, SEC Release No. 57741 (Apr. 30, 2008)). (That was why the SEC insisted that Judge Elliot make factual findings on all four allegations of Lucia’s deception. See supra, at 3.) And when factfinding derives from credibility judgments, as it frequently does, acceptance is near-automatic. Recognizing ALJs’ “personal experience with the witnesses,” the Commission adopts their “credibility finding[s] absent overwhelming evidence to the contrary.” App. to Pet. for Cert. 241a; In re Clawson, SEC Release No. 48143 (July 9, 2003). That practice erases the constitutional line the amicus proposes to draw. The only issue left is remedial. For all the reasons we have given, and all those Freytag gave before, the Commission’s ALJs are “Officers of the United States,” subject to the Appointments Clause. And as noted earlier, Judge Elliot heard and decided Lucia’s case without the kind of appointment the Clause requires. See supra, at 5. This Court has held that “one who makes a timely challenge to the constitutional validity of the appointment of an officer who adjudicates his case” is entitled to relief. Ryder v. United States, 515 U. S. 177, 182–183 (1995). Lucia made just such a timely challenge: He contested the validity of Judge Elliot’s appointment before the Commission, and continued pressing that claim in the Court of Appeals and this Court. So what relief follows? This Court has also held that the “appropriate” remedy for an adjudication tainted with an appointments violation is a new “hearing before a properly appointed” official. Id., at 183, 188. And we add today one thing more. That official cannot be Judge Elliot, even if he has by now received (or receives sometime in the future) a constitutional appointment. Judge Elliot has already both heard Lucia’s case and issued an initial decision on the merits. He cannot be expected to consider the matter as though he had not adjudicated it before.[5] To cure the constitutional error, another ALJ (or the Commission itself) must hold the new hearing to which Lucia is entitled.[6] We accordingly reverse the judgment of the Court of Appeals and remand the case for further proceedings consistent with this opinion. It is so ordered. Notes 1 In the same certiorari-stage brief, the Government asked us to add a second question presented: whether the statutory restrictions on removing the Commission’s ALJs are constitutional. See Brief in Response 21. When we granted certiorari, we chose not to take that step. See 583 U. S. ___ (2018). The Government’s merits brief now asks us again to address the removal issue. See Brief for United States 39–55. We once more decline. No court has addressed that question, and we ordinarily await “thorough lower court opinions to guide our analysis of the merits.” Zivotofsky v. Clinton, 566 U. S. 189, 201 (2012). 2 We appointed Anton Metlitsky to brief and argue the case, 583 U. S. ___ (2018), and he has ably discharged his responsibilities. 3 That statement elides a distinction, not at issue here, between “principal” and “inferior” officers. See Edmond v. United States, 520 U. S. 651, 659–660 (1997). Only the President, with the advice and consent of the Senate, can appoint a principal officer; but Congress (instead of relying on that method) may authorize the President alone, a court, or a department head to appoint an inferior officer. See ibid. Both the Government and Lucia view the SEC’s ALJs as inferior officers and acknowledge that the Commission, as a head of department, can constitutionally appoint them. See Brief for United States 38; Brief for Petitioners 50–51. 4 The Court also provided an alternative basis for viewing the STJs as officers. “Even if the duties of [STJs in major cases] were not as significant as we . . . have found them,” we stated, “our conclusion would be unchanged.” Freytag, 501 U. S., at 882. That was because the Government had conceded that in minor matters, where STJs could enter final decisions, they had enough “independent authority” to count as officers. Ibid. And we thought it made no sense to classify the STJs as officers for some cases and employees for others. See ibid. Justice Sotomayor relies on that back-up rationale in trying to reconcile Freytag with her view that “a prerequisite to officer status is the authority” to issue at least some “final decisions.” Post, at 5 (dissenting opinion). But Freytag has two parts, and its primary analysis explicitly rejects Justice Sotomayor’s theory that final decisionmaking authority is a sine qua non of officer status. See 501 U. S., at 881–882. As she acknowledges, she must expunge that reasoning to make her reading work. See post, at 5 (“That part of the opinion[ ] was unnecessary to the result”). 5 Justice Breyer disagrees with our decision to wrest further proceedings from Judge Elliot, arguing that “[f]or him to preside once again would not violate the structural purposes [of] the Appointments Clause.” Post, at 13 (opinion concurring in judgment in part and dissenting in part). But our Appointments Clause remedies are designed not only to advance those purposes directly, but also to create “[ ]incentive[s] to raise Appointments Clause challenges.” Ryder v. United States, 515 U. S. 177, 183 (1995). We best accomplish that goal by providing a successful litigant with a hearing before a new judge. That is especially so because (as Justice Breyer points out) the old judge would have no reason to think he did anything wrong on the merits, see post, at 13—and so could be expected to reach all the same judgments. But we do not hold that a new officer is required for every Appointments Clause violation. As Justice Breyer suggests, we can give that remedy here because other ALJs (and the Commission) are available to hear this case on remand. See ibid. If instead the Appointments Clause problem is with the Commission itself, so that there is no substitute decisionmaker, the rule of necessity would presumably kick in and allow the Commission to do the rehearing. See FTC v. Cement Institute, 333 U. S. 683, 700–703 (1948); 3 K. Davis, Administrative Law Treatise §19.9 (2d ed. 1980). 6 While this case was on judicial review, the SEC issued an order “ratif[ying]” the prior appointments of its ALJs. Order (Nov. 30, 2017), online at https://www.sec.gov/litigation/opinions/2017/33-10440.pdf (as last visited June 18, 2018). Lucia argues that the order is invalid. See Brief for Petitioners 50–56. We see no reason to address that issue. The Commission has not suggested that it intends to assign Lucia’s case on remand to an ALJ whose claim to authority rests on the ratification order. The SEC may decide to conduct Lucia’s rehearing itself. Or it may assign the hearing to an ALJ who has received a constitutional appointment independent of the ratification.

584.US.2017_16-1144

Between 2004 and 2009, the Internal Revenue Service (IRS) intermittently investigated petitioner Marinello’s tax activities. In 2012, the Government indicted Marinello for violating, among other criminal tax statutes, a provision in 26 U. S. C. §7212(a) known as the Omnibus Clause, which forbids “corruptly or by force or threats of force . . . obstruct[ing] or imped[ing], or endeavor[ing] to obstruct or impede, the due administration of [the Internal Revenue Code].” The judge instructed the jury that, to convict Marinello of an Omnibus Clause violation, it must find that he “corruptly” engaged in at least one of eight specified activities, but the jury was not told that it needed to find that Marinello knew he was under investigation and intended corruptly to interfere with that investigation. Marinello was convicted. The Second Circuit affirmed, rejecting his claim that an Omnibus Clause violation requires the Government to show the defendant tried to interfere with a pending IRS proceeding, such as a particular investigation. Held: To convict a defendant under the Omnibus Clause, the Government must prove the defendant was aware of a pending tax-related proceeding, such as a particular investigation or audit, or could reasonably foresee that such a proceeding would commence. Pp. 3–11. (a) In United States v. Aguilar, 515 U. S. 593 , this Court interpreted a similarly worded criminal statute—which made it a felony “corruptly or by threats or force . . . [to] influenc[e], obstruc[t], or imped[e], or endeavo[r] to influence, obstruct, or impede, the due administration of justice,” 18 U. S. C. §1503(a). There, the Court required the Government to show there was a “nexus” between the defendant’s obstructive conduct and a particular judicial proceeding. The Court said that the defendant’s “act must have a relationship in time, causation, or logic with the judicial proceedings.” 515 U. S., at 599. In reaching this conclusion, the Court emphasized that it has “traditionally exercised restraint in assessing the reach of a federal criminal statute, both out of deference to the prerogatives of Congress and out of concern that ‘a fair warning should be given to the world in language that the common world will understand, of what the law intends to do if a certain line is passed.’ ” Id., at 600. That reasoning applies here with similar strength. The verbs “obstruct” and “impede” require an object. The taxpayer must hinder a particular person or thing. The object in §7212(a) is the “due administration of [the Tax Code].” That phrase is best viewed, like the “due administration of justice” in Aguilar, as referring to discrete targeted administrative acts rather than every conceivable task involved in the Tax Code’s administration. Statutory context confirms this reading. The Omnibus Clause appears in the middle of a sentence that refers to efforts to “intimidate or impede any officer or employee of the United States acting in an official capacity.” §7212(a). The first part of the sentence also refers to “force or threats of force,” which the statute elsewhere defines as “threats of bodily harm to the officer or employee of the United States or to a member of his family.” Ibid. And §7212(b) refers to the “forcibl[e] rescu[e]” of “any property after it shall have been seized under” the Internal Revenue Code. Subsections (a) and (b) thus refer to corrupt or forceful actions taken against individual identifiable persons or property. In context, the Omnibus Clause logically serves as a “catchall” for the obstructive conduct the subsection sets forth, not for every violation that interferes with routine administrative procedures such as the processing of tax returns, receipt of tax payments, or issuance of tax refunds. The statute’s legislative history does not suggest otherwise. The broader context of the full Internal Revenue Code also counsels against a broad reading. Interpreting the Omnibus Clause to apply to all Code administration could transform the Code’s numerous misdemeanor provisions into felonies, making them redundant or perhaps the subject matter of plea bargaining. It could also result in a similar lack of fair warning and related kinds of unfairness that led this Court to “exercise” interpretive “restraint” in Aguilar. See 515 U. S., at 600. The Government claims that the “corrupt state of mind” requirement will cure any overbreadth problem, but it is difficult to imagine a scenario when that requirement will make a practical difference in the context of federal tax prosecutions. And to rely on prosecutorial discretion to narrow the otherwise wide-ranging scope of a criminal statute’s general language places too much power in the prosecutor’s hands. Pp. 3–9. (b) Following the same approach taken in similar cases, the Government here must show that there is a “nexus” between the defendant’s conduct and a particular administrative proceeding, such as an investigation, an audit, or other targeted administrative action. See Aguilar, supra, at 599. The term “particular administrative proceeding” does not mean every act carried out by IRS employees in the course of their administration of the Tax Code. Just because a taxpayer knows that the IRS will review her tax return annually does not transform every Tax Code violation into an obstruction charge. In addition to satisfying the nexus requirement, the Government must show that the proceeding was pending at the time the defendant engaged in the obstructive conduct or, at the least, was then reasonably foreseeable by the defendant. See Arthur Andersen LLP v. United States, 544 U. S. 696 –708. Pp. 9–11. 839 F. 3d 209, reversed and remanded. Breyer, J., delivered the opinion of the Court, in which Roberts, C. J., and Kennedy, Ginsburg, Sotomayor, Kagan, and Gorsuch, JJ., joined. Thomas, J., filed a dissenting opinion, in which Alito, J., joined.

A clause in §7212(a) of the Internal Revenue Code makes it a felony “corruptly or by force” to “endeavo[r] to obstruct or imped[e] the due administration of this title.” 26 U. S. C. §7212(a). The question here concerns the breadth of that statutory phrase. Does it cover virtually all governmental efforts to collect taxes? Or does it have a narrower scope? In our view, “due administration of [the Tax Code]” does not cover routine administrative procedures that are near-universally applied to all taxpayers, such as the ordinary processing of income tax returns. Rather, the clause as a whole refers to specific interference with targeted governmental tax-related proceedings, such as a particular investigation or audit. I The Internal Revenue Code provision at issue, §7212(a), has two substantive clauses. The first clause, which we shall call the “Officer Clause,” forbids “corruptly or by force or threats of force (including any threatening letter or communication) endeavor[ing] to intimidate or impede any officer or employee of the United States acting in an official capacity under [the Internal Revenue Code].” Ibid. (emphasis added). The second clause, which we shall call the “Omnibus Clause,” forbids “corruptly or by force or threats of force (including any threatening letter or communication) obstruct[ing] or imped[ing], or endeavor[ing] to obstruct or impede, the due administration of [the Internal Revenue Code].” Ibid. (emphasis added). As we said at the outset, we here consider the scope of the Omnibus Clause. (We have placed the full text of §7212 in the Appendix, infra.) Between 2004 and 2009, the Internal Revenue Service (IRS) opened, then closed, then reopened an investigation into the tax activities of Carlo Marinello, the petitioner here. In 2012 the Government indicted Marinello, charging him with violations of several criminal tax statutes including the Omnibus Clause. In respect to the Omnibus Clause the Government claimed that Marinello had engaged in at least one of eight different specified activities, including “failing to maintain corporate books and records,” “failing to provide” his tax accountant “with complete and accurate” tax “information,” “destroying . . . business records,” “hiding income,” and “paying employees . . . with cash.” 839 F. 3d 209, 213 (CA2 2016). Before the jury retired to consider the charges, the judge instructed it that, to convict Marinello of violating the Omnibus Clause, it must find unanimously that he engaged in at least one of the eight practices just mentioned, that the jurors need not agree on which one, and that he did so “corruptly,” meaning “with the intent to secure an unlawful advantage or benefit, either for [himself ] or for another.” App. in No. 15–2224 (CA2), p. 432. The judge, however, did not instruct the jury that it must find that Marinello knew he was under investigation and intended corruptly to interfere with that investigation. The jury subsequently convicted Marinello on all counts. Marinello appealed to the Court of Appeals for the Second Circuit. He argued, among other things, that a violation of the Omnibus Clause requires the Government to show that the defendant had tried to interfere with a “pending IRS proceeding,” such as a particular investigation. Brief for Appellant in No. 15–2224, pp. 23–25. The appeals court disagreed. It held that a defendant need not possess “ ‘an awareness of a particular [IRS] action or investigation.’ ” 839 F. 3d, at 221 (quoting United States v. Wood, 384 Fed. Appx. 698, 704 (CA2 2010); alteration in original). The full Court of Appeals rejected Marinello’s petition for rehearing, two judges dissenting. 855 F. 3d 455 (CA2 2017). Marinello then petitioned for certiorari, asking us to decide whether the Omnibus Clause requires the Government to prove the defendant was aware of “a pending IRS action or proceeding, such as an investigation or audit,” when he “engaged in the purportedly obstructive conduct.” Pet. for Cert. i. In light of a division of opinion among the Circuits on this point, we granted the petition. Compare United States v. Kassouf, 144 F. 3d 952 (CA6 1998) (requiring showing of a pending proceeding), with 839 F. 3d, at 221 (disagreeing with Kassouf ). II In United States v. Aguilar, 515 U. S. 593 (1995) , we interpreted a similarly worded criminal statute. That statute made it a felony “corruptly or by threats or force, or by any threatening letter or communication, [to] influenc[e], obstruc[t], or imped[e], or endeavo[r] to influence, obstruct, or impede, the due administration of justice.” 18 U. S. C. §1503(a). The statute concerned not (as here) “the due administration of” the Internal Revenue Code but rather “the due administration of justice.” (We have placed the full text of §1503 in the Appendix, infra.) In interpreting that statute we pointed to earlier cases in which courts had held that the Government must prove “an intent to influence judicial or grand jury proceedings.” Aguilar, supra, at 599 (citing United States v. Brown, 688 F. 2d 596, 598 (CA9 1982)). We noted that some courts had imposed a “ ‘nexus’ requirement”: that the defendant’s “act must have a relationship in time, causation, or logic with the judicial proceedings.” Aguilar, supra, at 599 (citing United States v. Wood, 6 F. 3d 692, 696 (CA10 1993), and United States v. Walasek, 527 F. 2d 676, 679, and n. 12 (CA3 1975)). And we adopted the same requirement. We set forth two important reasons for doing so. We wrote that we “have traditionally exercised restraint in assessing the reach of a federal criminal statute, both out of deference to the prerogatives of Congress and out of concern that ‘a fair warning should be given to the world in language that the common world will understand, of what the law intends to do if a certain line is passed.’ ” Aguilar, supra, at 600 (quoting McBoyle v. United States, 283 U. S. 25, 27 (1931) ; citation omitted). Both reasons apply here with similar strength. As to Congress’ intent, the literal language of the statute is neutral. The statutory words “obstruct or impede” are broad. They can refer to anything that “block[s],” “make[s] difficult,” or “hinder[s].” Black’s Law Dictionary 1246 (10th ed. 2014) (obstruct); Webster’s New International Dictionary (Webster’s) 1248 (2d ed. 1954) (impede); id., at 1682 (obstruct); accord, 5 Oxford English Dictionary 80 (1933) (impede); 7 id., at 36 (obstruct). But the verbs “obstruct” and “impede” suggest an object—the taxpayer must hinder a particular person or thing. Here, the object is the “due administration of this title.” The word “administration” can be read literally to refer to every “[a]ct or process of administering” including every act of “managing” or “conduct[ing]” any “office,” or “performing the executive duties of” any “institution, business, or the like.” Webster’s 34. But the whole phrase—the due administration of the Tax Code—is best viewed, like the due administration of justice, as referring to only some of those acts or to some separable parts of an institution or business. Cf. Aguilar, supra, at 600–601 (concluding false statements made to an investigating agent, rather than a grand jury, do not support a conviction for obstruction of justice). Here statutory context confirms that the text refers to specific, targeted acts of administration. The Omnibus Clause appears in the middle of a statutory sentence that refers specifically to efforts to “intimidate or impede any officer or employee of the United States acting in an official capacity.” 26 U. S. C. §7212(a) (emphasis added). The first part of the sentence also refers to “force or threats of force,” which the statute elsewhere defines as “threats of bodily harm to the officer or employee of the United States or to a member of his family.” Ibid. (emphasis added). The following subsection refers to the “forcibl[e] rescu[e]” of “any property after it shall have been seized under” the Internal Revenue Code. §7212(b) (emphasis added). Subsections (a) and (b) thus refer to corrupt or forceful actions taken against individual identifiable persons or property. And, in that context the Omnibus Clause logically serves as a “catchall” in respect to the obstructive conduct the subsection sets forth, not as a “catchall” for every violation that interferes with what the Government describes as the “continuous, ubiquitous, and universally known” administration of the Internal Revenue Code. Brief in Opposition 9. Those who find legislative history helpful can find confirmation of the more limited scope of the Omnibus Clause in the House and Senate Reports written when Congress first enacted the Omnibus Clause. See H. R. Rep. No. 1337, 83d Cong., 2d Sess. (1954); S. Rep. No. 1622, 83d Cong., 2d Sess. (1954). According to the House Report, §7212 “provides for the punishment of threats or threatening acts against agents of the Internal Revenue Service, or any other officer or employee of the United States, or members of the families of such persons, on account of the performance by such agents or officers or employees of their official duties” and “will also punish the corrupt solicitation of an internal revenue employee.” H. R. Rep. No. 1337, at A426 (emphasis added). The Senate Report also refers to the section as aimed at targeting officers and employees. It says that §7212 “covers all cases where the officer is intimidated or injured; that is, where corruptly, by force or threat of force, directly or by communication, an attempt is made to impede the administration of the internal-revenue laws.” S. Rep. No. 1622, at 147 (emphasis added). We have found nothing in the statute’s history suggesting that Congress intended the Omnibus Clause as a catchall applicable to the entire Code including the routine processing of tax returns, receipt of tax payments, and issuance of tax refunds. Viewing the Omnibus Clause in the broader statutory context of the full Internal Revenue Code also counsels against adopting the Government’s broad reading. That is because the Code creates numerous misdemeanors, ranging from willful failure to furnish a required statement to employees, §7204, to failure to keep required records, §7203, to misrepresenting the number of exemptions to which an employee is entitled on IRS Form W–4, §7205, to failure to pay any tax owed, however small the amount, §7203. To interpret the Omnibus Clause as applying to all Code administration would potentially transform many, if not all, of these misdemeanor provisions into felonies, making the specific provisions redundant, or perhaps the subject matter of plea bargaining. Some overlap in criminal provisions is, of course, inevitable. See, e.g., Sansone v. United States, 380 U. S. 343, 349 (1965) (affirming conviction for tax evasion despite overlap with other provisions). Indeed, as the dissent notes, post, at 8 (opinion of Thomas, J.), Marinello’s preferred reading of §7212 potentially overlaps with another provision of federal law that criminalizes the obstruction of the “due and proper administration of the law under which any pending proceeding is being had before any department or agency of the United States,” 18 U. S. C. §1505. But we have not found any case from this Court interpreting a statutory provision that would create overlap and redundancy to the degree that would result from the Government’s broad reading of §7212—particularly when it would “ ‘render superfluous other provisions in the same enactment.’ ” Freytag v. Commissioner, 501 U. S. 868, 877 (1991) (quoting Pennsylvania Dept. of Public Welfare v. Davenport, 495 U. S. 552, 562 (1990) ; see also Yates v. United States, 574 U. S. ___, ___ (2015) (plurality opinion) (slip op., at 13). A broad interpretation would also risk the lack of fair warning and related kinds of unfairness that led this Court in Aguilar to “exercise” interpretive “restraint.” See 515 U. S., at 600; see also Yates, supra, at ___–___ (slip op., at 18–19); Arthur Andersen LLP v. United States, 544 U. S. 696 –704 (2005). Interpreted broadly, the provision could apply to a person who pays a babysitter $41 per week in cash without withholding taxes, see 26 CFR §31.3102–1(a)(2017); IRS, Publication 926, pp. 5–6 (2018), leaves a large cash tip in a restaurant, fails to keep donation receipts from every charity to which he or she contributes, or fails to provide every record to an accountant. Such an individual may sometimes believe that, in doing so, he is running the risk of having violated an IRS rule, but we sincerely doubt he would believe he is facing a potential felony prosecution for tax obstruction. Had Congress intended that outcome, it would have spoken with more clarity than it did in §7212(a). The Government argues that the need to show the defendant’s obstructive conduct was done “corruptly” will cure any overbreadth problem. But we do not see how. The Government asserts that “corruptly” means acting with “the specific intent to obtain an unlawful advantage” for the defendant or another. See Tr. of Oral Arg. 37; accord, 839 F. 3d, at 218. Yet, practically speaking, we struggle to imagine a scenario where a taxpayer would “willfully” violate the Tax Code (the mens rea requirement of various tax crimes, including misdemeanors, see, e.g., 26 U. S. C. §§7203, 7204, 7207) without intending someone to obtain an unlawful advantage. See Cheek v. United States, 498 U. S. 192, 201 (1991) (“Willfulness . . . requires the Government to prove that the law imposed a duty on the defendant, that the defendant knew of this duty, and that he voluntarily and intentionally violated that duty”) A taxpayer may know with a fair degree of certainty that her babysitter will not declare a cash payment as income—and, if so, a jury could readily find that the taxpayer acted to obtain an unlawful benefit for another. For the same reason, we find unconvincing the dissent’s argument that the distinction between “willfully” and “corruptly”—at least as defined by the Government—reflects any meaningful difference in culpability. See post, at 6–7. Neither can we rely upon prosecutorial discretion to narrow the statute’s scope. True, the Government used the Omnibus Clause only sparingly during the first few decades after its enactment. But it used the clause more often after the early 1990’s. Brief for Petitioner 9. And, at oral argument the Government told us that, where more punitive and less punitive criminal provisions both apply to a defendant’s conduct, the Government will charge a violation of the more punitive provision as long as it can readily prove that violation at trial. Tr. of Oral Arg. 46–47, 55–57; see Office of the Attorney General, Department Charging and Sentencing Policy (May 10, 2017), online at http://www.justice.gov/opa/press-release/ file/965896/download (as last visited Mar. 16, 2018). Regardless, to rely upon prosecutorial discretion to narrow the otherwise wide-ranging scope of a criminal statute’s highly abstract general statutory language places great power in the hands of the prosecutor. Doing so risks allowing “policemen, prosecutors, and juries to pursue their personal predilections,” Smith v. Goguen, 415 U. S. 566, 575 (1974) , which could result in the nonuniform execution of that power across time and geographic location. And insofar as the public fears arbitrary prosecution, it risks undermining necessary confidence in the criminal justice system. That is one reason why we have said that we “cannot construe a criminal statute on the assumption that the Government will ‘use it responsibly.’ ” McDonnell v. United States, 579 U. S. ___, ___ (2016) (slip op., at 23) (quoting United States v. Stevens, 559 U. S. 460, 480 (2010) ). And it is why “[w]e have traditionally exercised restraint in assessing the reach of a federal criminal statute.” Aguilar, supra, at 600. III In sum, we follow the approach we have taken in similar cases in interpreting §7212(a)’s Omnibus Clause. To be sure, the language and history of the provision at issue here differ somewhat from that of other obstruction provisions we have considered in the past. See Aguilar, supra (interpreting a statute prohibiting the obstruction of “the due administration of justice”); Arthur Andersen, supra (interpreting a statute prohibiting the destruction of an object with intent to impair its integrity or availability for use in an official proceeding); Yates, supra (interpreting a statute prohibiting the destruction, concealment, or covering up of any “record, document, or tangible object with the intent to” obstruct the “investigation or proper administration of any matter within the jurisdiction of any department or agency of the United States”). The Government and the dissent urge us to ignore these precedents because of those differences. The dissent points out, for example, that the predecessor to the obstruction statute we interpreted in Aguilar, 18 U. S. C. §1503, prohibited influencing, intimidating, or impeding “any witness or officer in any court of the United States” or endeavoring “to obstruct or imped[e] the due administration of justice therein.” Pettibone v. United States, 148 U. S. 197, 202 (1893) (citing Rev. Stat. §5399; emphasis added); see post, at 9. But Congress subsequently deleted the word “therein,” leaving only a broadly worded prohibition against obstruction of “the due administration of justice.” Act of June 25, 1948, §1503, 62Stat. 769–770. Congress then used that same amended formulation when it enacted §7212, prohibiting the “obstruction of the due administration” of the Tax Code. Internal Revenue Code of 1954, 68A Stat. 855. Given this similarity, it is helpful to consider how we have interpreted §1503 and other obstruction statutes in considering §7212. The language of some and the underlying principles of all these cases are similar. We consequently find these precedents—though not controlling—highly instructive for use as a guide toward a proper resolution of the issue now before us. See Smith v. City of Jackson, 544 U. S. 228, 233 (2005) . We conclude that, to secure a conviction under the Omnibus Clause, the Government must show (among other things) that there is a “nexus” between the defendant’s conduct and a particular administrative proceeding, such as an investigation, an audit, or other targeted administrative action. That nexus requires a “relationship in time, causation, or logic with the [administrative] proceeding.” Aguilar, 515 U. S., at 599 (citing Wood, 6 F. 3d, at 696). By “particular administrative proceeding” we do not mean every act carried out by IRS employees in the course of their “continuous, ubiquitous, and universally known” administration of the Tax Code. Brief in Opposition 9. While we need not here exhaustively itemize the types of administrative conduct that fall within the scope of the statute, that conduct does not include routine, day-to-day work carried out in the ordinary course by the IRS, such as the review of tax returns. The Government contends the processing of tax returns is part of the administration of the Internal Revenue Code and any corrupt effort to interfere with that task can therefore serve as the basis of an obstruction conviction. But the same could have been said of the defendant’s effort to mislead the investigating agent in Aguilar. The agent’s investigation was, at least in some broad sense, a part of the administration of justice. But we nevertheless held the defendant’s conduct did not support an obstruction charge. 515 U. S., at 600. In light of our decision in Aguilar, we find it appropriate to construe §7212’s Omnibus Clause more narrowly than the Government proposes. Just because a taxpayer knows that the IRS will review her tax return every year does not transform every violation of the Tax Code into an obstruction charge. In addition to satisfying this nexus requirement, the Government must show that the proceeding was pending at the time the defendant engaged in the obstructive conduct or, at the least, was then reasonably foreseeable by the defendant. See Arthur Andersen, 544 U. S., at 703, 707–708 (requiring the Government to prove a proceeding was foreseeable in order to convict a defendant for persuading others to shred documents to prevent their “use in an official proceeding”). It is not enough for the Government to claim that the defendant knew the IRS may catch on to his unlawful scheme eventually. To use a maritime analogy, the proceeding must at least be in the offing. For these reasons, the Second Circuit’s judgment is reversed, and the case is remanded for further proceedings consistent with this opinion. It is so ordered. APPENDIX 26 U. S. C. §7212: “Attempts to interfere with administration of internal revenue laws “(a) Corrupt or forcible interference “Whoever corruptly or by force or threats of force (including any threatening letter or communication) endeavors to intimidate or impede any officer or employee of the United States acting in an official capacity under this title, or in any other way corruptly or by force or threats of force (including any threatening letter or communication) obstructs or impedes, or endeavors to obstruct or impede, the due administration of this title, shall, upon conviction thereof, be fined not more than $5,000, or imprisoned not more than 3 years, or both, except that if the offense is committed only by threats of force, the person convicted thereof shall be fined not more than $3,000, or imprisoned not more than 1 year, or both. The term ‘threats of force’, as used in this subsection, means threats of bodily harm to the officer or employee of the United States or to a member of his family. “(b) Forcible rescue of seized property “Any person who forcibly rescues or causes to be rescued any property after it shall have been seized under this title, or shall attempt or endeavor so to do, shall, excepting in cases otherwise provided for, for every such offense, be fined not more than $500, or not more than double the value of the property so rescued, whichever is the greater, or be imprisoned not more than 2 years.” 18 U. S. C. §1503: “Influencing or injuring officer or juror generally “(a) Whoever corruptly, or by threats or force, or by any threatening letter or communication, endeavors to influence, intimidate, or impede any grand or petit juror, or officer in or of any court of the United States, or officer who may be serving at any examination or other proceeding before any United States magistrate judge or other committing magistrate, in the discharge of his duty, or injures any such grand or petit juror in his person or property on account of any verdict or indictment assented to by him, or on account of his being or having been such juror, or injures any such officer, magistrate judge, or other committing magistrate in his person or property on account of the performance of his official duties, or corruptly or by threats or force, or by any threatening letter or communication, influences, obstructs, or impedes, or endeavors to influence, obstruct, or impede, the due administration of justice, shall be punished as provided in subsection (b). If the offense under this section occurs in connection with a trial of a criminal case, and the act in violation of this section involves the threat of physical force or physical force, the maximum term of imprisonment which may be imposed for the offense shall be the higher of that otherwise provided by law or the maximum term that could have been imposed for any offense charged in such case. “(b) The punishment for an offense under this section is— “(1) in the case of a killing, the punishment provided in sections 1111 and 1112; “(2) in the case of an attempted killing, or a case in which the offense was committed against a petit juror and in which a class A or B felony was charged, imprisonment for not more than 20 years, a fine under this title, or both; and “(3) in any other case, imprisonment for not more than 10 years, a fine under this title, or both.”

584.US.2017_16-111

Masterpiece Cakeshop, Ltd., is a Colorado bakery owned and operated by Jack Phillips, an expert baker and devout Christian. In 2012 he told a same-sex couple that he would not create a cake for their wedding celebration because of his religious opposition to same-sex marriages—marriages that Colorado did not then recognize—but that he would sell them other baked goods, e.g., birthday cakes. The couple filed a charge with the Colorado Civil Rights Commission (Commission) pursuant to the Colorado Anti-Discrimination Act (CADA), which prohibits, as relevant here, discrimination based on sexual orientation in a “place of business engaged in any sales to the public and any place offering services . . . to the public.” Under CADA’s administrative review system, the Colorado Civil Rights Division first found probable cause for a violation and referred the case to the Commission. The Commission then referred the case for a formal hearing before a state Administrative Law Judge (ALJ), who ruled in the couple’s favor. In so doing, the ALJ rejected Phillips’ First Amendment claims: that requiring him to create a cake for a same-sex wedding would violate his right to free speech by compelling him to exercise his artistic talents to express a message with which he disagreed and would violate his right to the free exercise of religion. Both the Commission and the Colorado Court of Appeals affirmed. Held: The Commission’s actions in this case violated the Free Exercise Clause. Pp. 9–18. (a) The laws and the Constitution can, and in some instances must, protect gay persons and gay couples in the exercise of their civil rights, but religious and philosophical objections to gay marriage are protected views and in some instances protected forms of expression. See Obergefell v. Hodges, 576 U. S. ___, ___. While it is unexceptional that Colorado law can protect gay persons in acquiring products and services on the same terms and conditions as are offered to other members of the public, the law must be applied in a manner that is neutral toward religion. To Phillips, his claim that using his artistic skills to make an expressive statement, a wedding endorsement in his own voice and of his own creation, has a significant First Amendment speech component and implicates his deep and sincere religious beliefs. His dilemma was understandable in 2012, which was before Colorado recognized the validity of gay marriages performed in the State and before this Court issued United States v. Windsor, 570 U. S. 744, or Obergefell. Given the State’s position at the time, there is some force to Phillips’ argument that he was not unreasonable in deeming his decision lawful. State law at the time also afforded storekeepers some latitude to decline to create specific messages they considered offensive. Indeed, while the instant enforcement proceedings were pending, the State Civil Rights Division concluded in at least three cases that a baker acted lawfully in declining to create cakes with decorations that demeaned gay persons or gay marriages. Phillips too was entitled to a neutral and respectful consideration of his claims in all the circumstances of the case. Pp. 9–12. (b) That consideration was compromised, however, by the Commission’s treatment of Phillips’ case, which showed elements of a clear and impermissible hostility toward the sincere religious beliefs motivating his objection. As the record shows, some of the commissioners at the Commission’s formal, public hearings endorsed the view that religious beliefs cannot legitimately be carried into the public sphere or commercial domain, disparaged Phillips’ faith as despicable and characterized it as merely rhetorical, and compared his invocation of his sincerely held religious beliefs to defenses of slavery and the Holocaust. No commissioners objected to the comments. Nor were they mentioned in the later state-court ruling or disavowed in the briefs filed here. The comments thus cast doubt on the fairness and impartiality of the Commission’s adjudication of Phillips’ case. Another indication of hostility is the different treatment of Phillips’ case and the cases of other bakers with objections to anti-gay messages who prevailed before the Commission. The Commission ruled against Phillips in part on the theory that any message on the requested wedding cake would be attributed to the customer, not to the baker. Yet the Division did not address this point in any of the cases involving requests for cakes depicting anti-gay marriage symbolism. The Division also considered that each bakery was willing to sell other products to the prospective customers, but the Commission found Phillips’ willingness to do the same irrelevant. The State Court of Appeals’ brief discussion of this disparity of treatment does not answer Phillips’ concern that the State’s practice was to disfavor the religious basis of his objection. Pp. 12–16. (c) For these reasons, the Commission’s treatment of Phillips’ case violated the State’s duty under the First Amendment not to base laws or regulations on hostility to a religion or religious viewpoint. The government, consistent with the Constitution’s guarantee of free exercise, cannot impose regulations that are hostile to the religious beliefs of affected citizens and cannot act in a manner that passes judgment upon or presupposes the illegitimacy of religious beliefs and practices. Church of Lukumi Babalu Aye, Inc. v. Hialeah, 508 U. S. 520. Factors relevant to the assessment of governmental neutrality include “the historical background of the decision under challenge, the specific series of events leading to the enactment or official policy in question, and the legislative or administrative history, including contemporaneous statements made by members of the decisionmaking body.” Id., at 540. In view of these factors, the record here demonstrates that the Commission’s consideration of Phillips’ case was neither tolerant nor respectful of his religious beliefs. The Commission gave “every appearance,” id., at 545, of adjudicating his religious objection based on a negative normative “evaluation of the particular justification” for his objection and the religious grounds for it, id., at 537, but government has no role in expressing or even suggesting whether the religious ground for Phillips’ conscience-based objection is legitimate or illegitimate. The inference here is thus that Phillips’ religious objection was not considered with the neutrality required by the Free Exercise Clause. The State’s interest could have been weighed against Phillips’ sincere religious objections in a way consistent with the requisite religious neutrality that must be strictly observed. But the official expressions of hostility to religion in some of the commissioners’ comments were inconsistent with that requirement, and the Commission’s disparate consideration of Phillips’ case compared to the cases of the other bakers suggests the same. Pp. 16–18. 370 P. 3d 272, reversed. Kennedy, J., delivered the opinion of the Court, in which Roberts, C. J., and Breyer, Alito, Kagan, and Gorsuch, JJ., joined. Kagan, J., filed a concurring opinion, in which Breyer, J., joined. Gorsuch, J., filed a concurring opinion, in which Alito, J., joined. Thomas, J., filed an opinion concurring in part and concurring in the judgment, in which Gorsuch, J., joined. Ginsburg, J., filed a dissenting opinion, in which Sotomayor, J., joined.

In 2012 a same-sex couple visited Masterpiece Cakeshop, a bakery in Colorado, to make inquiries about ordering a cake for their wedding reception. The shop’s owner told the couple that he would not create a cake for their wedding because of his religious opposition to same-sex marriages—marriages the State of Colorado itself did not recognize at that time. The couple filed a charge with the Colorado Civil Rights Commission alleging discrimination on the basis of sexual orientation in violation of the Colorado Anti-Discrimination Act. The Commission determined that the shop’s actions violated the Act and ruled in the couple’s favor. The Colorado state courts affirmed the ruling and its enforcement order, and this Court now must decide whether the Commission’s order violated the Constitution. The case presents difficult questions as to the proper reconciliation of at least two principles. The first is the authority of a State and its governmental entities to protect the rights and dignity of gay persons who are, or wish to be, married but who face discrimination when they seek goods or services. The second is the right of all persons to exercise fundamental freedoms under the First Amendment, as applied to the States through the Fourteenth Amendment. The freedoms asserted here are both the freedom of speech and the free exercise of religion. The free speech aspect of this case is difficult, for few persons who have seen a beautiful wedding cake might have thought of its creation as an exercise of protected speech. This is an instructive example, however, of the proposition that the application of constitutional freedoms in new contexts can deepen our understanding of their meaning. One of the difficulties in this case is that the parties disagree as to the extent of the baker’s refusal to provide service. If a baker refused to design a special cake with words or images celebrating the marriage—for instance, a cake showing words with religious meaning—that might be different from a refusal to sell any cake at all. In defining whether a baker’s creation can be protected, these details might make a difference. The same difficulties arise in determining whether a baker has a valid free exercise claim. A baker’s refusal to attend the wedding to ensure that the cake is cut the right way, or a refusal to put certain religious words or decorations on the cake, or even a refusal to sell a cake that has been baked for the public generally but includes certain religious words or symbols on it are just three examples of possibilities that seem all but endless. Whatever the confluence of speech and free exercise principles might be in some cases, the Colorado Civil Rights Commission’s consideration of this case was inconsistent with the State’s obligation of religious neutrality. The reason and motive for the baker’s refusal were based on his sincere religious beliefs and convictions. The Court’s precedents make clear that the baker, in his capacity as the owner of a business serving the public, might have his right to the free exercise of religion limited by generally applicable laws. Still, the delicate question of when the free exercise of his religion must yield to an otherwise valid exercise of state power needed to be determined in an adjudication in which religious hostility on the part of the State itself would not be a factor in the balance the State sought to reach. That requirement, however, was not met here. When the Colorado Civil Rights Commission considered this case, it did not do so with the religious neutrality that the Constitution requires. Given all these considerations, it is proper to hold that whatever the outcome of some future controversy involving facts similar to these, the Commission’s actions here violated the Free Exercise Clause; and its order must be set aside. I A Masterpiece Cakeshop, Ltd., is a bakery in Lakewood, Colorado, a suburb of Denver. The shop offers a variety of baked goods, ranging from everyday cookies and brownies to elaborate custom-designed cakes for birthday parties, weddings, and other events. Jack Phillips is an expert baker who has owned and operated the shop for 24 years. Phillips is a devout Christian. He has explained that his “main goal in life is to be obedient to” Jesus Christ and Christ’s “teachings in all aspects of his life.” App. 148. And he seeks to “honor God through his work at Masterpiece Cakeshop.” Ibid. One of Phillips’ religious beliefs is that “God’s intention for marriage from the beginning of history is that it is and should be the union of one man and one woman.” Id., at 149. To Phillips, creating a wedding cake for a same-sex wedding would be equivalent to participating in a celebration that is contrary to his own most deeply held beliefs. Phillips met Charlie Craig and Dave Mullins when they entered his shop in the summer of 2012. Craig and Mullins were planning to marry. At that time, Colorado did not recognize same-sex marriages, so the couple planned to wed legally in Massachusetts and afterwards to host a reception for their family and friends in Denver. To prepare for their celebration, Craig and Mullins visited the shop and told Phillips that they were interested in ordering a cake for “our wedding.” Id., at 152 (emphasis de- leted). They did not mention the design of the cake they envisioned. Phillips informed the couple that he does not “create” wedding cakes for same-sex weddings. Ibid. He explained, “I’ll make your birthday cakes, shower cakes, sell you cookies and brownies, I just don’t make cakes for same sex weddings.” Ibid. The couple left the shop without further discussion. The following day, Craig’s mother, who had accompanied the couple to the cakeshop and been present for their interaction with Phillips, telephoned to ask Phillips why he had declined to serve her son. Phillips explained that he does not create wedding cakes for same-sex weddings because of his religious opposition to same-sex marriage, and also because Colorado (at that time) did not recognize same-sex marriages. Id., at 153. He later explained his belief that “to create a wedding cake for an event that celebrates something that directly goes against the teachings of the Bible, would have been a personal endorsement and participation in the ceremony and relationship that they were entering into.” Ibid. (emphasis deleted). B For most of its history, Colorado has prohibited discrimination in places of public accommodation. In 1885, less than a decade after Colorado achieved statehood, the General Assembly passed “An Act to Protect All Citizens in Their Civil Rights,” which guaranteed “full and equal enjoyment” of certain public facilities to “all citizens,” “regardless of race, color or previous condition of servitude.” 1885 Colo. Sess. Laws pp. 132–133. A decade later, the General Assembly expanded the requirement to apply to “all other places of public accommodation.” 1895 Colo. Sess. Laws ch. 61, p. 139. Today, the Colorado Anti-Discrimination Act (CADA) carries forward the state’s tradition of prohibiting discrimination in places of public accommodation. Amended in 2007 and 2008 to prohibit discrimination on the basis of sexual orientation as well as other protected characteristics, CADA in relevant part provides as follows: “It is a discriminatory practice and unlawful for a person, directly or indirectly, to refuse, withhold from, or deny to an individual or a group, because of disability, race, creed, color, sex, sexual orientation, marital status, national origin, or ancestry, the full and equal enjoyment of the goods, services, facilities, privileges, advantages, or accommodations of a place of public accommodation.” Colo. Rev. Stat. §24–34–601(2)(a) (2017). The Act defines “public accommodation” broadly to include any “place of business engaged in any sales to the public and any place offering services . . . to the public,” but excludes “a church, synagogue, mosque, or other place that is principally used for religious purposes.” §24–34–601(1). CADA establishes an administrative system for the resolution of discrimination claims. Complaints of discrimination in violation of CADA are addressed in the first instance by the Colorado Civil Rights Division. The Division investigates each claim; and if it finds probable cause that CADA has been violated, it will refer the matter to the Colorado Civil Rights Commission. The Commission, in turn, decides whether to initiate a formal hearing before a state Administrative Law Judge (ALJ), who will hear evidence and argument before issuing a written decision. See §§24–34–306, 24–4–105(14). The decision of the ALJ may be appealed to the full Commission, a seven-member appointed body. The Commission holds a public hearing and deliberative session before voting on the case. If the Commission determines that the evidence proves a CADA violation, it may impose remedial measures as provided by statute. See §24–34–306(9). Available remedies include, among other things, orders to cease-and-desist a discriminatory policy, to file regular compliance reports with the Commission, and “to take affirmative action, including the posting of notices setting forth the substantive rights of the public.” §24–34–605. Colorado law does not permit the Commission to assess money damages or fines. §§24–34–306(9), 24–34–605. C Craig and Mullins filed a discrimination complaint against Masterpiece Cakeshop and Phillips in August 2012, shortly after the couple’s visit to the shop. App. 31. The complaint alleged that Craig and Mullins had been denied “full and equal service” at the bakery because of their sexual orientation, id., at 35, 48, and that it was Phillips’ “standard business practice” not to provide cakes for same-sex weddings, id., at 43. The Civil Rights Division opened an investigation. The investigator found that “on multiple occasions,” Phillips “turned away potential customers on the basis of their sexual orientation, stating that he could not create a cake for a same-sex wedding ceremony or reception” because his religious beliefs prohibited it and because the potential customers “were doing something illegal” at that time. Id., at 76. The investigation found that Phillips had declined to sell custom wedding cakes to about six other same-sex couples on this basis. Id., at 72. The investigator also recounted that, according to affidavits submitted by Craig and Mullins, Phillips’ shop had refused to sell cupcakes to a lesbian couple for their commitment celebration because the shop “had a policy of not selling baked goods to same-sex couples for this type of event.” Id., at 73. Based on these findings, the Division found probable cause that Phillips violated CADA and referred the case to the Civil Rights Commission. Id., at 69. The Commission found it proper to conduct a formal hearing, and it sent the case to a State ALJ. Finding no dispute as to material facts, the ALJ entertained cross-motions for summary judgment and ruled in the couple’s favor. The ALJ first rejected Phillips’ argument that declining to make or create a wedding cake for Craig and Mullins did not violate Colorado law. It was undisputed that the shop is subject to state public accommodations laws. And the ALJ determined that Phillips’ actions constituted prohibited discrimination on the basis of sex- ual orientation, not simply opposition to same-sex marriage as Phillips contended. App. to Pet. for Cert. 68a–72a. Phillips raised two constitutional claims before the ALJ. He first asserted that applying CADA in a way that would require him to create a cake for a same-sex wedding would violate his First Amendment right to free speech by compelling him to exercise his artistic talents to express a message with which he disagreed. The ALJ rejected the contention that preparing a wedding cake is a form of protected speech and did not agree that creating Craig and Mullins’ cake would force Phillips to adhere to “an ideological point of view.” Id., at 75a. Applying CADA to the facts at hand, in the ALJ’s view, did not interfere with Phillips’ freedom of speech. Phillips also contended that requiring him to create cakes for same-sex weddings would violate his right to the free exercise of religion, also protected by the First Amendment. Citing this Court’s precedent in Employment Div., Dept. of Human Resources of Ore. v. Smith, 494 U. S. 872 (1990), the ALJ determined that CADA is a “valid and neutral law of general applicability” and therefore that applying it to Phillips in this case did not violate the Free Exercise Clause. Id., at 879; App. to Pet. for Cert. 82a–83a. The ALJ thus ruled against Phillips and the cakeshop and in favor of Craig and Mullins on both constitutional claims. The Commission affirmed the ALJ’s decision in full. Id., at 57a. The Commission ordered Phillips to “cease and desist from discriminating against . . . same-sex couples by refusing to sell them wedding cakes or any product [they] would sell to heterosexual couples.” Ibid. It also ordered additional remedial measures, including “comprehensive staff training on the Public Accommodations section” of CADA “and changes to any and all company policies to comply with . . . this Order.” Id., at 58a. The Commission additionally required Phillips to prepare “quarterly compliance reports” for a period of two years documenting “the number of patrons denied service” and why, along with “a statement describing the remedial actions taken.” Ibid. Phillips appealed to the Colorado Court of Appeals, which affirmed the Commission’s legal determinations and remedial order. The court rejected the argument that the “Commission’s order unconstitutionally compels” Phillips and the shop “to convey a celebratory message about same sex marriage.” Craig v. Masterpiece Cakeshop, Inc., 370 P. 3d 272, 283 (2015). The court also rejected the argument that the Commission’s order violated the Free Exercise Clause. Relying on this Court’s precedent in Smith, supra, at 879, the court stated that the Free Exercise Clause “does not relieve an individual of the obligation to comply with a valid and neutral law of general applicability” on the ground that following the law would interfere with religious practice or belief. 370 P. 3d, at 289. The court concluded that requiring Phillips to comply with the statute did not violate his free exercise rights. The Colorado Supreme Court declined to hear the case. Phillips sought review here, and this Court granted certiorari. 582 U. S. ___ (2017). He now renews his claims under the Free Speech and Free Exercise Clauses of the First Amendment. II A Our society has come to the recognition that gay persons and gay couples cannot be treated as social outcasts or as inferior in dignity and worth. For that reason the laws and the Constitution can, and in some instances must, protect them in the exercise of their civil rights. The exercise of their freedom on terms equal to others must be given great weight and respect by the courts. At the same time, the religious and philosophical objections to gay marriage are protected views and in some instances protected forms of expression. As this Court observed in Obergefell v. Hodges, 576 U. S. ___ (2015), “[t]he First Amendment ensures that religious organizations and persons are given proper protection as they seek to teach the principles that are so fulfilling and so central to their lives and faiths.” Id., at ___ (slip op., at 27). Nevertheless, while those religious and philosophical objections are protected, it is a general rule that such objections do not allow business owners and other actors in the economy and in society to deny protected persons equal access to goods and services under a neutral and generally applicable public accommodations law. See Newman v. Piggy Park Enterprises, Inc., 390 U. S. 400, 402, n. 5 (1968) (per curiam); see also Hurley v. Irish-American Gay, Lesbian and Bisexual Group of Boston, Inc., 515 U. S. 557, 572 (1995) (“Provisions like these are well within the State’s usual power to enact when a legislature has reason to believe that a given group is the target of discrimination, and they do not, as a general matter, violate the First or Fourteenth Amendments”). When it comes to weddings, it can be assumed that a member of the clergy who objects to gay marriage on moral and religious grounds could not be compelled to perform the ceremony without denial of his or her right to the free exercise of religion. This refusal would be well understood in our constitutional order as an exercise of religion, an exercise that gay persons could recognize and accept without serious diminishment to their own dignity and worth. Yet if that exception were not confined, then a long list of persons who provide goods and services for marriages and weddings might refuse to do so for gay persons, thus resulting in a community-wide stigma inconsistent with the history and dynamics of civil rights laws that ensure equal access to goods, services, and public accommodations. It is unexceptional that Colorado law can protect gay persons, just as it can protect other classes of individuals, in acquiring whatever products and services they choose on the same terms and conditions as are offered to other members of the public. And there are no doubt innumerable goods and services that no one could argue implicate the First Amendment. Petitioners conceded, moreover, that if a baker refused to sell any goods or any cakes for gay weddings, that would be a different matter and the State would have a strong case under this Court’s precedents that this would be a denial of goods and services that went beyond any protected rights of a baker who offers goods and services to the general public and is subject to a neutrally applied and generally applicable public accommodations law. See Tr. of Oral Arg. 4–7, 10. Phillips claims, however, that a narrower issue is presented. He argues that he had to use his artistic skills to make an expressive statement, a wedding endorsement in his own voice and of his own creation. As Phillips would see the case, this contention has a significant First Amendment speech component and implicates his deep and sincere religious beliefs. In this context the baker likely found it difficult to find a line where the customers’ rights to goods and services became a demand for him to exercise the right of his own personal expression for their message, a message he could not express in a way consistent with his religious beliefs. Phillips’ dilemma was particularly understandable given the background of legal principles and administration of the law in Colorado at that time. His decision and his actions leading to the refusal of service all occurred in the year 2012. At that point, Colorado did not recognize the validity of gay marriages performed in its own State. See Colo. Const., Art. II, §31 (2012); 370 P. 3d, at 277. At the time of the events in question, this Court had not issued its decisions either in United States v. Windsor, 570 U. S. 744 (2013), or Obergefell. Since the State itself did not allow those marriages to be performed in Colorado, there is some force to the argument that the baker was not unreasonable in deeming it lawful to decline to take an action that he understood to be an expression of support for their validity when that expression was contrary to his sincerely held religious beliefs, at least insofar as his refusal was limited to refusing to create and express a message in support of gay marriage, even one planned to take place in another State. At the time, state law also afforded storekeepers some latitude to decline to create specific messages the storekeeper considered offensive. Indeed, while enforcement proceedings against Phillips were ongoing, the Colorado Civil Rights Division itself endorsed this proposition in cases involving other bakers’ creation of cakes, concluding on at least three occasions that a baker acted lawfully in declining to create cakes with decorations that demeaned gay persons or gay marriages. See Jack v. Gateaux, Ltd., Charge No. P20140071X (Mar. 24, 2015); Jack v. Le Bakery Sensual, Inc., Charge No. P20140070X (Mar. 24, 2015); Jack v. Azucar Bakery, Charge No. P20140069X (Mar. 24, 2015). There were, to be sure, responses to these arguments that the State could make when it contended for a different result in seeking the enforcement of its generally applicable state regulations of businesses that serve the public. And any decision in favor of the baker would have to be sufficiently constrained, lest all purveyors of goods and services who object to gay marriages for moral and religious reasons in effect be allowed to put up signs saying “no goods or services will be sold if they will be used for gay marriages,” something that would impose a serious stigma on gay persons. But, nonetheless, Phillips was entitled to the neutral and respectful consideration of his claims in all the circumstances of the case. B The neutral and respectful consideration to which Phillips was entitled was compromised here, however. The Civil Rights Commission’s treatment of his case has some elements of a clear and impermissible hostility toward the sincere religious beliefs that motivated his objection. That hostility surfaced at the Commission’s formal, public hearings, as shown by the record. On May 30, 2014, the seven-member Commission convened publicly to consider Phillips’ case. At several points during its meeting, commissioners endorsed the view that religious beliefs cannot legitimately be carried into the public sphere or commercial domain, implying that religious beliefs and persons are less than fully welcome in Colorado’s business community. One commissioner suggested that Phillips can believe “what he wants to believe,” but cannot act on his religious beliefs “if he decides to do business in the state.” Tr. 23. A few moments later, the commissioner restated the same position: “[I]f a businessman wants to do business in the state and he’s got an issue with the—the law’s impacting his personal belief system, he needs to look at being able to compromise.” Id., at 30. Standing alone, these statements are susceptible of different interpretations. On the one hand, they might mean simply that a business cannot refuse to provide services based on sexual orientation, regardless of the proprietor’s personal views. On the other hand, they might be seen as inappropriate and dismissive comments showing lack of due consideration for Phillips’ free exercise rights and the dilemma he faced. In view of the comments that followed, the latter seems the more likely. On July 25, 2014, the Commission met again. This meeting, too, was conducted in public and on the record. On this occasion another commissioner made specific reference to the previous meeting’s discussion but said far more to disparage Phillips’ beliefs. The commissioner stated: “I would also like to reiterate what we said in the hearing or the last meeting. Freedom of religion and religion has been used to justify all kinds of discrimination throughout history, whether it be slavery, whether it be the holocaust, whether it be—I mean, we—we can list hundreds of situations where freedom of religion has been used to justify discrimination. And to me it is one of the most despicable pieces of rhetoric that people can use to—to use their religion to hurt others.” Tr. 11–12. To describe a man’s faith as “one of the most despicable pieces of rhetoric that people can use” is to disparage his religion in at least two distinct ways: by describing it as despicable, and also by characterizing it as merely rhetorical—something insubstantial and even insincere. The commissioner even went so far as to compare Phillips’ invocation of his sincerely held religious beliefs to defenses of slavery and the Holocaust. This sentiment is inappropriate for a Commission charged with the solemn responsibility of fair and neutral enforcement of Colorado’s antidiscrimination law—a law that protects discrimination on the basis of religion as well as sexual orientation. The record shows no objection to these comments from other commissioners. And the later state-court ruling reviewing the Commission’s decision did not mention those comments, much less express concern with their content. Nor were the comments by the commissioners disavowed in the briefs filed in this Court. For these reasons, the Court cannot avoid the conclusion that these statements cast doubt on the fairness and impartiality of the Commission’s adjudication of Phillips’ case. Members of the Court have disagreed on the question whether statements made by lawmakers may properly be taken into account in determining whether a law intentionally discriminates on the basis of religion. See Church of Lukumi Babalu Aye, Inc. v. Hialeah, 508 U. S. 520, 540–542 (1993); id., at 558 (Scalia, J., concurring in part and concurring in judgment). In this case, however, the remarks were made in a very different context—by an adjudicatory body deciding a particular case. Another indication of hostility is the difference in treatment between Phillips’ case and the cases of other bakers who objected to a requested cake on the basis of conscience and prevailed before the Commission. As noted above, on at least three other occasions the Civil Rights Division considered the refusal of bakers to create cakes with images that conveyed disapproval of same-sex marriage, along with religious text. Each time, the Division found that the baker acted lawfully in refusing service. It made these determinations because, in the words of the Division, the requested cake included “wording and images [the baker] deemed derogatory,” Jack v. Gateaux, Ltd., Charge No. P20140071X, at 4; featured “language and images [the baker] deemed hateful,” Jack v. Le Bakery Sensual, Inc., Charge No. P20140070X, at 4; or displayed a message the baker “deemed as discriminatory, Jack v. Azucar Bakery, Charge No. P20140069X, at 4. The treatment of the conscience-based objections at issue in these three cases contrasts with the Commission’s treatment of Phillips’ objection. The Commission ruled against Phillips in part on the theory that any message the requested wedding cake would carry would be attributed to the customer, not to the baker. Yet the Division did not address this point in any of the other cases with respect to the cakes depicting anti-gay marriage symbolism. Additionally, the Division found no violation of CADA in the other cases in part because each bakery was willing to sell other products, including those depicting Christian themes, to the prospective customers. But the Commission dismissed Phillips’ willingness to sell “birthday cakes, shower cakes, [and] cookies and brownies,” App. 152, to gay and lesbian customers as irrelevant. The treatment of the other cases and Phillips’ case could reasonably be interpreted as being inconsistent as to the question of whether speech is involved, quite apart from whether the cases should ultimately be distinguished. In short, the Commission’s consideration of Phillips’ religious objection did not accord with its treatment of these other objections. Before the Colorado Court of Appeals, Phillips protested that this disparity in treatment reflected hostility on the part of the Commission toward his beliefs. He argued that the Commission had treated the other bakers’ conscience-based objections as legitimate, but treated his as illegitimate—thus sitting in judgment of his religious beliefs themselves. The Court of Appeals addressed the disparity only in passing and relegated its complete analysis of the issue to a footnote. There, the court stated that “[t]his case is distinguishable from the Colorado Civil Rights Division’s recent findings that [the other bakeries] in Denver did not discriminate against a Christian patron on the basis of his creed” when they refused to create the requested cakes. 370 P. 3d, at 282, n. 8. In those cases, the court continued, there was no impermissible discrimination because “the Division found that the bakeries . . . refuse[d] the patron’s request . . . because of the offensive nature of the requested message.” Ibid. A principled rationale for the difference in treatment of these two instances cannot be based on the government’s own assessment of offensiveness. Just as “no official, high or petty, can prescribe what shall be orthodox in politics, nationalism, religion, or other matters of opinion,” West Virginia Bd. of Ed. v. Barnette, 319 U. S. 624, 642 (1943), it is not, as the Court has repeatedly held, the role of the State or its officials to prescribe what shall be offensive. See Matal v. Tam, 582 U. S. ___, ___–___ (2017) (opinion of Alito, J.) (slip op., at 22–23). The Colorado court’s attempt to account for the difference in treatment elevates one view of what is offensive over another and itself sends a signal of official disapproval of Phillips’ religious beliefs. The court’s footnote does not, therefore, answer the baker’s concern that the State’s practice was to disfavor the religious basis of his objection. C For the reasons just described, the Commission’s treatment of Phillips’ case violated the State’s duty under the First Amendment not to base laws or regulations on hostility to a religion or religious viewpoint. In Church of Lukumi Babalu Aye, supra, the Court made clear that the government, if it is to respect the Constitution’s guarantee of free exercise, cannot impose regulations that are hostile to the religious beliefs of affected citizens and cannot act in a manner that passes judgment upon or presupposes the illegitimacy of religious beliefs and practices. The Free Exercise Clause bars even “subtle departures from neutrality” on matters of religion. Id., at 534. Here, that means the Commission was obliged under the Free Exercise Clause to proceed in a manner neutral toward and tolerant of Phillips’ religious beliefs. The Constitution “commits government itself to religious tolerance, and upon even slight suspicion that proposals for state intervention stem from animosity to religion or distrust of its practices, all officials must pause to remember their own high duty to the Constitution and to the rights it secures.” Id., at 547. Factors relevant to the assessment of governmental neutrality include “the historical background of the decision under challenge, the specific series of events leading to the enactment or official policy in question, and the legislative or administrative history, including contemporaneous statements made by members of the decisionmaking body.” Id., at 540. In view of these factors the record here demonstrates that the Commission’s consideration of Phillips’ case was neither tolerant nor respectful of Phillips’ religious beliefs. The Commission gave “every appearance,” id., at 545, of adjudicating Phillips’ religious objection based on a negative normative “evaluation of the particular justification” for his objection and the religious grounds for it. Id., at 537. It hardly requires restating that government has no role in deciding or even suggesting whether the religious ground for Phillips’ conscience-based objection is legitimate or illegitimate. On these facts, the Court must draw the inference that Phillips’ religious objection was not considered with the neutrality that the Free Exercise Clause requires. While the issues here are difficult to resolve, it must be concluded that the State’s interest could have been weighed against Phillips’ sincere religious objections in a way consistent with the requisite religious neutrality that must be strictly observed. The official expressions of hostility to religion in some of the commissioners’ comments—comments that were not disavowed at the Commission or by the State at any point in the proceedings that led to affirmance of the order—were inconsistent with what the Free Exercise Clause requires. The Commission’s disparate consideration of Phillips’ case compared to the cases of the other bakers suggests the same. For these reasons, the order must be set aside. III The Commission’s hostility was inconsistent with the First Amendment’s guarantee that our laws be applied in a manner that is neutral toward religion. Phillips was entitled to a neutral decisionmaker who would give full and fair consideration to his religious objection as he sought to assert it in all of the circumstances in which this case was presented, considered, and decided. In this case the adjudication concerned a context that may well be different going forward in the respects noted above. However later cases raising these or similar concerns are resolved in the future, for these reasons the rulings of the Commission and of the state court that enforced the Commission’s order must be invalidated. The outcome of cases like this in other circumstances must await further elaboration in the courts, all in the context of recognizing that these disputes must be resolved with tolerance, without undue disrespect to sincere religious beliefs, and without subjecting gay persons to indignities when they seek goods and services in an open market. The judgment of the Colorado Court of Appeals is reversed. It is so ordered.

584.US.2017_16-8255

Petitioner Robert McCoy was charged with murdering his estranged wife’s mother, stepfather, and son. McCoy pleaded not guilty to first-degree murder, insisting that he was out of State at the time of the killings and that corrupt police killed the victims when a drug deal went wrong. Although he vociferously insisted on his innocence and adamantly objected to any admission of guilt, the trial court permitted his counsel, Larry English, to tell the jury, during the trial’s guilt phase, McCoy “committed [the] three murders.” English’s strategy was to concede that McCoy committed the murders, but argue that McCoy’s mental state prevented him from forming the specific intent necessary for a first-degree murder conviction. Over McCoy’s repeated objection, English told the jury McCoy was the killer and that English “took [the] burden off of [the prosecutor]” on that issue. McCoy testified in his own defense, maintaining his innocence and pressing an alibi difficult to fathom. The jury found him guilty of all three first-degree murder counts. At the penalty phase, English again conceded McCoy’s guilt, but urged mercy in view of McCoy’s mental and emotional issues. The jury returned three death verdicts. Represented by new counsel, McCoy unsuccessfully sought a new trial. The Louisiana Supreme Court affirmed the trial court’s ruling that English had authority to concede guilt, despite McCoy’s opposition. Held: The Sixth Amendment guarantees a defendant the right to choose the objective of his defense and to insist that his counsel refrain from admitting guilt, even when counsel’s experienced-based view is that confessing guilt offers the defendant the best chance to avoid the death penalty. Pp. 5–13. (a) The Sixth Amendment guarantees to each criminal defendant “the Assistance of Counsel for his defence.” The defendant does not surrender control entirely to counsel, for the Sixth Amendment, in “grant[ing] to the accused personally the right to make his defense,” “speaks of the ‘assistance’ of counsel, and an assistant, however expert, is still an assistant.” Faretta v. California, 422 U. S. 806, 819–820. The lawyer’s province is trial management, but some decisions are reserved for the client—including whether to plead guilty, waive the right to a jury trial, testify in one’s own behalf, and forgo an appeal. Autonomy to decide that the objective of the defense is to assert innocence belongs in this reserved-for-the-client category. Refusing to plead guilty in the face of overwhelming evidence against her, rejecting the assistance of counsel, and insisting on maintaining her innocence at the guilt phase of a capital trial are not strategic choices; they are decisions about what the defendant’s objectives in fact are. See Weaver v. Massachusetts, 582 U. S. ___, ___. Counsel may reasonably assess a concession of guilt as best suited to avoiding the death penalty, as English did here. But the client may not share that objective. He may wish to avoid, above all else, the opprobrium attending admission that he killed family members, or he may hold life in prison not worth living and prefer to risk death for any hope, however small, of exoneration. See Tr. of Oral Arg. 21–22. Thus, when a client makes it plain that the objective of “his defence” is to maintain innocence of the charged criminal acts and pursue an acquittal, his lawyer must abide by that objective and may not override it by conceding guilt. Pp. 5–8. (b) Florida v. Nixon, 543 U. S. 175, is not to the contrary. Nixon’s attorney did not negate Nixon’s autonomy by overriding Nixon’s desired defense objective, for Nixon “was generally unresponsive” during discussions of trial strategy and “never verbally approved or protested” counsel’s proposed approach. Id., at 181. He complained about counsel’s admission of his guilt only after trial. Id., at 185. McCoy, in contrast, opposed English’s assertion of his guilt at every opportunity, before and during trial, both in conference with his lawyer and in open court. Citing Nix v. Whiteside, 475 U. S. 157, the Louisiana Supreme Court concluded that English’s refusal to maintain McCoy’s innocence was necessitated by a Louisiana Rule of Professional Conduct that prohibits counsel from suborning perjury. But in Nix, the defendant told his lawyer that he intended to commit perjury. Here, there was no avowed perjury. English harbored no doubt that McCoy believed what he was saying; English simply disbelieved that account in view of the prosecution’s evidence. Louisiana’s ethical rules might have stopped English from presenting McCoy’s alibi evidence if English knew perjury was involved, but Louisiana has identified no ethical rule requiring English to admit McCoy’s guilt over McCoy’s objection. Pp. 8–11. (c) The Court’s ineffective-assistance-of-counsel jurisprudence, see Strickland v. Washington, 466 U. S. 668, does not apply here, where the client’s autonomy, not counsel’s competence, is in issue. To gain redress for attorney error, a defendant ordinarily must show prejudice. See id., at 692. But here, the violation of McCoy’s protected autonomy right was complete when the court allowed counsel to usurp control of an issue within McCoy’s sole prerogative. Violation of a defendant’s Sixth Amendment-secured autonomy has been ranked “structural” error; when present, such an error is not subject to harmless-error review. See, e.g., McKaskle v. Wiggins, 465 U. S. 168, 177, n. 8; United States v. Gonzalez-Lopez, 548 U. S. 140; Waller v. Georgia, 467 U. S. 39. An error is structural if it is not designed to protect defendants from erroneous conviction, but instead protects some other interest, such as “the fundamental legal principle that a defendant must be allowed to make his own choices about the proper way to protect his own liberty.” Weaver, 582 U. S., at __ (citing Faretta, 422 U. S., at 834). Counsel’s admission of a client’s guilt over the client’s express objection is error structural in kind, for it blocks the defendant’s right to make a fundamental choice about his own defense. See Weaver, 582 U. S., at ___. McCoy must therefore be accorded a new trial without any need first to show prejudice. Pp. 11–12. 2014–1449 (La. 10/19/16), 218 So. 3d 535, reversed and remanded. Ginsburg, J., delivered the opinion of the Court, in which Roberts, C. J., and Kennedy, Breyer, Sotomayor, and Kagan, JJ., joined. Alito, J., filed a dissenting opinion, in which Thomas and Gorsuch, JJ., joined.

In Florida v. Nixon, 543 U. S. 175 (2004), this Court considered whether the Constitution bars defense counsel from conceding a capital defendant’s guilt at trial “when [the] defendant, informed by counsel, neither consents nor objects,” id., at 178. In that case, defense counsel had several times explained to the defendant a proposed guilt-phase concession strategy, but the defendant was unresponsive. Id., at 186. We held that when counsel confers with the defendant and the defendant remains silent, neither approving nor protesting counsel’s proposed concession strategy, id., at 181, “[no] blanket rule demand[s] the defendant’s explicit consent” to implementation of that strategy, id., at 192. In the case now before us, in contrast to Nixon, the defendant vociferously insisted that he did not engage in the charged acts and adamantly objected to any admission of guilt. App. 286–287, 505–506. Yet the trial court permitted counsel, at the guilt phase of a capital trial, to tell the jury the defendant “committed three murders. . . . [H]e’s guilty.” Id., at 509, 510. We hold that a defendant has the right to insist that counsel refrain from admitting guilt, even when counsel’s experienced-based view is that confessing guilt offers the defendant the best chance to avoid the death penalty. Guaranteeing a defendant the right “to have the Assistance of Counsel for his defence,” the Sixth Amendment so demands. With individual lib- erty—and, in capital cases, life—at stake, it is the defendant’s prerogative, not counsel’s, to decide on the objective of his defense: to admit guilt in the hope of gaining mercy at the sentencing stage, or to maintain his innocence, leaving it to the State to prove his guilt beyond a reason- able doubt. I On May 5, 2008, Christine and Willie Young and Gregory Colston were shot and killed in the Youngs’ home in Bossier City, Louisiana. The three victims were the mother, stepfather, and son of Robert McCoy’s estranged wife, Yolanda. Several days later, police arrested McCoy in Idaho. Extradited to Louisiana, McCoy was appointed counsel from the public defender’s office. A Bossier Parish grand jury indicted McCoy on three counts of first-degree murder, and the prosecutor gave notice of intent to seek the death penalty. McCoy pleaded not guilty. Throughout the proceedings, he insistently maintained he was out of State at the time of the killings and that corrupt police killed the victims when a drug deal went wrong. App. 284–286. At defense counsel’s request, a court-appointed sanity commission examined McCoy and found him competent to stand trial. In December 2009 and January 2010, McCoy told the court his relationship with assigned counsel had broken down irretrievably. He sought and gained leave to represent himself until his parents engaged new counsel for him. In March 2010, Larry English, engaged by McCoy’s parents, enrolled as McCoy’s counsel. English eventually concluded that the evidence against McCoy was overwhelming and that, absent a concession at the guilt stage that McCoy was the killer, a death sentence would be impossible to avoid at the penalty phase.[1] McCoy, English reported, was “furious” when told, two weeks before trial was scheduled to begin, that English would concede McCoy’s commission of the triple murders. Id., at 286.[2] McCoy told English “not to make that concession,” and English knew of McCoy’s “complet[e] oppos[ition] to [English] telling the jury that [McCoy] was guilty of killing the three victims”; instead of any concession, McCoy pressed English to pursue acquittal. Id., at 286–287. At a July 26, 2011 hearing, McCoy sought to terminate English’s representation, id., at 449, and English asked to be relieved if McCoy secured other counsel, id., at 458. With trial set to start two days later, the court refused to relieve English and directed that he remain as counsel of record. Id., at 461. “[Y]ou are the attorney,” the court told English when he expressed disagreement with McCoy’s wish to put on a defense case, and “you have to make the trial decision of what you’re going to proceed with.” Id., at 469. At the beginning of his opening statement at the guilt phase of the trial, English told the jury there was “no way reasonably possible” that they could hear the prosecution’s evidence and reach “any other conclusion than Robert McCoy was the cause of these individuals’ death.” Id., at 504. McCoy protested; out of earshot of the jury, McCoy told the court that English was “selling [him] out” by maintaining that McCoy “murdered [his] family.” Id., at 505–506. The trial court reiterated that English was “representing” McCoy and told McCoy that the court would not permit “any other outbursts.” Id., at 506. Continuing his opening statement, English told the jury the evidence is “unambiguous,” “my client committed three murders.” Id., at 509. McCoy testified in his own defense, maintaining his innocence and pressing an alibi difficult to fathom. In his closing argument, English reiterated that McCoy was the killer. On that issue, English told the jury that he “took [the] burden off of [the prosecutor].” Id., at 647. The jury then returned a unanimous verdict of guilty of first-degree murder on all three counts. At the penalty phase, English again conceded “Robert McCoy committed these crimes,” id., at 751, but urged mercy in view of McCoy’s “serious mental and emotional issues,” id., at 755. The jury returned three death verdicts. Represented by new counsel, McCoy unsuccessfully moved for a new trial, arguing that the trial court violated his constitutional rights by allowing English to concede McCoy “committed three murders,” id., at 509, over McCoy’s objection. The Louisiana Supreme Court affirmed the trial court’s ruling that defense counsel had authority so to concede guilt, despite the defendant’s opposition to any admission of guilt. See 2014–1449 (La. 10/19/16), 218 So. 3d 535. The concession was permissible, the court concluded, because counsel reasonably believed that admitting guilt afforded McCoy the best chance to avoid a death sentence. We granted certiorari in view of a division of opinion among state courts of last resort on the question whether it is unconstitutional to allow defense counsel to concede guilt over the defendant’s intransigent and unambiguous objection. 582 U. S. ___ (2017). Compare with the instant case, e.g., Cooke v. State, 977 A. 2d 803, 842–846 (Del. 2009) (counsel’s pursuit of a “guilty but mentally ill” verdict over defendant’s “vociferous and repeated protestations” of innocence violated defendant’s “constitutional right to make the fundamental decisions regarding his case”); State v. Carter, 270 Kan. 426, 440, 14 P. 3d 1138, 1148 (2000) (counsel’s admission of client’s involvement in murder when client adamantly maintained his innocence contravened Sixth Amendment right to counsel and due process right to a fair trial). II A The Sixth Amendment guarantees to each criminal defendant “the Assistance of Counsel for his defence.” At common law, self-representation was the norm. See Fa- retta v. California, 422 U. S. 806, 823 (1975) (citing 1 F. Pollock & F. Maitland, The History of English Law 211 (2d ed. 1909)). As the laws of England and the American Colonies developed, providing for a right to counsel in criminal cases, self-representation remained common and the right to proceed without counsel was recognized. Faretta, 422 U. S., at 824–828. Even now, when most defendants choose to be represented by counsel, see, e.g., Goldschmidt & Stemen, Patterns and Trends in Federal Pro Se Defense, 1996–2011: An Exploratory Study, 8 Fed. Cts. L. Rev. 81, 91 (2015) (0.2% of federal felony defendants proceeded pro se), an accused may insist upon representing herself—however counterproductive that course may be, see Faretta, 422 U. S., at 834. As this Court explained, “[t]he right to defend is personal,” and a defendant’s choice in exercising that right “must be honored out of ‘that respect for the individual which is the lifeblood of the law.’ ” Ibid. (quoting Illinois v. Allen, 397 U. S. 337, 350–351 (1970) (Brennan, J., concurring)); see McKaskle v. Wiggins, 465 U. S. 168, 176–177 (1984) (“The right to appear pro se exists to affirm the dignity and autonomy of the accused.”). The choice is not all or nothing: To gain assistance, a defendant need not surrender control entirely to counsel. For the Sixth Amendment, in “grant[ing] to the accused personally the right to make his defense,” “speaks of the ‘assistance’ of counsel, and an assistant, however expert, is still an assistant.” Faretta, 422 U. S., at 819–820; see Gannett Co. v. DePasquale, 443 U. S. 368, 382, n. 10 (1979) (the Sixth Amendment “contemplat[es] a norm in which the accused, and not a lawyer, is master of his own defense”). Trial management is the lawyer’s province: Counsel provides his or her assistance by making decisions such as “what arguments to pursue, what evidentiary objections to raise, and what agreements to conclude regarding the admission of evidence.” Gonzalez v. United States, 553 U. S. 242, 248 (2008) (internal quotation marks and citations omitted). Some decisions, however, are reserved for the client—notably, whether to plead guilty, waive the right to a jury trial, testify in one’s own behalf, and forgo an appeal. See Jones v. Barnes, 463 U. S. 745, 751 (1983). Autonomy to decide that the objective of the defense is to assert innocence belongs in this latter category. Just as a defendant may steadfastly refuse to plead guilty in the face of overwhelming evidence against her, or reject the assistance of legal counsel despite the defendant’s own inexperience and lack of professional qualifications, so may she insist on maintaining her innocence at the guilt phase of a capital trial. These are not strategic choices about how best to achieve a client’s objectives; they are choices about what the client’s objectives in fact are. See Weaver v. Massachusetts, 582 U. S. ___, ___ (2017) (slip op., at 6) (2017) (self-representation will often increase the likelihood of an unfavorable outcome but “is based on the fundamental legal principle that a defendant must be allowed to make his own choices about the proper way to protect his own liberty”); Martinez v. Court of Appeal of Cal., Fourth Appellate Dist., 528 U. S. 152, 165 (2000) (Scalia, J., concurring in judgment) (“Our system of laws generally presumes that the criminal defendant, after being fully informed, knows his own best interests and does not need them dictated by the State.”). Counsel may reasonably assess a concession of guilt as best suited to avoiding the death penalty, as English did in this case. But the client may not share that objective. He may wish to avoid, above all else, the opprobrium that comes with admitting he killed family members. Or he may hold life in prison not worth living and prefer to risk death for any hope, however small, of exoneration. See Tr. of Oral Arg. 21–22 (it is for the defendant to make the value judgment whether “to take a minuscule chance of not being convicted and spending a life in . . . prison”); Hashimoto, Resurrecting Autonomy: The Criminal Defendant’s Right to Control the Case, 90 B. U. L. Rev. 1147, 1178 (2010) (for some defendants, “the possibility of an acquittal, even if remote, may be more valuable than the difference between a life and a death sentence”); cf. Jae Lee v. United States, 582 U. S. ___, ___ (2017) (slip op., at 12) (recognizing that a defendant might reject a plea and prefer “taking a chance at trial” despite “[a]lmost certai[n]” conviction (emphasis deleted)). When a client expressly asserts that the objective of “his defence” is to maintain innocence of the charged criminal acts, his lawyer must abide by that objective and may not override it by conceding guilt. U. S. Const., Amdt. 6 (emphasis added); see ABA Model Rule of Professional Conduct 1.2(a) (2016) (a “lawyer shall abide by a client’s decisions concerning the objectives of the representation”). Preserving for the defendant the ability to decide whether to maintain his innocence should not displace counsel’s, or the court’s, respective trial management roles. See Gonzalez, 553 U. S., at 249 (“[n]umerous choices affecting conduct of the trial” do not require client consent, including “the objections to make, the witnesses to call, and the arguments to advance”); cf. post, at 8–9. Counsel, in any case, must still develop a trial strategy and discuss it with her client, see Nixon, 543 U. S., at 178, explaining why, in her view, conceding guilt would be the best option. In this case, the court had determined that McCoy was competent to stand trial, i.e., that McCoy had “sufficient present ability to consult with his lawyer with a reason- able degree of rational understanding.” Godinez v. Moran, 509 U. S. 389, 396 (1993) (quoting Dusky v. United States, 362 U. S. 402 (1960) ( per curiam)).[3] If, after consultations with English concerning the management of the defense, McCoy disagreed with English’s proposal to concede McCoy committed three murders, it was not open to English to override McCoy’s objection. English could not interfere with McCoy’s telling the jury “I was not the murderer,” although counsel could, if consistent with providing effective assistance, focus his own collaboration on urging that McCoy’s mental state weighed against conviction. See Tr. of Oral Arg. 21–23. B Florida v. Nixon, see supra, at 1–2, is not to the con- trary. Nixon’s attorney did not negate Nixon’s autonomy by overriding Nixon’s desired defense objective, for Nixon never asserted any such objective. Nixon “was generally unresponsive” during discussions of trial strategy, and “never verbally approved or protested” counsel’s proposed approach. 543 U. S., at 181. Nixon complained about the admission of his guilt only after trial. Id., at 185. McCoy, in contrast, opposed English’s assertion of his guilt at every opportunity, before and during trial, both in conference with his lawyer and in open court. See App. 286–287, 456, 505–506. See also Cooke, 977 A. 2d, at 847 (distinguishing Nixon because, “[i]n stark contrast to the defendant’s silence in that case, Cooke repeatedly objected to his counsel’s objective of obtaining a verdict of guilty but mentally ill, and asserted his factual innocence consistent with his plea of not guilty”). If a client declines to participate in his defense, then an attorney may permissibly guide the defense pursuant to the strategy she believes to be in the defendant’s best interest. Presented with express statements of the client’s will to maintain innocence, however, counsel may not steer the ship the other way. See Gonzalez, 553 U. S., at 254 (Scalia, J., concurring in judgment) (“[A]ction taken by counsel over his client’s objection . . . ha[s] the effect of revoking [counsel’s] agency with respect to the action in question.”). The Louisiana Supreme Court concluded that English’s refusal to maintain McCoy’s innocence was necessitated by Louisiana Rule of Professional Conduct 1.2(d) (2017), which provides that “[a] lawyer shall not counsel a client to engage, or assist a client, in conduct that the lawyer knows is criminal or fraudulent.” 218 So. 3d, at 564. Presenting McCoy’s alibi defense, the court said, would put English in an “ethical conundrum,” implicating English in perjury. Id., at 565 (citing Nix v. Whiteside, 475 U. S. 157, 173–176 (1986)). But McCoy’s case does not resemble Nix, where the defendant told his lawyer that he intended to commit perjury. There was no such avowed perjury here. Cf. ABA Model Rule of Professional Conduct 3.3, Comment 8 (“The prohibition against offering false evidence only applies if the lawyer knows that the evidence is false.”). English harbored no doubt that McCoy believed what he was saying, see App. 285–286; English simply disbelieved McCoy’s account in view of the prosecution’s evidence. English’s express motivation for conceding guilt was not to avoid suborning perjury, but to try to build credibility with the jury, and thus obtain a sentence lesser than death. Id., at 287. Louisiana’s ethical rules might have stopped English from presenting McCoy’s alibi evidence if English knew perjury was involved. But Louisiana has identified no ethical rule requiring English to admit McCoy’s guilt over McCoy’s objection. See 3 W. LaFave, J. Israel, N. King, & O. Kerr, Criminal Procedure §11.6(c), p. 935 (4th ed. 2015) (“A lawyer is not placed in a professionally embarrassing position when he is reluc- tantly required . . . to go to trial in a weak case, since that decision is clearly attributed to his client.”). The dissent describes the conflict between English and McCoy as “rare” and “unlikely to recur.” Post, at 2, 5–7, and n. 2. Yet the Louisiana Supreme Court parted ways with three other State Supreme Courts that have addressed this conflict in the past twenty years. People v. Bergerud, 223 P. 3d 686, 691 (Colo. 2010) (“Although defense counsel is free to develop defense theories based on reasonable assessments of the evidence, as guided by her professional judgment, she cannot usurp those fundamental choices given directly to criminal defendants by the United States and the Colorado Constitutions.”); Cooke, 977 A. 2d 803 (Del. 2009); Carter, 270 Kan. 426, 14 P. 3d 1138 (2000). In each of the three cases, as here, the defendant repeatedly and adamantly insisted on maintaining his factual innocence despite counsel’s preferred course: concession of the defendant’s commission of criminal acts and pursuit of diminished capacity, mental illness, or lack of premeditation defenses. See Bergerud, 223 P. 3d, at 690–691; Cooke, 977 A. 2d, at 814; Carter, 270 Kan., at 429, 14 P. 3d, at 1141. These were not strategic disputes about whether to concede an element of a charged offense, cf. post, at 8; they were intractable disagreements about the fundamental objective of the defendant’s representation. For McCoy, that objective was to maintain “I did not kill the members of my family.” Tr. of Oral Arg. 26. In this stark scenario, we agree with the majority of state courts of last resort that counsel may not admit her client’s guilt of a charged crime over the client’s intransigent objection to that admission. III Because a client’s autonomy, not counsel’s competence, is in issue, we do not apply our ineffective-assistance-of-counsel jurisprudence, Strickland v. Washington, 466 U. S. 668 (1984), or United States v. Cronic, 466 U. S. 648 (1984), to McCoy’s claim. See Brief for Petitioner 43–48; Brief for Respondent 46–52. To gain redress for attorney error, a defendant ordinarily must show prejudice. See Strickland, 466 U. S., at 692. Here, however, the violation of McCoy’s protected autonomy right was complete when the court allowed counsel to usurp control of an issue within McCoy’s sole prerogative. Violation of a defendant’s Sixth Amendment-secured autonomy ranks as error of the kind our decisions have called “structural”; when present, such an error is not subject to harmless-error review. See, e.g., McKaskle, 465 U. S., at 177, n. 8 (harmless-error analysis is inapplicable to deprivations of the self-representation right, because “[t]he right is either respected or denied; its deprivation cannot be harmless”); United States v. Gonzalez-Lopez, 548 U. S. 140, 150 (2006) (choice of counsel is structural); Waller v. Georgia, 467 U. S. 39, 49–50 (1984) (public trial is structural). Structural error “affect[s] the framework within which the trial proceeds,” as distinguished from a lapse or flaw that is “simply an error in the trial process itself.” Arizona v. Fulminante, 499 U. S. 279, 310 (1991). An error may be ranked structural, we have explained, “if the right at issue is not designed to protect the defendant from erroneous conviction but instead protects some other interest,” such as “the fundamental legal principle that a defendant must be allowed to make his own choices about the proper way to protect his own liberty.” Weaver, 582 U. S., at ___ (slip op., at 6) (citing Faretta, 422 U. S., at 834). An error might also count as structural when its effects are too hard to measure, as is true of the right to counsel of choice, or where the error will inevitably signal fundamental unfairness, as we have said of a judge’s failure to tell the jury that it may not convict unless it finds the defendant’s guilt beyond a reasonable doubt. 582 U. S., at ___–___ (slip op., at 6–7) (citing Gonzalez-Lopez, 548 U. S., at 149, n. 4, and Sullivan v. Louisiana, 508 U. S. 275, 279 (1993)). Under at least the first two rationales, counsel’s admission of a client’s guilt over the client’s express objection is error structural in kind. See Cooke, 977 A. 2d, at 849 (“Counsel’s override negated Cooke’s decisions regarding his constitutional rights, and created a structural defect in the proceedings as a whole.”). Such an admission blocks the defendant’s right to make the fundamental choices about his own defense. And the effects of the admission would be immeasurable, because a jury would almost certainly be swayed by a lawyer’s concession of his client’s guilt. McCoy must therefore be accorded a new trial without any need first to show prejudice.[4] * * * Larry English was placed in a difficult position; he had an unruly client and faced a strong government case. He reasonably thought the objective of his representation should be avoidance of the death penalty. But McCoy insistently maintained: “I did not murder my family.” App. 506. Once he communicated that to court and counsel, strenuously objecting to English’s proposed strategy, a concession of guilt should have been off the table. The trial court’s allowance of English’s admission of McCoy’s guilt despite McCoy’s insistent objections was incompatible with the Sixth Amendment. Because the error was structural, a new trial is the required corrective. For the reasons stated, the judgment of the Louisiana Supreme Court is reversed, and the case is remanded for further proceedings not inconsistent with this opinion. It is so ordered. Notes 1 Part of English’s strategy was to concede that McCoy committed the murders and to argue that he should be convicted only of second-degree murder, because his “mental incapacity prevented him from forming the requisite specific intent to commit first degree murder.” 2014–1449 (La. 10/19/16), 218 So. 3d 535, 570. But the second-degree strategy would have encountered a shoal, for Louisiana does not permit introduction of evidence of a defendant’s diminished capacity absent the entry of a plea of not guilty by reason of insanity. Ibid., and n. 35. 2 The dissent states that English told McCoy his proposed trial strategy eight months before trial. Post, at 3. English did encourage McCoy, “[a] couple of months before the trial,” to plead guilty rather than proceed to trial. App. 66–67. But English declared under oath that “the first time [he] told [McCoy] that [he] intended to concede to the jury that [McCoy] was the killer” was July 12, 2011, two weeks before trial commenced. Id., at 286. Encouraging a guilty plea pretrial, of course, is not equivalent to imparting to a defendant counsel’s strategic determination to concede guilt should trial occur. 3 Several times, English did express his view that McCoy was not, in fact, competent to stand trial. See App. 388, 436. 4 The dissent suggests that a remand would be in order, so that the Louisiana Supreme Court, in the first instance, could consider the structural-error question. See post, at 10–11. “[W]e did not grant certiorari to review” that question. Post, at 10. But McCoy raised his structural-error argument in his opening brief, see Brief for Petitioner 38–43, and Louisiana explicitly chose not to grapple with it, see Brief for Respondent 45, n. 5. In any event, “we have the authority to make our own assessment of the harmlessness of a constitutional error in the first instance.” Yates v. Evatt, 500 U. S. 391, 407 (1991) (citing Rose v. Clark, 478 U. S. 570, 584 (1986)).

583.US.2017_16-784

The Bankruptcy Code allows trustees to set aside and recover certain transfers for the benefit of the bankruptcy estate, including, as relevant here, certain fraudulent transfers “of an interest of the debtor in property.” 11 U. S. C. §548(a). It also sets out a number of limits on the exercise of these avoiding powers. Central here is the securities safe harbor, which, inter alia, provides that “the trustee may not avoid a transfer that is a . . . settlement payment . . . made by or to (or for the benefit of) a . . . financial institution . . . or that is a transfer made by or to (or for the benefit of) a . . . financial institution . . . in connection with a securities contract.” §546(e). Valley View Downs, LP, and Bedford Downs Management Corp. entered into an agreement under which Valley View, if it got the last harness-racing license in Pennsylvania, would purchase all of Bedford Downs’ stock for $55 million. Valley View was granted the license and arranged for the Cayman Islands branch of Credit Suisse to wire $55 million to third-party escrow agent Citizens Bank of Pennsylvania. The Bedford Downs shareholders, including petitioner Merit Management Group, LP, deposited their stock certificates into escrow. Citizens Bank disbursed the $55 million over two installments according to the agreement, of which petitioner Merit received $16.5 million. Although Valley View secured the harness-racing license, it was unable to achieve its goal of opening a racetrack casino. Valley View and its parent company, Centaur, LLC, filed for Chapter 11 bankruptcy. Respondent FTI Consulting, Inc., was appointed to serve as trustee of the Centaur litigation trust. FTI then sought to avoid the transfer from Valley View to Merit for the sale of Bedford Downs’ stock, arguing that it was constructively fraudulent under §548(a)(1)(B). Merit contended that the §546(e) safe harbor barred FTI from avoiding the transfer because it was a “settlement payment . . . made by or to (or for the benefit of)” two “financial institutions,” Credit Suisse and Citizens Bank. The District Court agreed with Merit, but the Seventh Circuit reversed, holding that §546(e) did not protect transfers in which financial institutions served as mere conduits. Held: The only relevant transfer for purposes of the §546(e) safe harbor is the transfer that the trustee seeks to avoid. Pp. 9–19. (a) Before a court can determine whether a transfer was “made by or to (or for the benefit of)” a covered entity, it must first identify the relevant transfer to test in that inquiry. Merit posits that the relevant transfer should include not only the Valley-View-to-Merit end-to-end transfer, but also all of its component parts, i.e., the Credit-Suisse-to-Citizens-Bank and the Citizens-Bank-to-Merit transfers. FTI maintains that the only relevant transfer is the transfer that it sought to avoid, specifically, the overarching transfer between Valley View and Merit. Pp. 9–14. (1) The language of §546(e) and the specific context in which that language is used support the conclusion that the relevant transfer for purposes of the safe-harbor inquiry is the transfer the trustee seeks to avoid. The first clause of the provision—“Notwithstanding sections 544, 545, 547, 548(a)(1)(B), and 548(b) of this title”—indicates that §546(e) operates as an exception to trustees’ avoiding powers granted elsewhere in the Code. The text makes clear that the starting point for the §546(e) inquiry is the expressly listed avoiding powers and, consequently, the transfer that the trustee seeks to avoid in exercising those powers. The last clause—“except under section 548(a)(1)(A) of this title”—also focuses on the transfer that the trustee seeks to avoid. Creating an exception to the exception for §548(a)(1)(A) transfers, the text refers back to a specific type of transfer that falls within the avoiding powers, signaling that the exception applies to the overarching transfer that the trustee seeks to avoid, not any component part of that transfer. This reading is reinforced by the §546 section heading, “Limitations on avoiding powers,” and is confirmed by the rest of the statutory text: The provision provides that “the trustee may not avoid” certain transfers, which naturally invites scrutiny of the transfers that “the trustee . . . may avoid,” the parallel language used in the avoiding powers provisions. The text further provides that the transfer that is saved from avoidance is one “that is” (not one that involves) a securities transaction covered under §546(e). In other words, to qualify for protection under the securities safe harbor, §546(e) provides that the otherwise avoidable transfer itself be a transfer that meets the safe-harbor criteria. Pp. 11–13. (2) The statutory structure also supports this reading of §546(e). The Code establishes a system for avoiding transfers as well as a safe harbor from avoidance. It is thus only logical to view the pertinent transfer under §546(e) as the same transfer that the trustee seeks to avoid pursuant to one of its avoiding powers. In an avoidance action, the trustee must establish that the transfer it seeks to set aside meets the carefully set out criteria under the substantive avoidance provisions of the Code. The defendant in that avoidance action is free to argue that the trustee failed to properly identify an avoidable transfer under the Code, including any available arguments concerning the role of component parts of the transfer. If a trustee properly identifies an avoidable transfer, however, the court has no reason to examine the relevance of component parts when considering a limit to the avoiding power, where that limit is defined by reference to an otherwise avoidable transfer, as is the case with §546(e). Pp. 13–14. (b) The primary argument Merit advances that is moored in the statutory text—concerning Congress’ 2006 addition of the parenthetical “(or for the benefit of)” to §546(e)—is unavailing. Merit contends that Congress meant to abrogate the Eleventh Circuit decision in In re Munford, Inc., 98 F. 3d 604, which held that §546(e) was inapplicable to transfers in which a financial institution acted only as an intermediary. However, Merit points to nothing in the text or legislative history to corroborate its argument. A simpler explanation rooted in the text of the statute and consistent with the interpretation of §546(e) adopted here is that Congress added the “or for the benefit of” language that is common in other substantive avoidance provisions to the §546(e) safe harbor to ensure that the scope of the safe harbor and scope of the avoiding powers matched. That reading would not, contrary to what Merit contends, render other provisions ineffectual or superfluous. Rather, it gives full effect to the text of §546(e). If the transfer the trustee seeks to avoid was made “by” or “to” a covered entity, then §546(e) will bar avoidance without regard to whether the entity acted only as an intermediary. It will also bar avoidance if the transfer was made “for the benefit of” that entity, even if it was not made “by” or “to” that entity. Finally, Merit argues that reading the safe harbor so that its application depends on the identity of the investor and the manner in which its investment is held rather than on the general nature of the transaction is incongruous with Congress’ purportedly “prophylactic” approach to §546(e). But this argument is nothing more than an attack on the text of the statute, which protects only certain transactions “made by or to (or for the benefit of)” certain covered entities. Pp. 14–18. (c) Applying this reading of the §546(e) safe harbor to this case yields a straightforward result. FTI sought to avoid the Valley-View-to-Merit transfer. When determining whether the §546(e) safe harbor saves that transfer from avoidance liability, the Court must look to that overarching transfer to evaluate whether it meets the safe-harbor criteria. Because the parties do not contend that either Valley View or Merit is a covered entity, the transfer falls outside of the §546(e) safe harbor. Pp. 18–19. 830 F. 3d 690, affirmed and remanded. Sotomayor, J., delivered the opinion for a unanimous Court.

To maximize the funds available for, and ensure equity in, the distribution to creditors in a bankruptcy proceeding, the Bankruptcy Code gives a trustee the power to invalidate a limited category of transfers by the debtor or transfers of an interest of the debtor in property. Those powers, referred to as “avoiding powers,” are not without limits, however, as the Code sets out a number of exceptions. The operation of one such exception, the securities safe harbor, 11 U. S. C. §546(e), is at issue in this case. Specifically, this Court is asked to determine how the safe harbor operates in the context of a transfer that was executed via one or more transactions, e.g., a transfer from A → D that was executed via B and C as intermediaries, such that the component parts of the transfer include A → B → C → D. If a trustee seeks to avoid the A → D transfer, and the §546(e) safe harbor is invoked as a defense, the question becomes: When determining whether the §546(e) securities safe harbor saves the transfer from avoidance, should courts look to the transfer that the trustee seeks to avoid (i.e., A → D) to determine whether that transfer meets the safe-harbor criteria, or should courts look also to any component parts of the overarching transfer (i.e., A → B → C → D)? The Court concludes that the plain meaning of §546(e) dictates that the only relevant transfer for purposes of the safe harbor is the transfer that the trustee seeks to avoid. I A Because the §546(e) safe harbor operates as a limit to the general avoiding powers of a bankruptcy trustee,[1] we begin with a review of those powers. Chapter 5 of the Bankruptcy Code affords bankruptcy trustees the authority to “se[t] aside certain types of transfers . . . and . . . recaptur[e] the value of those avoided transfers for the benefit of the estate.” Tabb §6.2, p. 474. These avoiding powers “help implement the core principles of bankruptcy.” Id., §6.1, at 468. For example, some “deter the race of diligence of creditors to dismember the debtor before bankruptcy” and promote “equality of distribution.” Union Bank v. Wolas, 502 U. S. 151, 162 (1991) (internal quotation marks omitted); see also Tabb §6.2. Others set aside transfers that “unfairly or improperly deplete . . . assets or . . . dilute the claims against those assets.” 5 Collier on Bankruptcy ¶548.01, p. 548–10 (16th ed. 2017); see also Tabb §6.2, at 475 (noting that some avoiding powers are designed “to ensure that the debtor deals fairly with its creditors”). Sections 544 through 553 of the Code outline the circumstances under which a trustee may pursue avoidance. See, e.g., 11 U. S. C. §544(a) (setting out circumstances under which a trustee can avoid unrecorded liens and conveyances); §544(b) (detailing power to avoid based on rights that unsecured creditors have under nonbankruptcy law); §545 (setting out criteria that allow a trustee to avoid a statutory lien); §547 (detailing criteria for avoidance of so-called “preferential transfers”). The particular avoidance provision at issue here is §548(a), which provides that a “trustee may avoid” certain fraudulent transfers “of an interest of the debtor in property.” §548(a)(1). Section 548(a)(1)(A) addresses so-called “actually” fraudulent transfers, which are “made . . . with actual intent to hinder, delay, or defraud any entity to which the debtor was or became . . . indebted.” Section 548(a)(1)(B) addresses “constructively” fraudulent transfers. See BFP v. Resolution Trust Corporation, 511 U. S. 531, 535 (1994) . As relevant to this case, the statute defines constructive fraud in part as when a debtor: “(B)(i) received less than a reasonably equivalent value in exchange for such transfer or obligation; and “(ii)(I) was insolvent on the date that such transfer was made or such obligation was incurred, or became insolvent as a result of such transfer or obligation. 11 U. S. C. §548(a)(1). If a transfer is avoided, §550 identifies the parties from whom the trustee may recover either the transferred property or the value of that property to return to the bankruptcy estate. Section 550(a) provides, in relevant part, that “to the extent that a transfer is avoided . . . the trustee may recover . . . the property transferred, or, if the court so orders, the value of such property” from “the initial transferee of such transfer or the entity for whose benefit such transfer was made,” or from “any immediate or mediate transferee of such initial transferee.” §550(a). B The Code sets out a number of limits on the exercise of these avoiding powers. See, e.g., §546(a) (setting statute of limitations for avoidance actions); §§546(c)–(d) (setting certain policy-based exceptions to avoiding powers); §548(a)(2) (setting limit to avoidance of “a charitable contribution to a qualified religious or charitable entity or organization”). Central to this case is the securities safe harbor set forth in §546(e), which provides (as presently codified and in full): “Notwithstanding sections 544, 545, 547, 548(a)(1)(B), and 548(b) of this title, the trustee may not avoid a transfer that is a margin payment, as defined in section 101, 741, or 761 of this title, or settlement payment, as defined in section 101 or 741 of this title, made by or to (or for the benefit of ) a commodity broker, forward contract merchant, stockbroker, financial institution, financial participant, or securities clearing agency, or that is a transfer made by or to (or for the benefit of ) a commodity broker, forward contract merchant, stockbroker, financial institution, financial participant, or securities clearing agency, in connection with a securities contract, as defined in section 741(7), commodity contract, as defined in section 761(4), or forward contract, that is made before the commencement of the case, except under section 548(a)(1)(A) of this title.” The predecessor to this securities safe harbor, formerly codified at 11 U. S. C. §764(c), was enacted in 1978 against the backdrop of a district court decision in a case called Seligson v. New York Produce Exchange, 394 F. Supp. 125 (SDNY 1975), which involved a transfer by a bankrupt commodity broker. See S. Rep. No. 95–989, pp. 8, 106 (1978); see also Brubaker, Understanding the Scope of the §546(e) Securities Safe Harbor Through the Concept of the “Transfer” Sought To Be Avoided, 37 Bkrtcy. L. Letter 11–12 (July 2017). The bankruptcy trustee in Seligson filed suit seeking to avoid over $12 million in margin payments made by the commodity broker debtor to a clearing association on the basis that the transfer was constructively fraudulent. The clearing association attempted to defend on the theory that it was a mere “conduit” for the transmission of the margin payments. 394 F. Supp., at 135. The District Court found, however, triable issues of fact on that question and denied summary judgment, leaving the clearing association exposed to the risk of significant liability. See id., at 135–136. Following that decision, Congress enacted the §764(c) safe harbor, providing that “the trustee may not avoid a transfer that is a margin payment to or deposit with a commodity broker or forward contract merchant or is a settlement payment made by a clearing organization.” 92Stat. 2619, codified at 11 U. S. C. §764(c) (repealed 1982). Congress amended the securities safe harbor exception over the years, each time expanding the categories of covered transfers or entities. In 1982, Congress expanded the safe harbor to protect margin and settlement payments “made by or to a commodity broker, forward contract merchant, stockbroker, or securities clearing agency.” §4, 96Stat. 236, codified at 11 U. S. C. §546(d). Two years later Congress added “financial institution” to the list of protected entities. See §461(d), 98Stat. 377, codified at 11 U. S. C. §546(e).[2] In 2005, Congress again expanded the list of protected entities to include a “financial participant” (defined as an entity conducting certain high-value transactions). See §907(b), 119Stat. 181–182; 11 U. S. C. §101(22A). And, in 2006, Congress amended the provision to cover transfers made in connection with securities contracts, commodity contracts, and forward contracts. §5(b)(1), 120Stat. 2697–2698. The 2006 amendment also modified the statute to its current form by adding the new parenthetical phrase “(or for the benefit of )” after “by or to,” so that the safe harbor now covers transfers made “by or to (or for the benefit of )” one of the covered entities. Id., at 2697. C With this background, we now turn to the facts of this case, which comes to this Court from the world of competitive harness racing (a form of horse racing). Harness racing is a closely regulated industry in Pennsylvania, and the Commonwealth requires a license to operate a racetrack. See Bedford Downs Management Corp. v. State Harness Racing Comm’n, 592 Pa. 475, 485–487, 926 A. 2d 908, 914–915 (2007) (per curiam). The number of avail- able licenses is limited, and in 2003 two companies, Valley View Downs, LP, and Bedford Downs Management Corporation, were in competition for the last harness-racing license in Pennsylvania. Valley View and Bedford Downs needed the harness-racing license to open a “ ‘racino,’ ” which is a clever moniker for racetrack casino, “a racing facility with slot machines.” Brief for Petitioner 8. Both companies were stopped before the finish line, because in 2005 the Pennsylvania State Harness Racing Commission denied both applications. The Pennsylvania Supreme Court upheld those denials in 2007, but allowed the companies to reapply for the license. See Bedford Downs, 592 Pa., at 478–479, 926 A. 2d, at 910. Instead of continuing to compete for the last available harness-racing license, Valley View and Bedford Downs entered into an agreement to resolve their ongoing feud. Under that agreement, Bedford Downs withdrew as a competitor for the harness-racing license, and Valley View was to purchase all of Bedford Downs’ stock for $55 million after Valley View obtained the license.[3] With Bedford Downs out of the race, the Pennsylvania Harness Racing Commission awarded Valley View the last harness-racing license. Valley View proceeded with the corporate acquisition required by the parties’ agreement and arranged for the Cayman Islands branch of Credit Suisse to finance the $55 million purchase price as part of a larger $850 million transaction. Credit Suisse wired the $55 million to Citizens Bank of Pennsylvania, which had agreed to serve as the third-party escrow agent for the transaction. The Bedford Downs shareholders, including petitioner Merit Management Group, LP, deposited their stock certificates into escrow as well. At closing, Valley View received the Bedford Downs stock certificates, and in October 2007 Citizens Bank disbursed $47.5 million to the Bedford Downs shareholders, with $7.5 million remaining in escrow at Citizens Bank under the multiyear indemnification holdback period provided for in the parties’ agreement. Citizens Bank disbursed that $7.5 million installment to the Bedford Downs shareholders in October 2010, after the holdback period ended. All told, Merit received approximately $16.5 million from the sale of its Bedford Downs stock to Valley View. Notably, the closing statement for the transaction reflected Valley View as the “Buyer,” the Bedford Downs shareholders as the “Sellers,” and $55 million as the “Purchase Price.” App. 30. In the end, Valley View never got to open its racino. Although it had secured the last harness-racing license, it was unable to secure a separate gaming license for the operation of the slot machines in the time set out in its financing package. Valley View and its parent company, Centaur, LLC, thereafter filed for Chapter 11 bankruptcy. The Bankruptcy Court confirmed a reorganization plan and appointed respondent FTI Consulting, Inc., to serve as trustee of the Centaur litigation trust. FTI filed suit against Merit in the Northern District of Illinois, seeking to avoid the $16.5 million transfer from Valley View to Merit for the sale of Bedford Downs’ stock. The complaint alleged that the transfer was constructively fraudulent under §548(a)(1)(B) of the Code because Valley View was insolvent when it purchased Bedford Downs and “significantly overpaid” for the Bedford Downs stock.[4] Merit moved for judgment on the pleadings under Federal Rule of Civil Procedure 12(c), contending that the §546(e) safe harbor barred FTI from avoiding the Valley View-to-Merit transfer. According to Merit, the safe harbor applied because the transfer was a “settlement payment . . . made by or to (or for the benefit of )” a covered “financial institution”—here, Credit Suisse and Citizens Bank. The District Court granted the Rule 12(c) motion, reasoning that the §546(e) safe harbor applied because the financial institutions transferred or received funds in connection with a “settlement payment” or “securities contract.” See 541 B. R. 850, 858 (ND Ill. 2015).[5] The Court of Appeals for the Seventh Circuit reversed, holding that the §546(e) safe harbor did not protect transfers in which financial institutions served as mere conduits. See 830 F. 3d 690, 691 (2016). This Court granted certiorari to resolve a conflict among the circuit courts as to the proper application of the §546(e) safe harbor.[6] 581 U. S. ___ (2017). II The question before this Court is whether the transfer between Valley View and Merit implicates the safe harbor exception because the transfer was “made by or to (or for the benefit of ) a . . . financial institution.” §546(e). The parties and the lower courts dedicate much of their attention to the definition of the words “by or to (or for the benefit of )” as used in §546(e), and to the question whether there is a requirement that the “financial institution” or other covered entity have a beneficial interest in or dominion and control over the transferred property in order to qualify for safe harbor protection. In our view, those inquiries put the proverbial cart before the horse. Before a court can determine whether a transfer was made by or to or for the benefit of a covered entity, the court must first identify the relevant transfer to test in that inquiry. At bottom, that is the issue the parties dispute in this case. On one side, Merit posits that the Court should look not only to the Valley View-to-Merit end-to-end transfer, but also to all its component parts. Here, those component parts include one transaction by Credit Suisse to Citizens Bank (i.e., the transmission of the $16.5 million from Credit Suisse to escrow at Citizens Bank), and two transactions by Citizens Bank to Merit (i.e., the transmission of $16.5 million over two installments by Citizens Bank as escrow agent to Merit). Because those component parts include transactions by and to financial institutions, Merit contends that §546(e) bars avoidance. FTI, by contrast, maintains that the only relevant transfer for purposes of the §546(e) safe-harbor inquiry is the overarching transfer between Valley View and Merit of $16.5 million for purchase of the stock, which is the transfer that the trustee seeks to avoid under §548(a)(1)(B). Because that transfer was not made by, to, or for the benefit of a financial institution, FTI contends that the safe harbor has no application. The Court agrees with FTI. The language of §546(e), the specific context in which that language is used, and the broader statutory structure all support the conclusion that the relevant transfer for purposes of the §546(e) safe-harbor inquiry is the overarching transfer that the trustee seeks to avoid under one of the substantive avoidance provisions. A Our analysis begins with the text of §546(e), and we look to both “the language itself [and] the specific context in which that language is used . . . .” Robinson v. Shell Oil Co., 519 U. S. 337, 341 (1997) . The pertinent language provides: “Notwithstanding sections 544, 545, 547, 548(a)(1)(B), and 548(b) of this title, the trustee may not avoid a transfer that is a . . . settlement payment . . . made by or to (or for the benefit of ) a . . . financial institution . . . or that is a transfer made by or to (or for the benefit of ) a . . . financial institution . . . in connection with a securities contract . . . , except under section 548(a)(1)(A) of this title.” The very first clause—“Notwithstanding sections 544, 545, 547, 548(a)(1)(B), and 548(b) of this title”—already begins to answer the question. It indicates that §546(e) operates as an exception to the avoiding powers afforded to the trustee under the substantive avoidance provisions. See A. Scalia & B. Garner, Reading Law: The Interpretation of Legal Texts 126 (2012) (“A dependent phrase that begins with notwithstanding indicates that the main clause that it introduces or follows derogates from the provision to which it refers”). That is, when faced with a transfer that is otherwise avoidable, §546(e) provides a safe harbor notwithstanding that avoiding power. From the outset, therefore, the text makes clear that the starting point for the §546(e) inquiry is the substantive avoiding power under the provisions expressly listed in the “notwithstanding” clause and, consequently, the transfer that the trustee seeks to avoid as an exercise of those powers. Then again in the very last clause—“except under section 548(a)(1)(A) of this title”—the text reminds us that the focus of the inquiry is the transfer that the trustee seeks to avoid. It does so by creating an exception to the exception, providing that “the trustee may not avoid a transfer” that meets the covered transaction and entity criteria of the safe harbor, “except” for an actually fraudulent transfer under §548(a)(1)(A). 11 U. S. C. §546(e). By referring back to a specific type of transfer that falls within the avoiding power, Congress signaled that the excep- tion applies to the overarching transfer that the trustee seeks to avoid, not any component part of that transfer. Reinforcing that reading of the safe-harbor provision, the section heading for §546—within which the securities safe harbor is found—is: “Limitations on avoiding powers.” Although section headings cannot limit the plain meaning of a statutory text, see Florida Dept. of Revenue v. Piccadilly Cafeterias, Inc., 554 U. S. 33, 47 (2008) , “they supply cues” as to what Congress intended, see Yates v. United States, 574 U. S. ___, ___ (2015) (slip op., at 10). In this case, the relevant section heading demonstrates the close connection between the transfer that the trustee seeks to avoid and the transfer that is exempted from that avoiding power pursuant to the safe harbor. The rest of the statutory text confirms what the “notwithstanding” and “except” clauses and the section heading begin to suggest. The safe harbor provides that “the trustee may not avoid” certain transfers. §546(e). Naturally, that text invites scrutiny of the transfers that “the trustee may avoid,” the parallel language used in the substantive avoiding powers provisions. See §544(a) (providing that “the trustee . . . may avoid” transfers falling under that provision); §545 (providing that “[t]he trustee may avoid” certain statutory liens); §547(b) (providing that “the trustee may avoid” certain preferential transfers); §548(a)(1) (providing that “[t]he trustee may avoid” certain fraudulent transfers). And if any doubt remained, the language that follows dispels that doubt: The transfer that the “the trustee may not avoid” is specified to be “a transfer that is” either a “settlement payment” or made “in connection with a securities contract.” §546(e) (emphasis added). Not a transfer that involves. Not a transfer that comprises. But a transfer that is a securities transaction covered under §546(e). The provision explicitly equates the transfer that the trustee may otherwise avoid with the transfer that, under the safe harbor, the trustee may not avoid. In other words, to qualify for protection under the securities safe harbor, §546(e) provides that the otherwise avoidable transfer itself be a transfer that meets the safe-harbor criteria. Thus, the statutory language and the context in which it is used all point to the transfer that the trustee seeks to avoid as the relevant transfer for consideration of the §546(e) safe-harbor criteria. B The statutory structure also reinforces our reading of §546(e). See Hall v. United States, 566 U. S. 506, 516 (2012) (looking to statutory structure in interpreting the Bankruptcy Code). As the Seventh Circuit aptly put it, the Code “creates both a system for avoiding transfers and a safe harbor from avoidance—logically these are two sides of the same coin.” 830 F. 3d, at 694; see also Fidelity Financial Services, Inc. v. Fink, 522 U. S. 211, 217 (1998) (“Section 546 of the Code puts certain limits on the avoidance powers set forth elsewhere”). Given that structure, it is only logical to view the pertinent transfer under §546(e) as the same transfer that the trustee seeks to avoid pursuant to one of its avoiding powers. As noted in Part I–A, supra, the substantive avoidance provisions in Chapter 5 of the Code set out in detail the criteria that must be met for a transfer to fall within the ambit of the avoiding powers. These provisions, as Merit admits, “focus mostly on the characteristics of the transfer that may be avoided.” Brief for Petitioner 28. The trustee, charged with exercising those avoiding powers, must establish to the satisfaction of a court that the transfer it seeks to set aside meets the characteristics set out under the substantive avoidance provisions. Thus, the trustee is not free to define the transfer that it seeks to avoid in any way it chooses. Instead, that transfer is necessarily defined by the carefully set out criteria in the Code. As FTI itself recognizes, its power as trustee to define the transfer is not absolute because “the transfer identified must sat- isfy the terms of the avoidance provision the trustee invokes.” Brief for Respondent 23. Accordingly, after a trustee files an avoidance action identifying the transfer it seeks to set aside, a defendant in that action is free to argue that the trustee failed to properly identify an avoidable transfer under the Code, including any available arguments concerning the role of component parts of the transfer. If a trustee properly identifies an avoidable transfer, however, the court has no reason to examine the relevance of component parts when considering a limit to the avoiding power, where that limit is defined by reference to an otherwise avoidable transfer, as is the case with §546(e), see Part II–A, supra. In the instant case, FTI identified the purchase of Bedford Downs’ stock by Valley View from Merit as the transfer that it sought to avoid. Merit does not contend that FTI improperly identified the Valley View-to-Merit transfer as the transfer to be avoided, focusing instead on whether FTI can “ignore” the component parts at the safe-harbor inquiry. Absent that argument, however, the Credit Suisse and Citizens Bank component parts are simply irrelevant to the analysis under §546(e). The focus must remain on the transfer the trustee sought to avoid. III A The primary argument Merit advances that is moored in the statutory text concerns the 2006 addition of the parenthetical “(or for the benefit of )” to §546(e). Merit contends that in adding the phrase “or for the benefit of” to the requirement that a transfer be “made by or to” a protected entity, Congress meant to abrogate the 1998 decision of the Court of Appeals for the Eleventh Circuit in In re Munford, Inc., 98 F.3d 604, 610 (1996) (per curiam), which held that the §546(e) safe harbor was inapplicable to transfers in which a financial institution acted only as an intermediary. Congress abrogated Munford, Merit reasons, by use of the disjunctive “or,” so that even if a beneficial interest, i.e., a transfer “for the benefit of” a financial institution or other covered entity, is sufficient to trigger safe harbor protection, it is not necessary for the financial institution to have a beneficial interest in the transfer for the safe harbor to apply. Merit thus argues that a transaction “by or to” a financial institution such as Credit Suisse or Citizens Bank would meet the requirements of §546(e), even if the financial institution is acting as an intermediary without a beneficial interest in the transfer. Merit points to nothing in the text or legislative history that corroborates the proposition that Congress sought to overrule Munford in its 2006 amendment. There is a simpler explanation for Congress’ addition of this language that is rooted in the text of the statute as a whole and consistent with the interpretation of §546(e) the Court adopts. A number of the substantive avoidance provisions include that language, thus giving a trustee the power to avoid a transfer that was made to “or for the benefit of” certain actors. See §547(b)(1) (avoiding power with respect to preferential transfers “to or for the benefit of a creditor”); §548(a)(1) (avoiding power with respect to certain fraudulent transfers “including any transfer to or for the benefit of an insider . . . ”). By adding the same language to the §546(e) safe harbor, Congress ensured that the scope of the safe harbor matched the scope of the avoiding powers. For example, a trustee seeking to avoid a preferential transfer under §547 that was made “for the benefit of a creditor,” where that creditor is a covered entity under §546(e), cannot now escape application of the §546(e) safe harbor just because the transfer was not “made by or to” that entity. Nothing in the amendment therefore changed the focus of the §546(e) safe-harbor inquiry on the transfer that is otherwise avoidable under the substantive avoiding powers. If anything, by tracking language already included in the substantive avoidance provisions, the amendment reinforces the connection between the inquiry under §546(e) and the otherwise avoidable transfer that the trustee seeks to set aside. Merit next attempts to bolster its reading of the safe harbor by reference to the inclusion of securities clearing agencies as covered entities under §546(e). Because a securities clearing agency is defined as, inter alia, an intermediary in payments or deliveries made in connection with securities transactions, see 15 U. S. C. §78c(23)(A) and 11 U. S. C. §101(48) (defining “securities clearing agency” by reference to the Securities Exchange Act of 1934), Merit argues that the §546(e) safe harbor must be read to protect intermediaries without reference to any beneficial interest in the transfer. The contrary interpretation, Merit contends, “would run afoul of the canon disfavoring an interpretation of a statute that renders a provision ineffectual or superfluous.” Brief for Petitioner 25. Putting aside the question whether a securities clearing agency always acts as an intermediary without a beneficial interest in a challenged transfer—a question that the District Court in Seligson found presented triable issues of fact in that case—the reading of the statute the Court adopts here does not yield any superfluity. Reading §546(e) to provide that the relevant transfer for purposes of the safe harbor is the transfer that the trustee seeks to avoid under a substantive avoiding power, the question then becomes whether that transfer was “made by or to (or for the benefit of )” a covered entity, including a securities clearing agency. If the transfer that the trustee seeks to avoid was made “by” or “to” a securities clearing agency (as it was in Seligson), then §546(e) will bar avoidance, and it will do so without regard to whether the entity acted only as an intermediary. The safe harbor will, in addition, bar avoidance if the transfer was made “for the benefit of” that securities clearing agency, even if it was not made “by” or “to” that entity. This reading gives full effect to the text of §546(e). B In a final attempt to support its proposed interpretation of §546(e), Merit turns to what it perceives was Congress’ purpose in enacting the safe harbor. Specifically, Merit contends that the broad language of §546(e) shows that Congress took a “comprehensive approach to securities and commodities transactions” that “was prophylactic, not surgical,” and meant to “advanc[e] the interests of parties in the finality of transactions.” Brief for Petitioner 41–43. Given that purported broad purpose, it would be incongruous, according to Merit, to read the safe harbor such that its application “would depend on the identity of the investor and the manner in which it held its investment” rather than “the nature of the transaction generally.” Id., at 33. Moreover, Merit posits that Congress’ concern was plainly broader than the risk that is posed by the imposition of avoidance liability on a securities industry entity because Congress provided a safe harbor not only for transactions “to” those entities (thus protecting the entities from direct financial liability), but also “by” these entities to non-covered entities. See Reply Brief 10–14. And, according to Merit, “[t]here is no reason to believe that Congress was troubled by the possibility that transfers by an industry hub could be unwound but yet was unconcerned about trustees’ pursuit of transfers made through industry hubs.” Id., at 12–13 (emphasis in original). Even if this were the type of case in which the Court would consider statutory purpose, see, e.g., Watson v. Philip Morris Cos., 551 U. S. 142 –152 (2007), here Merit fails to support its purposivist arguments. In fact, its perceived purpose is actually contradicted by the plain language of the safe harbor. Because, of course, here we do have a good reason to believe that Congress was concerned about transfers “by an industry hub” specifically: The safe harbor saves from avoidance certain securities transactions “made by or to (or for the benefit of )” covered entities. See §546(e). Transfers “through” a covered entity, conversely, appear nowhere in the statute. And although Merit complains that, absent its reading of the safe harbor, protection will turn “on the identity of the investor and the manner in which it held its investment,” that is nothing more than an attack on the text of the statute, which protects only certain transactions “made by or to (or for the benefit of )” certain covered entities. For these reasons, we need not deviate from the plain meaning of the language used in §546(e). IV For the reasons stated, we conclude that the relevant transfer for purposes of the §546(e) safe harbor is the same transfer that the trustee seeks to avoid pursuant to its substantive avoiding powers. Applying that understanding of the safe-harbor provision to this case yields a straightforward result. FTI, the trustee, sought to avoid the $16.5 million Valley View-to-Merit transfer. FTI did not seek to avoid the component transactions by which that overarching transfer was executed. As such, when determining whether the §546(e) safe harbor saves the transfer from avoidance liability, i.e., whether it was “made by or to (or for the benefit of ) a . . . financial institution,” the Court must look to the overarching transfer from Valley View to Merit to evaluate whether it meets the safe-harbor criteria. Because the parties do not contend that either Valley View or Merit is a “financial institution” or other covered entity, the transfer falls outside of the §546(e) safe harbor. The judgment of the Seventh Circuit is therefore affirmed, and the case is remanded for further proceedings consistent with this opinion. It is so ordered. Notes 1 Avoiding powers may be exercised by debtors, trustees, or creditors’ committees, depending on the circumstances of the case. See generally C. Tabb, Law of Bankruptcy §6.1 (4th ed. 2016) (Tabb). Because this case concerns an avoidance action brought by a trustee, we refer throughout to the trustee in discussing the avoiding power and avoidance action. The resolution of this case is not dependent on the identity of the actor exercising the avoiding power. 2 The term “financial institution” is defined as: “(A) a Federal reserve bank, or an entity that is a commercial or savings bank, industrial savings bank, savings and loan association, trust company, federally-insured credit union, or receiver, liquidating agent, or conservator for such entity and, when any such Federal reserve bank, receiver, liquidating agent, conservator or entity is acting as agent or custodian for a customer (whether or not a ‘customer’, as defined in section 741) in connection with a securities contract (as defined in section 741) such customer; or “(B) in connection with a securities contract (as defined in section 741) an investment company registered under the Investment Company Act of 1940.” 11 U. S. C. §101(22). The parties here do not contend that either the debtor or petitioner in this case qualified as a “financial institution” by virtue of its status as a “customer” under §101(22)(A). Petitioner Merit Management Group, LP, discussed this definition only in footnotes and did not argue that it somehow dictates the outcome in this case. See Brief for Petitioner 45, n. 14; Reply Brief 14, n. 6. We therefore do not address what impact, if any, §101(22)(A) would have in the application of the §546(e) safe harbor. 3 A separate provision of the agreement providing that Bedford Downs would sell land to Valley View for $20 million is not at issue in this case. 4 In its complaint, FTI also sought to avoid the transfer under §544(b). See App. 20–21. The District Court did not address the claim, see 541 B. R. 850, 852–853, n. 1 (ND Ill. 2015), and neither did the Court of Appeals for the Seventh Circuit. 5 The parties do not ask this Court to determine whether the transaction at issue in this case qualifies as a transfer that is a “settlement payment” or made in connection with a “securities contract” as those terms are used in §546(e), nor is that determination necessary for resolution of the question presented. 6 Compare In re Quebecor World (USA) Inc., 719 F. 3d 94, 99 (CA2 2013) (finding the safe harbor applicable where covered entity was intermediary); In re QSI Holdings, Inc., 571 F. 3d 545, 551 (CA6 2009) (same); Contemporary Indus. Corp. v. Frost, 564 F. 3d 981, 987 (CA8 2009) (same); In re Resorts Int’l, Inc., 181 F. 3d 505, 516 (CA3 1999) (same); In re Kaiser Steel Corp., 952 F. 2d 1230, 1240 (CA10 1991) (same), with In re Munford, Inc., 98 F. 3d 604, 610 (CA11 1996) ( per curiam) (rejecting applicability of safe harbor where covered entity was intermediary).

585.US.2017_16-1435

Minnesota law prohibits individuals, including voters, from wearing a “political badge, political button, or other political insignia” inside a polling place on Election Day. Minn. Stat. §211B.11(1) (Supp. 2017). This “political apparel ban” covers articles of clothing and accessories with political insignia upon them. State election judges have the authority to decide whether a particular item falls within the ban. Violators are subject to a civil penalty or prosecution for a petty misdemeanor. Days before the November 2010 election, petitioner Minnesota Voters Alliance (MVA) and other plaintiffs challenged the ban in Federal District Court on First Amendment grounds. In response to the lawsuit, the State distributed an Election Day Policy to election officials providing guidance on enforcement of the ban. The Election Day Policy specified examples of prohibited apparel to include items displaying the name of a political party, items displaying the name of a candidate, items supporting or opposing a ballot question, “[i]ssue oriented material designed to influence or impact voting,” and “[m]aterial promoting a group with recognizable political views.” App. to Pet. for Cert. I–1 to I–2. On Election Day, some voters ran into trouble with the ban, including petitioner Andrew Cilek, who allegedly was turned away from the polls for wearing a “Please I. D. Me” button and a T-shirt bearing the words “Don’t Tread on Me” and a Tea Party Patriots logo. MVA and the other plaintiffs argued that the ban was unconstitutional both on its face and as applied to their particular items of apparel. The District Court granted the State’s motion to dismiss, and the Eighth Circuit affirmed the dismissal of the facial challenge and remanded the case for further proceedings on the as-applied challenge. The District Court granted summary judgment to the State on the as-applied challenge, and the Eighth Circuit affirmed. MVA, Cilek, and petitioner Susan Jeffers (collectively MVA) petitioned for review of their facial First Amendment claim only. Held: Minnesota’s political apparel ban violates the Free Speech Clause of the First Amendment. Pp. 7–19. (a) Because the political apparel ban applies only in a specific location—the interior of a polling place—it implicates the Court’s “ ‘forum based’ approach for assessing restrictions that the government seeks to place on the use of its property.” International Soc. for Krishna Consciousness, Inc. v. Lee, 505 U. S. 672, 678. A polling place in Minnesota qualifies as a nonpublic forum under the Court’s precedents. As such it may be subject to content-based restrictions on speech, see, e.g., Cornelius v. NAACP Legal Defense & Ed. Fund, Inc., 473 U. S. 788, 806–811, so long as the restrictions are “reasonable and not an effort to suppress expression merely because public officials oppose the speaker’s view,” Perry Ed. Assn. v. Perry Local Educators’ Assn., 460 U. S. 37, 46. Because the text of the statute makes no distinction based on the speaker’s political persuasion, the question is whether the apparel ban is “reasonable in light of the purpose served by the forum”: voting. Cornelius, 473 U. S., at 806. Pp. 7–9. (b) Minnesota’s prohibition on political apparel serves a permissible objective. In Burson v. Freeman, 504 U. S. 191, the Court upheld a Tennessee law imposing a 100-foot zone around polling place entrances in which no person could solicit votes, distribute campaign materials, or “display . . . campaign posters, signs or other campaign materials.” 504 U. S., at 193–194 (plurality opinion). In finding that the law withstood even strict scrutiny, the Burson plurality—whose analysis was endorsed by Justice Scalia’s opinion concurring in the judgment—emphasized the problems of fraud, voter intimidation, confusion, and general disorder that had plagued polling places in the past. Against that historical backdrop, the plurality and Justice Scalia upheld Tennessee’s determination that a campaign-free zone outside the polls was necessary to secure the advantages of the secret ballot and protect the right to vote. MVA argues that Burson considered only active campaigning outside the polling place by campaign workers and others trying to engage voters approaching the polls, while Minnesota’s ban prohibits passive self-expression by voters themselves when voting. But although the plurality and Justice Scalia in Burson did not expressly address the application of the Tennessee law to apparel—or consider the interior of the polling place as opposed to its environs—the Tennessee law swept broadly to ban even the plain “display” of a campaign-related message, and the Burson Court upheld the law in full. The plurality’s conclusion that the State was warranted in designating an area for the voters as “their own” as they enter the polling place, id., at 210, suggests an interest more significant, not less, within that place. No basis exists for rejecting Minnesota’s determination that some forms of campaign advocacy should be excluded from the polling place in order to set it aside as “an island of calm in which voters can peacefully contemplate their choices.” Brief for Respondents 43. Casting a vote is a weighty civic act, and the State may reasonably decide that the interior of the polling place should reflect the distinction between voting and campaigning. And while the Court has noted the “nondisruptive” nature of expressive apparel in more mundane settings, see, e.g., Board of Airport Comm’rs of Los Angeles v. Jews for Jesus, Inc., 482 U. S. 569, 576, those observations do not speak to the unique context of a polling place on Election Day. Pp. 9–12. (c) But the line the State draws must be reasonable. The State therefore must be able to articulate some sensible basis for distinguishing what may come in from what must stay out. The unmoored use of the term “political” in the Minnesota law, combined with haphazard interpretations the State has provided in official guidance and representations to this Court, cause Minnesota’s restriction to fail this test. The statute does not define the term “political,” a word that can broadly encompass anything “of or relating to government, a government, or the conduct of governmental affairs.” Webster’s Third New International Dictionary 1755. The State argues that the apparel ban should be interpreted more narrowly to proscribe “only words and symbols that an objectively reasonable observer would perceive as conveying a message about the electoral choices at issue in [the] polling place.” Brief for Respondents 13. At the same time, the State argues that the category of “political” apparel is not limited to campaign apparel. The Court considers a State’s authoritative constructions in interpreting a state law. But far from clarifying the indeterminate scope of the provision, Minnesota’s “electoral choices” construction introduces confusing line-drawing problems. For specific examples of what messages are banned under that standard, the State points to the Election Day Policy. The first three categories of prohibited items in the Policy are clear. But the next category—“issue oriented material designed to influence or impact voting”—raises more questions than it answers. The State takes the position that any subject on which a political candidate or party has taken a stance qualifies as an “issue” within the meaning of that category. Such a rule—whose fair enforcement requires an election judge to maintain a mental index of the platforms and positions of every candidate and party on the ballot—is not reasonable. The next broad category in the Election Day Policy—any item “promoting a group with recognizable political views”—makes matters worse. The State does not confine that category to groups that have endorsed a candidate or taken a position on a ballot question. As a result, any number of associations, educational institutions, businesses, and religious organizations could have an opinion on an “issue confronting voters.” The State represents that the ban is limited to apparel promoting groups with “well-known” political positions. But that requirement only increases the potential for erratic application, as its enforcement may turn in significant part on the background knowledge of the particular election judge applying it. It is “self-evident” that an indeterminate prohibition carries with it “[t]he opportunity for abuse, especially where [it] has received a virtually open-ended interpretation.” Jews for Jesus, 482 U. S., at 576. The discretion election judges exercise in enforcing the ban must be guided by objective, workable standards. Without them, an election judge’s own politics may shape his views on what counts as “political.” And if voters experience or witness episodes of unfair or inconsistent enforcement of the ban, the State’s interest in maintaining a polling place free of distraction and disruption would be undermined by the very measure intended to further it. Thus, if a State wishes to set its polling places apart as areas free of partisan discord, it must employ a more discernible approach than the one offered by Minnesota here. Pp. 12–19. 849 F. 3d 749, reversed and remanded. Roberts, C. J., delivered the opinion of the Court, in which Kennedy, Thomas, Ginsburg, Alito, Kagan, and Gorsuch, JJ., joined. Sotomayor, J., filed a dissenting opinion, in which Breyer, J., joined.

Under Minnesota law, voters may not wear a political badge, political button, or anything bearing political insignia inside a polling place on Election Day. The question presented is whether this ban violates the Free Speech Clause of the First Amendment. I A Today, Americans going to their polling places on Election Day expect to wait in a line, briefly interact with an election official, enter a private voting booth, and cast an anonymous ballot. Little about this ritual would have been familiar to a voter in the mid-to-late nineteenth century. For one thing, voters typically deposited pri- vately prepared ballots at the polls instead of completing official ballots on-site. These pre-made ballots often took the form of “party tickets”—printed slates of candidate selections, often distinctive in appearance, that political parties distributed to their supporters and pressed upon others around the polls. See E. Evans, A History of the Australian Ballot System in the United States 6–11 (1917) (Evans); R. Bensel, The American Ballot Box in the Mid-Nineteenth Century 14–15 (2004) (Bensel). The physical arrangement confronting the voter was also different. The polling place often consisted simply of a “voting window” through which the voter would hand his ballot to an election official situated in a separate room with the ballot box. Bensel 11, 13; see, e.g., C. Rowell, Digest of Contested-Election Cases in the Fifty-First Congress 224 (1891) (report of Rep. Lacey) (considering whether “the ability to reach the window and actually tender the ticket to the [election] judges” is “essential in all cases to constitute a good offer to vote”); Holzer, Election Day 1860, Smithsonian Magazine (Nov. 2008), pp. 46, 52 (describing the interior voting window on the third floor of the Springfield, Illinois courthouse where Abraham Lincoln voted). As a result of this arrangement, “the actual act of voting was usually performed in the open,” frequently within view of interested onlookers. Rusk, The Effect of the Australian Ballot Reform on Split Ticket Voting: 1876–1908, Am. Pol. Sci. Rev. 1220, 1221 (1970) (Rusk); see Evans 11–13. As documented in Burson v. Freeman, 504 U. S. 191 (1992), “[a]pproaching the polling place under this system was akin to entering an open auction place.” Id., at 202 (plurality opinion). The room containing the ballot boxes was “usually quiet and orderly,” but “[t]he public space outside the window . . . was chaotic.” Bensel 13. Electioneering of all kinds was permitted. See id., at 13, 16–17; R. Dinkin, Election Day: A Documentary History 19 (2002). Crowds would gather to heckle and harass voters who appeared to be supporting the other side. Indeed, “[u]nder the informal conventions of the period, election etiquette required only that a ‘man of ordinary courage’ be able to make his way to the voting window.” Bensel 20–21. “In short, these early elections were not a very pleasant spectacle for those who believed in democratic government.” Burson, 504 U. S., at 202 (plurality opinion) (internal quotation marks omitted). By the late nineteenth century, States began implementing reforms to address these vulnerabilities and improve the reliability of elections. Between 1888 and 1896, nearly every State adopted the secret ballot. See id., at 203–205. Because voters now needed to mark their state-printed ballots on-site and in secret, voting moved into a sequestered space where the voters could “deliberate and make a decision in . . . privacy.” Rusk 1221; see Evans 35; 1889 Minn. Stat. ch. 3, §§27–28, p. 21 (regulating, as part of Minnesota’s secret ballot law, the arrangement of voting compartments inside the polling place). In addition, States enacted “viewpoint-neutral restrictions on election-day speech” in the immediate vicinity of the polls. Burson, 504 U. S., at 214–215 (Scalia, J., concurring in judgment) (by 1900, 34 of 45 States had such restrictions). Today, all 50 States and the District of Columbia have laws curbing various forms of speech in and around polling places on Election Day. Minnesota’s such law contains three prohibitions, only one of which is challenged here. See Minn. Stat. §211B.11(1) (Supp. 2017). The first sentence of §211B.11(1) forbids any person to “display campaign material, post signs, ask, solicit, or in any manner try to induce or persuade a voter within a polling place or within 100 feet of the building in which a polling place is situ- ated” to “vote for or refrain from voting for a candidate or ballot question.” The second sentence prohibits the distribution of “political badges, political buttons, or other political insignia to be worn at or about the polling place.” The third sentence—the “political apparel ban”—states that a “political badge, political button, or other political insignia may not be worn at or about the polling place.” Versions of all three prohibitions have been on the books in Minnesota for over a century. See 1893 Minn. Laws ch. 4, §108, pp. 51–52; 1912 Minn. Laws, 1st Spec. Sess., ch. 3, p. 24; 1988 Minn. Laws ch. 578, Art. 3, §11, p. 594 (reenacting the prohibitions as part of §211B.11). There is no dispute that the political apparel ban applies only within the polling place, and covers articles of clothing and accessories with “political insignia” upon them. Minnesota election judges—temporary government employees working the polls on Election Day—have the authority to decide whether a particular item falls within the ban. App. to Pet. for Cert. I–1. If a voter shows up wearing a prohibited item, the election judge is to ask the individual to conceal or remove it. Id., at I–2. If the individual refuses, the election judge must allow him to vote, while making clear that the incident “will be recorded and referred to appropriate authorities.” Ibid. Violators are subject to an administrative process before the Minnesota Office of Administrative Hearings, which, upon finding a violation, may issue a reprimand or impose a civil penalty. Minn. Stat. §§211B.32, 211B.35(2) (2014). That administrative body may also refer the complaint to the county attorney for prosecution as a petty misdemeanor; the maximum penalty is a $300 fine. §§211B.11(4) (Supp. 2017), 211B.35(2) (2014), 609.02(4a) (2016). B Petitioner Minnesota Voters Alliance (MVA) is a nonprofit organization that “seeks better government through election reforms.” Pet. for Cert. 5. Petitioner Andrew Cilek is a registered voter in Hennepin County and the executive director of MVA; petitioner Susan Jeffers served in 2010 as a Ramsey County election judge. Five days before the November 2010 election, MVA, Jeffers, and other likeminded groups and individuals filed a lawsuit in Federal District Court challenging the political apparel ban on First Amendment grounds. The groups—calling themselves “Election Integrity Watch” (EIW)—planned to have supporters wear buttons to the polls printed with the words “Please I. D. Me,” a picture of an eye, and a telephone number and web address for EIW. (Minnesota law does not require individuals to show identification to vote.) One of the individual plaintiffs also planned to wear a “Tea Party Patriots” shirt. The District Court denied the plaintiffs’ request for a temporary restraining order and preliminary injunction and allowed the apparel ban to remain in effect for the upcoming election. In response to the lawsuit, officials for Hennepin and Ramsey Counties distributed to election judges an “Election Day Policy,” providing guidance on the enforcement of the political apparel ban. The Minnesota Secretary of State also distributed the Policy to election officials throughout the State. The Policy specified that examples of apparel falling within the ban “include, but are not limited to”: “• Any item including the name of a political party in Minnesota, such as the Republican, [Democratic-Farmer-Labor], Independence, Green or Libertar- ian parties. • Any item including the name of a candidate at any election. • Any item in support of or opposition to a ballot question at any election. • Issue oriented material designed to influence or impact voting (including specifically the ‘Please I. D. Me’ buttons). • Material promoting a group with recognizable political views (such as the Tea Party, MoveOn.org, and so on).” App. to Pet. for Cert. I–1 to I–2. As alleged in the plaintiffs’ amended complaint and supporting declarations, some voters associated with EIW ran into trouble with the ban on Election Day. One individual was asked to cover up his Tea Party shirt. Another refused to conceal his “Please I. D. Me” button, and an election judge recorded his name and address for possible referral. And petitioner Cilek—who was wearing the same button and a T-shirt with the words “Don’t Tread on Me” and the Tea Party Patriots logo—was twice turned away from the polls altogether, then finally permitted to vote after an election judge recorded his information. Back in court, MVA and the other plaintiffs (now joined by Cilek) argued that the ban was unconstitutional both on its face and as applied to their apparel. The District Court granted the State’s motions to dismiss, and the Court of Appeals for the Eighth Circuit affirmed in part and reversed in part. Minnesota Majority v. Mansky, 708 F. 3d 1051 (2013). In evaluating MVA’s facial challenge, the Court of Appeals observed that this Court had previously upheld a state law restricting speech “related to a political campaign” in a 100-foot zone outside a polling place; the Court of Appeals determined that Minnesota’s law likewise passed constitutional muster. Id., at 1056–1058 (quoting Burson, 504 U. S., at 197 (plurality opinion)). The Court of Appeals reversed the dismissal of the plaintiffs’ as-applied challenge, however, finding that the District Court had improperly considered matters outside the pleadings. 708 F. 3d, at 1059. Judge Shepherd concurred in part and dissented in part. In his view, Minnesota’s broad restriction on political apparel did not “rationally and reasonably” serve the State’s asserted interests. Id., at 1062. On remand, the District Court granted summary judgment for the State on the as-applied challenge, and this time the Court of Appeals affirmed. Minnesota Majority v. Mansky, 849 F. 3d 749 (2017). MVA, Cilek, and Jeffers (hereinafter MVA) petitioned for review of their facial First Amendment claim only. We granted certiorari. 583 U. S. ___ (2017). II The First Amendment prohibits laws “abridging the freedom of speech.” Minnesota’s ban on wearing any “political badge, political button, or other political insignia” plainly restricts a form of expression within the protection of the First Amendment. But the ban applies only in a specific location: the interior of a polling place. It therefore implicates our “ ‘forum based’ approach for assessing restrictions that the government seeks to place on the use of its property.” International Soc. for Krishna Consciousness, Inc. v. Lee, 505 U. S. 672, 678 (1992) (ISKCON). Generally speaking, our cases recognize three types of government-controlled spaces: traditional public forums, designated public forums, and nonpublic forums. In a traditional public forum—parks, streets, sidewalks, and the like—the government may impose reasonable time, place, and manner restrictions on private speech, but restrictions based on content must satisfy strict scrutiny, and those based on viewpoint are prohibited. See Pleasant Grove City v. Summum, 555 U. S. 460, 469 (2009). The same standards apply in designated public forums—spaces that have “not traditionally been regarded as a public forum” but which the government has “intentionally opened up for that purpose.” Id., at 469–470. In a nonpublic forum, on the other hand—a space that “is not by tradition or designation a forum for public communication”—the government has much more flexibility to craft rules limiting speech. Perry Ed. Assn. v. Perry Local Educators’ Assn., 460 U. S. 37, 46 (1983). The government may reserve such a forum “for its intended purposes, communicative or otherwise, as long as the regulation on speech is reasonable and not an effort to suppress expression merely because public officials oppose the speaker’s view.” Ibid. This Court employs a distinct standard of review to assess speech restrictions in nonpublic forums because the government, “no less than a private owner of property,” retains the “power to preserve the property under its control for the use to which it is lawfully dedicated.” Adderley v. Florida, 385 U. S. 39, 47 (1966). “Nothing in the Constitution requires the Government freely to grant access to all who wish to exercise their right to free speech on every type of Government property without regard to the nature of the property or to the disruption that might be caused by the speaker’s activities.” Cornelius v. NAACP Legal Defense & Ed. Fund, Inc., 473 U. S. 788, 799–800 (1985). Accordingly, our decisions have long recognized that the government may impose some content-based restrictions on speech in nonpublic forums, including restrictions that exclude political advocates and forms of political advocacy. See id., at 806–811; Greer v. Spock, 424 U. S. 828, 831–833, 838–839 (1976); Lehman v. Shaker Heights, 418 U. S. 298, 303–304 (1974) (plurality opin- ion); id., at 307–308 (Douglas, J., concurring in judgment). A polling place in Minnesota qualifies as a nonpublic forum. It is, at least on Election Day, government-controlled property set aside for the sole purpose of voting. The space is “a special enclave, subject to greater restriction.” ISKCON, 505 U. S., at 680. Rules strictly govern who may be present, for what purpose, and for how long. See Minn. Stat. §204C.06 (2014). And while the four-Justice plurality in Burson and Justice Scalia’s concurrence in the judgment parted ways over whether the public sidewalks and streets surrounding a polling place qualify as a nonpublic forum, neither opinion suggested that the interior of the building was anything but. See 504 U. S., at 196–197, and n. 2 (plurality opinion); id., at 214–216 (opinion of Scalia, J.). We therefore evaluate MVA’s First Amendment challenge under the nonpublic forum standard. The text of the apparel ban makes no distinction based on the speaker’s political persuasion, so MVA does not claim that the ban discriminates on the basis of viewpoint on its face. The question accordingly is whether Minnesota’s ban on political apparel is “reasonable in light of the purpose served by the forum”: voting. Cornelius, 473 U. S., at 806. III A We first consider whether Minnesota is pursuing a permissible objective in prohibiting voters from wearing particular kinds of expressive apparel or accessories while inside the polling place. The natural starting point for evaluating a First Amendment challenge to such a restriction is this Court’s decision in Burson, which upheld a Tennessee law imposing a 100-foot campaign-free zone around polling place entrances. Under the Tennessee law—much like Minnesota’s buffer-zone provision—no person could solicit votes for or against a candidate, party, or ballot measure, distribute campaign materials, or “display . . . campaign posters, signs or other campaign materials” within the restricted zone. 504 U. S., at 193–194 (plurality opinion). The plurality found that the law withstood even the strict scrutiny applicable to speech restrictions in traditional public forums. Id., at 211. In his opinion concurring in the judgment, Justice Scalia argued that the less rigorous “reasonableness” standard of review should apply, and found the law “at least reasonable” in light of the plurality’s analysis. Id., at 216. That analysis emphasized the problems of fraud, voter intimidation, confusion, and general disorder that had plagued polling places in the past. See id., at 200–204 (plurality opinion). Against that historical backdrop, the plurality and Justice Scalia upheld Tennessee’s determination, supported by overwhelming consensus among the States and “common sense,” that a campaign-free zone outside the polls was “necessary” to secure the advantages of the secret ballot and protect the right to vote. Id., at 200, 206–208, 211. As the plurality explained, “[t]he State of Tennessee has decided that [the] last 15 seconds before its citizens enter the polling place should be their own, as free from interference as possible.” Id., at 210. That was not “an unconstitutional choice.” Ibid. MVA disputes the relevance of Burson to Minnesota’s apparel ban. On MVA’s reading, Burson considered only “active campaigning” outside the polling place by campaign workers and others trying to engage voters approaching the polls. Brief for Petitioners 36–37. Minne- sota’s law, by contrast, prohibits what MVA characterizes as “passive, silent” self-expression by voters themselves when voting. Reply Brief 17. MVA also points out that the plurality focused on the extent to which the restricted zone combated “voter intimidation and election fraud,” 504 U. S., at 208—concerns that, in MVA’s view, have little to do with a prohibition on certain types of voter apparel. Campaign buttons and apparel did come up in the Burson briefing and argument, but neither the plurality nor Justice Scalia expressly addressed such applications of the law.[1] Nor did either opinion specifically consider the interior of the polling place as opposed to its environs, and it is true that the plurality’s reasoning focused on campaign activities of a sort not likely to occur in an area where, for the most part, only voters are permitted while voting. At the same time, Tennessee’s law swept broadly to ban even the plain “display” of a campaign-related message, and the Court upheld the law in full. The plurality’s conclusion that the State was warranted in designating an area for the voters as “their own” as they enter the polling place suggests an interest more significant, not less, within that place. Id., at 210. In any event, we see no basis for rejecting Minnesota’s determination that some forms of advocacy should be excluded from the polling place, to set it aside as “an island of calm in which voters can peacefully contemplate their choices.” Brief for Respondents 43. Casting a vote is a weighty civic act, akin to a jury’s return of a verdict, or a representative’s vote on a piece of legislation. It is a time for choosing, not campaigning. The State may reasonably decide that the interior of the polling place should reflect that distinction. To be sure, our decisions have noted the “nondisruptive” nature of expressive apparel in more mundane settings. Board of Airport Comm’rs of Los Angeles v. Jews for Jesus, Inc., 482 U. S. 569, 576 (1987) (so characterizing “the wearing of a T-shirt or button that contains a political message” in an airport); Tinker v. Des Moines Independent Community School Dist., 393 U. S. 503, 508 (1969) (students wearing black armbands to protest the Vietnam War engaged in “silent, passive expression of opinion, unaccompanied by any disorder or disturbance”). But those observations do not speak to the unique context of a polling place on Election Day. Members of the public are brought together at that place, at the end of what may have been a divisive election season, to reach considered decisions about their government and laws. The State may reasonably take steps to ensure that partisan discord not follow the voter up to the voting booth, and distract from a sense of shared civic obligation at the moment it counts the most. That interest may be thwarted by displays that do not raise significant concerns in other situations. Other States can see the matter differently, and some do.[2] The majority, however, agree with Minnesota that at least some kinds of campaign-related clothing and accessories should stay outside. [3] That broadly shared judgment is entitled to respect. Cf. Burson, 504 U. S., at 206 (plurality opinion) (finding that a “widespread and time-tested consensus” supported the constitutionality of campaign buffer zones). Thus, in light of the special purpose of the polling place itself, Minnesota may choose to prohibit certain apparel there because of the message it conveys, so that voters may focus on the important decisions immediately at hand. B But the State must draw a reasonable line. Although there is no requirement of narrow tailoring in a nonpublic forum, the State must be able to articulate some sensible basis for distinguishing what may come in from what must stay out. See Cornelius, 473 U. S., at 808–809. Here, the unmoored use of the term “political” in the Minnesota law, combined with haphazard interpretations the State has provided in official guidance and representations to this Court, cause Minnesota’s restriction to fail even this forgiving test. Again, the statute prohibits wearing a “political badge, political button, or other political insignia.” It does not define the term “political.” And the word can be expansive. It can encompass anything “of or relating to government, a government, or the conduct of governmental affairs,” Webster’s Third New International Dictionary 1755 (2002), or anything “[o]f, relating to, or dealing with the structure or affairs of government, politics, or the state,” American Heritage Dictionary 1401 (3d ed. 1996). Under a literal reading of those definitions, a button or T-shirt merely imploring others to “Vote!” could qualify. The State argues that the apparel ban should not be read so broadly. According to the State, the statute does not prohibit “any conceivably ‘political’ message” or cover “all ‘political’ speech, broadly construed.” Brief for Respondents 21, 23. Instead, the State interprets the ban to proscribe “only words and symbols that an objectively reasonable observer would perceive as conveying a message about the electoral choices at issue in [the] polling place.” Id., at 13; see id., at 19 (the ban “applies not to any message regarding government or its affairs, but to messages relating to questions of governmental affairs facing voters on a given election day”). At the same time, the State argues that the category of “political” apparel is not limited to campaign apparel. After all, the reference to “campaign material” in the first sentence of the statute—describing what one may not “display” in the buffer zone as well as inside the polling place—implies that the distinct term “political” should be understood to cover a broader class of items. As the State’s counsel explained to the Court, Minnesota’s law “expand[s] the scope of what is prohibited from campaign speech to additional political speech.” Tr. of Oral Arg. 50. We consider a State’s “authoritative constructions” in interpreting a state law. Forsyth County v. Nationalist Movement, 505 U. S. 123, 131 (1992). But far from clarifying the indeterminate scope of the political apparel provision, the State’s “electoral choices” construction introduces confusing line-drawing problems. Cf. Jews for Jesus, 482 U. S., at 575–576 (a resolution banning all “ First Amendment activities” in an airport could not be saved by a “murky” construction excluding “airport-related” activity). For specific examples of what is banned under its standard, the State points to the 2010 Election Day Pol- icy—which it continues to hold out as authoritative guidance regarding implementation of the statute. See Brief for Respondents 22–23. The first three examples in the Policy are clear enough: items displaying the name of a political party, items displaying the name of a candidate, and items demonstrating “support of or opposition to a ballot question.” App. to Pet. for Cert. I–2. But the next example—“[i]ssue oriented material designed to influence or impact voting,” id., at I–2—raises more questions than it answers. What qualifies as an “issue”? The answer, as far as we can tell from the State’s briefing and argument, is any subject on which a political candidate or party has taken a stance. See Tr. of Oral Arg. 37 (explaining that the “electoral choices” test looks at the “issues that have been raised” in a campaign “that are relevant to the election”). For instance, the Election Day Policy specifically notes that the “Please I. D. Me” buttons are prohibited. App. to Pet. for Cert. I–2. But a voter identification requirement was not on the ballot in 2010, see Brief for Respondents 47, n. 24, so a Minnesotan would have had no explicit “electoral choice” to make in that respect. The buttons were nonetheless covered, the State tells us, because the Republican candidates for Governor and Secretary of State had staked out positions on whether photo identification should be required. Ibid.; see App. 58–60.[4] A rule whose fair enforcement requires an election judge to maintain a mental index of the platforms and positions of every candidate and party on the ballot is not reason- able. Candidates for statewide and federal office and major political parties can be expected to take positions on a wide array of subjects of local and national import. See, e.g., Democratic Platform Committee, 2016 Democratic Party Platform (approved July 2016) (stating positions on over 90 issues); Republican Platform Committee, Republican Platform 2016 (approved July 2016) (similar). Would a “Support Our Troops” shirt be banned, if one of the candidates or parties had expressed a view on military funding or aid for veterans? What about a “#MeToo” shirt, referencing the movement to increase awareness of sexual harassment and assault? At oral argument, the State indicated that the ban would cover such an item if a candidate had “brought up” the topic. Tr. of Oral Arg. 64–65. The next broad category in the Election Day Policy—any item “promoting a group with recognizable political views,” App. to Pet. for Cert. I–2—makes matters worse. The State construes the category as limited to groups with “views” about “the issues confronting voters in a given election.” Brief for Respondents 23. The State does not, however, confine that category to groups that have endorsed a candidate or taken a position on a ballot question. Any number of associations, educational institutions, businesses, and religious organizations could have an opinion on an “issue[ ] confronting voters in a given election.” For instance, the American Civil Liberties Union, the AARP, the World Wildlife Fund, and Ben & Jerry’s all have stated positions on matters of public concern.[5] If the views of those groups align or conflict with the position of a candidate or party on the ballot, does that mean that their insignia are banned? See id., at 24, n. 15 (representing that “AFL–CIO or Chamber of Commerce apparel” would be banned if those organizations “had objectively recognizable views on an issue in the election at hand”). Take another example: In the run-up to the 2012 election, Presidential candidates of both major parties issued public statements regarding the then-existing policy of the Boy Scouts of America to exclude members on the basis of sexual orientation.[6] Should a Scout leader in 2012 stopping to vote on his way to a troop meeting have been asked to cover up his uniform? The State emphasizes that the ban covers only apparel promoting groups whose political positions are sufficiently “well-known.” Tr. of Oral Arg. 37. But that requirement, if anything, only increases the potential for erratic application. Well known by whom? The State tells us the lodestar is the “typical observer” of the item. Brief for Respondents 21. But that measure may turn in significant part on the background knowledge and media consumption of the particular election judge applying it. The State’s “electoral choices” standard, considered together with the nonexclusive examples in the Election Day Policy, poses riddles that even the State’s top lawyers struggle to solve. A shirt declaring “All Lives Matter,” we are told, could be “perceived” as political. Tr. of Oral Arg. 41. How about a shirt bearing the name of the National Rifle Association? Definitely out. Id., at 39–40. That said, a shirt displaying a rainbow flag could be worn “unless there was an issue on the ballot” that “related somehow . . . to gay rights.” Id., at 38 (emphasis added). A shirt simply displaying the text of the Second Amendment? Prohibited. Id., at 40. But a shirt with the text of the First Amendment? “It would be allowed.” Ibid. “[P]erfect clarity and precise guidance have never been required even of regulations that restrict expressive activ- ity.” Ward v. Rock Against Racism, 491 U. S. 781, 794 (1989). But the State’s difficulties with its restriction go beyond close calls on borderline or fanciful cases. And that is a serious matter when the whole point of the exercise is to prohibit the expression of political views. It is “self-evident” that an indeterminate prohibition carries with it “[t]he opportunity for abuse, especially where [it] has received a virtually open-ended interpretation.” Jews for Jesus, 482 U. S., at 576; see Heffron v. International Soc. for Krishna Consciousness, Inc., 452 U. S. 640, 649 (1981) (warning of the “more covert forms of discrimination that may result when arbitrary discretion is vested in some governmental authority”). Election judges “have the authority to decide what is political” when screening individuals at the entrance to the polls. App. to Pet. for Cert. I–1. We do not doubt that the vast majority of election judges strive to enforce the statute in an evenhanded manner, nor that some degree of discretion in this setting is necessary. But that discretion must be guided by objective, workable standards. Without them, an election judge’s own politics may shape his views on what counts as “political.” And if voters experience or witness episodes of unfair or inconsistent enforcement of the ban, the State’s interest in maintaining a polling place free of distraction and disruption would be undermined by the very measure intended to further it. That is not to say that Minnesota has set upon an impossible task. Other States have laws proscribing displays (including apparel) in more lucid terms. See, e.g., Cal. Elec. Code Ann. §319.5 (West Cum. Supp. 2018) (prohibiting “the visible display . . . of information that advocates for or against any candidate or measure,” including the “display of a candidate’s name, likeness, or logo,” the “display of a ballot measure’s number, title, subject, or logo,” and “[b]uttons, hats,” or “shirts” containing such information); Tex. Elec. Code Ann. §61.010(a) (West 2010) (prohibiting the wearing of “a badge, insignia, emblem, or other similar communicative device relating to a candidate, measure, or political party appearing on the ballot, or to the conduct of the election”). We do not suggest that such provisions set the outer limit of what a State may proscribe, and do not pass on the constitutionality of laws that are not before us. But we do hold that if a State wishes to set its polling places apart as areas free of partisan discord, it must employ a more discernible approach than the one Minnesota has offered here.[7] * * * Cases like this “present[ ] us with a particularly difficult reconciliation: the accommodation of the right to engage in political discourse with the right to vote.” Burson, 504 U. S., at 198 (plurality opinion). Minnesota, like other States, has sought to strike the balance in a way that affords the voter the opportunity to exercise his civic duty in a setting removed from the clamor and din of electioneering. While that choice is generally worthy of our respect, Minnesota has not supported its good intentions with a law capable of reasoned application. The judgment of the Court of Appeals is reversed, and the case is remanded for further proceedings consistent with this opinion. It is so ordered. APPENDIX State Laws Prohibiting Accessories or Apparel in the Polling Place[8]* Notes 1 The State of Tennessee represented that its prohibition on campaign displays extended both to items of apparel and to voters. Tr. of Oral Arg. in No. 90–1056, p. 33 (argument of Atty. Gen. Burson) (explaining that the statute banned “[t]ee-shirts,” “campaign buttons,” and “hats” because such items “implicate and invite the same problems,” and that voters would be “asked to take campaign button[s] off as they go in”); see Brief for State of Tennessee et al. as Amici Curiae 3, 28–30, and n. 3 (making the same representation in the present case). The Burson plaintiff also emphasized that the Tennessee law would cover apparel, including apparel worn by voters, see Brief for Respondent in No. 90–1056, p. 3; Tr. of Oral Arg. in No. 90–1056, p. 21, and Justice Stevens in dissent referred to the application of the law to campaign buttons, see Burson, 504 U. S., at 218–219, 224. 2 See, e.g., Ala. Secretary of State, 2018 Alabama Voter Guide 14 (voters may wear “campaign buttons or T-shirts with political advertisements”); 2018 Va. Acts ch. 700, §1 (prohibitions on exhibiting campaign material “shall not be construed” to prohibit a voter “from wearing a shirt, hat, or other apparel on which a candidate’s name or a political slogan appears or from having a sticker or button attached to his apparel on which a candidate’s name or a political slogan appears”); R. I. Bd. of Elections, Rules and Regulations for Polling Place Conduct 3 (2016) (voters may “display or wear any campaign or political party button, badge or other document or item designed or tending to aid, injure or defeat any candidate for public office or any political party or any question,” but they must “immediately exit the polling location without unreasonable delay” after voting). 3 See Appendix, infra. 4 The State also maintains that the “Please I. D. Me” buttons were properly banned because the buttons were designed to confuse other voters about whether they needed photo identification to vote. Brief for Respondents 46–47. We do not doubt that the State may prohibit messages intended to mislead voters about voting requirements and procedures. But that interest does not align with the State’s construction of “political” to refer to messages “about the electoral choices at issue in [the] polling place.” Id., at 13. 5 See, e.g., American Civil Liberties Union, Campaign for Smart Justice (2018), online at http://www.aclu.org/issues/mass-incarceration/smart-justice/campaign-smart-justice (taking positions on criminal justice reform) (all Internet materials as last visited June 11, 2018); AARP, Government & Elections, online at https://www.aarp.org/politics-society/government-elections/ (listing positions on Social Secu-rity and health care); World Wildlife Fund, A Win on Capitol Hill (Apr. 17, 2018), online at https://www.worldwildlife.org/stories/a-win-on-capitol-hill (describing the organization’s position on federal funding for international conservation programs); Ben & Jerry’s, Issues We Care About, online at https://www.benjerry.com/values/issues-we-care-about (sharing the corporation’s views on campaign finance reform, international conflict, and civil rights). 6 C. Camia, Obama, Romney Opposed to Boy Scouts Ban on Gays, USA Today OnPolitics (updated Aug. 08, 2012), online at http : / / content.usatoday.com/communities/onpolitics/post/2012/08/barack-obama-boy-scouts-gays-mitt-romney-/1. 7 The State argues that, in the event this Court concludes that there is a “substantial question” about the proper interpretation of §211B.11(1), we should postpone our decision and certify that issue to the Minnesota Supreme Court. Brief for Respondents 57; see Minn. Stat. §480.065(3) (2016). The dissent takes up this cause as well. See post, at 1 (opinion of Sotomayor, J.). The decision to certify, however, “rests in the sound discretion of the federal court.” Expressions Hair Design v. Schneiderman, 581 U. S. ___, ___ (2017) (Sotomayor, J., concurring in judgment) (slip op., at 8). We decline to exercise that discretion in this instance. Minnesota’s request for certification comes very late in the day: This litigation had been ongoing in the federal courts for over seven years before the State made its certification request in its merits brief before this Court. See Stenberg v. Carhart, 530 U. S. 914, 945 (2000) (noting, in denying certification, that the State had never asked the lower federal courts to certify). And the State has not offered sufficient reason to believe that certification would obviate the need to address the constitutional question. Our analysis today reflects the State’s proffered interpretation; nothing in that analysis would change if the State’s interpretation were also adopted by the Minnesota Supreme Court. Nor has the State (or the dissent) suggested a viable alternative construction that the Minnesota Supreme Court might adopt instead. See Brief for Respondents 56–58; post, at 5–8. 8 * Based on statutory or regulatory language and official resources, where available.

583.US.2017_16-1067

Petitioner Charles Murphy was awarded a judgment in his federal civil rights suit against two of his prison guards, including an award of attorney’s fees. Pursuant to 42 U. S. C. §1997e(d)(2), which provides that in such cases “a portion of the [prisoner’s] judgment (not to exceed 25 percent) shall be applied to satisfy the amount of attorney’s fees awarded against the defendant,” the district court ordered Mr. Murphy to pay 10% of his judgment toward the fee award, leaving defendants responsible for the remainder. The Seventh Circuit reversed, holding that §1997e(d)(2) required the district court to exhaust 25% of the prisoner’s judgment before demanding payment from the defendants. Held: In cases governed by §1997e(d), district courts must apply as much of the judgment as necessary, up to 25%, to satisfy an award of attorney’s fees. The specific statutory language supports the Seventh Circuit’s interpretation. First, the mandatory phrase “shall be applied” suggests that the district court has some nondiscretionary duty to perform. Second, the infinitival phrase “to satisfy the amount of attorney’s fees awarded” specifies the purpose or aim of the preceding verb’s nondiscretionary duty. Third, “to satisfy” an obligation, especially a financial obligation, usually means to discharge the obligation in full. Together, these three clues suggest that a district court (1) must act (2) with the purpose of (3) fully discharging the fee award. And the district court must use as much of the judgment as necessary to satisfy the fee award without exceeding the 25% cap. Contrary to Mr. Murphy’s suggestion, the district court does not have wide discretion to pick any “portion” that does not exceed the 25% cap. The larger statutory scheme supports the Seventh Circuit’s interpretation. The previously governing provision, 42 U. S. C. §1988(b), granted district courts discretion to award fees in unambiguous terms. It is doubtful that Congress, had it wished to confer the same sort of discretion in §1997e(d), would have bothered to write a new law for prisoner civil rights suits alone; omit all of the words that afforded discretion in the old law; and then replace those old discretionary words with new mandatory ones. This conclusion is reinforced by §1997e(d)’s surrounding provisions, which like paragraph (2), also limit the district court’s pre-existing discretion under §1988(b). See, e.g., §§1997e(d)(1)(A) and (B)(ii). The discretion urged by Mr. Murphy is exactly the sort of unguided and freewheeling choice that this Court has sought to expunge from practice under §1988. And his suggested cure for rudderless discretion—to have district courts apportion fees in proportion to the defendant’s culpability—has no basis in the statutory text or roots in the law. Pp. 2–9. 844 F. 3d 653, affirmed. Gorsuch, J., delivered the opinion of the Court, in which Roberts, C. J., and Kennedy, Thomas, and Alito, JJ., joined. Sotomayor, J., filed a dissenting opinion, in which Ginsburg, Breyer, and Kagan, JJ., joined.

This is a case about how much prevailing prisoners must pay their lawyers. When a prisoner wins a civil rights suit and the district court awards fees to the prisoner’s attorney, a federal statute says that “a portion of the [prisoner’s] judgment (not to exceed 25 percent) shall be applied to satisfy the amount of attorney’s fees awarded against the defendant. If the award of attorney’s fees is not greater than 150 percent of the judgment, the excess shall be paid by the defendant.” 42 U. S. C. §1997e(d)(2). Whatever else you might make of this, the first sentence pretty clearly tells us that the prisoner has to pay some part of the attorney’s fee award before financial responsibility shifts to the defendant. But how much is enough? Does the first sentence allow the district court discretion to take any amount it wishes from the plaintiff’s judgment to pay the attorney, from 25% down to a penny? Or does the first sentence instead mean that the court must pay the attorney’s entire fee award from the plaintiff’s judgment until it reaches the 25% cap and only then turn to the defendant? The facts of our case illustrate the problem we face. After a jury trial, the district court entered judgment for Charles Murphy in the amount of $307,733.82 against two of his prison guards, Officer Robert Smith and Lieutenant Gregory Fulk. The court also awarded Mr. Murphy’s attorney $108,446.54 in fees. So far, so good. But then came the question who should pay what portion of the fee award. The defendants argued that, under the statute’s terms, the court had to take 25% (or about $77,000) from Mr. Murphy’s judgment before taxing them for the balance of the fee award. The court, however, refused that request. Instead, it ordered that Mr. Murphy “shall pay 10% of [his] judgment” (or about $31,000) toward the fee award, with the defendants responsible for the rest. In support of this allocation, the district court explained that it commonly varied the amount prisoners pay, though the court offered no explanation for choosing 10% instead of some other number. On appeal, a unanimous panel reversed, explaining its view that the language of §1997e(d)(2) requires a district court to exhaust 25% of the prisoner’s judgment before demanding payment from the defendants. 844 F. 3d 653, 660 (CA7 2016). So there we have both sides of the debate, and our question, in a nutshell: did the district court have latitude to apply 10% (or some other discretionary amount) of the plaintiff’s judgment to his attorney’s fee award instead of 25%? See 582 U. S. ___ (2017) (granting certiorari to resolve this question). As always, we start with the specific statutory language in dispute. That language (again) says “a portion of the judgment (not to exceed 25 percent) shall be applied to satisfy the amount of attorney’s fees awarded.” §1997e(d)(2). And we think this much tells us a few things. First, the word “shall” usually creates a mandate, not a liberty, so the verb phrase “shall be applied” tells us that the district court has some nondiscretionary duty to perform. See Lexecon Inc. v. Milberg Weiss Bershad Hynes & Lerach, 523 U. S. 26, 35 (1998) (“[T]he mandatory ‘shall’ . . . normally creates an obligation impervious to judicial discretion”). Second, immediately following the verb we find an infinitival phrase (“to satisfy the amount of attorney’s fees awarded”) that specifies the purpose or aim of the verb’s non-discretionary duty. Cf. R. Huddleston & G. Pullum, Cambridge Grammar of the English Language, ch. 8, §§1, 12.2, pp. 669, 729–730 (2002). Third, we know that when you purposefully seek or aim “to satisfy” an obligation, especially a financial obligation, that usually means you intend to discharge the obligation in full.[1] Together, then, these three clues suggest that the court (1) must apply judgment funds toward the fee award (2) with the purpose of (3) fully discharging the fee award. And to meet that duty, a district court must apply as much of the judgment as necessary to satisfy the fee award, without of course exceeding the 25% cap. If Congress had wished to afford the judge more discretion in this area, it could have easily substituted “may” for “shall.” And if Congress had wished to prescribe a different purpose for the judge to pursue, it could have easily replaced the infinitival phrase “to satisfy . . . ” with “to reduce . . . ” or “against . . . .” But Congress didn’t choose those other words. And respect for Congress’s prerogatives as policymaker means carefully attending to the words it chose rather than replacing them with others of our own. Mr. Murphy’s reply does more to hurt than help his cause. Consider, he says, college math credits that the college prospectus says shall be “applied to satisfy” a chemistry degree. No one, the argument goes, would understand that phrase to suggest a single math course will fully discharge all chemistry degree requirements. We quite agree, but that is beside the point. In Mr. Murphy’s example, as in our statute, the word “satisfy” does not suggest some hidden empirical judgment about how often a math class will satisfy a chemistry degree. Instead it serves to tell the college registrar what purpose he must pursue when handed the student’s transcript: the registrar must, without discretion, apply those credits toward the satisfaction or discharge of the student’s credit obligations. No doubt a college student needing three credits to graduate who took a three-credit math course would be bewildered to learn the registrar thought he had discretion to count only two of those credits toward her degree. So too here. It doesn’t matter how many fee awards will be fully satisfied from a judgment without breaking the 25% cap, or whether any particular fee award could be. The statute’s point is to instruct the judge about the purpose he must pursue—to discharge the fee award using judgment funds to the extent possible, subject to the 25% cap. Retreating now, Mr. Murphy contends that whatever the verb and the infinitival phrase mean, the subject of the sentence—“a portion of the judgment (not to exceed 25 percent)”—necessarily suggests wide judicial discretion. This language, he observes, anticipates a range of amounts (some “portion” up to 25%) that can be taken from his judgment. And the existence of the range, Mr. Murphy contends, necessarily means that the district court must enjoy discretion to pick any “portion” so long as it doesn’t exceed the 25% cap. But that does not logically follow. Under either side’s reading of the statute the portion of fees taken from the plaintiff’s judgment will vary over a range—whether because of the district court’s discretionary choice (as Mr. Murphy contends), or because of the variance in the size of fee awards themselves, which sometimes will be less than 25% of the judgment (as Officer Smith and Lieutenant Fulk suggest). If the police have two suspects in a robbery committed with a red getaway car, the fact that one suspect drives a red sedan proves nothing if the other does too. The fact that the statute contemplates a range of possible “portion[s]” to be paid out of the judgment, thus, just doesn’t help identify which of the two proposed interpretations we should adopt for both bear that feature. Nor does the word “portion” necessarily denote unfettered discretion. If someone told you to follow a written recipe but double the portion of sugar, you would know precisely how much sugar to put in—twice whatever’s on the page. And Congress has certainly used the word “portion” in just that way. Take 16 U. S. C. §673b, which defines the National Elk Refuge to include the “[e]ntire portion now in Jackson Hole National Monument except that portion in section 2 lying west of the east right-of-way line of United States Highway Numbered 187,” among other similar plots—descriptions sufficiently determinate that the statute itself can later give the total number of acres of covered land (“six thousand three hundred and seventy-six acres, more or less”). So the question is how has Congress used the word “portion” in this statute? And as we have explained, the text persuades us that, subject to the 25% cap, the size of the relevant “portion” here is fixed by reference to the size of the attorney’s fee award, not left to a district court’s unguided choice. Even if the interpretive race in this case seems close at this point, close races still have winners. Besides, stepping back to take in the larger statutory scheme surrounding the specific language before us reveals that this case isn’t quite as close as it might first appear. In 1976, Congress enacted what is now 42 U. S. C. §1988(b) to authorize discretionary fee shifting in civil rights suits. Civil Rights Attorney’s Fees Awards Act, 90Stat. 2641. For years that statute governed the award of attorney’s fees in a large variety of civil rights actions, including prisoner civil rights lawsuits like this one. But in the Prison Litigation Reform Act of 1995, Congress reentered the field and adopted §1997e’s new and specialized fee shifting rule for prisoner civil rights suits alone. See 110Stat. 1321–71. Comparing the terms of the old and new statutes helps to shed a good deal of light on the parties’ positions. Section 1988(b) confers discretion on district courts in unambiguous terms: “[T]he court, in its discretion, may allow the prevailing party . . . a reasonable attorney’s fee as part of the costs” against the defendant. (Emphasis added.) Meanwhile, §1997e(d) expressly qualifies the usual operation of §1988(b) in prisoner cases. See §1997e(d)(1) (providing that “[i]n any action brought by a prisoner . . . in which attorney’s fees are authorized under section 1988 . . . such fees shall not be awarded, except” under certain conditions). And as we’ve seen §1997e(d)(2) proceeds to use very different language to describe the district court’s job in awarding fees. It does not say “may,” it does not say “reasonable,” and it certainly does not say anything about “discretion.” If Congress had wished to confer the same discretion in §1997e(d) that it conferred in §1988(b), we very much doubt it would have bothered to write a new law; omit all the words that afforded discretion in the old law; and then replace those old discretionary words with new mandatory ones. See Russello v. United States, 464 U. S. 16, 23 (1983) (refusing to conclude that “the differing language” in two statutory provisions “has the same meaning in each”). The surrounding statutory structure of §1997e(d) reinforces this conclusion. Like paragraph (2), the other provisions of §1997e(d) also limit the district court’s pre-existing discretion under §1988(b). These provisions limit the fees that would otherwise be available under §1988 to cover only certain kinds of lawyerly tasks, see §§1997e(d)(1)(A) and (B)(ii); they require proportionality between fee awards and the relief ordered, see §1997e(d)(1)(B)(i); and they restrict the hourly rate of the prisoner’s lawyer, see §1997e(d)(3). All this suggests a statute that seeks to restrain, rather than replicate, the discretion found in §1988(b). Notably, too, the discretion Mr. Murphy would have us introduce into §1997e doesn’t even sit easily with our precedent under §1988. Our cases interpreting §1988 establish “[a] strong presumption that the lodestar figure—the product of reasonable hours times a reasonable rate—represents a ‘reasonable’ fee.” Pennsylvania v. Delaware Valley Citizens’ Council for Clean Air, 478 U. S. 546, 565 (1986) . To be sure, before the lodestar became “the guiding light of our fee shifting jurisprudence,” Burlington v. Dague, 505 U. S. 557, 562 (1992) , many lower courts used one of your classic 12-factor balancing tests. See Delaware Valley, 478 U. S., at 562, and n. 7. Ultimately, though, this Court rejected undue reliance on the 12-factor test because it “gave very little actual guidance to district courts[,] . . . placed unlimited discretion in trial judges[,] and produced disparate results.” Id., at 563. Yet, despite this guidance, Mr. Murphy effectively seeks to (re)introduce into §1997e(d)(2) exactly the sort of unguided and freewheeling choice—and the disparate results that come with it—that this Court has sought to expunge from practice under §1988. And he seeks to achieve all this on the basis of considerably less helpful statutory language. To state the suggestion is to reveal its defect. Nor does Mr. Murphy’s proposed cure solve his problem. To avoid reading §1997e(d)(2) as affording entirely rudderless discretion, Mr. Murphy contends that district courts should apportion fees in proportion to the defendant’s culpability. When a defendant has acted egregiously, he says, the court should lower the plaintiff’s responsibility for the fee award and increase the defendant’s—even if that means applying only a “nominal” amount of the plaintiff’s judgment toward the fee. But precisely none of this appears in §1997e(d)(2) or, for that matter, enjoys any analogue in §1988’s lodestar analysis or even the old 12-factor approach. Whatever you might have to say about Mr. Murphy’s culpability formula as a matter of policy, it has no roots in the law. Nor is it clear, for what it’s worth, that the culpability approach would even help him. The district court never cited the defendants’ culpability (or any other reason) to justify taking only 10% rather than 25% from Mr. Murphy’s judgment. And it’s tough to see what the choice of 10% might have had to do with the defendant’s culpability in this case. The district court actually remitted the jury’s punitive damages award— suggesting that, if anything, the defendants’ culpability had been already amply addressed. At the end of the day, what may have begun as a close race turns out to have a clear winner. Now with a view of the full field of textual, contextual, and precedential evidence, we think the interpretation the court of appeals adopted prevails. In cases governed by §1997e(d), we hold that district courts must apply as much of the judgment as necessary, up to 25%, to satisfy an award of attorney’s fees.[2] The judgment is Affirmed. Notes 1 See Black’s Law Dictionary 1543 (10th ed. 2014) (defining “satisfaction” as “[t]he fulfillment of an obligation; esp., the payment in full of a debt”); 14 Oxford English Dictionary 504 (2d ed. 1989) (defining “sat-isfy” as “[t]o pay off or discharge fully; to liquidate (a debt); to fulfil completely (an obligation), comply with (a demand)”); Webster’s New International Dictionary 2220 (2d ed. 1950) (defining “satisfy” as “1. In general, to fill up to the measure of a want of (a person or a thing); hence, to gratify fully the desire of . . . . 2. a To pay to the extent of claims or deserts; to give what is due to; as, to satisfy a creditor. b To answer or discharge, as a claim, debt, legal demand, or the like; . . . to pay off ”). 2 Even for those of us who might be inclined to entertain it, Mr. Murphy’s legislative history argument fails to overcome the textual, contextual, and precedential evidence before us. He points to an early draft of §1997e(d)(2) that read: “Whenever a monetary judgment is awarded in an action described in paragraph (1), a portion of the judgment (not to exceed 25 percent) shall be applied to satisfy the amount of attorney’s fees awarded against the defendant. If the award of attorney’s fees is greater than 25 percent of the judgment, the excess shall be paid by the defendant.” Prison Litigation Reform Act of 1995, S. 1279, 104th Cong., 1st Sess., §3(d), p. 16 (1995) (emphasis added). Mr. Murphy admits that the italicized language in the second sentence suggests that it is the size of the attorney’s fees award, not some invisible discretion, that determines what the defendant must pay. Yet, he notes, the second sentence was revised in the legislative process and now reads: “If the award of attorney’s fees is not greater than 150 percent of the judgment, the excess shall be paid by the defendant.” 42 U. S. C. §1997e(d)(2) (emphasis added).

583.US.2017_16-299

The Clean Water Act (Act) generally prohibits “the discharge of any pollutant by any person,” except in express circumstances. 33 U. S. C. §1311(a). A “discharge of a pollutant” includes “any addition of any pollutant to navigable waters from any point source,” §1362(12), and the statutory term “navigable waters,” in turn, means “the waters of the United States,” §1362(7). Section §1311(a) contains important exceptions to the general prohibition on discharge of pollutants, including two permitting schemes that authorize certain entities to discharge pollutants into navigable waters: the National Pollutant Discharge Elimination System (NPDES) program administered by the Environmental Protection Agency (EPA) under §1342, and a program administered by the Army Corps of Engineers (Corps) under §1344. The statutory term “waters of the United States” delineates the geographic reach of those permitting programs as well as other substantive provisions of the Act. In 2015, the EPA and the Corps proffered a definition of that term through an agency regulation dubbed the Waters of the United States Rule (WOTUS Rule or Rule). The WOTUS Rule “imposes no enforceable duty on any state, local, or tribal governments, or the private sector.” 80 Fed. Reg. 37102. As stated in its preamble, the Rule “does not establish any regulatory requirements” and is instead “a definitional rule that clarifies the scope of” the statutory term “waters of the United States.” Id., at 37054. There are two principal avenues of judicial review of an EPA action. Generally, parties may file challenges to final EPA actions in federal district courts, typically under the Administrative Procedure Act. But the Clean Water Act enumerates seven categories of EPA actions for which review lies directly and exclusively in the federal courts of appeals, including, as relevant here, EPA actions “approving or promulgating any effluent limitation or other limitation under section 1311, 1312, 1316, or 1345,” §1369(b)(1)(E), and EPA actions “issuing or denying any permit under section 1342,” §1369(b)(1)(F). Several parties, including petitioner National Association of Manufacturers (NAM), challenged the Rule in United States District Courts across the country. Many parties, but not NAM, filed “protective” petitions for review in various Courts of Appeals to preserve their challenges should their District Court lawsuits be dismissed for lack of jurisdiction under §1369(b). The circuit-court actions were consolidated and transferred to the Court of Appeals for the Sixth Circuit. Meanwhile, the parallel actions in the District Courts continued. NAM intervened as a respondent in the Sixth Circuit and, along with several other parties, moved to dismiss for lack of jurisdiction. The Government opposed those motions, arguing that the challenges must be brought first in the Court of Appeals because the WOTUS Rule fell within subparagraphs (E) and (F) of §1369(b)(1). The Sixth Circuit denied the motions to dismiss. Held: Because the WOTUS Rule falls outside the ambit of §1369(b)(1), challenges to the Rule must be filed in federal district courts. Pp. 9–20. (a) Neither subparagraph (E) nor subparagraph (F) of §1369(b)(1) grants courts of appeals exclusive jurisdiction to review the WOTUS Rule in the first instance. Pp. 9–17. (1) Subparagraph (E) grants courts of appeals exclusive jurisdiction to review any EPA action “in approving or promulgating any effluent limitation or other limitation under section 1311, 1312, 1316, or 1345.” 33 U. S. C. §1369(b)(1)(E). The WOTUS Rule does not fall within that provision. To begin, the Rule is not an “effluent limitation,” which the Act defines as “any restriction . . . on quantities, rates, and concentrations” of certain pollutants “which are discharged from point sources into navigable waters.” §1362(11). The WOTUS Rule imposes no such restriction; instead, it announces a regulatory definition for a statutory term. Nor does the Rule fit within subparagraph (E)’s “other limitation” language. Congress’ use of the phrase “effluent limitation or other limitation” suggests that an “other limitation” must be similar in kind to an “effluent limitation”: that is, a limitation related to the discharge of pollutants. This natural reading is reinforced by subparagraph (E)’s cross-references to §§1311, 1312, 1316, and 1345, which each impose restrictions on the discharge of certain pollutants. The statutory structure thus confirms that an “other limitation” must also be some type of restriction on the discharge of pollutants. Because the WOTUS Rule does no such thing, it falls outside the scope of subparagraph (E). Even if the Government’s reading of “effluent limitation or other limitation” were accepted, however, the Rule still does not fall within subparagraph (E) because it is not a limitation promulgated or approved “under section 1311.” As subparagraph (E)’s statutory context makes clear, this phrase is most naturally read to mean that the effluent or other limitation must be approved or promulgated “pursuant to” or “by reason of the authority of” §1311. But the EPA did not promulgate or approve the WOTUS Rule under §1311, which neither directs nor authorizes the EPA to define a statutory phrase appearing elsewhere in the Act. Rather, the WOTUS Rule was promulgated or approved under §1361(a), which grants the EPA general rulemaking authority “to prescribe such regulations as are necessary to carry out [its] functions under” the Act. The Government contends that the statutory language “under section 1311” poses no barrier to its reading of subparagraph (E) because the WOTUS Rule’s practical effect is to make §1311’s limitations applicable to the waters covered by the Rule. But the Government’s “practical effects” test is not grounded in the statute, renders other statutory language superfluous, and ignores Congress’ decision to grant courts of appeals exclusive jurisdiction only over seven enumerated types of EPA actions set forth in §1369(b)(1). Pp. 9–15. (2) The Government fares no better under subparagraph (F), which grants courts of appeals exclusive and original jurisdiction to review any EPA action “in issuing or denying any permit under section 1342.” §1369(b)(1)(F). That provision does not cover the WOTUS Rule, which neither issues nor denies NPDES permits issued under §1342. Seeking to avoid that conclusion, the Government invokes this Court’s decision in Crown Simpson Pulp Co. v. Costle, 445 U. S. 193 , and argues that the WOTUS Rule falls under subparagraph (F) because it is “functionally similar” to issuing or denying a permit. But that construction misconstrues Crown Simpson, is unmoored from the statutory text, and would create surplusage in other parts of the statute. Pp. 15–17. (b) The Government’s policy arguments provide no basis to depart from the statute’s plain language. First, the Government contends that initial circuit-court review of the WOTUS Rule would avoid a bifurcated judicial-review scheme under which courts of appeals would review individual actions issuing or denying permits, whereas district courts would review broader regulations governing those actions. But, as explained, Congress has made clear that rules like the WOTUS Rule must be reviewed first in federal district courts. Crown Simpson, 445 U. S., at 197, distinguished. Moreover, the bifurcation that the Government bemoans is no more irrational than Congress’ choice to assign challenges to NPDES permits to circuit courts and challenges to §1344 permits to district courts, see §1369(b)(1)(E). And many of this Court’s recent decisions regarding the agencies’ application and definition of “waters of the United States” have originated in district courts, not the courts of appeals. Second, the Court acknowledges that, as the Government argues, routing WOTUS Rule challenges directly to the courts of appeals may improve judicial efficiency. But efficiency was not Congress’ only consideration. Had Congress wanted to prioritize efficiency, it could have authorized direct circuit-court review of all nationally applicable regulations, as it did under the Clean Air Act, instead of structuring judicial review as it did in §1369(b)(1). Third, the Government argues that initial review in the courts of appeals promotes the important goal of national uniformity with regard to broad regulations. Although that argument carries some logical force, Congress did not pursue that end at all costs. Finally, contrary to the Government’s contention, the presumption favoring court-of-appeals review of administrative action does not apply here, for the scope of subparagraphs (E) and (F) is set forth clearly in the statute. Florida Power & Light Co. v. Lorion, 470 U. S. 729 , distinguished. Pp. 17–20. 817 F. 3d 261, reversed and remanded. Sotomayor, J., delivered the opinion for a unanimous Court.

What are the “waters of the United States”? As it turns out, defining that statutory phrase—a central component of the Clean Water Act—is a contentious and difficult task. In 2015, the Environmental Protection Agency (EPA) and the Army Corps of Engineers (Corps) tried their hand at proffering a definition through an agency regulation dubbed the Waters of the United States Rule (WOTUS Rule or Rule).[1] The WOTUS Rule prompted several parties, including petitioner National Association of Manufacturers (NAM), to challenge the regulation in federal court. This case, however, is not about the substantive challenges to the WOTUS Rule. Rather, it is about in which federal court those challenges must be filed. There are two principal avenues of judicial review of an action by the EPA. Generally, parties may file challenges to final EPA actions in federal district courts, ordinarily under the Administrative Procedure Act (APA). But the Clean Water Act (or Act) enumerates seven categories of EPA actions for which review lies directly and exclusively in the federal courts of appeals. See 86Stat. 892, as amended, 33 U. S. C. §1369(b)(1). The Government contends that the WOTUS Rule fits within two of those enumerated categories: (1) EPA actions “in approving or promulgating any effluent limitation or other limitation under section 1311, 1312, 1316, or 1345,” 33 U. S. C. §1369(b)(1)(E), and (2) EPA actions “in issuing or denying any permit under section 1342,” §1369(b)(1)(F). We disagree. The WOTUS Rule falls outside the ambit of §1369(b)(1), and any challenges to the Rule therefore must be filed in federal district courts. I A Although the jurisdictional question in this case is a discrete issue of statutory interpretation, it unfolds against the backdrop of a complex administrative scheme. The Court reviews below the aspects of that scheme that are relevant to the question at hand. 1 Congress enacted the Clean Water Act in 1972 “to restore and maintain the chemical, physical, and biological integrity of the Nation’s waters.” §1251(a). One of the Act’s principal tools in achieving that objective is §1311(a), which prohibits “the discharge of any pollutant by any person,” except in express circumstances. A “discharge of a pollutant” is defined broadly to include “any addition of any pollutant to navigable waters from any point source,” such as a pipe, ditch, or other “discernible, confined and discrete conveyance.” §§1362(12), (14). And “navigable waters,” in turn, means “the waters of the United States, including the territorial seas.” §1362(7). Because many of the Act’s substantive provisions apply to “navigable waters,” the statutory phrase “waters of the United States” circumscribes the geographic scope of the Act in certain respects. Section 1311(a) contains important exceptions to the prohibition on discharge of pollutants. Among them are two permitting schemes that authorize certain entities to discharge pollutants into navigable waters. See Rapanos v. United States, 547 U. S. 715, 723 (2006) (plurality opinion). The first is the National Pollutant Discharge Elimination System (NPDES) program, which is administered by the EPA under §1342. Under that program, the EPA issues permits allowing persons to discharge pollutants that can wash downstream “upon [the] condition that such discharge will meet . . . all applicable requirements under sections 1311, 1312, 1316, 1317, 1318, and 1343.” §1342(a)(1). “NPDES permits impose limitations on the discharge of pollutants, and establish related monitoring and reporting requirements, in order to improve the cleanliness and safety of the Nation’s waters.” Friends of the Earth, Inc. v. Laidlaw Environmental Services (TOC), Inc., 528 U. S. 167, 174 (2000) . One such limitation is an “effluent limitation,” defined in the Act as a “restriction . . . on quantities, rates, and concentrations” of specified pollutants “discharged from point sources into navigable waters, the waters of the contiguous zone, or the ocean, including schedules of compliance.” §1362(11). The second permitting program, administered by the Corps under §1344, authorizes discharges of “ ‘dredged or fill material,’ ” which “are solids that do not readily wash downstream.” Rapanos, 547 U. S., at 723 (plurality opinion). Although the Corps bears primary responsibility in determining whether to issue a §1344 permit, the EPA retains authority to veto the specification of a site for discharge of fill material. See §1344(c).[2] 2 The statutory term “waters of the United States” delineates the geographic reach of many of the Act’s substantive provisions, including the two permitting programs outlined above. In decades past, the EPA and the Corps (collectively, the agencies) have struggled to define and apply that statutory term. See, e.g., 42 Fed. Reg. 37124, 37127 (1977); 51 Fed. Reg. 41216–41217 (1986). And this Court, in turn, has considered those regulatory efforts on several occasions, upholding one such effort as a permissible interpretation of the statute but striking down two others as overbroad. Compare United States v. Riverside Bayview Homes, Inc., 474 U. S. 121 (1985) (upholding the Corps’ interpretation that “waters of the United States” include wetlands adjacent to navigable waters), with Solid Waste Agency of Northern Cook Cty. v. Army Corps of Engineers, 531 U. S. 159 (2001) (rejecting application of the Corps’ interpretation of “waters of the United States” as applied to sand and gravel pit); and Rapanos, 547 U. S., at 729, 757 (plurality opinion) (remanding for further review the Corps’ application of the Act to wetlands lying “near ditches or man-made drains that eventually empty into traditional navigable waters”). In 2015, responding to repeated calls for a more precise definition of “waters of the United States,” the agencies jointly promulgated the WOTUS Rule. 80 Fed. Reg. 37054 (final rule). The WOTUS Rule was intended to “provid[e] simpler, clearer, and more consistent approaches for iden- tifying the geographic scope of the [Act].” Id., at 37057. To that end, the Rule separates waters into three jurisdictional groups—waters that are categorically jurisdictional (e.g., interstate waters); those that require a case-specific showing of their significant nexus to traditionally covered waters (e.g., waters lying in the flood plain of interstate waters); and those that are categorically excluded from jurisdiction (e.g., swimming pools and puddles). See 33 CFR §328.3 (2017); 80 Fed. Reg. 37057. Although the revised regulatory definition “applies broadly to [the Act’s] programs,” the WOTUS Rule itself states that it “imposes no enforceable duty on any state, local, or tribal governments, or the private sector.” 80 Fed. Reg. 37102. Indeed, the Rule’s preamble states that it “does not establish any regulatory requirements” and is instead “a definitional rule that clarifies the scope of” the statutory term “waters of the United States.” Id., at 37054. B As noted above, the Act contemplates two primary avenues for judicial review of EPA actions, each with its own unique set of procedural provisions and statutes of limitations. For “certain suits challenging some agency actions,” the Act grants the federal courts of appeals original and “exclusive” jurisdiction. Decker v. Northwest Environmental Defense Center, 568 U. S. 597 (2013). Seven categories of EPA actions fall within that jurisdictional provision; they include actions of the EPA Administrator— “(A) in promulgating any standard of performance under section 1316 of this title, (B) in making any determination pursuant to section 1316(b)(1)(C) of this title, (C) in promulgating any effluent standard, prohibition, or pretreatment standard under section 1317 of this title, (D) in making any determination as to a State permit program submitted under section 1342(b) of this title, (E) in approving or promulgating any effluent limitation or other limitation under section 1311, 1312, 1316, or 1345 of this title, (F) in issuing or denying any permit under section 1342 of this title, and (G) in promulgating any individual control strategy under section 1314(l) of this title.” 33 U. S. C. §1369(b)(1). To challenge those types of actions, a party must file a petition for review in the court of appeals for the “judicial district in which [the party] resides or transacts business which is directly affected by” the challenged action. Ibid. Any such petition must be filed within 120 days after the date of the challenged action. Ibid. If there are multiple petitions challenging the same EPA action, those petitions are consolidated in one circuit, chosen randomly from among the circuits in which the petitions were filed. See 28 U. S. C. §2112(a)(3). Section 1369(b) also contains a preclusion-of-review provision, which mandates that any agency action reviewable under §1369(b)(1) “shall not be subject to judicial review in any civil or criminal proceeding for enforcement.” 33 U. S. C. §1369(b)(2). The second avenue for judicial review covers final EPA actions falling outside the scope of §1369(b)(1). Those actions are typically governed by the APA.[3] Under the APA, an aggrieved party may file suit in a federal district court to obtain review of any “final agency action for which there is no other adequate remedy in a court.” See 5 U. S. C. §704. Those suits generally must be filed within six years after the claim accrues. 28 U. S. C. §2401(a). C Soon after the agencies promulgated the WOTUS Rule, several parties, including NAM, challenged the Rule in United States District Courts across the country. The Judicial Panel on Multidistrict Litigation (JPML) denied the Government’s request to consolidate and transfer those actions to a single district court. See Order Denying Transfer in In re Clean Water Rule, MDL No. 2663, Doc. 163 (Oct. 13, 2015). Uncertainty surrounding the scope of the Act’s judicial-review provision had also prompted many parties—but not NAM—to file “protective” petitions for review in various Courts of Appeals to preserve their challenges in the event that their District Court lawsuits were dismissed for lack of jurisdiction under §1369(b). The JPML consolidated these appellate-court actions and transferred them to the Court of Appeals for the Sixth Circuit. See Consolidation Order in In re EPA and Dept. of Defense Final Rule, MCP No. 135, Doc. 3 (July 28, 2015). The Court of Appeals thereafter issued a nationwide stay of the WOTUS Rule pending further proceedings. See In re EPA and Dept. of Defense Final Rule, 803 F. 3d 804 (CA6 2015). Meanwhile, parallel litigation continued in the District Courts. Some District Courts dismissed the pending lawsuits, concluding that the courts of appeals had exclusive jurisdiction over challenges to the Rule. See Murray Energy Corp. v. EPA, 2015 WL 5062506, *6 (ND W. Va., Aug. 26, 2015) (dismissing for lack of jurisdiction); Georgia v. McCarthy, 2015 WL 5092568, *3 (SD Ga., Aug. 27, 2015) (concluding that court lacked jurisdiction to enter preliminary injunction). One District Court, by contrast, held that it had jurisdiction to review the WOTUS Rule. See North Dakota v. EPA, 127 F. Supp. 3d 1047, 1052–1053 (ND 2015). NAM intentionally did not file a protective petition in any court of appeals to “ensure that [it] could challenge the Sixth Circuit’s jurisdiction.” Brief for Petitioner 1, n. 1. Instead, NAM intervened as a respondent in the Sixth Circuit and, along with several other parties, moved to dismiss for lack of jurisdiction.[4] The Government opposed those motions, arguing that challenges to the WOTUS Rule must be brought first in the Court of Appeals because the Rule fell within subparagraphs (E) and (F) of §1369(b)(1). The Court of Appeals denied the motions to dismiss in a fractured decision that resulted in three separate opinions. In re Dept. of Defense, 817 F. 3d 261 (2016). The Court of Appeals denied rehearing en banc. We granted certiorari, 580 U. S. ___ (2017), and now reverse.[5] II As noted, §1369(b)(1) enumerates seven categories of EPA actions that must be challenged directly in the fed- eral courts of appeals. Of those seven, only two are at issue in this case: subparagraph (E), which encompasses actions “approving or promulgating any effluent limitation or other limitation under section 1311, 1312, 1316, or 1345,” §1369(b)(1)(E), and subparagraph (F), which covers actions “issuing or denying any [NPDES] permit,” §1369(b)(1)(F).[6] We address each of those statutory provisions in turn. A Subparagraph (E) grants courts of appeals exclusive jurisdiction to review any EPA action “in approving or promulgating any effluent limitation or other limitation under section 1311, 1312, 1316, or 1345.” 33 U. S. C. §1369(b)(1)(E). The Government contends that “EPA’s action in issuing the” WOTUS Rule “readily qualifies as an action promulgating or approving an ‘other limitation’ under section 1311,” because the Rule establishes the “geographic scope of limitations promulgated under Section 1311.” Brief for Federal Respondents 18–19. We disagree. To begin, the WOTUS Rule is not an “effluent limitation”—a conclusion the Government does not meaningfully dispute. An “effluent limitation” is “any restriction . . . on quantities, rates, and concentrations” of certain pollutants “which are discharged from point sources into navigable waters.” §1362(11). The WOTUS Rule imposes no such restriction. Rather, the Rule announces a regulatory definition for a statutory term and “imposes no enforceable duty” on the “private sector.” See 80 Fed. Reg. 37102. The Government instead maintains that the WOTUS Rule is an “other limitation” under subparagraph (E). Although the Act provides no express definition of that residual phrase, the text and structure of subparagraph (E) tell us what that language means. And it is not as broad as the Government insists. For starters, Congress’ use of the phrase “effluent limitation or other limitation” in subparagraph (E) suggests that an “other limitation” must be similar in kind to an “effluent limitation”: that is, a limitation related to the discharge of pollutants. An “other limitation,” for instance, could be a non-numerical operational practice or an equipment specification that, like an “effluent limitation,” restricts the discharge of pollutants, even though such a limitation would not fall within the precise statu- tory definition of “effluent limitation.” That subparagraph (E) cross-references §§1311, 1312, 1316, and 1345 reinforces this natural reading. The unifying feature among those cross-referenced sections is that they impose restrictions on the discharge of certain pollutants. See, e.g., 33 U. S. C. §1311 (imposing general prohibition on “the discharge of any pollutant by any person”); §1312 (governing “water quality related effluent limitations”); §1316 (governing national performance standards for new sources of discharges); §1345 (restricting discharges and use of sewage sludge). In fact, some of those sections give us concrete examples of the type of “other limitation” Congress had in mind. Section 1311(b)(1)(C) allows the EPA to issue “any more stringent limitation[s]” if technology-based effluent limitations cannot “meet water quality standards, treatment standards, or schedules of compliance.” And §1345(d)(3) provides that, if “it is not feasible to prescribe or enforce a numerical limitation” on pollutants in sewage sludge, the EPA may “promulgate a design, equipment, management practice, or operational standard.” All of this demonstrates that an “other limitation,” at a minimum, must also be some type of restriction on the discharge of pollutants. Because the WOTUS Rule does no such thing, it does not fit within the “other limitation” language of subparagraph (E). The Government tries to escape this conclusion by arguing that subparagraph (E) expressly covers “any effluent limitation or other limitation,” §1369(b)(1)(E) (emphasis added), and that the use of the word “any” makes clear that Congress intended subparagraph (E) to sweep broadly and encompass all EPA actions imposing limitations of any sort under the cross-referenced sections. True, use of the word “any” will sometimes indicate that Congress intended particular statutory text to sweep broadly. See, e.g., Ali v. Federal Bureau of Prisons, 552 U. S. 214, 220 (2008) (“Congress’ use of ‘any’ to modify ‘other law enforcement officer’ is most naturally read to mean law enforcement officers of whatever kind”). But whether it does so necessarily depends on the statutory context, and the word “any” in this context does not bear the heavy weight the Government puts upon it. Contrary to the Government’s assertion, the word “any” cannot expand the phrase “other limitation” beyond those limitations that, like effluent limitations, restrict the discharge of pollutants. In urging otherwise, the Government reads the words “effluent limitation and other” completely out of the statute and insists that what Congress really meant to say is “any limitation” under the cross-referenced sections. Of course, those are not the words that Congress wrote, and this Court is not free to “rewrite the statute” to the Government’s liking. Puerto Rico v. Franklin Cal. Tax-Free Trust, 579 U. S. ___, ___ (2016) (slip op., at 14) (“[O]ur constitutional structure does not permit this Court to rewrite the statute that Congress has enacted” (internal quotation marks omitted)). Even if the Court accepted the Government’s reading of “effluent limitation or other limitation,” however, the Rule still does not fall within subparagraph (E) because it is not a limitation promulgated or approved “under section 1311.” [7] §1369(b)(1)(E). This Court has acknowledged that the word “under” is a “chameleon” that “must draw its meaning from its context.” Kucana v. Holder, 558 U. S. 233, 245 (2010) (internal quotation marks omitted). With respect to subparagraph (E), the statutory context makes clear that the prepositional phrase—“under section 1311”—is most naturally read to mean that the effluent limitation or other limitation must be approved or promulgated “pursuant to” or “by reason of the authority of” §1311. See St. Louis Fuel and Supply Co., Inc. v. FERC, 890 F. 2d 446, 450 (CADC 1989) (R. B. Ginsburg, J.) (“ ‘under’ means ‘subject [or pursuant] to’ or ‘by reason of the authority of’ ”); cf. Black’s Law Dictionary 1368 (5th ed. 1979) (defining “under” as “according to”). Here, the EPA did not promulgate or approve the WOTUS Rule under §1311. As noted above, §1311 generally bans the discharge of pollutants into navigable waters absent a permit. Nowhere does that provision direct or authorize the EPA to define a statutory phrase appearing elsewhere in the Act. In fact, the phrase “waters of the United States” does not appear in §1311 at all. Rather, the WOTUS Rule was promulgated or approved under §1361(a), which grants the EPA general rulemaking authority “to prescribe such regulations as are necessary to carry out [its] functions under” the Act. Proving the point, the Government’s own brief cites §1361(a) as the statutory provision that “authorized the [EPA] to issue the [WOTUS] Rule.” Brief for Federal Respondents 17, n. 3.[8] The Government nonetheless insists that the language “under section 1311” poses no barrier to its reading of subparagraph (E) because the “[WOTUS] Rule’s legal and practical effect is to make effluent and other limitations under Section 1311 applicable to the waters that the Rule covers.” Id., at 28. But the Government’s “practical-effects” test is not grounded in the statutory text. Subparagraph (E) encompasses EPA actions that “approv[e] or promulgat[e] any effluent limitation or other limitation under section 1311,” not EPA actions that have the “legal or practical effect” of making such limitations applicable to certain waters. Tellingly, the Government offers no textual basis to read its “practical-effects” test into subparagraph (E). Beyond disregarding the statutory text, the Government’s construction also renders other statutory language superfluous. Take, for instance, subparagraph (E)’s cross-references to §§1312 and 1316. See §1369(b)(1)(E) (covering EPA action “in approving or promulgating any effluent limitation or other limitation under section 1311, 1312, 1316, or 1345” (emphasis added)). Section 1311(a) authorizes discharges that comply with those two cross-referenced sections. See §1311(a) (prohibiting discharge of pollutants “[e]xcept as in compliance with . . . sections 1312, 1316 . . . ”). Thus, EPA actions under §§1312 and 1316 also would have a “legal and practical effect” on the scope of §1311’s general prohibition of discharges, as the Government contends is the case with the WOTUS Rule. If, on the Government’s reading, EPA actions under §§1312 and 1316 would count as actions “under section 1311” sufficient to trigger subparagraph (E), Congress would not have needed to cross-reference §§1312 and 1316 again in subparagraph (E). That Congress did so undercuts the Government’s proposed “practical-effects” test. Similarly, the Government’s “practical-effects” test ignores Congress’ decision to grant appellate courts exclusive jurisdiction only over seven enumerated types of EPA actions set forth in §1369(b)(1). Section 1313, which governs the EPA’s approval and promulgation of state water-quality standards, is a prime example. Approving or promulgating state water-quality standards under §1313 also has the “legal and practical effect” of requiring that effluent limitations be tailored to meet those standards. Under the Government’s reading, subparagraph (E) would encompass EPA actions taken under §1313, even though such actions are nowhere listed in §1369(b)(1). Courts are required to give effect to Congress’ express inclusions and exclusions, not disregard them. See Russello v. United States, 464 U. S. 16, 23 (1983) (“Where Congress includes particular language in one section of a statute but omits it in another section of the same Act, it is generally presumed that Congress acts intentionally and purposely in the disparate inclusion or exclusion” (internal quotation marks and brackets omitted)). Accordingly, subparagraph (E) does not confer original and exclusive jurisdiction on courts of appeals to review the WOTUS Rule. B The Government fares no better under subparagraph (F). That provision grants courts of appeals exclusive and original jurisdiction to review any EPA action “in issu- ing or denying any permit under section 1342.” §1369(b)(1)(F). As explained above, NPDES permits issued under §1342 “authoriz[e] the discharge of pollutants” into certain waters “in accordance with specified conditions.” Gwaltney of Smithfield, Ltd. v. Chesapeake Bay Foundation, Inc., 484 U. S. 49, 52 (1987) . The WOTUS Rule neither issues nor denies a permit under the NPDES permitting program. Because the plain language of subparagraph (F) is “unambiguous,” “our inquiry begins with the statutory text, and ends there as well.” BedRoc Limited, LLC v. United States, 541 U. S. 176, 183 (2004) (plurality opinion). Rather than confront that statutory text, the Government asks us to ignore it altogether. To that end, the Government urges us to apply the “functional interpretive approach” that it purports the Court employed in Crown Simpson Pulp Co. v. Costle, 445 U. S. 193 (1980) (per curiam). Brief for Federal Respondents 31. Crown Simpson, the Government says, broadens the statutory inquiry under subparagraph (F) by directing courts to ask whether agency actions are “ ‘functionally similar’ ” to permit issuances or denials. Brief for Federal Respondents 33 (quoting Crown Simpson, 445 U. S., at 196). According to the Government, the WOTUS Rule is “functionally similar” to issuing or denying a permit because it establishes the geographical bounds of EPA’s permitting authority and thereby dictates whether permits may or may not be issued. We reject this approach because it misconstrues Crown Simpson and ignores the statutory text. First, Crown Simpson provides scant support for the Government’s atextual construction of subparagraph (F). In that case, the Court held that subparagraph (F) conferred jurisdiction over the EPA’s veto of a state-issued permit. See 445 U. S., at 196. The Court explained that “[w]hen [the] EPA . . . objects to effluent limitations contained in a state-issued permit, the precise effect of its action is to ‘den[y]’ a permit within the meaning of [subparagraph F].” Ibid. Contrary to the Government’s suggestion, the WOTUS Rule in no way resembles the EPA’s veto of a state-issued permit addressed in Crown Simpson. Although the WOTUS Rule may define a jurisdictional prerequisite of the EPA’s authority to issue or deny a permit, the Rule itself makes no decision whatsoever on individual permit applications. Crown Simpson is therefore inapposite. In addition, the Government’s proposed “functional interpretive approach” is completely unmoored from the statutory text. As explained above, subparagraph (F) applies only to EPA actions “issuing or denying” a permit “under section 1342.” The Government invites us to broaden that narrow language to cover any agency action that dictates whether a permit is issued or denied. Congress easily could have drafted subparagraph (F) in that broad manner. Indeed, Congress could have said that subparagraph (F) covers EPA actions “relating to whether a permit is issued or denied,” or, alternatively, EPA actions “establishing the boundaries of EPA’s permitting authority.” But Congress chose not to do so. The Court declines the Government’s invitation to override Congress’ considered choice by rewriting the words of the statute. See Franklin Cal. Tax-Free Trust, 579 U. S., at ___ (slip op., at 14). Finally, the Government’s interpretation of subparagraph (F) would create surplusage in other parts of §1369(b)(1). Subparagraph (D) is one example. That provision gives federal appellate courts original jurisdiction to review EPA actions “making any determination as to a State permit program submitted under section 1342(b).” Put differently, subparagraph (D) establishes the boundaries of EPA’s permitting authority vis-à-vis the States. Under the Government’s functional interpretive approach, however, subparagraph (F) would already reach actions delineating the boundaries of EPA’s permitting authority, thus rendering subparagraph (D) unnecessary. Absent clear evidence that Congress intended this surplusage, the Court rejects an interpretation of the statute that would render an entire subparagraph meaningless. As this Court has noted time and time again, the Court is “obliged to give effect, if possible, to every word Congress used.” Reiter v. Sonotone Corp., 442 U. S. 330, 339 (1979) . For these reasons, subparagraph (F) does not grant courts of appeals exclusive jurisdiction to review the WOTUS Rule in the first instance. III A Unable to anchor its preferred reading in the statutory text, the Government seeks refuge in a litany of extratextual considerations that it believes support direct circuit-court review of the WOTUS Rule. Those considerations—alone and in combination—provide no basis to depart from the statute’s plain language. First, the Government contends that initial circuit-court review of the WOTUS Rule would avoid an irrational bifurcated judicial-review scheme under which federal courts of appeals would review individual actions issuing or denying permits, whereas district courts would review broader regulations governing those actions. In E. I. du Pont de Nemours & Co. v. Train, 430 U. S. 112 (1977) , the Court described such a bifurcated regime as a “truly perverse situation.” Id., at 136. And a few years later, in Crown Simpson, the Court declared that “[a]bsent a far clearer expression of congressional intent, we are unwilling to read the Act as creating such a seemingly irrational bifurcated system.” 445 U. S., at 197. Unlike in Crown Simpson, however, here the Court perceives such a “clea[r] expression of congressional intent.” Ibid. Even if the Court might draft the statute differently, Congress made clear that rules like the WOTUS Rule must be reviewed first in federal district courts. Moreover, the bifurcation that the Government bemoans is no more irrational than Congress’ choice to assign challenges to NPDES permits to circuit courts, and challenges to §1344 permits to district courts. See 33 U. S. C. §1369(b)(1)(E). And notably, many of this Court’s recent decisions regarding the agencies’ application and definition of the term “waters of the United States” have originated in district courts, not the courts of appeals. See, e.g., Army Corps of Engineers v. Hawkes Co., 578 U. S. ___ (2016); Sackett v. EPA, 566 U. S. 120 (2012) ; Rapanos, 547 U. S., at 729 (plurality opinion). Second, and relatedly, the Government argues that immediate court-of-appeals review facilitates quick and orderly resolution of disputes about the WOTUS Rule. We acknowledge that routing WOTUS Rule challenges di- rectly to the courts of appeals may improve judicial efficiency. See Crown Simpson, 445 U. S., at 197 (noting that “the additional level of judicial review” that would occur in district courts “would likely cause delays in resolving disputes under the Act”); see also Harrison v. PPG Industries, Inc., 446 U.S. 578, 593 (1980) (“The most obvious advantage of direct review by a court of appeals is the time saved compared to review by a district court, followed by a second review on appeal”). But efficiency was not Congress’ only consideration. Had Congress wanted to prioritize efficiency, it could have authorized direct circuit-court review of all nationally applicable regulations, as it did under the Clean Air Act. See 42 U. S. C. §7607(b)(1) (granting the D. C. Circuit original jurisdiction to review “any other nationally applicable regulations promulgated, or final action taken, by the Administrator under this chapter” and granting regional circuits jurisdiction to review “any other final action of the Administrator under this chapter . . . which is locally or regionally applicable”). That Congress structured judicial review under the Act differently confirms what the text makes clear—that §1369(b)(1) does not grant courts of appeals original jurisdiction to review many types of EPA action, including the WOTUS Rule. Third, the Government contends that “initial review in a court of appeals” promotes “ ‘[n]ational uniformity, an important goal in dealing with broad regulations.’ ” Brief for Federal Respondents 35 (quoting National Resources Defense Council v. EPA, 673 F.2d 400, 405, n. 15 (CADC 1982) (R. B. Ginsburg, J.)). That argument carries some logical force. After all, the numerous challenges to the WOTUS Rule in this very case were consolidated in one Court of Appeals, avoiding any risk of conflict among other courts of appeals, whereas the same was not true for the challenges filed in district courts, leading to some conflicting outcomes. But even if Congress sought to ensure national uniformity, it did not pursue that end at all costs. Although §1369(b)(1) does not authorize immediate circuit-court review of all national rules under the Act, it does permit federal appellate courts to review directly certain effluent and other limitations and individual permit decisions. See, e.g., §§1369(b)(1)(E), (F). It is true that Congress could have funneled all challenges to national rules to the courts of appeals, but it chose a different tack here: It carefully enumerated the seven categories of EPA action for which it wanted immediate circuit- court review and relegated the rest to the jurisdiction of the federal district courts. Ultimately, the Government’s policy arguments do not obscure what the statutory language makes clear: Subparagraphs (E) and (F) do not grant courts of appeals exclusive jurisdiction to review the WOTUS Rule. B In a final effort to bolster its preferred reading of the Act, the Government invokes the presumption favoring court-of-appeals review of administrative action. According to the Government, when a direct-review provision like §1369(b)(1) exists, this Court “will not presume that Congress intended to depart from the sound policy of placing initial . . . review in the courts of appeals” “[a]bsent a firm indication that Congress intended to locate initial APA review of agency action in the district courts.” Florida Power & Light Co. v. Lorion, 470 U. S. 729, 745 (1985) . But the Government’s reliance on Florida Power is misplaced. Unlike the “ambiguous” judicial review provisions at issue in Florida Power, id., at 737, the scope of subparagraphs (E) and (F) is set forth clearly in the statute. As the Court recognized in Florida Power, jurisdiction is “governed by the intent of Congress and not by any views we may have about sound policy.” Id., at 746. Here, Congress’ intent is clear from the statutory text.[9] IV For the foregoing reasons, we reverse the judgment of the Court of Appeals and remand the case with in- structions to dismiss the petitions for review for lack of jurisdiction. It is so ordered. Notes 1 We note that some of the parties and the Court of Appeals below refer to the WOTUS Rule as the “Clean Water Rule.” Throughout this opinion, we have opted to use the former term in lieu of the latter. 2 Both permitting programs allow the States to operate their own permitting schemes to govern waters within their borders. See 33 U. S. C. §§1342(b), 1344(g). Many States have opted to operate an NPDES permitting program under §1342(b), and two have done so under §1344(g). 3 The Act also grants federal district courts jurisdiction over certain kinds of citizen enforcement actions. See 33 U. S. C. §1365(a); Decker, 568 U. S., at 607–08. 4 Some of the parties who filed protective petitions moved to dismiss those same petitions, agreeing with NAM that direct review of the WOTUS Rule belonged in the United States District Courts. Many of those parties, though nominally respondents before this Court, filed briefs in support of NAM. 5 There have been a number of developments since the Court granted review in this case. In February 2017, the President issued an Executive Order directing the agencies to propose a rule rescinding or revising the WOTUS Rule. See Exec. Order No. 13778, 82 Fed. Reg. 12497. On July 27, 2017, the agencies responded to that directive by issuing a proposed rule. See Definition of “Waters of the United States”—Recodification of Pre-Existing Rules, 82 Fed. Reg. 34899, 34901–34902. That proposed rule, once implemented, would rescind the WOTUS Rule and recodify the pre-2015 regulatory definition of “waters of the United States.” See ibid. Then, in November 2017, following oral argument in this case, the agencies issued a second proposed rule establishing a new effective date for the WOTUS Rule. Definition of “Waters of the United States”—Addition of an Applicability Date to 2015 Clean Water Rule, 82 Fed. Reg. 55542 (explaining that the 2015 WOTUS Rule had an original effective date of Aug. 28, 2015). That November 2017 proposed rule sets a new effective date of “two years from the date of final action on [the agencies’] proposal,” to “ensure that there is sufficient time for the regulatory process for reconsidering the definition of ‘waters of the United States’ to be fully completed.” Id., at 55542–55544. 6 It is undisputed that the WOTUS Rule does not fall within the remaining five categories set forth in §1369(b)(1). 7 Because no party argues that the WOTUS Rule is an EPA action approving or promulgating an effluent limitation or other limitation under §1312, §1316, or §1345, the Court confines its analysis to §1311. 8 It is true that the agencies cited §1311 among the provisions under which they purported to have issued the Rule. See 80 Fed. Reg. 37055. They also cited other provisions, including §§1314, 1321, 1341, 1342, and 1344. Ibid. As noted, however, §1311 grants the EPA no authority to clarify the regulatory definition of “waters of the United States.” Furthermore, the agencies’ passing invocation of §1311 does not control our interpretive inquiry. See Adamo Wrecking Co. v. United States, 434 U. S. 275, 283 (1978) (Congress “did not empower the Adminis-trator . . . to make a regulation an ‘emission standard’ by his mere designation”). 9 Although the parties paint dueling portraits of the legislative his-tory, the murky waters of the Congressional Record do not provide helpful guidance in illuminating Congress’ intent in this case. Even for “[t]hose of us who make use of legislative history,” “ambiguous legislative history” cannot trump “clear statutory language.” Milner v. Department of Navy, 562 U. S. 562, 572 (2011) . Just so here.

585.US.2017_16-1140

The California Reproductive Freedom, Accountability, Comprehensive Care, and Transparency Act (FACT Act) was enacted to regulate crisis pregnancy centers—pro-life centers that offer pregnancy-related services. The FACT Act requires clinics that primarily serve pregnant women to provide certain notices. Clinics that are licensed must notify women that California provides free or low-cost services, including abortions, and give them a phone number to call. Its stated purpose is to make sure that state residents know their rights and what health care services are available to them. Unlicensed clinics must notify women that California has not licensed the clinics to provide medical services. Its stated purpose is to ensure that pregnant women know when they are receiving health care from licensed professionals. Petitioners—two crisis pregnancy centers, one licensed and one unlicensed, and an organization of crisis pregnancy centers—filed suit. They alleged that both the licensed and the unlicensed notices abridge the freedom of speech protected by the First Amendment. The District Court denied their motion for a preliminary injunction, and the Ninth Circuit affirmed. Holding that petitioners could not show a likelihood of success on the merits, the court concluded that the licensed notice survived a lower level of scrutiny applicable to regulations of “professional speech,” and that the unlicensed notice satisfied any level of scrutiny. Held: 1. The licensed notice likely violates the First Amendment. Pp. 6–17. (a) Content-based laws “target speech based on its communicative content” and “are presumptively unconstitutional and may be justified only if the government proves that they are narrowly tailored to serve compelling state interests.” Reed v. Town of Gilbert, 576 U. S. ___, ___. The licensed notice is a content-based regulation. By compelling petitioners to speak a particular message, it “alters the content of [their] speech.” Riley v. National Federation of Blind of N. C., Inc., 487 U. S. 781, 795. For example, one of the state-sponsored services that the licensed notice requires petitioners to advertise is abortion—the very practice that petitioners are devoted to opposing. Pp. 6–7. (b) Although the licensed notice is content-based, the Ninth Circuit did not apply strict scrutiny because it concluded that the notice regulates “professional speech.” But this Court has never recognized “professional speech” as a separate category of speech subject to different rules. Speech is not unprotected merely because it is uttered by professionals. The Court has afforded less protection for professional speech in two circumstances—where a law requires professionals to disclose factual, noncontroversial information in their “commercial speech,” see, e.g., Zauderer v. Office of Disciplinary Counsel of Supreme Court of Ohio, 471 U. S. 626, 651, and where States regulate professional conduct that incidentally involves speech, see, e.g., Ohralik v. Ohio State Bar Assn., 436 U. S. 447, 456. Neither line of precedents is implicated here. Pp. 7–14. (1) Unlike the rule in Zauderer, the licensed notice is not limited to “purely factual and uncontroversial information about the terms under which . . . services will be available,” 471 U. S., at 651. California’s notice requires covered clinics to disclose information about state-sponsored services—including abortion, hardly an “uncontroversial” topic. Accordingly, Zauderer has no application here. P. 9. (2) Nor is the licensed notice a regulation of professional conduct that incidentally burdens speech. The Court’s precedents have long drawn a line between speech and conduct. In Planned Parenthood of Southeastern Pa. v. Casey, 505 U. S. 833, for example, the joint opinion rejected a free-speech challenge to an informed-consent law requiring physicians to “give a woman certain information as part of obtaining her consent to an abortion,” id., at 884. But the licensed notice is neither an informed-consent requirement nor any other regulation of professional conduct. It applies to all interactions between a covered facility and its clients, regardless of whether a medical procedure is ever sought, offered, or performed. And many other facilities providing the exact same services, such as general practice clinics, are not subject to the requirement. Pp. 10–11. (3) Outside of these two contexts, the Court’s precedents have long protected the First Amendment rights of professionals. The Court has applied strict scrutiny to content-based laws regulating the noncommercial speech of lawyers, see Reed, supra, at ___, professional fundraisers, see Riley, supra, at 798, and organizations providing specialized advice on international law, see Holder v. Humanitarian Law Project, 561 U. S. 1, 27–28. And it has stressed the danger of content-based regulations “in the fields of medicine and public health, where information can save lives.” Sorrell v. IMS Health Inc., 564 U. S. 552, 566. Such dangers are also present in the context of professional speech, where content-based regulation poses the same “risk that the Government seeks not to advance a legitimate regulatory goal, but to suppress unpopular ideas or information,” Turner Broadcasting Systems, Inc. v. FCC, 512 U. S. 622, 641. When the government polices the content of professional speech, it can fail to “ ‘preserve an uninhibited marketplace of ideas in which truth will ultimately prevail.’ ” McCullen v. Coakley, 573 U. S. ___, ___–___. Professional speech is also a difficult category to define with precision. See Brown v. Entertainment Merchants Assn., 564 U. S. 786, 791. If States could choose the protection that speech receives simply by requiring a license, they would have a powerful tool to impose “invidious discrimination of disfavored subjects.” Cincinnati v. Discovery Network, Inc., 507 U. S. 410, 423, n. 19. Pp. 11–14. (c) Although neither California nor the Ninth Circuit have advanced a persuasive reason to apply different rules to professional speech, the Court need not foreclose the possibility that some such reason exists because the licensed notice cannot survive even intermediate scrutiny. Assuming that California’s interest in providing low-income women with information about state-sponsored service is substantial, the licensed notice is not sufficiently drawn to promote it. The notice is “wildly underinclusive,” Entertainment Merchants Assn., supra, at 802, because it applies only to clinics that have a “primary purpose” of “providing family planning or pregnancy-related services” while excluding several other types of clinics that also serve low-income women and could educate them about the State’s services. California could also inform the women about its services “without burdening a speaker with unwanted speech,” Riley, supra, at 800, most obviously through a public-information campaign. Petitioners are thus likely to succeed on the merits of their challenge. Pp. 14–17. 2. The unlicensed notice unduly burdens protected speech. It is unnecessary to decide whether Zauderer’s standard applies here, for even under Zauderer, a disclosure requirement cannot be “unjustified or unduly burdensome.” 471 U. S., at 651. Disclosures must remedy a harm that is “potentially real not purely hypothetical,” Ibanez v. Florida Dept. of Business and Professional Regulation, Bd. of Accountancy, 512 U. S. 136, 146, and can extend “no broader than reasonably necessary,” In re R. M. J., 455 U. S. 191, 203. California has not demonstrated any justification for the unlicensed notice that is more than “purely hypothetical.” The only justification put forward by the state legislature was ensuring that pregnant women know when they are receiving medical care from licensed professionals, but California denied that the justification for the law was that women did not know what kind of facility they are entering when they go to a crisis pregnancy center. Even if the State had presented a nonhypothetical justification, the FACT Act unduly burdens protected speech. It imposes a government-scripted, speaker-based disclosure requirement that is wholly disconnected from the State’s informational interest. It requires covered facilities to post California’s precise notice, no matter what the facilities say on site or in their advertisements. And it covers a curiously narrow subset of speakers: those that primarily provide pregnancy-related services, but not those that provide, e.g., nonprescription birth control. Such speaker-based laws run the risk that “the State has left unburdened those speakers whose messages are in accord with its own views.” Sorrell, supra, at 580. For these reasons, the unlicensed notice does not satisfy Zauderer, assuming that standard applies. Pp. 17–20. 839 F. 3d 823, reversed and remanded. Thomas, J., delivered the opinion of the Court, in which Roberts, C. J., and Kennedy, Alito, and Gorsuch, JJ., joined. Kennedy, J., filed a concurring opinion, in which Roberts, C. J., and Alito and Gorsuch, JJ., joined. Breyer, J., filed dissenting opinion, in which Ginsburg, Sotomayor, and Kagan, JJ., joined.

The California Reproductive Freedom, Accountability, Comprehensive Care, and Transparency Act (FACT Act) requires clinics that primarily serve pregnant women to provide certain notices. Cal. Health & Safety Code Ann. §123470 et seq. (West 2018). Licensed clinics must notify women that California provides free or low-cost services, including abortions, and give them a phone number to call. Unlicensed clinics must notify women that California has not licensed the clinics to provide medical services. The question in this case is whether these notice requirements violate the First Amendment. I A The California State Legislature enacted the FACT Act to regulate crisis pregnancy centers. Crisis pregnancy centers—according to a report commissioned by the California State Assembly, App. 86—are “pro-life (largely Christian belief-based) organizations that offer a limited range of free pregnancy options, counseling, and other services to individuals that visit a center.” Watters et al., Pregnancy Resource Centers: Ensuring Access and Accuracy of Information 4 (2011). “[U]nfortunately,” the author of the FACT Act stated, “there are nearly 200 licensed and unlicensed” crisis pregnancy centers in California. App. 84. These centers “aim to discourage and prevent women from seeking abortions.” Id., at 85. The author of the FACT Act observed that crisis pregnancy centers “are commonly affiliated with, or run by organizations whose stated goal” is to oppose abortion—including “the National Institute of Family and Life Advocates,” one of the petitioners here. Ibid. To address this perceived problem, the FACT Act imposes two notice requirements on facilities that provide pregnancy-related services—one for licensed facilities and one for unlicensed facilities. 1 The first notice requirement applies to “licensed covered facilit[ies].” Cal. Health & Safety Code Ann. §123471(a). To fall under the definition of “licensed covered facility,” a clinic must be a licensed primary care or specialty clinic or qualify as an intermittent clinic under California law. Ibid. (citing §§1204, 1206(h)). A licensed covered facility also must have the “primary purpose” of “providing fam- ily planning or pregnancy-related services.” §123471(a). And it must satisfy at least two of the following six requirements: “(1) The facility offers obstetric ultrasounds, obstetric sonograms, or prenatal care to pregnant women. “(2) The facility provides, or offers counseling about, contraception or contraceptive methods. “(3) The facility offers pregnancy testing or pregnancy diagnosis. “(4) The facility advertises or solicits patrons with offers to provide prenatal sonography, pregnancy tests, or pregnancy options counseling. “(5) The facility offers abortion services. “(6) The facility has staff or volunteers who collect health information from clients.” Ibid. The FACT Act exempts several categories of clinics that would otherwise qualify as licensed covered facilities. Clinics operated by the United States or a federal agency are excluded, as are clinics that are “enrolled as a Medi-Cal provider” and participate in “the Family Planning, Access, Care, and Treatment Program” (Family PACT program). §123471(c). To participate in the Family PACT program, a clinic must provide “the full scope of family planning . . . services specified for the program,” Cal. Welf. & Inst. Code Ann. §24005(c) (West 2018), including sterilization and emergency contraceptive pills, §§24007(a)(1), (2). If a clinic is a licensed covered facility, the FACT Act requires it to disseminate a government-drafted notice on site. Cal. Health & Safety Code Ann. §123472(a)(1). The notice states that “California has public programs that provide immediate free or low-cost access to comprehensive family planning services (including all FDA-approved methods of contraception), prenatal care, and abortion for eligible women. To determine whether you qualify, contact the county social services office at [insert the telephone number].” Ibid. This notice must be posted in the waiting room, printed and distributed to all clients, or provided digitally at check-in. §123472(a)(2). The notice must be in English and any additional languages identified by state law. §123472(a). In some counties, that means the notice must be spelled out in 13 different languages. See State of Cal., Dept. of Health Care Services, Frequency of Threshold Language Speakers in the Medi-Cal Population by County for Jan. 2015, pp. 4–5 (Sept. 2016) (identifying the required languages for Los Angeles County as English, Spanish, Armenian, Mandarin, Cantonese, Korean, Vietnamese, Farsi, Tagalog, Russian, Cambodian, Other Chinese, and Arabic). The stated purpose of the FACT Act, including its licensed notice requirement, is to “ensure that California residents make their personal reproductive health care decisions knowing their rights and the health care services available to them.” 2015 Cal. Legis. Serv. Ch. 700, §2 (A. B. 775) (West) (Cal. Legis. Serv.). The Legislature posited that “thousands of women remain unaware of the public programs available to provide them with contraception, health education and counseling, family planning, prenatal care, abortion, or delivery.” §1(b). Citing the “time sensitive” nature of pregnancy-related decisions, §1(c), the Legislature concluded that requiring licensed facilities to inform patients themselves would be “[t]he most effective” way to convey this information, §1(d). 2 The second notice requirement in the FACT Act applies to “unlicensed covered facilit[ies].” §123471(b). To fall under the definition of “unlicensed covered facility,” a facility must not be licensed by the State, not have a licensed medical provider on staff or under contract, and have the “primary purpose” of “providing pregnancy-related services.” Ibid. An unlicensed covered facility also must satisfy at least two of the following four requirements: “(1) The facility offers obstetric ultrasounds, obstetric sonograms, or prenatal care to pregnant women. “(2) The facility offers pregnancy testing or pregnancy diagnosis. “(3) The facility advertises or solicits patrons with offers to provide prenatal sonography, pregnancy tests, or pregnancy options counseling. “(4) The facility has staff or volunteers who collect health information from clients.” Ibid. Clinics operated by the United States and licensed primary care clinics enrolled in Medi-Cal and Family PACT are excluded. §123471(c). Unlicensed covered facilities must provide a government-drafted notice stating that “[t]his facility is not li- censed as a medical facility by the State of California and has no licensed medical provider who provides or directly supervises the provision of services.” Cal. Health & Safety Code Ann. §123472(b)(1). This notice must be provided on site and in all advertising materials. §§123472(b)(2), (3). Onsite, the notice must be posted “conspicuously” at the entrance of the facility and in at least one waiting area. §123472(b)(2). It must be “at least 8.5 inches by 11 inches and written in no less than 48-point type.” Ibid. In advertisements, the notice must be in the same size or larger font than the surrounding text, or otherwise set off in a way that draws attention to it. §123472(b)(3). Like the licensed notice, the unlicensed notice must be in English and any additional languages specified by state law. §123471(b). Its stated purpose is to ensure “that pregnant women in California know when they are getting medical care from licensed professionals.” Cal. Legis. Serv., §1(e). B After the Governor of California signed the FACT Act, petitioners—a licensed pregnancy center, an unlicensed pregnancy center, and an organization composed of crisis pregnancy centers—filed this suit. Petitioners alleged that the licensed and unlicensed notices abridge the freedom of speech protected by the First Amendment. The District Court denied their motion for a preliminary injunction. The Court of Appeals for the Ninth Circuit affirmed. National Institute of Family and Life Advocates v. Harris, 839 F. 3d 823, 845 (2016). After concluding that petitioners’ challenge to the FACT Act was ripe,[1] id., at 833, the Ninth Circuit held that petitioners could not show a likelihood of success on the merits. It concluded that the licensed notice survives the “lower level of scrutiny” that applies to regulations of “professional speech.” Id., at 833–842. And it concluded that the unlicensed notice satisfies any level of scrutiny. See id., at 843–844. We granted certiorari to review the Ninth Circuit’s decision. 583 U. S. ___ (2017). We reverse with respect to both notice requirements. II We first address the licensed notice.[2] A The First Amendment, applicable to the States through the Fourteenth Amendment, prohibits laws that abridge the freedom of speech. When enforcing this prohibition, our precedents distinguish between content-based and content-neutral regulations of speech. Content-based regulations “target speech based on its communicative content.” Reed v. Town of Gilbert, 576 U. S. ___, ___ (2015) (slip op., at 6). As a general matter, such laws “are presumptively unconstitutional and may be justified only if the government proves that they are narrowly tailored to serve compelling state interests.” Ibid. This stringent standard reflects the fundamental principle that governments have “ ‘no power to restrict expression because of its message, its ideas, its subject matter, or its content.’ ” Ibid. (quoting Police Dept. of Chicago v. Mosley, 408 U. S. 92, 95 (1972)). The licensed notice is a content-based regulation of speech. By compelling individuals to speak a particular message, such notices “alte[r] the content of [their] speech.” Riley v. National Federation of Blind of N. C., Inc., 487 U. S. 781, 795 (1988); accord, Turner Broadcasting System, Inc. v. FCC, 512 U. S. 622, 642 (1994); Miami Herald Publishing Co. v. Tornillo, 418 U. S. 241, 256 (1974). Here, for example, licensed clinics must provide a government-drafted script about the availability of state-sponsored services, as well as contact information for how to obtain them. One of those services is abortion—the very practice that petitioners are devoted to opposing. By requiring petitioners to inform women how they can obtain state-subsidized abortions—at the same time petitioners try to dissuade women from choosing that option—the licensed notice plainly “alters the content” of petitioners’ speech. Riley, supra, at 795. B Although the licensed notice is content based, the Ninth Circuit did not apply strict scrutiny because it concluded that the notice regulates “professional speech.” 839 F. 3d, at 839. Some Courts of Appeals have recognized “professional speech” as a separate category of speech that is subject to different rules. See, e.g., King v. Governors of New Jersey, 767 F. 3d 216, 232 (CA3 2014); Pickup v. Brown, 740 F. 3d 1208, 1227–1229 (CA9 2014); Moore-King v. County of Chesterfield, 708 F. 3d 560, 568–570 (CA4 2014). These courts define “professionals” as individuals who provide personalized services to clients and who are subject to “a generally applicable licensing and regulatory regime.” Id., at 569; see also, King, supra, at 232; Pickup, supra, at 1230. “Professional speech” is then defined as any speech by these individuals that is based on “[their] expert knowledge and judgment,” King, supra, at 232, or that is “within the confines of [the] professional relationship,” Pickup, supra, at 1228. So defined, these courts except professional speech from the rule that content-based regulations of speech are subject to strict scru- tiny. See King, supra, at 232; Pickup, supra, at 1053–1056; Moore-King, supra, at 569. But this Court has not recognized “professional speech” as a separate category of speech. Speech is not unpro- tected merely because it is uttered by “professionals.” This Court has “been reluctant to mark off new categories of speech for diminished constitutional protection.” Denver Area Ed. Telecommunications Consortium, Inc. v. FCC, 518 U. S. 727, 804 (1996) (Kennedy, J., concurring in part, concurring in judgment in part, and dissenting in part). And it has been especially reluctant to “exemp[t] a category of speech from the normal prohibition on content-based restrictions.” United States v. Alvarez, 567 U. S. 709, 722 (2012) (plurality opinion). This Court’s precedents do not permit governments to impose content-based restrictions on speech without “ ‘persuasive evidence . . . of a long (if heretofore unrecognized) tradition’ ” to that effect. Ibid. (quoting Brown v. Entertainment Merchants Assn., 564 U. S. 786, 792 (2011)). This Court’s precedents do not recognize such a tradition for a category called “professional speech.” This Court has afforded less protection for professional speech in two circumstances—neither of which turned on the fact that professionals were speaking. First, our precedents have applied more deferential review to some laws that require professionals to disclose factual, noncontroversial information in their “commercial speech.” See, e.g., Zauderer v. Office of Disciplinary Counsel of Supreme Court of Ohio, 471 U. S. 626, 651 (1985); Milavetz, Gallop & Milavetz, P. A. v. United States, 559 U. S. 229, 250 (2010); Ohralik v. Ohio State Bar Assn., 436 U. S. 447, 455–456 (1978). Second, under our precedents, States may regulate professional conduct, even though that conduct incidentally involves speech. See, e.g., id., at 456; Planned Parenthood of Southeastern Pa. v. Casey, 505 U. S. 833, 884 (1992) (opinion of O’Connor, Kennedy, and Souter, JJ.). But neither line of precedents is implicated here. 1 This Court’s precedents have applied a lower level of scrutiny to laws that compel disclosures in certain contexts. In Zauderer, for example, this Court upheld a rule requiring lawyers who advertised their services on a contingency-fee basis to disclose that clients might be re- quired to pay some fees and costs. 471 U. S., at 650–653. Noting that the disclosure requirement governed only “commercial advertising” and required the disclosure of “purely factual and uncontroversial information about the terms under which . . . services will be available,” the Court explained that such requirements should be upheld unless they are “unjustified or unduly burdensome.” Id., at 651. The Zauderer standard does not apply here. Most obviously, the licensed notice is not limited to “purely factual and uncontroversial information about the terms under which . . . services will be available.” 471 U. S., at 651; see also Hurley v. Irish-American Gay, Lesbian and Bisexual Group of Boston, Inc., 515 U. S. 557, 573 (1995) (explaining that Zauderer does not apply outside of these circumstances). The notice in no way relates to the services that licensed clinics provide. Instead, it requires these clinics to disclose information about state-sponsored services—including abortion, anything but an “uncontroversial” topic. Accordingly, Zauderer has no application here. 2 In addition to disclosure requirements under Zauderer, this Court has upheld regulations of professional conduct that incidentally burden speech. “[T]he First Amendment does not prevent restrictions directed at commerce or conduct from imposing incidental burdens on speech,” Sorrell v. IMS Health Inc., 564 U. S. 552, 567 (2011), and professionals are no exception to this rule, see Ohralik, supra, at 456. Longstanding torts for professional malpractice, for example, “fall within the traditional purview of state regulation of professional conduct.” NAACP v. Button, 371 U. S. 415, 438 (1963); but cf. id., at 439 (“[A] State may not, under the guise of prohibiting professional misconduct, ignore constitutional rights”). While drawing the line between speech and conduct can be difficult, this Court’s precedents have long drawn it, see, e.g., Sorrell, supra, at 567; Giboney v. Empire Storage & Ice Co., 336 U. S. 490, 502 (1949), and the line is “ ‘long familiar to the bar,’ ” United States v. Stevens, 559 U. S. 460, 468 (2010) (quoting Simon & Schuster, Inc. v. Members of N. Y State Crime Victims Bd., 502 U. S. 105, 127 (1991) (Kennedy, J., concurring in judgment)). In Planned Parenthood of Southeastern Pa. v. Casey, for example, this Court upheld a law requiring physicians to obtain informed consent before they could perform an abortion. 505 U. S., at 884 (joint opinion of O’Connor, Kennedy, and Souter, JJ.). Pennsylvania law required physicians to inform their patients of “the nature of the procedure, the health risks of the abortion and childbirth, and the ‘probable gestational age of the unborn child.’ ” Id., at 881. The law also required physicians to inform patients of the availability of printed materials from the State, which provided information about the child and various forms of assistance. Ibid. The joint opinion in Casey rejected a free-speech challenge to this informed-consent requirement. Id., at 884. It described the Pennsylvania law as “a requirement that a doctor give a woman certain information as part of obtaining her consent to an abortion,” which “for constitutional purposes, [was] no different from a requirement that a doctor give certain specific information about any medical procedure.” Ibid. The joint opinion explained that the law regulated speech only “as part of the practice of medicine, subject to reasonable licensing and regulation by the State.” Ibid. (emphasis added). Indeed, the requirement that a doctor obtain informed consent to perform an operation is “firmly entrenched in American tort law.” Cruzan v. Director, Mo. Dept. of Health, 497 U. S. 261, 269 (1990); see, e.g., Schloendorff v. Society of N. Y. Hospital, 211 N. Y. 125, 129–130, 105 N. E. 92, 93 (1914) (Cardozo, J.) (explaining that “a surgeon who performs an operation without his patient’s consent commits an assault”). The licensed notice at issue here is not an informed-consent requirement or any other regulation of professional conduct. The notice does not facilitate informed consent to a medical procedure. In fact, it is not tied to a procedure at all. It applies to all interactions between a covered facility and its clients, regardless of whether a medical procedure is ever sought, offered, or performed. If a covered facility does provide medical procedures, the notice provides no information about the risks or benefits of those procedures. Tellingly, many facilities that provide the exact same services as covered facilities—such as general practice clinics, see §123471(a)—are not required to provide the licensed notice. The licensed notice regulates speech as speech. 3 Outside of the two contexts discussed above—disclosures under Zauderer and professional conduct—this Court’s precedents have long protected the First Amendment rights of professionals. For example, this Court has applied strict scrutiny to content-based laws that regulate the noncommercial speech of lawyers, see Reed, 576 U. S., at ___ (slip op., at 10) (discussing Button, supra, at 438); In re Primus, 436 U. S. 412, 432 (1978); professional fundraisers, see Riley, 487 U. S., at 798; and organizations that provided specialized advice about international law, see Holder v. Humanitarian Law Project, 561 U. S. 1, 27–28 (2010). And the Court emphasized that the lawyer’s statements in Zauderer would have been “fully protected” if they were made in a context other than advertising. 471 U. S., at 637, n. 7. Moreover, this Court has stressed the danger of content-based regulations “in the fields of medicine and public health, where information can save lives.” Sorrell, supra, at 566. The dangers associated with content-based regulations of speech are also present in the context of professional speech. As with other kinds of speech, regulating the content of professionals’ speech “pose[s] the inherent risk that the Government seeks not to advance a legitimate regulatory goal, but to suppress unpopular ideas or information.” Turner Broadcasting, 512 U. S., at 641. Take medicine, for example. “Doctors help patients make deeply personal decisions, and their candor is crucial.” Wollschlaeger v. Governor of Florida, 848 F. 3d 1293, 1328 (CA11 2017) (en banc) (W. Pryor, J. concurring). Throughout history, governments have “manipulat[ed] the content of doctor-patient discourse” to increase state power and suppress minorities: “For example, during the Cultural Revolution, Chinese physicians were dispatched to the countryside to convince peasants to use contraception. In the 1930s, the Soviet government expedited completion of a construction project on the Siberian railroad by ordering doctors to both reject requests for medical leave from work and conceal this government order from their patients. In Nazi Germany, the Third Reich systematically violated the separation between state ideology and medical discourse. German physicians were taught that they owed a higher duty to the ‘health of the Volk’ than to the health of individual patients. Recently, Nicolae Ceausescu’s strategy to increase the Romanian birth rate included prohibitions against giving advice to patients about the use of birth control devices and disseminating information about the use of condoms as a means of preventing the transmission of AIDS.” Berg, Toward a First Amendment Theory of Doctor-Patient Discourse and the Right To Receive Unbiased Medical Advice, 74 B. U. L. Rev. 201, 201–202 (1994) (footnotes omitted). Further, when the government polices the content of professional speech, it can fail to “ ‘preserve an uninhibited marketplace of ideas in which truth will ultimately prevail.’ ” McCullen v. Coakley, 573 U. S. ___, ___–___ (2014) (slip op., at 8–9). Professionals might have a host of good-faith disagreements, both with each other and with the government, on many topics in their respective fields. Doctors and nurses might disagree about the ethics of assisted suicide or the benefits of medical marijuana; lawyers and marriage counselors might disagree about the prudence of prenuptial agreements or the wisdom of divorce; bankers and accountants might disagree about the amount of money that should be devoted to savings or the benefits of tax reform. “[T]he best test of truth is the power of the thought to get itself accepted in the competition of the market,” Abrams v. United States, 250 U. S. 616, 630 (1919) (Holmes, J., dissenting), and the people lose when the government is the one deciding which ideas should prevail. “Professional speech” is also a difficult category to define with precision. See Entertainment Merchants Assn., 564 U. S., at 791. As defined by the courts of appeals, the professional-speech doctrine would cover a wide array of individuals—doctors, lawyers, nurses, physical therapists, truck drivers, bartenders, barbers, and many others. See Smolla, Professional Speech and the First Amendment, 119 W. Va. L. Rev. 67, 68 (2016). One court of appeals has even applied it to fortune tellers. See Moore-King, 708 F. 3d, at 569. All that is required to make something a “profession,” according to these courts, is that it involves personalized services and requires a professional license from the State. But that gives the States unfettered power to reduce a group’s First Amendment rights by simply imposing a licensing requirement. States cannot choose the protection that speech receives under the First Amendment, as that would give them a powerful tool to impose “invidious discrimination of disfavored subjects.” Cincinnati v. Discovery Network, Inc., 507 U. S. 410, 423–424, n. 19 (1993); see also Riley, 487 U. S., at 796 (“[S]tate labels cannot be dispositive of [the] degree of First Amendment protection” (citing Bigelow v. Virginia, 421 U. S. 809, 826 (1975)). C In sum, neither California nor the Ninth Circuit has identified a persuasive reason for treating professional speech as a unique category that is exempt from ordinary First Amendment principles. We do not foreclose the possibility that some such reason exists. We need not do so because the licensed notice cannot survive even intermediate scrutiny. California asserts a single interest to justify the licensed notice: providing low-income women with information about state-sponsored services. Assuming that this is a substantial state interest, the licensed notice is not sufficiently drawn to achieve it. If California’s goal is to educate low-income women about the services it provides, then the licensed notice is “wildly underinclusive.” Entertainment Merchants Assn., supra, at 802. The notice applies only to clinics that have a “primary purpose” of “providing family planning or pregnancy-related services” and that provide two of six categories of specific services. §123471(a). Other clinics that have another primary purpose, or that provide only one category of those services, also serve low-income women and could educate them about the State’s services. According to the legislative record, California has “nearly 1,000 community clinics”—including “federally designated community health centers, migrant health centers, rural health centers, and frontier health centers”—that “serv[e] more than 5.6 million patients . . . annually through over 17 million patient encounters.” App. 58. But most of those clinics are excluded from the licensed notice requirement without explanation. Such “[u]nderinclusive- ness raises serious doubts about whether the government is in fact pursuing the interest it invokes, rather than disfavoring a particular speaker or viewpoint.” Entertainment Merchants Assn., 564 U. S., at 802. The FACT Act also excludes, without explanation, federal clinics and Family PACT providers from the licensed-notice requirement. California notes that those clinics can enroll women in California’s programs themselves, but California’s stated interest is informing women that these services exist in the first place. California has identified no evidence that the exempted clinics are more likely to provide this information than the covered clinics. In fact, the exempted clinics have long been able to enroll women in California’s programs, but the FACT Act was premised on the notion that “thousands of women remain unaware of [them].” Cal. Legis. Serv., §1(b). If the goal is to maximize women’s awareness of these programs, then it would seem that California would ensure that the places that can immediately enroll women also provide this information. The FACT Act’s exemption for these clinics, which serve many women who are pregnant or could become pregnant in the future, demonstrates the disconnect between its stated purpose and its actual scope. Yet “[p]recision . . . must be the touchstone” when it comes to regulations of speech, which “so closely touc[h] our most precious freedoms.” Button, 371 U. S., at 438. Further, California could inform low-income women about its services “without burdening a speaker with unwanted speech.” Riley, 487 U. S., at 800. Most obvi- ously, it could inform the women itself with a public-information campaign. See ibid. (concluding that a compelled disclosure was unconstitutional because the government could “itself publish . . . the disclosure”). Califor- nia could even post the information on public property near crisis pregnancy centers. California argues that it has already tried an advertising campaign, and that many women who are eligible for publicly-funded healthcare have not enrolled. But California has identified no evidence to that effect. And regardless, a “tepid response” does not prove that an advertising campaign is not a sufficient alternative. United States v. Playboy Entertainment Group, Inc., 529 U. S. 803, 816 (2000). Here, for example, individuals might not have enrolled in California’s services because they do not want them, or because California spent insufficient resources on the advertising campaign. Either way, California cannot co-opt the licensed facilities to deliver its message for it. “[T]he First Amendment does not permit the State to sacrifice speech for efficiency.” Riley, supra, at 795; accord, Arizona Free Enterprise Club’s Freedom Club PAC v. Bennett, 564 U. S. 721, 747 (2011). In short, petitioners are likely to succeed on the merits of their challenge to the licensed notice. Contrary to the suggestion in the dissent, post, at 3–4 (opinion of Breyer, J.), we do not question the legality of health and safety warnings long considered permissible, or purely factual and uncontroversial disclosures about commercial products. III We next address the unlicensed notice. The parties dispute whether the unlicensed notice is subject to deferential review under Zauderer.[3] We need not decide whether the Zauderer standard applies to the unlicensed notice. Even under Zauderer, a disclosure requirement cannot be “unjustified or unduly burdensome.” 471 U. S., at 651. Our precedents require disclosures to remedy a harm that is “potentially real not purely hypothetical,” Ibanez v. Florida Dept. of Business and Professional Regulation, Bd. of Accountancy, 512 U. S. 136, 146 (1994), and to extend “no broader than reasonably necessary,” In re R. M. J., 455 U. S. 191, 203 (1982); accord, Virginia Bd. of Pharmacy v. Virginia Citizens Consumer Council, Inc., 425 U. S. 748, 772, n. 24 (1976); Bates v. State Bar of Ariz., 433 U. S. 350, 384 (1977); cf. Zauderer, 471 U. S., at 649 (rejecting “broad prophylactic rules” in this area). Otherwise, they risk “chilling” protected speech.” Id., at 651. Importantly, California has the burden to prove that the unlicensed notice is neither unjustified nor unduly burdensome. See Ibanez, 512 U. S., at 146. It has not met its burden. We need not decide what type of state interest is sufficient to sustain a disclosure requirement like the unlicensed notice. California has not demonstrated any justification for the unlicensed notice that is more than “purely hypothetical.” Ibid. The only justification that the California Legislature put forward was ensuring that “pregnant women in California know when they are getting medical care from licensed professionals.” 2015 Cal. Legis. Serv., §1(e). At oral argument, however, California denied that the justification for the FACT Act was that women “go into [crisis pregnancy centers] and they don’t realize what they are.” See Tr. of Oral Arg. at 44–45. Indeed, California points to nothing suggesting that pregnant women do not already know that the covered facilities are staffed by unlicensed medical professionals. The services that trigger the unlicensed notice—such as having “volunteers who collect health information from clients,” “advertis[ing] . . . pregnancy options counseling,” and offering over-the-counter “pregnancy testing,” §123471(b)—do not require a medical license. And California already makes it a crime for individuals without a medical license to practice medicine. See Cal. Bus. & Prof. Code Ann. §2052. At this preliminary stage of the litigation, we agree that petitioners are likely to prevail on the question whether California has proved a justification for the unlicensed notice.[4] Even if California had presented a nonhypothetical justification for the unlicensed notice, the FACT Act unduly burdens protected speech. The unlicensed notice imposes a government-scripted, speaker-based disclosure requirement that is wholly disconnected from California’s informational interest. It requires covered facilities to post California’s precise notice, no matter what the facilities say on site or in their advertisements. And it covers a curiously narrow subset of speakers. While the licensed notice applies to facilities that provide “family planning” services and “contraception or contraceptive methods,” §123471(a), the California Legislature dropped these triggering conditions for the unlicensed notice. The unlicensed notice applies only to facilities that primarily provide “pregnancy-related” services. §123471(b). Thus, a facility that advertises and provides pregnancy tests is covered by the unlicensed notice, but a facility across the street that advertises and provides nonprescription contraceptives is excluded—even though the latter is no less likely to make women think it is licensed. This Court’s precedents are deeply skeptical of laws that “distinguis[h] among different speakers, allowing speech by some but not others.” Citizens United v. Federal Election Comm’n, 558 U. S. 310, 340 (2010). Speaker-based laws run the risk that “the State has left unburdened those speakers whose messages are in accord with its own views.” Sorrell, 564 U. S., at 580. The application of the unlicensed notice to advertisements demonstrates just how burdensome it is. The notice applies to all “print and digital advertising materials” by an unlicensed covered facility. §123472(b). These materials must include a government-drafted statement that “[t]his facility is not licensed as a medical facility by the State of California and has no licensed medical provider who provides or directly supervises the provision of services.” §123472(b)(1). An unlicensed facility must call attention to the notice, instead of its own message, by some method such as larger text or contrasting type or color. See §§123472(b)(2)–(3). This scripted language must be posted in English and as many other languages as California chooses to require. As California conceded at oral argument, a billboard for an unlicensed facility that says “Choose Life” would have to surround that two-word statement with a 29-word statement from the government, in as many as 13 different languages. In this way, the unlicensed notice drowns out the facility’s own message. More likely, the “detail required” by the unlicensed notice “effectively rules out” the possibility of having such a billboard in the first place. Ibanez, supra, at 146. For all these reasons, the unlicensed notice does not satisfy Zauderer, assuming that standard applies. California has offered no justification that the notice plausibly furthers. It targets speakers, not speech, and imposes an unduly burdensome disclosure requirement that will chill their protected speech. Taking all these circumstances together, we conclude that the unlicensed notice is unjustified and unduly burdensome under Zauderer. We express no view on the legality of a similar disclosure requirement that is better supported or less burdensome. IV We hold that petitioners are likely to succeed on the merits of their claim that the FACT Act violates the First Amendment. We reverse the judgment of the Court of Appeals and remand the case for further proceedings consistent with this opinion. It is so ordered. Notes 1 We agree with the Ninth Circuit’s ripeness determination. 2 Petitioners raise serious concerns that both the licensed and unlicensed notices discriminate based on viewpoint. Because the notices are unconstitutional either way, as explained below, we need not reach that issue. 3 Other than a conclusory assertion that the unlicensed notice satisfies any standard of review, see Brief for Respondents 19, California does not explain how the unlicensed notice could satisfy any standard other than Zauderer. 4 Nothing in our opinion should be read to foreclose the possibility that California will gather enough evidence in later stages of this litigation.

585.US.2017_16-1454

Respondent credit-card companies American Express Company and American Express Travel Related Services Company (collectively, Amex) operate what economists call a “two-sided platform,” providing services to two different groups (cardholders and merchants) who depend on the platform to intermediate between them. Because the interaction between the two groups is a transaction, credit-card networks are a special type of two-sided platform known as a “transaction” platform. The key feature of transaction platforms is that they cannot make a sale to one side of the platform without simultaneously making a sale to the other. Unlike traditional markets, two-sided platforms exhibit “indirect network effects,” which exist where the value of the platform to one group depends on how many members of another group participate. Two-sided platforms must take these effects into account before making a change in price on either side, or they risk creating a feedback loop of declining demand. Thus, striking the optimal balance of the prices charged on each side of the platform is essential for two-sided platforms to maximize the value of their services and to compete with their rivals. Visa and MasterCard—two of the major players in the credit-card market—have significant structural advantages over Amex. Amex competes with them by using a different business model, which focuses on cardholder spending rather than cardholder lending. To encourage cardholder spending, Amex provides better rewards than the other credit-card companies. Amex must continually invest in its cardholder rewards program to maintain its cardholders’ loyalty. But to fund those investments, it must charge merchants higher fees than its rivals. Although this business model has stimulated competitive innovations in the credit-card market, it sometimes causes friction with merchants. To avoid higher fees, merchants sometimes attempt to dissuade cardholders from using Amex cards at the point of sale—a practice known as “steering.” Amex places antisteering provisions in its contracts with merchants to combat this. In this case, the United States and several States (collectively, plaintiffs) sued Amex, claiming that its antisteering provisions violate §1 of the Sherman Antitrust Act. The District Court agreed, finding that the credit-card market should be treated as two separate markets—one for merchants and one for cardholders—and that Amex’s antisteering provisions are anticompetitive because they result in higher merchant fees. The Second Circuit reversed. It determined that the credit-card market is one market, not two. And it concluded that Amex’s antisteering provisions did not violate §1. Held: Amex’s antisteering provisions do not violate federal antitrust law. Pp. 8–20. (a) Section 1 of the Sherman Act prohibits “unreasonable restraints” of trade. State Oil Co. v. Khan, 522 U. S. 3, 10. Restraints may be unreasonable in one of two ways—unreasonable per se or unreasonable as judged under the “rule of reason.” Business Electronics Corp. v. Sharp Electronics Corp., 485 U. S. 717, 723. The parties agree that Amex’s antisteering provisions should be judged under the rule of reason using a three-step burden-shifting framework. They ask this Court to decide whether the plaintiffs have satisfied the first step in that framework—i.e., whether they have proved that Amex’s antisteering provisions have a substantial anticompetitive effect that harms consumers in the relevant market. Pp. 8–10. (b) Applying the rule of reason generally requires an accurate definition of the relevant market. In this case, both sides of the two-sided credit-card market—cardholders and merchants—must be considered. Only a company with both cardholders and merchants willing to use its network could sell transactions and compete in the credit-card market. And because credit-card networks cannot make a sale unless both sides of the platform simultaneously agree to use their services, they exhibit more pronounced indirect network effects and interconnected pricing and demand. Indeed, credit-card networks are best understood as supplying only one product—the transaction—that is jointly consumed by a cardholder and a merchant. Accordingly, the two-sided market for credit-card transactions should be analyzed as a whole. Pp. 10–15. (c) The plaintiffs have not carried their burden to show anticompetitive effects. Their argument—that Amex’s antisteering provisions increase merchant fees—wrongly focuses on just one side of the market. Evidence of a price increase on one side of a two-sided transaction platform cannot, by itself, demonstrate an anticompetitive exercise of market power. Instead, plaintiffs must prove that Amex’s antisteering provisions increased the cost of credit-card transactions above a competitive level, reduced the number of credit-card transactions, or otherwise stifled competition in the two-sided credit-card market. They failed to do so. Pp. 15–20. (1) The plaintiffs offered no evidence that the price of credit-card transactions was higher than the price one would expect to find in a competitive market. Amex’s increased merchant fees reflect increases in the value of its services and the cost of its transactions, not an ability to charge above a competitive price. It uses higher merchant fees to offer its cardholders a more robust rewards program, which is necessary to maintain cardholder loyalty and encourage the level of spending that makes it valuable to merchants. In addition, the evidence that does exist cuts against the plaintiffs’ view that Amex’s antisteering provisions are the cause of any increases in merchant fees: Visa and MasterCard’s merchant fees have continued to increase, even at merchant locations where Amex is not accepted. Pp. 16–17. (2) The plaintiffs’ evidence that Amex’s merchant-fee increases between 2005 and 2010 were not entirely spent on cardholder rewards does not prove that Amex’s antisteering provisions gave it the power to charge anticompetitive prices. This Court will “not infer competitive injury from price and output data absent some evidence that tends to prove that output was restricted or prices were above a competitive level.” Brooke Group Ltd. v. Brown & Williamson Tobacco Corp., 509 U. S. 209, 237. There is no such evidence here. Output of credit-card transactions increased during the relevant period, and the plaintiffs did not show that Amex charged more than its competitors. P. 17. (3) The plaintiffs also failed to prove that Amex’s antisteering provisions have stifled competition among credit-card companies. To the contrary, while they have been in place, the market experienced expanding output and improved quality. Nor have Amex’s antisteering provisions ended competition between credit-card networks with respect to merchant fees. Amex’s competitors have exploited its higher merchant fees to their advantage. Lastly, there is nothing inherently anticompetitive about the provisions. They actually stem negative externalities in the credit-card market and promote interbrand competition. And they do not prevent competing credit-card networks from offering lower merchant fees or promoting their broader merchant acceptance. Pp. 18–20. 838 F. 3d 179, affirmed. Thomas, J., delivered the opinion of the Court, in which Roberts, C. J., and Kennedy, Alito, and Gorsuch, JJ., joined. Breyer, J., filed a dissenting opinion, in which Ginsburg, Sotomayor, and Kagan, JJ., joined.

American Express Company and American Express Travel Related Services Company (collectively, Amex) provide credit-card services to both merchants and cardholders. When a cardholder buys something from a merchant who accepts Amex credit cards, Amex processes the transaction through its network, promptly pays the merchant, and subtracts a fee. If a merchant wants to accept Amex credit cards—and attract Amex cardholders to its business—Amex requires the merchant to agree to an antisteering contractual provision. The antisteering provision prohibits merchants from discouraging customers from using their Amex card after they have already entered the store and are about to buy something, thereby avoiding Amex’s fee. In this case, we must decide whether Amex’s antisteering provisions violate federal antitrust law. We conclude they do not. I A Credit cards have become a primary way that consumers in the United States purchase goods and services. When a cardholder uses a credit card to buy something from a merchant, the transaction is facilitated by a credit-card network. The network provides separate but inter- related services to both cardholders and merchants. For cardholders, the network extends them credit, which allows them to make purchases without cash and to defer payment until later. Cardholders also can receive rewards based on the amount of money they spend, such as airline miles, points for travel, or cash back. For merchants, the network allows them to avoid the cost of processing transactions and offers them quick, guaranteed payment. This saves merchants the trouble and risk of extending credit to customers, and it increases the number and value of sales that they can make. By providing these services to cardholders and merchants, credit-card companies bring these parties together, and therefore operate what economists call a “two-sided platform.” As the name implies, a two-sided platform offers different products or services to two different groups who both depend on the platform to intermediate between them. See Evans & Schmalensee, Markets With Two-Sided Platforms, 1 Issues in Competition L. & Pol’y 667 (2008) (Evans & Schmalensee); Evans & Noel, Defining Antitrust Markets When Firms Operate Two-Sided Platforms, 2005 Colum. Bus. L. Rev. 667, 668 (Evans & Noel); Filistrucchi, Geradin, Van Damme, & Affeldt, Market Definition in Two-Sided Markets: Theory and Practice, 10 J. Competition L. & Econ. 293, 296 (2014) (Filistrucchi). For credit cards, that interaction is a transaction. Thus, credit-card networks are a special type of two-sided platform known as a “transaction” platform. See id., at 301, 304, 307; Evans & Noel 676–678. The key feature of transaction platforms is that they cannot make a sale to one side of the platform without simultaneously making a sale to the other. See Klein, Lerner, Murphy, & Plache, Competition in Two-Sided Markets: The Antitrust Economics of Payment Card Interchange Fees, 73 Antitrust L. J. 571, 580, 583 (2006) (Klein). For example, no credit-card transaction can occur unless both the merchant and the cardholder simultaneously agree to use the same credit-card network. See Filistrucchi 301. Two-sided platforms differ from traditional markets in important ways. Most relevant here, two-sided platforms often exhibit what economists call “indirect network effects.” Evans & Schmalensee 667. Indirect network effects exist where the value of the two-sided platform to one group of participants depends on how many members of a different group participate. D. Evans & R. Schmalensee, Matchmakers: The New Economics of Multisided Platforms 25 (2016). In other words, the value of the services that a two-sided platform provides increases as the number of participants on both sides of the platform increases. A credit card, for example, is more valuable to cardholders when more merchants accept it, and is more valuable to merchants when more cardholders use it. See Evans & Noel 686–687; Klein 580, 584. To ensure sufficient participation, two-sided platforms must be sensitive to the prices that they charge each side. See Evans & Schmalensee 675; Evans & Noel 680; Muris, Payment Card Regulation and the (Mis)Application of the Economics of Two-Sided Markets, 2005 Colum. Bus. L. Rev. 515, 532–533 (Muris); Rochet & Tirole, Platform Competition in Two-Sided Markets, 1 J. Eur. Econ. Assn. 990, 1013 (2003). Raising the price on side A risks losing participation on that side, which decreases the value of the platform to side B. If participants on side B leave due to this loss in value, then the platform has even less value to side A—risking a feedback loop of declining demand. See Evans & Schmalensee 675; Evans & Noel 680–681. Two-sided platforms therefore must take these indirect network effects into account before making a change in price on either side. See Evans & Schmalensee 675; Evans & Noel 680–681.[1] Sometimes indirect network effects require two-sided platforms to charge one side much more than the other. See Evans & Schmalensee 667, 675, 681, 690–691; Evans & Noel 668, 691; Klein 585; Filistrucchi 300. For two-sided platforms, “ ‘the [relative] price structure matters, and platforms must design it so as to bring both sides on board.’ ” Evans & Schmalensee 669 (quoting Rochet & Tirole, Two-Sided Markets: A Progress Report, 37 RAND J. Econ. 645, 646 (2006)). The optimal price might require charging the side with more elastic demand a below-cost (or even negative) price. See Muris 519, 550; Klein 579; Evans & Schmalensee 675; Evans & Noel 681. With credit cards, for example, networks often charge cardholders a lower fee than merchants because cardholders are more price sensitive.[2] See Muris 522; Klein 573–574, 585, 595. In fact, the network might well lose money on the cardholder side by offering rewards such as cash back, airline miles, or gift cards. See Klein 587; Evans & Schmalensee 672. The network can do this because increasing the number of cardholders increases the value of accepting the card to merchants and, thus, increases the number of merchants who accept it. Muris 522; Evans & Schmalensee 692. Networks can then charge those merchants a fee for every transaction (typically a percentage of the purchase price). Striking the optimal balance of the prices charged on each side of the platform is essential for two-sided platforms to maximize the value of their services and to compete with their rivals. B Amex, Visa, MasterCard, and Discover are the four dominant participants in the credit-card market. Visa, which is by far the largest, has 45% of the market as measured by transaction volume.[3] Amex and MasterCard trail with 26.4% and 23.3%, respectively, while Discover has just 5.3% of the market. Visa and MasterCard have significant structural advantages over Amex. Visa and MasterCard began as bank cooperatives and thus almost every bank that offers credit cards is in the Visa or MasterCard network. This makes it very likely that the average consumer carries, and the average merchant accepts, Visa or MasterCard. As a result, the vast majority of Amex cardholders have a Visa or MasterCard, but only a small number of Visa and Master-Card cardholders have an Amex. Indeed, Visa and MasterCard account for more than 432 million cards in circulation in the United States, while Amex has only 53 million. And while 3.4 million merchants at 6.4 million locations accept Amex, nearly three million more locations accept Visa, MasterCard, and Discover.[4] Amex competes with Visa and MasterCard by using a different business model. While Visa and MasterCard earn half of their revenue by collecting interest from their cardholders, Amex does not. Amex instead earns most of its revenue from merchant fees. Amex’s business model thus focuses on cardholder spending rather than card- holder lending. To encourage cardholder spending, Amex provides better rewards than other networks. Due to its superior rewards, Amex tends to attract cardholders who are wealthier and spend more money. Merchants place a higher value on these cardholders, and Amex uses this advantage to recruit merchants. Amex’s business model has significantly influenced the credit-card market. To compete for the valuable cardholders that Amex attracts, both Visa and MasterCard have introduced premium cards that, like Amex, charge merchants higher fees and offer cardholders better rewards. To maintain their lower merchant fees, Visa and MasterCard have created a sliding scale for their various cards—charging merchants less for low-reward cards and more for high-reward cards. This differs from Amex’s strategy, which is to charge merchants the same fee no matter the rewards that its card offers. Another way that Amex has influenced the credit-card market is by making banking and card-payment services available to low-income individuals, who otherwise could not qualify for a credit card and could not afford the fees that traditional banks charge. See 2 Record 3835–3837, 4527–4529. In sum, Amex’s business model has stimulated competitive innovations in the credit-card market, increasing the volume of transactions and improving the quality of the services. Despite these improvements, Amex’s business model sometimes causes friction with merchants. To maintain the loyalty of its cardholders, Amex must continually invest in its rewards program. But, to fund those investments, Amex must charge merchants higher fees than its rivals. Even though Amex’s investments benefit merchants by encouraging cardholders to spend more money, merchants would prefer not to pay the higher fees. One way that merchants try to avoid them, while still enticing Amex’s cardholders to shop at their stores, is by dissuading cardholders from using Amex at the point of sale. This practice is known as “steering.” Amex has prohibited steering since the 1950s by placing antisteering provisions in its contracts with merchants. These antisteering provisions prohibit merchants from implying a preference for non-Amex cards; dissuading customers from using Amex cards; persuading customers to use other cards; imposing any special restrictions, conditions, disadvantages, or fees on Amex cards; or promoting other cards more than Amex. The antisteering provisions do not, however, prevent merchants from steering customers toward debit cards, checks, or cash. C In October 2010, the United States and several States (collectively, plaintiffs) sued Amex, claiming that its antisteering provisions violate §1 of the Sherman Act, 26Stat. 209, as amended, 15 U. S. C. §1.[5] After a 7-week trial, the District Court agreed that Amex’s antisteering provisions violate §1. United States v. American Express Co., 88 F. Supp. 3d 143, 151–152 (EDNY 2015). It found that the credit-card market should be treated as two separate markets—one for merchants and one for cardholders. See id., at 171–175. Evaluating the effects on the merchant side of the market, the District Court found that Amex’s antisteering provisions are anticompetitive because they result in higher merchant fees. See id., at 195–224. The Court of Appeals for the Second Circuit reversed. United States v. American Express Co., 838 F. 3d 179, 184 (2016). It concluded that the credit-card market is one market, not two. Id., at 196–200. Evaluating the credit-card market as a whole, the Second Circuit concluded that Amex’s antisteering provisions were not anticompetitive and did not violate §1. See id., at 200–206. We granted certiorari, 583 U. S. ___ (2017), and now affirm. II Section 1 of the Sherman Act prohibits “[e]very contract, combination in the form of trust or otherwise, or conspir- acy, in restraint of trade or commerce among the several States.” 15 U. S. C. §1. This Court has long recognized that, “[i]n view of the common law and the law in this country” when the Sherman Act was passed, the phrase “restraint of trade” is best read to mean “undue restraint.” Standard Oil Co. of N. J. v. United States, 221 U. S. 1, 59–60 (1911). This Court’s precedents have thus understood §1 “to outlaw only unreasonable restraints.” State Oil Co. v. Khan, 522 U. S. 3, 10 (1997) (emphasis added). Restraints can be unreasonable in one of two ways. A small group of restraints are unreasonable per se because they “ ‘ “always or almost always tend to restrict competition and decrease output.” ’ ” Business Electronics Corp. v. Sharp Electronics Corp., 485 U. S. 717, 723 (1988). Typically only “horizontal” restraints—restraints “imposed by agreement between competitors”—qualify as unreasonable per se. Id., at 730. Restraints that are not unreasonable per se are judged under the “rule of reason.” Id., at 723. The rule of reason requires courts to conduct a fact-specific assessment of “market power and market structure . . . to assess the [restraint]’s actual effect” on competition. Copperweld Corp. v. Independence Tube Corp., 467 U. S. 752, 768 (1984). The goal is to “distinguis[h] between restraints with anticompetitive effect that are harmful to the consumer and restraints stimulating competition that are in the consumer’s best interest.” Leegin Creative Leather Products, Inc. v. PSKS, Inc., 551 U. S. 877, 886 (2007). In this case, both sides correctly acknowledge that Amex’s antisteering provisions are vertical restraints—i.e., restraints “imposed by agreement between firms at different levels of distribution.” Business Electronics, supra, at 730. The parties also correctly acknowledge that, like nearly every other vertical restraint, the anti- steering provisions should be assessed under the rule of reason. See Leegin, supra, at 882; State Oil, supra, at 19; Business Electronics, supra, at 726; Continental T. V., Inc. v. GTE Sylvania Inc., 433 U. S. 36, 57 (1977). To determine whether a restraint violates the rule of reason, the parties agree that a three-step, burden-shifting framework applies. Under this framework, the plaintiff has the initial burden to prove that the challenged restraint has a substantial anticompetitive effect that harms consumers in the relevant market. See 1 J. Kalinowski, Antitrust Laws and Trade Regulation §12.02[1] (2d ed. 2017) (Kalinowski); P. Areeda & H. Hovenkamp, Fundamentals of Antitrust Law §15.02[B] (4th ed. 2017) (Areeda & Hovenkamp); Capital Imaging Assoc., P. C. v. Mohawk Valley Medical Associates, Inc., 996 F. 2d 537, 543 (CA2 1993). If the plaintiff carries its burden, then the burden shifts to the defendant to show a procompetitive rationale for the restraint. See 1 Kalinowski §12.02[1]; Areeda & Hovenkamp §15.02[B]; Capital Imaging Assoc., supra, at 543. If the defendant makes this showing, then the burden shifts back to the plaintiff to demonstrate that the procompetitive efficiencies could be reasonably achieved through less anticompetitive means. See 1 Kalinowski §12.02[1]; Capital Imaging Assoc., supra, at 543. Here, the parties ask us to decide whether the plaintiffs have carried their initial burden of proving that Amex’s antisteering provisions have an anticompetitive effect. The plaintiffs can make this showing directly or indirectly. Direct evidence of anticompetitive effects would be “ ‘proof of actual detrimental effects [on competition],’ ” FTC v. Indiana Federation of Dentists, 476 U. S. 447, 460 (1986), such as reduced output, increased prices, or decreased quality in the relevant market, see 1 Kalinowski §12.02[2]; Craftsman Limousine, Inc. v. Ford Motor Co., 491 F. 3d 381, 390 (CA8 2007); Virginia Atlantic Airways Ltd. v. British Airways PLC, 257 F. 3d 256, 264 (CA2 2001). Indirect evidence would be proof of market power plus some evidence that the challenged restraint harms competition. See 1 Kalinowski §12.02[2]; Tops Markets, Inc. v. Quality Markets, Inc., 142 F. 3d 90, 97 (CA2 1998); Spanish Broadcasting System of Fla. v. Clear Channel Communications, Inc., 376 F. 3d 1065, 1073 (CA11 2004). Here, the plaintiffs rely exclusively on direct evidence to prove that Amex’s antisteering provisions have caused anticompetitive effects in the credit-card market.[6] To assess this evidence, we must first define the relevant market. Once defined, it becomes clear that the plaintiffs’ evidence is insufficient to carry their burden. A Because “[l]egal presumptions that rest on formalistic distinctions rather than actual market realities are generally disfavored in antitrust law,” Eastman Kodak Co. v. Image Technical Services, Inc., 504 U. S. 451, 466–467 (1992), courts usually cannot properly apply the rule of reason without an accurate definition of the relevant market.[7] “Without a definition of [the] market there is no way to measure [the defendant’s] ability to lessen or destroy competition.” Walker Process Equipment, Inc. v. Food Machinery & Chemical Corp., 382 U. S. 172, 177 (1965); accord, 2 Kalinowski §24.01[4][a]. Thus, the relevant market is defined as “the area of effective competition.” Ibid. Typically this is the “arena within which significant substitution in consumption or production occurs.” Areeda & Hovenkamp §5.02; accord, 2 Kalinowski §24.02[1]; United States v. Grinnell Corp., 384 U. S. 563, 571 (1966). But courts should “combin[e]” different products or services into “a single market” when “that combination reflects commercial realities.” Id., at 572; see also Brown Shoe Co. v. United States, 370 U. S. 294, 336–337 (1962) (pointing out that “the definition of the relevant market” must “ ‘correspond to the commercial realities’ of the industry”). As explained, credit-card networks are two-sided platforms. Due to indirect network effects, two-sided platforms cannot raise prices on one side without risking a feedback loop of declining demand. See Evans & Schmalensee 674–675; Evans & Noel 680–681. And the fact that two-sided platforms charge one side a price that is below or above cost reflects differences in the two sides’ demand elasticity, not market power or anticompetitive pricing. See Klein 574, 595, 598, 626. Price increases on one side of the platform likewise do not suggest anticompetitive effects without some evidence that they have increased the overall cost of the platform’s services. See id., at 575, 594, 626. Thus, courts must include both sides of the platform—merchants and cardholders—when defining the credit-card market. To be sure, it is not always necessary to consider both sides of a two-sided platform. A market should be treated as one sided when the impacts of indirect network effects and relative pricing in that market are minor. See Filistrucchi 321–322. Newspapers that sell advertisements, for example, arguably operate a two-sided platform because the value of an advertisement increases as more people read the newspaper. Id., at 297, 315; Klein 579. But in the newspaper-advertisement market, the indirect networks effects operate in only one direction; newspaper readers are largely indifferent to the amount of advertising that a newspaper contains. See Filistrucchi 321, 323, and n. 99; Klein 583. Because of these weak indirect network effects, the market for newspaper advertising behaves much like a one-sided market and should be analyzed as such. See Filistrucchi 321; Times-Picayune Publishing Co. v. United States, 345 U. S. 594, 610 (1953). But two-sided transaction platforms, like the credit-card market, are different. These platforms facilitate a single, simultaneous transaction between participants. For credit cards, the network can sell its services only if a mer- chant and cardholder both simultaneously choose to use the network. Thus, whenever a credit-card network sells one transaction’s worth of card-acceptance services to a merchant it also must sell one transaction’s worth of card-payment services to a cardholder. It cannot sell transaction services to either cardholders or merchants individually. See Klein 583 (“Because cardholders and merchants jointly consume a single product, payment card transactions, their consumption of payment card transactions must be directly proportional”). To optimize sales, the network must find the balance of pricing that encourages the greatest number of matches between cardholders and merchants. Because they cannot make a sale unless both sides of the platform simultaneously agree to use their services, two-sided transaction platforms exhibit more pronounced indirect network effects and interconnected pricing and demand. Transaction platforms are thus better understood as “suppl[ying] only one product”—transactions. Klein 580. In the credit-card market, these transactions “are jointly consumed by a cardholder, who uses the payment card to make a transaction, and a merchant, who accepts the payment card as a method of payment.” Ibid. Tellingly, credit cards determine their market share by measuring the volume of transactions they have sold.[8] Evaluating both sides of a two-sided transaction platform is also necessary to accurately assess competition. Only other two-sided platforms can compete with a two-sided platform for transactions. See Filistrucchi 301. A credit-card company that processed transactions for merchants, but that had no cardholders willing to use its card, could not compete with Amex. See ibid. Only a company that had both cardholders and merchants willing to use its network could sell transactions and compete in the credit-card market. Similarly, if a merchant accepts the four major credit cards, but a cardholder only uses Visa or Amex, only those two cards can compete for the particular transaction. Thus, competition cannot be accurately assessed by looking at only one side of the platform in isolation.[9] For all these reasons, “[i]n two-sided transaction markets, only one market should be defined.” Id., at 302; see also Evans & Noel 671 (“[F]ocusing on one dimension of . . . competition tends to distort the competition that actually exists among [two-sided platforms]”). Any other analysis would lead to “ ‘ “mistaken inferences” ’ ” of the kind that could “ ‘ “chill the very conduct the antitrust laws are designed to protect.” ’ ” Brooke Group Ltd. v. Brown & Williamson Tobacco Corp., 509 U. S. 209, 226 (1993); see also Matsushita Elec. Industrial Co. v. Zenith Radio Corp., 475 U. S. 574, 594 (1986) (“ ‘[W]e must be concerned lest a rule or precedent that authorizes a search for a particular type of undesirable pricing behavior end up by discouraging legitimate price competition’ ”); Leegin, 551 U. S., at 895 (noting that courts should avoid “increas[ing] the total cost of the antitrust system by prohibiting procompetitive conduct the antitrust laws should encourage”). Accordingly, we will analyze the two-sided market for credit-card transactions as a whole to determine whether the plaintiffs have shown that Amex’s antisteering provisions have anticompetitive effects. B The plaintiffs have not carried their burden to prove anticompetitive effects in the relevant market. The plaintiffs stake their entire case on proving that Amex’s agreements increase merchant fees. We find this argument unpersuasive. As an initial matter, the plaintiffs’ argument about merchant fees wrongly focuses on only one side of the two-sided credit-card market. As explained, the credit-card market must be defined to include both merchants and cardholders. Focusing on merchant fees alone misses the mark because the product that credit-card companies sell is transactions, not services to merchants, and the competitive effects of a restraint on transactions cannot be judged by looking at merchants alone. Evidence of a price increase on one side of a two-sided transaction platform cannot by itself demonstrate an anticompetitive exercise of market power. To demonstrate anticompetitive effects on the two-sided credit-card market as a whole, the plaintiffs must prove that Amex’s antisteering provisions increased the cost of credit-card transactions above a competitive level, reduced the number of credit-card transactions, or otherwise stifled competition in the credit-card market. See 1 Kalinowski §12.02[2]; Craftsman Limousine, Inc., 491 F. 3d, at 390; Virginia Atlantic Airways Ltd., 257 F. 3d, at 264. They failed to do so. 1 The plaintiffs did not offer any evidence that the price of credit-card transactions was higher than the price one would expect to find in a competitive market. As the District Court found, the plaintiffs failed to offer any reliable measure of Amex’s transaction price or profit margins. 88 F. Supp. 3d, at 198, 215. And the evidence about whether Amex charges more than its competitors was ultimately inconclusive. Id., at 199, 202, 215. Amex’s increased merchant fees reflect increases in the value of its services and the cost of its transactions, not an ability to charge above a competitive price. Amex began raising its merchant fees in 2005 after Visa and MasterCard raised their fees in the early 2000s. Id., at 195, 199–200. As explained, Amex has historically charged higher merchant fees than these competitors because it delivers wealthier cardholders who spend more money. Id., at 200–201. Amex’s higher merchant fees are based on a careful study of how much additional value its cardholders offer merchants. See id., at 192–193. On the other side of the market, Amex uses its higher merchant fees to offer its cardholders a more robust rewards program, which is necessary to maintain cardholder loyalty and encourage the level of spending that makes Amex valuable to merchants. Id., at 160, 191–195. That Amex allocates prices between merchants and cardholders differently from Visa and MasterCard is simply not evidence that it wields market power to achieve anticompetitive ends. See Evans & Noel 670–671; Klein 574–575, 594–595, 598, 626. In addition, the evidence that does exist cuts against the plaintiffs’ view that Amex’s antisteering provisions are the cause of any increases in merchant fees. Visa and Master- Card’s merchant fees have continued to increase, even at merchant locations where Amex is not accepted and, thus, Amex’s antisteering provisions do not apply. See 88 F. Supp. 3d, at 222. This suggests that the cause of increased merchant fees is not Amex’s antisteering provisions, but rather increased competition for cardholders and a corresponding marketwide adjustment in the relative price charged to merchants. See Klein 575, 609. 2 The plaintiffs did offer evidence that Amex increased the percentage of the purchase price that it charges merchants by an average of 0.09% between 2005 and 2010 and that this increase was not entirely spent on cardholder rewards. See 88 F. Supp. 3d, at 195–197, 215. The plaintiffs believe that this evidence shows that the price of Amex’s transactions increased. Even assuming the plaintiffs are correct, this evidence does not prove that Amex’s antisteering provisions gave it the power to charge anticompetitive prices. “Market power is the ability to raise price profitably by restricting output.” Areeda & Hovenkamp §5.01 (emphasis added); accord, Kodak, 504 U. S., at 464; Business Electronics, 485 U. S., at 723. This Court will “not infer competitive injury from price and output data absent some evidence that tends to prove that output was restricted or prices were above a competitive level.” Brooke Group Ltd., 509 U. S., at 237. There is no such evidence in this case. The output of credit-card transactions grew dramatically from 2008 to 2013, increasing 30%. See 838 F. 3d, at 206. “Where . . . output is expanding at the same time prices are increasing, rising prices are equally consistent with growing product demand.” Brooke Group Ltd., supra, at 237. And, as previously explained, the plaintiffs did not show that Amex charged more than its competitors. 3 The plaintiffs also failed to prove that Amex’s antisteering provisions have stifled competition among credit-card companies. To the contrary, while these agreements have been in place, the credit-card market experienced expanding output and improved quality. Amex’s business model spurred Visa and MasterCard to offer new premium card categories with higher rewards. And it has increased the availability of card services, including free banking and card-payment services for low-income customers who otherwise would not be served. Indeed, between 1970 and 2001, the percentage of households with credit cards more than quadrupled, and the proportion of households in the bottom-income quintile with credit cards grew from just 2% to over 38%. See D. Evans & R. Schmalensee, Paying With Plastic: The Digital Revolution in Buying and Borrowing 88–89 (2d ed. 2005) (Paying With Plastic). Nor have Amex’s antisteering provisions ended competition between credit-card networks with respect to merchant fees. Instead, fierce competition between networks has constrained Amex’s ability to raise these fees and has, at times, forced Amex to lower them. For instance, when Amex raised its merchant prices between 2005 and 2010, some merchants chose to leave its network. 88 F. Supp. 3d, at 197. And when its remaining merchants complained, Amex stopped raising its merchant prices. Id., at 198. In another instance in the late 1980s and early 1990s, competition forced Amex to offer lower merchant fees to “everyday spend” merchants—supermarkets, gas stations, pharmacies, and the like—to persuade them to accept Amex. See id., at 160–161, 202. In addition, Amex’s competitors have exploited its higher merchant fees to their advantage. By charging lower merchant fees, Visa, MasterCard, and Discover have achieved broader merchant acceptance—approximately 3 million more locations than Amex. Id., at 204. This broader merchant acceptance is a major advantage for these networks and a significant challenge for Amex, since consumers prefer cards that will be accepted everywhere. Ibid. And to compete even further with Amex, Visa and MasterCard charge different merchant fees for different types of cards to maintain their comparatively lower merchant fees and broader acceptance. Over the long run, this competition has created a trend of declining merchant fees in the credit-card market. In fact, since the first credit card was introduced in the 1950s, merchant fees—including Amex’s merchant fees—have decreased by more than half. See id., at 202–203; Paying With Plastic 54, 126, 152. Lastly, there is nothing inherently anticompetitive about Amex’s antisteering provisions. These agreements actually stem negative externalities in the credit-card market and promote interbrand competition. When merchants steer cardholders away from Amex at the point of sale, it undermines the cardholder’s expectation of “welcome acceptance”—the promise of a frictionless transaction. 88 F. Supp. 3d, at 156. A lack of welcome acceptance at one merchant makes a cardholder less likely to use Amex at all other merchants. This externality endangers the viability of the entire Amex network. And it undermines the investments that Amex has made to encourage increased cardholder spending, which discourages investments in rewards and ultimately harms both cardholders and merchants. Cf. Leegin, 551 U. S., at 890–891 (recognizing that vertical restraints can prevent retailers from free riding and thus increase the availability of “tangible or intangible services or promotional efforts” that enhance competition and consumer welfare). Perhaps most importantly, antisteering provisions do not prevent Visa, MasterCard, or Discover from competing against Amex by offering lower merchant fees or promoting their broader merchant acceptance.[10] In sum, the plaintiffs have not satisfied the first step of the rule of reason. They have not carried their burden of proving that Amex’s antisteering provisions have anticompetitive effects. Amex’s business model has spurred robust interbrand competition and has increased the quality and quantity of credit-card transactions. And it is “[t]he promotion of interbrand competition,” after all, that “is . . . ‘the primary purpose of the antitrust laws.’ ” Id., at 890. * * * Because Amex’s antisteering provisions do not unreasonably restrain trade, we affirm the judgment of the Court of Appeals. It is so ordered. Notes 1 In a competitive market, indirect network effects also encourage companies to take increased profits from a price increase on side A and spend them on side B to ensure more robust participation on that side and to stem the impact of indirect network effects. See Evans & Schmalensee 688; Evans & Noel 670–671, 695. Indirect network effects thus limit the platform’s ability to raise overall prices and impose a check on its market power. See Evans & Schmalensee 688; Evans & Noel 695. 2 “Cardholders are more price-sensitive because many consumers have multiple payment methods, including alternative payment cards. Most merchants, by contrast, cannot accept just one major card because they are likely to lose profitable incremental sales if they do not take [all] the major payment cards. Because most consumers do not carry all of the major payment cards, refusing to accept a major card may cost the merchant substantial sales.” Muris 522. 3 All figures are accurate as of 2013. 4 Discover entered the credit-card market several years after Amex, Visa, and MasterCard. It nonetheless managed to gain a foothold because Sears marketed Discover to its already significant base of private-label cardholders. Discover’s business model shares certain features with Amex, Visa, and MasterCard. Like Amex, Discover interacts directly with its cardholders. But like Visa and MasterCard, Discover uses banks that cooperate with its network to interact with merchants. 5 Plaintiffs also sued Visa and MasterCard, claiming that their anti-steering provisions violated §1. But Visa and MasterCard voluntarily revoked their antisteering provisions and are no longer parties to this case. 6 Although the plaintiffs relied on indirect evidence below, they have abandoned that argument in this Court. See Brief for United States 23, n. 4 (citing Pet. for Cert. i, 18–25). 7 The plaintiffs argue that we need not define the relevant market in this case because they have offered actual evidence of adverse effects on competition—namely, increased merchant fees. See Brief for United States 40–41 (citing FTC v. Indiana Federation of Dentists, 476 U. S. 447 (1986), and Catalano, Inc. v. Target Sales, Inc., 446 U. S. 643 (1980) (per curiam)). We disagree. The cases that the plaintiffs cite for this proposition evaluated whether horizontal restraints had an adverse effect on competition. See Indiana Federation of Dentists, supra, at 450–451, 459 (agreement between competing dentists not to share X rays with insurance companies); Catalano, supra, at 644–645, 650 (agreement among competing wholesalers not to compete on extending credit to retailers). Given that horizontal restraints involve agreements between competitors not to compete in some way, this Court concluded that it did not need to precisely define the relevant market to conclude that these agreements were anticompetitive. See Indiana Federation of Dentists, supra, at 460–461; Catalano, supra, at 648–649. But vertical restraints are different. See Arizona v. Maricopa County Medical Soc., 457 U. S. 332, 348, n. 18 (1982); Leegin Creative Leather Products, Inc. v. PSKS, Inc., 551 U. S. 877, 888 (2007). Vertical restraints often pose no risk to competition unless the entity imposing them has market power, which cannot be evaluated unless the Court first defines the relevant market. See id., at 898 (noting that a vertical restraint “may not be a serious concern unless the relevant entity has market power”); Easterbrook, Vertical Arrangements and the Rule of Reason, 53 Antitrust L. J. 135, 160 (1984) (“[T]he possibly anticompetitive manifestations of vertical arrangements can occur only if there is market power”). 8 Contrary to the dissent’s assertion, post, at 11–12, merchant services and cardholder services are not complements. See Filistrucchi 297 (“[A] two-sided market [is] different from markets for complementary products, in which both products are bought by the same buyers, who, in their buying decisions, can therefore be expected to take into account both prices”). As already explained, credit-card companies are best understood as supplying only one product—transactions—which is jointly consumed by a cardholder and a merchant. See Klein 580. Merchant services and cardholder services are both inputs to this single product. See ibid. 9 Nontransaction platforms, by contrast, often do compete with companies that do not operate on both sides of their platform. A newspaper that sells advertising, for example, might have to compete with a television network, even though the two do not meaningfully compete for viewers. See Filistrucchi 301. 10 The plaintiffs argue that United States v. Topco Associates, Inc., 405 U. S. 596, 610 (1972), forbids any restraint that would restrict competition in part of the market—here, for example, merchant steering. See Brief for Petitioners and Respondents Nebraska, Tennessee, and Texas 30, 42. Topco does not stand for such a broad proposition. Topco concluded that a horizontal agreement between competitors was unreasonable per se, even though the agreement did not extend to every competitor in the market. See 405 U. S., at 599, 608. A horizontal agreement between competitors is markedly different from a vertical agreement that incidentally affects one particular method of competition. See Leegin, 551 U. S., at 888; Maricopa County Medical Soc., 457 U. S., at 348, n. 18.

584.US.2017_16-712

564 U. S. 462, 484 . Another means of canceling a patent at that time—a petition to the Privy Council to vacate a patent—closely resembles inter partes review. The parties have cited nothing to suggest that the Framers were not aware of this common practice when writing the Patent Clause, or that they excluded the practice from the scope of the Clause. Relatedly, the fact that American courts have traditionally adjudicated patent validity in this country does not mean that they must forever do so. See post, at 8–10. Historical practice is not decisive here because matters governed by the public-rights doctrine may be assigned to the Legislature, the Executive, or the Judiciary. Ex parte Bakelite Corp., supra, at 451. That Congress chose the courts in the past does not foreclose its choice of the PTO today. Pp. 12–15. (d) Finally, the similarities between the various procedures used in inter partes review and procedures typically used in courts does not lead to the conclusion that inter partes review violates Article III. This Court has never adopted a “looks like” test to determine if an adjudication has improperly occurred outside an Article III court. See, e.g., Williams v. United States, 289 U. S. 553, 563 . Pp. 15–16. (e) This holding is narrow. The Court addresses only the constitutionality of inter partes review and the precise constitutional challenges that Oil States raised here. The decision should not be misconstrued as suggesting that patents are not property for purposes of the Due Process Clause or the Takings Clause. Pp. 16–17. 2. Inter partes review does not violate the Seventh Amendment. When Congress properly assigns a matter to adjudication in a non-Article III tribunal, “the Seventh Amendment poses no independent bar to the adjudication of that action by a nonjury factfinder.” Granfinanciera, S. A. v. Nordberg, 492 U. S. 33, 52 –53. Thus, the rejection of Oil States’ Article III challenge also resolves its Seventh Amendment challenge. P. 17. 639 Fed. Appx. 639, affirmed. Thomas, J., delivered the opinion of the Court, in which Kennedy, Ginsburg, Breyer, Alito, Sotomayor, and Kagan, JJ., joined. Breyer, J., filed a concurring opinion, in which Ginsburg and Sotomayor, JJ., joined. Gorsuch, J., filed a dissenting opinion, in which Roberts, C. J., joined.

The Leahy-Smith America Invents Act, 35 U. S. C. §100 et seq., establishes a process called “inter partes review.” Under that process, the United States Patent and Trademark Office (PTO) is authorized to reconsider and to cancel an issued patent claim in limited circumstances. In this case, we address whether inter partes review violates Article III or the Seventh Amendment of the Constitution. We hold that it violates neither. I A Under the Patent Act, the PTO is “responsible for the granting and issuing of patents.” 35 U. S. C. §2(a)(1). When an inventor applies for a patent, an examiner reviews the proposed claims and the prior art to determine if the claims meet the statutory requirements. See §§112, 131. Those requirements include utility, novelty, and nonobviousness based on the prior art. §§101, 102, 103. The Director of the PTO then approves or rejects the application. See §§131, 132(a). An applicant can seek judicial review of a final rejection. §§141(a), 145. B Over the last several decades, Congress has created administrative processes that authorize the PTO to reconsider and cancel patent claims that were wrongly issued. In 1980, Congress established “ex parte reexamination,” which still exists today. See Act To Amend the Patent and Trademark Laws, 35 U. S. C. §301 et seq. Ex parte re- examination permits “[a]ny person at any time” to “file a request for reexamination.” §302. If the Director determines that there is “a substantial new question of patentability” for “any claim of the patent,” the PTO can reexamine the patent. §§303(a), 304. The reexamination process follows the same procedures as the initial examination. §305. In 1999, Congress added a procedure called “inter partes reexamination.” See American Inventors Protection Act, §§4601–4608, 113Stat. 1501A–567 to 1501A–572. Under this procedure, any person could file a request for reexamination. 35 U. S. C. §311(a) (2006 ed.). The Director would determine if the request raised “a substantial new question of patentability affecting any claim of the patent” and, if so, commence a reexamination. §§312(a), 313 (2006 ed.). The reexamination would follow the general procedures for initial examination, but would allow the third-party requester and the patent owner to participate in a limited manner by filing responses and replies. §§314(a), (b) (2006 ed.). Inter partes reexamination was phased out when the America Invents Act went into effect in 2012. See §6, 125Stat. 299–305. C The America Invents Act replaced inter partes reexamination with inter partes review, the procedure at issue here. See id., at 299. Any person other than the patent owner can file a petition for inter partes review. 35 U. S. C. §311(a) (2012 ed.). The petition can request cancellation of “1 or more claims of a patent” on the grounds that the claim fails the novelty or nonobviousness standards for patentability. §311(b). The challenges must be made “only on the basis of prior art consisting of patents or printed publications.” Ibid. If a petition is filed, the patent owner has the right to file a preliminary response explaining why inter partes review should not be insti- tuted. §313. Before he can institute inter partes review, the Director must determine “that there is a reasonable likelihood that the petitioner would prevail with respect to at least 1 of the claims challenged.” §314(a). The decision whether to institute inter partes review is committed to the Director’s discretion. See Cuozzo Speed Technologies, LLC v. Lee, 579 U. S. ___, ___ (2016) (slip op., at 9). The Director’s decision is “final and nonappealable.” §314(d).[1] Once inter partes review is instituted, the Patent Trial and Appeal Board—an adjudicatory body within the PTO created to conduct inter partes review—examines the patent’s validity. See 35 U. S. C. §§6, 316(c). The Board sits in three-member panels of administrative patent judges. See §6(c). During the inter partes review, the petitioner and the patent owner are entitled to certain discovery, §316(a)(5); to file affidavits, declarations, and written memoranda, §316(a)(8); and to receive an oral hearing before the Board, §316(a)(10). The petitioner has the burden of proving unpatentability by a preponderance of the evidence. §316(e). The owner can file a motion to amend the patent by voluntarily canceling a claim or by “propos[ing] a reasonable number of substitute claims.” §316(d)(1)(B). The owner can also settle with the peti- tioner by filing a written agreement prior to the Board’s final decision, which terminates the proceedings with respect to that petitioner. §317. If the settlement results in no petitioner remaining in the inter partes review, the Board can terminate the proceeding or issue a final written decision. §317(a). If the proceeding does not terminate, the Board must issue a final written decision no later than a year after it notices the institution of inter partes review, but that deadline can be extended up to six months for good cause. §§316(a)(11), 318(a). If the Board’s decision becomes final, the Director must “issue and publish a certificate.” §318(b). The certificate cancels patent claims “finally determined to be unpatentable,” confirms patent claims “determined to be patentable,” and incorporates into the patent “any new or amended claim determined to be patentable.” Ibid. A party dissatisfied with the Board’s decision can seek judicial review in the Court of Appeals for the Federal Circuit. §319. Any party to the inter partes review can be a party in the Federal Circuit. Ibid. The Director can intervene to defend the Board’s decision, even if no party does. See §143; Cuozzo, supra, at ___ (slip op., at 15). When reviewing the Board’s decision, the Federal Circuit assesses “the Board’s compliance with governing legal standards de novo and its underlying factual determinations for substantial evidence.” Randall Mfg. v. Rea, 733 F. 3d 1355, 1362 (CA Fed. 2013). II Petitioner Oil States Energy Services, LLC, and respondent Greene’s Energy Group, LLC, are both oilfield services companies. In 2001, Oil States obtained a patent relating to an apparatus and method for protecting wellhead equipment used in hydraulic fracturing. In 2012, Oil States sued Greene’s Energy in Federal District Court for infringing that patent. Greene’s Energy responded by challenging the patent’s validity. Near the close of discovery, Greene’s Energy also petitioned the Board to institute inter partes review. It argued that two of the patent’s claims were unpatentable because they were anticipated by prior art not mentioned by Oil States in its original patent application. Oil States filed a response opposing review. The Board found that Greene’s Energy had established a reasonable likelihood that the two claims were unpatentable and, thus, instituted inter partes review. The proceedings before the District Court and the Board progressed in parallel. In June 2014, the District Court issued a claim-construction order. The order construed the challenged claims in a way that foreclosed Greene’s Energy’s arguments about the prior art. But a few months later, the Board issued a final written decision concluding that the claims were unpatentable. The Board acknowledged the District Court’s contrary decision, but nonetheless concluded that the claims were anticipated by the prior art. Oil States sought review in the Federal Circuit. In addition to its arguments about patentability, Oil States challenged the constitutionality of inter partes review. Specifically, it argued that actions to revoke a patent must be tried in an Article III court before a jury. While Oil States’ case was pending, the Federal Circuit issued an opinion in a different case, rejecting the same constitutional arguments. MCM Portfolio LLC v. Hewlett-Packard Co., 812 F. 3d 1284, 1288–1293 (2015). The Federal Circuit summarily affirmed the Board’s decision in this case. 639 Fed. Appx. 639 (2016). We granted certiorari to determine whether inter partes review violates Article III or the Seventh Amendment. 582 U. S. ___ (2017). We address each issue in turn. III Article III vests the judicial power of the United States “in one supreme Court, and in such inferior Courts as the Congress may from time to time ordain and establish.” §1. Consequently, Congress cannot “confer the Government’s ‘judicial Power’ on entities outside Article III.” Stern v. Marshall, 564 U. S. 462, 484 (2011) . When determining whether a proceeding involves an exercise of Article III judicial power, this Court’s precedents have distinguished between “public rights” and “private rights.” Executive Benefits Ins. Agency v. Arkison, 573 U. S. ___, ___ (2014) (slip op., at 6) (internal quotation marks omitted). Those precedents have given Congress significant latitude to assign adjudication of public rights to entities other than Article III courts. See ibid.; Stern, supra, at 488–492. This Court has not “definitively explained” the distinction between public and private rights, Northern Pipeline Constr. Co. v. Marathon Pipe Line Co., 458 U. S. 50, 69 (1982) , and its precedents applying the public-rights doctrine have “not been entirely consistent,” Stern, 564 U. S., at 488. But this case does not require us to add to the “various formulations” of the public-rights doctrine. Ibid. Our precedents have recognized that the doctrine covers matters “which arise between the Government and persons subject to its authority in connection with the performance of the constitutional functions of the executive or legislative departments.” Crowell v. Benson, 285 U. S. 22 , 50 (1932). In other words, the public-rights doctrine applies to matters “ ‘arising between the government and others, which from their nature do not require judicial determination and yet are susceptible of it.’ ” Ibid. (quoting Ex parte Bakelite Corp., 279 U. S. 438, 451 (1929) ). Inter partes review involves one such matter: reconsid- eration of the Government’s decision to grant a public franchise. A Inter partes review falls squarely within the public-rights doctrine. This Court has recognized, and the parties do not dispute, that the decision to grant a patent is a matter involving public rights—specifically, the grant of a public franchise. Inter partes review is simply a reconsideration of that grant, and Congress has permissibly reserved the PTO’s authority to conduct that reconsideration. Thus, the PTO can do so without violating Article III. 1 This Court has long recognized that the grant of a patent is a “ ‘matte[r] involving public rights.’ ” United States v. Duell, 172 U. S. 576, 582 –583 (1899) (quoting Murray’s Lessee v. Hoboken Land & Improvement Co., 18 How. 272, 284 (1856)). It has the key features to fall within this Court’s longstanding formulation of the public-rights doctrine. Ab initio, the grant of a patent involves a matter “arising between the government and others.” Ex parte Bakelite Corp., supra, at 451. As this Court has long recognized, the grant of a patent is a matter between “ ‘the public, who are the grantors, and . . . the patentee.’ ” Duell, supra, at 586 (quoting Butterworth v. United States ex rel. Hoe, 112 U. S. 50, 59 (1884) ). By “issuing patents,” the PTO “take[s] from the public rights of immense value, and bestow[s] them upon the patentee.” United States v. American Bell Telephone Co., 128 U. S. 315, 370 (1888) . Specifically, patents are “public franchises” that the Government grants “to the inventors of new and useful improvements.” Seymour v. Osborne, 11 Wall. 516, 533 (1871); accord, Pfaff v. Wells Electronics, Inc., 525 U. S. 55, 63 –64 (1998). The franchise gives the patent owner “the right to exclude others from making, using, offering for sale, or selling the invention throughout the United States.” 35 U. S. C. §154(a)(1). That right “did not exist at common law.” Gayler v. Wilder, 10 How. 477, 494 (1851). Rather, it is a “creature of statute law.” Crown Die & Tool Co. v. Nye Tool & Machine Works, 261 U. S. 24, 40 (1923) . Additionally, granting patents is one of “the constitutional functions” that can be carried out by “the executive or legislative departments” without “ ‘judicial determination.’ ” Crowell, supra, at 50–51 (quoting Ex parte Bakelite Corp., supra, at 452). Article I gives Congress the power “[t]o promote the Progress of Science and useful Arts, by securing for limited Times to Authors and Inventors the exclusive Right to their respective Writings and Discoveries.” §8, cl. 8. Congress can grant patents itself by statute. See, e.g., Bloomer v. McQuewan, 14 How. 539, 548–550 (1853). And, from the founding to today, Congress has authorized the Executive Branch to grant patents that meet the statutory requirements for patentability. See 35 U. S. C. §§2(a)(1), 151; see also Act of July 8, 1870, §31, 16Stat. 202; Act of July 4, 1836, §7, 5Stat. 119–120; Act of Apr. 10, 1790, ch. 7, §1, 1Stat. 109–110. When the PTO “adjudicate[s] the patentability of inventions,” it is “exercising the executive power.” Freytag v. Commissioner, 501 U. S. 868, 910 (1991) (Scalia, J., concurring in part and concurring in judgment) (emphasis deleted). Accordingly, the determination to grant a patent is a “matte[r] involving public rights.” Murray’s Lessee, supra, at 284. It need not be adjudicated in Article III court. 2 Inter partes review involves the same basic matter as the grant of a patent. So it, too, falls on the public-rights side of the line. Inter partes review is “a second look at an earlier administrative grant of a patent.” Cuozzo, 579 U. S., at ___ (slip op., at 16). The Board considers the same statutory requirements that the PTO considered when granting the patent. See 35 U. S. C. §311(b). Those statutory requirements prevent the “issuance of patents whose effects are to remove existent knowledge from the public domain.” Graham v. John Deere Co. of Kansas City, 383 U. S. 1, 6 (1966) . So, like the PTO’s initial review, the Board’s inter partes review protects “the public’s paramount interest in seeing that patent monopolies are kept within their legitimate scope,” Cuozzo, supra, at ___ (slip op., at 16) (internal quotation marks and alterations omitted). Thus, inter partes review involves the same interests as the determination to grant a patent in the first instance. See Duell, supra, at 586. The primary distinction between inter partes review and the initial grant of a patent is that inter partes review occurs after the patent has issued. But that distinction does not make a difference here. Patent claims are granted subject to the qualification that the PTO has “the au- thority to reexamine—and perhaps cancel—a patent claim” in an inter partes review. See Cuozzo, supra, at ___ (slip op., at 3). Patents thus remain “subject to [the Board’s] authority” to cancel outside of an Article III court. Crowell, 285 U. S., at 50. This Court has recognized that franchises can be qualified in this manner. For example, Congress can grant a franchise that permits a company to erect a toll bridge, but qualify the grant by reserving its authority to revoke or amend the franchise. See, e.g., Louisville Bridge Co. v. United States, 242 U. S. 409, 421 (1917) (collecting cases). Even after the bridge is built, the Government can exercise its reserved authority through legislation or an administrative proceeding. See, e.g., id., at 420–421; Hannibal Bridge Co. v. United States, 221 U. S. 194, 205 (1911) ; Bridge Co. v. United States, 105 U. S. 470, 478 –482 (1882). The same is true for franchises that permit companies to build railroads or telegraph lines. See, e.g., United States v. Union Pacific R. Co., 160 U. S. 1, 24 –25, 37–38 (1895). Thus, the public-rights doctrine covers the matter resolved in inter partes review. The Constitution does not prohibit the Board from resolving it outside of an Article III court. B Oil States challenges this conclusion, citing three decisions that recognize patent rights as the “private property of the patentee.” American Bell Telephone Co., 128 U. S., at 370; see also McCormick Harvesting Machine Co. v. Aultman, 169 U. S. 606, 609 (1898) (“[A granted patent] has become the property of the patentee”); Brown v. Du- chesne, 19 How. 183, 197 (1857) (“[T]he rights of a party under a patent are his private property”). But those cases do not contradict our conclusion. Patents convey only a specific form of property right—a public franchise. See Pfaff, 525 U. S., at 63–64. And patents are “entitled to protection as any other property, consisting of a franchise.” Seymour, 11 Wall. at 533 (emphasis added). As a public franchise, a patent can confer only the rights that “the statute prescribes.” Gayler, supra, at 494; Wheaton v. Peters, 8 Pet. 591, 663–664 (1834) (noting that Congress has “the power to prescribe the conditions on which such right shall be enjoyed”). It is noteworthy that one of the precedents cited by Oil States acknowledges that the patentee’s rights are “derived altogether” from statutes, “are to be regulated and measured by these laws, and cannot go beyond them.” Brown, supra, at 195.[2] One such regulation is inter partes review. See Cuozzo, 579 U. S., at ___ (slip op., at 3). The Patent Act provides that, “[s]ubject to the provisions of this title, patents shall have the attributes of personal property.” 35 U. S. C. §261. This provision qualifies any property rights that a patent owner has in an issued patent, subjecting them to the express provisions of the Patent Act. See eBay Inc. v. MercExchange, L. L. C., 547 U. S. 388, 392 (2006) . Those provisions include inter partes review. See §§311–319. Nor do the precedents that Oil States cites foreclose the kind of post-issuance administrative review that Congress has authorized here. To be sure, two of the cases make broad declarations that “[t]he only authority competent to set a patent aside, or to annul it, or to correct it for any reason whatever, is vested in the courts of the United States, and not in the department which issued the pat- ent.” McCormick Harvesting Machine Co., supra, at 609; accord, American Bell Telephone Co., 128 U. S., at 364. But those cases were decided under the Patent Act of 1870. See id., at 371; McCormick Harvesting Machine Co., supra, at 611. That version of the Patent Act did not include any provision for post-issuance administrative review. Those precedents, then, are best read as a description of the statutory scheme that existed at that time. They do not resolve Congress’ authority under the Constitution to establish a different scheme.[3] C Oil States and the dissent contend that inter partes review violates the “general” principle that “Congress may not ‘withdraw from judicial cognizance any matter which, from its nature, is the subject of a suit at the common law, or in equity, or admiralty.’ ” Stern, 564 U. S., at 484 (quoting Murray’s Lessee, 18 How., at 284). They argue that this is so because patent validity was often decided in English courts of law in the 18th century. For example, if a patent owner brought an infringement action, the defendant could challenge the validity of the patent as an affirmative defense. See Lemley, Why Do Juries Decide If Patents Are Valid? 99 Va. L. Rev. 1673, 1682, 1685–1686, and n. 52 (2013). Or, an individual could challenge the validity of a patent by filing a writ of scire facias in the Court of Chancery, which would sit as a law court when adjudicating the writ. See id., at 1683–1685, and n. 44; Bottomley, Patent Cases in the Court of Chancery, 1714–58, 35 J. Legal Hist. 27, 36–37, 41–43 (2014). But this history does not establish that patent validity is a matter that, “from its nature,” must be decided by a court. Stern, supra, at 484 (quoting Murray’s Lessee, supra, at 284). The aforementioned proceedings were between private parties. But there was another means of canceling a patent in 18th-century England, which more closely resembles inter partes review: a petition to the Privy Council to vacate a patent. See Lemley, supra, at 1681–1682; Hulme, Privy Council Law and Practice of Letters Patent for Invention From the Restoration to 1794, 33 L. Q. Rev. 63 (1917). The Privy Council was composed of the Crown’s advisers. Lemley, supra, at 1681. From the 17th through the 20th centuries, English patents had a standard revocation clause that permitted six or more Privy Counsellors to declare a patent void if they determined the invention was contrary to law, “prejudicial” or “inconvenient,” not new, or not invented by the patent owner. See 11 W. Holdsworth, A History of English Law 426–427, and n. 6 (1938); Davies, The Early History of the Patent Specification, 50 L. Q. Rev. 86, 102–106 (1934). Individuals could petition the Council to revoke a patent, and the petition was referred to the Attorney General. The Attorney General examined the petition, considered affidavits from the petitioner and patent owner, and heard from counsel. See, e.g., Bull v. Lydall, PC2/81, pp. 180–181 (1706). Depending on the Attorney General’s conclusion, the Council would either void the patent or dismiss the petition. See, e.g., Darby v. Betton, PC2/99, pp. 358–359 (1745–1746) (voiding the patent); Baker v. James, PC2/103, pp. 320–321, 346–347 (1752) (dismissing the petition). The Privy Council was a prominent feature of the English system. It had exclusive authority to revoke patents until 1753, and after that, it had concurrent jurisdiction with the courts. See Hulme, 33 L. Q. Rev., at 189–191, 193–194. The Privy Council continued to consider revocation claims and to revoke patents throughout the 18th century. Its last revocation was in 1779. See id., at 192–193. It considered, but did not act on, revocation claims in 1782, 1794, and 1810. See ibid.; Board of Ordinance v. Parr, PC1/3919 (1810). The Patent Clause in our Constitution “was written against the backdrop” of the English system. Graham, 383 U. S., at 5. Based on the practice of the Privy Council, it was well understood at the founding that a patent system could include a practice of granting patents subject to potential cancellation in the executive proceeding of the Privy Council. The parties have cited nothing in the text or history of the Patent Clause or Article III to suggest that the Framers were not aware of this common practice. Nor is there any reason to think they excluded this practice during their deliberations. And this Court has recognized that, “[w]ithin the scope established by the Constitution, Congress may set out conditions and tests for patentability.” Id., at 6. We conclude that inter partes review is one of those conditions.[4] For similar reasons, we disagree with the dissent’s assumption that, because courts have traditionally adjudicated patent validity in this country, courts must forever continue to do so. See post, at 8–10. Historical practice is not decisive here because matters governed by the public-rights doctrine “from their nature” can be resolved in multiple ways: Congress can “reserve to itself the power to decide,” “delegate that power to executive officers,” or “commit it to judicial tribunals.” Ex parte Bakelite Corp., 279 U. S., at 451. That Congress chose the courts in the past does not foreclose its choice of the PTO today. D Finally, Oil States argues that inter partes review violates Article III because it shares “every salient characteristic associated with the exercise of the judicial power.” Brief for Petitioner 20. Oil States highlights various procedures used in inter partes review: motion practice before the Board; discovery, depositions, and cross-examination of witnesses; introduction of evidence and objections based on the Federal Rules of Evidence; and an adversarial hearing before the Board. See 35 U. S. C. §316(a); 77 Fed. Reg. 48758, 48761–48763 (2012). Similarly, Oil States cites PTO regulations that use terms typically associated with courts—calling the hearing a “trial,” id., at 48758; the Board members “judges,” id., at 48763; and the Board’s final decision a “judgment,” id., at 48761, 48766–48767. But this Court has never adopted a “looks like” test to determine if an adjudication has improperly occurred outside of an Article III court. The fact that an agency uses court-like procedures does not necessarily mean it is exercising the judicial power. See Freytag, 501 U. S., at 910 (opinion of Scalia, J.). This Court has rejected the notion that a tribunal exercises Article III judicial power simply because it is “called a court and its decisions called judgments.” Williams v. United States, 289 U. S. 553, 563 (1933) . Nor does the fact that an administrative adjudication is final and binding on an individual who acquiesces in the result necessarily make it an exercise of the judicial power. See, e.g., Murray’s Lessee, 18 How., at 280–281 (permitting the Treasury Department to conduct “final and binding” audits outside of an Article III court). Al- though inter partes review includes some of the features of adversarial litigation, it does not make any binding determination regarding “the liability of [Greene’s Energy] to [Oil States] under the law as defined.” Crowell, 285 U. S., at 51. It remains a matter involving public rights, one “between the government and others, which from [its] nature do[es] not require judicial determination.” Ex parte Bakelite Corp., 279 U. S., at 451.[5] E We emphasize the narrowness of our holding. We address the constitutionality of inter partes review only. We do not address whether other patent matters, such as infringement actions, can be heard in a non-Article III forum. And because the Patent Act provides for judicial review by the Federal Circuit, see 35 U. S. C. §319, we need not consider whether inter partes review would be constitutional “without any sort of intervention by a court at any stage of the proceedings,” Atlas Roofing Co. v. Occupational Safety and Health Review Comm’n, 430 U. S. 442, 455 , n. 13 (1977). Moreover, we address only the precise constitutional challenges that Oil States raised here. Oil States does not challenge the retroactive application of inter partes review, even though that procedure was not in place when its patent issued. Nor has Oil States raised a due process challenge. Finally, our decision should not be misconstrued as suggesting that patents are not property for purposes of the Due Process Clause or the Takings Clause. See, e.g., Florida Prepaid Postsecondary Ed. Expense Bd. v. College Savings Bank, 527 U. S. 627, 642 (1999) ; James v. Campbell, 104 U. S. 356, 358 (1882) . IV In addition to Article III, Oil States challenges inter partes review under the Seventh Amendment. The Seventh Amendment preserves the “right of trial by jury” in “Suits at common law, where the value in controversy shall exceed twenty dollars.” This Court’s precedents establish that, when Congress properly assigns a matter to adjudication in a non-Article III tribunal, “the Seventh Amendment poses no independent bar to the adjudication of that action by a nonjury factfinder.” Granfinanciera, S. A. v. Nordberg, 492 U. S. 33, 53 –54 (1989); accord, Atlas Roofing Co., supra, at 450–455. No party challenges or attempts to distinguish those precedents. Thus, our rejection of Oil States’ Article III challenge also resolves its Seventh Amendment challenge. Because inter partes review is a matter that Congress can properly assign to the PTO, a jury is not necessary in these proceedings. V Because inter partes review does not violate Article III or the Seventh Amendment, we affirm the judgment of the Court of Appeals. It is so ordered. Notes 1 The Director has delegated his authority to the Patent Trial and Appeal Board. See 37 CFR §42.108(c) (2017). 2 This Court has also recognized this dynamic for state-issued franchises. For instance, States often reserve the right to alter or revoke a corporate charter either “in the act of incorporation or in some general law of the State which was in operation at the time the charter was granted.” Pennsylvania College Cases, 13 Wall. 190, 214, and n. † (1872). That reservation remains effective even after the corporation comes into existence, and such alterations do not offend the Contracts Clause of Article I, §10. See Pennsylvania College Cases, supra, at 212–214; e.g., Miller v. State, 15 Wall. 478, 488–489 (1873). 3 The dissent points to McCormick’s statement that the Patent Office Commissioner could not invalidate the patent at issue because it would “ ‘deprive the applicant of his property without due process of law, and would be in fact an invasion of the judicial branch.’ ” Post, at 10 (quoting McCormick Harvesting Machine Co. v. Aultman, 169 U. S. 606, 612 (1898) ). But that statement followed naturally from the Court’s determination that, under the Patent Act of 1870, the Commissioner “was functus officio” and “had no power to revoke, cancel, or annul” the patent at issue. 169 U. S., at 611–612. Nor is it significant that the McCormick Court “equated invention patents with land patents.” Post, at 10. McCormick itself makes clear that the analogy between the two depended on the particulars of the Patent Act of 1870. See 169 U. S., at 609–610. Modern invention patents, by contrast, are meaningfully different from land patents. The land-patent cases invoked by the dissent involved a “transaction [in which] ‘all authority or control’ over the lands has passed from ‘the Executive Department.’ ” Boesche v. Udall, 373 U. S. 472, 477 (1963) (quoting Moore v. Robbins, 96 U. S. 530, 533 (1878) ). Their holdings do not apply when “the Government continues to possess some measure of control over” the right in question. Boesche, 373 U. S., at 477; see id., at 477–478 (affirming administrative cancellations of public-land leases). And that is true of modern invention patents under the current Patent Act, which gives the PTO continuing authority to review and potentially cancel patents after they are issued. See 35 U. S. C. §§261, 311–319. 4 Oil States also suggests that inter partes review could be an unconstitutional condition because it conditions the benefit of a patent on accepting the possibility of inter partes review. Cf. Koontz v. St. Johns River Water Management Dist., 570 U. S. 595, 604 (2013) (“[T]he government may not deny a benefit to a person because he exercises a constitutional right” (internal quotation marks omitted)). Even assuming a patent is a “benefit” for purposes of the unconstitutional-conditions doctrine, that doctrine does not apply here. The doctrine prevents the Government from using conditions “to produce a result which it could not command directly.” Perry v. Sindermann, 408 U. S. 593, 597 (1972) (internal quotation marks and alterations omitted). But inter partes review is consistent with Article III, see Part III–A, supra, and falls within Congress’ Article I authority, see Part III–C, supra, so it is something Congress can “command directly,” Perry, supra, at 597. 5 Oil States also points out that inter partes review “is initiated by private parties and implicates no waiver of sovereign immunity.” Brief for Petitioner 30–31. But neither of those features takes inter partes review outside of the public-rights doctrine. That much is clear from United States v. Duell, 172 U. S. 576 (1899) , which held that the doctrine covers interference proceedings—a procedure to “determin[e] which of two claimants is entitled to a patent”—even though interference proceedings were initiated by “ ‘private interests compet[ing] for preference’ ” and did not involve a waiver of sovereign immunity. Id., at 582, 586 (quoting Butterworth v. United States ex rel. Hoe, 112 U. S. 50, 59 (1884) ). Also, inter partes review is not initiated by private parties in the way that a common-law cause of action is. To be sure, a private party files the petition for review. 35 U. S. C. §311(a). But the decision to institute review is made by the Director and committed to his unreviewable discretion. See Cuozzo Speed Technologies, LLC v. Lee, 579 U. S. ___, ___ (2016) (slip op., at 9).

583.US.2017_16-498

Petitioner David Patchak filed suit challenging the authority of the Secretary of the Interior to invoke the Indian Reorganization Act, 25 U. S. C. §5108, and take into trust a property (Bradley Property) on which respondent Match-E-Be-Nash-She-Wish Band of Pottawatomi Indians wished to build a casino. In Match-E-Be-Nash-She-Wish Band of Pottawatomi Indians v. Patchak, 567 U. S. 209 (Patchak I), this Court held that the Secretary lacked sovereign immunity and that Patchak had standing, and it remanded the case for further proceedings. While the suit was back in District Court, Congress enacted the Gun Lake Act, 128Stat. 1913, which “reaffirmed as trust land” the Bradley Property, §2(a), and provided that “an action . . . relating to [that] land shall not be filed or maintained in a Federal court and shall be promptly dismissed,” §2(b). In response, the District Court dismissed Patchak’s suit, and the D. C. Circuit affirmed. Held: The judgment is affirmed. 828 F. 3d 995, affirmed. Justice Thomas, joined by Justice Breyer, Justice Alito, and Justice Kagan, concluded that §2(b) of the Gun Lake Act does not violate Article III of the Constitution. Pp. 4–16. (a) Congress may not exercise the judicial power, see Plaut v. Spendthrift Farm, Inc., 514 U. S. 211 , but the legislative power permits Congress to make laws that apply retroactively to pending lawsuits, even when it effectively ensures that one side will win, Bank Markazi v. Peterson, 578 U. S. ___, ___–___. Permissible exercises of the legislative power and impermissible infringements of the judicial power are distinguished by the following rule: Congress violates Article III when it “compel[s] . . . findings or results under old law,” Robertson v. Seattle Audubon Soc., 503 U. S. 429 , but not when it “changes the law,” Plaut, supra, at 218. Pp. 4–6. (b) By stripping federal courts of jurisdiction over actions “relating to” the Bradley Property, §2(b) changes the law. Pp. 6–10. (1) Section 2(b) is best read as a jurisdiction-stripping statute. It uses jurisdictional language, imposes jurisdictional consequences, and applies “[n]otwithstanding any other provision of law,” including the general grant of federal-question jurisdiction, 28 U. S. C. §1331. And while §2(b) does not use the word “jurisdiction,” jurisdictional statutes are not required to do so. See Sebelius v. Auburn Regional Medical Center, 568 U. S. 145 . Indeed, §2(b) uses language similar to language used in other jurisdictional statutes. See, e.g., Gonzalez v. Thaler, 565 U. S. 134 . And §2(b) cannot plausibly be read as anything other than jurisdictional. Pp. 6–7. (2) When Congress strips federal courts of jurisdiction, it exercises a valid legislative power. This Court has held that Congress generally does not violate Article III when it strips federal jurisdiction over a class of cases, see Ex parte McCardle, 7 Wall. 506, 514, and has reaffirmed these principles on many occasions, see, e.g., Steel Co. v. Citizens for Better Environment, 523 U. S. 83 –95. Pp. 7–10. (b) Patchak’s two arguments for why §2(b) violates Article III even if it does strip jurisdiction—that the provision flatly directs federal courts to dismiss lawsuits without allowing them to interpret or apply any new law, and that it attempts to interfere with this Court’s decision in Patchak I that his suit “may proceed,” 567 U. S., at 212—are unpersuasive. Pp. 10–15. Justice Ginsburg, joined by Justice Sotomayor, concluded that Congress’ authority to forgo or retain the Government’s sovereign immunity from suit suffices to decide this case. With Patchak I in mind, Congress acted effectively to displace the Administrative Procedure Act’s waiver of immunity for suits against the United States—which enabled Patchak to launch this litigation—with a contrary command applicable to the Bradley Property. Pp. 1–2. Justice Sotomayor concluded that §2(b) of the Gun Lake Act is most naturally read as having restored the Federal Government’s sovereign immunity from Patchak’s suit challenging the trust status of the Bradley Property. Pp. 1–2. Thomas, J., announced the judgment of the Court and delivered an opinion, in which Breyer, Alito, and Kagan, JJ., joined. Breyer, J., filed a concurring opinion. Ginsburg, J., filed an opinion concurring in the judgment, in which Sotomayor, J., joined. Sotomayor, J., filed an opinion concurring in the judgment. Roberts, C. J., filed a dissenting opinion, in which Kennedy and Gorsuch, JJ., joined.

in which Justice Breyer, Justice Alito, and Justice Kagan join. Petitioner, David Patchak, sued the Secretary of the Interior for taking land into trust on behalf of an Indian Tribe. While his suit was pending in the District Court, Congress enacted the Gun Lake Trust Land Reaffirmation Act (Gun Lake Act or Act), Pub. L. 113–179, 128Stat. 1913, which provides that suits relating to the land “shall not be filed or maintained in a Federal court and shall be promptly dismissed.” Patchak contends that, in enacting this statute, Congress impermissibly infringed the judicial power that Article III of the Constitution vests exclusively in the Judicial Branch. Because we disagree, we affirm the judgment of the United States Court of Appeals for the District of Columbia Circuit. I The Match-E-Be-Nash-She-Wish Band of Pottawatomi Indians (Band) resides in southwestern Michigan, near the township of Wayland. The Band traces its relationship with the United States back hundreds of years, pointing to treaties it negotiated with the Federal Government as early as 1795. But the Secretary of the Interior did not formally recognize the Band until 1999. See 63 Fed. Reg. 56936 (1998); 65 Fed. Reg. 13298 (2000). After obtaining formal recognition, the Band identified a 147-acre parcel of land in Wayland, known as the Bradley Property, where it wanted to build a casino. The Band asked the Secretary to invoke the Indian Reorganization Act, §5, 48Stat. 985, 25 U. S. C. §5108, and take the Bradley Property into trust.[1] In 2005, the Secretary agreed and posted a notice informing the public that the Bradley Property would be taken into trust for the Band. See 70 Fed. Reg. 25596 (2005). The Michigan Gambling Opposition (MichGO) sued, alleging that the Secretary’s decision violated federal environmental and gaming laws. After several years of litigation, the D. C. Circuit affirmed the dismissal of MichGo’s claims, and this Court denied certiorari. Michigan Gambling Opposition v. Kempthorne, 525 F. 3d 23 (2008), cert. denied, 555 U. S. 1137 (2009) . In January 2009, the Secretary formally took the Bradley Property into trust. And in February 2011, the Band opened its casino. Before the Secretary formally took the land into trust, a nearby landowner, David Patchak, filed another lawsuit challenging the Secretary’s decision. Invoking the Administrative Procedure Act, 5 U. S. C. §§702, 706(2), Patchak alleged that the Secretary lacked statutory authority to take the Bradley Property into trust for the Band. The Indian Reorganization Act does not allow the Secretary to take land into trust for tribes that were not under federal jurisdiction when the statute was enacted in 1934. See Carcieri v. Salazar, 555 U. S. 379 –383 (2009). The Band was not federally recognized until 1999, which Patchak argued was more than 65 years too late. Based on this alleged statutory violation, Patchak sought to reverse the Secretary’s decision to take the Bradley Property into trust. The Secretary raised preliminary objections to Patchak’s suit, contending that it was barred by sovereign immunity and that Patchak lacked prudential standing to bring it. The District Court granted the Secretary’s motion to dismiss, but the D. C. Circuit reversed. Patchak v. Salazar, 646 F. Supp. 2d 72 (DC 2009), rev’d, 632 F. 3d 702 (2011). This Court granted certiorari and affirmed the D. C. Circuit. Match-E-Be-Nash-She-Wish Band of Pottawatomi Indians v. Patchak, 567 U. S. 209 (2012) (Patchak I ). This Court’s decision in Patchak I held that Congress had waived the Secretary’s sovereign immunity from suits like Patchak’s. Id., at 215–224. It also held that Patchak had prudential standing because his suit arguably fell within the “zone of interests” protected by the Indian Reorganization Act. Id., at 224–228. Because Patchak had standing and the Secretary lacked immunity, this Court concluded that “Patchak’s suit may proceed,” id., at 212, and remanded for further proceedings, id., at 228. In September 2014, while Patchak’s suit was back in the District Court, Congress enacted the Gun Lake Act, 128Stat. 1913. Section 2(a) of the Act states that the Bradley Property “is reaffirmed as trust land, and the actions of the Secretary of the Interior in taking that land into trust are ratified and confirmed.” Section 2(b) then provides the following: “No Claims.—Notwithstanding any other provision of law, an action (including an action pending in a Federal court as of the date of enactment of this Act) relating to the land described in subsection (a) shall not be filed or maintained in a Federal court and shall be promptly dismissed.” Based on §2(b), the District Court entered summary judgment against Patchak and dismissed his suit for lack of jurisdiction. 109 F. Supp. 3d 152 (DC 2015). The D. C. Circuit affirmed. Patchak v. Jewell, 828 F. 3d 995 (2016). It held that “[t]he language of the Gun Lake Act makes plain that Congress has stripped federal courts of subject matter jurisdiction” over suits, like Patchak’s, that relate to the Bradley Property. Id., at 1001. The D. C. Circuit rejected Patchak’s argument that §2(b) violates Article III of the Constitution. Id., at 1001–1003. Article III prohibits Congress from “direct[ing] the result of pending litigation,” the D. C. Circuit reasoned, but it does not prohibit Congress from “ ‘supply[ing] new law.’ ” Id., at 1002 (quoting Robertson v. Seattle Audubon Soc., 503 U. S. 429, 439 (1992) ). Section 2(b) supplies new law: “[I]f an action relates to the Bradley Property, it must promptly be dismissed.” 828 F. 3d, at 1003. We granted certiorari to review whether §2(b) violates Article III of the Constitution.[2] See 581 U. S. ___ (2017). Because it does not, we now affirm. II A The Constitution creates three branches of Government and vests each branch with a different type of power. See Art. I, §1; Art. II, §1, cl. 1; Art. III, §1. “To the legislative department has been committed the duty of making laws; to the executive the duty of executing them; and to the judiciary the duty of interpreting and applying them in cases properly brought before the courts.” Massachusetts v. Mellon, 262 U. S. 447, 488 (1923) ; see also Wayman v. Southard, 10 Wheat. 1, 46 (1825) (Marshall, C. J.) (“[T]he legislature makes, the executive executes, and the judiciary construes the law”). By vesting each branch with an exclusive form of power, the Framers kept those powers separate. See INS v. Chadha, 462 U. S. 919, 946 (1983) . Each branch “exercise[s] . . . the powers appropriate to its own department,” and no branch can “encroach upon the powers confided to the others.” Kilbourn v. Thompson, 103 U. S. 168, 191 (1881) . This system prevents “[t]he accumulation of all powers, legislative, executive, and judiciary, in the same hands,” The Federalist No. 47, p. 301 (C. Rossiter ed. 1961) (J. Madison)—an accumulation that would pose an inherent “threat to liberty,” Clinton v. City of New York, 524 U. S. 417, 450 (1998) (Kennedy, J., concurring). The separation of powers, among other things, prevents Congress from exercising the judicial power. See Plaut v. Spendthrift Farm, Inc., 514 U. S. 211, 218 (1995) . One way that Congress can cross the line from legislative power to judicial power is by “usurp[ing] a court’s power to interpret and apply the law to the [circumstances] before it.” Bank Markazi v. Peterson, 578 U. S. ___, ___ (2016) (slip op., at 12). The simplest example would be a statute that says, “In Smith v. Jones, Smith wins.” See id., at ___–___, n. 17 (slip op., at 12–13, n. 17). At the same time, the legislative power is the power to make law, and Congress can make laws that apply retroactively to pending lawsuits, even when it effectively ensures that one side wins. See id., at ___–___ (slip op., at 15–19). To distinguish between permissible exercises of the legislative power and impermissible infringements of the judicial power, this Court’s precedents establish the following rule: Congress violates Article III when it “compel[s] . . . findings or results under old law.” Seattle Audubon, supra, at 438. But Congress does not violate Article III when it “changes the law.” Plaut, supra, at 218. B Section 2(b) changes the law. Specifically, it strips federal courts of jurisdiction over actions “relating to” the Bradley Property. Before the Gun Lake Act, federal courts had jurisdiction to hear these actions. See 28 U. S. C. §1331. Now they do not. This kind of legal change is well within Congress’ authority and does not violate Article III. 1 Section 2(b) strips federal jurisdiction over suits relating to the Bradley Property. The statute uses jurisdictional language. It states that an “action” relating to the Bradley Property “shall not be filed or maintained in a Federal court.” It imposes jurisdictional consequences: Actions relating to the Bradley Property “shall be promptly dismissed.” See Ex parte McCardle, 7 Wall. 506, 514 (1869) (“[W]hen [jurisdiction] ceases to exist, the only function remaining to the court is that of announcing the fact and dismissing the cause”). Section 2(b) has no exceptions. Cf. Reed Elsevier, Inc. v. Muchnick, 559 U. S. 154, 165 (2010) . And it applies “[n]otwithstanding any other provision of law,” including the general grant of federal-question jurisdiction, 28 U. S. C. §1331. Although §2(b) does not use the word “jurisdiction,” this Court does not require jurisdictional statutes to “incant magic words.” Sebelius v. Auburn Regional Medical Center, 568 U. S. 145, 153 (2013) . Indeed, §2(b) uses language similar to other statutes that this Court has deemed jurisdictional. See, e.g., Gonzalez v. Thaler, 565 U. S. 134, 142 (2012) (“ ‘an appeal may not be taken’ ” (quoting 28 U. S. C. §2253(c)(1))); Keene Corp. v. United States, 508 U. S. 200 –209 (1993) (“ ‘[n]o person shall file or prosecute’ ” (quoting 36Stat. 1138)); Weinberger v. Salfi, 422 U. S. 749, 756 (1975) (“ ‘[n]o action . . . shall be brought under [ 28 U. S. C. §1331]’ ” (quoting 42 U. S. C. §405(h))). Our conclusion that §2(b) is jurisdictional is bolstered by the fact that it cannot plausibly be read as anything else. Section 2(b) is not one of the nonjurisdictional rules that this Court’s precedents have identified as “important and mandatory” but not governing “a court’s adjudicatory capacity.” Henderson v. Shinseki, 562 U. S. 428, 435 (2011) . Section 2(b) does not identify an “element of [the] plaintiff ’s claim for relief ” or otherwise define its “substantive adequacy.” Arbaugh v. Y & H Corp., 546 U. S. 500, 516, 504 (2006) . Nor is it a “claim-processing rule,” like a filing deadline or an exhaustion requirement, that requires the parties to “take certain procedural steps at certain specified times.” Henderson, supra, at 435. Instead, §2(b) completely prohibits actions relating to the Bradley Property. Because §2(b) addresses “a court’s competence to adjudicate a particular category of cases,” Wachovia Bank, N. A. v. Schmidt, 546 U. S. 303, 316 (2006) , it is best read as a jurisdiction-stripping statute. 2 Statutes that strip jurisdiction “chang[e] the law” for the purpose of Article III, Plaut, supra, at 218, just as much as other exercises of Congress’ legislative authority. Article I permits Congress “[t]o constitute Tribunals inferior to the supreme Court,” §8, and Article III vests the judicial power “in one supreme Court, and in such inferior Courts as the Congress may from time to time ordain and establish,” §1. These provisions reflect the so-called Madison- ian Compromise, which resolved the Framers’ disagreement about creating lower federal courts by leaving that decision to Congress. See Printz v. United States, 521 U. S. 898, 907 (1997) ; 1 Records of the Federal Convention of 1787, p. 125 (M. Farrand ed. 1911). Congress’ greater power to create lower federal courts includes its lesser power to “limit the jurisdiction of those Courts.” United States v. Hudson, 7 Cranch 32, 33 (1812); accord, Lockerty v. Phillips, 319 U. S. 182, 187 (1943) . So long as Congress does not violate other constitutional provisions, its “control over the jurisdiction of the federal courts” is “plenary.” Trainmen v. Toledo, P. & W. R. Co., 321 U. S. 50 –64 (1944); see also Bowles v. Russell, 551 U. S. 205, 212 (2007) (“Within constitutional bounds, Congress decides what cases the federal courts have jurisdiction to consider”). Thus, when Congress strips federal courts of jurisdic- tion, it exercises a valid legislative power no less than when it lays taxes, coins money, declares war, or invokes any other power that the Constitution grants it. Indeed, this Court has held that Congress generally does not violate Article III when it strips federal jurisdiction over a class of cases.[3] Shortly after the Civil War, for example, Congress repealed this Court’s appellate jurisdiction over certain habeas corpus cases. See Act of Mar. 27, 1868, ch. 34, §2, 15Stat. 44; see also U. S. Const., Art. III, §2 (permitting Congress to make “Exceptions” to this Court’s appellate jurisdiction). William McCardle, a military prisoner whose appeal was pending at the time, argued that the repealing statute was “an exercise by the Congress of judicial power.” 7 Wall., at 510. This Court disagreed. Jurisdiction-stripping statutes, the Court explained, do not involve “the exercise of judicial power” or “legislative interference with courts in the exercising of continuing jurisdiction.” Id., at 514. Because jurisdiction is the “power to declare the law” in the first place, “judicial duty is not less fitly performed by declining ungranted jurisdiction than in exercising firmly that which the Constitution and the laws confer.” Id., at 514–515.[4] This Court has reaffirmed these principles on many occasions. Congress generally does not infringe the judicial power when it strips jurisdiction because, with limited exceptions, a congressional grant of jurisdiction is a prerequisite to the exercise of judicial power. See Steel Co. v. Citizens for Better Environment, 523 U. S. 83 –95 (1998) (“The requirement that jurisdiction be established as a threshold matter ‘spring[s] from the nature and limits of the judicial power of the United States’ ” (quoting Mansfield, C. & L. M. R. Co. v. Swan, 111 U. S. 379, 382 (1884) )); Cary v. Curtis, 3 How. 236, 245 (1845) (“[T]he judicial power of the United States . . . is (except in enumerated instances, applicable exclusively to this court) dependent . . . entirely upon the action of Congress”); Hudson, supra, at 33 (similar). “To deny this position” would undermine the separation of powers by “elevat[ing] the judicial over the legislative branch.” Cary, supra, at 245. Congress’ power over federal jurisdiction is “an essential ingredient of separation and equilibration of powers, restraining the courts from acting at certain times, and even restraining them from acting permanently regarding certain subjects.” Steel Co., supra, at 101. In sum, §2(b) strips jurisdiction over suits relating to the Bradley Property. It is a valid exercise of Congress’ legislative power. And because it changes the law, it does not infringe the judicial power. The constitutionality of jurisdiction-stripping statutes like this one is well established. III Patchak does not dispute Congress’ power to withdraw jurisdiction from the federal courts. He instead raises two arguments why §2(b) violates Article III, even if it strips jurisdiction. First, relying on United States v. Klein, 13 Wall. 128 (1872), Patchak argues that §2(b) flatly directs federal courts to dismiss lawsuits without allowing them to interpret or apply any new law. Second, relying on Plaut, 514 U. S. 211 , Patchak argues that §2(b) attempts to interfere with this Court’s decision in Patchak I—specifically, its conclusion that his suit “may proceed,” 567 U. S., at 212. We reject both arguments. A Section 2(b) does not flatly direct federal courts to dismiss lawsuits under old law. It creates new law for suits relating to the Bradley Property, and the District Court interpreted and applied that new law in Patchak’s suit. Section 2(b)’s “relating to” standard effectively guaranteed that Patchak’s suit would be dismissed. But “a statute does not impinge on judicial power when it directs courts to apply a new legal standard to undisputed facts.” Bank Markazi, 578 U. S., at ___ (slip op., at 17). “[I]t is not any the less a case or controversy upon which a court possessing the federal judicial power may rightly give judgment” when the arguments before the court are “uncontested or incontestable.” Pope v. United States, 323 U. S. 1, 11 (1944) . Patchak argues that the last four words of §2(b)—“shall be promptly dismissed”—direct courts to reach a particular outcome. But a statute does not violate Article III merely because it uses mandatory language. See Seattle Audubon, 503 U. S., at 439. Instead of directing outcomes, the mandatory language in §2(b) “simply imposes the consequences” of a court’s determination that it lacks jurisdiction because a suit relates to the Bradley Property. Miller v. French, 530 U. S. 327, 349 (2000) ; see McCardle, 7 Wall., at 514.[5] Patchak compares §2(b) to the statute this Court held unconstitutional in Klein. In that case, the administrator of the estate of V. F. Wilson, a former Confederate soldier, sued to recover the value of some cotton that the Government had seized during the war. 13 Wall., at 132. The relevant statute required claimants to prove their loyalty in order to reclaim their property. Ch. 120, §3, 12Stat. 820. Wilson had received a pardon before he died, 13 Wall., at 132, and this Court had held that pardons conclusively prove loyalty under the statute, see United States v. Padelford, 9 Wall. 531, 543 (1870). But after Wilson’s administrator secured a judgment in his favor, 13 Wall., at 132, Congress passed a statute making pardons proof of disloyalty and declaring that, if a claimant had accepted one, “the jurisdiction of the court in the case shall cease, and the court shall forthwith dismiss the suit of such claimant.” Act of July 12, 1870, 16Stat. 235. If the court had already entered judgment in favor of a pardoned claimant and the Government had appealed, the statute instructed this Court to dismiss the whole suit for lack of jurisdiction. See ibid. Klein held that this statute infringed the executive power by attempting to “change the effect of . . . a pardon.” Id., at 148. Klein also held that the statute infringed the judicial power, see id., at 147, although its reasons for this latter holding were not entirely clear. This Court has since explained that “the statute in Klein infringed the judicial power, not because it left too little for courts to do, but because it attempted to direct the result without altering the legal standards governing the effect of a pardon—standards Congress was powerless to prescribe.” Bank Markazi, supra, at ___ (slip op., at 15). Congress had no authority to declare that pardons are not evidence of loyalty, so it could not achieve the same result by stripping jurisdiction whenever claimants cited pardons as evidence of loyalty. See Klein, 13 Wall., at 147–148. Nor could Congress confer jurisdiction to a federal court but then strip jurisdiction from that same court once the court concluded that a pardoned claimant should prevail under the statute. See id., at 146–147. Patchak’s attempts to compare §2(b) to the statute in Klein are unpersuasive. Section 2(b) does not attempt to exercise a power that the Constitution vests in another branch. And unlike the selective jurisdiction-stripping statute in Klein, §2(b) strips jurisdiction over every suit relating to the Bradley Property. Indeed, Klein itself explained that statutes that do “nothing more” than strip jurisdiction over “a particular class of cases” are constitutional. Id., at 145. That is precisely what §2(b) does. B Section 2(b) does not unconstitutionally interfere with this Court’s decision in Patchak I. Article III, this Court explained in Plaut, prohibits Congress from “retroactively commanding the federal courts to reopen final judgments.” 514 U. S., at 219. But Patchak I did not finally conclude Patchak’s case. See Bradley v. School Bd. of Richmond, 416 U. S. 696 , n. 14 (1974). When this Court said that his suit “may proceed,” 567 U. S., at 212, it meant that the Secretary’s preliminary defenses lacked merit and that Patchak could return to the District Court for further proceedings. It did not mean that Congress was powerless to change the law that governs his case. As this Court emphasized in Plaut, Article III does not prohibit Congress from enacting new laws that apply to pending civil cases. See 514 U. S., at 226–227. When a new law clearly governs pending cases, Article III does not prevent courts from applying it because “each court, at every level, must ‘decide according to existing laws.’ ” Ibid. (quoting United States v. Schooner Peggy, 1 Cranch 103, 109 (1801)). This principle applies equally to statutes that strip jurisdiction. See Landgraf v. USI Film Products, 511 U. S. 244, 274 (1994) ; Kline v. Burke Constr. Co., 260 U. S. 226, 234 (1922) ; Hallowell v. Commons, 239 U. S. 506, 509 (1916) . Because §2(b) expressly references “pending” cases, it applies to Patchak’s suit. And because Patchak’s suit is not final, applying §2(b) here does not offend Article III.[6] Of course, we recognize that the Gun Lake Act was a response to this Court’s decision in Patchak I. The text of the Act, after all, cites both the administrative decision and the property at issue in that case. See §§2(a)–(b). And we understand why Patchak would view the Gun Lake Act as unfair. By all accounts, the Band exercised its political influence to persuade Congress to enact a narrow jurisdiction-stripping provision that effectively ends all lawsuits threatening its casino, including Patchak’s. But the question in this case is “[n]ot favoritism, nor even corruption, but power.” Plaut, supra, at 228; see also McCardle, 7 Wall., at 514 (“We are not at liberty to inquire into the motives of the legislature. We can only examine into its power under the Constitution”). Under this Court’s precedents, Congress has the power to “apply newly enacted, outcome-altering legislation in pending civil cases,” Bank Markazi, 578 U. S., at ___ (slip op., at 16), even when the legislation “govern[s] one or a very small number of specific subjects,” id., at ___ (slip op., at 21). For example, this Court has upheld narrow statutes that identified specific cases by caption and docket number in their text. See id., at ___ (slip op., at 19); Seattle Audubon, 503 U. S., at 440. And this Court has approvingly cited a D. C. Circuit decision, which upheld a statute that retroactively stripped jurisdiction over suits challenging “a single memorial.” Bank Markazi, supra, at ___ (slip op., at 22) (citing National Coalition To Save Our Mall v. Norton, 269 F. 3d 1092 (2001)). If these targeted statutes did not cross the line from legislative to judicial power, then §2(b) does not either. IV The dissent offers a different theory for why §2(b) violates Article III. A statute impermissibly exercises the judicial power, the dissent contends, when it “targets” a particular suit and “manipulates” jurisdiction to direct the outcome, “practical[ly] operat[es]” to affect only one suit, and announces a legal standard that does not “imply some measure of generality” or “preserv[e] . . . an adjudicative role for the courts.” Post, at 8, 11. We doubt that the constitutional line separating the legislative and judicial powers turns on factors such as a court’s doubts about Congress’ unexpressed motives, the number of “cases [that] were pending when the provision was enacted,” or the time left on the statute of limitations. Post, at 8. But even if it did, we disagree with the dissent’s characterization of §2(b). Nothing on the face of §2(b) is limited to Patchak’s case, or even to his challenge under the Indian Reorganization Act. Instead, the text extends to all suits “relating to” the Bradley Property. Thus, §2(b) survives even under the dissent’s theory: It “prospectively govern[s] an open-ended class of disputes,” post, at 6, and its “relating to” standard “preserv[es] . . . an adjudicative role for the courts,” post, at 11. While §2(b)’s “relating to” standard is not difficult to interpret or apply, this Court’s precedents encourage Congress to draft jurisdictional statutes in this manner. See Hertz Corp. v. Friend, 559 U. S. 77, 94 (2010) (“[A]dministrative simplic- ity is a major virtue in a jurisdictional statute. . . . [C]ourts benefit from straightforward rules under which they can readily assure themselves of their power to hear a case”).[7] * * * We conclude that §2(b) of the Gun Lake Act does not violate Article III of the Constitution. The judgment of the Court of Appeals is, therefore, affirmed. It is so ordered. Notes 1 Federal law allows Indian tribes to operate casinos on “Indian lands,” 25 U. S. C. §2710, which includes lands “held in trust by the United States for the benefit of any Indian tribe,” §2703(4)(B). 2 Patchak does not challenge the constitutionality of §2(a) of the Gun Lake Act. See Reply Brief 7; Tr. of Oral Arg. 5. We thus limit our analysis to §2(b). 3 Jurisdiction-stripping statutes can violate other provisions of the Constitution. And, under our precedents, jurisdiction-stripping statutes can violate Article III if they “attemp[t] to direct the result” by effectively altering legal standards that Congress is “powerless to prescribe.” Bank Markazi v. Peterson, 578 U. S. ___, ___ (2016) (slip op., at 15) (citing United States v. Klein, 13 Wall. 128, 146–147 (1872)). 4 The dissent appears to disagree with McCardle, questions the motives of the unanimous Court that decided it, asserts that it is “inconsistent” with Klein, and distinguishes it on the ground that the statute there “did not foreclose all avenues of judicial review.” Post, at 12–13 (opinion of Roberts, C. J.). But the core holding of McCardle—that Congress does not exercise the judicial power when it strips jurisdiction over a class of cases—has never been questioned, has been repeatedly reaffirmed, and was reaffirmed in Klein itself. See 13 Wall., at 145 (“[T]here could be no doubt” that Congress can “den[y] the right of appeal in a particular class of cases”). And if there is any inconsistency between the two, this Court has said that it is Klein—not McCardle—that “cannot [be] take[n] . . . ‘at face value.’ ” Bank Markazi, 578 U. S., at ___ (slip op., at 15) (quoting R. Fallon, J. Manning, D. Meltzer, & D. Shapiro, Hart and Wechsler’s The Federal Courts and the Federal System 324 (7th ed. 2015)). Moreover, it is true that McCardle emphasized that the statute there did not withdraw “the whole appellate power of the court, in cases of habeas corpus.” 7 Wall., at 515 (emphasis added). But McCardle’s reservation, this Court later explained, was responding to a potential problem under the Suspension Clause, not a potential problem under Article III. See Ex parte Yerger, 8 Wall. 85, 102–103 (1869) (“We agree that [jurisdiction] is given subject to exception and regulation by Congress; but it is too plain for argument that the denial to this court of appellate jurisdiction in this class of cases must greatly weaken the efficacy of the writ”); id., at 96 (“It would have been . . . a remarkable anomaly if this court . . . had been denied, under a constitution which absolutely prohibits suspension of the writ, except under extraordinary exigencies, that power in cases of alleged unlawful restraint”). 5 To prove that it does not change the law, Patchak repeatedly asserts that §2(b) does not amend any “generally applicable” statute. Brief for Petitioner 11; Reply Brief 1, 4, 9. But this Court rejected that same argument in Seattle Audubon. Congress can change a law “directly,” or it can change a law indirectly by passing “an entirely separate statute.” 503 U. S., at 439–440. Either way, it changes the law. The same is true for jurisdictional statutes. See Insurance Co. v. Ritchie, 5 Wall. 541 (1867). 6 Retroactive legislation can violate other provisions of the Constitution, such as the Ex Post Facto Clause and the Bills of Attainder Clause. See Bank Markazi, 578 U. S., at ___ (slip op., at 16). But Patchak’s Article III claim is the only challenge to §2(b) before us. 7 We also doubt that the statute this Court upheld in Bank Markazi would survive under the dissent’s theory. The dissent notes that the statute there affected “16 different actions,” while the statute here affects “a single pending case.” Post, at 8. But if the problem is Congress “pick[ing] winners and losers in pending litigation,” post, at 14, then it seems backwards to conclude that Congress is on stronger constitutional footing when it picks winners and losers in more pending cases.

585.US.2017_17-459

Under the Illegal Immigration Reform and Immigrant Responsibility Act of 1996 (IIRIRA), nonpermanent residents who are subject to removal proceedings may be eligible for cancellation of removal if, among other things, they have “been physically present in the United States for a continuous period of not less than 10 years immediately preceding the date of [an] application” for cancellation. 8 U. S. C. §1229(b)(1)(A). Under the stop-time rule, however, the period of continuous presence is “deemed to end . . . when the alien is served a notice to appear under section 1229(a).” §1229(d)(1)(A). Section 1229(a), in turn, provides that the Government shall serve noncitizens in removal proceedings with a written “ ‘notice to appear,’ ” specifying, among other things, “[t]he time and place at which the [removal] proceedings will be held.” §1229(a)(1)(G)(i). Per a 1997 regulation stating that a “notice to appear” served on a noncitizen need only provide “the time, place and date of the initial removal hearing, where practicable,” 62 Fed. Reg. 10332, the Department of Homeland Security (DHS), at least in recent years, almost always serves noncitizens with notices that fail to specify the time, place, or date of initial removal hearings whenever the agency deems it impracticable to include such information. The Board of Immigration Appeals (BIA) has held that such notices trigger the stop-time rule even if they do not specify the time and date of the removal proceedings. Petitioner Wescley Fonseca Pereira is a native and citizen of Brazil who came to the United States in 2000 and remained after his visa expired. Following a 2006 arrest for operating a vehicle while under the influence of alcohol, DHS served Pereira with a document titled “notice to appear” that did not specify the date and time of his initial removal hearing, instead ordering him to appear at a time and date to be set in the future. More than a year later, in 2007, the Immigration Court mailed Pereira a more specific notice setting the date and time for his initial hearing, but the notice was sent to the wrong address and was returned as undeliverable. As a result, Pereira failed to appear, and the Immigration Court ordered him removed in absentia. In 2013, Pereira was arrested for a minor motor vehicle violation and detained by DHS. The Immigration Court reopened the removal proceedings after Pereira demonstrated that he never received the 2007 notice. Pereira then applied for cancellation of removal, arguing that he had been continuously present in the United States for more than 10 years and that the stop-time rule was not triggered by DHS’ initial 2006 notice because the document lacked information about the time and date of his removal hearing. The Immigration Court disagreed and ordered Pereira removed. The BIA agreed with the Immigration Court that the 2006 notice triggered the stop-time rule, even though it failed to specify the time and date of Pereira’s initial removal hearing. The Court of Appeals for the First Circuit denied Pereira’s petition for review of the BIA’s order. Applying the framework set forth in Chevron U. S. A. Inc. v. Natural Resources Defense Council, Inc., 467 U. S. 837, it held that the stop-time rule is ambiguous and that the BIA’s interpretation of the rule was a permissible reading of the statute. Held: A putative notice to appear that fails to designate the specific time or place of the noncitizen’s removal proceedings is not a “notice to appear under §1229(a),” and so does not trigger the stop-time rule. Pp. 7–20. (a) The Court need not resort to Chevron deference, for the unambiguous statutory text alone is enough to resolve this case. Under the stop-time rule, “any period of . . . continuous physical presence” is “deemed to end . . . when the alien is served a notice to appear under section 1229(a).” 8 U. S. C. §1229b(d)(1). By expressly referencing §1229(a), the statute specifies where to look to find out what “notice to appear” means. Section 1229(a), in turn, clarifies that the type of notice “referred to as a ‘notice to appear’ ” throughout the statutory section is a “written notice . . . specifying,” as relevant here, “[t]he time and place at which the [removal] proceedings will be held.” §1229(a)(1)(G)(i). Thus, to trigger the stop-time rule, the Government must serve a notice to appear that, at the very least, “specif[ies]” the “time and place” of the removal hearing. The Government and dissent point out that the stop-time rule refers broadly to a notice to appear under “§1229(a)”—which includes paragraph (1), as well as paragraphs (2) and (3). But that does not matter, because only paragraph (1) bears on the meaning of a “notice to appear.” If anything, paragraph (2), which allows for a “change or postponement” of the proceedings to a “new time and place,” §1229(a)(2)(A)(i), bolsters the Court’s interpretation of the statute because the provision presumes that the Government has already served a “notice to appear” that specified a time and place as required by §1229(a)(1)(G)(i). Another neighboring provision, §1229(b)(1), lends further support for the view that a “notice to appear” must specify the time and place of removal proceedings to trigger the stop-time rule. Section 1229(b)(1) gives a noncitizen “the opportunity to secure counsel before the first [removal] hearing date” by mandating that such “hearing date shall not be scheduled earlier than 10 days after the service of the notice to appear.” For that provision to have any meaning, the “notice to appear” must specify the time and place that the noncitizen, and his counsel, must appear at the removal proceedings. Finally, common sense reinforces the conclusion that a notice that does not specify when and where to appear for a removal proceeding is not a “notice to appear” that triggers the stop-time rule. After all, an essential function of a “notice to appear” is to provide noncitizens “notice” of the information (i.e., the “time” and “place”) that would enable them “to appear” at the removal hearing in the first place. Without conveying such information, the Government cannot reasonably expect noncitizens to appear for their removal proceedings. Pp. 7–13. (b) The Government and the dissent advance a litany of counterarguments, all of which are unpersuasive. To begin, the Government mistakenly argues that §1229(a) is not definitional. That is wrong. Section 1229(a) speaks in definitional terms, requiring that a notice to appear specify, among other things, the “time and place at which the proceedings will be held.” As such, the dissent is misguided in arguing that a defective notice to appear, which fails to specify time-and-place information, is still a notice to appear for purposes of the stop-time rule. Equally unavailing is the Government’s (and the dissent’s) attempt to generate ambiguity in the statute based on the word “under.” In light of the plain language and statutory context, the word “under,” as used in the stop-time rule, clearly means “in accordance with” or “according to” because it connects the stop-time trigger in §1229b(d)(1) to a “notice to appear” that specifies the enumerated time-and-place information. The Government fares no better in arguing that surrounding statutory provisions reinforce its preferred reading of the stop-time rule, as none of those provisions supports its atextual interpretation. Unable to root its reading in the statutory text, the Government and dissent raise a number of practical concerns, but those concerns are meritless and do not justify departing from the statute’s clear text. In a final attempt to salvage its atextual interpretation, the Government turns to the alleged statutory purpose and legislative history of the stop-time rule. Even for those who consider statutory purpose and legislative history, however, neither supports the Government’s position. Requiring the Government to furnish time-and-place information in a notice to appear is entirely consistent with Congress’ stated objective of preventing noncitizens from exploiting administrative delays to accumulate lengthier periods of continuous precedent. Pp. 13–20. 866 F. 3d 1, reversed and remanded. Sotomayor, J., delivered the opinion of the Court, in which Roberts, C. J., and Kennedy, Thomas, Ginsburg, Breyer, Kagan, and Gorsuch, JJ., joined. Kennedy, J., filed a concurring opinion. Alito, J., filed a dissenting opinion.

Nonpermanent residents, like petitioner here, who are subject to removal proceedings and have accrued 10 years of continuous physical presence in the United States, may be eligible for a form of discretionary relief known as cancellation of removal. 8 U. S. C. §1229b(b)(1). Under the so-called “stop-time rule” set forth in §1229b(d)(1)(A), however, that period of continuous physical presence is “deemed to end . . . when the alien is served a notice to appear under section 1229(a).” Section 1229(a), in turn, provides that the Government shall serve noncitizens in removal proceedings with “written notice (in this section referred to as a ‘notice to appear’) . . . specifying” several required pieces of information, including “[t]he time and place at which the [removal] proceedings will be held.” §1229(a)(1)(G)(i).[1] The narrow question in this case lies at the intersection of those statutory provisions. If the Government serves a noncitizen with a document that is labeled “notice to appear,” but the document fails to specify either the time or place of the removal proceedings, does it trigger the stop-time rule? The answer is as obvious as it seems: No. A notice that does not inform a noncitizen when and where to appear for removal proceedings is not a “notice to appear under section 1229(a)” and therefore does not trigger the stop-time rule. The plain text, the statutory context, and common sense all lead inescapably and unambiguously to that conclusion. I A Under the Illegal Immigration Reform and Immigrant Responsibility Act of 1996 (IIRIRA), 110Stat. 3009–546, the Attorney General of the United States has discretion to “cancel removal” and adjust the status of certain nonpermanent residents. §1229b(b). To be eligible for such relief, a nonpermanent resident must meet certain enumerated criteria, the relevant one here being that the noncitizen must have “been physically present in the United States for a continuous period of not less than 10 years immediately preceding the date of [an] application” for cancellation of removal. §1229b(b)(1)(A).[2] IIRIRA also established the stop-time rule at issue in this case. Under that rule, “any period of . . . continuous physical presence in the United States shall be deemed to end . . . when the alien is served a notice to appear under section 1229(a) of this title.”[3] §1229b(d)(1)(A). Section 1229(a), in turn, provides that “written notice (in this section referred to as a ‘notice to appear’) shall be given . . . to the alien . . . specifying”: “(A) The nature of the proceedings against the alien. “(B) The legal authority under which the proceedings are conducted. “(C) The acts or conduct alleged to be in violation of law. “(D) The charges against the alien and the statutory provisions alleged to have been violated. “(E) The alien may be represented by counsel and the alien will be provided (i) a period of time to secure counsel under subsection (b)(1) of this section and (ii) a current list of counsel prepared under subsection (b)(2) of this section. “(F)(i) The requirement that the alien must immediately provide (or have provided) the Attorney General with a written record of an address and telephone number (if any) at which the alien may be contacted respecting proceedings under section 1229a of this title. “(ii) The requirement that the alien must provide the Attorney General immediately with a written record of any change of the alien’s address or telephone number. “(iii) The consequences under section 1229a(b)(5) of this title of failure to provide address and telephone information pursuant to this subparagraph. “(G)(i) The time and place at which the [removal] proceedings will be held. “(ii) The consequences under section 1229a(b)(5) of this title of the failure, except under exceptional cir- cumstances, to appear at such proceedings.” §1229(a)(1) (boldface added). The statute also enables the Government to “change or postpon[e] . . . the time and place of [the removal] proceedings.” §1229(a)(2)(A). To do so, the Government must give the noncitizen “a written notice . . . specifying . . . the new time or place of the proceedings” and “the consequences . . . of failing, except under exceptional circumstances, to attend such proceedings.” Ibid. The Government is not required to provide written notice of the change in time or place of the proceedings if the noncitizen is “not in detention” and “has failed to provide [his] address” to the Government. §1229(a)(2)(B). The consequences of a noncitizen’s failure to appear at a removal proceeding can be quite severe. If a noncitizen who has been properly served with the “written notice required under paragraph (1) or (2) of section 1229(a)” fails to appear at a removal proceeding, he “shall be ordered removed in absentia” if the Government “establishes by clear, unequivocal, and convincing evidence that the written notice was so provided and that the alien is removable.” §1229a(b)(5)(A). Absent “exceptional circumstances,” a noncitizen subject to an in absentia removal order is ineligible for some forms of discretionary relief for 10 years if, “at the time of the notice described in paragraph (1) or (2) of section 1229(a),” he “was provided oral notice . . . of the time and place of the proceedings and of the consequences” of failing to appear. §1229a(b)(7). In certain limited circumstances, however, a removal order entered in absentia may be rescinded—e.g., when the noncitizen “demonstrates that [he] did not receive notice in accordance with paragraph (1) or (2) of section 1229(a).” §1229a(b)(5)(C)(ii). B In 1997, shortly after Congress passed IIRIRA, the Attorney General promulgated a regulation stating that a “notice to appear” served on a noncitizen need only provide “the time, place and date of the initial removal hearing, where practicable.” 62 Fed. Reg. 10332 (1997). Per that regulation, the Department of Homeland Security (DHS), at least in recent years, almost always serves noncitizens with notices that fail to specify the time, place, or date of initial removal hearings whenever the agency deems it impracticable to include such information. See Brief for Petitioner 14; Brief for Respondent 48–49; Tr. of Oral Arg. 52–53 (Government’s admission that “almost 100 percent” of “notices to appear omit the time and date of the proceeding over the last three years”). Instead, these notices state that the times, places, or dates of the initial hearings are “to be determined.” Brief for Petitioner 14. In Matter of Camarillo, 25 I. & N. Dec. 644 (2011), the Board of Immigration Appeals (BIA) addressed whether such notices trigger the stop-time rule even if they do not specify the time and date of the removal proceedings. The BIA concluded that they do. Id., at 651. It reasoned that the statutory phrase “notice to appear ‘under section [1229](a)’ ” in the stop-time rule “merely specifies the document the DHS must serve on the alien to trigger the ‘stop-time’ rule,” but otherwise imposes no “substantive requirements” as to what information that document must include to trigger the stop-time rule. Id., at 647. C Petitioner Wescley Fonseca Pereira is a native and citizen of Brazil. In 2000, at age 19, he was admitted to the United States as a temporary “non-immigrant visitor.” App. to Pet. for Cert. 3a. After his visa expired, he remained in the United States. Pereira is married and has two young daughters, both of whom are United States citizens. He works as a handyman and, according to submissions before the Immigration Court, is a well-respected member of his community. In 2006, Pereira was arrested in Massachusetts for operating a vehicle while under the influence of alcohol. On May 31, 2006, while Pereira was detained, DHS served him (in person) with a document labeled “Notice to Appear.” App. 7–13. That putative notice charged Pereira as removable for overstaying his visa, informed him that “removal proceedings” were being initiated against him, and provided him with information about the “[c]onduct of the hearing” and the consequences for failing to appear. Id., at 7, 10–12. Critical here, the notice did not specify the date and time of Pereira’s removal hearing. Instead, it ordered him to appear before an Immigration Judge in Boston “on a date to be set at a time to be set.” Id., at 9 (underlining in original). More than a year later, on August 9, 2007, DHS filed the 2006 notice with the Boston Immigration Court. The Immigration Court thereafter attempted to mail Pereira a more specific notice setting the date and time for his initial removal hearing for October 31, 2007, at 9:30 a.m. But that second notice was sent to Pereira’s street address rather than his post office box (which he had provided to DHS), so it was returned as undeliverable. Because Pe- reira never received notice of the time and date of his re- moval hearing, he failed to appear, and the Immigration Court ordered him removed in absentia. Unaware of that re- moval order, Pereira remained in the United States. In 2013, after Pereira had been in the country for more than 10 years, he was arrested for a minor motor vehicle violation (driving without his headlights on) and was subsequently detained by DHS. The Immigration Court reopened the removal proceedings after Pereira demonstrated that he never received the Immigration Court’s 2007 notice setting out the specific date and time of his hearing. Pereira then applied for cancellation of removal, arguing that the stop-time rule was not triggered by DHS’ initial 2006 notice because the document lacked information about the time and date of his removal hearing. The Immigration Court disagreed, finding the law “quite settled that DHS need not put a date certain on the Notice to Appear in order to make that document effective.” App. to Pet. for Cert. 23a. The Immigration Court therefore concluded that Pereira could not meet the 10-year physical-presence requirement under §1229b(b), thereby render- ing him statutorily ineligible for cancellation of removal, and ordered Pereira removed from the country. The BIA dismissed Pereira’s appeal. Adhering to its precedent in Camarillo, the BIA agreed with the Immigration Court that the 2006 notice triggered the stop-time rule and that Pereira thus failed to satisfy the 10-year physical-presence requirement and was ineligible for cancellation of removal. The Court of Appeals for the First Circuit denied Pereira’s petition for review of the BIA’s order. 866 F. 3d 1 (2017). Applying the framework set forth in Chevron U. S. A. Inc. v. Natural Resources Defense Council, Inc., 467 U. S. 837 (1984), the Court of Appeals first found that the stop-time rule in §1229b(d)(1) is ambiguous because it “does not explicitly state that the date and time of the hearing must be included in a notice to appear in order to cut off an alien’s period of continuous physical presence.” 866 F. 3d, at 5. Then, after reviewing the statutory text and structure, the administrative context, and pertinent legislative history, the Court of Appeals held that the BIA’s interpretation of the stop-time rule was a permissible reading of the statute. Id., at 6–8. II A The Court granted certiorari in this case, 583 U. S. ___ (2018), to resolve division among the Courts of Appeals on a simple, but important, question of statutory interpretation: Does service of a document styled as a “notice to appear” that fails to specify “the items listed” in §1229(a)(1) trigger the stop-time rule?[4] Pet. for Cert. i. As a threshold matter, the Court notes that the question presented by Pereira, which focuses on all “items listed” in §1229(a)(1), sweeps more broadly than necessary to resolve the particular case before us. Although the time-and-place information in a notice to appear will vary from case to case, the Government acknowledges that “[m]uch of the information Section 1229(a)(1) calls for does not” change and is therefore “included in standardized language on the I–862 notice-to-appear form.” Brief for Respondent 36 (referencing 8 U. S. C. §§1229(a)(1)(A)–(B), (E)–(F), and (G)(ii)). In fact, the Government’s 2006 notice to Pereira included all of the information required by §1229(a)(1), except it failed to specify the date and time of Pereira’s removal proceedings. See App. 10–12. Accordingly, the dispositive question in this case is much narrower, but no less vital: Does a “notice to appear” that does not specify the “time and place at which the proceedings will be held,” as required by §1229(a)(1)(G)(i), trigger the stop-time rule?[5] In addressing that narrower question, the Court need not resort to Chevron deference, as some lower courts have done, for Congress has supplied a clear and unambiguous answer to the interpretive question at hand. See 467 U. S., at 842–843 (“If the intent of Congress is clear, that is the end of the matter; for the court, as well as the agency, must give effect to the unambiguously expressed intent of Congress”). A putative notice to appear that fails to designate the specific time or place of the noncitizen’s removal proceedings is not a “notice to appear under section 1229(a),” and so does not trigger the stop-time rule. B The statutory text alone is enough to resolve this case. Under the stop-time rule, “any period of . . . continuous physical presence” is “deemed to end . . . when the alien is served a notice to appear under section 1229(a).” 8 U. S. C. §1229b(d)(1). By expressly referencing §1229(a), the statute specifies where to look to find out what “notice to appear” means. Section 1229(a), in turn, clarifies that the type of notice “referred to as a ‘notice to appear’ ” throughout the statutory section is a “written notice . . . specifying,” as relevant here, “[t]he time and place at which the [removal] proceedings will be held.” §1229(a)(1)(G)(i). Thus, based on the plain text of the statute, it is clear that to trigger the stop-time rule, the Government must serve a notice to appear that, at the very least, “specif[ies]” the “time and place” of the removal proceedings. It is true, as the Government and dissent point out, that the stop-time rule makes broad reference to a notice to appear under “section 1229(a),” which includes paragraph (1), as well as paragraphs (2) and (3). See Brief for Respondent 27–28; post, at 5–6 (opinion of Alito, J.). But the broad reference to §1229(a) is of no consequence, because, as even the Government concedes, only paragraph (1) bears on the meaning of a “notice to appear.” Brief for Respondent 27. By contrast, paragraph (2) governs the “[n]otice of change in time or place of proceedings,” and paragraph (3) provides for a system to record noncitizens’ addresses and phone numbers. Nowhere else within §1229(a) does the statute purport to delineate the requirements of a “notice to appear.” In fact, the term “notice to appear” appears only in paragraph (1) of §1229(a). If anything, paragraph (2) of §1229(a) actually bolsters the Court’s interpretation of the statute. Paragraph (2) provides that, “in the case of any change or postponement in the time and place of [removal] proceedings,” the Government shall give the noncitizen “written notice . . . specifying . . . the new time or place of the proceedings.” §1229(a)(2)(A)(i). By allowing for a “change or postponement” of the proceedings to a “new time or place,” paragraph (2) presumes that the Government has already served a “notice to appear under section 1229(a)” that specified a time and place as required by §1229(a)(1)(G)(i). Otherwise, there would be no time or place to “change or postpon[e ].” §1229(a)(2). Notably, the dissent concedes that paragraph (2) confirms that a notice to appear must “state the ‘time and place’ of the removal proceeding as required by §1229(a)(1).’ ” Post, at 13. The dissent nevertheless retorts that this point is “entirely irrelevant.” Ibid. Not so. Paragraph (2) clearly reinforces the conclusion that “a notice to appear under section 1229(a),” §1229b(d)(1), must include at least the time and place of the removal proceedings to trigger the stop-time rule. Another neighboring statutory provision lends further contextual support for the view that a “notice to appear” must include the time and place of the removal proceedings to trigger the stop-time rule. Section 1229(b)(1) gives a noncitizen “the opportunity to secure counsel before the first [removal] hearing date” by mandating that such “hearing date shall not be scheduled earlier than 10 days after the service of the notice to appear.” For §1229(b)(1) to have any meaning, the “notice to appear” must specify the time and place that the noncitizen, and his counsel, must appear at the removal hearing. Otherwise, the Government could serve a document labeled “notice to appear” without listing the time and location of the hearing and then, years down the line, provide that information a day before the removal hearing when it becomes available. Under that view of the statute, a noncitizen theoretically would have had the “opportunity to secure counsel,” but that opportunity will not be meaningful if, given the absence of a specified time and place, the noncitizen has minimal time and incentive to plan accordingly, and his counsel, in turn, receives limited notice and time to prepare adequately. It therefore follows that, if a “notice to appear” for purposes of §1229(b)(1) must include the time-and-place information, a “notice to appear” for purposes of the stop-time rule under §1229b(d)(1) must as well. After all, “it is a normal rule of statutory construction that identical words used in different parts of the same act are intended to have the same meaning.” Taniguchi v. Kan Pacific Saipan, Ltd., 566 U. S. 560, 571 (2012) (internal quotation marks omitted).[6] Finally, common sense compels the conclusion that a notice that does not specify when and where to appear for a removal proceeding is not a “notice to appear” that triggers the stop-time rule. If the three words “notice to appear” mean anything in this context, they must mean that, at a minimum, the Government has to provide noncitizens “notice” of the information, i.e., the “time” and “place,” that would enable them “to appear” at the removal hearing in the first place. Conveying such time-and-place information to a noncitizen is an essential function of a notice to appear, for without it, the Government cannot reason- ably expect the noncitizen to appear for his removal proceedings. To hold otherwise would empower the Government to trigger the stop-time rule merely by sending noncitizens a barebones document labeled “Notice to Appear,” with no mention of the time and place of the removal proceedings, even though such documents would do little if anything to facilitate appearance at those proceedings.[7] “ ‘We are not willing to impute to Congress . . . such [a] contradictory and absurd purpose,’ ” United States v. Bryan, 339 U. S. 323, 342 (1950), particularly where doing so has no basis in the statutory text. III Straining to inject ambiguity into the statute, the Government and the dissent advance several overlapping arguments. None is persuasive. A First, the Government posits that §1229(a) “is not worded in the form of a definition” and thus cannot circum- scribe what type of notice counts as a “notice to appear” for purposes of the stop-time rule. Brief for Respondent 32. Section 1229(a), however, does speak in definitional terms, at least with respect to the “time and place at which the proceedings will be held”: It specifically provides that the notice described under paragraph (1) is “referred to as a ‘notice to appear,’ ” which in context is quintessential definitional language.[8] It then defines that term as a “written notice” that, as relevant here, “specif[ies] . . . [t]he time and place at which the [removal] proceedings will be held.” §1229(a)(1)(G)(i). Thus, when the term “notice to appear” is used elsewhere in the statutory section, including as the trigger for the stop-time rule, it carries with it the substantive time-and-place criteria required by §1229(a). Resisting this straightforward understanding of the text, the dissent posits that “§1229(a)(1)’s language can be understood to define what makes a notice to appear complete.” Post, at 10 (emphasis in original). In the dissent’s view, a defective notice to appear is still a “notice to appear” even if it is incomplete—much like a three-wheeled Chevy is still a car. Post, at 10–11. The statutory text proves otherwise. Section 1229(a)(1) does not say a “notice to appear” is “complete” when it specifies the time and place of the removal proceedings. Rather, it defines a “notice to appear” as a “written notice” that “specif[ies],” at a minimum, the time and place of the removal proceedings. §1229(a)(1)(G)(i). Moreover, the omission of time-and-place information is not, as the dissent asserts, some trivial, ministerial defect, akin to an unsigned notice of appeal. Cf. Becker v. Montgomery, 532 U. S. 757, 763, 768 (2001). Failing to specify integral information like the time and place of removal proceedings unquestionably would “deprive [the notice to appear] of its essential character.” Post, at 12, n. 5; see supra, at 12–13, n. 7.[9] B The Government and the dissent next contend that Congress’ use of the word “under” in the stop-time rule renders the statute ambiguous. Brief for Respondent 22–23; post, at 4–5. Recall that the stop-time rule provides that “any period of . . . continuous physical presence” is “deemed to end . . . when the alien is served a notice to appear under section 1229(a).” §1229b(d)(1)(A). According to the Government, the word “under” in that provision means “subject to,” “governed by,” or “issued under the authority of.” Brief for Respondent 24. The dissent offers yet another alternative, insisting that “under” can also mean “authorized by.” Post, at 4. Those definitions, the Government and dissent maintain, support the BIA’s view that the stop-time rule applies so long as DHS serves a notice that is “authorized by,” or “subject to or governed by, or issued under the authority of” §1229(a), even if the notice bears none of the time-and-place information required by that provision. See Brief for Respondent 24; post, at 4–5. We disagree. It is, of course, true that “[t]he word ‘under’ is [a] chameleon ” that “ ‘must draw its meaning from its context.’ ” Kucana v. Holder, 558 U. S. 233, 245 (2010) (quoting Ardestani v. INS, 502 U. S. 129, 135 (1991)). But nothing in the text or context here supports either the Government’s or the dissent’s preferred definition of “under.” Based on the plain language and statutory context discussed above, we think it obvious that the word “un- der,” as used in the stop-time rule, can only mean “in accordance with” or “according to,” for it connects the stop-time trigger in §1229b(d)(1) to a “notice to appear” that contains the enumerated time-and-place information described in §1229(a)(1)(G)(i). See 18 Oxford English Dictionary 950 (2d ed. 1989) (defining “under” as “[i]n accordance with”); Black’s Law Dictionary 1525 (6th ed. 1990) (defining “under” as “according to”). So construed, the stop-time rule applies only if the Government serves a “notice to appear” “[i]n accordance with” or “according to” the substantive time-and-place requirements set forth in §1229(a). See Kirtsaeng v. John Wiley & Sons, Inc., 568 U. S. 519, 530 (2013) (internal quotation marks omitted). Far from generating any “degree of ambiguity,” post, at 4, the word “under” provides the glue that bonds the stop-time rule to the substantive time-and-place requirements mandated by §1229(a). C The Government argues that surrounding statutory provisions reinforce its preferred reading. See Brief for Respondent 25–27. It points, for instance, to two separate provisions relating to in absentia removal orders: §1229a(b)(5)(A), which provides that a noncitizen may be removed in absentia if the Government has provided “written notice required under paragraph (1) or (2) of section 1229(a)”; and §1229a(b)(5)(C)(ii), which provides that, once an in absentia removal order has been entered, the noncitizen may seek to reopen the proceeding if, inter alia, he “demonstrates that [he] did not receive notice in accordance with paragraph (1) or (2) of section 1229(a).” According to the Government, those two provisions use the distinct phrases “required under” and “in accordance with” as shorthand for a notice that satisfies §1229(a)(1)’s requirements, whereas the stop-time rule uses the phrase “under section 1229(a)” to encompass a different type of notice that does not necessarily include the information outlined in §1229(a)(1). See Brief for Respondent 25–26. That logic is unsound. The Government essentially argues that phrase 1 (“written notice required under paragraph (1) . . . of section 1229(a)”) and phrase 2 (“notice in accordance with paragraph (1) . . . of section 1229(a)”) can refer to the same type of notice even though they use entirely different words, but that phrase 3 (“notice to appear under section 1229(a)”) cannot refer to that same type of notice because it uses words different from phrases 1 and 2. But the Government offers no convincing reason why that is so. The far simpler explanation, and the one that comports with the actual statutory language and context, is that each of these three phrases refers to notice satisfying, at a minimum, the time-and-place criteria defined in §1229(a)(1). Equally unavailing is the Government’s invocation of §1229a(b)(7). Brief for Respondent 26–27. Under that provision, a noncitizen who is ordered removed in absentia is ineligible for various forms of discretionary relief for a 10-year period if the noncitizen, “at the time of the notice described in paragraph (1) or (2) of section 1229(a) of [Title 8], was provided oral notice . . . of the time and place of the proceedings” and “of the consequences . . . of failing, other than because of exceptional circumstances,” to appear. §1229a(b)(7). The Government argues that the express reference to “the time and place of the proceedings” in §1229a(b)(7) shows that, when Congress wants to attach substantive significance to whether a noncitizen is given information about the specific “time and place” of a removal proceeding, it knows exactly how to do so. Brief for Respondent 26–27. But even if §1229a(b)(7) may impose harsher consequences on noncitizens who fail to appear at removal proceedings after having specifically received oral notice of the time and place of such proceedings, that reveals nothing about the distinct question here—i.e., whether Congress intended the stop-time rule to apply when the Government fails to provide written notice of the time and place of removal proceedings. As to that question, the statute makes clear that Congress fully intended to attach substantive significance to the requirement that noncitizens be given notice of at least the time and place of their removal proceedings. A document that fails to include such information is not a “notice to appear under section 1229(a)” and thus does not trigger the stop-time rule. D Unable to find sure footing in the statutory text, the Government and the dissent pivot away from the plain language and raise a number of practical concerns. These practical considerations are meritless and do not justify departing from the statute’s clear text. See Burrage v. United States, 571 U. S. 204, 218 (2014). The Government, for its part, argues that the “administrative realities of removal proceedings” render it difficult to guarantee each noncitizen a specific time, date, and place for his removal proceedings. See Brief for Respondent 48. That contention rests on the misguided premise that the time-and-place information specified in the notice to appear must be etched in stone. That is incorrect. As noted above, §1229(a)(2) expressly vests the Government with power to change the time or place of a noncitizen’s removal proceedings so long as it provides “written notice . . . specifying . . . the new time or place of the proceedings” and the consequences of failing to appear. See §1229(a)(2); Tr. of Oral Arg. 16–19. Nothing in our decision today inhibits the Government’s ability to exercise that statu- tory authority after it has served a notice to appear specify- ing the time and place of the removal proceedings. The dissent raises a similar practical concern, which is similarly misplaced. The dissent worries that requiring the Government to specify the time and place of removal proceedings, while allowing the Government to change that information, might encourage DHS to provide “arbitrary dates and times that are likely to confuse and confound all who receive them.” Post, at 8. The dissent’s argument wrongly assumes that the Government is ut- terly incapable of specifying an accurate date and time on a notice to appear and will instead engage in “arbitrary” behavior. See ibid. The Court does not embrace those unsupported assumptions. As the Government concedes, “a scheduling system previously enabled DHS and the immigration court to coordinate in setting hearing dates in some cases.” Brief for Respondent 50, n. 15; Brief for National Immigrant Justice Center as Amicus Curiae 30–31. Given today’s advanced software capabilities, it is hard to imagine why DHS and immigration courts could not again work together to schedule hearings before sending notices to appear. Finally, the dissent’s related contention that including a changeable date would “mislead” and “prejudice” noncitizens is unfounded. Post, at 8. As already explained, if the Government changes the date of the removal proceedings, it must provide written notice to the noncitizen, §1229(a)(2). This notice requirement mitigates any potential confusion that may arise from altering the hearing date. In reality, it is the dissent’s interpretation of the statute that would “confuse and confound” noncitizens, post, at 8, by authorizing the Government to serve notices that lack any information about the time and place of the removal proceedings. E In a last ditch effort to salvage its atextual interpretation, the Government invokes the alleged purpose and legislative history of the stop-time rule. Brief for Respondent 37–40. Even for those who consider statutory purpose and legislative history, however, neither supports the Government’s atextual position that Congress intended the stop-time rule to apply when a noncitizen has been deprived notice of the time and place of his removal proceedings. By the Government’s own account, Congress enacted the stop-time rule to prevent noncitizens from exploiting administrative delays to “buy time” during which they accumulate periods of continuous presence. Id., at 37–38 (citing H. R. Rep. No. 104–469, pt. 1, p. 122 (1996)). Requiring the Government to furnish time-and-place information in a notice to appear, however, is en- tirely consistent with that objective because, once a proper notice to appear is served, the stop-time rule is triggered, and a noncitizen would be unable to manipulate or delay removal proceedings to “buy time.” At the end of the day, given the clarity of the plain language, we “apply the statute as it is written.” Burrage, 571 U. S., at 218. IV For the foregoing reasons, the judgment of the Court of Appeals for the First Circuit is reversed, and the case is remanded for further proceedings consistent with this opinion. It is so ordered. Notes 1 The Court uses the term “noncitizen” throughout this opinion to refer to any person who is not a citizen or national of the United States. See 8 U. S. C. §1101(a)(3). 2 Lawful permanent residents also may be eligible for cancellation of removal if, inter alia, they have continuously resided in the United States for at least seven years. §1229b(a)(2). 3 The period of continuous physical presence also stops if and when “the alien has committed” certain enumerated offenses that would constitute grounds for removal or inadmissibility. §1229b(d)(1)(B). That provision is not at issue here. 4 Compare Orozco-Velasquez v. Attorney General United States, 817 F. 3d 78, 83–84 (CA3 2016) (holding that the stop-time rule unambiguously requires service of a “notice to appear” that meets §1229(a)(1)’s requirements), with Moscoso-Castellanos v. Lynch, 803 F. 3d 1079, 1083 (CA9 2015) (finding the statute ambiguous and deferring to the BIA’s interpretation); O’Garro v. United States Atty. Gen., 605 Fed. Appx. 951, 953 (CA11 2015) (per curiam) (same); Guaman-Yuqui v. Lynch, 786 F. 3d 235, 239–240 (CA2 2015) (per curiam) (same); Gonzalez-Garcia v. Holder, 770 F. 3d 431, 434–435 (CA6 2014) (same); Yi Di Wang v. Holder, 759 F. 3d 670, 674–675 (CA7 2014) (same); Urbina v. Holder, 745 F. 3d 736, 740 (CA4 2014) (same). 5 The Court leaves for another day whether a putative notice to appear that omits any of the other categories of information enumerated in §1229(a)(1) triggers the stop-time rule. Contrary to the dissent’s assertion, this exercise of judicial restraint is by no means “tantamount to admitting” that the Government’s (and dissent’s) atextual interpre- tation is a permissible construction of the statute. Post, at 10 (opinion of Alito, J.). 6 The dissent argues that, if a notice to appear must furnish time-and-place information, the Government “may be forced by the Court’s interpretation to guess that the hearing will take place far in the future, only to learn shortly afterwards that the hearing is in fact imminent.” Post, at 14. In such a scenario, the dissent hypothesizes, a noncitizen would be “lulled into a false sense of security” and thus would have little meaningful opportunity to secure counsel and prepare adequately. Ibid. But nothing in our interpretation of the statute “force[s]” the Government to guess when and where a hearing will take place, ibid., nor does our interpretation prevent DHS and the Immigration Courts from working together to streamline the scheduling of removal proceedings, see infra, at 18–19. Far from “lull[ing]” noncitizens into a false sense of security, post, at 14, our reading (unlike the Government’s and the dissent’s) still gives meaning to a noncitizen’s “opportunity to secure counsel before the first [removal] hearing date,” §1229(b)(1), by informing the noncitizen that the Government is committed to moving forward with removal proceedings at a specific time and place. Equipped with that knowledge, a noncitizen has an incentive to obtain counsel and prepare for his hearing. 7 At oral argument, the Government conceded that a blank piece of paper would not suffice to trigger the stop-time rule because (in its view) such a hypothetical notice would fail to specify the charges against the noncitizen. Tr. of Oral Arg. 39–40 (arguing that notice to appear must “tell the alien what proceedings he must appear for and why he must appear for them”). The dissent also endorses the view that a notice to appear “can also be understood to serve primarily as a charging document.” Post, at 14–15. But neither the Government nor the dissent offers any convincing basis, much less one rooted in the statutory text, for treating time-and-place information as any less crucial than charging information for purposes of triggering the stop-time rule. Furthermore, there is no reason why a notice to appear should have only one essential function. Even if a notice to appear functions as a “charging document,” that is not mutually exclusive with the conclusion that a notice to appear serves another equally integral function: telling a noncitizen when and where to appear. At bottom, the Government’s self-serving position that a notice to appear must specify charging information, but not the time-and-place information, reveals the arbitrariness inherent in its atextual approach to the stop-time rule. 8 Congress has employed similar definitional language in other statutory schemes. See, e.g., 21 U. S. C. §356(b)(1) (creating new class of “fast track product[s]” by setting out drug requirements and providing: “In this section, such a drug is referred to as a ‘fast track product’ ”); §356(a)(1) (“In this section, such a drug is referred to as a ‘breakthrough therapy’ ”); 38 U. S. C. §7451(a)(2) (“hereinafter in this section referred to as ‘covered positions’ ”); 42 U. S. C. §285g–4(b) (“hereafter in this section referred to as ‘medical rehabilitation’ ”). 9 The dissent maintains that Congress’ decision to make the stop-time rule retroactive to certain pre-IIRIRA “orders to show cause” “sheds considerable light on the question presented” because orders to show cause did not necessarily include time-and-place information. Post, at 6–7. That argument compares apples to oranges. Even if the stop-time rule sometimes applies retroactively to an order to show cause, that provides scant support for the dissent’s view that, under the new post-IIRIRA statutory regime, an entirely different document called a “notice to appear,” which, by statute, must specify the time and place of removal proceedings, see §1229(a)(1)(G)(i), need not include such information to trigger the stop-time rule.

585.US.2017_16-9493

Each year, district courts sentence thousands of individuals to imprisonment for violations of federal law. To help ensure certainty and fairness in those sentences, federal district courts are required to consider the advisory United States Sentencing Guidelines. Prior to sentencing, the United States Probation Office prepares a presentence investigation report to help the court determine the applicable Guidelines range. Ultimately, the district court is responsible for ensuring the Guidelines range it considers is correct. At times, however, an error in the calculation of the Guidelines range goes unnoticed by the court and the parties. On appeal, such errors not raised in the district court may be remedied under Federal Rule of Criminal Procedure 52(b), provided that, as established in United States v. Olano, 507 U. S. 725: (1) the error was not “intentionally relinquished or abandoned,” (2) the error is plain, and (3) the error “affected the defendant’s substantial rights,” Molina-Martinez v. United States, 578 U. S. ___, ___. If those conditions are met, “the court of appeals should exercise its discretion to correct the forfeited error if the error ‘ “seriously affects the fairness, integrity or public reputation of judicial proceedings.” ’ ” Id., at ___. This last consideration is often called Olano’s fourth prong. The issue here is when a Guidelines error that satisfies Olano’s first three conditions warrants relief under the fourth prong. Petitioner Florencio Rosales-Mireles pleaded guilty to illegal reentry into the United States. In calculating the Guidelines range, the Probation Office’s presentence report mistakenly counted a state misdemeanor conviction twice. As a result, the report yielded a Guidelines range of 77 to 96 months, when the correctly calculated range would have been 70 to 87 months. Rosales-Mireles did not object to the error in the District Court, which relied on the miscalculated Guidelines range and sentenced him to 78 months of imprisonment. On appeal, Rosales-Mireles challenged the incorrect Guidelines range for the first time. The Fifth Circuit found that the Guidelines error was plain and that it affected Rosales-Mireles’ substantial rights because there was a “reasonable probability that he would have been subject to a different sentence but for the error.” The Fifth Circuit nevertheless declined to remand the case for resentencing, concluding that Rosales-Mireles had not established that the error would seriously affect the fairness, integrity, or public reputation of judicial proceedings because neither the error nor the resulting sentence “would shock the conscience.” Held: A miscalculation of a Guidelines sentencing range that has been determined to be plain and to affect a defendant’s substantial rights calls for a court of appeals to exercise its discretion under Rule 52(b) to vacate the defendant’s sentence in the ordinary case. Pp. 6–15. (a) Although “Rule 52(b) is permissive, not mandatory,” Olano, 507 U. S., at 735, it is well established that courts “should” correct a forfeited plain error affecting substantial rights “if the error ‘seriously affect[s] the fairness, integrity or public reputation of judicial proceedings,’ ” id., at 736. Like the narrow rule rejected in Olano, which would have called for relief only for a miscarriage of justice, the Fifth Circuit’s shock-the-conscience standard too narrowly confines the extent of the court of appeals’ discretion. It is not reflected in Rule 52(b), nor in how the plain-error doctrine has been applied by this Court, which has reversed judgments for plain error based on inadvertent or unintentional errors by the court or the parties below and has remanded cases involving such errors, including sentencing errors, for consideration of Olano’s fourth prong. The errors are not required to amount to a “powerful indictment” of the system. The Fifth Circuit’s emphasis on the district judge’s “competence or integrity” also unnecessarily narrows Olano’s instruction to correct an error if it seriously affects “judicial proceedings.” Pp. 6–8. (b) The effect of the Fifth Circuit’s heightened standard is especially pronounced in cases like this one. An error resulting in a higher range than the Guidelines provide usually establishes a reasonable probability that a defendant will serve a prison sentence greater than “necessary” to fulfill the purposes of incarceration, 18 U. S. C. §3553(a). See Molina-Martinez, 578 U. S., at ___. That risk of unnecessary deprivation of liberty particularly undermines the fairness, integrity, or public reputation of judicial proceedings in the context of a plain Guidelines error because Guidelines miscalculations ultimately result from judicial error, as the district court is charged in the first instance with ensuring the Guidelines range it considers is correct. Moreover, remands for resentencing are relatively inexpensive proceedings compared to remands for retrial. Ensuring the accuracy of Guidelines determinations also furthers the Sentencing Commission’s goal of achieving uniformity and proportionality in sentencing more broadly, since including uncorrected sentences based on incorrect Guidelines ranges in the data the Commission collects could undermine the Commission’s ability to make appropriate revisions to the Guidelines. Because any exercise of discretion at the fourth prong of Olano inherently requires “a case-specific and fact-intensive” inquiry, Puckett v. United States, 556 U. S. 129, 142, countervailing factors may satisfy the court of appeals that the fairness, integrity, and public reputation of the proceedings will be preserved absent correction. But there are no such factors in this case. Pp. 8–11. (c) The Government and dissent maintain that even though the Fifth Circuit’s standard was inaccurate, Rosales-Mireles is still not entitled to relief. But their arguments are unpersuasive. They caution that granting this type of relief would be inconsistent with the Court’s statements that discretion under Rule 52(b) should be exercised “sparingly,” Jones v. United States, 527 U. S. 373, 389, and reserved for “exceptional circumstances,” Meyer v. Kenmore Granville Hotel Co., 297 U. S. 160. In contrast to the Jones remand, however, no additional jury proceedings would be required in a remand for resentencing based on a Guidelines miscalculation. Plus, the circumstances of Rosales-Mireles’ case are exceptional under this Court’s precedent, as they are reasonably likely to have resulted in a longer prison sentence than necessary and there are no countervailing factors that otherwise further the fairness, integrity, or public reputation of judicial proceedings. The Government and dissent also assert that Rosales-Mireles’ sentence is presumptively reasonable because it falls within the corrected Guidelines range. But a court of appeals can consider a sentence’s substantive reasonableness only after it ensures “that the district court committed no significant procedural error, such as failing to calculate (or improperly calculating) the Guidelines range.” Gall v. United States, 552 U. S. 38, 51. If a district court cannot properly determine whether, considering all sentencing factors, including the correct Guidelines range, a sentence is “sufficient, but not greater than necessary,” 18 U. S. C. §3553(a), the resulting sentence would not bear the reliability that would support a “presumption of reasonableness” on review. See 552 U. S., at 51. And regardless of its ultimate reasonableness, a sentence that lacks reliability because of unjust procedures may well undermine public perception of the proceedings. Finally, the Government and dissent maintain that the Court’s decision will create an opportunity for “sandbagging” that Rule 52(b) is supposed to prevent. But that concern fails to account for the realities at play in sentencing proceedings, where it is highly speculative that a defendant would benefit from a strategy of deliberately forgoing an objection in the district court, with hopes of arguing for reversal under plain-error review later. Pp. 12–14. 850 F. 3d 246, reversed and remanded. Sotomayor, J., delivered the opinion of the Court, in which Roberts, C. J., and Kennedy, Ginsburg, Breyer, Kagan, and Gorsuch, JJ., joined. Thomas, J., filed a dissenting opinion, in which Alito, J., joined.

Federal Rule of Criminal Procedure 52(b) provides that a court of appeals may consider errors that are plain and affect substantial rights, even though they are raised for the first time on appeal. This case concerns the bounds of that discretion, and whether a miscalculation of the United States Sentencing Guidelines range, that has been determined to be plain and to affect a defendant’s substantial rights, calls for a court of appeals to exercise its discretion under Rule 52(b) to vacate the defendant’s sentence. The Court holds that such an error will in the ordinary case, as here, seriously affect the fairness, integrity, or public reputation of judicial proceedings, and thus will warrant relief. I A Each year, thousands of individuals are sentenced to terms of imprisonment for violations of federal law. District courts must determine in each case what constitutes a sentence that is “sufficient, but not greater than necessary,” 18 U. S. C. §3553(a), to achieve the overarching sentencing purposes of “retribution, deterrence, incapacitation, and rehabilitation.” Tapia v. United States, 564 U. S. 319, 325 (2011); 18 U. S. C. §§3551(a), 3553(a)(2). Those decisions call for the district court to exercise discretion. Yet, to ensure “ ‘certainty and fairness’ ” in sentencing, district courts must operate within the framework established by Congress. United States v. Booker, 543 U. S. 220, 264 (2005) (quoting 28 U. S. C. §991(b)(1)(B)). The Sentencing Guidelines serve an important role in that framework. “ ‘[D]istrict courts must begin their analysis with the Guidelines and remain cognizant of them throughout the sentencing process.’ ” Peugh v. United States, 569 U. S. 530, 541 (2013) (quoting Gall v. United States, 552 U. S. 38, 50, n. 6 (2007); emphasis in original). Courts are not bound by the Guidelines, but even in an advisory capacity the Guidelines serve as “a meaningful benchmark” in the initial determination of a sentence and “through the process of appellate review.” 569 U. S., at 541. Of course, to consult the applicable Guidelines range, a district court must first determine what that range is. This can be a “complex” undertaking. Molina-Martinez v. United States, 578 U. S. ___, ___ (2016) (slip op., at 4). The United States Probation Office, operating as an arm of the district court, first creates a presentence investigation report, “which includes a calculation of the advisory Guidelines range it considers to be applicable.” Id., at ___ (slip op., at 3); see Fed. Rules Crim. Proc. 32(c)(1)(A), (d)(1); United States Sentencing Commission, Guidelines Manual §1B1.1(a) (Nov. 2016) (USSG). That calculation derives from an assessment of the “offense characteristics, offender characteristics, and other matters that might be relevant to the sentence.” Rita v. United States, 551 U.S. 338, 342 (2007) (internal quotation marks omitted). Specifically, an offense level is calculated by identifying a base level for the offense of conviction and adjusting that level to account for circumstances specific to the defendant’s case, such as how the crime was committed and whether the defendant accepted responsibility. See USSG §§1B1.1(a)(1)–(5). A numerical value is then attributed to any prior offenses committed by the defendant, which are added together to generate a criminal history score that places the defendant within a particular criminal history category. §§1B1.1(a)(6), 4A1.1. Together, the offense level and the criminal history category identify the applicable Guidelines range. §1B1.1(a)(7). B The district court has the ultimate responsibility to ensure that the Guidelines range it considers is correct, and the “[f]ailure to calculate the correct Guidelines range constitutes procedural error.” Peugh, 569 U. S., at 537. Given the complexity of the calculation, however, district courts sometimes make mistakes. It is unsurprising, then, that “there will be instances when a district court’s sentencing of a defendant within the framework of an incorrect Guidelines range goes unnoticed” by the parties as well, which may result in a defendant raising the error for the first time on appeal. Molina-Martinez, 578 U. S., at ___ (slip op., at 4). Those defendants are not entirely without recourse. Federal Rule of Criminal Procedure 52(b) provides that “[a] plain error that affects substantial rights may be considered even though it was not brought to the [district] court’s attention.” In United States v. Olano, 507 U. S. 725 (1993), the Court established three conditions that must be met before a court may consider exercising its discretion to correct the error. “First, there must be an error that has not been intentionally relinquished or abandoned. Second, the error must be plain—that is to say, clear or obvious. Third, the error must have affected the defendant’s substantial rights.” Molina-Martinez, 578 U. S., at ___ (slip op., at 4) (citations omitted). To satisfy this third condition, the defendant ordinarily must “ ‘show a reasonable probability that, but for the error,’ the outcome of the proceeding would have been different.” Ibid. (quoting United States v. Dominguez Benitez, 542 U. S. 74, 76, 82 (2004)). Once those three conditions have been met, “the court of appeals should exercise its discretion to correct the forfeited error if the error seriously affects the fairness, integrity or public reputation of judicial proceedings.” Molina-Martinez, 578 U. S., at ___ (slip op., at 4–5) (internal quotation marks omitted). It is this last consideration, often called Olano’s fourth prong, that we are asked to clarify and apply in this case. C Petitioner Florencio Rosales-Mireles pleaded guilty to illegal reentry in violation of 8 U. S. C. §§1326(a), (b)(2). The Probation Office in its presentence investigation report mistakenly counted a 2009 state conviction of misdemeanor assault twice. This double counting resulted in a criminal history score of 13, which placed Rosales-Mireles in criminal history category VI. Combined with his offense level of 21, that yielded a Guidelines range of 77 to 96 months. Had the criminal history score been calculated correctly, Rosales-Mireles would have been in criminal history category V, and the resulting Guidelines range would have been 70 to 87 months. See USSG ch. 5, pt. A (sentencing table). Rosales-Mireles did not object to the double-counting error before the District Court. Relying on the erroneous presentence investigation report, and after denying Rosales-Mireles’ request for a downward departure, the District Court sentenced Rosales-Mireles to 78 months of imprisonment, one month above the lower end of the Guidelines range that everyone thought applied. On appeal, Rosales-Mireles argued for the first time that his criminal history score and the resulting Guidelines range were incorrect because of the double counting of his 2009 conviction. Because he had not objected in the District Court, the Court of Appeals for the Fifth Circuit reviewed for plain error. 850 F. 3d 246, 248 (2017). Applying the Olano framework, the Fifth Circuit concluded that Rosales-Mireles had established that the Guidelines miscalculation constituted an error that was plain, satisfying Olano’s first two conditions. It also held that the error affected Rosales-Mireles’ substantial rights, thus satisfying the third condition, because there was “a reasonable probability that he would have been subject to a different sentence but for the error.” 850 F. 3d, at 249. In reaching that conclusion, the Fifth Circuit rejected the Government’s argument that Rosales-Mireles would have received the same sentence regardless of the Guidelines error, because the District Court had denied a downward departure “based, in part, on Rosales-Mireles’ criminal history,” which “erroneously included an extra conviction.” Ibid. The Fifth Circuit nevertheless declined to exercise its discretion to vacate and remand the case for resentencing because it concluded that Rosales-Mireles failed to establish that the error would seriously affect the fairness, integrity, or public reputation of judicial proceedings. In its view, “the types of errors that warrant reversal are ones that would shock the conscience of the common man, serve as a powerful indictment against our system of justice, or seriously call into question the competence or integrity of the district judge.” Id., at 250 (internal quotation marks and alterations omitted). Because Rosales-Mireles’ sentence of 78 months fell within the correct range of 70 to 87 months, the Fifth Circuit held that neither the error nor the resulting sentence “would shock the conscience.” Ibid. The Fifth Circuit’s articulation of Olano’s fourth prong is out of step with the practice of other Circuits.[1] We granted certiorari to resolve that conflict, 582 U. S. ___ (2017), and now reverse. II A Although “Rule 52(b) is permissive, not mandatory,” Olano, 507 U. S., at 735, it is well established that courts “should” correct a forfeited plain error that affects substantial rights “if the error ‘seriously affects the fairness, integrity or public reputation of judicial proceedings.’ ” Id., at 736 (quoting United States v. Atkinson, 297 U. S. 157, 160 (1936); alteration omitted); see also Molina-Martinez, 578 U. S., at ___–___ (slip op., at 4–5). The Court in Olano rejected a narrower rule that would have called for relief only “ ‘in those circumstances in which a miscarriage of justice would otherwise result,’ ” that is to say, where a defendant is actually innocent. 507 U. S., at 736 (quoting United States v. Young, 470 U. S. 1, 15 (1985)). By focusing instead on principles of fairness, integrity, and public reputation, the Court recognized a broader category of errors that warrant correction on plain-error review. See 507 U. S., at 736–737. Like the miscarriage-of-justice rule that the Court rejected in Olano, the Fifth Circuit’s standard is unduly restrictive. To be sure, a conclusion that an error “shock[s] the conscience of the common man, serve[s] as a powerful indictment against our system of justice, or seriously call[s] into question the competence or integrity of the district judge,” 850 F. 3d, at 250 (internal quotation marks omitted), would demand an exercise of discretion to correct the error. Limiting relief only to those circumstances, however, too narrowly confines the extent of a court of appeals’ discretion. The “shock the conscience” standard typically is employed when determining whether governmental action violates due process rights under the Fifth and Fourteenth Amendments. See County of Sacramento v. Lewis, 523 U. S. 833, 847, n. 8 (1998) (“[I]n a due process challenge to executive action, the threshold question is whether the behavior of the governmental officer is so egregious, so outrageous, that it may fairly be said to shock the contemporary conscience”). This Court has said that the “shock the conscience” standard is satisfied where the conduct was “intended to injure in some way unjustifiable by any government interest,” or in some circumstances if it resulted from deliberate indifference. Id., at 849–850. That standard is not reflected in Rule 52(b) itself, nor in how this Court has applied the plain-error doctrine. The Court repeatedly has reversed judgments for plain error on the basis of inadvertent or unintentional errors of the court or the parties below. See, e.g., Silber v. United States, 370 U. S. 717, 717–718 (1962) (per curiam) (reversing judgment for plain error as a result of insufficient indictment); Brasfield v. United States, 272 U. S. 448, 449–450 (1926) (reversing judgment for plain error where the trial judge improperly inquired of a jury’s numerical division); Clyatt v. United States, 197 U. S. 207, 222 (1905) (reversing judgment for plain error where the Government presented insufficient evidence to sustain conviction). The Court also “routinely remands” cases involving inadvertent or unintentional errors, including sentencing errors, for consideration of Olano’s fourth prong with the understanding that such errors may qualify for relief. Hicks v. United States, 582 U. S. ___, ___ (2017) (Gorsuch, J., concurring) (slip op., at 2). The Fifth Circuit’s additional focus on errors that “serve as a powerful indictment against our system of justice, or seriously call into question the competence or integrity of the district judge,” 850 F. 3d, at 250 (internal quotation marks omitted), similarly alters the Rule 52(b) standard. The Court has never said that errors must amount to a “powerful indictment” of the system, a phrase which implies by its terms that the only errors worthy of correction are those that rise to the level of grossly serious misconduct. Similarly, the Fifth Circuit’s emphasis on the “competence or integrity of the district judge” narrows Olano’s instruction that an error should be corrected if it seriously affects “judicial proceedings.” In articulating such a high standard, the Fifth Circuit substantially changed Olano’s fourth prong. B The effect of the Fifth Circuit’s heightened standard is especially pronounced in a case like this one. A plain Guidelines error that affects a defendant’s substantial rights is precisely the type of error that ordinarily warrants relief under Rule 52(b). In Molina-Martinez, the Court recognized that “[w]hen a defendant is sentenced under an incorrect Guidelines range—whether or not the defendant’s ultimate sentence falls within the correct range—the error itself can, and most often will, be sufficient to show a reasonable probability of a different outcome absent the error.” 578 U. S., at ___ (slip op., at 9). In other words, an error resulting in a higher range than the Guidelines provide usually establishes a reasonable probability that a defendant will serve a prison sentence that is more than “necessary” to fulfill the purposes of incarceration. 18 U. S. C. §3553(a); Tapia, 564 U. S., at 325. “To a prisoner,” this prospect of additional “time behind bars is not some theoretical or mathematical concept.” Barber v. Thomas, 560 U. S. 474, 504 (2010) (Kennedy, J., dissenting). “[A]ny amount of actual jail time” is significant, Glover v. United States, 531 U. S. 198, 203 (2001), and “ha[s] exceptionally severe consequences for the incarcerated individual [and] for society which bears the direct and indirect costs of incarceration,” United States v. Jenkins, 854 F. 3d 181, 192 (CA2 2017). The possibility of additional jail time thus warrants serious consideration in a determination whether to exercise discretion under Rule 52(b). It is crucial in maintaining public perception of fairness and integrity in the justice system that courts exhibit regard for fundamental rights and respect for prisoners “as people.” T. Tyler, Why People Obey the Law 164 (2006). The risk of unnecessary deprivation of liberty par- ticularly undermines the fairness, integrity, or public rep- utation of judicial proceedings in the context of a plain Guidelines error because of the role the district court plays in calculating the range and the relative ease of correcting the error. Unlike “case[s] where trial strategies, in retrospect, might be criticized for leading to a harsher sentence,” Guidelines miscalculations ultimately result from judicial error. Glover, 531 U. S., at 204; see also Peugh, 569 U. S., at 537. That was especially so here where the District Court’s error in imposing Rosales-Mireles’ sentence was based on a mistake made in the presentence investigation report by the Probation Office, which works on behalf of the District Court. Moreover, “a remand for resentencing, while not costless, does not invoke the same difficulties as a remand for retrial does.” Molina-Martinez, 578 U. S., at ___ (slip op., at 15) (internal quotation marks omitted). “A resentencing is a brief event, normally taking less than a day and requiring the attendance of only the defendant, counsel, and court personnel.” United States v. Williams, 399 F. 3d 450, 456 (CA2 2005). Ensuring the accuracy of Guidelines determinations also serves the purpose of “providing certainty and fairness in sentencing” on a greater scale. 28 U. S. C. §994(f); see also §991(b)(1)(B); Booker, 543 U. S., at 264. The Guidelines assist federal courts across the country in achieving uniformity and proportionality in sentencing. See Rita, 551 U. S., at 349. To realize those goals, it is important that sentencing proceedings actually reflect the nature of the offense and criminal history of the defendant, because the United States Sentencing Commission relies on data developed during sentencing proceedings, including information in the presentence investigation report, to determine whether revisions to the Guidelines are necessary. See id., at 350. When sentences based on incorrect Guidelines ranges go uncorrected, the Commission’s ability to make appropriate amendments is undermined.[2] In broad strokes, the public legitimacy of our justice system relies on procedures that are “neutral, accurate, consistent, trustworthy, and fair,” and that “provide opportunities for error correction.” Bowers & Robinson, Perceptions of Fairness and Justice: The Shared Aims and Occasional Conflicts of Legitimacy and Moral Credibility, 47 Wake Forest L. Rev. 211, 215–216 (2012). In considering claims like Rosales-Mireles’, then, “what reasonable citizen wouldn’t bear a rightly diminished view of the judicial process and its integrity if courts refused to correct obvious errors of their own devise that threaten to require individuals to linger longer in federal prison than the law demands?” United States v. Sabillon-Umana, 772 F. 3d 1328, 1333–1334 (CA10 2014) (Gorsuch, J.). In the context of a plain Guidelines error that affects substantial rights, that diminished view of the proceedings ordinarily will satisfy Olano’s fourth prong, as it does in this case.[3] As the Fifth Circuit itself concluded, there is a reasonable probability that, without correction of the Guidelines error, Rosales-Mireles will spend more time in prison than the District Court otherwise would have considered necessary. 850 F. 3d, at 249. That error was based on a mistake by the Probation Office, a mistake that can be remedied through a relatively inexpensive resentencing proceeding. Of course, any exercise of discretion at the fourth prong of Olano inherently requires “a case-specific and fact-intensive” inquiry. Puckett v. United States, 556 U. S. 129, 142 (2009); see also Young, 470 U. S., at 16–17, n. 14. There may be instances where countervailing factors satisfy the court of appeals that the fairness, integrity, and public reputation of the proceedings will be preserved absent correction. But on the facts of this case, there are no such factors.[4] III The United States and the dissent agree with Rosales-Mireles that the Fifth Circuit’s formulation of the standard for the exercise of discretion under Rule 52(b) “is an inaccurate description” of Olano’s fourth prong. Brief for United States 34; post, at 1, n. 1 (opinion of Thomas, J.) (“[T]he Fifth Circuit’s standard is higher than the one articulated in this Court’s precedents”). They nevertheless maintain that Rosales-Mireles is not entitled to relief. We are unpersuaded, though a few points merit brief discussion. First, the United States and the dissent caution that a grant of relief in Rosales-Mireles’ case and in others like his would be inconsistent with the Court’s statements that discretion under Rule 52(b) should be exercised “sparingly,” Jones v. United States, 527 U. S. 373, 389 (1999), and reserved for “exceptional circumstances,” Atkinson, 297 U. S., at 160. As an initial matter, Jones and the cases it relies on for the point that discretion should be exercised “sparingly” would have required additional jury proceedings on remand, either at resentencing or retrial. See 527 U. S., at 384, 389; see also Young, 470 U. S. 1; United States v. Frady, 456 U. S. 152 (1982); Henderson v. Kibbe, 431 U. S. 145 (1977). As we have explained, a decision remanding a case to the district court for resentencing on the basis of a Guidelines miscalculation is far less burdensome than a retrial, or other jury proceedings, and thus does not demand such a high degree of caution. In any event, the circumstances surrounding Rosales-Mireles’ case are exceptional within the meaning of the Court’s precedent on plain-error review, as they are reasonably likely to have resulted in a longer prison sentence than necessary and there are no countervailing factors that otherwise further the fairness, integrity, or public reputation of judicial proceedings. The fact that, as a result of the Court’s holding, most defendants in Rosales-Mireles’ situation will be eligible for relief under Rule 52(b) does not justify a decision that ignores the harmful effects of allowing the error to persist. Second, the United States and the dissent assert that, because Rosales-Mireles’ sentence falls within the cor- rected Guidelines range, the sentence is presumptively rea- sonable and “less likely to indicate a serious injury to the fairness, integrity, or public reputation of judicial proceedings.” Brief for United States 20–21; see also post, at 10. A substantive reasonableness determination, however, is an entirely separate inquiry from whether an error warrants correction under plain-error review. Before a court of appeals can consider the substantive reasonableness of a sentence, “[i]t must first ensure that the district court committed no significant procedural error, such as failing to calculate (or improperly calculating) the Guidelines range.” Gall, 552 U. S., at 51. This makes eminent sense, for the district court is charged in the first instance with determining whether, taking all sentencing factors into consideration, including the correct Guidelines range, a sentence is “sufficient, but not greater than necessary.” 18 U. S. C. §3553(a). If the district court is unable properly to undertake that inquiry because of an error in the Guidelines range, the resulting sentence no longer bears the reliability that would support a “presumption of reasonableness” on review. See Gall, 552 U. S., at 51. Likewise, regardless of its ultimate reason- ableness, a sentence that lacks reliability because of unjust procedures may well undermine public perception of the proceedings. See Hollander-Blumoff, The Psychology of Procedural Justice in the Federal Courts, 63 Hastings L. J. 127, 132–134 (2011) (compilation of psychology research showing that the fairness of procedures influences perceptions of outcomes). The mere fact that Rosales-Mireles’ sentence falls within the corrected Guidelines range does not preserve the fairness, integrity, or public reputation of the proceedings.[5] Third, the United States and the dissent contend that our decision “creates the very opportunity for ‘sandbagging’ that Rule 52(b) is supposed to prevent.” Post, at 5; Brief for United States 17–18, 27. But that concern fails to account for the realities at play in sentencing proceedings. As this Court repeatedly has explained, “the Guidelines are ‘the starting point for every sentencing calculation in the federal system,’ ” Hughes v. United States, 584 U. S. ___, ___ (2018) (slip op., at 9) (quoting Peugh, 569 U. S., at 542). It is hard to imagine that defense counsel would “deliberately forgo objection now” to a plain Guidelines error that would subject her client to a higher Guidelines range, “because [counsel] perceives some slightly expanded chance to argue for ‘plain error’ later.” Henderson v. United States, 568 U. S. 266, 276 (2013) (emphasis in original). Even setting aside the conflict such a strat- egy would create with defense counsel’s ethical obligations to represent her client vigorously and her duty of candor toward the court, any benefit from such a strategy is highly speculative. There is no guarantee that a court of appeals would agree to a remand, and no basis to believe that a district court would impose a lower sentence upon resentencing than the court would have imposed at the original sentencing proceedings had it been aware of the plain Guidelines error. IV For the foregoing reasons, we conclude that the Fifth Circuit abused its discretion in applying an unduly burdensome articulation of Olano’s fourth prong and declining to remand Rosales-Mireles’ case for resentencing. In the ordinary case, as here, the failure to correct a plain Guidelines error that affects a defendant’s substantial rights will seriously affect the fairness, integrity, and public reputation of judicial proceedings. The judgment of the Court of Appeals is therefore reversed, and the case is remanded for further proceedings consistent with this opinion. It is so ordered. Notes 1 Compare 850 F. 3d 246, 250 (CA5 2017), with United States v. Dahl, 833 F. 3d 345, 357, 359 (CA3 2016); United States v. Figueroa-Ocasio, 805 F. 3d 360, 367–368, 373–374 (CA1 2015); United States v. Sabillon-Umana, 772 F. 3d 1328, 1333–1334 (CA10 2014) (Gorsuch, J.); United States v. Joseph, 716 F. 3d 1273, 1281 (CA9 2013); United States v. Garrett, 528 F. 3d 525, 527, 529–530 (CA7 2008). 2 Similarly, the work of the Federal Bureau of Prisons is hindered by uncorrected Guidelines errors, because the Bureau relies, in part, on aspects of the Guidelines calculation in designating and classifying prisoners based on security and program needs. See Federal Bureau of Prisons, Program Statement No. P5100.08, Subject: Inmate Security Designation and Custody Classification, ch. 2, p. 1, ch. 4, p. 8, ch. 6, p. 5 (Sept. 12, 2006). 3 The dissent maintains that “adhering to procedure” does not have “prime importance for purposes of the fourth prong” because the Court has held in some instances, where the error was not likely to have affected the substantive outcome, that the procedural error alone did not satisfy Olano’s fourth prong. Post, at 7 (opinion of Thomas, J.) (citing Johnson v. United States, 520 U. S. 461 (1997); United States v. Cotton, 535 U. S. 625 (2002); United States v. Marcus, 560 U. S. 258 (2010)). The cases on which the dissent relies do not stand for the view, however, that procedural errors are unimportant or could never satisfy Olano’s fourth prong, especially where, as here, the defendant has shown a likelihood that the error affected the substantive outcome. 4 As the dissent points out, post, at 8–9, a defendant bears the “burden to persuade the court that the error seriously affect[ed] the fairness, integrity or public reputation of judicial proceedings.” United States v. Vonn, 535 U. S. 55, 63 (2002) (internal quotation marks omitted). In the ordinary case, proof of a plain Guidelines error that affects the defendant’s substantial rights is sufficient to meet that burden. 5 The dissent’s discussion of Rosales-Mireles’ criminal history, post, at 9–10, misses the point. That history is relevant to the District Court’s determination of an appropriate sentence under 18 U. S. C. §3553(a). It does not help explain whether the plain procedural error in Rosales-Mireles’ sentencing proceedings, which may have resulted in a longer sentence than is justified in light of that history, seriously affects the fairness, integrity, or public reputation of judicial proceedings.

583.US.2017_16-534

The Foreign Sovereign Immunities Act of 1976 (FSIA) grants foreign states and their agencies and instrumentalities immunity from suit in the United States and grants their property immunity from attachment and execution in satisfaction of judgments against them, see 28 U. S. C. §§1604, 1609, but with some exceptions. Petitioners hold a judgment against respondent Islamic Republic of Iran pursuant to an exception that applies to foreign states designated as state sponsors of terrorism with respect to claims arising out of acts of terrorism. See §1605A. To enforce that judgment, petitioners filed an action in the District Court to attach and execute against certain Iranian assets—a collection of ancient clay tablets and fragments housed at respondent University of Chicago. The District Court concluded that §1610(g)—which provides that certain property will be “subject to attachment in aid of execution, and execution, upon [a §1605A] judgment as provided in this section”—does not deprive the collection of the immunity typically afforded the property of a foreign sovereign. The Seventh Circuit affirmed. Held: Section 1610(g) does not provide a freestanding basis for parties holding a judgment under §1605A to attach and execute against the property of a foreign state; rather, for §1610(g) to apply, the immunity of the property at issue must be rescinded under a separate provision within §1610. Pp. 4–15. (a) Congress enacted the FSIA in an effort to codify the careful balance between respecting the immunity historically afforded to foreign sovereigns and holding them accountable, in certain circumstances, for their actions. As a default, foreign states have immunity “from the jurisdiction of the courts of the United States and of the States,” §1604, but there are express exceptions, including the one at issue here, for state sponsors of terrorism, see §1605A(a). The FSIA similarly provides as a default that “the property in the United States of a foreign state shall be immune from attachment arrest and execution.” §1609. But §1610 outlines certain exceptions to this immunity. For example, §1610(a)(7) provides that property in the United States of a foreign state that is used for a commercial activity in the United States shall not be immune from attachment and execution where the plaintiff holds a §1605A judgment against the foreign state. Before 2008, the FSIA did not expressly address under which circumstances a foreign state’s agencies or instrumentalities could be held liable for judgments against the state. The Court had addressed that question in First Nat. City Bank v. Banco Para el Comercio Exterior de Cuba, 462 U. S. 611 (Bancec), and held that, as a default, agencies and instrumentalities of a foreign state are separate legal entities that cannot be held liable. It recognized the availability of exceptions, however, and left the lower courts to determine whether an exception applied on a case-by-case basis. The lower courts coalesced around five relevant factors (the Bancec factors) to assist in those determinations. In 2008, Congress amended the FSIA, adding §1610(g). Subparagraphs (A) through (E) incorporate almost verbatim the Bancec factors, leaving no dispute that, at a minimum, §1610(g) serves to abrogate Bancec where a §1605A judgment holder seeks to satisfy a judgment held against the foreign state. The question here is whether, in addition to abrogating Bancec, it provides a freestanding exception to property immunity in the context of a §1605A judgment. Pp. 4–8. (b) The most natural reading of §1610(g)(1)’s phrase “as provided in this section” is that it refers to §1610 as a whole, so that §1610(g)(1) will apply to property that is exempted from the grant of immunity as provided elsewhere in §1610. Those §1610 provisions that do unambiguously revoke the immunity of a foreign state’s property employ phrases such as “shall not be immune,” see §1610(a)(7), and “[n]otwithstanding any other provision of law,” see §1610(f)(1)(A). Such textual markers are conspicuously absent from §1610(g). Thus, its phrase “as provided in this section” is best read to signal only that a judgment holder seeking to take advantage of §1610(g)(1) must identify a basis under one of §1610’s express immunity-abrogating provisions to attach and execute against a relevant property. This reading provides relief to judgment holders who previously would not have been able to attach and execute against property of an agency or instrumentality of a foreign state in light of Bancec. It is also consistent with the basic interpretive canon to construe a statute so as to give effect to all of its provisions, see Corley v. United States, 556 U. S. 303 , and with the historical practice of rescinding attachment and execution immunity primarily in the context of a foreign state’s commercial acts, see Verlinden B. V. v. Central Bank of Nigeria, 461 U. S. 480 –488. Pp. 8–11. (c) Petitioners’ counterarguments are unpersuasive. They assert that the phrase “as provided in this section” might refer to the procedures in §1610(f)(1), which permits §1605A judgment holders to attach and execute against property associated with certain prohibited financial transactions, but which was waived by the President before it could take effect. However, it is not logical to read the phrase as indicating a congressional intent to create §1610(g) as an alternative to §1610(f)(1), particularly since Congress knows how to make clear when it is rescinding immunity. Nor could Congress have intended “as provided in this section” to refer only to §1610(f)(2)’s instruction that the Federal Government assist in identifying assets, since that provision does not provide for attachment or execution at all. Finally, there is no basis to conclude that “this section” in §1610(g) reflects a mere drafting error. The words “property of a foreign state,” which appear in the first substantive clause of §1610(g), are not rendered superfluous under the Court’s reading. Section 1610(g) serves to identify in one place all the categories of property that will be available to §1605A judgment holders for attachment and execution, and commands that the availability of such property will not be limited by the Bancec factors. Also, without the opening clause, §1610(g) would abrogate the Bancec presumption of separateness in all cases, not just those involving terrorism judgments under §1605A. Although petitioners contend that any uncertainty in §1610(g) should be resolved by giving full effect to the legislative purpose behind its enactment—removing obstacles to enforcing terrorism judgments—they offer no real support for their position that §1610(g) was intended to divest all property of a foreign state or its agencies or instrumentalities of immunity. Bank Markazi v. Peterson, 578 U. S. ___, ___, n. 2, distinguished. Pp. 12–15. 830 F. 3d 470, affirmed. Sotomayor, J., delivered the opinion of the Court, in which all other Members joined, except Kagan, J., who took no part in the consideration or decision of the case.

The Foreign Sovereign Immunities Act of 1976 (FSIA) grants foreign states and their agencies and instrumentalities immunity from suit in the United States (called jurisdictional immunity) and grants their property immunity from attachment and execution in satisfaction of judgments against them. See 28 U. S. C. §§1604, 1609. But those grants of immunity are subject to exception. Petitioners hold a judgment against respondent Islamic Republic of Iran pursuant to one such exception to jurisdictional immunity, which applies where the foreign state is designated as a state sponsor of terrorism and the claims arise out of acts of terrorism. See §1605A. The issue presented in this case is whether certain property of Iran, specifically, a collection of antiquities owned by Iran but in the possession of respondent University of Chicago, is subject to attachment and execution by petitioners in satisfaction of that judgment. Petitioners contend that the property is stripped of its immunity by another provision of the FSIA, §1610(g), which they maintain provides a blanket exception to the immunity typically afforded to the property of a foreign state where the party seeking to attach and execute holds a §1605A judgment. We disagree. Section 1610(g) serves to identify property that will be available for attachment and execution in satisfaction of a §1605A judgment, but it does not in itself divest property of immunity. Rather, the provision’s language “as provided in this section” shows that §1610(g) operates only when the property at issue is exempt from immunity as provided elsewhere in §1610. Petitioners cannot invoke §1610(g) to attach and execute against the antiquities at issue here, which petitioners have not established are exempt from immunity under any other provision in §1610. I A On September 4, 1997, Hamas carried out three suicide bombings on a crowded pedestrian mall in Jerusalem, resulting in the deaths of 5 people and injuring nearly 200 others. Petitioners are United States citizens who were either wounded in the attack or are the close relatives of those who were injured. In an attempt to recover for their harm, petitioners sued Iran in the District Court for the District of Columbia, alleging that Iran was responsible for the bombing because it provided material support and training to Hamas. At the time of that action, Iran was subject to the jurisdiction of the federal courts pursuant to 28 U. S. C. §1605(a)(7) (1994 ed., Supp. II), which rescinded the immunity of foreign states designated as state sponsors of terrorism with respect to claims arising out of acts of terrorism. Iran did not appear in the action, and the District Court entered a default judgment in favor of petitioners in the amount of $71.5 million.[1] When Iran did not pay the judgment, petitioners brought this action in the District Court for the Northern District of Illinois to attach and execute against certain Iranian assets located in the United States in satisfaction of their judgment. Those assets—a collection of approximately 30,000 clay tablets and fragments containing ancient writings, known as the Persepolis Collection—are in the possession of the University of Chicago, housed at its Oriental Institute. University archeologists recovered the artifacts during an excavation of the old city of Persepolis in the 1930’s. In 1937, Iran loaned the collection to the Oriental Institute for research, translation, and cataloging.[2] Petitioners maintained in the District Court, inter alia, that §1610(g) of the FSIA renders the Persepolis Collection subject to attachment and execution. The District Court concluded otherwise and held that §1610(g) does not deprive the Persepolis Collection of the immunity typically afforded the property of a foreign sovereign. The Court of Appeals for the Seventh Circuit affirmed. 830 F. 3d 470 (2016). As relevant, the Seventh Circuit held that the text of §1610(g) demonstrates that the provision serves to identify the property of a foreign state or its agencies or instrumentalities that are subject to attachment and execution, but it does not in itself divest that property of immunity. The Court granted certiorari to resolve a split among the Courts of Appeals regarding the effect of §1610(g).[3] 582 U. S. ___ (2017). We agree with the conclusion of the Seventh Circuit, and therefore affirm. B We start with a brief review of the historical development of foreign sovereign immunity law and the statutory framework at issue here, as it provides a helpful guide to our decision. This Court consistently has recognized that foreign sovereign immunity “is a matter of grace and comity on the part of the United States.” Verlinden B. V. v. Central Bank of Nigeria, 461 U. S. 480, 486 (1983) ; Schooner Exchange v. McFaddon, 7 Cranch 116, 136 (1812). In determining whether to exercise jurisdiction over suits against foreign sovereigns, courts traditionally “deferred to the decisions of the political branches . . . on whether to take jurisdiction over actions against foreign sovereigns.” Verlinden, 461 U. S., at 486. Prior to 1952, the State Department generally held the position that foreign states enjoyed absolute immunity from all actions in the United States. See ibid. But, as foreign states became more involved in commercial activity in the United States, the State Department recognized that such participation “makes necessary a practice which will enable persons doing business with them to have their rights determined in the courts.” J. Tate, Changed Policy Concerning the Granting of Sovereign Immunity to Foreign Governments, 26 Dept. State Bull. 984, 985 (1952). The Department began to follow the “restrictive” theory of foreign sovereign immunity in advising courts whether they should take jurisdiction in any given case. Immunity typically was afforded in cases involving a foreign sovereign’s public acts, but not in “cases arising out of a foreign state’s strictly commercial acts.” Verlinden, 461 U. S., at 487. In 1976, Congress enacted the FSIA in an effort to codify this careful balance between respecting the immunity historically afforded to foreign sovereigns and holding them accountable, in certain circumstances, for their actions. 90Stat. 2891, as amended, 28 U. S. C. §1602 et seq. “For the most part, the Act” tracks “the restrictive theory of sovereign immunity.” Verlinden, 461 U. S., at 488. As a default, foreign states enjoy immunity “from the jurisdiction of the courts of the United States and of the States.” §1604. But this immunity is subject to certain express exceptions. For example, in line with the restrictive theory, a foreign sovereign will be stripped of jurisdictional immunity when a claim is based upon commercial activity it carried out in the United States. See, e.g., §1605(a)(2). The FSIA also provides that a foreign state will be subject to suit when it is designated as a state sponsor of terrorism and damages are sought as a result of acts of terrorism. See §1605A(a). With respect to the immunity of property, the FSIA similarly provides as a default that “the property in the United States of a foreign state shall be immune from attachment arrest and execution.” §1609. But, again, there are exceptions, and §1610 outlines the circumstances under which property will not be immune. See §1610. For example, subsection (a) expressly provides that property “shall not be immune” from attachment and execution where, inter alia, it is “used for a commercial activity in the United States” and the “judgment relates to a claim for which the foreign state is not immune under section 1605A or section 1605(a)(7) (as such section was in effect on January 27, 2008), regardless of whether the property is or was involved with the act upon which the claim is based.” §1610(a)(7). Prior to 2008, the FSIA did not address expressly under what circumstances, if any, the agencies or instrumentalities of a foreign state could be held liable for judgments against the state. Faced with that question in First Nat. City Bank v. Banco Para el Comercio Exterior de Cuba, 462 U. S. 611 (1983) (Bancec), this Court held that “government instrumentalities established as juridical entities distinct and independent from their sovereign should normally be treated as such.” Id., at 626–627. Thus, as a default, those agencies and instrumentalities of a foreign state were to be considered separate legal entities that cannot be held liable for acts of the foreign state. See id., at 628. Nevertheless, the Court recognized that such a stringent rule should not be without exceptions. The Court suggested that liability would be warranted, for example, “where a corporate entity is so extensively controlled by [the state] that a relationship of principal and agent is created,” id., at 629, or where recognizing the state and its agency or instrumentality as distinct entities “would work fraud or injustice,” ibid. (internal quotation marks omitted). See id., at 630. But the Court declined to develop a “mechanical formula for determining” when these exceptions should apply, id., at 633, leaving lower courts with the task of assessing the availability of exceptions on a case-by-case basis. Over time, the Courts of Appeals coalesced around the following five factors (referred to as the Bancec factors) to aid in this analysis: “(1) the level of economic control by the government; “(2) whether the entity’s profits go to the government; “(3) the degree to which government officials manage the entity or otherwise have a hand in its daily affairs; “(4) whether the government is the real beneficiary of the entity’s conduct; and “(5) whether adherence to separate identities would entitle the foreign state to benefits in United States courts while avoiding its obligations.” Walter Fuller Aircraft Sales, Inc. v. Republic of Philippines, 965 F. 2d 1375, 1380, n. 7 (CA5 1992); see also Flatow v. Islamic Republic of Iran, 308 F. 3d 1065, 1071, n. 9 (CA9 2002). In 2008, Congress amended the FSIA and added §1610(g). See NDAA §1083(b)(3)(D), 122Stat. 341–342. Section 1610(g)(1) provides: “(g) Property in Certain Actions.— “(1) In general. [T]he property of a foreign state against which a judgment is entered under section 1605A, and the property of an agency or instrumentality of such a state, including property that is a separate juridical entity or is an interest held directly or indirectly in a separate juridical entity, is subject to attachment in aid of execution, and execution, upon that judgment as provided in this section, regardless of— “(A) the level of economic control over the property by the government of the foreign state; “(B) whether the profits of the property go to that government; “(C) the degree to which officials of that government manage the property or otherwise control its daily affairs; “(D) whether that government is the sole beneficiary in interest of the property; or “(E) whether establishing the property as a separate entity would entitle the foreign state to benefits in United States courts while avoiding its obligations.” Subparagraphs (A) through (E) incorporate almost verbatim the five Bancec factors, leaving no dispute that, at a minimum, §1610(g) serves to abrogate Bancec with respect to the liability of agencies and instrumentalities of a foreign state where a §1605A judgment holder seeks to satisfy a judgment held against the foreign state. The issue at hand is whether §1610(g) does something more; whether, like the commercial activity exception in §1610(a)(7), it provides an independent exception to immunity so that it allows a §1605A judgment holder to attach and execute against any property of the foreign state, regardless of whether the property is deprived of immunity elsewhere in §1610. II We turn first to the text of the statute. Section 1610(g)(1) provides that certain property will be “subject to attachment in aid of execution, and execution, upon [a §1605A] judgment as provided in this section.” (Emphasis added.) The most natural reading is that “this section” refers to §1610 as a whole, so that §1610(g)(1) will govern the attachment and execution of property that is exempted from the grant of immunity as provided elsewhere in §1610. Cf. Reno v. American-Arab Anti-Discrimination Comm., 525 U. S. 471, 487 (1999) (noting that the phrase “[e]xcept as provided in this section” in one subsection serves to incorporate “the rest of” the section in which the subsection appears). Other provisions of §1610 unambiguously revoke the immunity of property of a foreign state, including specifically where a plaintiff holds a judgment under §1605A, provided certain express conditions are satisfied. For example, subsection (a) provides that “property in the United States . . . used for a commercial activity in the United States . . . shall not be immune” from attachment and execution in seven enumerated circumstances, including when “the judgment relates to a claim for which the foreign state is not immune under section 1605A . . . .” §1610(a)(7). Subsections (b), (d), and (e) similarly set out circumstances in which certain property of a foreign state “shall not be immune.”[4] And two other provisions within §1610 specifically allow §1605A judgment holders to attach and execute against property of a foreign state, “[n]otwithstanding any other provision of law,” including those provisions otherwise granting immunity, but only with respect to assets associated with certain regulated and prohibited financial transactions. See §1610(f )(1)(A); Terrorism Risk Insurance Act of 2002 (TRIA), §201(a), 116Stat. 2337, note following 28 U. S. C. §1610. Section 1610(g) conspicuously lacks the textual markers, “shall not be immune” or “notwithstanding any other provision of law,” that would have shown that it serves as an independent avenue for abrogation of immunity. In fact, its use of the phrase “as provided in this section” signals the opposite: A judgment holder seeking to take advantage of §1610(g)(1) must identify a basis under one of §1610’s express immunity-abrogating provisions to attach and execute against a relevant property. Reading §1610(g) in this way still provides relief to judgment holders who previously would not have been able to attach and execute against property of an agency or instrumentality of a foreign state in light of this Court’s decision in Bancec. Suppose, for instance, that plaintiffs obtain a §1605A judgment against a foreign state and seek to collect against the assets located in the United States of a state-owned telecommunications company. Cf. Alejandre v. Telefonica Larga Distancia de Puerto Rico, Inc., 183 F. 3d 1277 (CA11 1999). Prior to the enactment of §1610(g), the plaintiffs would have had to establish that the Bancec factors favor holding the agency or instrumentality liable for the foreign state’s misconduct. With §1610(g), however, the plaintiffs could attach and execute against the property of the state-owned entity regardless of the Bancec factors, so long as the plaintiffs can establish that the property is otherwise not immune (e.g., pursuant to §1610(a)(7) because it is used in commercial activity in the United States). Moreover, our reading of §1610(g)(1) is consistent “with one of the most basic interpretive canons, that [a] statute should be construed so that effect is given to all its provisions, so that no part will be inoperative or superfluous, void or insignificant.” Corley v. United States, 556 U. S. 303, 314 (2009) (internal quotation marks omitted). Section 1610 expressly references §1605A judgments in its immunity-abrogating provisions, such as 28 U. S. C. §§1610(a)(7), (b)(3), (f )(1), and §201 of the TRIA, showing that those provisions extend to §1605A judgment holders’ ability to attach and execute against property. If the Court were to conclude that §1610(g) establishes a basis for the withdrawal of property immunity any time a plaintiff holds a judgment under §1605A, each of those provisions would be rendered superfluous because a judgment holder could always turn to §1610(g), regardless of whether the conditions of any other provision were met.[5] The Court’s interpretation of §1610(g) is also consistent with the historical practice of rescinding attachment and execution immunity primarily in the context of a foreign state’s commercial acts. See Verlinden, 461 U. S., at 487–488. Indeed, the FSIA expressly provides in its findings and declaration of purpose that “[u]nder international law, states are not immune from the jurisdiction of foreign courts insofar as their commercial activities are concerned, and their commercial property may be levied upon for the satisfaction of judgments rendered against them in connection with their commercial activities.” §1602. This focus of the FSIA is reflected within §1610, as subsections (a), (b), and (d) all outline exceptions to immunity of property when that property is used for commercial activity. The Court’s reading of §1610(g) means that individuals with §1605A judgments against a foreign state must primarily invoke other provisions revoking the grant of immunity for property related to commercial activity, including §1610(a)(7), unless the property is expressly carved out in an exception that applies “[n]otwithstanding any other provision of law,” §1610(f )(1)(A); §201(a) of the TRIA. That result is consistent with the history and structure of the FSIA. Throughout the FSIA, special avenues of relief to victims of terrorism exist, even absent a nexus to commercial activity. Where the FSIA goes so far as to divest a foreign state or property of immunity in relation to terrorism-related judgments, however, it does so expressly. See §§1605A, 1610(a)(7), (b)(3), (f )(1)(A); §201(a) of the TRIA. Out of respect for the delicate balance that Congress struck in enacting the FSIA, we decline to read into the statute a blanket abrogation of attachment and execution immunity for §1605A judgment holders absent a clearer indication of Congress’ intent. III A Petitioners resist that the phrase “as provided in this section” refers to §1610 as a whole and contend that Congress more likely was referencing a specific provision within §1610 or a section in the NDAA. That explanation is unpersuasive. Petitioners first assert that “this section” might refer to procedures contained in §1610(f ). Section 1610(f ) permits §1605A judgment holders to attach and execute against property associated with certain regulated and prohibited financial transactions, §1610(f )(1), and it provides that the United States Secretary of State and Secretary of the Treasury will make every effort to assist in “identifying, locating, and executing against the property of [a] foreign state or any agency or instrumentality of such state,” §1610(f )(2). Petitioners point out that paragraph (1) of subsection (f ) has never come into effect because it was immediately waived by the President after it was enacted, pursuant to §1610(f )(3).[6] So, the argument goes, it would make sense that Congress created §1610(g) as an alternative mechanism to achieve a similar result.[7] This is a strained and unnatural reading of the phrase “as provided in this section.” In enacting §201(a) of the TRIA, which, similar to 28 U. S. C. §1610(f ), permits attachment and execution against blocked assets, Congress signaled that it was rescinding immunity by permitting attachment and execution “[n]otwithstanding any other provision of law.” See §201(a) of the TRIA. Had Congress likewise intended §1610(g) to have such an effect, it knew how to say so. Cf. Bank Markazi v. Peterson, 578 U. S. ___, ___, n. 2 (2016) (slip op., at 4, n. 2) (noting that “[s]ection 1610(g) does not take precedence over ‘any other provision of law,’ as the TRIA does”). Petitioners fare no better in arguing that Congress may have intended “this section” to refer only to the instruction in §1610(f )(2) that the United States Government assist in identifying assets. Section 1610(f )(2) does not provide for attachment or execution at all, so petitioners’ argument does not account for the lack of textual indicators that exist in provisions like §§1610(a)(7) and (f )(1) that unambiguously abrogate immunity and permit attachment and execution. Finally, petitioners assert that “this section” could possibly reflect a drafting error that was intended to actually refer to §1083 of the NDAA, the Public Law in which §1610(g) was enacted. This interpretation would require not only a stark deviation from the plain text of §1610(g), but also a departure from the clear text of the NDAA. Section 1083(b)(3) of the NDAA provides that “Section 1610 of title 28, United States Code, is amended . . . by adding at the end” the new subsection “(g).” 122Stat. 341. The language “this section” within (g), then, clearly and expressly incorporates the NDAA’s reference to “Section 1610” as a whole. There is no basis to conclude that Congress’ failure to change “this section” in §1610(g) was the result of a mere drafting error. B In an effort to show that §1610(g) does much more than simply abrogate the Bancec factors, petitioners argue that the words “property of a foreign state,” which appear in the first substantive clause of §1610(g), would otherwise be rendered superfluous because the property of a foreign state will never be subject to a Bancec inquiry. By its plain text, §1610(g)(1) permits enforcement of a §1605A judgment against both the property of a foreign state and the property of the agencies or instrumentalities of that foreign state. Because the Bancec factors would never have applied to the property of a foreign state, petitioners contend, those words must signal something else: that §1610(g) provides an independent basis for the withdrawal of immunity. The words “property of a foreign state” accomplish at least two things, however, that are consistent with the Court’s understanding of the effect of §1610(g). First, §1610(g) serves to identify in one place all the categories of property that will be available to §1605A judgment holders for attachment and execution, whether it is “property of the foreign state” or property of its agencies or instrumentalities, and commands that the availability of such property will not be limited by the Bancec factors. So long as the property is deprived of its immunity “as provided in [§1610],” all of the types of property identified in §1610(g) will be available to §1605A judgment holders. Second, in the context of the entire phrase, “the property of a foreign state against which a judgment is entered under section 1605A,” the words “foreign state” identify the type of judgment that will invoke application of §1610(g); specifically, a judgment held against a foreign state and entered under §1605A. Without this opening phrase, §1610(g) would abrogate the Bancec presumption of separateness in all cases, not just those involving terrorism judgments under §1605A. The words, “property of a foreign state,” thus, are not rendered superfluous under the Court’s reading because they do not merely identify a category of property that is subject to §1610(g) but also help inform when §1610(g) will apply in the first place. Indeed, §1610(g) would make no sense if those words were removed. C All else aside, petitioners contend that any uncertainty in §1610(g) should be resolved by giving full effect to the legislative purpose behind its enactment. Petitioners posit that Congress enacted §1610(g) “with the specific purpose of removing the remaining obstacles to terrorism judgment enforcement.” Brief for Petitioners 26. In support of that position, they reference a brief discussion of §1610(g) in a footnote to the Court’s decision in Bank Markazi, 578 U. S. ___, that notes that Congress “expand[ed] the availability of assets for postjudgment execution” when it added §1610(g) by making “available for execution the property (whether or not blocked) of a foreign state sponsor of terrorism, or its agency or instrumentality, to satisfy a judgment against that state.” Id., at ___, n. 2 (slip op., at 4, n. 2). But Bank Markazi’s characterization of §1610(g) simply mirrors the text of §1610(g) and is entirely consistent with the Court’s holding today that §1610(g) expands the assets available for attachment and execution by abrogating this Court’s decision in Bancec with respect to judgments held under §1605A. Beyond their citation to Bank Markazi, petitioners have not directed us to any evidence that supports their position that §1610(g) was intended to divest all property of a foreign state or its agencies or instrumentalities of immunity. IV For the foregoing reasons, we conclude that 28 U. S. C. §1610(g) does not provide a freestanding basis for parties holding a judgment under §1605A to attach and execute against the property of a foreign state, where the immunity of the property is not otherwise rescinded under a separate provision within §1610. The judgment of the Seventh Circuit is affirmed. It is so ordered. Justice Kagan took no part in the consideration or decision of this case. Notes 1 Congress amended the FSIA in 2008 and replaced 28 U. S. C. §1605(a)(7) with a separate, more expansive provision addressing the foreign sovereign immunity of foreign states that are designated as state sponsors of terrorism, §1605A. See National Defense Authorization Act for Fiscal Year 2008 (NDAA), §1083(a), 122Stat. 338–341. Shortly thereafter, petitioners moved in the District Court for an order converting their judgment under §1605(a)(7) to one under the new provision, §1605A, which the District Court granted. See Rubin v. Islamic Republic of Iran, 563 F. Supp. 2d 38, 39, n. 3 (DC 2008). 2 Petitioners also sought to execute the judgment against three other collections that are no longer at issue in this case: the Chogha Mish Collection, the Oriental Institute Collection, and the Herzfeld Collection. The Chogha Mish Collection has been removed from the territorial jurisdiction of the federal courts, and the Court of Appeals for the Seventh Circuit determined that the Oriental Institute Collection and Herzfeld Collection are not property of Iran. See 830 F. 3d 470, 475–476 (2016). Petitioners do not challenge that decision here. 3 Compare Bennett v. Islamic Republic of Iran, 825 F.3d 949, 959 (CA9 2016) (holding that §1610(g) provides a freestanding exception to attachment and execution immunity); Weinstein v. Islamic Republic of Iran, 831 F. 3d 470, 483 (CADC 2016) (same); Kirschenbaum v. 650 Fifth Avenue and Related Properties, 830 F. 3d 107, 123 (CA2 2016) (same), with 830 F. 3d, at 481 (concluding that §1610(g) does not create a freestanding exception to immunity). 4 Section 1610(b), for example, provides that “any property . . . of [the] agency or instrumentality of a foreign state engaged in commercial activity in the United States shall not be immune” from attachment and execution in satisfaction of a judgment on a claim for which the agency or instrumentality is not immune under §1605A. §1610(b)(3). 5 To the extent petitioners suggest that those references to §1605A were inadvertent, see Brief for Petitioners 41–44, the statutory history further supports the conclusion that §1610(a)(7) applies to §1605A judgment holders, as the reference to §1605A was added to §1610(a)(7) in the same Act that created §§1605A and 1610(g). See NDAA §§1083(a), (b)(3), 122Stat. 338–342. 6 Section 1610(f )(3) authorizes the President to waive paragraph (1) of subsection (f ) “in the interest of national security.” President Clinton immediately waived the provision, and the waiver has never been withdrawn. See Pres. Determ. No. 99–1, 63 Fed. Reg. 59201 (1998); Pres. Determ. No. 2001–03, 65 Fed. Reg. 66483 (2000). 7 Petitioners reference the decision of the Court of Appeals for the Ninth Circuit in Bennett, 825 F.3d 949, in support of this position.

584.US.2017_16-969

Inter partes review allows private parties to challenge previously issued patent claims in an adversarial process before the Patent Office. At the outset, a party must file a petition to institute review, 35 U. S. C. §311(a), that identifies the challenged claims and the grounds for challenge with particularity, §312(a)(3). The patent owner, in turn, may file a response. §313. If the Director of the Patent Office determines “there is a reasonable likelihood that the petitioner would prevail with respect to at least 1 of the claims challenged in the petition,” §314(a), he decides “whether to institute . . . review . . . pursuant to [the] petition,” §314(b). “If . . . review is instituted and not dismissed,” at the end of the litigation the Patent Trial and Appeal Board “shall issue a final written decision with respect to the patentability of any patent claim challenged by the petitioner.” §318(a). Petitioner SAS sought review of respondent ComplementSoft’s software patent, alleging that all 16 of the patent’s claims were unpatentable. Relying on a Patent Office regulation recognizing a power of “partial institution,” 37 CFR §42.108(a), the Director instituted review on some of the claims and denied review on the rest. The Board’s final decision addressed only the claims on which the Director had instituted review. On appeal, the Federal Circuit rejected SAS’s argument that §318(a) required the Board to decide the patentability of every claim challenged in the petition. Held: When the Patent Office institutes an inter partes review, it must decide the patentability of all of the claims the petitioner has challenged. The plain text of §318(a) resolves this case. Its directive is both mandatory and comprehensive. The word “shall” generally imposes a nondiscretionary duty, and the word “any” ordinarily implies every member of a group. Thus, §318(a) means that the Board must address every claim the petitioner has challenged. The Director’s “partial institution” power appears nowhere in the statutory text. And both text and context strongly counsel against inferring such a power. The statute envisions an inter partes review guided by the initial petition. See §312(a)(3). Congress structured the process such that the petitioner, not the Director, defines the proceeding’s contours. The ex parte reexamination statute shows that Congress knew exactly how to authorize the Director to investigate patentability questions “[o]n his own initiative, and at any time,” §303(a). The inter partes review statute indicates that the Director’s decision “whether” to institute review “pursuant to [the] petition” is a yes-or-no choice. §314(b). Section 314(a)’s requirement that the Director find “a reasonable likelihood” that the petitioner will prevail on “at least 1 of the claims challenged in the petition” suggests, if anything, a regime where a reasonable prospect of success on a single claim justifies review of them all. Again, if Congress had wanted to adopt the Director’s claim-by-claim approach, it knew how to do so. See §304. Nor does it follow that, because §314(a) invests the Director with discretion on the question whether to institute review, it also invests him with discretion regarding what claims that review will encompass. The rest of the statute confirms, too, that the petitioner’s petition, not the Director’s discretion, should guide the life of the litigation. See, e.g., §316(a)(8). The Director suggests that a textual discrepancy between §314(a)—which addresses whether to institute review based on claims found “in the petition”—and §318(a)—which addresses the Board’s final resolution of the claims challenged “by the petitioner”—means that the Director enjoys the power to institute a review covering fewer than all of the claims challenged in the petition. However, the statute’s winnowing mechanism—which allows a patent owner to concede one part of a petitioner’s challenge and “[c]ancel any challenged patent claim,” §316(d)(1)(A)—fully explains why Congress adopted the slightly different language. The Director’s policy argument—that partial institution is efficient because it permits the Board to focus on the most promising challenges and avoid spending time and resources on others—is properly addressed to Congress, not this Court. And the Director’s asserted “partial institution” power, which is wholly unmentioned in the statute, is not entitled to deference under Chevron U. S. A. Inc. v. Natural Resources Defense Council, Inc., 467 U. S. 837 . Finally, notwithstanding §314(d)—which makes the Director’s determination whether to institute an inter partes review “final and nonappealable”—judicial review remains available consistent with the Administrative Procedure Act to ensure that the Patent Office does not exceed its statutory bounds. Cuozzo Speed Technologies, LLC v. Lee, 579 U. S. ___, distinguished. Pp. 4–14. 825 F. 3d 1341, reversed and remanded. Gorsuch, J., delivered the opinion of the Court, in which Roberts, C. J., and Kennedy, Thomas, and Alito, JJ., joined. Ginsburg, J., filed a dissenting opinion, in which Breyer, Sotomayor, and Kagan, JJ., joined. Breyer, J., filed a dissenting opinion, in which Ginsburg and Sotomayor, JJ., joined, and in which Kagan, J., joined except as to Part III–A.

A few years ago Congress created “inter partes review.” The new procedure allows private parties to challenge previously issued patent claims in an adversarial process before the Patent Office that mimics civil litigation. Recently, the Court upheld the inter partes review statute against a constitutional challenge. Oil States Energy Services, LLC v. Greene’s Energy Group, LLC, ante, p. ___. Now we take up a question concerning the statute’s operation. When the Patent Office initiates an inter partes review, must it resolve all of the claims in the case, or may it choose to limit its review to only some of them? The statute, we find, supplies a clear answer: the Patent Office must “issue a final written decision with respect to the patentability of any patent claim challenged by the petitioner.” 35 U. S. C. §318(a) (emphasis added). In this context, as in so many others, “any” means “every.” The agency cannot curate the claims at issue but must decide them all. “To promote the Progress of Science and useful Arts,” Congress long ago created a patent system granting inventors rights over the manufacture, sale, and use of their inventions. U. S. Const., Art. I, §8, cl. 8; see 35 U. S. C. §154(a)(1). To win a patent, an applicant must (among other things) file “claims” that describe the invention and establish to the satisfaction of the Patent Office the invention’s novelty and nonobviousness. See §§102, 103, 112(b), 131; Cuozzo Speed Technologies, LLC v. Lee, 579 U. S. ___, ___–___ (2016) (slip op., at 2–3). Sometimes, though, bad patents slip through. Maybe the invention wasn’t novel, or maybe it was obvious all along, and the patent owner shouldn’t enjoy the special privileges it has received. To remedy these sorts of problems, Congress has long permitted parties to challenge the validity of patent claims in federal court. See §§282(b)(2)–(3). More recently, Congress has supplemented litigation with various administrative remedies. The first of these was ex parte reexamination. Anyone, including the Director of the Patent Office, can seek ex parte reexamination of a patent claim. §§302, 303(a). Once instituted, though, an ex parte reexamination follows essentially the same inquisitorial process between patent owner and examiner as the initial Patent Office examination. §305. Later, Congress supplemented ex parte reexamination with inter partes reexamination. Inter partes reexamination (since repealed) provided a slightly more adversarial process, allowing a third party challenger to submit comments throughout the proceeding. §314(b)(2) (2006 ed.) (repealed). But otherwise it too followed a more or less inquisitorial course led by the Patent Office. §314(a). Apparently unsatisfied with this approach, in 2011 Congress repealed inter partes reexamination and replaced it with inter partes review. See 35 U. S. C. §§311–319 (2012 ed.). The new inter partes review regime looks a good deal more like civil litigation. At its outset, a party must file “a petition to institute an inter partes review of [a] patent.” §311(a). The petition “may request to cancel as unpatentable 1 or more claims of [the] patent” on the ground that the claims are obvious or not novel. §311(b); see §§102 and 103. In doing so, the petition must identify “each claim challenged,” the grounds for the challenge, and the evidence supporting the challenge. §312(a)(3). The patent owner, in turn, may respond with “a preliminary response to the petition” explaining “why no inter partes review should be instituted.” §313. With the parties’ submissions before him, the Director then decides “whether to institute an inter partes review . . . pursuant to [the] petition.” §314(b). (In practice, the agency’s Patent Trial and Appeal Board exercises this authority on behalf of the Director, see 37 CFR §42.4(a) (2017).) Before instituting review, the Director must determine, based on the parties’ papers, “that there is a reasonable likelihood that the petitioner would prevail with respect to at least 1 of the claims challenged in the petition.” 35 U. S. C. §314(a). Once the Director institutes an inter partes review, the matter proceeds before the Board with many of the usual trappings of litigation. The parties conduct discovery and join issue in briefing and at an oral hearing. §§316(a)(5), (6), (8), (10), (13). During the course of the case, the patent owner may seek to amend its patent or to cancel one or more of its claims. §316(d). The parties may also settle their differences and seek to end the review. §317. But “[i]f an inter partes review is instituted and not dismissed,” at the end of the litigation the Board “shall issue a final written decision with respect to the patentability of any patent claim challenged by the petitioner.” §318(a). Our case arose when SAS sought an inter partes review of ComplementSoft’s software patent. In its petition, SAS alleged that all 16 of the patent’s claims were unpatent- able for various reasons. The Director (in truth the Board acting on the Director’s behalf) concluded that SAS was likely to succeed with respect to at least one of the claims and that an inter partes review was therefore warranted. But instead of instituting review on all of the claims challenged in the petition, the Director instituted review on only some (claims 1 and 3–10) and denied review on the rest. The Director did all this on the strength of a Patent Office regulation that purported to recognize a power of “partial institution,” claiming that “[w]hen instituting inter partes review, the [Director] may authorize the review to proceed on all or some of the challenged claims and on all or some or the grounds of unpatentability asserted for each claim.” 37 CFR §42.108(a). At the end of litigation, the Board issued a final written decision finding claims 1, 3, and 5–10 to be unpatentable while upholding claim 4. But the Board’s decision did not address the remaining claims on which the Director had refused review. That last fact led SAS to seek review in the Federal Circuit. There SAS argued that 35 U. S. C. §318(a) required the Board to decide the patentability of every claim SAS challenged in its petition, not just some. For its part, the Federal Circuit rejected SAS’s argument over a vigorous dissent by Judge Newman. SAS Institute, Inc. v. ComplementSoft, LLC, 825 F. 3d 1341 (2016). We granted certiorari to decide the question ourselves. 581 U. S. ___ (2017). We find that the plain text of §318(a) supplies a ready answer. It directs that “[i]f an inter partes review is instituted and not dismissed under this chapter, the [Board] shall issue a final written decision with respect to the patentability of any patent claim challenged by the petitioner . . . .” §318(a) (emphasis added). This directive is both mandatory and comprehensive. The word “shall” generally imposes a nondiscretionary duty. See Lexecon Inc. v. Milberg Weiss Bershad Hynes & Lerach, 523 U. S. 26, 35 (1998) . And the word “any” naturally carries “an expansive meaning.” United States v. Gonzales, 520 U. S. 1, 5 (1997). When used (as here) with a “singular noun in affirmative contexts,” the word “any” ordinarily “refer[s] to a member of a particular group or class without distinction or limitation” and in this way “impl[ies] every member of the class or group.” Oxford English Dictionary (3d ed., Mar. 2016), www.oed.com/view/Entry/8973 (OED) (emphasis added) (all Internet materials as last visited Apr. 20, 2018). So when §318(a) says the Board’s final written decision “shall” resolve the patentability of “any patent claim challenged by the petitioner,” it means the Board must address every claim the petitioner has challenged. That would seem to make this an easy case. Where a statute’s language carries a plain meaning, the duty of an administrative agency is to follow its commands as written, not to supplant those commands with others it may prefer. Social Security Bd. v. Nierotko, 327 U. S. 358, 369 (1946) . Because SAS challenged all 16 claims of ComplementSoft’s patent, the Board in its final written decision had to address the patentability of all 16 claims. Much as in the civil litigation system it mimics, in an inter partes review the petitioner is master of its complaint and normally entitled to judgment on all of the claims it raises, not just those the decisionmaker might wish to address. The Director replies that things are not quite as simple as they seem. Maybe the Board has to decide every claim challenged by the petitioner in an inter partes review. But, he says, that doesn’t mean every challenged claim gains admission to the review process. In the Director’s view, he retains discretion to decide which claims make it into an inter partes review and which don’t. The trouble is, nothing in the statute says anything like that. The Director’s claimed “partial institution” power appears nowhere in the text of §318, or anywhere else in the statute for that matter. And what can be found in the statutory text and context strongly counsels against the Director’s view. Start where the statute does. In its very first provision, the statute says that a party may seek inter partes review by filing “a petition to institute an inter partes review.” §311(a). This language doesn’t authorize the Director to start proceedings on his own initiative. Nor does it contemplate a petition that asks the Director to initiate whatever kind of inter partes review he might choose. Instead, the statute envisions that a petitioner will seek an inter partes review of a particular kind—one guided by a petition describing “each claim challenged” and “the grounds on which the challenge to each claim is based.” §312(a)(3). From the outset, we see that Congress chose to structure a process in which it’s the petitioner, not the Director, who gets to define the contours of the proceeding. And “[ j]ust as Congress’ choice of words is presumed to be deliberate” and deserving of judicial respect, “so too are its structural choices.” University of Tex. Southwestern Medical Center v. Nassar, 570 U. S. 338, 353 (2013) . It’s telling, too, to compare this structure with what came before. In the ex parte reexamination statute, Congress embraced an inquisitorial approach, authorizing the Director to investigate a question of patentability “[o]n his own initiative, and at any time.” §303(a). If Congress had wanted to give the Director similar authority over the institution of inter partes review, it knew exactly how to do so—it could have simply borrowed from the statute next door. But rather than create (another) agency-led, inquisitorial process for reconsidering patents, Congress opted for a party-directed, adversarial process. Congress’s choice to depart from the model of a closely related statute is a choice neither we nor the agency may disregard. See Nassar, supra, at 353–354. More confirmation comes as we move to the point of institution. Here the statute says the Director must decide “whether to institute an inter partes review . . . pursuant to a petition.” §314(b). The Director, we see, is given only the choice “whether” to institute an inter partes review. That language indicates a binary choice—either institute review or don’t. And by using the term “pursuant to,” Congress told the Director what he must say yes or no to: an inter partes review that proceeds “[i]n accordance with” or “in conformance to” the petition. OED, www.oed.com/view/Entry/155073. Nothing suggests the Director enjoys a license to depart from the petition and institute a different inter partes review of his own design. To this the Director replies by pointing to another part of §314. Section 314(a) provides that the Director may not authorize an inter partes review unless he determines “there is a reasonable likelihood” the petitioner will prevail on “at least 1 of the claims challenged in the petition.” The Director argues that this language requires him to “evaluate claims individually” and so must allow him to institute review on a claim-by-claim basis as well. Brief for Federal Respondent 28. But this language, if anything, suggests just the opposite. Section 314(a) does not require the Director to evaluate every claim individually. Instead, it simply requires him to decide whether the petitioner is likely to succeed on “at least 1” claim. Once that single claim threshold is satisfied, it doesn’t matter whether the petitioner is likely to prevail on any additional claims; the Director need not even consider any other claim before instituting review. Rather than contemplate claim-by-claim institution, then, the language anticipates a regime where a reasonable prospect of success on a single claim justifies review of all. Here again we know that if Congress wanted to adopt the Director’s approach it knew exactly how to do so. The ex parte reexamination statute allows the Director to assess whether a request raises “a substantial new question of patentability affecting any claim” and (if so) to institute reexamination limited to “resolution of the question.” §304 (emphasis added). In other words, that statute allows the Director to institute proceedings on a claim-by-claim and ground-by-ground basis. But Congress didn’t choose to pursue that known and readily available approach here. And its choice to try something new must be given effect rather than disregarded in favor of the comfort of what came before. See Nassar, supra, at 353–354. Faced with this difficulty, the Director tries another tack. He points to the fact that §314(a) doesn’t require him to institute an inter partes review even after he finds the “reasonable likelihood” threshold met with respect to one claim. Whether to institute proceedings upon such a finding, he says, remains a matter left to his discretion. See Cuozzo, 579 U. S., at ___ (slip op., at 9). But while §314(a) invests the Director with discretion on the question whether to institute review, it doesn’t follow that the statute affords him discretion regarding what claims that review will encompass. The text says only that the Director can decide “whether” to institute the requested review—not “whether and to what extent” review should proceed. §314(b). The rest of the statute confirms, too, that the petitioner’s petition, not the Director’s discretion, is supposed to guide the life of the litigation. For example, §316(a)(8) tells the Director to adopt regulations ensuring that, “after an inter partes review has been instituted,” the patent owner will file “a response to the petition.” Surely it would have made little sense for Congress to insist on a response to the petition if, in truth, the Director enjoyed the discretion to limit the claims under review. What’s the point, after all, of answering claims that aren’t in the proceeding? If Congress had meant to afford the Director the power he asserts, we would have expected it to instruct him to adopt regulations requiring the patent owner to file a response to the Director’s institution notice or to the claims on which the Director instituted review. Yet we have nothing like that here. And then and again there is §318(a). At the end of the proceeding, §318(a) categorically commands the Board to address in its final written decision “any patent claim challenged by the petitioner.” In all these ways, the statute tells us that the petitioner’s contentions, not the Director’s discretion, define the scope of the litigation all the way from institution through to conclusion. The Director says we can find at least some hint of the discretion he seeks by comparing §314(a) and §318(a). He notes that, when addressing whether to institute review at the beginning of the litigation, §314(a) says he must focus on the claims found “in the petition”; but when addressing what claims the Board must address at the end of the litigation, §318(a) says it must resolve the claims challenged “by the petitioner.” According to the Director, this (slight) linguistic discrepancy means the claims the Board must address in its final decision are not necessarily the same as those identified in the petition. And the only possible explanation for this arrangement, the Director submits, is that he must enjoy the (admittedly implicit) power to institute an inter partes review that covers fewer than all of the claims challenged in the petition. We just don’t see it. Whatever differences they might display, §314(a) and §318(a) both focus on the petitioner’s contentions and, given that, it’s difficult to see how they might be read to give the Director power to decide what claims are at issue. Particularly when there’s a much simpler and sounder explanation for the statute’s wording. As we’ve seen, a patent owner may move to “[c]ancel any challenged patent claim” during the course of an inter partes review, effectively conceding one part of a petitioner’s challenge. §316(d)(1)(A). Naturally, then, the claims challenged “in the petition” will not always survive to the end of the case; some may drop out thanks to the patent owner’s actions. And in that light it is plain enough why Congress provided that only claims still challenged “by the petitioner” at the litigation’s end must be addressed in the Board’s final written decision. The statute’s own winnowing mechanism fully explains why Congress adopted slightly different language in §314(a) and §318(a). We need not and will not invent an atextual explanation for Congress’s drafting choices when the statute’s own terms supply an answer. See United States v. Ron Pair Enterprises, Inc., 489 U. S. 235, 240 –241 (1989) (“[A]s long as the statutory scheme is coherent and consistent, there generally is no need for a court to inquire beyond the plain language of the statute”). Moving past the statute’s text and context, the Director attempts a policy argument. He tells us that partial institution is efficient because it permits the Board to focus on the most promising challenges and avoid spending time and resources on others. Brief for Federal Respondent 35–36; see also post, at 1 (Ginsburg, J., dissenting); post, at 7–8 (Breyer, J., dissenting). SAS responds that all patent challenges usually end up being litigated somewhere, and that partial institution creates inefficiency by requiring the parties to litigate in two places instead of one—the Board for claims the Director chooses to entertain and a federal court for claims he refuses. Indeed, SAS notes, the government itself once took the same view, arguing that partial institution “ ‘undermine[s] the Congressional efficiency goal’ ” for this very reason. Brief for Petitioner 30. Each side offers plausible reasons why its approach might make for the more efficient policy. But who should win that debate isn’t our call to make. Policy arguments are properly addressed to Congress, not this Court. It is Congress’s job to enact policy and it is this Court’s job to follow the policy Congress has prescribed. And whatever its virtues or vices, Congress’s prescribed policy here is clear: the petitioner in an inter partes review is entitled to a decision on all the claims it has challenged.[1] That leaves the Director to suggest that, however this Court might read the statute, he should win anyway because of Chevron U. S. A. Inc. v. Natural Resources Defense Council, Inc., 467 U. S. 837 (1984) . Even though the statute says nothing about his asserted “partial institution” power, the Director says the statute is at least ambiguous on the propriety of the practice and so we should leave the matter to his judgment. For its part, SAS replies that we might use this case as an opportunity to abandon Chevron and embrace the “ ‘impressive body’ ” of pre-Chevron law recognizing that “ ‘the meaning of a statutory term’ ” is properly a matter for “ ‘judicial [rather than] administrative judgment.’ ” Brief for Petitioner 41 (quoting Pittston Stevedoring Corp. v. Dellaventura, 544 F. 2d 35, 49 (CA2 1976) (Friendly, J.)). But whether Chevron should remain is a question we may leave for another day. Even under Chevron, we owe an agency’s interpretation of the law no deference unless, after “employing traditional tools of statutory construction,” we find ourselves unable to discern Congress’s meaning. 467 U. S., at 843, n. 9. And after applying traditional tools of interpretation here, we are left with no uncertainty that could warrant deference. The statutory provisions before us deliver unmistakable commands. The statute hinges inter partes review on the filing of a petition challenging specific patent claims; it makes the petition the centerpiece of the proceeding both before and after institution; and it requires the Board’s final written decision to address every claim the petitioner presents for review. There is no room in this scheme for a wholly unmentioned “partial institution” power that lets the Director select only some challenged claims for decision. The Director may (today) think his approach makes for better policy, but policy considerations cannot create an ambiguity when the words on the page are clear. See SEC v. Sloan, 436 U. S. 103, 116 –117 (1978). Neither may we defer to an agency official’s preferences because we imagine some “hypothetical reasonable legislator” would have favored that approach. Post, at 9 (Breyer, J., dissenting). Our duty is to give effect to the text that 535 actual legislators (plus one President) enacted into law. At this point, only one final question remains to resolve. Even if the statute forbids his partial institution practice, the Director suggests we lack the power to say so. By way of support, he points to §314(d) and our decision in Cuozzo, 579 U. S. ___. Section 314(d) says that the “determination by the Director whether to institute an inter partes review under this section shall be final and nonappealable.” In Cuozzo, we held that this provision prevented courts from entertaining an argument that the Director erred in instituting an inter partes review of certain patent claims. Id., at ___–___ (slip op., at 7–12). The Director reads these authorities as foreclosing judicial review of any legal question bearing on the institution of inter partes review—including whether the statute permits his “partial institution” practice. But this reading overreads both the statute and our precedent. As Cuozzo recognized, we begin with “the ‘strong presumption’ in favor of judicial review.” Id., at ___ (slip op., at 9). To overcome that presumption, Cuozzo explained, this Court’s precedents require “clear and convincing indications” that Congress meant to foreclose review. Id., at ___ (slip op., at 10) (internal quotation marks omitted). Given the strength of this presumption and the statute’s text, Cuozzo concluded that §314(d) precludes judicial review only of the Director’s “initial determination” under §314(a) that “there is a ‘reasonable likelihood’ that the claims are unpatentable on the grounds asserted” and review is therefore justified. Id., at ___ (slip op., at 9); see id., at ___ (slip op., at 12) (review unavailable “where a patent holder merely challenges the Patent Office’s ‘determin[ation] that the information presented in the petition . . . shows that there is a reasonable likelihood’ of success ‘with respect to at least 1 of the claims challenged’ ”); ibid. (claim that a “petition was not pleaded ‘with particularity’ under §312 is little more than a challenge to the Patent Office’s conclusion, under §314(a), that the ‘information presented in the petition’ warranted review”). In fact, Cuozzo proceeded to emphasize that §314(d) does not “enable the agency to act outside its statutory limits.” Id., at ___ (slip op., at 11). If a party believes the Patent Office has engaged in “ ‘shenanigans’ ” by exceeding its statutory bounds, judicial review remains available consistent with the Administrative Procedure Act, which directs courts to set aside agency action “not in accordance with law” or “in excess of statutory jurisdiction, authority, or limitations.” Ibid.; 5 U. S. C. §§706(2)(A), (C). And that, of course, is exactly the sort of question we are called upon to decide today. SAS does not seek to challenge the Director’s conclusion that it showed a “reason- able likelihood” of success sufficient to warrant “institut[ing] an inter partes review.” 35 U. S. C. §§314(a), (d). No doubt SAS remains very pleased with the Director’s judgment on that score. Instead, SAS contends that the Director exceeded his statutory authority by limiting the review to fewer than all of the claims SAS challenged. And nothing in §314(d) or Cuozzo withdraws our power to ensure that an inter partes review proceeds in accordance with the law’s demands. Because everything in the statute before us confirms that SAS is entitled to a final written decision addressing all of the claims it has challenged and nothing suggests we lack the power to say so, the judgment of the Federal Circuit is reversed and the case is remanded for further proceedings consistent with this opinion. So ordered. Notes 1 Justice Ginsburg suggests the Director might yet avoid this command by refusing to review a petition he thinks too broad while signaling his willingness to entertain one more tailored to his sympathies. Post, at 1 (dissenting opinion). We have no occasion today to consider whether this stratagem is consistent with the statute’s demands. See Cuozzo Speed Technologies, LLC v. Lee, 579 U. S. ___, ___ (2016) (slip op., at 11) (noting that courts may invalidate “ ‘shenanigans’ ” by the Director that are “outside [his] statutory limits”); CAB v. Delta Air Lines, Inc., 367 U. S. 316, 328 (1961) (questioning an agency’s “power to do indirectly what it cannot do directly”). But even assuming (without granting) the law would tolerate this tactic, it would show only that a lawful means exists for the Director to achieve his policy aims—not that he “should be allowed to improvise on the powers granted by Congress” by devising an extralegal path to the same goal. Id., at 330. That an agency’s improvisation might be thought by some more expedient than what the law allows, post, at 1, does nothing to commend it either, for lawful ends do not justify unlawful means.

585.US.2017_17-494

South Dakota, like many States, taxes the retail sales of goods and services in the State. Sellers are required to collect and remit the tax to the State, but if they do not then in-state consumers are responsible for paying a use tax at the same rate. Under National Bellas Hess, Inc. v. Department of Revenue of Ill., 386 U. S. 753, and Quill Corp. v. North Dakota, 504 U. S. 298, South Dakota may not require a business that has no physical presence in the State to collect its sales tax. Consumer compliance rates are notoriously low, however, and it is estimated that Bellas Hess and Quill cause South Dakota to lose between $48 and $58 million annually. Concerned about the erosion of its sales tax base and corresponding loss of critical funding for state and local services, the South Dakota Legislature enacted a law requiring out-of-state sellers to collect and remit sales tax “as if the seller had a physical presence in the State.” The Act covers only sellers that, on an annual basis, deliver more than $100,000 of goods or services into the State or engage in 200 or more separate transactions for the delivery of goods or services into the State. Respondents, top online retailers with no employees or real estate in South Dakota, each meet the Act’s minimum sales or transactions requirement, but do not collect the State’s sales tax. South Dakota filed suit in state court, seeking a declaration that the Act’s requirements are valid and applicable to respondents and an injunction requiring respondents to register for licenses to collect and remit the sales tax. Respondents sought summary judgment, arguing that the Act is unconstitutional. The trial court granted their motion. The State Supreme Court affirmed on the ground that Quill is controlling precedent. Held: Because the physical presence rule of Quill is unsound and incorrect, Quill Corp. v. North Dakota, 504 U. S. 298, and National Bellas Hess, Inc. v. Department of Revenue of Ill., 386 U. S. 753, are overruled. Pp. 5–24. (a) An understanding of this Court’s Commerce Clause principles and their application to state taxes is instructive here. Pp. 5–9. (1) Two primary principles mark the boundaries of a State’s authority to regulate interstate commerce: State regulations may not discriminate against interstate commerce; and States may not impose undue burdens on interstate commerce. These principles guide the courts in adjudicating challenges to state laws under the Commerce Clause. Pp. 5–7. (2) They also animate Commerce Clause precedents addressing the validity of state taxes, which will be sustained so long as they (1) apply to an activity with a substantial nexus with the taxing State, (2) are fairly apportioned, (3) do not discriminate against interstate commerce, and (4) are fairly related to the services the State provides. See Complete Auto Transit, Inc. v. Brady, 430 U. S. 274, 279. Before Complete Auto, the Court held in Bellas Hess that a “seller whose only connection with customers in the State is by common carrier or . . . mail” lacked the requisite minimum contacts with the State required by the Due Process Clause and the Commerce Clause, and that unless the retailer maintained a physical presence in the State, the State lacked the power to require that retailer to collect a local tax. 386 U. S., at 758. In Quill, the Court overruled the due process holding, but not the Commerce Clause holding, grounding the physical presence rule in Complete Auto’s requirement that a tax have a “substantial nexus” with the activity being taxed. Pp. 7–9. (b) The physical presence rule has long been criticized as giving out-of-state sellers an advantage. Each year, it becomes further removed from economic reality and results in significant revenue losses to the States. These critiques underscore that the rule, both as first formulated and as applied today, is an incorrect interpretation of the Commerce Clause. Pp. 9–17. (1) Quill is flawed on its own terms. First, the physical presence rule is not a necessary interpretation of Complete Auto’s nexus requirement. That requirement is “closely related,” Bellas Hess, 386 U. S. at 756, to the due process requirement that there be “some definite link, some minimum connection, between a state and the person, property or transaction it seeks to tax.” Miller Brothers Co. v. Maryland, 347 U. S. 340, 344–345. And, as Quill itself recognized, a business need not have a physical presence in a State to satisfy the demands of due process. When considering whether a State may levy a tax, Due Process and Commerce Clause standards, though not identical or coterminous, have significant parallels. The reasons given in Quill for rejecting the physical presence rule for due process purposes apply as well to the question whether physical presence is a requisite for an out-of-state seller’s liability to remit sales taxes. Other aspects of the Court’s doctrine can better and more accurately address potential burdens on interstate commerce, whether or not Quill’s physical presence rule is satisfied. Second, Quill creates rather than resolves market distortions. In effect, it is a judicially created tax shelter for businesses that limit their physical presence in a State but sell their goods and services to the State’s consumers, something that has become easier and more prevalent as technology has advanced. The rule also produces an incentive to avoid physical presence in multiple States, affecting development that might be efficient or desirable. Third, Quill imposes the sort of arbitrary, formalistic distinction that the Court’s modern Commerce Clause precedents disavow in favor of “a sensitive, case-by-case analysis of purposes and effects,” West Lynn Creamery, Inc. v. Healy, 512 U. S. 186, 201. It treats economically identical actors differently for arbitrary reasons. For example, a business that maintains a few items of inventory in a small warehouse in a State is required to collect and remit a tax on all of its sales in the State, while a seller with a pervasive Internet presence cannot be subject to the same tax for the sales of the same items. Pp. 10–14. (2) When the day-to-day functions of marketing and distribution in the modern economy are considered, it becomes evident that Quill’s physical presence rule is artificial, not just “at its edges,” 504 U. S. at 315, but in its entirety. Modern e-commerce does not align analytically with a test that relies on the sort of physical presence defined in Quill. And the Court should not maintain a rule that ignores substantial virtual connections to the State. Pp. 14–15. (3) The physical presence rule of Bellas Hess and Quill is also an extraordinary imposition by the Judiciary on States’ authority to collect taxes and perform critical public functions. Forty-one States, two Territories, and the District of Columbia have asked the Court to reject Quill’s test. Helping respondents’ customers evade a lawful tax unfairly shifts an increased share of the taxes to those consumers who buy from competitors with a physical presence in the State. It is essential to public confidence in the tax system that the Court avoid creating inequitable exceptions. And it is also essential to the confidence placed in the Court’s Commerce Clause decisions. By giving some online retailers an arbitrary advantage over their competitors who collect state sales taxes, Quill’s physical presence rule has limited States’ ability to seek long-term prosperity and has prevented market participants from competing on an even playing field. Pp. 16–17. (c) Stare decisis can no longer support the Court’s prohibition of a valid exercise of the States’ sovereign power. If it becomes apparent that the Court’s Commerce Clause decisions prohibit the States from exercising their lawful sovereign powers, the Court should be vigilant in correcting the error. It is inconsistent with this Court’s proper role to ask Congress to address a false constitutional premise of this Court’s own creation. The Internet revolution has made Quill’s original error all the more egregious and harmful. The Quill Court did not have before it the present realities of the interstate marketplace, where the Internet’s prevalence and power have changed the dynamics of the national economy. The expansion of e-commerce has also increased the revenue shortfall faced by States seeking to collect their sales and use taxes, leading the South Dakota Legislature to declare an emergency. The argument, moreover, that the physical presence rule is clear and easy to apply is unsound, as attempts to apply the physical presence rule to online retail sales have proved unworkable. Because the physical presence rule as defined by Quill is no longer a clear or easily applicable standard, arguments for reliance based on its clarity are misplaced. Stare decisis may accommodate “legitimate reliance interest[s],” United States v. Ross, 456 U. S. 798, 824, but a business “is in no position to found a constitutional right . . . on the practical opportunities for tax avoidance,” Nelson v. Sears, Roebuck & Co., 312 U. S. 359, 366. Startups and small businesses may benefit from the physical presence rule, but here South Dakota affords small merchants a reasonable degree of protection. Finally, other aspects of the Court’s Commerce Clause doctrine can protect against any undue burden on interstate commerce, taking into consideration the small businesses, startups, or others who engage in commerce across state lines. The potential for such issues to arise in some later case cannot justify retaining an artificial, anachronistic rule that deprives States of vast revenues from major businesses. Pp. 17–22. (d) In the absence of Quill and Bellas Hess, the first prong of the Complete Auto test simply asks whether the tax applies to an activity with a substantial nexus with the taxing State, 430 U. S., at 279. Here, the nexus is clearly sufficient. The Act applies only to sellers who engage in a significant quantity of business in the State, and respondents are large, national companies that undoubtedly maintain an extensive virtual presence. Any remaining claims regarding the Commerce Clause’s application in the absence of Quill and Bellas Hess may be addressed in the first instance on remand. Pp. 22–23. 2017 S.D. 56, 901 N. W. 2d 754, vacated and remanded. Kennedy, J., delivered the opinion of the Court, in which Thomas, Ginsburg, Alito, and Gorsuch, JJ., joined. Thomas, J., and Gorsuch, J., filed concurring opinions. Roberts, C. J., filed a dissenting opinion, in which Breyer, Sotomayor, and Kagan, JJ., joined.

When a consumer purchases goods or services, the consumer’s State often imposes a sales tax. This case requires the Court to determine when an out-of-state seller can be required to collect and remit that tax. All concede that taxing the sales in question here is lawful. The question is whether the out-of-state seller can be held responsible for its payment, and this turns on a proper interpretation of the Commerce Clause, U. S. Const., Art. I, §8, cl. 3. In two earlier cases the Court held that an out-of-state seller’s liability to collect and remit the tax to the consumer’s State depended on whether the seller had a physical presence in that State, but that mere shipment of goods into the consumer’s State, following an order from a catalog, did not satisfy the physical presence requirement. National Bellas Hess, Inc. v. Department of Revenue of Ill., 386 U. S. 753 (1967); Quill Corp. v. North Dakota, 504 U. S. 298 (1992). The Court granted certiorari here to reconsider the scope and validity of the physical presence rule mandated by those cases. I Like most States, South Dakota has a sales tax. It taxes the retail sales of goods and services in the State. S. D. Codified Laws §§10–45–2, 10–45–4 (2010 and Supp. 2017). Sellers are generally required to collect and remit this tax to the Department of Revenue. §10–45–27.3. If for some reason the sales tax is not remitted by the seller, then in-state consumers are separately responsible for paying a use tax at the same rate. See §§10–46–2, 10–46–4, 10–46–6. Many States employ this kind of complementary sales and use tax regime. Under this Court’s decisions in Bellas Hess and Quill, South Dakota may not require a business to collect its sales tax if the business lacks a physical presence in the State. Without that physical presence, South Dakota instead must rely on its residents to pay the use tax owed on their purchases from out-of-state sellers. “[T]he impracticability of [this] collection from the multitude of individual purchasers is obvious.” National Geographic Soc. v. California Bd. of Equalization, 430 U. S. 551, 555 (1977). And consumer compliance rates are notoriously low. See, e.g., GAO, Report to Congressional Requesters: Sales Taxes, States Could Gain Revenue from Expanded Authority, but Businesses Are Likely to Experience Compliance Costs 5 (GAO–18–114, Nov. 2017) (Sales Taxes Report); California State Bd. of Equalization, Revenue Estimate: Electronic Commerce and Mail Order Sales 7 (2013) (Table 3) (estimating a 4 percent collection rate). It is estimated that Bellas Hess and Quill cause the States to lose between $8 and $33 billion every year. See Sales Taxes Report, at 11–12 (estimating $8 to $13 billion); Brief for Petitioner 34–35 (citing estimates of $23 and $33.9 billion). In South Dakota alone, the Department of Revenue estimates revenue loss at $48 to $58 million annually. App. 24. Particularly because South Dakota has no state income tax, it must put substantial reliance on its sales and use taxes for the revenue necessary to fund essential services. Those taxes account for over 60 percent of its general fund. In 2016, South Dakota confronted the serious inequity Quill imposes by enacting S. 106—“An Act to provide for the collection of sales taxes from certain remote sellers, to establish certain Legislative findings, and to declare an emergency.” S. 106, 2016 Leg. Assembly, 91st Sess. (S. D. 2016) (S. B. 106). The legislature found that the inability to collect sales tax from remote sellers was “seriously eroding the sales tax base” and “causing revenue losses and imminent harm . . . through the loss of critical funding for state and local services.” §8(1). The legislature also declared an emergency: “Whereas, this Act is necessary for the support of the state government and its existing public institutions, an emergency is hereby declared to exist.” §9. Fearing further erosion of the tax base, the legislature expressed its intention to “apply South Dakota’s sales and use tax obligations to the limit of federal and state constitutional doctrines” and noted the urgent need for this Court to reconsider its precedents. §§8(11), (8). To that end, the Act requires out-of-state sellers to collect and remit sales tax “as if the seller had a physical presence in the state.” §1. The Act applies only to sellers that, on an annual basis, deliver more than $100,000 of goods or services into the State or engage in 200 or more separate transactions for the delivery of goods or services into the State. Ibid. The Act also forecloses the retroactive application of this requirement and provides means for the Act to be appropriately stayed until the constitutionality of the law has been clearly established. §§5, 3, 8(10). Respondents Wayfair, Inc., Overstock.com, Inc., and Newegg, Inc., are merchants with no employees or real estate in South Dakota. Wayfair, Inc., is a leading online retailer of home goods and furniture and had net revenues of over $4.7 billion last year. Overstock.com, Inc., is one of the top online retailers in the United States, selling a wide variety of products from home goods and furniture to clothing and jewelry; and it had net revenues of over $1.7 billion last year. Newegg, Inc., is a major online retailer of consumer electronics in the United States. Each of these three companies ships its goods directly to purchasers throughout the United States, including South Dakota. Each easily meets the minimum sales or transactions requirement of the Act, but none collects South Dakota sales tax. 2017 S. D. 56, ¶¶ 10–11, 901 N. W. 2d 754, 759–760. Pursuant to the Act’s provisions for expeditious judicial review, South Dakota filed a declaratory judgment action against respondents in state court, seeking a declaration that the requirements of the Act are valid and applicable to respondents and an injunction requiring respondents to register for licenses to collect and remit sales tax. App. 11, 30. Respondents moved for summary judgment, arguing that the Act is unconstitutional. 901 N. W. 2d, at 759–760. South Dakota conceded that the Act cannot survive under Bellas Hess and Quill but asserted the importance, indeed the necessity, of asking this Court to review those earlier decisions in light of current economic realities. 901 N. W. 2d, at 760; see also S. B. 106, §8. The trial court granted summary judgment to respondents. App. to Pet. for Cert. 17a. The South Dakota Supreme Court affirmed. It stated: “However persuasive the State’s arguments on the merits of revisiting the issue, Quill has not been overruled [and] remains the controlling precedent on the issue of Commerce Clause limitations on interstate collection of sales and use taxes.” 901 N. W. 2d, at 761. This Court granted certiorari. 583 U. S. ___ (2018). II The Constitution grants Congress the power “[t]o regulate Commerce . . . among the several States.” Art. I, §8, cl. 3. The Commerce Clause “reflect[s] a central concern of the Framers that was an immediate reason for calling the Constitutional Convention: the conviction that in order to succeed, the new Union would have to avoid the tendencies toward economic Balkanization that had plagued relations among the Colonies and later among the States under the Articles of Confederation.” Hughes v. Oklahoma, 441 U. S. 322, 325–326 (1979). Although the Commerce Clause is written as an affirmative grant of authority to Congress, this Court has long held that in some instances it imposes limitations on the States absent congressional action. Of course, when Congress exercises its power to regulate commerce by enacting legislation, the legislation controls. Southern Pacific Co. v. Arizona ex rel. Sullivan, 325 U. S. 761, 769 (1945). But this Court has observed that “in general Congress has left it to the courts to formulate the rules” to preserve “the free flow of interstate commerce.” Id., at 770. To understand the issue presented in this case, it is instructive first to survey the general development of this Court’s Commerce Clause principles and then to review the application of those principles to state taxes. A From early in its history, a central function of this Court has been to adjudicate disputes that require interpretation of the Commerce Clause in order to determine its meaning, its reach, and the extent to which it limits state regulations of commerce. Gibbons v. Ogden, 9 Wheat. 1 (1824), began setting the course by defining the meaning of commerce. Chief Justice Marshall explained that commerce included both “the interchange of commodities” and “commercial intercourse.” Id., at 189, 193. A concurring opinion further stated that Congress had the exclusive power to regulate commerce. See id., at 236 (opinion of Johnson, J.). Had that latter submission prevailed and States been denied the power of concurrent regulation, history might have seen sweeping federal regulations at an early date that foreclosed the States from experimentation with laws and policies of their own, or, on the other hand, proposals to reexamine Gibbons’ broad definition of commerce to accommodate the necessity of allowing States the power to enact laws to implement the political will of their people. Just five years after Gibbons, however, in another opinion by Chief Justice Marshall, the Court sustained what in substance was a state regulation of interstate commerce. In Willson v. Black Bird Creek Marsh Co., 2 Pet. 245 (1829), the Court allowed a State to dam and bank a stream that was part of an interstate water system, an action that likely would have been an impermissible intrusion on the national power over commerce had it been the rule that only Congress could regulate in that sphere. See id., at 252. Thus, by implication at least, the Court indicated that the power to regulate commerce in some circumstances was held by the States and Congress concurrently. And so both a broad interpretation of interstate commerce and the concurrent regulatory power of the States can be traced to Gibbons and Willson. Over the next few decades, the Court refined the doctrine to accommodate the necessary balance between state and federal power. In Cooley v. Board of Wardens of Port of Philadelphia ex rel. Soc. for Relief of Distressed Pilots, 12 How. 299 (1852), the Court addressed local laws regulating river pilots who operated in interstate waters and guided many ships on interstate or foreign voyages. The Court held that, while Congress surely could regulate on this subject had it chosen to act, the State, too, could regulate. The Court distinguished between those subjects that by their nature “imperatively deman[d] a single uniform rule, operating equally on the commerce of the United States,” and those that “deman[d] th[e] diversity, which alone can meet . . . local necessities.” Id., at 319. Though considerable uncertainties were yet to be overcome, these precedents still laid the groundwork for the analytical framework that now prevails for Commerce Clause cases. This Court’s doctrine has developed further with time. Modern precedents rest upon two primary principles that mark the boundaries of a State’s authority to regulate interstate commerce. First, state regulations may not discriminate against interstate commerce; and second, States may not impose undue burdens on interstate commerce. State laws that discriminate against interstate commerce face “a virtually per se rule of invalidity.” Granholm v. Heald, 544 U. S. 460, 476 (2005) (internal quotation marks omitted). State laws that “regulat[e] even-handedly to effectuate a legitimate local public interest . . . will be upheld unless the burden imposed on such commerce is clearly excessive in relation to the putative local benefits.” Pike v. Bruce Church, Inc., 397 U. S. 137, 142 (1970); see also Southern Pacific, supra, at 779. Al- though subject to exceptions and variations, see, e.g., Hughes v. Alexandria Scrap Corp., 426 U. S. 794 (1976); Brown-Forman Distillers Corp. v. New York State Liquor Authority, 476 U. S. 573 (1986), these two principles guide the courts in adjudicating cases challenging state laws under the Commerce Clause. B These principles also animate the Court’s Commerce Clause precedents addressing the validity of state taxes. The Court explained the now-accepted framework for state taxation in Complete Auto Transit, Inc. v. Brady, 430 U. S. 274 (1977). The Court held that a State “may tax exclusively interstate commerce so long as the tax does not create any effect forbidden by the Commerce Clause.” Id., at 285. After all, “interstate commerce may be required to pay its fair share of state taxes.” D. H. Holmes Co. v. McNamara, 486 U. S. 24, 31 (1988). The Court will sustain a tax so long as it (1) applies to an activity with a substantial nexus with the taxing State, (2) is fairly apportioned, (3) does not discriminate against interstate commerce, and (4) is fairly related to the services the State provides. See Complete Auto, supra, at 279. Before Complete Auto, the Court had addressed a challenge to an Illinois tax that required out-of-state retailers to collect and remit taxes on sales made to consumers who purchased goods for use within Illinois. Bellas Hess, 386 U. S., at 754–755. The Court held that a mail-order company “whose only connection with customers in the State is by common carrier or the United States mail” lacked the requisite minimum contacts with the State required by both the Due Process Clause and the Commerce Clause. Id., at 758. Unless the retailer maintained a physical presence such as “retail outlets, solicitors, or property within a State,” the State lacked the power to require that retailer to collect a local use tax. Ibid. The dissent dis- agreed: “There should be no doubt that this large-scale, systematic, continuous solicitation and exploitation of the Illinois consumer market is a sufficient ‘nexus’ to require Bellas Hess to collect from Illinois customers and to remit the use tax.” Id., at 761–762 (opinion of Fortas, J., joined by Black and Douglas, JJ.). In 1992, the Court reexamined the physical presence rule in Quill. That case presented a challenge to North Dakota’s “attempt to require an out-of-state mail-order house that has neither outlets nor sales representatives in the State to collect and pay a use tax on goods purchased for use within the State.” 504 U. S., at 301. Despite the fact that Bellas Hess linked due process and the Commerce Clause together, the Court in Quill overruled the due process holding, but not the Commerce Clause holding; and it thus reaffirmed the physical presence rule. 504 U. S., at 307–308, 317–318. The Court in Quill recognized that intervening precedents, specifically Complete Auto, “might not dictate the same result were the issue to arise for the first time today.” 504 U. S., at 311. But, nevertheless, the Quill majority concluded that the physical presence rule was necessary to prevent undue burdens on interstate commerce. Id., at 313, and n. 6. It grounded the physical presence rule in Complete Auto’s requirement that a tax have a “ ‘substantial nexus’ ” with the activity being taxed. 504 U. S., at 311. Three Justices based their decision to uphold the physical presence rule on stare decisis alone. Id., at 320 (Scalia, J., joined by Kennedy and Thomas, JJ., concurring in part and concurring in judgment). Dissenting in relevant part, Justice White argued that “there is no relationship between the physical-presence/nexus rule the Court retains and Commerce Clause considerations that allegedly justify it.” Id., at 327 (opinion concurring in part and dissenting in part). III The physical presence rule has “been the target of criticism over many years from many quarters.” Direct Marketing Assn. v. Brohl, 814 F. 3d 1129, 1148, 1150–1151 (CA10 2016) (Gorsuch, J., concurring). Quill, it has been said, was “premised on assumptions that are unfounded” and “riddled with internal inconsistencies.” Rothfeld, Quill: Confusing the Commerce Clause, 56 Tax Notes 487, 488 (1992). Quill created an inefficient “online sales tax loophole” that gives out-of-state businesses an advantage. A. Laffer & D. Arduin, Pro-Growth Tax Reform and E-Fairness 1, 4 (July 2013). And “while nexus rules are clearly necessary,” the Court “should focus on rules that are appropriate to the twenty-first century, not the nineteenth.” Hellerstein, Deconstructing the Debate Over State Taxation of Electronic Commerce, 13 Harv. J. L. & Tech. 549, 553 (2000). Each year, the physical presence rule becomes further removed from economic reality and results in significant revenue losses to the States. These critiques underscore that the physical presence rule, both as first formulated and as applied today, is an incorrect interpretation of the Commerce Clause. A Quill is flawed on its own terms. First, the physical presence rule is not a necessary interpretation of the requirement that a state tax must be “applied to an activ- ity with a substantial nexus with the taxing State.” Complete Auto, 430 U. S., at 279. Second, Quill creates rather than resolves market distortions. And third, Quill im- poses the sort of arbitrary, formalistic distinction that the Court’s modern Commerce Clause precedents disavow. 1 All agree that South Dakota has the authority to tax these transactions. S. B. 106 applies to sales of “tangible personal property, products transferred electronically, or services for delivery into South Dakota.” §1 (emphasis added). “It has long been settled” that the sale of goods or services “has a sufficient nexus to the State in which the sale is consummated to be treated as a local transaction taxable by that State.” Oklahoma Tax Comm’n v. Jefferson Lines, Inc., 514 U. S. 175, 184 (1995); see also 2 C. Trost & P. Hartman, Federal Limitations on State and Local Taxation 2d §11:1, p. 471 (2003) (“Generally speaking, a sale is attributable to its destination”). The central dispute is whether South Dakota may require remote sellers to collect and remit the tax without some additional connection to the State. The Court has previously stated that “[t]he imposition on the seller of the duty to insure collection of the tax from the purchaser does not violate the [C]ommerce [C]lause.” McGoldrick v. Berwind-White Coal Mining Co., 309 U. S. 33, 50, n. 9 (1940). It is a “ ‘familiar and sanctioned device.’ ” Scripto, Inc. v. Carson, 362 U. S. 207, 212 (1960). There just must be “a substantial nexus with the taxing State.” Complete Auto, supra, at 279. This nexus requirement is “closely related,” Bellas Hess, 386 U. S., at 756, to the due process requirement that there be “some definite link, some minimum connection, between a state and the person, property or transaction it seeks to tax,” Miller Brothers Co. v. Maryland, 347 U. S. 340, 344–345 (1954). It is settled law that a business need not have a physical presence in a State to satisfy the demands of due process. Burger King Corp. v. Rudzewicz, 471 U. S. 462, 476 (1985). Although physical presence “ ‘frequently will enhance’ ” a business’ connection with a State, “ ‘it is an inescapable fact of modern commercial life that a substantial amount of business is transacted . . . [with no] need for physical presence within a State in which business is conducted.’ ” Quill, 504 U. S., at 308. Quill itself recognized that “[t]he requirements of due process are met irrespective of a corporation’s lack of physical presence in the taxing State.” Ibid. When considering whether a State may levy a tax, Due Process and Commerce Clause standards may not be identical or coterminous, but there are significant parallels. The reasons given in Quill for rejecting the physical presence rule for due process purposes apply as well to the question whether physical presence is a requisite for an out-of-state seller’s liability to remit sales taxes. Physical presence is not necessary to create a substantial nexus. The Quill majority expressed concern that without the physical presence rule “a state tax might unduly burden interstate commerce” by subjecting retailers to tax-collection obligations in thousands of different taxing jurisdictions. Id., at 313, n. 6. But the administrative costs of compliance, especially in the modern economy with its Internet technology, are largely unrelated to whether a company happens to have a physical presence in a State. For example, a business with one salesperson in each State must collect sales taxes in every jurisdiction in which goods are delivered; but a business with 500 salespersons in one central location and a website accessible in every State need not collect sales taxes on otherwise identical nationwide sales. In other words, under Quill, a small company with diverse physical presence might be equally or more burdened by compliance costs than a large remote seller. The physical presence rule is a poor proxy for the compliance costs faced by companies that do business in multiple States. Other aspects of the Court’s doctrine can better and more accurately address any potential burdens on interstate commerce, whether or not Quill’s physical presence rule is satisfied. 2 The Court has consistently explained that the Commerce Clause was designed to prevent States from engaging in economic discrimination so they would not divide into isolated, separable units. See Philadelphia v. New Jersey, 437 U. S. 617, 623 (1978). But it is “not the purpose of the [C]ommerce [C]lause to relieve those engaged in interstate commerce from their just share of state tax burden.” Complete Auto, supra, at 288 (internal quotation marks omitted). And it is certainly not the purpose of the Commerce Clause to permit the Judiciary to create market distortions. “If the Commerce Clause was intended to put businesses on an even playing field, the [physical presence] rule is hardly a way to achieve that goal.” Quill, supra, at 329 (opinion of White, J.). Quill puts both local businesses and many interstate businesses with physical presence at a competitive disadvantage relative to remote sellers. Remote sellers can avoid the regulatory burdens of tax collection and can offer de facto lower prices caused by the widespread failure of consumers to pay the tax on their own. This “guarantees a competitive benefit to certain firms simply because of the organizational form they choose” while the rest of the Court’s jurisprudence “is all about preventing discrimination between firms.” Direct Marketing, 814 F. 3d, at 1150–1151 (Gorsuch, J., concurring). In effect, Quill has come to serve as a judicially created tax shelter for businesses that decide to limit their physical presence and still sell their goods and services to a State’s consumers—something that has become easier and more prevalent as technology has advanced. Worse still, the rule produces an incentive to avoid physical presence in multiple States. Distortions caused by the desire of businesses to avoid tax collection mean that the market may currently lack storefronts, distribution points, and employment centers that otherwise would be efficient or desirable. The Commerce Clause must not prefer interstate commerce only to the point where a merchant physically crosses state borders. Rejecting the physical presence rule is necessary to ensure that artificial competitive advantages are not created by this Court’s precedents. This Court should not prevent States from collecting lawful taxes through a physical presence rule that can be satisfied only if there is an employee or a building in the State. 3 The Court’s Commerce Clause jurisprudence has “eschewed formalism for a sensitive, case-by-case analysis of purposes and effects.” West Lynn Creamery, Inc. v. Healy, 512 U. S. 186, 201 (1994). Quill, in contrast, treats economically identical actors differently, and for arbitrary reasons. Consider, for example, two businesses that sell furniture online. The first stocks a few items of inventory in a small warehouse in North Sioux City, South Dakota. The second uses a major warehouse just across the border in South Sioux City, Nebraska, and maintains a sophisticated website with a virtual showroom accessible in every State, including South Dakota. By reason of its physical presence, the first business must collect and remit a tax on all of its sales to customers from South Dakota, even those sales that have nothing to do with the warehouse. See National Geographic, 430 U. S., at 561; Scripto, Inc., 362 U. S., at 211–212. But, under Quill, the second, hypothetical seller cannot be subject to the same tax for the sales of the same items made through a pervasive Internet presence. This distinction simply makes no sense. So long as a state law avoids “any effect forbidden by the Commerce Clause,” Complete Auto, 430 U. S., at 285, courts should not rely on anachronistic formalisms to invalidate it. The basic principles of the Court’s Commerce Clause jurisprudence are grounded in functional, marketplace dynamics; and States can and should consider those realities in enacting and enforcing their tax laws. B The Quill Court itself acknowledged that the physical presence rule is “artificial at its edges.” 504 U. S., at 315. That was an understatement when Quill was decided; and when the day-to-day functions of marketing and distribution in the modern economy are considered, it is all the more evident that the physical presence rule is artificial in its entirety. Modern e-commerce does not align analytically with a test that relies on the sort of physical presence defined in Quill. In a footnote, Quill rejected the argument that “title to ‘a few floppy diskettes’ present in a State” was sufficient to constitute a “substantial nexus,” id., at 315, n. 8. But it is not clear why a single employee or a single warehouse should create a substantial nexus while “physical” aspects of pervasive modern technology should not. For example, a company with a website accessible in South Dakota may be said to have a physical presence in the State via the customers’ computers. A website may leave cookies saved to the customers’ hard drives, or customers may download the company’s app onto their phones. Or a company may lease data storage that is permanently, or even occasionally, located in South Dakota. Cf. United States v. Microsoft Corp., 584 U. S. ___ (2018) (per curiam). What may have seemed like a “clear,” “bright-line tes[t]” when Quill was written now threatens to compound the arbitrary consequences that should have been apparent from the outset. 504 U. S., at 315. The “dramatic technological and social changes” of our “increasingly interconnected economy” mean that buyers are “closer to most major retailers” than ever before—“regardless of how close or far the nearest storefront.” Direct Marketing Assn. v. Brohl, 575 U. S. ___, ___, ___ (2015) (Kennedy, J., concurring) (slip op., at 2, 3). Between targeted advertising and instant access to most consumers via any internet-enabled device, “a business may be present in a State in a meaningful way without” that presence “being physical in the traditional sense of the term.” Id., at ___ (slip op., at 3). A virtual showroom can show far more inventory, in far more detail, and with greater opportunities for consumer and seller interaction than might be possible for local stores. Yet the continuous and pervasive virtual presence of retailers today is, under Quill, simply irrelevant. This Court should not maintain a rule that ignores these substantial virtual connections to the State. C The physical presence rule as defined and enforced in Bellas Hess and Quill is not just a technical legal problem—it is an extraordinary imposition by the Judiciary on States’ authority to collect taxes and perform critical public functions. Forty-one States, two Territories, and the District of Columbia now ask this Court to reject the test formulated in Quill. See Brief for Colorado et al. as Amici Curiae. Quill’s physical presence rule intrudes on States’ reasonable choices in enacting their tax systems. And that it allows remote sellers to escape an obligation to remit a lawful state tax is unfair and unjust. It is unfair and unjust to those competitors, both local and out of State, who must remit the tax; to the consumers who pay the tax; and to the States that seek fair enforcement of the sales tax, a tax many States for many years have considered an indispensable source for raising revenue. In essence, respondents ask this Court to retain a rule that allows their customers to escape payment of sales taxes—taxes that are essential to create and secure the active market they supply with goods and services. An example may suffice. Wayfair offers to sell a vast selection of furnishings. Its advertising seeks to create an image of beautiful, peaceful homes, but it also says that “ ‘[o]ne of the best things about buying through Wayfair is that we do not have to charge sales tax.’ ” Brief for Petitioner 55. What Wayfair ignores in its subtle offer to assist in tax evasion is that creating a dream home assumes solvent state and local governments. State taxes fund the police and fire departments that protect the homes containing their customers’ furniture and ensure goods are safely delivered; maintain the public roads and municipal services that allow communication with and access to customers; support the “sound local banking institutions to support credit transactions [and] courts to ensure collection of the purchase price,” Quill, 504 U. S., at 328 (opinion of White, J.); and help create the “climate of consumer confidence” that facilitates sales, see ibid. According to respondents, it is unfair to stymie their tax-free solicitation of customers. But there is nothing unfair about requiring companies that avail themselves of the States’ benefits to bear an equal share of the burden of tax collection. Fairness dictates quite the opposite result. Helping respondents’ customers evade a lawful tax unfairly shifts to those consumers who buy from their competitors with a physical presence that satisfies Quill—even one warehouse or one salesperson—an increased share of the taxes. It is essential to public confidence in the tax system that the Court avoid creating inequitable exceptions. This is also essential to the confidence placed in this Court’s Commerce Clause decisions. Yet the physical presence rule undermines that necessary confidence by giving some online retailers an arbitrary advantage over their competitors who collect state sales taxes. In the name of federalism and free markets, Quill does harm to both. The physical presence rule it defines has limited States’ ability to seek long-term prosperity and has prevented market participants from competing on an even playing field. IV “Although we approach the reconsideration of our decisions with the utmost caution, stare decisis is not an inexorable command.” Pearson v. Callahan, 555 U. S. 223, 233 (2009) (quoting State Oil Co. v. Khan, 522 U. S. 3, 20 (1997); alterations and internal quotation marks omitted). Here, stare decisis can no longer support the Court’s prohibition of a valid exercise of the States’ sovereign power. If it becomes apparent that the Court’s Commerce Clause decisions prohibit the States from exercising their lawful sovereign powers in our federal system, the Court should be vigilant in correcting the error. While it can be conceded that Congress has the authority to change the physical presence rule, Congress cannot change the constitutional default rule. It is inconsistent with the Court’s proper role to ask Congress to address a false constitutional premise of this Court’s own creation. Courts have acted as the front line of review in this limited sphere; and hence it is important that their principles be accurate and logical, whether or not Congress can or will act in response. It is currently the Court, and not Congress, that is limiting the lawful prerogatives of the States. Further, the real world implementation of Commerce Clause doctrines now makes it manifest that the physical presence rule as defined by Quill must give way to the “far-reaching systemic and structural changes in the economy” and “many other societal dimensions” caused by the Cyber Age. Direct Marketing, 575 U. S., at ___ (Kennedy, J., concurring) (slip op., at 3). Though Quill was wrong on its own terms when it was decided in 1992, since then the Internet revolution has made its earlier error all the more egregious and harmful. The Quill Court did not have before it the present realities of the interstate marketplace. In 1992, less than 2 percent of Americans had Internet access. See Brief for Retail Litigation Center, Inc., et al. as Amici Curiae 11, and n. 10. Today that number is about 89 percent. Ibid., and n. 11. When it decided Quill, the Court could not have envisioned a world in which the world’s largest retailer would be a remote seller, S. Li, Amazon Overtakes Wal-Mart as Biggest Retailer, L. A. Times, July 24, 2015, http://www. latimes.com/business/la-fi-amazon-walmart-20150724-story.html (all Internet materials as last visited June 18, 2018). The Internet’s prevalence and power have changed the dynamics of the national economy. In 1992, mail-order sales in the United States totaled $180 billion. 504 U. S., at 329 (opinion of White, J.). Last year, e-commerce retail sales alone were estimated at $453.5 billion. Dept. of Commerce, U. S. Census Bureau News, Quarterly Retail E-Commerce Sales: 4th Quarter 2017 (CB18–21, Feb. 16, 2018). Combined with traditional remote sellers, the total exceeds half a trillion dollars. Sales Taxes Report, at 9. Since the Department of Commerce first began tracking e-commerce sales, those sales have increased tenfold from 0.8 percent to 8.9 percent of total retail sales in the United States. Compare Dept. of Commerce, U. S. Census Bureau, Retail E-Commerce Sales in Fourth Quarter 2000 (CB01–28, Feb. 16, 2001), https://www.census.gov/mrts/ www/data/pdf/00Q4.pdf, with U. S. Census Bureau News, Quarterly Retail E-Commerce Sales: 4th Quarter 2017. And it is likely that this percentage will increase. Last year, e-commerce grew at four times the rate of traditional retail, and it shows no sign of any slower pace. See ibid. This expansion has also increased the revenue shortfall faced by States seeking to collect their sales and use taxes. In 1992, it was estimated that the States were losing between $694 million and $3 billion per year in sales tax revenues as a result of the physical presence rule. Brief for Law Professors et al. as Amici Curiae 11, n. 7. Now estimates range from $8 to $33 billion. Sales Taxes Report, at 11–12; Brief for Petitioner 34–35. The South Dakota Legislature has declared an emergency, S. B. 106, §9, which again demonstrates urgency of overturning the physical presence rule. The argument, moreover, that the physical presence rule is clear and easy to apply is unsound. Attempts to apply the physical presence rule to online retail sales are proving unworkable. States are already confronting the complexities of defining physical presence in the Cyber Age. For example, Massachusetts proposed a regulation that would have defined physical presence to include making apps available to be downloaded by in-state residents and placing cookies on in-state residents’ web browsers. See 830 Code Mass. Regs. 64H.1.7 (2017). Ohio recently adopted a similar standard. See Ohio Rev. Code Ann. §5741.01(I)(2)(i) (Lexis Supp. 2018). Some States have enacted so-called “click through” nexus statutes, which define nexus to include out-of-state sellers that contract with in-state residents who refer customers for compensation. See e.g., N. Y. Tax Law Ann. §1101(b)(8)(vi) (West 2017); Brief for Tax Foundation as Amicus Curiae 20–22 (listing 21 States with similar statutes). Others still, like Colorado, have imposed notice and reporting requirements on out-of-state retailers that fall just short of actually collecting and remitting the tax. See Direct Marketing, 814 F. 3d, at 1133 (discussing Colo. Rev. Stat. §39–21–112(3.5)); Brief for Tax Foundation 24–26 (listing nine States with similar statutes). Statutes of this sort are likely to embroil courts in technical and arbitrary disputes about what counts as physical presence. Reliance interests are a legitimate consideration when the Court weighs adherence to an earlier but flawed precedent. See Kimble v. Marvel Entertainment, LLC, 576 U. S. ___, ___–___ (2015) (slip op., at 9–10). But even on its own terms, the physical presence rule as defined by Quill is no longer a clear or easily applicable standard, so arguments for reliance based on its clarity are misplaced. And, importantly, stare decisis accommodates only “legitimate reliance interest[s].” United States v. Ross, 456 U. S. 798, 824 (1982). Here, the tax distortion created by Quill exists in large part because consumers regularly fail to comply with lawful use taxes. Some remote retailers go so far as to advertise sales as tax free. See S. B. 106, §8(3); see also Brief for Petitioner 55. A business “is in no position to found a constitutional right on the practical opportunities for tax avoidance.” Nelson v. Sears, Roebuck & Co., 312 U. S. 359, 366 (1941). Respondents argue that “the physical presence rule has permitted start-ups and small businesses to use the Internet as a means to grow their companies and access a national market, without exposing them to the daunting complexity and business-development obstacles of nationwide sales tax collection.” Brief for Respondents 29. These burdens may pose legitimate concerns in some instances, particularly for small businesses that make a small volume of sales to customers in many States. State taxes differ, not only in the rate imposed but also in the categories of goods that are taxed and, sometimes, the relevant date of purchase. Eventually, software that is available at a reasonable cost may make it easier for small businesses to cope with these problems. Indeed, as the physical presence rule no longer controls, those systems may well become available in a short period of time, either from private providers or from state taxing agencies themselves. And in all events, Congress may legislate to address these problems if it deems it necessary and fit to do so. In this case, however, South Dakota affords small merchants a reasonable degree of protection. The law at issue requires a merchant to collect the tax only if it does a considerable amount of business in the State; the law is not retroactive; and South Dakota is a party to the Streamlined Sales and Use Tax Agreement, see infra at 23. Finally, other aspects of the Court’s Commerce Clause doctrine can protect against any undue burden on interstate commerce, taking into consideration the small businesses, startups, or others who engage in commerce across state lines. For example, the United States argues that tax-collection requirements should be analyzed under the balancing framework of Pike v. Bruce Church, Inc., 397 U. S. 137. Others have argued that retroactive liability risks a double tax burden in violation of the Court’s apportionment jurisprudence because it would make both the buyer and the seller legally liable for collecting and remitting the tax on a transaction intended to be taxed only once. See Brief for Law Professors et al. as Amici Curiae 7, n. 5. Complex state tax systems could have the effect of discriminating against interstate commerce. Concerns that complex state tax systems could be a burden on small business are answered in part by noting that, as discussed below, there are various plans already in place to simplify collection; and since in-state businesses pay the taxes as well, the risk of discrimination against out-of-state sellers is avoided. And, if some small businesses with only de minimis contacts seek relief from collection systems thought to be a burden, those entities may still do so under other theories. These issues are not before the Court in the instant case; but their potential to arise in some later case cannot justify retaining this artificial, anachronistic rule that deprives States of vast revenues from major businesses. For these reasons, the Court concludes that the physical presence rule of Quill is unsound and incorrect. The Court’s decisions in Quill Corp. v. North Dakota, 504 U. S. 298 (1992), and National Bellas Hess, Inc. v. Department of Revenue of Ill., 386 U. S. 753 (1967), should be, and now are, overruled. V In the absence of Quill and Bellas Hess, the first prong of the Complete Auto test simply asks whether the tax applies to an activity with a substantial nexus with the taxing State. 430 U. S., at 279. “[S]uch a nexus is established when the taxpayer [or collector] ‘avails itself of the substantial privilege of carrying on business’ in that jurisdiction.” Polar Tankers, Inc. v. City of Valdez, 557 U. S. 1, 11 (2009). Here, the nexus is clearly sufficient based on both the economic and virtual contacts respondents have with the State. The Act applies only to sellers that deliver more than $100,000 of goods or services into South Dakota or engage in 200 or more separate transactions for the delivery of goods and services into the State on an annual basis. S. B. 106, §1. This quantity of business could not have occurred unless the seller availed itself of the substantial privilege of carrying on business in South Dakota. And respondents are large, national companies that undoubtedly maintain an extensive virtual presence. Thus, the substantial nexus requirement of Complete Auto is satisfied in this case. The question remains whether some other principle in the Court’s Commerce Clause doctrine might invalidate the Act. Because the Quill physical presence rule was an obvious barrier to the Act’s validity, these issues have not yet been litigated or briefed, and so the Court need not resolve them here. That said, South Dakota’s tax system includes several features that appear designed to prevent discrimination against or undue burdens upon interstate commerce. First, the Act applies a safe harbor to those who transact only limited business in South Dakota. Second, the Act ensures that no obligation to remit the sales tax may be applied retroactively. S. B. 106, §5. Third, South Dakota is one of more than 20 States that have adopted the Streamlined Sales and Use Tax Agreement. This system standardizes taxes to reduce administrative and compliance costs: It requires a single, state level tax administration, uniform definitions of products and services, simplified tax rate structures, and other uniform rules. It also provides sellers access to sales tax administration software paid for by the State. Sellers who choose to use such software are immune from audit liability. See App. 26–27. Any remaining claims regarding the application of the Commerce Clause in the absence of Quill and Bellas Hess may be addressed in the first instance on remand. The judgment of the Supreme Court of South Dakota is vacated, and the case is remanded for further proceedings not inconsistent with this opinion. It is so ordered.

584.US.2017_16-1432

The legal system has long used default rules to resolve estate litigation in a way that conforms to decedents’ presumed intent. In 2002, Minnesota enacted a statute establishing one such default rule. The statute provides that “the dissolution or annulment of a marriage revokes any revocable . . . beneficiary designation . . . made by an individual to the individual’s former spouse.” Minn. Stat. §524.2–804, subd. 1. Under the statute, if one spouse has made the other the beneficiary of a life insurance policy or similar asset, their divorce automatically revokes that designation so that the insurance proceeds will instead go to the contingent beneficiary or the policyholder’s estate upon his death. The law does this on the theory that the policyholder would want that result. But if he does not, he may rename the ex-spouse as beneficiary. Mark Sveen and respondent Kaye Melin were married in 1997. The next year, Sveen purchased a life insurance policy, naming Melin as the primary beneficiary and designating his two children from a prior marriage, petitioners Ashley and Antone Sveen, as contingent beneficiaries. The Sveen-Melin marriage ended in 2007, but the divorce decree made no mention of the insurance policy and Sveen took no action to revise his beneficiary designations. After Sveen passed away in 2011, Melin and the Sveen children made competing claims to the insurance proceeds. The Sveens argued that under Minnesota’s revocation-on-divorce law, their father’s divorce canceled Melin’s beneficiary designation, leaving them as the rightful recipients. Melin claimed that because the law did not exist when the policy was purchased and she was named as the primary beneficiary, applying the later-enacted law to the policy violates the Constitution’s Contracts Clause. The District Court awarded the insurance money to the Sveens, but the Eighth Circuit reversed, holding that the retroactive application of Minnesota’s law violates the Contracts Clause. Held: The retroactive application of Minnesota’s statute does not violate the Contracts Clause. That Clause restricts the power of States to disrupt contractual arrangements, but it does not prohibit all laws affecting pre-existing contracts, see El Paso v. Simmons, 379 U. S. 497, 506–507. The two-step test for determining when such a law crosses the constitutional line first asks whether the state law has “operated as a substantial impairment of a contractual relationship.” Allied Structural Steel Co. v. Spannaus, 438 U. S. 234, 244. In answering that question, the Court has considered the extent to which the law undermines the contractual bargain, interferes with a party’s reasonable expectations, and prevents the party from safeguarding or reinstating his rights. See id., at 246; El Paso, 379 U. S., at 514–515; Texaco, Inc. v. Short, 454 U. S. 516, 531. If such factors show a substantial impairment, the inquiry turns to whether the state law is drawn in an “appropriate” and “reasonable” way to advance “a significant and legitimate public purpose.” Energy Reserves Group, Inc. v. Kansas Power & Light Co., 459 U. S. 400, 411–412. The Court stops after the first step here, because three aspects of Minnesota’s law, taken together, show that the law does not substantially impair pre-existing contractual arrangements. First, the law is designed to reflect a policyholder’s intent—and so to support, rather than impair, the contractual scheme. It applies a prevalent legislative presumption that a divorcee would not want his former partner to benefit from his life insurance policy and other will substitutes. Thus the law often honors, not undermines, the intent of the only contracting party to care about the beneficiary term. Second, the law is unlikely to disturb any policyholder’s expectations at the time of contracting, because an insured cannot reasonably rely on a beneficiary designation staying in place after a divorce. Divorce courts have wide discretion to divide property upon dissolution of a marriage, including by revoking spousal beneficiary designations in life insurance policies or by mandating that such designations remain. Because a life insurance purchaser cannot know what will happen to that policy in the event of a divorce, his reliance interests are next to nil. And that fact cuts against providing protection under the Contracts Clause. Last, the law supplies a mere default rule, which the policyholder can undo in a moment. If the law’s presumption about what an insured wants after divorcing is wrong, the insured may overthrow it simply by sending a change-of-beneficiary form to his insurer. This Court has long held that laws imposing such minimal paperwork burdens do not violate the Contracts Clause. It has repeatedly sustained so-called recording statutes, which extinguish contractual interests unless timely recorded at government offices. See Jackson v. Lamphire, 3 Pet. 280; Vance v. Vance, 108 U. S. 514; Texaco, Inc. v. Short, 454 U. S. 516. The Court has also upheld laws mandating other kinds of notifications or filings against Contracts Clause attack. See Curtis v. Whitney, 13 Wall. 68; Gilfillan v. Union Canal Co. of Pa., 109 U. S. 401; Conley v. Barton, 260 U. S. 677. The Minnesota law places no greater obligation on a contracting party than these laws—while imposing a lesser penalty for noncompliance. Filing a change-of-beneficiary form is as easy as satisfying the paperwork requirements that this Court’s prior cases approved. And if an insured wants his ex-spouse to stay as beneficiary but does not send in his form, the result is only that the insurance money is redirected to his contingent beneficiaries, not that his contractual rights are extinguished. Pp. 6–14. 853 F. 3d 410, reversed and remanded. Kagan, J., delivered the opinion of the Court, in which Roberts, C. J., and Kennedy, Thomas, Ginsburg, Breyer, Alito, and Sotomayor, JJ., joined. Gorsuch, J., filed a dissenting opinion.

A Minnesota law provides that “the dissolution or annulment of a marriage revokes any revocable[ ] beneficiary designation[ ] made by an individual to the individual’s former spouse.” Minn. Stat. §524.2–804, subd. 1 (2016). That statute establishes a default rule for use when Minnesotans divorce. If one spouse has made the other the beneficiary of a life insurance policy or similar asset, their divorce automatically revokes that designation—on the theory that the policyholder would want that result. But if he does not, the policyholder may rename the ex-spouse as beneficiary. We consider here whether applying Minnesota’s automatic-revocation rule to a beneficiary designation made before the statute’s enactment violates the Contracts Clause of the Constitution. We hold it does not. I All good trust-and-estate lawyers know that “[d]eath is not the end; there remains the litigation over the estate.” 8 The Collected Works of Ambrose Bierce 365 (1911). That epigram, beyond presaging this case, helps explain the statute at its center. The legal system has long used default rules to resolve estate litigation in a way that conforms to decedents’ presumed intent. At common law, for example, marriage automatically revoked a woman’s prior will, while marriage and the birth of a child revoked a man’s. See 4 J. Kent, Commentaries on American Law 507, 512 (1830). The testator could then revive the old will or execute a new one. But if he (or she) did neither, the laws of intestate succession (generally prioritizing children and current spouses) would control the estate’s distribution. See 95 C. J. S., Wills §448, pp. 409–410 (2011); R. Sitkoff & J. Dukeminier, Wills, Trusts, and Estates 63 (10th ed. 2017). Courts reasoned that the average person would prefer that allocation to the one in the old will, given the intervening life events. See T. Atkinson, Handbook of the Law of Wills 423 (2d ed. 1953). If he’d only had the time, the thought went, he would have replaced that will himself. Changes in society brought about changes in the laws governing revocation of wills. In addition to removing gender distinctions, most States abandoned the common-law rule canceling whole wills executed before a marriage or birth. In its place, they enacted statutes giving a new spouse or child a specified share of the decedent’s estate while leaving the rest of his will intact. See Sitkoff & Dukeminier, Wills, Trusts, and Estates, at 240. But more important for our purposes, climbing divorce rates led almost all States by the 1980s to adopt another kind of automatic-revocation law. So-called revocation-on-divorce statutes treat an individual’s divorce as voiding a testamentary bequest to a former spouse. Like the old common-law rule, those laws rest on a “judgment about the typical testator’s probable intent.” Id., at 239. They presume, in other words, that the average Joe does not want his ex inheriting what he leaves behind. Over time, many States extended their revocation-on-divorce statutes from wills to “will substitutes,” such as revocable trusts, pension accounts, and life insurance policies. See Langbein, The Nonprobate Revolution and the Future of the Law of Succession, 97 Harv. L. Rev. 1108, 1109 (1984) (describing nonprobate assets). In doing so, States followed the lead of the Uniform Probate Code, a model statute amended in 1990 to include a provision revoking on divorce not just testamentary bequests but also beneficiary designations to a former spouse. See §§2–804(a)(1), (b)(1), 8 U. L. A. 330, 330–331 (2013). The new section, the drafters wrote, aimed to “unify the law of probate and nonprobate transfers.” §2–804, Comment, id., at 333. The underlying idea was that the typical decedent would no more want his former spouse to benefit from his pension plan or life insurance than to inherit under his will. A wealth transfer was a wealth transfer—and a former spouse (as compared with, say, a current spouse or child) was not likely to be its desired recipient. So a decedent’s failure to change his beneficiary probably resulted from “inattention,” not “intention.” Statement of the Joint Editorial Bd. for Uniform Probate Code, 17 Am. College Trust & Est. Counsel 184 (1991). Agreeing with that assumption, 26 States have by now adopted revocation-on-divorce laws substantially similar to the Code’s.[1] Minnesota is one. Under prior Minnesota law, a divorce alone did not affect a beneficiary designation—but a particular divorce decree could do so. Take first the simple case: Joe names his wife Ann as beneficiary of his insurance policy, later gets divorced, but never changes the designation. Upon his death, Ann would receive the insurance proceeds—even if Joe had just forgotten to redirect the money. In other words, the insurance contract’s beneficiary provision would govern after the divorce, exactly as it would have before. See Larsen v. Northwestern Nat. Life Ins. Co., 463 N. W. 2d 777, 779 (Minn. App. 1990). But now introduce a complication, in the form of a court addressing a spousal designation in a divorce decree. In Minnesota, as across the nation, divorce courts have always had “broad discretion in dividing property upon dissolution of a marriage.” Maurer v. Maurer, 623 N. W. 2d 604, 606 (Minn. 2001); see 24 Am. Jur. 2d, Divorce and Separation §456 (2008). In exercising that power, a court could revoke a beneficiary designation to a soon-to-be ex-spouse; or conversely, a court could mandate that the old designation remain. See, e.g., Paul v. Paul, 410 N. W. 2d 329, 330 (Minn. App. 1987); O’Brien v. O’Brien, 343 N. W. 2d 850, 853 (Minn. 1984). Either way, the court, rather than the insured, would decide whether the ex-spouse would stay the beneficiary. In contrast to the old law, Minnesota’s new revocation-on-divorce statute starts from another baseline: the cancellation, rather than continuation, of a beneficiary designation. Enacted in 2002 to track the Code, the law provides that “the dissolution or annulment of a marriage revokes any revocable[ ] disposition, beneficiary designation, or appointment of property made by an individual to the individual’s former spouse in a governing instrument.” Minn. Stat. §524.2–804, subd. 1. The term “governing instrument” is defined to include an “insurance or annuity policy,” along with a will and other will substitutes. §524.1–201. So now when Joe and Ann divorce, the clause naming Ann as Joe’s insurance beneficiary is automatically revoked. If nothing else occurs before Joe’s death, his insurance proceeds go to any contingent beneficiary named in the policy (perhaps his daughter Emma) or, failing that, to his estate. See §524.2–804, subd. 2. Something else, however, may well happen. As under Minnesota’s former law, a divorce decree may alter the natural state of things. So in our example, the court could direct that Ann remain as Joe’s insurance beneficiary, despite the normal revocation rule. See §524.2–804, subd. 1 (providing that a “court order” trumps the rule). And just as important, the policyholder himself may step in to override the revocation. Joe, for example, could agree to a marital settlement ensuring Ann’s continued status as his beneficiary. See ibid. (providing that such an agreement controls). Or else, and more simply, he could notify his insurance company at any time that he wishes to restore Ann to that position. But enough of our hypothetical divorcees: It is time they give way to Mark Sveen and Kaye Melin, whose marriage and divorce led to this case. In 1997, Sveen and Melin wed. The next year, Sveen purchased a life insurance policy. He named Melin as the primary beneficiary, while designating his two children from a prior marriage, Ashley and Antone Sveen, as the contingent beneficiaries. The Sveen-Melin marriage ended in 2007. The divorce decree made no mention of the insurance policy. And Sveen took no action, then or later, to revise his beneficiary designations. In 2011, he passed away. In this action, petitioners the Sveen children and respondent Melin make competing claims to the insurance proceeds. The Sveens contend that under Minnesota’s revocation-on-divorce law, their father’s divorce canceled Melin’s beneficiary designation and left the two of them as the rightful recipients. Melin notes in reply that the Minnesota law did not yet exist when her former husband bought his insurance policy and named her as the primary beneficiary. And she argues that applying the later-enacted law to the policy would violate the Constitution’s Contracts Clause, which prohibits any state “Law impairing the Obligation of Contracts.” Art. I, §10, cl. 1. The District Court rejected Melin’s argument and awarded the insurance money to the Sveens. See Civ. No. 14–5015 (D Minn., Jan. 7, 2016), App. to Pet. for Cert. 9a–16a. But the Court of Appeals for the Eighth Circuit reversed. It held that a “revocation-upon-divorce statute like [Minnesota’s] violates the Contract Clause when applied retroactively.” 853 F. 3d 410, 412 (2017). We granted certiorari, 583 U. S. ___ (2017), to resolve a split of authority over whether the Contracts Clause prevents a revocation-on-divorce law from applying to a pre-existing agreement’s beneficiary designation.[2] We now reverse the decision below. II The Contracts Clause restricts the power of States to disrupt contractual arrangements. It provides that “[n]o state shall . . . pass any . . . Law impairing the Obligation of Contracts.” U. S. Const., Art. I, §10, cl. 1. The origins of the Clause lie in legislation enacted after the Revolutionary War to relieve debtors of their obligations to creditors. See Keystone Bituminous Coal Assn. v. DeBenedictis, 480 U. S. 470, 502–503 (1987). But the Clause applies to any kind of contract. See Allied Structural Steel Co. v. Spannaus, 438 U. S. 234, 244–245, n. 16 (1978). That includes, as here, an insurance policy. At the same time, not all laws affecting pre-existing contracts violate the Clause. See El Paso v. Simmons, 379 U. S. 497, 506–507 (1965). To determine when such a law crosses the constitutional line, this Court has long applied a two-step test. The threshold issue is whether the state law has “operated as a substantial impairment of a contractual relationship.” Allied Structural Steel Co., 438 U. S., at 244. In answering that question, the Court has considered the extent to which the law undermines the contractual bargain, interferes with a party’s reasonable expectations, and prevents the party from safeguarding or reinstating his rights. See id., at 246; El Paso, 379 U. S., at 514–515; Texaco, Inc. v. Short, 454 U. S. 516, 531 (1982). If such factors show a substantial impairment, the inquiry turns to the means and ends of the legislation. In particular, the Court has asked whether the state law is drawn in an “appropriate” and “reasonable” way to advance “a significant and legitimate public purpose.” Energy Reserves Group, Inc. v. Kansas Power & Light Co., 459 U. S. 400, 411–412 (1983). Here, we may stop after step one because Minnesota’s revocation-on-divorce statute does not substantially impair pre-existing contractual arrangements. True enough that in revoking a beneficiary designation, the law makes a significant change. As Melin says, the “whole point” of buying life insurance is to provide the proceeds to the named beneficiary. Brief for Respondent 16. But three aspects of Minnesota’s law, taken together, defeat Melin’s argument that the change it effected “severely impaired” her ex-husband’s contract. Ibid. First, the statute is designed to reflect a policyholder’s intent—and so to support, rather than impair, the contractual scheme. Second, the law is unlikely to disturb any policyholder’s expectations because it does no more than a divorce court could always have done. And third, the statute supplies a mere default rule, which the policyholder can undo in a moment. Indeed, Minnesota’s revocation statute stacks up well against laws that this Court upheld against Contracts Clause challenges as far back as the early 1800s.[3] We now consider in detail each of the features that make this so. To begin, the Minnesota statute furthers the policyholder’s intent in many cases—indeed, the drafters reasonably thought in the typical one. As earlier described, legislatures have long made judgments about a decedent’s likely testamentary intent after large life changes—a marriage, a birth, or a divorce. See supra, at 2. And on that basis, they have long enacted statutes revoking earlier-made wills by operation of law. Legislative presumptions about divorce are now especially prevalent—probably because they accurately reflect the intent of most divorcing parties. Although there are exceptions, most divorcees do not aspire to enrich their former partners. (And that is true even when an ex-spouse has custody of shared children, given the many ways to provide them with independent support.) The Minnesota statute (like the model code it tracked) applies that understanding to beneficiary designations in life insurance policies and other will substitutes. See supra, at 3–5. Melin rightly notes that this extension raises a brand-new constitutional question because “an insurance policy is a contract under the Contracts Clause, and a will is not.” Brief for Respondent 44 (internal quotation marks omitted). But in answering that question, it matters that the old legislative presumption equally fits the new context: A person would as little want his ex-spouse to benefit from his insurance as to collect under his will. Or said otherwise, the insured’s failure to change the beneficiary after a divorce is more likely the result of neglect than choice. And that means the Minnesota statute often honors, not undermines, the intent of the only contracting party to care about the beneficiary term. The law no doubt changes how the insurance contract operates. But does it impair the contract? Quite the opposite for lots of policyholders. And even when presumed and actual intent diverge, the Minnesota law is unlikely to upset a policyholder’s expectations at the time of contracting. That is because an insured cannot reasonably rely on a beneficiary designation remaining in place after a divorce. As noted above, divorce courts have wide discretion to divide property between spouses when a marriage ends. See supra, at 4. The house, the cars, the sporting equipment are all up for grabs. See Judgment and Decree in 14–cv–5015 (D Minn.), p. 51 (awarding Melin, among other things, a snowmobile and all-terrain vehicle). And (what matters here) so too are the spouses’ life insurance policies, with their beneficiary provisions. Although not part of the Sveen-Melin divorce decree, they could have been; as Melin acknowledges, they sometimes are. See supra, at 4; Brief for Respondent 38. Melin counters that the Contracts Clause applies only to legislation, not to judicial decisions. See id., at 38–39; see also post, at 9 (Gorsuch, J., dissenting). That is true, but of no moment. The power of divorce courts over insurance policies is relevant here because it affects whether a party can reasonably expect a beneficiary designation to survive a marital breakdown. We venture to guess that few people, when purchasing life insurance, give a thought to what will happen in the event of divorce. But even if someone out there does, he can conclude only that . . . he cannot possibly know. So his reliance interests are next to nil. And as this Court has held before, that fact cuts against providing protection under the Contracts Clause. See, e.g., El Paso, 379 U. S., at 514–515. Finally, a policyholder can reverse the effect of the Minnesota statute with the stroke of a pen. The law puts in place a presumption about what an insured wants after divorcing. But if the presumption is wrong, the insured may overthrow it. And he may do so by the simple act of sending a change-of-beneficiary form to his insurer. (Or if he wants to commit himself forever, like Ulysses binding himself to the mast, he may agree to a divorce settlement continuing his ex-spouse’s beneficiary status. See supra, at 5.) That action restores his former spouse to the position she held before the divorce—and in so doing, cancels the state law’s operation. The statute thus reduces to a paperwork requirement (and a fairly painless one, at that): File a form and the statutory default rule gives way to the original beneficiary designation. In cases going back to the 1800s, this Court has held that laws imposing such minimal paperwork burdens do not violate the Contracts Clause. One set of decisions addresses so-called recording statutes, which extinguish contractual interests unless timely recorded at government offices. In Jackson v. Lamphire, 3 Pet. 280 (1830), for example, the Court rejected a Contracts Clause challenge to a New York law granting title in property to a later rather than earlier purchaser whenever the earlier had failed to record his deed. It made no difference, the Court held, whether the unrecorded deed was “dated before or after the passage” of the statute; in neither event did the law’s modest recording condition “impair[ ] the obligation of contracts.” Id., at 290. Likewise, in Vance v. Vance, 108 U. S. 514 (1883), the Court upheld a statute rendering unrecorded mortgages unenforceable against third parties—even when the mortgages predated the law. We reasoned that the law gave “due regard to existing contracts” because it demanded only that the mortgagee make a “public registration,” and gave him several months to do so. Id., at 517, 518. And more recently, in Texaco, Inc. v. Short, 454 U. S. 516 (1982), the Court held that a statute terminating pre-existing mineral interests unless the owner filed a “statement of claim” in a county office did not “unconstitutionally impair” a contract. Id., at 531. The filing requirement was “minimal,” we explained, and compliance with it would effectively “safeguard any contractual obligations or rights.” Ibid. So too, the Court has long upheld against Contracts Clause attack laws mandating other kinds of notifications or filings. In Curtis v. Whitney, 13 Wall. 68 (1872), for example, the Court approved a statute retroactively affecting buyers of “certificates” for land offered at tax sales. The law required the buyer to notify the tax-delinquent property owner, who could then put up the funds necessary to prevent the land’s final sale. If the buyer failed to give the notice, he could not take the land—and if he provided the notice, his chance of gaining the land declined. Still, the Court made short work of the Contracts Clause claim. Not “every statute which affects the value of a contract,” the Court stated, “impair[s] its obligation.” Id., at 70. Because the law’s notice rule was “easy [to] compl[y] with,” it did not raise a constitutional problem. Id., at 71. Similarly, in Gilfillan v. Union Canal Co. of Pa., 109 U. S. 401 (1883), the Court sustained a state law providing that an existing bondholder’s failure to reject a settlement proposal in writing would count as consent to the deal. The law operated to reduce the interest received by an investor who did not respond. Yet the Court rebuffed the ensuing Contracts Clause suit. “If [the bondholder did] not wish to abandon his old rights and accept the new,” the Court explained, “all he ha[d] to do [was] to say so in writing.” Id., at 406. And one last: In Conley v. Barton, 260 U. S. 677 (1923), the Court held that the Contracts Clause did not bar a State from compelling existing mortgagees to complete affidavits before finally foreclosing on properties. The law effectively added a paperwork requirement to the mortgage contracts’ foreclosure terms. But the Court said it was “only [a] condition, easily complied with, which the law, for its purposes, requires.” Id., at 681. The Minnesota statute places no greater obligation on a contracting party—while imposing a lesser penalty for noncompliance. Even supposing an insured wants his life insurance to benefit his ex-spouse, filing a change-of-beneficiary form with an insurance company is as “easy” as, say, providing a landowner with notice or recording a deed. Curtis, 13 Wall., at 71. Here too, with only “minimal” effort, a person can “safeguard” his contractual preferences. Texaco, 454 U. S., at 531. And here too, if he does not “wish to abandon his old rights and accept the new,” he need only “say so in writing.” Gilfillan, 109 U. S., at 406. What’s more, if the worst happens—if he wants his ex-spouse to stay as beneficiary but does not send in his form—the consequence pales in comparison with the losses incurred in our earlier cases. When a person ignored a recording obligation, for example, he could forfeit the sum total of his contractual rights—just ask the plaintiffs in Jackson and Vance. But when a policyholder in Minnesota does not redesignate his ex-spouse as beneficiary, his right to insurance does not lapse; the upshot is just that his contingent beneficiaries (here, his children) receive the money. See supra, at 5. That redirection of proceeds is not nothing; but under our precedents, it gives the policyholder—who, again, could have “easily” and entirely escaped the law’s effect—no right to complain of a Contracts Clause violation. Conley, 260 U. S., at 681. In addressing those precedents, Melin mainly urges us to distinguish between two ways a law can affect a contract. The Minnesota law, Melin claims, “operate[s] on the contract itself” by “directly chang[ing] an express term” (the insured’s beneficiary designation). Brief for Respondent 51; Tr. of Oral Arg. 57. In contrast, Melin continues, the recording statutes “impose[ ] a consequence” for failing to abide by a “procedural” obligation extraneous to the agreement (the State’s recording or notification rule). Brief for Respondent 51; Tr. of Oral Arg. 58. The difference, in her view, parallels the line between rights and remedies: The Minnesota law explicitly alters a person’s entitlement under the contract, while the recording laws interfere with his ability to enforce that entitlement against others. See Tr. of Oral Arg. 57–59; see also post, at 9–10 (Gorsuch, J., dissenting). But we see no meaningful distinction among all these laws. The old statutes also “act[ed] on the contract” in a significant way. Tr. of Oral Arg. 59. They added a paperwork obligation nowhere found in the original agreement—“record the deed,” say, or “notify the landowner.” And they informed a contracting party that unless he complied, he could not gain the benefits of his bargain. Or viewed conversely, the Minnesota statute also “impose[s] a consequence” for not satisfying a burden outside the contract. Brief for Respondent 51. For as we have shown, that law overrides a beneficiary designation only when the insured fails to send in a form to his insurer. See supra, at 10. Of course, the statutes (both old and new) vary in their specific mechanisms. But they all make contract benefits contingent on some simple filing—or more positively spun, enable a party to safeguard those benefits by taking an action. And that feature is what the Court, again and again, has found dispositive. Nor does Melin’s attempt to distinguish the cases gain force when framed in terms of rights and remedies. First, not all the old statutes, as a formal matter, confined the consequence of noncompliance to the remedial sphere. In Gilfillan, for example, the result of failing to give written consent to a settlement was to diminish the interest rate a bondholder got, not to prevent him from enforcing a claim against others. And second, even when the consequence formally related to enforcement—for example, precluding an earlier purchaser from contesting a later one’s title—the laws in fact wiped out substantive rights. Failure to record or notify, as noted earlier, would mean that the contracting party lost what (according to his agreement) was his land or mortgage or mineral interest. See supra, at 12–13. In Texaco, we replied to an argument like Melin’s by saying that when the results of “eliminating a remedy” and “extinguishing a right” are “identical,” the Contracts Clause “analysis is the same.” 454 U. S., at 528; see El Paso, 379 U. S., at 506–507. That statement rebuts Melin’s claim too. Once again: Just like Minnesota’s statute, the laws discussed above hinged core contractual benefits on compliance with noncontractual paperwork burdens. When all is said and done, that likeness controls. For those reasons, we reverse the judgment of the Court of Appeals and remand the case for further proceedings consistent with this opinion. It is so ordered. Notes 1 See Ala. Code §30–4–17 (2016); Alaska Stat. §13.12.804 (2016); Ariz. Rev. Stat. Ann. §14–2804 (2012); Colo. Rev. Stat. §15–11–804 (2017); Fla. Stat. §732.703 (2017); Haw. Rev. Stat. §560:2–804 (2006); Idaho Code Ann. §15–2–804 (2017 Cum. Supp.); Iowa Code §598.20A (2017); Mass. Gen. Laws, ch. 190B, §2–804 (2016); Mich. Comp. Laws Ann. §700.2807 (West 2018 Cum. Supp.); Minn. Stat. §524.2–804 subd. 1 (2016); Mont. Code Ann. §72–2–814 (2017); Nev. Rev. Stat. §111.781 (2015); N. J. Stat. Ann. §3B:3–14 (West 2007); N. M. Stat. Ann. §45–2–804 (2014); N. Y. Est., Powers & Trusts Law Ann. §5–1.4 (West 2018 Cum. Supp.); N. D. Cent. Code Ann. §30.1–10–04 (2010); Ohio Rev. Code Ann. §5815.33 (Lexis 2017); 20 Pa. Stat. and Cons. Stat. Ann. §6111.2 (2010); S. C. Code Ann. §62–2–507 (2017 Cum. Supp.); S. D. Codified Laws §29A–2–804 (2004); Tex. Fam. Code Ann. §9.301 (West 2006); Utah Code §75–2–804 (Supp. 2017); Va. Code Ann. §20–111.1 (2016); Wash. Rev. Code §11.07.010 (2016); Wis. Stat. §854.15 (2011). 2 Compare 853 F.3d 410, 414 (CA8 2017) (case below) (yes, it does); Parsonese v. Midland Nat. Ins. Co., 550 Pa. 423, 434, 706 A. 2d 814, 819 (1998) (same), with Lazar v. Kroncke, 862 F. 3d 1186, 1199–1200 (CA9 2017) (no, it does not); Stillman v. Teachers Ins. & Annuity Assn. College Retirement Equities Fund, 343 F.3d 1311, 1322 (CA10 2003) (same); In re Estate of DeWitt, 54 P. 3d 849, 859–860 (Colo. 2002) (same). 3 Because that is true, we have no occasion to address Melin’s contention that we should abandon our two-step Contracts Clause test to whatever extent it departs from the Clause’s original meaning and earliest applications. See Brief for Respondent 6–10, 18–33. Part of Melin’s argument focuses on the back half of the test, which we do not reach today. Another part claims that the front half goes wrong in exempting insubstantial impairments from the Clause’s reach. But as we explain below, see infra, at 10–12, the Court has always recognized that some laws affect contracts without violating the Contracts Clause. See, e.g., Curtis v. Whitney, 13 Wall. 68, 70 (1872) (“No[t] every statute which affects the value of a contract impair[s] its obligation”). And in particular, the Court has always approved statutes like this one, which enable a party with only minimal effort to protect his original contract rights against the law’s operation. See, e.g., Jackson v. Lamphire, 3 Pet. 280, 290 (1830). So this case presents no clash, of the kind Melin says we should resolve, between the Court’s two-step test and any older approach to applying the Contracts Clause.

585.US.2017_17-965

In September 2017, the President issued Proclamation No. 9645, seeking to improve vetting procedures for foreign nationals traveling to the United States by identifying ongoing deficiencies in the information needed to assess whether nationals of particular countries present a security threat. The Proclamation placed entry restrictions on the nationals of eight foreign states whose systems for managing and sharing information about their nationals the President deemed inadequate. Foreign states were selected for inclusion based on a review undertaken pursuant to one of the President’s earlier Executive Orders. As part of that review, the Department of Homeland Security (DHS), in consultation with the State Department and intelligence agencies, developed an information and risk assessment “baseline.” DHS then collected and evaluated data for all foreign governments, identifying those having deficient information-sharing practices and presenting national security concerns, as well as other countries “at risk” of failing to meet the baseline. After a 50-day period during which the State Department made diplomatic efforts to encourage foreign governments to improve their practices, the Acting Secretary of Homeland Security concluded that eight countries—Chad, Iran, Iraq, Libya, North Korea, Syria, Venezuela, and Yemen—remained deficient. She recommended entry restrictions for certain nationals from all of those countries but Iraq, which had a close cooperative relationship with the U. S. She also recommended including Somalia, which met the information-sharing component of the baseline standards but had other special risk factors, such as a significant terrorist presence. After consulting with multiple Cabinet members, the President adopted the recommendations and issued the Proclamation. Invoking his authority under 8 U. S. C. §§1182(f) and 1185(a), he determined that certain restrictions were necessary to “prevent the entry of those foreign nationals about whom the United States Government lacks sufficient information” and “elicit improved identity-management and information-sharing protocols and practices from foreign governments.” The Proclamation imposes a range of entry restrictions that vary based on the “distinct circumstances” in each of the eight countries. It exempts lawful permanent residents and provides case-by-case waivers under certain circumstances. It also directs DHS to assess on a continuing basis whether the restrictions should be modified or continued, and to report to the President every 180 days. At the completion of the first such review period, the President determined that Chad had sufficiently improved its practices, and he accordingly lifted restrictions on its nationals. Plaintiffs—the State of Hawaii, three individuals with foreign relatives affected by the entry suspension, and the Muslim Association of Hawaii—argue that the Proclamation violates the Immigration and Nationality Act (INA) and the Establishment Clause. The District Court granted a nationwide preliminary injunction barring enforcement of the restrictions. The Ninth Circuit affirmed, concluding that the Proclamation contravened two provisions of the INA: §1182(f), which authorizes the President to “suspend the entry of all aliens or any class of aliens” whenever he “finds” that their entry “would be detrimental to the interests of the United States,” and §1152(a)(1)(A), which provides that “no person shall . . . be discriminated against in the issuance of an immigrant visa because of the person’s race, sex, nationality, place of birth, or place of residence.” The court did not reach the Establishment Clause claim. Held: 1. This Court assumes without deciding that plaintiffs’ statutory claims are reviewable, notwithstanding consular nonreviewability or any other statutory nonreviewability issue. See Sale v. Haitian Centers Council, Inc., 509 U. S. 155. Pp. 8–9. 2. The President has lawfully exercised the broad discretion granted to him under §1182(f) to suspend the entry of aliens into the United States. Pp. 9–24. (a) By its terms, §1182(f) exudes deference to the President in every clause. It entrusts to the President the decisions whether and when to suspend entry, whose entry to suspend, for how long, and on what conditions. It thus vests the President with “ample power” to impose entry restrictions in addition to those elsewhere enumerated in the INA. Sale, 509 U. S., at 187. The Proclamation falls well within this comprehensive delegation. The sole prerequisite set forth in §1182(f) is that the President “find[ ]” that the entry of the covered aliens “would be detrimental to the interests of the United States.” The President has undoubtedly fulfilled that requirement here. He first ordered DHS and other agencies to conduct a comprehensive evaluation of every single country’s compliance with the information and risk assessment baseline. He then issued a Proclamation with extensive findings about the deficiencies and their impact. Based on that review, he found that restricting entry of aliens who could not be vetted with adequate information was in the national interest. Even assuming that some form of inquiry into the persuasiveness of the President’s findings is appropriate, but see Webster v. Doe, 486 U. S. 592, 600, plaintiffs’ attacks on the sufficiency of the findings cannot be sustained. The 12-page Proclamation is more detailed than any prior order issued under §1182(f). And such a searching inquiry is inconsistent with the broad statutory text and the deference traditionally accorded the President in this sphere. See, e.g., Sale, 509 U. S., at 187–188. The Proclamation comports with the remaining textual limits in §1182(f). While the word “suspend” often connotes a temporary deferral, the President is not required to prescribe in advance a fixed end date for the entry restriction. Like its predecessors, the Proclamation makes clear that its “conditional restrictions” will remain in force only so long as necessary to “address” the identified “inadequacies and risks” within the covered nations. Finally, the Proclamation properly identifies a “class of aliens” whose entry is suspended, and the word “class” comfortably encompasses a group of people linked by nationality. Pp. 10–15. (b) Plaintiffs have not identified any conflict between the Proclamation and the immigration scheme reflected in the INA that would implicitly bar the President from addressing deficiencies in the Nation’s vetting system. The existing grounds of inadmissibility and the narrow Visa Waiver Program do not address the failure of certain high-risk countries to provide a minimum baseline of reliable information. Further, neither the legislative history of §1182(f) nor historical practice justifies departing from the clear text of the statute. Pp. 15–20. (c) Plaintiffs’ argument that the President’s entry suspension violates §1152(a)(1)(A) ignores the basic distinction between admissibility determinations and visa issuance that runs throughout the INA. Section 1182 defines the universe of aliens who are admissible into the United States (and therefore eligible to receive a visa). Once §1182 sets the boundaries of admissibility, §1152(a)(1)(A) prohibits discrimination in the allocation of immigrant visas based on nationality and other traits. Had Congress intended in §1152(a)(1)(A) to constrain the President’s power to determine who may enter the country, it could have chosen language directed to that end. Common sense and historical practice confirm that §1152(a)(1)(A) does not limit the President’s delegated authority under §1182(f). Presidents have repeatedly exercised their authority to suspend entry on the basis of nationality. And on plaintiffs’ reading, the President would not be permitted to suspend entry from particular foreign states in response to an epidemic, or even if the United States were on the brink of war. Pp. 20–24. 3. Plaintiffs have not demonstrated a likelihood of success on the merits of their claim that the Proclamation violates the Establishment Clause. Pp. 24–38. (a) The individual plaintiffs have Article III standing to challenge the exclusion of their relatives under the Establishment Clause. A person’s interest in being united with his relatives is sufficiently concrete and particularized to form the basis of an Article III injury in fact. Cf., e.g., Kerry v. Din, 576 U. S. ___, ___. Pp. 24–26. (b) Plaintiffs allege that the primary purpose of the Proclamation was religious animus and that the President’s stated concerns about vetting protocols and national security were but pretexts for discriminating against Muslims. At the heart of their case is a series of statements by the President and his advisers both during the campaign and since the President assumed office. The issue, however, is not whether to denounce the President’s statements, but the significance of those statements in reviewing a Presidential directive, neutral on its face, addressing a matter within the core of executive responsibility. In doing so, the Court must consider not only the statements of a particular President, but also the authority of the Presidency itself. Pp. 26–29. (c) The admission and exclusion of foreign nationals is a “fundamental sovereign attribute exercised by the Government’s political departments largely immune from judicial control.” Fiallo v. Bell, 430 U. S. 787, 792. Although foreign nationals seeking admission have no constitutional right to entry, this Court has engaged in a circumscribed judicial inquiry when the denial of a visa allegedly burdens the constitutional rights of a U. S. citizen. That review is limited to whether the Executive gives a “facially legitimate and bona fide” reason for its action, Kleindienst v. Mandel, 408 U. S. 753, 769, but the Court need not define the precise contours of that narrow inquiry in this case. For today’s purposes, the Court assumes that it may look behind the face of the Proclamation to the extent of applying rational basis review, i.e., whether the entry policy is plausibly related to the Government’s stated objective to protect the country and improve vetting processes. Plaintiffs’ extrinsic evidence may be considered, but the policy will be upheld so long as it can reasonably be understood to result from a justification independent of unconstitutional grounds. Pp. 30–32. (d) On the few occasions where the Court has struck down a policy as illegitimate under rational basis scrutiny, a common thread has been that the laws at issue were “divorced from any factual context from which [the Court] could discern a relationship to legitimate state interests.” Romer v. Evans, 517 U. S. 620, 635. The Proclamation does not fit that pattern. It is expressly premised on legitimate purposes and says nothing about religion. The entry restrictions on Muslim-majority nations are limited to countries that were previously designated by Congress or prior administrations as posing national security risks. Moreover, the Proclamation reflects the results of a worldwide review process undertaken by multiple Cabinet officials and their agencies. Plaintiffs challenge the entry suspension based on their perception of its effectiveness and wisdom, but the Court cannot substitute its own assessment for the Executive’s predictive judgments on such matters. See Holder v. Humanitarian Law Project, 561 U. S. 1, 33–34. Three additional features of the entry policy support the Government’s claim of a legitimate national security interest. First, since the President introduced entry restrictions in January 2017, three Muslim-majority countries—Iraq, Sudan, and Chad—have been removed from the list. Second, for those countries still subject to entry restrictions, the Proclamation includes numerous exceptions for various categories of foreign nationals. Finally, the Proclamation creates a waiver program open to all covered foreign nationals seeking entry as immigrants or nonimmigrants. Under these circumstances, the Government has set forth a sufficient national security justification to survive rational basis review. Pp. 33–38. 878 F. 3d 662, reversed and remanded. Roberts, C. J., delivered the opinion of the Court, in which Kennedy, Thomas, Alito, and Gorsuch, JJ., joined. Kennedy, J., and Thomas, J., filed concurring opinions. Breyer, J., filed a dissenting opinion, in which Kagan, J., joined. Sotomayor, J., filed a dissenting opinion, in which Ginsburg, J., joined.

Under the Immigration and Nationality Act, foreign nationals seeking entry into the United States undergo a vetting process to ensure that they satisfy the numerous requirements for admission. The Act also vests the President with authority to restrict the entry of aliens when- ever he finds that their entry “would be detrimental to the interests of the United States.” 8 U. S. C. §1182(f). Relying on that delegation, the President concluded that it was necessary to impose entry restrictions on nationals of countries that do not share adequate information for an informed entry determination, or that otherwise present national security risks. Presidential Proclamation No. 9645, 82 Fed. Reg. 45161 (2017) (Proclamation). The plaintiffs in this litigation, respondents here, challenged the application of those entry restrictions to certain aliens abroad. We now decide whether the President had authority under the Act to issue the Proclamation, and whether the entry policy violates the Establishment Clause of the First Amendment. I A Shortly after taking office, President Trump signed Executive Order No. 13769, Protecting the Nation From Foreign Terrorist Entry Into the United States. 82 Fed. Reg. 8977 (2017) (EO–1). EO–1 directed the Secretary of Homeland Security to conduct a review to examine the adequacy of information provided by foreign governments about their nationals seeking to enter the United States. §3(a). Pending that review, the order suspended for 90 days the entry of foreign nationals from seven countries—Iran, Iraq, Libya, Somalia, Sudan, Syria, and Yemen—that had been previously identified by Congress or prior administrations as posing heightened terrorism risks. §3(c). The District Court for the Western District of Washington entered a temporary restraining order blocking the entry restrictions, and the Court of Appeals for the Ninth Circuit denied the Government’s request to stay that order. Washington v. Trump, 847 F. 3d 1151 (2017) (per curiam). In response, the President revoked EO–1, replacing it with Executive Order No. 13780, which again directed a worldwide review. 82 Fed. Reg. 13209 (2017) (EO–2). Citing investigative burdens on agencies and the need to diminish the risk that dangerous individuals would enter without adequate vetting, EO–2 also temporarily restricted the entry (with case-by-case waivers) of foreign nationals from six of the countries covered by EO–1: Iran, Libya, Somalia, Sudan, Syria, and Yemen. §§2(c), 3(a). The order explained that those countries had been selected because each “is a state sponsor of terrorism, has been significantly compromised by terrorist organizations, or contains active conflict zones.” §1(d). The entry restriction was to stay in effect for 90 days, pending completion of the worldwide review. These interim measures were immediately challenged in court. The District Courts for the Districts of Maryland and Hawaii entered nationwide preliminary injunctions barring enforcement of the entry suspension, and the respective Courts of Appeals upheld those injunctions, albeit on different grounds. International Refugee Assistance Project (IRAP) v. Trump, 857 F. 3d 554 (CA4 2017); Hawaii v. Trump, 859 F. 3d 741 (CA9 2017) (per curiam). This Court granted certiorari and stayed the injunctions—allowing the entry suspension to go into effect—with respect to foreign nationals who lacked a “credible claim of a bona fide relationship” with a person or entity in the United States. Trump v. IRAP, 582 U. S. ___, ___ (2017) (per curiam) (slip op., at 12). The temporary restrictions in EO–2 expired before this Court took any action, and we vacated the lower court decisions as moot. Trump v. IRAP, 583 U. S. ___ (2017); Trump v. Hawaii, 583 U. S. ___ (2017). On September 24, 2017, after completion of the worldwide review, the President issued the Proclamation before us—Proclamation No. 9645, Enhancing Vetting Capabilities and Processes for Detecting Attempted Entry Into the United States by Terrorists or Other Public-Safety Threats. 82 Fed. Reg. 45161. The Proclamation (as its title indicates) sought to improve vetting procedures by identifying ongoing deficiencies in the information needed to assess whether nationals of particular countries present “public safety threats.” §1(a). To further that purpose, the Proclamation placed entry restrictions on the nationals of eight foreign states whose systems for managing and sharing information about their nationals the President deemed inadequate. The Proclamation described how foreign states were selected for inclusion based on the review undertaken pursuant to EO–2. As part of that review, the Department of Homeland Security (DHS), in consultation with the State Department and several intelligence agencies, developed a “baseline” for the information required from foreign governments to confirm the identity of individuals seeking entry into the United States, and to determine whether those individuals pose a security threat. §1(c). The baseline included three components. The first, “identity-management information,” focused on whether a foreign government ensures the integrity of travel documents by issuing electronic passports, reporting lost or stolen passports, and making available additional identity-related information. Second, the agencies considered the extent to which the country discloses information on criminal history and suspected terrorist links, provides travel document exemplars, and facilitates the U. S. Government’s receipt of information about airline passengers and crews traveling to the United States. Finally, the agencies weighed various indicators of national security risk, including whether the foreign state is a known or potential terrorist safe haven and whether it regularly declines to receive returning nationals following final orders of removal from the United States. Ibid. DHS collected and evaluated data regarding all foreign governments. §1(d). It identified 16 countries as having deficient information-sharing practices and presenting national security concerns, and another 31 countries as “at risk” of similarly failing to meet the baseline. §1(e). The State Department then undertook diplomatic efforts over a 50-day period to encourage all foreign governments to improve their practices. §1(f ). As a result of that effort, numerous countries provided DHS with travel document exemplars and agreed to share information on known or suspected terrorists. Ibid. Following the 50-day period, the Acting Secretary of Homeland Security concluded that eight countries—Chad, Iran, Iraq, Libya, North Korea, Syria, Venezuela, and Yemen—remained deficient in terms of their risk profile and willingness to provide requested information. The Acting Secretary recommended that the President impose entry restrictions on certain nationals from all of those countries except Iraq. §§1(g), (h). She also concluded that although Somalia generally satisfied the information-sharing component of the baseline standards, its “identity-management deficiencies” and “significant terrorist presence” presented special circumstances justifying additional limitations. She therefore recommended entry limitations for certain nationals of that country. §1(i). As for Iraq, the Acting Secretary found that entry limitations on its nationals were not warranted given the close cooperative relationship between the U. S. and Iraqi Governments and Iraq’s commitment to combating ISIS. §1(g). After consulting with multiple Cabinet members and other officials, the President adopted the Acting Secretary’s recommendations and issued the Proclamation. Invoking his authority under 8 U. S. C. §§1182(f ) and 1185(a), the President determined that certain entry restrictions were necessary to “prevent the entry of those foreign nationals about whom the United States Government lacks sufficient information”; “elicit improved identity- management and information-sharing protocols and practices from foreign governments”; and otherwise “advance [the] foreign policy, national security, and counterterrorism objectives” of the United States. Proclamation §1(h). The President explained that these restrictions would be the “most likely to encourage cooperation” while “protect[ing] the United States until such time as improvements occur.” Ibid. The Proclamation imposed a range of restrictions that vary based on the “distinct circumstances” in each of the eight countries. Ibid. For countries that do not cooperate with the United States in identifying security risks (Iran, North Korea, and Syria), the Proclamation suspends entry of all nationals, except for Iranians seeking nonimmigrant student and exchange-visitor visas. §§2(b)(ii), (d)(ii), (e)(ii). For countries that have information-sharing deficiencies but are nonetheless “valuable counterterrorism partner[s]” (Chad, Libya, and Yemen), it restricts entry of nationals seeking immigrant visas and nonimmigrant business or tourist visas. §§2(a)(i), (c)(i), (g)(i). Because Somalia generally satisfies the baseline standards but was found to present special risk factors, the Proclamation suspends entry of nationals seeking immigrant visas and requires additional scrutiny of nationals seeking nonimmigrant visas. §2(h)(ii). And for Venezuela, which refuses to cooperate in information sharing but for which alternative means are available to identify its nationals, the Proclamation limits entry only of certain government officials and their family members on nonimmigrant business or tourist visas. §2(f )(ii). The Proclamation exempts lawful permanent residents and foreign nationals who have been granted asylum. §3(b). It also provides for case-by-case waivers when a foreign national demonstrates undue hardship, and that his entry is in the national interest and would not pose a threat to public safety. §3(c)(i); see also §3(c)(iv) (listing examples of when a waiver might be appropriate, such as if the foreign national seeks to reside with a close family member, obtain urgent medical care, or pursue significant business obligations). The Proclamation further directs DHS to assess on a continuing basis whether entry restrictions should be modified or continued, and to report to the President every 180 days. §4. Upon completion of the first such review period, the President, on the recommendation of the Secretary of Homeland Security, determined that Chad had sufficiently improved its practices, and he accordingly lifted restrictions on its nationals. Presidential Proclamation No. 9723, 83 Fed. Reg. 15937 (2018). B Plaintiffs in this case are the State of Hawaii, three individuals (Dr. Ismail Elshikh, John Doe #1, and John Doe #2), and the Muslim Association of Hawaii. The State operates the University of Hawaii system, which recruits students and faculty from the designated countries. The three individual plaintiffs are U. S. citizens or lawful permanent residents who have relatives from Iran, Syria, and Yemen applying for immigrant or nonimmigrant visas. The Association is a nonprofit organization that operates a mosque in Hawaii. Plaintiffs challenged the Proclamation—except as applied to North Korea and Venezuela—on several grounds. As relevant here, they argued that the Proclamation contravenes provisions in the Immigration and Nationality Act (INA), 66Stat. 187, as amended. Plaintiffs further claimed that the Proclamation violates the Establishment Clause of the First Amendment, because it was motivated not by concerns pertaining to national security but by animus toward Islam. The District Court granted a nationwide preliminary injunction barring enforcement of the entry restrictions. The court concluded that the Proclamation violated two provisions of the INA: §1182(f ), because the President did not make sufficient findings that the entry of the covered foreign nationals would be detrimental to the national interest, and §1152(a)(1)(A), because the policy discriminates against immigrant visa applicants on the basis of nationality. 265 F. Supp. 3d 1140, 1155–1159 (Haw. 2017). The Government requested expedited briefing and sought a stay pending appeal. The Court of Appeals for the Ninth Circuit granted a partial stay, permitting enforcement of the Proclamation with respect to foreign nationals who lack a bona fide relationship with the United States. This Court then stayed the injunction in full pending disposition of the Government’s appeal. 583 U. S. ___ (2017). The Court of Appeals affirmed. The court first held that the Proclamation exceeds the President’s authority under §1182(f ). In its view, that provision authorizes only a “temporary” suspension of entry in response to “exigencies” that “Congress would be ill-equipped to address.” 878 F. 3d 662, 684, 688 (2017). The court further reasoned that the Proclamation “conflicts with the INA’s finely reticulated regulatory scheme” by addressing “matters of immigration already passed upon by Congress.” Id., at 685, 690. The Ninth Circuit then turned to §1152(a)(1)(A) and determined that the entry restrictions also contravene the prohibition on nationality-based discrimination in the issuance of immigrant visas. The court did not reach plaintiffs’ Establishment Clause claim. We granted certiorari. 583 U. S. ___ (2018). II Before addressing the merits of plaintiffs’ statutory claims, we consider whether we have authority to do so. The Government argues that plaintiffs’ challenge to the Proclamation under the INA is not justiciable. Relying on the doctrine of consular nonreviewability, the Government contends that because aliens have no “claim of right” to enter the United States, and because exclusion of aliens is “a fundamental act of sovereignty” by the political branches, review of an exclusion decision “is not within the province of any court, unless expressly authorized by law.” United States ex rel. Knauff v. Shaughnessy, 338 U. S. 537, 542–543 (1950). According to the Government, that principle barring review is reflected in the INA, which sets forth a comprehensive framework for review of orders of removal, but authorizes judicial review only for aliens physically present in the United States. See Brief for Petitioners 19–20 (citing 8 U. S. C. §1252). The justiciability of plaintiffs’ challenge under the INA presents a difficult question. The Government made similar arguments that no judicial review was available in Sale v. Haitian Centers Council, Inc., 509 U. S. 155 (1993). The Court in that case, however, went on to consider on the merits a statutory claim like the one before us without addressing the issue of reviewability. The Government does not argue that the doctrine of consular nonreview- ability goes to the Court’s jurisdiction, see Tr. of Oral Arg. 13, nor does it point to any provision of the INA that expressly strips the Court of jurisdiction over plaintiffs’ claims, see Sebelius v. Auburn Regional Medical Center, 568 U. S. 145, 153 (2013) (requiring Congress to “clearly state[ ]” that a statutory provision is jurisdictional). As a result, we may assume without deciding that plaintiffs’ statutory claims are reviewable, notwithstanding consular nonreviewability or any other statutory nonreviewability issue, and we proceed on that basis. III The INA establishes numerous grounds on which an alien abroad may be inadmissible to the United States and ineligible for a visa. See, e.g., 8 U. S. C. §§1182(a)(1) (health-related grounds), (a)(2) (criminal history), (a)(3)(B) (terrorist activities), (a)(3)(C) (foreign policy grounds). Congress has also delegated to the President authority to suspend or restrict the entry of aliens in certain circumstances. The principal source of that authority, §1182(f ), enables the President to “suspend the entry of all aliens or any class of aliens” whenever he “finds” that their entry “would be detrimental to the interests of the United States.”[1] Plaintiffs argue that the Proclamation is not a valid exercise of the President’s authority under the INA. In their view, §1182(f ) confers only a residual power to temporarily halt the entry of a discrete group of aliens engaged in harmful conduct. They also assert that the Proclamation violates another provision of the INA— 8 U. S. C. §1152(a)(1)(A)—because it discriminates on the basis of nationality in the issuance of immigrant visas. By its plain language, §1182(f ) grants the President broad discretion to suspend the entry of aliens into the United States. The President lawfully exercised that discretion based on his findings—following a worldwide, multi-agency review—that entry of the covered aliens would be detrimental to the national interest. And plaintiffs’ attempts to identify a conflict with other provisions in the INA, and their appeal to the statute’s purposes and legislative history, fail to overcome the clear statutory language. A The text of §1182(f ) states: “Whenever the President finds that the entry of any aliens or of any class of aliens into the United States would be detrimental to the interests of the United States, he may by proclamation, and for such period as he shall deem necessary, suspend the entry of all aliens or any class of aliens as immigrants or nonimmigrants, or impose on the entry of aliens any restrictions he may deem to be appropriate.” By its terms, §1182(f ) exudes deference to the President in every clause. It entrusts to the President the decisions whether and when to suspend entry (“[w]henever [he] finds that the entry” of aliens “would be detrimental” to the national interest); whose entry to suspend (“all aliens or any class of aliens”); for how long (“for such period as he shall deem necessary”); and on what conditions (“any restrictions he may deem to be appropriate”). It is therefore unsurprising that we have previously observed that §1182(f ) vests the President with “ample power” to impose entry restrictions in addition to those elsewhere enumerated in the INA. Sale, 509 U. S., at 187 (finding it “perfectly clear” that the President could “establish a naval blockade” to prevent illegal migrants from entering the United States); see also Abourezk v. Reagan, 785 F. 2d 1043, 1049, n. 2 (CADC 1986) (describing the “sweeping proclamation power” in §1182(f ) as enabling the President to supplement the other grounds of inadmissibility in the INA). The Proclamation falls well within this comprehensive delegation. The sole prerequisite set forth in §1182(f ) is that the President “find[ ]” that the entry of the covered aliens “would be detrimental to the interests of the United States.” The President has undoubtedly fulfilled that requirement here. He first ordered DHS and other agencies to conduct a comprehensive evaluation of every single country’s compliance with the information and risk assessment baseline. The President then issued a Proclamation setting forth extensive findings describing how deficiencies in the practices of select foreign governments—several of which are state sponsors of terrorism—deprive the Government of “sufficient information to assess the risks [those countries’ nationals] pose to the United States.” Proclamation §1(h)(i). Based on that review, the President found that it was in the national interest to restrict entry of aliens who could not be vetted with adequate information—both to protect national security and public safety, and to induce improvement by their home countries. The Proclamation therefore “craft[ed] . . . country-specific restrictions that would be most likely to encourage cooperation given each country’s distinct circumstances,” while securing the Nation “until such time as improvements occur.” Ibid.[2] Plaintiffs believe that these findings are insufficient. They argue, as an initial matter, that the Proclamation fails to provide a persuasive rationale for why nationality alone renders the covered foreign nationals a security risk. And they further discount the President’s stated concern about deficient vetting because the Proclamation allows many aliens from the designated countries to enter on nonimmigrant visas. Such arguments are grounded on the premise that §1182(f ) not only requires the President to make a finding that entry “would be detrimental to the interests of the United States,” but also to explain that finding with sufficient detail to enable judicial review. That premise is questionable. See Webster v. Doe, 486 U. S. 592, 600 (1988) (concluding that a statute authorizing the CIA Director to terminate an employee when the Director “shall deem such termination necessary or advisable in the interests of the United States” forecloses “any meaningful judicial standard of review”). But even assuming that some form of review is appropriate, plaintiffs’ attacks on the sufficiency of the President’s findings cannot be sustained. The 12-page Proclamation—which thoroughly describes the process, agency evaluations, and recommendations underlying the President’s chosen restrictions—is more detailed than any prior order a President has issued under §1182(f ). Contrast Presidential Proclamation No. 6958, 3 CFR 133 (1996) (President Clinton) (explaining in one sentence why suspending entry of members of the Sudanese government and armed forces “is in the foreign policy interests of the United States”); Presidential Proclamation No. 4865, 3 CFR 50–51 (1981) (President Reagan) (explaining in five sentences why measures to curtail “the continuing illegal migration by sea of large numbers of undocumented aliens into the southeastern United States” are “necessary”). Moreover, plaintiffs’ request for a searching inquiry into the persuasiveness of the President’s justifications is inconsistent with the broad statutory text and the deference traditionally accorded the President in this sphere. “Whether the President’s chosen method” of addressing perceived risks is justified from a policy perspective is “irrelevant to the scope of his [§1182(f )] authority.” Sale, 509 U. S., at 187–188. And when the President adopts “a preventive measure . . . in the context of international affairs and national security,” he is “not required to conclusively link all of the pieces in the puzzle before [courts] grant weight to [his] empirical conclusions.” Holder v. Humanitarian Law Project, 561 U. S. 1, 35 (2010). The Proclamation also comports with the remaining textual limits in §1182(f ). We agree with plaintiffs that the word “suspend” often connotes a “defer[ral] till later,” Webster’s Third New International Dictionary 2303 (1966). But that does not mean that the President is required to prescribe in advance a fixed end date for the entry restrictions. Section 1182(f ) authorizes the President to suspend entry “for such period as he shall deem necessary.” It follows that when a President suspends entry in response to a diplomatic dispute or policy concern, he may link the duration of those restrictions, implicitly or explicitly, to the resolution of the triggering condition. See, e.g., Presidential Proclamation No. 5829, 3 CFR 88 (1988) (President Reagan) (suspending the entry of certain Panamanian nationals “until such time as . . . democracy has been restored in Panama”); Presidential Proclamation No. 8693, 3 CFR 86–87 (2011) (President Obama) (suspending the entry of individuals subject to a travel restriction under United Nations Security Council resolutions “until such time as the Secretary of State determines that [the suspension] is no longer necessary”). In fact, not one of the 43 suspension orders issued prior to this litigation has specified a precise end date. Like its predecessors, the Proclamation makes clear that its “conditional restrictions” will remain in force only so long as necessary to “address” the identified “inadequacies and risks” within the covered nations. Proclamation Preamble, and §1(h); see ibid. (explaining that the aim is to “relax[ ] or remove[ ]” the entry restrictions “as soon as possible”). To that end, the Proclamation establishes an ongoing process to engage covered nations and assess every 180 days whether the entry restrictions should be modified or terminated. §§4(a), (b). Indeed, after the initial review period, the President determined that Chad had made sufficient improvements to its identity-management protocols, and he accordingly lifted the entry suspension on its nationals. See Proclamation No. 9723, 83 Fed. Reg. 15937. Finally, the Proclamation properly identifies a “class of aliens”—nationals of select countries—whose entry is suspended. Plaintiffs argue that “class” must refer to a well-defined group of individuals who share a common “characteristic” apart from nationality. Brief for Respondents 42. But the text of §1182(f ), of course, does not say that, and the word “class” comfortably encompasses a group of people linked by nationality. Plaintiffs also contend that the class cannot be “overbroad.” Brief for Respondents 42. But that simply amounts to an unspoken tailoring requirement found nowhere in Congress’s grant of authority to suspend entry of not only “any class of aliens” but “all aliens.” In short, the language of §1182(f ) is clear, and the Proclamation does not exceed any textual limit on the President’s authority. B Confronted with this “facially broad grant of power,” 878 F. 3d, at 688, plaintiffs focus their attention on statutory structure and legislative purpose. They seek support in, first, the immigration scheme reflected in the INA as a whole, and, second, the legislative history of §1182(f ) and historical practice. Neither argument justifies departing from the clear text of the statute. 1 Plaintiffs’ structural argument starts with the premise that §1182(f ) does not give the President authority to countermand Congress’s considered policy judgments. The President, they say, may supplement the INA, but he cannot supplant it. And in their view, the Proclamation falls in the latter category because Congress has already specified a two-part solution to the problem of aliens seeking entry from countries that do not share sufficient information with the United States. First, Congress designed an individualized vetting system that places the burden on the alien to prove his admissibility. See §1361. Second, instead of banning the entry of nationals from particular countries, Congress sought to encourage information sharing through a Visa Waiver Program offering fast-track admission for countries that cooperate with the United States. See §1187. We may assume that §1182(f ) does not allow the President to expressly override particular provisions of the INA. But plaintiffs have not identified any conflict between the statute and the Proclamation that would implicitly bar the President from addressing deficiencies in the Nation’s vetting system. To the contrary, the Proclamation supports Congress’s individualized approach for determining admissibility. The INA sets forth various inadmissibility grounds based on connections to terrorism and criminal history, but those provisions can only work when the consular officer has sufficient (and sufficiently reliable) information to make that determination. The Proclamation promotes the effectiveness of the vetting process by helping to ensure the availability of such information. Plaintiffs suggest that the entry restrictions are unnecessary because consular officers can simply deny visas in individual cases when an alien fails to carry his burden of proving admissibility—for example, by failing to produce certified records regarding his criminal history. Brief for Respondents 48. But that misses the point: A critical finding of the Proclamation is that the failure of certain countries to provide reliable information prevents the Government from accurately determining whether an alien is inadmissible or poses a threat. Proclamation §1(h). Unless consular officers are expected to apply categorical rules and deny entry from those countries across the board, fraudulent or unreliable documentation may thwart their review in individual cases. And at any rate, the INA certainly does not require that systemic problems such as the lack of reliable information be addressed only in a progression of case-by-case admissibility determinations. One of the key objectives of the Proclamation is to encourage foreign governments to improve their practices, thus facilitating the Government’s vetting process overall. Ibid. Nor is there a conflict between the Proclamation and the Visa Waiver Program. The Program allows travel without a visa for short-term visitors from 38 countries that have entered into a “rigorous security partnership” with the United States. DHS, U. S. Visa Waiver Program (Apr. 6, 2016), http://www.dhs.gov/visa-waiver-program (as last visited June 25, 2018). Eligibility for that partnership involves “broad and consequential assessments of [the country’s] foreign security standards and operations.” Ibid. A foreign government must (among other things) undergo a comprehensive evaluation of its “counterterrorism, law enforcement, immigration enforcement, passport security, and border management capabilities,” often including “operational site inspections of airports, seaports, land borders, and passport production and issuance facilities.” Ibid. Congress’s decision to authorize a benefit for “many of America’s closest allies,” ibid., did not implicitly foreclose the Executive from imposing tighter restrictions on nationals of certain high-risk countries. The Visa Waiver Program creates a special exemption for citizens of countries that maintain exemplary security standards and offer “reciprocal [travel] privileges” to United States citizens. 8 U. S. C. §1187(a)(2)(A). But in establishing a select partnership covering less than 20% of the countries in the world, Congress did not address what requirements should govern the entry of nationals from the vast majority of countries that fall short of that gold standard—particularly those nations presenting heightened terrorism concerns. Nor did Congress attempt to determine—as the multi-agency review process did—whether those high-risk countries provide a minimum baseline of information to adequately vet their nationals. Once again, this is not a situation where “Congress has stepped into the space and solved the exact problem.” Tr. of Oral Arg. 53. Although plaintiffs claim that their reading preserves for the President a flexible power to “supplement” the INA, their understanding of the President’s authority is remarkably cramped: He may suspend entry by classes of aliens “similar in nature” to the existing categories of inadmissibility—but not too similar—or only in response to “some exigent circumstance” that Congress did not already touch on in the INA. Brief for Respondents 31, 36, 50; see also Tr. of Oral Arg. 57 (“Presidents have wide berth in this area . . . if there’s any sort of emergency.”). In any event, no Congress that wanted to confer on the President only a residual authority to address emergency situations would ever use language of the sort in §1182(f ). Fairly read, the provision vests authority in the President to impose additional limitations on entry beyond the grounds for exclusion set forth in the INA—including in response to circumstances that might affect the vetting system or other “interests of the United States.” Because plaintiffs do not point to any contradiction with another provision of the INA, the President has not exceeded his authority under §1182(f ). 2 Plaintiffs seek to locate additional limitations on the scope of §1182(f ) in the statutory background and legislative history. Given the clarity of the text, we need not consider such extra-textual evidence. See State Farm Fire & Casualty Co. v. United States ex rel. Rigsby, 580 U. S. ___, ___ (2016) (slip op., at 9). At any rate, plaintiffs’ evidence supports the plain meaning of the provision. Drawing on legislative debates over §1182(f ), plaintiffs suggest that the President’s suspension power should be limited to exigencies where it would be difficult for Congress to react promptly. Precursor provisions enacted during the First and Second World Wars confined the President’s exclusion authority to times of “war” and “national emergency.” See Act of May 22, 1918, §1(a), 40Stat. 559; Act of June 21, 1941, ch. 210, §1, 55Stat. 252. When Congress enacted §1182(f ) in 1952, plaintiffs note, it borrowed “nearly verbatim” from those predecessor statutes, and one of the bill’s sponsors affirmed that the provision would apply only during a time of crisis. According to plaintiffs, it therefore follows that Congress sought to delegate only a similarly tailored suspension power in §1182(f ). Brief for Respondents 39–40. If anything, the drafting history suggests the opposite. In borrowing “nearly verbatim” from the pre-existing statute, Congress made one critical alteration—it removed the national emergency standard that plaintiffs now seek to reintroduce in another form. Weighing Congress’s conscious departure from its wartime statutes against an isolated floor statement, the departure is far more probative. See NLRB v. SW General, Inc., 580 U. S. ___, ___ (2017) (slip op., at 16) (“[F]loor statements by individual legislators rank among the least illuminating forms of legislative history.”). When Congress wishes to condition an exercise of executive authority on the President’s finding of an exigency or crisis, it knows how to say just that. See, e.g., 16 U. S. C. §824o–1(b); 42 U. S. C. §5192; 50 U. S. C. §§1701, 1702. Here, Congress instead chose to condition the President’s exercise of the suspension authority on a different finding: that the entry of an alien or class of aliens would be “detrimental to the interests of the United States.” Plaintiffs also strive to infer limitations from executive practice. By their count, every previous suspension order under §1182(f ) can be slotted into one of two categories. The vast majority targeted discrete groups of foreign nationals engaging in conduct “deemed harmful by the immigration laws.” And the remaining entry restrictions that focused on entire nationalities—namely, President Carter’s response to the Iran hostage crisis and President Reagan’s suspension of immigration from Cuba—were, in their view, designed as a response to diplomatic emergencies “that the immigration laws do not address.” Brief for Respondents 40–41. Even if we were willing to confine expansive language in light of its past applications, the historical evidence is more equivocal than plaintiffs acknowledge. Presidents have repeatedly suspended entry not because the covered nationals themselves engaged in harmful acts but instead to retaliate for conduct by their governments that conflicted with U. S. foreign policy interests. See, e.g., Exec. Order No. 13662, 3 CFR 233 (2014) (President Obama) (suspending entry of Russian nationals working in the financial services, energy, mining, engineering, or defense sectors, in light of the Russian Federation’s “annexation of Crimea and its use of force in Ukraine”); Presidential Proclamation No. 6958, 3 CFR 133 (1997) (President Clinton) (suspending entry of Sudanese governmental and military personnel, citing “foreign policy interests of the United States” based on Sudan’s refusal to comply with United Nations resolution). And while some of these reprisals were directed at subsets of aliens from the countries at issue, others broadly suspended entry on the basis of nationality due to ongoing diplomatic disputes. For example, President Reagan invoked §1182(f ) to suspend entry “as immigrants” by almost all Cuban nationals, to apply pressure on the Cuban Government. Presidential Proclamation No. 5517, 3 CFR 102 (1986). Plaintiffs try to fit this latter order within their carve-out for emergency action, but the proclamation was based in part on Cuba’s decision to breach an immigration agreement some 15 months earlier. More significantly, plaintiffs’ argument about historical practice is a double-edged sword. The more ad hoc their account of executive action—to fit the history into their theory—the harder it becomes to see such a refined delegation in a statute that grants the President sweeping authority to decide whether to suspend entry, whose entry to suspend, and for how long. C Plaintiffs’ final statutory argument is that the President’s entry suspension violates §1152(a)(1)(A), which provides that “no person shall . . . be discriminated against in the issuance of an immigrant visa because of the person’s race, sex, nationality, place of birth, or place of residence.” They contend that we should interpret the provision as prohibiting nationality-based discrimination throughout the entire immigration process, despite the reference in §1152(a)(1)(A) to the act of visa issuance alone. Specifically, plaintiffs argue that §1152(a)(1)(A) applies to the predicate question of a visa applicant’s eligibility for admission and the subsequent question whether the holder of a visa may in fact enter the country. Any other conclusion, they say, would allow the President to circumvent the protections against discrimination enshrined in §1152(a)(1)(A). As an initial matter, this argument challenges only the validity of the entry restrictions on immigrant travel. Section 1152(a)(1)(A) is expressly limited to the issuance of “immigrant visa[s]” while §1182(f ) allows the Presi- dent to suspend entry of “immigrants or nonimmigrants.” At a minimum, then, plaintiffs’ reading would not affect any of the limitations on nonimmigrant travel in the Proclamation. In any event, we reject plaintiffs’ interpretation because it ignores the basic distinction between admissibility determinations and visa issuance that runs throughout the INA.[3] Section 1182 defines the pool of individuals who are admissible to the United States. Its restrictions come into play at two points in the process of gaining entry (or admission)[4] into the United States. First, any alien who is inadmissible under §1182 (based on, for example, health risks, criminal history, or foreign policy consequences) is screened out as “ineligible to receive a visa.” 8 U. S. C. §1201(g). Second, even if a consular officer issues a visa, entry into the United States is not guaranteed. As every visa application explains, a visa does not entitle an alien to enter the United States “if, upon arrival,” an immigration officer determines that the applicant is “inadmissible under this chapter, or any other provision of law”—including §1182(f ). §1201(h). Sections 1182(f ) and 1152(a)(1)(A) thus operate in different spheres: Section 1182 defines the universe of aliens who are admissible into the United States (and therefore eligible to receive a visa). Once §1182 sets the boundaries of admissibility into the United States, §1152(a)(1)(A) prohibits discrimination in the allocation of immigrant visas based on nationality and other traits. The distinction between admissibility—to which §1152(a)(1)(A) does not apply—and visa issuance—to which it does—is apparent from the text of the provision, which specifies only that its protections apply to the “issuance” of “immigrant visa[s],” without mentioning admissibility or entry. Had Congress instead intended in §1152(a)(1)(A) to constrain the President’s power to determine who may enter the country, it could easily have chosen language directed to that end. See, e.g., §§1182(a)(3)(C)(ii), (iii) (providing that certain aliens “shall not be excludable or subject to restrictions or conditions on entry . . . because of the alien’s past, current, or expected beliefs, statements, or associations” (emphasis added)). “The fact that [Congress] did not adopt [a] readily available and apparent alternative strongly supports” the conclusion that §1152(a)(1)(A) does not limit the President’s delegated authority under §1182(f ). Knight v. Commissioner, 552 U. S. 181, 188 (2008). Common sense and historical practice confirm as much. Section 1152(a)(1)(A) has never been treated as a constraint on the criteria for admissibility in §1182. Presidents have repeatedly exercised their authority to suspend entry on the basis of nationality. As noted, President Reagan relied on §1182(f ) to suspend entry “as immigrants by all Cuban nationals,” subject to exceptions. Proclamation No. 5517, 51 Fed. Reg. 30470 (1986). Likewise, President Carter invoked §1185(a)(1) to deny and revoke visas to all Iranian nationals. See Exec. Order No. 12172, 3 CFR 461 (1979), as amended by Exec. Order No. 12206, 3 CFR 249 (1980); Public Papers of the Presidents, Jimmy Carter, Sanctions Against Iran, Vol. 1, Apr. 7, 1980, pp. 611–612 (1980); see also n. 1, supra. On plaintiffs’ reading, those orders were beyond the President’s authority. The entry restrictions in the Proclamation on North Korea (which plaintiffs do not challenge in this litigation) would also be unlawful. Nor would the President be permitted to suspend entry from particular foreign states in response to an epidemic confined to a single region, or a verified terrorist threat involving nationals of a specific foreign nation, or even if the United States were on the brink of war. In a reprise of their §1182(f ) argument, plaintiffs attempt to soften their position by falling back on an implicit exception for Presidential actions that are “closely drawn” to address “specific fast-breaking exigencies.” Brief for Respondents 60–61. Yet the absence of any textual basis for such an exception more likely indicates that Congress did not intend for §1152(a)(1)(A) to limit the President’s flexible authority to suspend entry based on foreign policy interests. In addition, plaintiffs’ proposed exigency test would require courts, rather than the President, to determine whether a foreign government’s conduct rises to the level that would trigger a supposed implicit exception to a federal statute. See Reno v. American-Arab Anti-Discrimination Comm., 525 U. S. 471, 491 (1999) (explaining that even if the Executive “disclose[d] its . . . reasons for deeming nationals of a particular country a special threat,” courts would be “unable to assess their adequacy”). The text of §1152(a)(1)(A) offers no standards that would enable courts to assess, for example, whether the situation in North Korea justifies entry restrictions while the terrorist threat in Yemen does not. * * * The Proclamation is squarely within the scope of Presidential authority under the INA. Indeed, neither dissent even attempts any serious argument to the contrary, despite the fact that plaintiffs’ primary contention below and in their briefing before this Court was that the Proclamation violated the statute. IV A We now turn to plaintiffs’ claim that the Proclamation was issued for the unconstitutional purpose of excluding Muslims. Because we have an obligation to assure ourselves of jurisdiction under Article III, we begin by addressing the question whether plaintiffs have standing to bring their constitutional challenge. Federal courts have authority under the Constitution to decide legal questions only in the course of resolving “Cases” or “Controversies.” Art. III, §2. One of the essential elements of a legal case or controversy is that the plaintiff have standing to sue. Standing requires more than just a “keen interest in the issue.” Hollingsworth v. Perry, 570 U. S. 693, 700 (2013). It requires allegations—and, eventually, proof—that the plaintiff “personal[ly]” suffered a concrete and particularized injury in connection with the conduct about which he complains. Spokeo, Inc. v. Robins, 578 U. S. ___, ___ (2016) (slip op., at 7). In a case arising from an alleged violation of the Establishment Clause, a plaintiff must show, as in other cases, that he is “directly affected by the laws and practices against which [his] complaints are directed.” School Dist. of Abington Township v. Schempp, 374 U. S. 203, 224, n. 9 (1963). That is an issue here because the entry restrictions apply not to plaintiffs themselves but to others seeking to enter the United States. Plaintiffs first argue that they have standing on the ground that the Proclamation “establishes a disfavored faith” and violates “their own right to be free from federal [religious] establishments.” Brief for Respondents 27–28 (emphasis deleted). They describe such injury as “spirit- ual and dignitary.” Id., at 29. We need not decide whether the claimed dignitary interest establishes an adequate ground for standing. The three individual plaintiffs assert another, more concrete injury: the alleged real-world effect that the Proclamation has had in keeping them separated from certain relatives who seek to enter the country. See ibid.; Town of Chester v. Laroe Estates, Inc., 581 U. S. ___, ___–___ (2017) (slip op., at 5–6) (“At least one plaintiff must have standing to seek each form of relief requested in the complaint.”). We agree that a person’s interest in being united with his relatives is sufficiently concrete and particularized to form the basis of an Article III injury in fact. This Court has previously considered the merits of claims asserted by United States citizens regarding violations of their personal rights allegedly caused by the Government’s exclusion of particular foreign nationals. See Kerry v. Din, 576 U. S. ___, ___ (2015) (plurality opinion) (slip op., at 15); id., at ___ (Kennedy, J., concurring in judgment) (slip op., at 1); Kleindienst v. Mandel, 408 U. S. 753, 762 (1972). Likewise, one of our prior stay orders in this litigation recognized that an American individual who has “a bona fide relationship with a particular person seeking to enter the country . . . can legitimately claim concrete hardship if that person is excluded.” Trump v. IRAP, 582 U. S., at ___ (slip op., at 13). The Government responds that plaintiffs’ Establishment Clause claims are not justiciable because the Clause does not give them a legally protected interest in the admission of particular foreign nationals. But that argument—which depends upon the scope of plaintiffs’ Establishment Clause rights—concerns the merits rather than the justiciability of plaintiffs’ claims. We therefore conclude that the individual plaintiffs have Article III standing to challenge the exclusion of their relatives under the Establishment Clause. B The First Amendment provides, in part, that “Congress shall make no law respecting an establishment of religion, or prohibiting the free exercise thereof.” Our cases recognize that “[t]he clearest command of the Establishment Clause is that one religious denomination cannot be officially preferred over another.” Larson v. Valente, 456 U. S. 228, 244 (1982). Plaintiffs believe that the Proclamation violates this prohibition by singling out Muslims for disfavored treatment. The entry suspension, they contend, operates as a “religious gerrymander,” in part because most of the countries covered by the Proclamation have Muslim-majority populations. And in their view, deviations from the information-sharing baseline criteria suggest that the results of the multi-agency review were “foreordained.” Relying on Establishment Clause precedents concerning laws and policies applied domestically, plaintiffs allege that the primary purpose of the Proclamation was religious animus and that the President’s stated concerns about vetting protocols and national security were but pretexts for discriminating against Muslims. Brief for Respondents 69–73. At the heart of plaintiffs’ case is a series of statements by the President and his advisers casting doubt on the official objective of the Proclamation. For example, while a candidate on the campaign trail, the President published a “Statement on Preventing Muslim Immigration” that called for a “total and complete shutdown of Muslims entering the United States until our country’s representatives can figure out what is going on.” App. 158. That statement remained on his campaign website until May 2017. Id., at 130–131. Then-candidate Trump also stated that “Islam hates us” and asserted that the United States was “having problems with Muslims coming into the country.” Id., at 120–121, 159. Shortly after being elected, when asked whether violence in Europe had affected his plans to “ban Muslim immigration,” the President replied, “You know my plans. All along, I’ve been proven to be right.” Id., at 123. One week after his inauguration, the President issued EO–1. In a television interview, one of the President’s campaign advisers explained that when the President “first announced it, he said, ‘Muslim ban.’ He called me up. He said, ‘Put a commission together. Show me the right way to do it legally.’ ” Id., at 125. The adviser said he assembled a group of Members of Congress and lawyers that “focused on, instead of religion, danger. . . . [The order] is based on places where there [is] substantial evidence that people are sending terrorists into our country.” Id., at 229. Plaintiffs also note that after issuing EO–2 to replace EO–1, the President expressed regret that his prior order had been “watered down” and called for a “much tougher version” of his “Travel Ban.” Shortly before the release of the Proclamation, he stated that the “travel ban . . . should be far larger, tougher, and more specific,” but “stupidly that would not be politically correct.” Id., at 132–133. More recently, on November 29, 2017, the President retweeted links to three anti-Muslim propaganda videos. In response to questions about those videos, the President’s deputy press secretary denied that the President thinks Muslims are a threat to the United States, explaining that “the President has been talking about these security issues for years now, from the campaign trail to the White House” and “has addressed these issues with the travel order that he issued earlier this year and the companion proclamation.” IRAP v. Trump, 883 F.3d 233, 267 (CA4 2018). The President of the United States possesses an extraordinary power to speak to his fellow citizens and on their behalf. Our Presidents have frequently used that power to espouse the principles of religious freedom and tolerance on which this Nation was founded. In 1790 George Washington reassured the Hebrew Congregation of Newport, Rhode Island that “happily the Government of the United States . . . gives to bigotry no sanction, to persecution no assistance [and] requires only that they who live under its protection should demean themselves as good citizens.” 6 Papers of George Washington 285 (D. Twohig ed. 1996). President Eisenhower, at the opening of the Islamic Center of Washington, similarly pledged to a Muslim audience that “America would fight with her whole strength for your right to have here your own church,” declaring that “[t]his concept is indeed a part of America.” Public Papers of the Presidents, Dwight D. Eisenhower, June 28, 1957, p. 509 (1957). And just days after the attacks of September 11, 2001, President George W. Bush returned to the same Islamic Center to implore his fellow Americans—Muslims and non-Muslims alike—to remember during their time of grief that “[t]he face of terror is not the true faith of Islam,” and that America is “a great country because we share the same values of respect and dignity and human worth.” Public Papers of the Presidents, George W. Bush, Vol. 2, Sept. 17, 2001, p. 1121 (2001). Yet it cannot be denied that the Federal Government and the Presidents who have carried its laws into effect have—from the Nation’s earliest days—performed unevenly in living up to those inspiring words. Plaintiffs argue that this President’s words strike at fundamental standards of respect and tolerance, in violation of our constitutional tradition. But the issue before us is not whether to denounce the statements. It is instead the significance of those statements in reviewing a Presidential directive, neutral on its face, addressing a matter within the core of executive responsibility. In doing so, we must consider not only the statements of a particular President, but also the authority of the Presidency itself. The case before us differs in numerous respects from the conventional Establishment Clause claim. Unlike the typical suit involving religious displays or school prayer, plaintiffs seek to invalidate a national security directive regulating the entry of aliens abroad. Their claim accordingly raises a number of delicate issues regarding the scope of the constitutional right and the manner of proof. The Proclamation, moreover, is facially neutral toward religion. Plaintiffs therefore ask the Court to probe the sincerity of the stated justifications for the policy by reference to extrinsic statements—many of which were made before the President took the oath of office. These various aspects of plaintiffs’ challenge inform our standard of review. C For more than a century, this Court has recognized that the admission and exclusion of foreign nationals is a “fundamental sovereign attribute exercised by the Government’s political departments largely immune from judicial control.” Fiallo v. Bell, 430 U. S. 787, 792 (1977); see Harisiades v. Shaughnessy, 342 U. S. 580, 588–589 (1952) (“[A]ny policy toward aliens is vitally and intricately interwoven with contemporaneous policies in regard to the conduct of foreign relations [and] the war power.”). Because decisions in these matters may implicate “relations with foreign powers,” or involve “classifications defined in the light of changing political and economic circumstances,” such judgments “are frequently of a character more appropriate to either the Legislature or the Executive.” Mathews v. Diaz, 426 U. S. 67, 81 (1976). Nonetheless, although foreign nationals seeking admission have no constitutional right to entry, this Court has engaged in a circumscribed judicial inquiry when the denial of a visa allegedly burdens the constitutional rights of a U. S. citizen. In Kleindienst v. Mandel, the Attorney General denied admission to a Belgian journalist and self-described “revolutionary Marxist,” Ernest Mandel, who had been invited to speak at a conference at Stanford University. 408 U. S., at 756–757. The professors who wished to hear Mandel speak challenged that decision under the First Amendment, and we acknowledged that their constitutional “right to receive information” was implicated. Id., at 764–765. But we limited our review to whether the Executive gave a “facially legitimate and bona fide” reason for its action. Id., at 769. Given the authority of the political branches over admission, we held that “when the Executive exercises this [delegated] power negatively on the basis of a facially legitimate and bona fide reason, the courts will neither look behind the exercise of that discretion, nor test it by balancing its justification” against the asserted constitutional interests of U. S. citizens. Id., at 770. The principal dissent suggests that Mandel has no bearing on this case, post, at 14, and n. 5 (opinion of Sotomayor, J.) (hereinafter the dissent), but our opinions have reaffirmed and applied its deferential standard of review across different contexts and constitutional claims. In Din, Justice Kennedy reiterated that “respect for the political branches’ broad power over the creation and administration of the immigration system” meant that the Government need provide only a statutory citation to explain a visa denial. 576 U. S., at ___ (opinion concurring in judgment) (slip op., at 6). Likewise in Fiallo, we applied Mandel to a “broad congressional policy” giving immigration preferences to mothers of illegitimate children. 430 U. S., at 795. Even though the statute created a “categorical” entry classification that discriminated on the basis of sex and legitimacy, post, at 14, n. 5, the Court concluded that “it is not the judicial role in cases of this sort to probe and test the justifications” of immigration policies. 430 U. S., at 799 (citing Mandel, 408 U. S., at 770). Lower courts have similarly applied Mandel to broad executive action. See Rajah v. Mukasey, 544 F.3d 427, 433, 438–439 (CA2 2008) (upholding National Security Entry-Exit Registration System instituted after September 11, 2001). Mandel’s narrow standard of review “has particular force” in admission and immigration cases that overlap with “the area of national security.” Din, 576 U. S., at ___ (Kennedy, J., concurring in judgment) (slip op., at 3). For one, “[j]udicial inquiry into the national-security realm raises concerns for the separation of powers” by intruding on the President’s constitutional responsibilities in the area of foreign affairs. Ziglar v. Abbasi, 582 U. S. ___, ___ (2017) (slip op., at 19) (internal quotation marks omitted). For another, “when it comes to collecting evidence and drawing inferences” on questions of national security, “the lack of competence on the part of the courts is marked.” Humanitarian Law Project, 561 U. S., at 34. The upshot of our cases in this context is clear: “Any rule of constitutional law that would inhibit the flexibility” of the President “to respond to changing world conditions should be adopted only with the greatest caution,” and our inquiry into matters of entry and national security is highly constrained. Mathews, 426 U. S., at 81–82. We need not define the precise contours of that inquiry in this case. A conventional application of Mandel, asking only whether the policy is facially legitimate and bona fide, would put an end to our review. But the Government has suggested that it may be appropriate here for the inquiry to extend beyond the facial neutrality of the order. See Tr. of Oral Arg. 16–17, 25–27 (describing Mandel as “the starting point” of the analysis). For our purposes today, we assume that we may look behind the face of the Proclamation to the extent of applying rational basis review. That standard of review considers whether the entry policy is plausibly related to the Government’s stated objective to protect the country and improve vetting processes. See Railroad Retirement Bd. v. Fritz, 449 U. S. 166, 179 (1980). As a result, we may consider plaintiffs’ extrinsic evidence, but will uphold the policy so long as it can reasonably be understood to result from a justification independent of unconstitutional grounds.[5] D Given the standard of review, it should come as no surprise that the Court hardly ever strikes down a policy as illegitimate under rational basis scrutiny. On the few occasions where we have done so, a common thread has been that the laws at issue lack any purpose other than a “bare . . . desire to harm a politically unpopular group.” Department of Agriculture v. Moreno, 413 U. S. 528, 534 (1973). In one case, we invalidated a local zoning ordinance that required a special permit for group homes for the intellectually disabled, but not for other facilities such as fraternity houses or hospitals. We did so on the ground that the city’s stated concerns about (among other things) “legal responsibility” and “crowded conditions” rested on “an irrational prejudice” against the intellectually dis- abled. Cleburne v. Cleburne Living Center, Inc., 473 U. S. 432, 448–450 (1985) (internal quotation marks omitted). And in another case, this Court overturned a state constitutional amendment that denied gays and lesbians access to the protection of antidiscrimination laws. The amendment, we held, was “divorced from any factual context from which we could discern a relationship to legitimate state interests,” and “its sheer breadth [was] so discontinuous with the reasons offered for it” that the initiative seemed “inexplicable by anything but animus.” Romer v. Evans, 517 U. S. 620, 632, 635 (1996). The Proclamation does not fit this pattern. It cannot be said that it is impossible to “discern a relationship to legitimate state interests” or that the policy is “inexplicable by anything but animus.” Indeed, the dissent can only attempt to argue otherwise by refusing to apply anything resembling rational basis review. But because there is persuasive evidence that the entry suspension has a legitimate grounding in national security concerns, quite apart from any religious hostility, we must accept that independent justification. The Proclamation is expressly premised on legitimate purposes: preventing entry of nationals who cannot be adequately vetted and inducing other nations to improve their practices. The text says nothing about religion. Plaintiffs and the dissent nonetheless emphasize that five of the seven nations currently included in the Proclamation have Muslim-majority populations. Yet that fact alone does not support an inference of religious hostility, given that the policy covers just 8% of the world’s Muslim population and is limited to countries that were previously designated by Congress or prior administrations as posing national security risks. See 8 U. S. C. §1187(a)(12)(A) (identifying Syria and state sponsors of terrorism such as Iran as “countr[ies] or area[s] of concern” for purposes of administering the Visa Waiver Program); Dept. of Homeland Security, DHS Announces Further Travel Restrictions for the Visa Waiver Program (Feb. 18, 2016) (designating Libya, Somalia, and Yemen as additional countries of concern); see also Rajah, 544 F. 3d, at 433, n. 3 (describing how nonimmigrant aliens from Iran, Libya, Somalia, Syria, and Yemen were covered by the National Security Entry-Exit Registration System). The Proclamation, moreover, reflects the results of a worldwide review process undertaken by multiple Cabinet officials and their agencies. Plaintiffs seek to discredit the findings of the review, pointing to deviations from the review’s baseline criteria resulting in the inclusion of Somalia and omission of Iraq. But as the Proclamation explains, in each case the determinations were justified by the distinct conditions in each country. Although Somalia generally satisfies the information-sharing component of the baseline criteria, it “stands apart . . . in the degree to which [it] lacks command and control of its territory.” Proclamation §2(h)(i). As for Iraq, the Secretary of Homeland Security determined that entry restrictions were not warranted in light of the close cooperative relationship between the U. S. and Iraqi Governments and the country’s key role in combating terrorism in the region. §1(g). It is, in any event, difficult to see how exempting one of the largest predominantly Muslim countries in the region from coverage under the Proclamation can be cited as evidence of animus toward Muslims. The dissent likewise doubts the thoroughness of the multi-agency review because a recent Freedom of Information Act request shows that the final DHS report “was a mere 17 pages.” Post, at 19. Yet a simple page count offers little insight into the actual substance of the final report, much less predecisional materials underlying it. See 5 U. S. C. §552(b)(5) (exempting deliberative materials from FOIA disclosure). More fundamentally, plaintiffs and the dissent challenge the entry suspension based on their perception of its effectiveness and wisdom. They suggest that the policy is overbroad and does little to serve national security interests. But we cannot substitute our own assessment for the Executive’s predictive judgments on such matters, all of which “are delicate, complex, and involve large elements of prophecy.” Chicago & Southern Air Lines, Inc. v. Waterman S. S. Corp., 333 U. S. 103, 111 (1948); see also Regan v. Wald, 468 U. S. 222, 242–243 (1984) (declining invitation to conduct an “independent foreign policy analysis”). While we of course “do not defer to the Government’s reading of the First Amendment,” the Executive’s evaluation of the underlying facts is entitled to appropriate weight, particularly in the context of litigation involving “sensitive and weighty interests of national security and foreign affairs.” Humanitarian Law Project, 561 U. S., at 33–34.[6] Three additional features of the entry policy support the Government’s claim of a legitimate national security interest. First, since the President introduced entry restrictions in January 2017, three Muslim-majority countries—Iraq, Sudan, and Chad—have been removed from the list of covered countries. The Proclamation emphasizes that its “conditional restrictions” will remain in force only so long as necessary to “address” the identified “inadequacies and risks,” Proclamation Preamble, and §1(h), and establishes an ongoing process to engage covered nations and assess every 180 days whether the entry restrictions should be terminated, §§4(a), (b). In fact, in announcing the termination of restrictions on nationals of Chad, the President also described Libya’s ongoing engagement with the State Department and the steps Libya is taking “to improve its practices.” Proclamation No. 9723, 83 Fed. Reg. 15939. Second, for those countries that remain subject to entry restrictions, the Proclamation includes significant exceptions for various categories of foreign nationals. The policy permits nationals from nearly every covered country to travel to the United States on a variety of nonimmigrant visas. See, e.g., §§2(b)–(c), (g), (h) (permitting student and exchange visitors from Iran, while restricting only business and tourist nonimmigrant entry for nationals of Libya and Yemen, and imposing no restrictions on nonimmigrant entry for Somali nationals). These carveouts for nonimmigrant visas are substantial: Over the last three fiscal years—before the Proclamation was in effect—the majority of visas issued to nationals from the covered countries were nonimmigrant visas. Brief for Petitioners 57. The Proclamation also exempts permanent resi- dents and individuals who have been granted asylum. §§3(b)(i), (vi). Third, the Proclamation creates a waiver program open to all covered foreign nationals seeking entry as immigrants or nonimmigrants. According to the Proclamation, consular officers are to consider in each admissibility determination whether the alien demonstrates that (1) denying entry would cause undue hardship; (2) entry would not pose a threat to public safety; and (3) entry would be in the interest of the United States. §3(c)(i); see also §3(c)(iv) (listing examples of when a waiver might be appropriate, such as if the foreign national seeks to reside with a close family member, obtain urgent medical care, or pursue significant business obligations). On its face, this program is similar to the humanitarian exceptions set forth in President Carter’s order during the Iran hostage crisis. See Exec. Order No. 12206, 3 CFR 249; Public Papers of the Presidents, Jimmy Carter, Sanctions Against Iran, at 611–612 (1980) (outlining exceptions). The Proclamation also directs DHS and the State Department to issue guidance elaborating upon the circumstances that would justify a waiver.[7] Finally, the dissent invokes Korematsu v. United States, 323 U. S. 214 (1944). Whatever rhetorical advantage the dissent may see in doing so, Korematsu has nothing to do with this case. The forcible relocation of U. S. citizens to concentration camps, solely and explicitly on the basis of race, is objectively unlawful and outside the scope of Presidential authority. But it is wholly inapt to liken that morally repugnant order to a facially neutral policy denying certain foreign nationals the privilege of admission. See post, at 26–28. The entry suspension is an act that is well within executive authority and could have been taken by any other President—the only question is evaluating the actions of this particular President in promulgating an otherwise valid Proclamation. The dissent’s reference to Korematsu, however, affords this Court the opportunity to make express what is already obvious: Korematsu was gravely wrong the day it was decided, has been overruled in the court of history, and—to be clear—“has no place in law under the Constitution.” 323 U. S., at 248 (Jackson, J., dissenting). * * * Under these circumstances, the Government has set forth a sufficient national security justification to survive rational basis review. We express no view on the soundness of the policy. We simply hold today that plaintiffs have not demonstrated a likelihood of success on the merits of their constitutional claim. V Because plaintiffs have not shown that they are likely to succeed on the merits of their claims, we reverse the grant of the preliminary injunction as an abuse of discretion. Winter v. Natural Resources Defense Council, Inc., 555 U. S. 7, 32 (2008). The case now returns to the lower courts for such further proceedings as may be appropriate. Our disposition of the case makes it unnecessary to consider the propriety of the nationwide scope of the injunction issued by the District Court. The judgment of the Court of Appeals is reversed, and the case is remanded for further proceedings consistent with this opinion. It is so ordered. Notes 1 The President also invoked his power under 8 U. S. C. §1185(a)(1), which grants the President authority to adopt “reasonable rules, regulations, and orders” governing entry or removal of aliens, “subject to such limitations and exceptions as [he] may prescribe.” Because this provision “substantially overlap[s]” with §1182(f ), we agree with the Government that we “need not resolve . . . the precise relationship between the two statutes” in evaluating the validity of the Proclamation. Brief for Petitioners 32–33. 2 The Proclamation states that it does not disclose every ground for the country-specific restrictions because “[d]escribing all of those reasons publicly . . . would cause serious damage to the national security of the United States, and many such descriptions are classified.” §1(j). 3 The Act is rife with examples distinguishing between the two concepts. See, e.g., 8 U. S. C. §1101(a)(4) (“The term ‘application for admission’ has reference to the application for admission into the United States and not to the application for the issuance of an immigrant or nonimmigrant visa.”); §1182(a) (“ineligible to receive visas and ineligible to be admitted”); §1182(a)(3)(D)(iii) (“establishes to the satisfaction of the consular officer when applying for a visa . . . or to the satisfaction of the Attorney General when applying for admission”); §1182(h)(1)(A)(i) (“alien’s application for a visa, admission, or adjustment of status”); §1187 (permitting entry without a visa); §1361 (establishing burden of proof for when a person “makes application for a visa . . . , or makes application for admission, or otherwise attempts to enter the United States”). 4 The concepts of entry and admission—but not issuance of a visa—are used interchangeably in the INA. See §1101(a)(13)(A) (defining “admission” as the “lawful entry of the alien into the United States”). 5 The dissent finds “perplexing” the application of rational basis review in this context. Post, at 15. But what is far more problematic is the dissent’s assumption that courts should review immigration policies, diplomatic sanctions, and military actions under the de novo “reasonable observer” inquiry applicable to cases involving holiday displays and graduation ceremonies. The dissent criticizes application of a more constrained standard of review as “throw[ing] the Establishment Clause out the window.” Post, at 16, n. 6. But as the numerous precedents cited in this section make clear, such a circumscribed inquiry applies to any constitutional claim concerning the entry of foreign nationals. See Part IV–C, supra. The dissent can cite no authority for its proposition that the more free-ranging inquiry it proposes is appropriate in the national security and foreign affairs context. 6 The dissent recycles much of plaintiffs’ §1182(f ) argument to assert that “Congress has already erected a statutory scheme that fulfills” the President’s stated concern about deficient vetting. Post, at 19–21. But for the reasons set forth earlier, Congress has not in any sense “stepped into the space and solved the exact problem.” Tr. of Oral Arg. 53. Neither the existing inadmissibility grounds nor the narrow Visa Waiver Program address the failure of certain high-risk countries to provide a minimum baseline of reliable information. See Part III–B–1, supra. 7 Justice Breyer focuses on only one aspect of our consideration—the waiver program and other exemptions in the Proclamation. Citing selective statistics, anecdotal evidence, and a declaration from unre-lated litigation, Justice Breyer suggests that not enough individuals are receiving waivers or exemptions. Post, at 4–8 (dissenting opinion). Yet even if such an inquiry were appropriate under rational basis review, the evidence he cites provides “but a piece of the picture,” post, at 6, and does not affect our analysis.

583.US.2017_15-1509

Respondent Lakeridge is a corporate entity with a single owner, MBP Equity Partners. When Lakeridge filed for Chapter 11 bankruptcy, it had a pair of substantial debts: It owed petitioner U. S. Bank over $10 million and MBP another $2.76 million. Lakeridge submitted a reorganization plan, proposing to impair the interests of both U. S. Bank and MBP. U. S. Bank refused the offer, thus blocking Lake- ridge’s option for reorganization through a fully consensual plan. See 11 U. S. C. §1129(a)(8). Lakeridge then turned to the so-called “cramdown” plan option for imposing a plan impairing the interests of a non-consenting class of creditors. See §1129(b). Among the prerequisites for judicial approval of such a plan is that another impaired class of creditors has consented to it. See §1129(a)(10). But crucially here, the consent of a creditor who is also an “insider” of the debtor does not count for that purpose. Ibid. The Bankruptcy Code’s definition of an insider “includes” any director, officer, or “person in control” of the entity. §101(31)(B)(i)–(iii). Courts have devised tests for identifying other, so-called “non-statutory” insiders, focusing, in whole or in part, on whether a person’s transactions with the debtor were at arm’s length. Here, MBP (an insider of Lakeridge) could not provide the partial agreement needed for a cramdown plan, and Lakeridge’s reorganization was thus impeded. MBP sought to transfer its claim against Lakeridge to a non-insider who could agree to the cramdown plan. Kathleen Bartlett, an MBP board member and Lakeridge officer, offered MBP’s claim to Robert Rabkin, a retired surgeon, for $5,000. Rabkin purchased the claim and consented to Lakeridge’s proposed reorganization. U. S. Bank objected, arguing that Rabkin was a non-statutory insider because he had a “romantic” relationship with Bartlett and the purchase was not an arm’s-length transaction. The Bankruptcy Court rejected U. S. Bank’s argument. The Ninth Circuit affirmed. Viewing the Bankruptcy Court’s decision as one based on a finding that the relevant transaction was conducted at arm’s length, the Ninth Circuit held that that finding was entitled to clear-error review, and could not be reversed under that deferential standard. Held: The Ninth Circuit was right to review the Bankruptcy Court’s determination for clear error (rather than de novo). At the heart of this case is a so-called “mixed question” of law and fact—whether the Bankruptcy Court’s findings of fact satisfy the legal test chosen for conferring non-statutory insider status. U. S. Bank contends that the Bankruptcy Court’s resolution of this mixed question must be reviewed de novo, while Lakeridge (joined by the Federal Government) argues for a clear-error standard. For all their differences, both parties rightly point to the same query: What is the nature of the mixed question here and which kind of court (bankruptcy or appellate) is better suited to resolve it? Mixed questions are not all alike. Some require courts to expound on the law, and should typically be reviewed de novo. Others immerse courts in case-specific factual issues, and should usually be reviewed with deference. In short, the standard of review for a mixed question depends on whether answering it entails primarily legal or factual work. Here, the Bankruptcy Court confronted the question whether the basic facts it had discovered (concerning Rabkin’s relationships, motivations, etc.) were sufficient to make Rabkin a non-statutory insider. Using the transactional prong of the Ninth Circuit’s legal test for identifying such insiders (whether the transaction was conducted at arm’s length, i.e., as though the two parties were strangers) the mixed question became: Given all the basic facts found, was Rabkin’s purchase of MBP’s claim conducted as if the two were strangers to each other? That is about as factual sounding as any mixed question gets. Such an inquiry primarily belongs in the court that has presided over the presentation of evidence, that has heard all the witnesses, and that has both the closest and deepest understanding of the record—i.e., the bankruptcy court. One can arrive at the same point by asking how much legal work applying the arm’s-length test requires. It is precious little—as shown by judicial opinions applying the familiar legal term without further elaboration. Appellate review of the arm’s-length issue—even if conducted de novo—will not much clarify legal principles or provide guidance to other courts resolving other disputes. The issue is therefore one that primarily rests with a bankruptcy court, subject only to review for clear error. Pp. 5–11. 814 F. 3d 993, affirmed. Kagan, J., delivered the opinion for a unanimous Court. Kennedy, J., filed a concurring opinion. Sotomayor, J., filed a concurring opinion, in which Kennedy, Thomas, and Gorsuch, JJ., joined.

The Bankruptcy Code places various restrictions on anyone who qualifies as an “insider” of a debtor. The statutory definition of that term lists a set of persons related to the debtor in particular ways. See 11 U. S. C. §101(31). Courts have additionally recognized as insiders some persons not on that list—commonly known as “non-statutory insiders.” The conferral of that status often turns on whether the person’s transactions with the debtor (or another of its insiders) were at arm’s length. In this case, we address how an appellate court should review that kind of determination: de novo or for clear error? We hold that a clear-error standard should apply. I Chapter 11 of the Bankruptcy Code enables a debtor company to reorganize its business under a court-approved plan governing the distribution of assets to creditors. See 11 U. S. C. §1101 et seq. The plan divides claims against the debtor into discrete “classes” and specifies the “treatment” each class will receive. §1123; see §1122. Usually, a bankruptcy court may approve such a plan only if every affected class of creditors agrees to its terms. See §1129(a)(8). But in certain circumstances, the court may confirm what is known as a “cramdown” plan—that is, a plan impairing the interests of some non-consenting class. See §1129(b). Among the prerequisites for judicial approval of a cramdown plan is that another impaired class of creditors has consented to it. See §1129(a)(10). But crucially for this case, the consent of a creditor who is also an “insider” of the debtor does not count for that purpose. See ibid. (requiring “at least one” impaired class to have “accepted the plan, determined with-out including any acceptance of the plan by any insider”). The Code enumerates certain insiders, but courts have added to that number. According to the Code’s defini- tional section, an insider of a corporate debtor “includes” any director, officer, or “person in control” of the entity. §§101(31)(B)(i)–(iii). Because of the word “includes” in that section, courts have long viewed its list of insiders as non-exhaustive. See §102(3) (stating as one of the Code’s “[r]ules of construction” that “ ‘includes’ and ‘including’ are not limiting”); 2 A. Resnick & H. Sommer, Collier on Bankruptcy ¶101.31, p. 101–142 (16th ed. 2016) (discussing cases). Accordingly, courts have devised tests for identifying other, so-called “non-statutory” insiders. The decisions are not entirely uniform, but many focus, in whole or in part, on whether a person’s “transaction of business with the debtor is not at arm’s length.” Ibid. (quoting In re U. S. Medical, Inc., 531 F. 3d 1272, 1280 (CA10 2008)). This case came about because the Code’s list of insiders placed an obstacle in the way of respondent Lakeridge’s attempt to reorganize under Chapter 11. Lakeridge is a corporate entity which, at all relevant times, had a single owner, MBP Equity Partners, and a pair of substantial debts. The company owed petitioner U. S. Bank over $10 million for the balance due on a loan. And it owed MBP another $2.76 million. In 2011, Lakeridge filed for Chapter 11 bankruptcy. The reorganization plan it submitted placed its two creditors in separate classes and proposed to impair both of their interests. U. S. Bank refused that offer, thus taking a fully consensual plan off the table. But likewise, a cramdown plan based only on MBP’s consent could not go forward. Recall that an insider cannot provide the partial agreement needed for a cramdown plan. See supra, at 2; §1129(a)(10). And MBP was the consummate insider: It owned Lakeridge and so was—according to the Code’s definition—“in control” of the debtor. §101(31)(B)(iii). The path to a successful reorganization was thus impeded, and Lakeridge was faced with liquidation. Unless . . . Unless MBP could transfer its claim against Lakeridge to a non-insider who would then agree to the reorganization plan. So that was what MBP attempted. Kathleen Bartlett, a member of MBP’s board and an officer of Lake- ridge, approached Robert Rabkin, a retired surgeon, and offered to sell him MBP’s $2.76 million claim for $5,000. Rabkin took the deal. And as the new holder of MBP’s old loan, he consented to Lakeridge’s proposed reorganization. As long as he was not himself an insider, Rabkin’s agreement would satisfy one of the prerequisites for a cramdown plan. See §1129(a)(10); supra, at 2. That would bring Lakeridge a large step closer to reorganizing its business over U. S. Bank’s objection. Hence commenced this litigation about whether Rabkin, too, was an insider. U. S. Bank argued that he qualified as a non-statutory insider because he had a “romantic” relationship with Bartlett and his purchase of MBP’s loan “was not an arm’s-length transaction.” Motion to Designate Claim of Robert Rabkin as an Insider Claim in No. 11–51994 (Bkrtcy. Ct. Nev.), Doc. 194, p. 11 (Motion).[1] At an evidentiary hearing, both Rabkin and Bartlett testified that their relationship was indeed “romantic.” App. 128, 142–143.[2] But the Bankruptcy Court still rejected U. S. Bank’s view that Rabkin was a non-statutory insider. See App. to Pet. for Cert. 66a. The court found that Rabkin purchased the MBP claim as a “speculative investment” for which he did adequate due diligence. Id., at 67a. And it noted that Rabkin and Bartlett, for all their dating, lived in separate homes and managed their finances independently. See id., at 66a. The Court of Appeals for the Ninth Circuit affirmed by a divided vote. According to the court, a creditor qualifies as a non-statutory insider if two conditions are met: “(1) the closeness of its relationship with the debtor is comparable to that of the enumerated insider classifications in [the Code], and (2) the relevant transaction is negotiated at less than arm’s length.” In re Village at Lakeridge, LLC, 814 F. 3d 993, 1001 (2016). The majority viewed the Bankruptcy Court’s decision as based on a finding that the relevant transaction here (Rabkin’s purchase of MBP’s claim) “was conducted at arm’s length.” Id., at 1003, n. 15. That finding, the majority held, was entitled to clear-error review, and could not be reversed under that deferential standard. See id., at 1001–1003. Rabkin’s consent could therefore support the cramdown plan. See id., at 1003. Judge Clifton dissented. He would have applied de novo review, but in any event thought the Bankruptcy Court committed clear error in declining to classify Rabkin as an insider. See id., at 1006. This Court granted certiorari to decide a single question: Whether the Ninth Circuit was right to review for clear error (rather than de novo) the Bankruptcy Court’s determination that Rabkin does not qualify as a non-statutory insider because he purchased MBP’s claim in an arm’s-length transaction. 580 U. S. ___ (2017). II To decide whether a particular creditor is a non-statutory insider, a bankruptcy judge must tackle three kinds of issues—the first purely legal, the next purely factual, the last a combination of the other two. And to assess the judge’s decision, an appellate court must consider all its component parts, each under the appropriate standard of review. In this case, only the standard for the final, mixed question is contested. But to resolve that dispute, we begin by describing the unalloyed legal and factual questions that both kinds of courts have to address along the way, as well as the answers that the courts below provided. Initially, a bankruptcy court must settle on a legal test to determine whether someone is a non-statutory insider (again, a person who should be treated as an insider even though he is not listed in the Bankruptcy Code). But that choice of standard really resides with the next court: As all parties agree, an appellate panel reviews such a legal conclusion without the slightest deference. See Highmark Inc. v. Allcare Health Management. System, Inc., 572 U. S. ___, ___ (2014) (slip op., at 4) (“Traditionally, decisions on questions of law are reviewable de novo” (internal quotation marks omitted)); Tr. of Oral Arg. 29–30, 33. The Ninth Circuit here, as noted earlier, endorsed a two-part test for non-statutory insider status, asking whether the person’s relationship with the debtor was similar to those of listed insiders and whether the relevant prior transaction was at “less than arm’s length.” 814 F. 3d, at 1001; see supra, at 4–5. And the Ninth Circuit held that the Bankruptcy Court had used just that standard—more specifically, that it had denied insider status under the test’s second, transactional prong. See 814 F. 3d, at 1002–1003, and n. 15; supra, at 4–5. We do not address the correctness of the Ninth Circuit’s legal test; indeed, we specifically rejected U. S. Bank’s request to include that question in our grant of certiorari. See 580 U. S. ___; Pet. for Cert. i. We simply take that test as a given in deciding the standard-of-review issue we chose to resolve. Along with adopting a legal standard, a bankruptcy court evaluating insider status must make findings of what we have called “basic” or “historical” fact—addressing questions of who did what, when or where, how or why. Thompson v. Keohane, 516 U. S. 99, 111 (1995) . The set of relevant historical facts will of course depend on the legal test used: So under the Ninth Circuit’s test, the facts found may relate to the attributes of a particular relationship or the circumstances and terms of a prior transaction. By well-settled rule, such factual findings are reviewable only for clear error—in other words, with a serious thumb on the scale for the bankruptcy court. See Fed. Rule Civ. Proc. 52(a)(6) (clear-error standard); Fed. Rules Bkrtcy. Proc. 7052 and 9014(c) (applying Rule 52 to various bankruptcy proceedings). Accordingly, as all parties again agree, the Ninth Circuit was right to review deferentially the Bankruptcy Court’s findings about Rabkin’s relationship with Bartlett (e.g., that they did not “cohabitate” or pay each other’s “bills or living expenses”) and his motives for purchasing MBP’s claim (e.g., to make a “speculative investment”). App. to Pet. for Cert. 66a–67a; see Tr. of Oral Arg. 8, 39. What remains for a bankruptcy court, after all that, is to determine whether the historical facts found satisfy the legal test chosen for conferring non-statutory insider status. We here arrive at the so-called “mixed question” of law and fact at the heart of this case. Pullman-Standard v. Swint, 456 U.S. 273, 289, n. 19 (1982) (A mixed question asks whether “the historical facts . . . satisfy the statutory standard, or to put it another way, whether the rule of law as applied to the established facts is or is not violated”). As already described, the Bankruptcy Court below had found a set of basic facts about Rabkin; and it had adopted a legal test for non-statutory insider status that requires (as one of its two prongs) a less-than-arm’s-length transaction. See supra, at 4, 6. As its last move, the court compared the one to the other—and determined that the facts found did not show the kind of preferential transaction necessary to turn a creditor into a non-statutory insider. For that decisive determination, what standard of review should apply? The parties, after traveling so far together, part ways at this crucial point. U. S. Bank contends that the Bankruptcy Court’s resolution of the mixed question must be reviewed de novo. That is because, U. S. Bank claims, application of the Ninth Circuit’s “very general” standard to a set of basic facts requires the further elaboration of legal principles—a task primarily for appellate courts. Brief for Petitioner 35; see id., at 53 (The “open-ended nature of the Ninth Circuit’s standard” compels courts to “develop the norms and criteria they deem most appropriate” and so should be viewed as “quasi-legal”). By contrast, Lakeridge (joined by the Federal Government as amicus curiae) thinks a clear-error standard should apply. In Lakeridge’s view, the ultimate law-application question is all “bound up with the case-specific details of the highly factual circumstances below”—and thus falls naturally within the domain of bankruptcy courts. Brief for Respondent 17; see Brief for United States 21 (similarly describing the mixed question as “fact-intensive”). For all their differences, both parties rightly point us to the same query: What is the nature of the mixed question here and which kind of court (bankruptcy or appellate) is better suited to resolve it? See Miller v. Fenton, 474 U. S. 104, 114 (1985) (When an “issue falls somewhere between a pristine legal standard and a simple historical fact,” the standard of review often reflects which “judicial actor is better positioned” to make the decision).[3] Mixed questions are not all alike. As U. S. Bank suggests, some require courts to expound on the law, particularly by amplifying or elaborating on a broad legal standard. When that is so—when applying the law involves developing auxiliary legal principles of use in other cases—appellate courts should typically review a decision de novo. See Salve Regina College v. Russell, 499 U. S. 225 –233 (1991) (discussing appellate courts’ “institutional advantages” in giving legal guidance). But as Lakeridge replies, other mixed questions immerse courts in case-specific factual issues—compelling them to marshal and weigh evidence, make credibility judgments, and otherwise address what we have (emphatically if a tad redundantly) called “multifarious, fleeting, special, narrow facts that utterly resist generalization.” Pierce v. Underwood, 487 U. S. 552 –562 (1988) (internal quotation marks omitted). And when that is so, appellate courts should usually review a decision with deference. See Anderson v. Bessemer City, 470 U. S. 564 –576 (1985) (discussing trial courts’ “superiority” in resolving such issues).[4] In short, the standard of review for a mixed question all depends—on whether answering it entails primarily legal or factual work. Now again, recall the mixed question the Bankruptcy Court confronted in this case. See supra, at 7. At a high level of generality, the court needed to determine whether the basic facts it had discovered (concerning Rabkin’s relationships, motivations, and so on) were sufficient to make Rabkin a non-statutory insider. But the court’s use of the Ninth Circuit’s legal test for identifying such in- siders reduced that question to a more particular one: whether the facts found showed an arm’s-length transaction between Rabkin and MBP. See ibid.[5] And still, we can further delineate that issue just by plugging in the widely (universally?) understood definition of an arm’s-length transaction: a transaction conducted as though the two parties were strangers. See, e.g., Black’s Law Dictionary 1726 (10th ed. 2014). Thus the mixed question becomes: Given all the basic facts found, was Rabkin’s purchase of MBP’s claim conducted as if the two were strangers to each other? That is about as factual sounding as any mixed question gets. Indeed, application of the Ninth Circuit’s arm’s-length legal standard really requires what we have previously described as a “factual inference[ ] from undisputed basic facts.” Commissioner v. Duberstein, 363 U. S. 278, 291 (1960) (holding that clear-error review applied to a decision that a particular transfer was a statutory “gift”). The court takes a raft of case-specific historical facts,[6] considers them as a whole, balances them one against another—all to make a determination that when two particular persons entered into a particular transaction, they were (or were not) acting like strangers. Just to describe that inquiry is to indicate where it (primarily) belongs: in the court that has presided over the presentation of evidence, that has heard all the witnesses, and that has both the closest and the deepest understanding of the record—i.e., the bankruptcy court. And we can arrive at the same point from the opposite direction—by asking how much legal work applying the arm’s-length test requires. Precious little, in our view—as shown by judicial opinions addressing that concept. Our own decisions, arising in a range of contexts, have never tried to elaborate on the established idea of a transaction conducted as between strangers; nor, to our knowledge, have lower courts. See, e.g., Jones v. Harris Associates L. P., 559 U. S. 335, 346 (2010) ; Commissioner v. Wemyss, 324 U. S. 303, 307 (1945) ; Pepper v. Litton, 308 U. S. 295 –307 (1939). The stock judicial method is merely to state the requirement of such a transaction and then to do the fact-intensive job of exploring whether, in a particular case, it occurred. See, e.g., Wemyss, 324 U. S., at 307. Contrary to U. S. Bank’s view, there is no apparent need to further develop “norms and criteria,” or to devise a supplemental multi-part test, in order to apply the familiar term. Brief for Petitioner 53; see Tr. of Oral Arg. 18; supra, at 7. So appellate review of the arm’s-length issue—even if conducted de novo—will not much clarify legal principles or provide guidance to other courts resolving other disputes. And that means the issue is not of the kind that appellate courts should take over.[7] The Court of Appeals therefore applied the appropriate standard in reviewing the Bankruptcy Court’s determination that Rabkin did not qualify as an insider because his transaction with MBP was conducted at arm’s length. A conclusion of that kind primarily rests with a bankruptcy court, subject only to review for clear error. We accordingly affirm the judgment below. It is so ordered. Notes 1 U. S. Bank also contended that Rabkin automatically inherited MBP’s statutory insider status when he purchased its loan. See Motion, p. 10 (“[A]n entity which acquires a claim steps into the shoes of that claimant” (internal quotation marks omitted)). We did not grant review of that question and therefore do not address it in this opinion. 2 Perhaps Bartlett expressed some ambivalence on that score. The transcript of her direct examination reads: “Q. Okay. And I think the term has been a romantic relationship—you have a romantic relationship? A. I guess. Q. Why do you say I guess? A. Well, no—yes.” App. 142–143. One hopes Rabkin was not listening. 3 In selecting standards of review, our decisions have also asked whether a “long history of appellate practice” supplies the answer. Pierce v. Underwood, 487 U. S. 552, 558 (1988) . But we cannot find anything resembling a “historical tradition” to provide a standard for reviewing the mixed question here. Ibid. 4 Usually but not always: In the constitutional realm, for example, the calculus changes. There, we have often held that the role of appellate courts “in marking out the limits of [a] standard through the process of case-by-case adjudication” favors de novo review even when answering a mixed question primarily involves plunging into a factual record. Bose Corp. v. Consumers Union of United States, Inc., 466 U. S. 485, 503 (1984) ; see Ornelas v. United States, 517 U. S. 690, 697 (1996) (reasonable suspicion and probable cause under the Fourth Amendment); Hurley v. Irish-American Gay, Lesbian and Bisexual Group of Boston, Inc., 515 U. S. 557, 567 (1995) (expression under the First Amendment); Miller v. Fenton, 474 U. S. 104 –116 (1985) (voluntariness of confession under the Fourteenth Amendment’s Due Process Clause). 5 A bankruptcy court applying the Ninth Circuit’s test might, in another case, reach its separate, non-transactional prong: whether “the closeness of [a person’s] relationship with the debtor is comparable to that of the enumerated insider classifications” in the Code. In re Village at Lakeridge, LLC, 814 F. 3d 993, 1001 (2016); see supra, at 4. We express no opinion on how an appellate court should review a bankruptcy court’s application of that differently framed standard to a set of established facts. 6 Or, to use the more abundant description we quoted above, “multifarious, fleeting, special, narrow facts that utterly resist generalization.” Pierce, 487 U. S., at 561–562 (internal quotation marks omitted); see supra, at 8. 7 That conclusion still leaves some role for appellate courts in this area. They of course must decide whether a bankruptcy court committed clear error in finding that a transaction was arm’s length (or not). (We express no view of that aspect of the Ninth Circuit’s decision because we did not grant certiorari on the question. See supra, at 5.) In addition, an appellate court must correct any legal error infecting a bankruptcy court’s decision. So if the bankruptcy court somehow misunderstood the nature of the arm’s-length query—or if it devised some novel multi-factor test for addressing that issue—an appellate court should apply de novo review. And finally, if an appellate court someday finds that further refinement of the arm’s-length standard is necessary to maintain uniformity among bankruptcy courts, it may step in to perform that legal function. By contrast, what it may not do is review independently a garden-variety decision, as here, that the various facts found amount to an arm’s-length (or a non-arm’s-length) transaction and so do not (or do) confer insider status.

584.US.2017_17-312

The judges of the United States District Court for the Southern District of California adopted a districtwide policy permitting the use of full restraints—handcuffs connected to a waist chain, with legs shackled—on most in-custody defendants produced in court for nonjury proceedings by the United States Marshals Service. Respondents Jasmin Morales, Rene Sanchez-Gomez, Moises Patricio-Guzman, and Mark Ring challenged the use of such restraints in their respective cases and the restraint policy as a whole. The District Court denied their challenges, and respondents appealed to the Court of Appeals for the Ninth Circuit. Before that court could issue a decision, respondents’ underlying criminal cases ended. The court—viewing the case as a “functional class action” involving “class-like claims” seeking “class-like relief,” 859 F. 3d 649, 655, 657–658—held that this Court’s civil class action precedents saved the case from mootness. On the merits, the Court of Appeals held the policy unconstitutional. Held: This case is moot. Pp. 3–12. (a) The federal judiciary may adjudicate only “actual and concrete disputes, the resolutions of which have direct consequences on the parties involved.” Genesis HealthCare Corp. v. Symczyk, 569 U. S. 66, 71. Such a dispute “must be extant at all stages of review, not merely at the time the complaint is filed.” Preiser v. Newkirk, 422 U. S. 395, 401. A case that becomes moot at any point during the proceedings is thus outside the jurisdiction of the federal courts. See Already, LLC v. Nike, Inc., 568 U. S. 85, 91. Pp. 3–4. (b) In concluding that this case was not moot, the Court of Appeals relied upon this Court’s class action precedents, most prominently Gerstein v. Pugh, 420 U. S. 103. That reliance was misplaced. Gerstein was a class action respecting pretrial detention brought under Federal Rule of Civil Procedure 23. The named class representatives’ individual claims had apparently become moot before class certification. This Court held that the case could nonetheless proceed, explaining that due to the inherently temporary nature of pretrial detention, no named representative might be in custody long enough for a class to be certified. Gerstein does not support a freestanding exception to mootness outside the class action context. It belongs to a line of cases that this Court has described as turning on the particular traits of Rule 23 class actions. See, e.g., Sosna v. Iowa, 419 U. S. 393; United States Parole Comm’n v. Geraghty, 445 U. S. 388; Genesis HealthCare, 569 U. S. 66. The Federal Rules of Criminal Procedure establish for criminal cases no vehicle comparable to the civil class action, and this Court has never permitted criminal defendants to band together to seek prospective relief in their individual cases on behalf of a class. Here, the mere presence of allegations that might, if resolved in respondents’ favor, benefit other similarly situated individuals cannot save their case from mootness. See id., at 73. That conclusion is unaffected by the Court of Appeals’ decision to recast respondents’ appeals as petitions for supervisory mandamus. Pp. 4–9. (c) Respondents do not defend the reasoning of the Court of Appeals, and instead argue that the claims of two respondents—Sanchez-Gomez and Patricio-Guzman—fall within the “exception to the mootness doctrine for a controversy that is capable of repetition, yet evading review.” Kingdomware Technologies, Inc. v. United States, 579 U. S. ___, ___ (internal quotation marks omitted). Respondents claim that the exception applies because Sanchez-Gomez and Patricio-Guzman will again violate the law, be apprehended, and be returned to pretrial custody. But this Court has consistently refused to “conclude that the case-or-controversy requirement is satisfied by” the possibility that a party “will be prosecuted for violating valid criminal laws.” O’Shea v. Littleton, 414 U. S. 488, 497. Respondents argue that this usual refusal to assume future criminal conduct is unwarranted here given the particular circumstances of Sanchez-Gomez’s and Patricio-Guzman’s offenses. They cite two civil cases—Honig v. Doe, 484 U. S. 305, and Turner v. Rogers, 564 U. S. 431—in which this Court concluded that the expectation that a litigant would repeat the misconduct that gave rise to his claims rendered those claims capable of repetition. But Honig and Turner are inapposite because they concerned litigants unable, for reasons beyond their control, to prevent themselves from transgressing and avoid recurrence of the challenged conduct. Sanchez-Gomez and Patricio-Guzman, in contrast, are “able—and indeed required by law”—to refrain from further criminal conduct. Lane v. Williams, 455 U. S. 624, 633, n. 13. No departure from the settled rule is warranted. Pp. 9–12. 859 F. 3d 649, vacated and remanded. Roberts, C. J., delivered the opinion for a unanimous Court.

Four criminal defendants objected to being bound by full restraints during pretrial proceedings in their cases, but the District Court denied relief. On appeal, the Court of Appeals for the Ninth Circuit held that the use of such restraints was unconstitutional, even though each of the four criminal cases had ended prior to its decision. The question presented is whether the appeals were saved from mootness either because the defendants sought “class-like relief” in a “functional class action,” or because the challenged practice was “capable of repetition, yet evading review.” I It is the responsibility of the United States Marshals Service to “provide for the security . . . of the United States District Courts.” 28 U. S. C. §566(a). To fulfill that duty, the United States Marshal for the Southern District of California requested that the judges of that district permit the use of full restraints on all in-custody defendants during nonjury proceedings. When “full restraints” are applied, “a defendant’s hands are closely handcuffed together, these handcuffs are connected by chain to another chain running around the defendant’s waist, and the defendant’s feet are shackled and chained together.” 859 F. 3d 649, 653 (CA9 2017) (en banc). In support of his proposal, the Marshal cited safety concerns arising from understaffing, past incidents of violence, and the high volume of in-custody defendants produced in the Southern District. The judges agreed to the Marshal’s request, with modifications providing that a district or magistrate judge may require a defendant to be produced without restraints, and that a defendant can request that this be done. See App. 78–79. Respondents Jasmin Morales, Rene Sanchez-Gomez, Moises Patricio-Guzman, and Mark Ring were among the defendants produced by the Marshals Service for pretrial proceedings in full restraints. They raised constitutional objections to the use of such restraints in their respective cases, and to the restraint policy as a whole. They noted that the policy had resulted in the imposition of full restraints on, for example, a woman with a fractured wrist, a man with a severe leg injury, a blind man, and a wheelchair-bound woman. The District Court denied their challenges. Respondents appealed to the Court of Appeals for the Ninth Circuit, but before the court could issue a decision, their underlying criminal cases came to an end. Morales, Sanchez-Gomez, and Patricio-Guzman each pled guilty to the offense for which they were charged: Morales, to felony importation of a controlled substance, in violation of 21 U. S. C. §§952 and 960; Sanchez-Gomez, to felony misuse of a passport, in violation of 18 U. S. C. §1544; and Patricio-Guzman, to misdemeanor illegal entry into the United States, in violation of 8 U. S. C. §1325. The charges against Ring—for making an interstate threat in violation of 18 U. S. C. §875(c)—were dismissed pursuant to a deferred-prosecution agreement. A panel of the Court of Appeals nonetheless concluded that respondents’ claims were not moot, and went on to strike down the restraint policy as violating the Due Process Clause of the Fifth Amendment. 798 F. 3d 1204 (CA9 2015). Those rulings were reaffirmed on rehearing en banc. 859 F. 3d 649. The en banc court understood the “main dispute” before it to be a challenge to the policy itself, not just to the application of that policy to respondents. Id., at 655. The court then construed respondents’ notices of appeal as petitions for mandamus, which invoked the court’s supervisory authority over the Southern District. Id., at 657. The case was, in the court’s view, a “functional class action” involving “class-like claims” seeking “class-like relief.” Id., at 655, 657–658. In light of that understanding, the Court of Appeals held that this Court’s civil class action precedents kept the case alive, even though respondents were no longer subject to the restraint policy. Id., at 657–659 (citing Gerstein v. Pugh, 420 U. S. 103, 110–111, n. 11 (1975)). On the merits, the Court of Appeals concluded that the restraint policy violated the Constitution. 859 F. 3d, at 666. Judge Ikuta, writing in dissent for herself and four colleagues, rejected the majority’s application of class action precedents to the individual criminal cases before the court and would have held the case moot. Id., at 675. She also disagreed with the majority on the merits, concluding that the restraint policy did not violate the Constitution. Id., at 683. We granted certiorari. 583 U. S. ___ (2017). II To invoke federal jurisdiction, a plaintiff must show a “personal stake” in the outcome of the action. Genesis HealthCare Corp. v. Symczyk, 569 U. S. 66, 71 (2013). “This requirement ensures that the Federal Judiciary confines itself to its constitutionally limited role of adjudicating actual and concrete disputes, the resolutions of which have direct consequences on the parties involved.” Ibid. Such a dispute “must be extant at all stages of review, not merely at the time the complaint is filed.” Preiser v. Newkirk, 422 U. S. 395, 401 (1975). A case that becomes moot at any point during the proceedings is “no longer a ‘Case’ or ‘Controversy’ for purposes of Article III,” and is outside the jurisdiction of the federal courts. Already, LLC v. Nike, Inc., 568 U. S. 85, 91 (2013). A In concluding that this case was not moot, the Court of Appeals relied upon our class action precedents, most prominently Gerstein v. Pugh. That reliance was misplaced.[1]* Gerstein, a class action brought under Federal Rule of Civil Procedure 23, involved a certified class of detainees raising claims concerning their pretrial detention. 420 U. S., at 106–107. By the time this Court heard the case, the named representatives’ claims were moot, and the record suggested that their interest might have lapsed even before the District Court certified the class. See id., at 110–111, n. 11. Normally a class action would be moot if no named class representative with an unexpired claim remained at the time of class certification. See ibid. (citing Sosna v. Iowa, 419 U. S. 393, 402, n. 11 (1975)). The Court nevertheless held that the case remained live. As we explained, pretrial custody was inherently temporary and of uncertain length, such that we could not determine “that any given individual, named as plaintiff, would be in pretrial custody long enough for a district judge to certify the class.” Gerstein, 420 U. S., at 110–111, n. 11. At the same time, it was certain that there would always be some group of detainees subject to the challenged practice. Ibid. Given these circumstances, the Court determined that the class action could proceed. Ibid.; see Swisher v. Brady, 438 U. S. 204, 213–214, n. 11 (1978) (employing same analysis in a class action challenging juvenile court procedures). The Court of Appeals interpreted Gerstein to cover all “cases sufficiently similar to class actions” in which, “because of the inherently transitory nature of the claims,” the claimant’s “interests would expire before litigation could be completed.” 859 F. 3d, at 658. Gerstein was an action brought under Federal Rule of Civil Procedure 23, but the Court of Appeals decided that such “a procedural mechanism to aggregate the claims” was not a “necessary prerequisite” for application of the Gerstein rule. 859 F. 3d, at 659 (alteration omitted). Respondents, the court noted, sought “relief [from the restraint policy] not merely for themselves, but for all in-custody defendants in the district.” Id., at 655. Those “class-like claims” seeking “class-like relief” were sufficient to trigger the application of Gerstein and save the case from mootness, despite the termination of respondents’ criminal cases. 859 F. 3d, at 655. We reject the notion that Gerstein supports a freestanding exception to mootness outside the class action context. The class action is a creature of the Federal Rules of Civil Procedure. See generally 7A C. Wright, A. Miller & M. Kane, Federal Practice and Procedure §1751 et seq. (3d ed. 2005). It is an “exception to the usual rule that litigation is conducted by and on behalf of the individual named parties only,” and “provides a procedure by which the court may exercise . . . jurisdiction over the various individual claims in a single proceeding.” Califano v. Yamasaki, 442 U. S. 682, 700–701 (1979). “The certification of a suit as a class action has important consequences for the unnamed members of the class.” Sosna, 419 U. S., at 399, n. 8. Those class members may be “bound by the judgment” and are considered parties to the litigation in many important respects. Devlin v. Scardelletti, 536 U. S. 1, 7, 9–10 (2002). A certified class thus “acquires a legal status separate from the interest asserted by the named plaintiff.” Genesis HealthCare, 569 U. S., at 74 (quoting Sosna, 419 U. S., at 399; alterations omitted). Gerstein belongs to a line of cases that we have described as turning on the particular traits of civil class actions. The first case in this line, Sosna v. Iowa, held that when the claim of the named plaintiff becomes moot after class certification, a “live controversy may continue to exist” based on the ongoing interests of the remaining unnamed class members. Genesis HealthCare, 569 U. S., at 74 (citing Sosna, 419 U. S., at 399–402); see Franks v. Bowman Transp. Co., 424 U. S. 747, 755–756 (1976). The “fact that a putative class acquires an independent legal status once it is certified” was, we later explained, “essential to our decision[ ] in Sosna.” Genesis HealthCare, 569 U. S., at 75; see Kremens v. Bartley, 431 U. S. 119, 131–133 (1977) (explaining that, under Sosna’s rule, “only a ‘properly certified’ class . . . may succeed to the adversary position of a named representative whose claim becomes moot”); Alvarez v. Smith, 558 U. S. 87, 92–93 (2009) (same). Gerstein, announced one month after Sosna, provides a limited exception to Sosna’s requirement that a named plaintiff with a live claim exist at the time of class certification. The exception applies when the pace of litigation and the inherently transitory nature of the claims at issue conspire to make that requirement difficult to fulfill. See Sosna, 419 U. S., at 402, n. 11 (anticipating the Gerstein rule as an exception); Gerstein, 420 U. S., at 110–111, n. 11 (describing its holding as “a suitable exception” to Sosna). We have repeatedly tied Gerstein’s rule to the class action setting from which it emerged. See, e.g., Genesis HealthCare, 569 U. S., at 71, n. 2 (describing Gerstein’s rule as “developed in the context of class actions under Rule 23 to address the circumstance in which a named plaintiff’s claim becomes moot prior to certification of the class”); United States Parole Comm’n v. Geraghty, 445 U. S. 388, 397–399 (1980) (highlighting Gerstein as an example of the Court “consider[ing] the application of the ‘personal stake’ requirement in the class-action context”). In concluding that Gerstein reaches further, the Court of Appeals looked to our recent decision in Genesis HealthCare Corp. v. Symczyk. But in that case the Court refused to extend Gerstein beyond the class action context, even with respect to a procedural device bearing many features similar to a class action. Genesis HealthCare addressed whether a “collective action” brought under the Fair Labor Standards Act (FLSA) by a plaintiff on behalf of herself “and other ‘similarly situated’ employees” remained “justiciable when the lone plaintiff’s individual claim bec[ame] moot.” 569 U. S., at 69. In an effort to continue her case on behalf of others, the plaintiff turned to Sosna and its progeny, including Gerstein. But those cases, we explained, were “inapposite,” not least because “Rule 23 actions are fundamentally different from collective actions under the FLSA.” Genesis HealthCare, 569 U. S., at 74. Such collective actions, we stressed, do not “produce a class with an independent legal status, or join additional parties to the action.” Id., at 75. This case, which does not involve any formal mechanism for aggregating claims, is even further removed from Rule 23 and Gerstein. The Federal Rules of Criminal Procedure establish for criminal cases no vehicle comparable to the FLSA collective action, much less the class action. And we have never permitted criminal defendants to band to- gether to seek prospective relief in their individual criminal cases on behalf of a class. As we said when declining to apply nonparty preclusion outside the formal class action context, courts may not “recognize . . . a common-law kind of class action” or “create de facto class actions at will.” Taylor v. Sturgell, 553 U. S. 880, 901 (2008) (alterations omitted); see Smith v. Bayer Corp., 564 U. S. 299, 315–316 (2011) (same); Pasadena City Bd. of Ed. v. Spangler, 427 U. S. 424, 430 (1976) (rejecting in mootness context the idea that “the failure to obtain the class certification required under Rule 23 is merely the absence of a meaningless ‘verbal recital’ ”). The court below designated respondents’ case a “functional class action” because respondents were pursuing relief “not merely for themselves, but for all in-custody defendants in the district.” 859 F. 3d, at 655, 657–658. But as explained in Genesis HealthCare, the “mere presence of . . . allegations” that might, if resolved in respondents’ favor, benefit other similarly situated individuals cannot “save [respondents’] suit from mootness once the[ir] individual claim[s]” have dissipated. 569 U. S., at 73. Our conclusion is unaffected by the decision of the court below to recast respondents’ appeals as petitions for “supervisory mandamus.” See 859 F. 3d, at 659 (viewing such a petition, like the civil class action, as a procedural vehicle to which the Gerstein rule applies). Supervisory mandamus refers to the authority of the Courts of Appeals to exercise “supervisory control of the District Courts” through their “discretionary power to issue writs of mandamus.” La Buy v. Howes Leather Co., 352 U. S. 249, 259–260 (1957). There is no sign in our scant supervisory mandamus precedents that such cases are exempt from the normal mootness rules. See generally Will v. United States, 389 U. S. 90 (1967); Schlagenhauf v. Holder, 379 U. S. 104 (1964); La Buy, 352 U. S. 249. Indeed, as the court below acknowledged, “[s]upervisory mandamus cases require live controversies.” 859 F. 3d, at 657. B Respondents do not defend the reasoning of the Court of Appeals. See Brief for Respondents 58 (arguing that this Court need not reach the functional class action issue and should “discard[ ]” that label); Tr. of Oral Arg. 43 (respondents’ counsel agreeing that they “have not made any effort to defend” the functional class action approach). In respondents’ view, functional class actions and Gerstein’s rule are beside the point because two respondents—Sanchez-Gomez and Patricio-Guzman—retain a personal stake in the outcome of their appeals. Sanchez-Gomez and Patricio-Guzman are no longer in pretrial custody. Their criminal cases, arising from their illegal entry into the United States, ended in guilty pleas well before the Court of Appeals issued its decision. Respondents contend, however, that the claims brought by Sanchez-Gomez and Patricio-Guzman fall within the “exception to the mootness doctrine for a controversy that is capable of repetition, yet evading review.” Kingdomware Technologies, Inc. v. United States, 579 U. S. ___, ___ (2016) (slip op., at 7) (internal quotation marks omitted). A dispute qualifies for that exception only “if (1) the challenged action is in its duration too short to be fully liti- gated prior to its cessation or expiration, and (2) there is a reasonable expectation that the same complaining party will be subjected to the same action again.” Turner v. Rogers, 564 U. S. 431, 439–440 (2011) (alterations and internal quotation marks omitted). The parties do not contest that the claims at issue satisfy the first prong of that test, but they sharply disagree as to the second. Respondents argue that Sanchez-Gomez and Patricio-Guzman meet the second prong because they will again violate the law, be apprehended, and be returned to pretrial custody. But we have consistently refused to “conclude that the case-or-controversy requirement is satisfied by” the possibility that a party “will be prosecuted for violating valid criminal laws.” O’Shea v. Littleton, 414 U. S. 488, 497 (1974). We have instead “assume[d] that [litigants] will conduct their activities within the law and so avoid prosecution and conviction as well as exposure to the challenged course of conduct.” Ibid.; see, e.g., Spencer v. Kemna, 523 U. S. 1, 15 (1998) (reasoning that a claim regarding a parole revocation order was moot following release from custody because any continuing consequences of the order were “contingent upon [the claimant] violating the law, getting caught, and being convicted”); Honig v. Doe, 484 U. S. 305, 320 (1988) (“[W]e generally have been unwilling to assume that the party seeking relief will repeat the type of misconduct that would once again place him or her at risk of that injury.”); Lane v. Williams, 455 U. S. 624, 632–633, n. 13 (1982) (concluding that case was moot where the challenged parole revocation could not “affect a subsequent parole determination unless respondents again violate state law, are returned to prison, and become eligible for parole”). Respondents argue that this usual refusal to assume future criminal conduct is unwarranted here given the particular circumstances of Sanchez-Gomez’s and Patricio-Guzman’s offenses. They cite two civil cases—Honig v. Doe and Turner v. Rogers—in which this Court concluded that the expectation that a litigant would repeat the misconduct that gave rise to his claims rendered those claims capable of repetition. Neither case, however, supports a departure from the settled rule. Honig involved a disabled student’s challenge to his suspension from school for disruptive behavior. We found that given his “inability to conform his conduct to socially acceptable norms” or “govern his aggressive, impulsive behavior,” it was “reasonable to expect that [the student would] again engage in the type of misconduct that precipitated this suit” and “be subjected to the same unilateral school action for which he initially sought relief.” 484 U. S., at 320–321. In Turner, we determined that an indigent person repeatedly held in civil contempt for failing to make child support payments, who was at the time over $13,000 in arrears, and whose next hearing was only five months away, was destined to find himself in civil contempt proceedings again. The challenged denial of appointed counsel at his contempt hearing was thus capable of repetition. See 564 U. S., at 440. Respondents contend that Sanchez-Gomez and Patricio-Guzman, like the challengers in Honig and Turner, are likely to find themselves right back where they started if we dismiss their case as moot. Respondents cite a Sentencing Commission report finding that in 2013 thirty-eight percent of those convicted and sentenced for an illegal entry or illegal reentry offense “were deported and subsequently illegally reentered at least one time.” United States Sentencing Commission, Illegal Reentry Offenses 15 (2015) (cited by Brief for Respondents 51). Respondents emphasize the economic and familial pressures that often compel individuals such as Sanchez-Gomez and Patricio-Guzman to repeatedly attempt to enter the United States. And respondents note that both men, after their release, actually did cross the border into the United States, were apprehended again, and were charged with new illegal entry offenses. All this, respondents say, adds up to a sufficient showing that Sanchez-Gomez and Patricio-Guzman satisfy the “capable of repetition” require- ment. Because the Court of Appeals was not aware that Sanchez-Gomez and Patricio-Guzman had subsequently reentered the United States illegally, respondents invite us to remand this case for further proceedings. We decline to do so because Honig and Turner are inapposite. Our decisions in those civil cases rested on the litigants’ inability, for reasons beyond their control, to prevent themselves from transgressing and avoid recurrence of the challenged conduct. In Honig, such incapacity was the very reason the school sought to expel the student. And in Turner, the indigent individual’s large outstanding debt made him effectively incapable of satisfying his imminent support obligations. Sanchez-Gomez and Patricio-Guzman, in contrast, are “able—and indeed required by law”—to refrain from further criminal conduct. Lane, 455 U. S., at 633, n. 13. Their personal incentives to return to the United States, plus the elevated rate of recidivism associated with illegal entry offenses, do not amount to an inability to obey the law. We have consistently refused to find the case or controversy requirement satisfied where, as here, the litigants simply “anticipate violating lawful criminal statutes.” O’Shea, 414 U. S., at 496. III None of this is to say that those who wish to challenge the use of full physical restraints in the Southern District lack any avenue for relief. In the course of this litigation the parties have touched upon several possible options. See, e.g., Tr. of Oral Arg. 12 (indicating circumstances under which detainees could bring a civil suit). Because we hold this case moot, we take no position on the question. * * * We vacate the judgment of the Court of Appeals for the Ninth Circuit and remand the case to that court with instructions to dismiss as moot. It is so ordered. Notes 1 * Shortly after the panel decision in this case, the Southern District altered its policy to eliminate the routine use of full restraints in pretrial proceedings. The Government represents, however, that the Southern District intends to reinstate its policy once it is no longer bound by the decision of the Court of Appeals. Tr. of Oral Arg. 29. We agree with the Court of Appeals that the rescission of the policy does not render this case moot. A party “cannot automatically moot a case simply by ending its unlawful conduct once sued,” else it “could engage in unlawful conduct, stop when sued to have the case declared moot, then pick up where [it] left off, repeating this cycle until [it] achieves all [its] unlawful ends.” Already, LLC v. Nike, Inc., 568 U. S. 85, 91 (2013).

584.US.2017_17-387

The Upper Skagit Indian Tribe purchased a roughly 40-acre plot of land and then commissioned a boundary survey. The survey convinced the Tribe that about an acre of its land lay on the other side of a boundary fence between its land and land owned by Sharline and Ray Lundgren. The Lundgrens filed a quiet title action in Washington state court, invoking the doctrines of adverse possession and mutual acquiescence, but the Tribe asserted sovereign immunity from the suit. Ultimately, the State Supreme Court rejected the Tribe’s immunity claim and ruled for the Lundgrens, reasoning that, under County of Yakima v. Confederated Tribes and Bands of Yakima Nation, 502 U. S. 251, tribal sovereign immunity does not apply to in rem suits. Held: Yakima addressed not the scope of tribal sovereign immunity, but a question of statutory interpretation of the Indian General Allotment Act of 1887. That Act authorized the President to allot parcels of reservation land to individual tribal members and directed the United States eventually to issue fee patents to the allottees as private individuals. In 1934, Congress reversed course but made no attempt to withdraw the lands already conveyed. As a result, Indian reservations sometimes contain both trust land held by the United States and fee-patented land held by private parties. Yakima concerned the tax consequences of this intermixture. This Court had previously held that §6 of the General Allotment Act could no longer be read as allowing States to impose in personam taxes on transactions between Indians on fee-patented land within a reservation. Moe v. Confederated Salish and Kootenai Tribes of Flathead Reservation, 425 U. S. 463, 479–481. The Court reached a different conclusion in Yakima with respect to in rem state taxes, holding that the state collection of property taxes on fee-patented land within reservations was still allowed under §6. 502 U. S., at 265. In short, Yakima sought only to interpret a relic of a statute in light of a distinguishable precedent; it resolved nothing about the law of sovereign immunity. Acknowledging this, the Lundgrens now ask the Court to affirm on an alternative, common-law ground: that the Tribe cannot assert sovereign immunity because this suit relates to immovable property located in Washington State, purchased by the Tribe in the same manner as a private individual. Because this alternative argument did not emerge until late in this case, the Washington Supreme Court should address it in the first instance. Pp. 3–7. 187 Wash. 2d 857, 389 P. 3d 569, vacated and remanded. Gorsuch, J., delivered the opinion of the Court, in which Roberts, C. J., and Kennedy, Ginsburg, Breyer, Sotomayor, and Kagan, JJ., joined. Roberts, C. J., filed a concurring opinion, in which Kennedy, J., joined. Thomas, J., filed a dissenting opinion, in which Alito, J., joined.

Lower courts disagree about the significance of our decision in County of Yakima v. Confederated Tribes and Bands of Yakima Nation, 502 U. S. 251 (1992). Some think it means Indian tribes lack sovereign immunity in in rem lawsuits like this one; others don’t read it that way at all.[1]* We granted certiorari to set things straight. 583 U. S. ___ (2017). Ancestors of the Upper Skagit Tribe lived for centuries along the Skagit River in northwestern Washington State. But as settlers moved across the Cascades and into the region, the federal government sought to make room for them by displacing native tribes. In the treaty that followed with representatives of the Skagit people and others, the tribes agreed to “cede, relinquish, and convey” their lands to the United States in return for $150,000 and other promises. Treaty of Point Elliott, Jan. 22, 1855, 12Stat. 927; see Washington v. Washington State Commercial Passenger Fishing Vessel Assn., 443 U. S. 658, 676 (1979); United States v. Washington, 384 F. Supp. 312, 333 (WD Wash. 1974). Today’s dispute stems from the Upper Skagit Tribe’s efforts to recover a portion of the land it lost. In 1981, the federal government set aside a small reservation for the Tribe. 46 Fed. Reg. 46681. More recently, the Tribe has sought to purchase additional tracts in market transactions. In 2013, the Tribe bought roughly 40 acres where, it says, tribal members who died of smallpox are buried. The Tribe bought the property with an eye to asking the federal government to take the land into trust and add it to the existing reservation next door. See 25 U. S. C. §5108; 25 CFR §151.4 (2013). Toward that end, the Tribe commissioned a survey of the plot so it could confirm the property’s boundaries. But then a question arose. The problem was a barbed wire fence. The fence runs some 1,300 feet along the boundary separating the Tribe’s land from land owned by its neighbors, Sharline and Ray Lundgren. The survey convinced the Tribe that the fence is in the wrong place, leaving about an acre of its land on the Lundgrens’ side. So the Tribe informed its new neighbors that it intended to tear down the fence; clearcut the intervening acre; and build a new fence in the right spot. In response, the Lundgrens filed this quiet title action in Washington state court. Invoking the doctrines of adverse possession and mutual acquiescence, the Lundgrens offered evidence showing that the fence has stood in the same place for years, that they have treated the disputed acre as their own, and that the previous owner of the Tribe’s tract long ago accepted the Lundrens’ claim to the land lying on their side of the fence. For its part, the Tribe asserted sovereign immunity from the suit. It relied upon the many decisions of this Court recognizing the sovereign authority of Native American tribes and their right to “the common-law immunity from suit traditionally enjoyed by sovereign powers.” Michigan v. Bay Mills Indian Community, 572 U. S. ___, ___ (2014) (slip op., at 5) (internal quotation marks omitted). Ultimately, the Supreme Court of Washington rejected the Tribe’s claim of immunity and ruled for the Lundgrens. The court reasoned that sovereign immunity does not apply to cases where a judge “exercis[es] in rem jurisdiction” to quiet title in a parcel of land owned by a Tribe, but only to cases where a judge seeks to exercise in personam jurisdiction over the Tribe itself. 187 Wash. 2d 857, 867, 389 P. 3d 569, 573 (2017). In coming to this conclusion, the court relied in part on our decision in Yakima. Like some courts before it, the Washington Supreme Court read Yakima as distinguishing in rem from in personam lawsuits and “establish[ing] the principle that . . . courts have subject matter jurisdiction over in rem proceedings in certain situations where claims of sovereign immunity are asserted.” 187 Wash. 2d, at 868, 389 P. 3d, at 574. That was error. Yakima did not address the scope of tribal sovereign immunity. Instead, it involved only a much more prosaic question of statutory interpretation concerning the Indian General Allotment Act of 1887. See 24Stat. 388. Some background helps dispel the misunderstanding. The General Allotment Act represented part of Congress’s late Nineteenth Century Indian policy: “to extinguish tribal sovereignty, erase reservation boundaries, and force the assimilation of Indians into the society at large.” Yakima, supra, at 254; In re Heff, 197 U. S. 488, 499 (1905). It authorized the President to allot parcels of reservation land to individual tribal members. The law then directed the United States to hold the allotted parcel in trust for some years, and afterwards issue a fee patent to the allottee. 24Stat. 389. Section 6 of the Act, as amended, provided that once a fee patent issued, “each and every allottee shall have the benefit of and be subject to the laws, both civil and criminal, of the State or Territory in which they may reside” and “all restrictions as to sale, incumbrance, or taxation of said land shall be removed.” 25 U. S. C. §349. In 1934, Congress reversed course. It enacted the Indian Reorganization Act, 48Stat. 984, to restore “the principles of tribal self-determination and self-governance” that prevailed before the General Allotment Act. Yakima, 502 U. S., at 255. “Congress halted further allotments and extended indefinitely the existing periods of trust applicable to” parcels that were not yet fee patented. Ibid.; see 25 U. S. C. §§461–462. But the Legislature made no attempt to withdraw lands already conveyed to private persons through fee patents (and by now sometimes conveyed to non-Indians). As a result, Indian reservations today sometimes contain two kinds of land intermixed in a kind of checkerboard pattern: trust land held by the United States and fee-patented land held by private parties. See Yakima, supra, at 256. Yakima concerned the tax consequences of this checkerboard. Recall that the amended version of §6 of the General Allotment Act rendered allottees and their fee-patented land subject to state regulations and taxes. 25 U. S. C. §349. Despite that, in Moe v. Confederated Salish and Kootenai Tribes of Flathead Reservation, 425 U. S. 463 (1976), this Court held that §6 could no longer be read as allowing States to impose in personam taxes (like those on cigarette sales) on transactions between Indians on fee-patented land within a reservation. Id., at 479–481. Among other things, the Court pointed to the impracticality of using the ownership of a particular parcel within a reservation to determine the law governing transactions taking place upon it. See id., at 478–479. Despite Moe and some years later, this Court in Yakima reached a different conclusion with respect to in rem state taxes. The Court held that allowing States to collect property taxes on fee-patented land within reservations was still allowed by §6. Yakima, supra, at 265. Unlike the in personam taxes condemned in Moe, the Court held that imposing in rem taxes only on the fee-patented squares of the checkerboard was “not impracticable” because property tax assessors make “parcel-by-parcel determinations” about property tax liability all the time. Yakima, supra, at 265. In short, Yakima sought only to interpret a relic of a statute in light of a distinguishable precedent; it resolved nothing about the law of sovereign immunity. Commendably, the Lundgrens acknowledged all this at oral argument. Tr. of Oral Arg. 36. Instead of seeking to defend the Washington Supreme Court’s reliance on Yakima, they now ask us to affirm their judgment on an entirely distinct alternative ground. At common law, they say, sovereigns enjoyed no immunity from actions involving immovable property located in the territory of another sovereign. As our cases have put it, “[a] prince, by acquiring private property in a foreign country, . . . may be considered as so far laying down the prince, and assuming the character of a private individual.” Schooner Exchange v. McFaddon, 7 Cranch 116, 145 (1812). Relying on this line of reasoning, the Lundgrens argue, the Tribe cannot assert sovereign immunity because this suit relates to immovable property located in the State of Washington that the Tribe purchased in the “the character of a private individual.” The Tribe and the federal government disagree. They note that immunity doctrines lifted from other contexts do not always neatly apply to Indian tribes. See Kiowa Tribe of Okla. v. Manufacturing Technologies, Inc., 523 U. S. 751, 756 (1998) (“[T]he immunity possessed by Indian tribes is not coextensive with that of the States”). And since the founding, they say, the political branches rather than judges have held primary responsibility for determining when foreign sovereigns may be sued for their activities in this country. Verlinden B. V. v. Central Bank of Nigeria, 461 U. S. 480, 486 (1983); Ex parte Peru, 318 U. S. 578, 588 (1943). We leave it to the Washington Supreme Court to address these arguments in the first instance. Although we have discretion to affirm on any ground supported by the law and the record that will not expand the relief granted below, Thigpen v. Roberts, 468 U. S. 27, 30 (1984), in this case we think restraint is the best use of discretion. Determining the limits on the sovereign immunity held by Indian tribes is a grave question; the answer will affect all tribes, not just the one before us; and the alternative argument for affirmance did not emerge until late in this case. In fact, it appeared only when the United States filed an amicus brief in this case—after briefing on certiorari, after the Tribe filed its opening brief, and after the Tribe’s other amici had their say. This Court has often declined to take a “first view” of questions that make their appearance in this posture, and we think that course the wise one today. Cutter v. Wilkinson, 544 U. S. 709, 718, n. 7 (2005). The dissent is displeased with our decision on this score, but a contradiction lies at the heart of its critique. First, the dissent assures us that the immovable property exception applies with irresistible force—nothing more than a matter of “hornbook law.” Post, at 3–10 (opinion of Thomas, J.). But then, the dissent claims that allowing the Washington Supreme Court to address that exception is a “grave” decision that “casts uncertainty” over the law and leaves lower courts with insufficient “guidance.” Post, at 3, 13–14. Both cannot be true. If the immovable property exception presents such an easy question, then it’s hard to see what terrible things could happen if we allow the Washington Supreme Court to answer it. Surely our state court colleagues are no less versed than we in “hornbook law,” and we are confident they can and will faithfully apply it. And what if, instead, the question turns out to be more complicated than the dissent promises? In that case the virtues of inviting full adversarial testing will have proved themselves once again. Either way, we remain sanguine about the consequences. The dissent’s other objection to a remand rests on a belief that the immovable property exception was the source of “the disagreement that led us to take this case.” Post, at 1. But this too is mistaken. As we’ve explained, the courts below and the certiorari-stage briefs before us said precisely nothing on the subject. Nor do we understand how the dissent might think otherwise—for its essential premise is that no disagreement exists, or is even possible, about the exception’s scope. The source of confusion in the lower courts that led to our review was the one about Yakima, see supra, at 1, n., and we have dispelled it. That is work enough for the day. We vacate the judgment and remand the case for further proceedings not inconsistent with this opinion. It is so ordered. Notes 1 * Compare 187 Wash. 2d 857, 865–869, 389 P.3d 569, 573–574 (2017) (case below); Cass County Joint Water Resource Dist. v. 1.43 Acres of Land in Highland Twp., 2002 ND 83, 643 N.W.2d 685, 691–693 (2002) (conforming to the Washington Supreme Court’s interpretation of Yakima), with Hamaatsa, Inc. v. Pueblo of San Felipe, 2017–NMSC–007, 388 P.3d 977, 986 (2016) (disagreeing); Cayuga Indian Nation of N. Y. v. Seneca County, 761 F.3d 218, 221 (CA2 2014) (same).

585.US.2017_16-1011

Petitioner WesternGeco LLC owns patents for a system used to survey the ocean floor. Respondent ION Geophysical Corp. began selling a competing system that was built from components manufactured in the United States, shipped to companies abroad, and assembled there into a system indistinguishable from WesternGeco’s. WesternGeco sued for patent infringement under 35 U. S. C. §§271(f)(1) and (f)(2). The jury found ION liable and awarded WesternGeco damages in royalties and lost profits under §284. ION moved to set aside the verdict, arguing that WesternGeco could not recover damages for lost profits because §271(f) does not apply extraterritorially. The District Court denied the motion, but the Federal Circuit reversed. ION was liable for infringement under §271(f)(2), the court reasoned, but §271(f) does not allow patent owners to recover for lost foreign profits On remand from this Court in light of Halo Electronics, Inc. v. Pulse Electronics, Inc., 579 U. S. ____, the Federal Circuit reinstated the portion of its decision regarding §271(f)’s extraterritoriality. Held: WesternGeco’s award for lost profits was a permissible domestic application of §284 of the Patent Act. Pp. 4–10. (a) The presumption against extraterritoriality assumes that federal statutes “apply only within the territorial jurisdiction of the United States.” Foley Bros., Inc. v. Filardo, 336 U. S. 281, 285. The two-step framework for deciding extraterritoriality questions asks, first, “whether the presumption . . . has been rebutted.” RJR Nabisco, Inc. v. European Community, 579 U. S. ___, ___. If not, the second step asks “whether the case involves a domestic application of the statute.” Id., at ___. Courts make the second determination by identifying “the statute’s ‘focus’ ” and then asking whether the conduct relevant to that focus occurred in United States territory. Ibid. If so, the case involves a permissible domestic application of the statute. It is “usually . . . preferable” to begin with step one, but courts have the discretion to begin with step two “in appropriate cases.” Id., at ___, n. 5. The Court exercises that discretion here. Pp. 4–5. (b) When determining “the statute’s ‘focus’ ”—i.e., “the objec[t] of [its] solicitude,” Morrison v. National Australia Bank Ltd., 561 U. S. 247, 267—the provision at issue is not analyzed in a vacuum. If it works in tandem with other provisions, it must be assessed in concert with those provisions. Section 284, the Patent Act’s general damages provision, states that “the court shall award the claimant damages adequate to compensate for the infringement.” The focus of that provision is “the infringement.” The “overriding purpose” of §284 is to “affor[d] patent owners complete compensation” for infringements. General Motors Corp. v. Devex Corp., 461 U. S. 648, 655. Section 271 identifies several ways that a patent can be infringed. Thus, to determine §284’s focus in a given case, the type of infringement that occurred must be identified. Here, §271(f)(2) was the basis for WesternGeco’s infringement claim and the lost-profits damages that it received. That provision regulates the domestic act of “suppl[ying] in or from the United States,” and this Court has acknowledged that it vindicates domestic interests, see, e.g., Microsoft Corp. v. AT&T Corp., 550 U. S. 437, 457. In sum, the focus of §284 in a case involving infringement under §271(f)(2) is on the act of exporting components from the United States. So the conduct in this case that is relevant to the statutory focus clearly occurred in the United States. Pp. 5–8. (c) ION’s contrary arguments are unpersuasive. The award of damages is not the statutory focus here. The damages themselves are merely the means by which the statute achieves its end of remedying infringements, and the overseas events giving rise to the lost-profit damages here were merely incidental to the infringement. In asserting that damages awards for foreign injuries are always an extraterritorial application of a damages provision, ION misreads a portion of RJR Nabisco that interpreted a substantive element of a cause of action, not a remedial damages provision. See 579 U. S., at ___. Pp. 8–9. 837 F. 3d 1358, reversed and remanded. Thomas, J., delivered the opinion of the Court, in which Roberts, C. J., and Kennedy, Ginsburg, Alito, Sotomayor, and Kagan, JJ., joined. Gorsuch, J., filed a dissenting opinion, in which, Breyer, J., joined.

Under the Patent Act, a company can be liable for patent infringement if it ships components of a patented invention overseas to be assembled there. See 35 U. S. C. §271(f )(2). A patent owner who proves infringement under this provision is entitled to recover damages. §284. The question in this case is whether these statutes allow the patent owner to recover for lost foreign profits. We hold that they do. I The Patent Act gives patent owners a “civil action for infringement.” §281. Section 271 outlines several types of infringement. The general infringement provision, §271(a), covers most infringements that occur “within the United States.” The subsection at issue in this case, §271(f ), “expands the definition of infringement to include supplying from the United States a patented invention’s components.” Microsoft Corp. v. AT&T Corp., 550 U. S. 437, 444–445 (2007). It contains two provisions that “work in tandem” by addressing “different scenarios.” Life Technologies Corp. v. Promega Corp., 580 U. S. ___, ___ (2017) (slip op., at 9). Section 271(f )(1) addresses the act of exporting a substantial portion of an invention’s components: “Whoever without authority supplies or causes to be supplied in or from the United States all or a substantial portion of the components of a patented invention, where such components are uncombined in whole or in part, in such manner as to actively induce the combination of such components outside of the United States in a manner that would infringe the patent if such combination occurred within the United States, shall be liable as an infringer.” Section 271(f )(2), the provision at issue here, addresses the act of exporting components that are specially adapted for an invention: “Whoever without authority supplies or causes to be supplied in or from the United States any component of a patented invention that is especially made or especially adapted for use in the invention and not a staple article or commodity of commerce suitable for substantial noninfringing use, where such component is uncombined in whole or in part, knowing that such component is so made or adapted and intending that such component will be combined outside of the United States in a manner that would infringe the patent if such combination occurred within the United States, shall be liable as an infringer.” Patent owners who prove infringement under §271 are entitled to relief under §284, which authorizes “damages adequate to compensate for the infringement, but in no event less than a reasonable royalty for the use made of the invention by the infringer.” II Petitioner WesternGeco LLC owns four patents relating to a system that it developed for surveying the ocean floor. The system uses lateral-steering technology to produce higher quality data than previous survey systems. Western- Geco does not sell its technology or license it to compet- itors. Instead, it uses the technology itself, performing surveys for oil and gas companies. For several years, WesternGeco was the only surveyor that used such lateral-steering technology. In late 2007, respondent ION Geophysical Corporation began selling a competing system. It manufactured the components for its competing system in the United States and then shipped them to companies abroad. Those companies combined the components to create a surveying system indistinguishable from WesternGeco’s and used the system to compete with WesternGeco. WesternGeco sued for patent infringement under §§271(f )(1) and (f )(2). At trial, WesternGeco proved that it had lost 10 specific survey contracts due to ION’s infringement. The jury found ION liable and awarded WesternGeco damages of $12.5 million in royalties and $93.4 million in lost profits. ION filed a post-trial motion to set aside the verdict, arguing that WesternGeco could not recover damages for lost profits because §271(f ) does not apply extraterritorially. The District Court denied the motion. 953 F. Supp. 2d 731, 755–756 (SD Tex. 2013). On appeal, the Court of Appeals for the Federal Circuit reversed the award of lost-profits damages. WesternGeco LLC v. ION Geophysical Corp., 791 F. 3d 1340, 1343 (2015).[1] The Federal Circuit had previously held that §271(a), the general infringement provision, does not allow patent owners to recover for lost foreign sales. See id., at 1350–1351 (citing Power Integrations, Inc. v. Fairchild Semiconductor Int’l, Inc., 711 F. 3d 1348 (CA Fed. 2013)). Section 271(f ) should be interpreted the same way, the Federal Circuit reasoned, because it was “designed” to put patent infringers “in a similar position.” WesternGeco, 791 F. 3d, at 1351. Judge Wallach dissented. See id., at 1354–1364. WesternGeco petitioned for review in this Court. We granted the petition, vacated the Federal Circuit’s judgment, and remanded for further consideration in light of our decision in Halo Electronics, Inc. v. Pulse Electronics, Inc., 579 U. S. ___ (2016). WesternGeco LLC v. ION Geophysical Corp., 579 U. S. ___ (2016). On remand, the panel majority reinstated the portion of its decision regarding the extraterritoriality of §271(f ). 837 F. 3d 1358, 1361, 1364 (CA Fed. 2016). Judge Wallach dissented again, id., at 1364–1369, and we granted certiorari again, 583 U. S. ___ (2018). We now reverse. III Courts presume that federal statutes “apply only within the territorial jurisdiction of the United States.” Foley Bros., Inc. v. Filardo, 336 U. S. 281, 285 (1949). This principle, commonly called the presumption against extraterritoriality, has deep roots. See A. Scalia & B. Garner, Reading Law: The Interpretation of Legal Texts §43, p. 268 (2012) (tracing it to the medieval maxim Statuta suo clauduntur territorio, nec ultra territorium disponunt); e.g., United States v. Palmer, 3 Wheat. 610, 631 (1818) (Marshall, C. J.) (“[G]eneral words must . . . be limited to cases within the jurisdiction of the state”). The presumption rests on “the commonsense notion that Congress generally legislates with domestic concerns in mind.” Smith v. United States, 507 U. S. 197, 204, n. 5 (1993). And it prevents “unintended clashes between our laws and those of other nations which could result in international discord.” EEOC v. Arabian American Oil Co., 499 U. S. 244, 248 (1991). This Court has established a two-step framework for deciding questions of extraterritoriality. The first step asks “whether the presumption against extraterritoriality has been rebutted.” RJR Nabisco, Inc. v. European Community, 579 U. S. ___, ___ (2016) (slip op., at 9). It can be rebutted only if the text provides a “clear indication of an extraterritorial application.” Morrison v. National Australia Bank Ltd., 561 U. S. 247, 255 (2010). If the presumption against extraterritoriality has not been rebutted, the second step of our framework asks “whether the case involves a domestic application of the statute.” RJR Nabisco, 579 U. S., at ___ (slip op., at 9). Courts make this determination by identifying “the statute’s ‘focus’ ” and asking whether the conduct relevant to that focus occurred in United States territory. Ibid. If it did, then the case involves a permissible domestic application of the statute. See ibid. We resolve this case at step two. While “it will usually be preferable” to begin with step one, courts have the discretion to begin at step two “in appropriate cases.” See id., at ___, n. 5 (slip op., at 10, n. 5) (citing Pearson v. Callahan, 555 U. S. 223, 236–243 (2009)). One reason to exercise that discretion is if addressing step one would require resolving “difficult questions” that do not change “the outcome of the case,” but could have far-reaching effects in future cases. See id., at 236–237. That is true here. WesternGeco argues that the presumption against extraterritoriality should never apply to statutes, such as §284, that merely provide a general damages remedy for conduct that Congress has declared unlawful. Resolving that question could implicate many other statutes besides the Patent Act. We therefore exercise our discretion to forgo the first step of our extraterritoriality framework. A Under the second step of our framework, we must identify “the statute’s ‘focus.’ ” RJR Nabisco, supra, at ___ (slip op., at 9). The focus of a statute is “the objec[t] of [its] solicitude,” which can include the conduct it “seeks to ‘regulate,’ ” as well as the parties and interests it “seeks to ‘protec[t]’ ” or vindicate. Morrison, supra, at 267 (quoting Superintendent of Ins. of N. Y. v. Bankers Life & Casualty Co., 404 U. S. 6, 12, 10 (1971)). “If the conduct relevant to the statute’s focus occurred in the United States, then the case involves a permissible domestic application” of the statute, “even if other conduct occurred abroad.” RJR Nabisco, 579 U. S., at ___ (slip op., at 9). But if the relevant conduct occurred in another country, “then the case involves an impermissible extraterritorial application regardless of any other conduct that occurred in U. S. territory.” Ibid. When determining the focus of a statute, we do not analyze the provision at issue in a vacuum. See Morrison, supra, at 267–269. If the statutory provision at issue works in tandem with other provisions, it must be assessed in concert with those other provisions. Otherwise, it would be impossible to accurately determine whether the application of the statute in the case is a “domestic application.” RJR Nabisco, 579 U. S., at ___ (slip op., at 9). And determining how the statute has actually been applied is the whole point of the focus test. See ibid. Applying these principles here, we conclude that the conduct relevant to the statutory focus in this case is domestic. We begin with §284. It provides a general damages remedy for the various types of patent infringement identified in the Patent Act. The portion of §284 at issue here states that “the court shall award the claimant damages adequate to compensate for the infringement.” We conclude that “the infringement” is the focus of this statute. As this Court has explained, the “overriding purpose” of §284 is to “affor[d] patent owners complete compensation” for infringements. General Motors Corp. v. Devex Corp., 461 U. S. 648, 655 (1983). “The question” posed by the statute is “ ‘how much ha[s] the Patent Holder . . . suffered by the infringement.’ ” Aro Mfg. Co. v. Convertible Top Replacement Co., 377 U. S. 476, 507 (1964). Accordingly, the infringement is plainly the focus of §284. But that observation does not fully resolve this case, as the Patent Act identifies several ways that a patent can be infringed. See §271. To determine the focus of §284 in a given case, we must look to the type of infringement that occurred. We thus turn to §271(f )(2), which was the basis for WesternGeco’s infringement claim and the lost-profits damages that it received.[2] Section 271(f )(2) focuses on domestic conduct. It provides that a company “shall be liable as an infringer” if it “supplies” certain components of a patented invention “in or from the United States” with the intent that they “will be combined outside of the United States in a manner that would infringe the patent if such combination occurred within the United States.” The conduct that §271(f )(2) regulates—i.e., its focus—is the domestic act of “suppl[ying] in or from the United States.” As this Court has acknowledged, §271(f ) vindicates domestic interests: It “was a direct response to a gap in our patent law,” Microsoft Corp., 550 U. S., at 457, and “reach[es] components that are manufactured in the United States but assembled overseas,” Life Technologies, 580 U. S., at ___ (slip op., at 11). As the Federal Circuit explained, §271(f )(2) protects against “domestic entities who export components . . . from the United States.” 791 F. 3d, at 1351. In sum, the focus of §284, in a case involving infringement under §271(f )(2), is on the act of exporting components from the United States. In other words, the domestic infringement is “the objec[t] of the statute’s solicitude” in this context. Morrison, 561 U. S., at 267. The conduct in this case that is relevant to that focus clearly occurred in the United States, as it was ION’s domestic act of supplying the components that infringed WesternGeco’s pat- ents. Thus, the lost-profits damages that were awarded to WesternGeco were a domestic application of §284. B ION’s arguments to the contrary are not persuasive. ION contends that the statutory focus here is “self-evidently on the award of damages.” Brief for Respondent 22. While §284 does authorize damages, what a statute authorizes is not necessarily its focus. Rather, the focus is “the objec[t] of the statute’s solicitude”—which can turn on the “conduct,” “parties,” or interests that it regulates or protects. Morrison, supra, at 267. Here, the damages themselves are merely the means by which the statute achieves its end of remedying infringements. Similarly, ION is mistaken to assert that this case involves an extraterritorial application of §284 simply because “lost-profits damages occurred extraterritorially, and foreign conduct subsequent to [ION’s] infringement was necessary to give rise to the injury.” Brief for Respondent 22. Those overseas events were merely incidental to the infringement. In other words, they do not have “primacy” for purposes of the extraterritoriality analysis. Morrison, supra, at 267. ION also draws on the conclusion in RJR Nabisco that “RICO damages claims” based “entirely on injury suffered abroad” involve an extraterritorial application of 18 U. S. C. §1964(c). 579 U. S., at ___ (slip op., at 27). From this principle, ION extrapolates a general rule that damages awards for foreign injuries are always an extraterritorial application of a damages provision. This argument misreads RJR Nabisco. That portion of RJR Nabisco interpreted a substantive element of a cause of action, not a remedial damages provision. See id., at ___ (slip op., at 18). It explained that a plaintiff could not bring a damages claim under §1964(c) unless he could prove that he was “ ‘injured in his business or property,’ ” which required proof of “a domestic injury.” Ibid. Thus, RJR Nabisco was applying the presumption against extraterritoriality to interpret the scope of §1964(c)’s injury requirement; it did not make any statements about damages—a separate legal concept. Two of our colleagues contend that the Patent Act does not permit damages awards for lost foreign profits. Post, at 1 (Gorsuch, J., joined by Breyer, J., dissenting). Their position wrongly conflates legal injury with the damages arising from that injury. See post, at 2–3. And it is not the better reading of “the plain text of the Patent Act.” Post, at 9. Taken together, §271(f )(2) and §284 allow the patent owner to recover for lost foreign profits. Under §284, damages are “adequate” to compensate for infringement when they “plac[e] [the patent owner] in as good a position as he would have been in” if the patent had not been infringed. General Motors Corp., supra, at 655. Specifically, a patent owner is entitled to recover “ ‘the difference between [its] pecuniary condition after the infringement, and what [its] condition would have been if the infringement had not occurred.’ ” Aro Mfg. Co., supra, at 507. This recovery can include lost profits. See Yale Lock Mfg. Co. v. Sargent, 117 U. S. 536, 552–553 (1886). And, as we hold today, it can include lost foreign profits when the patent owner proves infringement under §271(f )(2).[3] * * * We hold that WesternGeco’s damages award for lost profits was a permissible domestic application of §284. The judgment of the Federal Circuit is reversed, and the case is remanded for further proceedings consistent with this opinion. It is so ordered. Notes 1 The Federal Circuit held that ION was liable for infringement under §271(f )(2). WesternGeco, 791 F. 3d, at 1347–1349. It did not address whether ION was liable under §271(f )(1). Id., at 1348. 2 Because the Federal Circuit did not address §271(f )(1), see n. 1, supra, we limit our analysis to §271(f )(2). 3 In reaching this holding, we do not address the extent to which other doctrines, such as proximate cause, could limit or preclude damages in particular cases.

584.US.2017_16-6855

Petitioner Marion Wilson was convicted of murder and sentenced to death. He sought habeas relief in Georgia Superior Court, claiming that his counsel’s ineffectiveness during sentencing violated the Sixth Amendment. The court denied the petition, in relevant part, because it concluded that counsel’s performance was not deficient and had not prejudiced Wilson. The Georgia Supreme Court summarily denied his application for a certificate of probable cause to appeal. Wilson subsequently filed a federal habeas petition, raising the same ineffective-assistance claim. The District Court assumed that his counsel was deficient but deferred to the state habeas court’s conclusion that any deficiencies did not prejudice Wilson. The Eleventh Circuit affirmed. First, however, the panel concluded that the District Court was wrong to “look though” the State Supreme Court’s unexplained decision and assume that it rested on the grounds given in the state habeas court’s opinion, rather than ask what arguments “could have supported” the State Supreme Court’s summary decision. The en banc court agreed with the panel’s methodology. Held: A federal habeas court reviewing an unexplained state-court decision on the merits should “look through” that decision to the last related state-court decision that provides a relevant rationale and presume that the unexplained decision adopted the same reasoning. The State may rebut the presumption by showing that the unexplained decision most likely relied on different grounds than the reasoned decision below. Pp. 5–11. (a) In Ylst v. Nunnemaker, 501 U. S. 797 , the Court held that where there has been one reasoned state judgment rejecting a federal claim, later unexplained orders upholding that judgment or rejecting the same claim are presumed to rest upon the same ground. In Ylst, where the last reasoned opinion on the claim explicitly imposed a procedural default, the Court presumed that a later decision rejecting the claim did not silently disregard that bar and consider the merits. Since Ylst, every Circuit to have considered the matter, but for the Eleventh Circuit, has applied a “look through” presumption even where the state courts did not apply a procedural bar to review, and most Circuits applied the presumption prior to Ylst. The presumption is often realistic, for state higher courts often issue summary decisions when they have examined the lower court’s reasoning and found nothing significant with which they disagree. The presumption also is often more efficiently applied than a contrary approach that would require a federal court to imagine what might have been the state court’s supportive reasoning. The State argues that Harrington v. Richter, 562 U. S. 86 , controls here and that Ylst should apply, at most, where the federal habeas court is trying to determine whether a state-court decision without opinion rested on a state procedural ground or whether the state court reached the merits of a federal issue. Richter, however, did not directly concern the issue in this case—whether to “look through” the silent state higher court opinion to the lower court’s reasoned opinion in order to determine the reasons for the higher court’s decision. In Richter, there was no lower court opinion to look to. And Richter does not say that Ylst’s reasoning does not apply in the context of an unexplained decision on the merits. Indeed, this Court has “looked though” to lower court decisions in cases involving the merits. See, e.g., Premo v. Moore, 562 U. S. 115, 123 –133. Pp. 5–9. (b) The State’s further arguments are unconvincing. It points out that the “look though” presumption may not accurately identify the grounds for a higher court’s decision. But the “look through” presumption is not an absolute rule. Additional evidence that might not be sufficient to rebut the presumption in a case like Ylst, where the lower court rested on a state-law procedural ground, would allow a federal court to conclude that counsel has rebutted the presumption in a case decided on the merits. For instance, a federal court may conclude that the presumption is rebutted where counsel identifies convincing alternative arguments for affirmance that were made to the State’s highest court, or equivalent evidence such as an alternative ground that is obvious in the state-court record. The State also argues that this Court does not necessarily presume that a federal court of appeals’ silent opinion adopts the reasoning of the court below, but that is a different context. Were there to be a “look through” approach as a general matter in that context, judges and lawyers might read those decisions as creating, through silence, binding circuit precedent. Here, a federal court “looks through” the silent decision for a specific and narrow purpose, to identify the grounds for the higher court’s decision as the Antiterrorism and Effective Death Penalty Act requires. Nor does the “look through” approach show disrespect for the States; rather, it seeks to replicate the grounds for the higher state court’s decision. Finally, the “look though” approach is unlikely to lead state courts to write full opinions where they would have preferred to decide summarily, at least not to any significant degree. Pp. 9–11. 834 F. 3d 1227, reversed and remanded. Breyer, J., delivered the opinion of the Court, in which Roberts, C. J., and Kennedy, Ginsburg, Sotomayor, and Kagan, JJ., joined. Gorsuch, J., filed a dissenting opinion, in which Thomas and Alito, JJ., joined.

The Antiterrorism and Effective Death Penalty Act of 1996 (AEDPA) requires a prisoner who challenges (in a federal habeas court) a matter “adjudicated on the merits in State court” to show that the relevant state-court “decision” (1) “was contrary to, or involved an unreasonable application of, clearly established Federal law,” or (2) “was based on an unreasonable determination of the facts in light of the evidence presented in the State court proceeding.” 28 U. S. C. §2254(d). Deciding whether a state court’s decision “involved” an unreasonable application of federal law or “was based on” an unreasonable determination of fact requires the federal habeas court to “train its attention on the particular reasons—both legal and fac- tual—why state courts rejected a state prisoner’s federal claims,” Hittson v. Chatman, 576 U. S. ___, ___ (2015) (Ginsburg, J., concurring in denial of certiorari) (slip op., at 1), and to give appropriate deference to that decision, Harrington v. Richter, 562 U. S. 86, 101 –102 (2011). This is a straightforward inquiry when the last state court to decide a prisoner’s federal claim explains its decision on the merits in a reasoned opinion. In that case, a federal habeas court simply reviews the specific reasons given by the state court and defers to those reasons if they are reasonable. We have affirmed this approach time and again. See, e.g., Porter v. McCollum, 558 U. S. 30, 39 –44 (2009) (per curiam); Rompilla v. Beard, 545 U. S. 374, 388 –392 (2005); Wiggins v. Smith, 539 U. S. 510, 523 –538 (2003). The issue before us, however, is more difficult. It concerns how a federal habeas court is to find the state court’s reasons when the relevant state-court decision on the merits, say, a state supreme court decision, does not come accompanied with those reasons. For instance, the decision may consist of a one-word order, such as “affirmed” or “denied.” What then is the federal habeas court to do? We hold that the federal court should “look through” the unexplained decision to the last related state-court decision that does provide a relevant rationale. It should then presume that the unexplained decision adopted the same reasoning. But the State may rebut the presumption by showing that the unexplained affirmance relied or most likely did rely on different grounds than the lower state court’s decision, such as alternative grounds for affirmance that were briefed or argued to the state supreme court or obvious in the record it reviewed. I In 1997 a Georgia jury convicted petitioner, Marion Wilson, of murder and related crimes. After a sentencing hearing, the jury sentenced Wilson to death. In 1999 the Georgia Supreme Court affirmed Wilson’s conviction and sentence, Wilson v. State, 271 Ga. 811, 525 S.E.2d 339 (1999), and this Court denied his petition for certiorari, Wilson v. Georgia, 531 U. S. 838 (2000) . Wilson then filed a petition for habeas corpus in a state court, the Superior Court for Butts County. Among other things, he claimed that his counsel was “ineffective” during his sentencing, in violation of the Sixth Amendment. See Strickland v. Washington, 466 U. S. 668, 687 (1984) (setting forth “two components” of an ineffective-assistance-of-counsel claim: “that counsel’s performance was deficient” and “that the deficient performance prejudiced the defense”). Wilson identified new evidence that he argued trial counsel should have introduced at sentencing, namely, testimony from various witnesses about Wilson’s childhood and the impairment of the frontal lobe of Wilson’s brain. After a hearing, the state habeas court denied the petition in relevant part because it thought Wilson’s evidence did not show that counsel was “deficient,” and, in any event, counsel’s failure to find and present the new evidence that Wilson offered had not prejudiced Wilson. Wilson v. Terry, No. 2001–v–38 (Super. Ct. Butts Cty., Ga., Dec. 1, 2008), App. 60–61. In the court’s view, that was because the new evidence was “inadmissible on evidentiary grounds,” was “cumulative of other testimony,” or “otherwise would not have, in reasonable probability, changed the outcome of the trial.” Id., at 61. Wilson applied to the Georgia Supreme Court for a certificate of probable cause to appeal the state habeas court’s decision. But the Georgia Supreme Court denied the application without any explanatory opinion. Wilson v. Terry, No. 2001–v–38 (May 3, 2010), App. 87, cert. denied, 562 U. S. 1093 (2010) . Wilson subsequently filed a petition for habeas corpus in the United States District Court for the Middle District of Georgia. He made what was essentially the same “ineffective assistance” claim. After a hearing, the District Court denied Wilson’s petition. Wilson v. Humphrey, No. 5:10–cv–489 (Dec. 19, 2013), App. 88–89. The court assumed that Wilson’s counsel had indeed been “deficient” in failing adequately to investigate Wilson’s background and physical condition for mitigation evidence and to present what he likely would have found at the sentencing hearing. Id., at 144. But, the court nonetheless deferred to the state habeas court’s conclusion that these deficiencies did not “prejudice” Wilson, primarily because the testimony of many witnesses was “cumulative,” and because the evidence of physical impairments did not include any physical examination or other support that would have shown the state-court determination was “unreasonable.” Id., at 187; see Richter, 562 U. S., at 111–112. Wilson appealed to the Court of Appeals for the Eleventh Circuit. Wilson v. Warden, 774 F. 3d 671 (2014). The panel first held that the District Court had used the wrong method for determining the reasoning of the relevant state court, namely, that of the Georgia Supreme Court (the final and highest state court to decide the merits of Wilson’s claims). Id., at 678. That state-court decision, the panel conceded, was made without an opinion. But, the federal court was wrong to “look through” that decision and assume that it rested on the grounds given in the lower court’s decision. Instead of “looking through” the decision to the state habeas court’s opinion, the federal court should have asked what arguments “could have supported” the Georgia Supreme Court’s refusal to grant permission to appeal. The panel proceeded to identify a number of bases that it believed reasonably could have supported the decision. Id., at 678–681. The Eleventh Circuit then granted Wilson rehearing en banc so that it could consider the matter of methodology. Wilson v. Warden, 834 F. 3d 1227 (2016). Ultimately six judges (a majority) agreed with the panel and held that its “could have supported” approach was correct. Id., at 1235. Five dissenting judges believed that the District Court should have used the methodology it did use, namely, the “look through” approach. Id., at 1242–1247, 1247–1269. Wilson then sought certiorari here. Because the Eleventh Circuit’s opinion creates a split among the Circuits, we granted the petition. Compare id., at 1285 (applying “could have supported” approach), with Grueninger v. Director, Va. Dept. of Corrections, 813 F. 3d 517, 525–526 (CA4 2016) (applying “look through” presumption post-Richter), and Cannedy v. Adams, 706 F. 3d 1148, 1156–1159 (CA9 2013) (same); see also Clements v. Clarke, 592 F. 3d 45, 52 (CA1 2010) (applying “look through” presumption pre-Richter); Bond v. Beard, 539 F. 3d 256, 289–290 (CA3 2008) (same); Mark v. Ault, 498 F. 3d 775, 782–783 (CA8 2007) (same); Joseph v. Coyle, 469 F. 3d 441, 450 (CA6 2006) (same). II We conclude that federal habeas law employs a “look through” presumption. That conclusion has parallels in this Court’s precedent. In Ylst v. Nunnemaker, a defendant, convicted in a California state court of murder, appealed his conviction to the state appeals court where he raised a constitutional claim based on Miranda v. Arizona, 384 U. S. 436 (1966) . 501 U. S. 797, 799 –800 (1991). The appeals court rejected that claim, writing that “ ‘an objection based upon a Miranda violation cannot be raised for the first time on appeal.’ ” Id., at 799. The defendant then similarly challenged his conviction in the California Supreme Court and on collateral review in several state courts (including once again the California Supreme Court). In each of these latter instances the state court denied the defendant relief (or review). In each instance the court did so without an opinion or other explanation. Id., at 799–800. Subsequently, the defendant asked a federal habeas court to review his constitutional claim. Id., at 800. The higher state courts had given no reason for their decision. And this Court ultimately had to decide how the federal court was to find the state court’s reasoning in those circumstances. Should it have “looked through” the unreasoned decisions to the state procedural ground articulated in the appeals court or should it have used a different method? In answering that question Justice Scalia wrote the following for the Court: “The problem we face arises, of course, because many formulary orders are not meant to convey anything as to the reason for the decision. Attributing a reason is therefore both difficult and artificial. We think that the attribution necessary for federal habeas purposes can be facilitated, and sound results more often assured, by applying the following presumption: Where there has been one reasoned state judgment rejecting a federal claim, later unexplained orders upholding that judgment or rejecting the same claim rest upon the same ground. If an earlier opinion ‘fairly appear[s] to rest primarily upon federal law,’ we will presume that no procedural default has been invoked by a subsequent unexplained order that leaves the judgment or its consequences in place. Similarly where, as here, the last reasoned opinion on the claim explicitly imposes a procedural default, we will presume that a later decision rejecting the claim did not silently disregard that bar and consider the merits.” Id., at 803 (citation omitted). Since Ylst, every Circuit to have considered the matter has applied this presumption, often called the “look through” presumption, but for the Eleventh Circuit—even where the state courts did not apply a procedural bar to review. See supra, at 4–5. And most Federal Circuits applied it prior to Ylst. See Ylst, supra, at 803 (citing Prihoda v. McCaughtry, 910 F. 2d 1379, 1383 (CA7 1990); Harmon v. Barton, 894 F. 2d 1268, 1272 (CA11 1990); Evans v. Thompson, 881 F. 2d 117, 123, n. 2 (CA4 1989); Ellis v. Lynaugh, 873 F. 2d 830, 838 (CA5 1989)). That is not surprising in light of the fact that the “look through” presumption is often realistic, for state higher courts often (but certainly not always, see Redmon v. Johnson, 2018 WL 415714 (Ga., Jan. 16, 2018)) write “denied” or “affirmed” or “dismissed” when they have examined the lower court’s reasoning and found nothing significant with which they disagree. Moreover, a “look through” presumption is often (but not always) more efficiently applied than a contrary approach—an approach, for example, that would require a federal habeas court to imagine what might have been the state court’s supportive reasoning. The latter task may prove particularly difficult where the issue involves state law, such as state procedural rules that may constrain the scope of a reviewing court’s summary decision, a matter in which a federal judge often lacks comparative expertise. See Ylst, supra, at 805. The State points to a later case, Harrington v. Richter, 562 U. S. 86 (2011) , which, it says, controls here instead of Ylst. In its view, Ylst should apply, at most, to cases in which the federal habeas court is trying to determine whether a state-court decision without opinion rested on a state procedural ground (for example, a procedural default) or whether the state court has reached the merits of a federal issue. In support, it notes that Richter held that the state-court decisions to which AEDPA refers include summary dispositions, i.e., decisions without opinion. Richter added that “determining whether a state court’s decision resulted from an unreasonable legal or factual conclusion does not require that there be an opinion from the state court explaining the state court’s reasoning.” 562 U. S., at 98. Richter then said that, where “a state court’s decision is unaccompanied by an explanation, the habeas petitioner’s burden still must be met by showing there was no reasonable basis for the state court to deny relief.” Ibid. And the Court concluded that, when “a federal claim has been presented to a state court and the state court has denied relief, it may be presumed that the state court adjudicated the claim on the merits in the absence of any indication or state-law procedural principles to the contrary.” Id., at 99. In our view, however, Richter does not control here. For one thing, Richter did not directly concern the issue before us—whether to “look through” the silent state higher court opinion to the reasoned opinion of a lower court in order to determine the reasons for the higher court’s decision. Indeed, it could not have considered that matter, for in Richter, there was no lower court opinion to look to. That is because the convicted defendant sought to raise his federal constitutional claim for the first time in the California Supreme Court (via a direct petition for habeas corpus, as California law permits). Id., at 96. For another thing, Richter does not say the reasoning of Ylst does not apply in the context of an unexplained decision on the merits. To the contrary, the Court noted that it was setting forth a presumption, which “may be overcome when there is reason to think some other explanation for the state court’s decision is more likely.” Richter, supra, at 99–100. And it referred in support to Ylst, 501 U. S., at 803. Further, we have “looked through” to lower court decisions in cases involving the merits. See, e.g., Premo v. Moore, 562 U. S. 115, 123 –133 (2011); Sears v. Upton, 561 U. S. 945, 951 –956 (2010) (per curiam). Indeed, we de- cided one of those cases, Premo, on the same day we decided Richter. And in our opinion in Richter we referred to Premo. 562 U. S., at 91. Had we intended Richter’s “could have supported” framework to apply even where there is a reasoned decision by a lower state court, our opinion in Premo would have looked very different. We did not even cite the reviewing state court’s summary affirmance. Instead, we focused exclusively on the actual reasons given by the lower state court, and we deferred to those reasons under AEDPA. 562 U. S., at 132 (“The state postconviction court’s decision involved no unreasonable application of Supreme Court precedent”). III The State’s further arguments do not convince us. The State points out that there could be many cases in which a “look through” presumption does not accurately identify the grounds for the higher court’s decision. And we agree. We also agree that it is more likely that a state supreme court’s single word “affirm” rests upon alternative grounds where the lower state court decision is unreasonable than, e.g., where the lower court rested on a state-law proce- dural ground, as in Ylst. But that is why we have set forth a presumption and not an absolute rule. And the unreasonableness of the lower court’s decision itself provides some evidence that makes it less likely the state supreme court adopted the same reasoning. Thus, additional evidence that might not be sufficient to rebut the presumption in a case like Ylst would allow a federal court to conclude that counsel has rebutted the presumption in a case like this one. For instance, a federal habeas court may conclude that counsel has rebutted the presumption on the basis of convincing alternative arguments for affirmance made to the State’s highest court or equivalent evidence presented in its briefing to the federal court similarly establishing that the State’s highest court relied on a different ground than the lower state court, such as the existence of a valid ground for affirmance that is obvious from the state-court record. The dissent argues that the Georgia Supreme Court’s recent decision in Redmon v. Johnson rebuts the presumption in Georgia because that court indicated its summary decisions should not be read to adopt the lower court’s reasoning. Post, at 6–8, 10–11 (opinion of Gorsuch, J.). This misses the point. A presumption that can be rebutted by evidence of, for instance, an alternative ground that was argued or that is clear in the record was the likely basis for the decision is in accord with full and proper respect for state courts, like those in Georgia, which have well-established systems and procedures in place in order to ensure proper consideration to the arguments and contention in the many cases they must process to determine whether relief should be granted when a criminal conviction or its ensuing sentence is challenged. The State also points out that we do not necessarily presume that a silent opinion of a federal court of appeals adopts the reasoning of the court below. The dissent similarly invokes these “traditional rules of appellate practice.” See post, at 5–6, 10. But neither the State nor the dissent provides examples of similar context. Were we to adopt a “look through” approach in respect to silent federal appeals court decisions as a general matter in other contexts, we would risk judges and lawyers reading those decisions as creating, through silence, a precedent that could be read as binding throughout the circuit—just what a silent decision may be thought not to do. Here, however, we “look through” the silent decision for a spe- cific and narrow purpose—to identify the grounds for the higher court’s decision, as AEDPA directs us to do. See supra, at 1–2. We see no reason why the federal court’s interpretation of the state court’s silence should be taken as binding precedent outside this context, for example, as a statewide binding interpretation of state law. Further, the State argues that the “look through” approach shows disrespect for the States. See Brief for Respondent 39 (“Wilson’s approach to summary decisions reflects an utter lack of faith in the ability of the highest state courts to adjudicate constitutional rights”). We do not believe this is so. Rather the presumption seeks to replicate the grounds for the higher state court’s decision. Where there are convincing grounds to believe the silent court had a different basis for its decision than the analysis followed by the previous court, the federal habeas court is free, as we have said, to find to the contrary. In our view, this approach is more likely to respect what the state court actually did, and easier to apply in practice, than to ask the federal court to substitute for silence the federal court’s thought as to more supportive reasoning. Finally, the State argues that the “look through” approach will lead state courts to believe they must write full opinions where, given the workload, they would have preferred to have decided summarily. Though the matter is empirical, given the narrowness of the context, we do not believe that they will feel compelled to do so—at least not to any significant degree. The State offers no such evidence in the many Circuits that have applied Ylst outside the procedural context. See supra, at 5. For these reasons, we reverse the Eleventh Circuit’s judgment and remand the case for further proceedings consistent with this opinion. It is so ordered.

585.US.2017_17-530

As the Great Depression took its toll, struggling railroad pension funds reached the brink of insolvency. During that time before the rise of the modern interstate highway system, privately owned railroads employed large numbers of Americans and provided services vital to the nation’s commerce. To address the emergency, Congress adopted the Railroad Retirement Tax Act of 1937. That legislation federalized private railroad pension plans and it remains in force even today. Under the law’s terms, private railroads and their employees pay a tax based on employees’ incomes. In return, the federal government provides employees a pension often more generous than the social security system supplies employees in other industries. This case arises from a peculiar feature of the statute and its history. At the time of the Act’s adoption, railroads compensated employees not just with money but also with food, lodging, railroad tickets, and the like. Because railroads typically didn’t count these in-kind benefits when calculating an employee’s pension on retirement, neither did Congress in its new statutory pension scheme. Nor did Congress seek to tax these in-kind benefits. Instead, it limited its levies to employee “compensation,” and defined that term to capture only “any form of money remuneration.” It’s this limitation that poses today’s question. To encourage employee performance and to align employee and corporate goals, some railroads have (like employers in many fields) adopted employee stock option plans. The government argues that these stock options qualify as a form of “compensation” subject to taxation under the Act. In its view, stock options can easily be converted into money and so qualify as “money remuneration.” The railroads and their employees reply that stock options aren’t “money remuneration” and remind the Court that when Congress passed the Act it sought to mimic existing industry pension practices that generally took no notice of in-kind benefits. Who has the better of it? Held: Employee stock options are not taxable “compensation” under the Railroad Retirement Tax Act because they are not “money remuneration.” When Congress adopted the Act in 1937, “money” was understood as currency “issued by [a] recognized authority as a medium of exchange.” Pretty obviously, stock options do not fall within that definition. While stock can be bought or sold for money, it isn’t usually considered a medium of exchange. Few people value goods and services in terms of stock, or buy groceries and pay rent with stock. Adding the word “remuneration” also does not alter the meaning of the phrase. When the statute speaks of taxing “any form of money remuneration,” it indicates Congress wanted to tax monetary compensation in any of the many forms an employer might choose. It does not prove that Congress wanted to tax things, like stock, that are not money at all. The broader statutory context points to this conclusion. For example, the 1939 Internal Revenue Code, adopted just two years later, also treated “money” and “stock” as different things. See, e.g., §27(d). And a companion statute enacted by the same Congress, the Federal Insurance Contributions Act, taxes “all remuneration,” including benefits “paid in any medium other than cash.” §3121(a). The Congress that enacted both of these pension schemes knew well the difference between “money” and “all” forms of remuneration and its choice to use the narrower term in the context of railroad pensions alone requires respect, not disregard. Even the IRS (then the Bureau of Internal Revenue) seems to have understood all this back in 1938. Shortly after the Railroad Retirement Tax Act’s enactment, the IRS issued a regulation explaining that the Act taxes “all remuneration in money, or in something which may be used in lieu of money (scrip and merchandise orders, for example).” The regulation said the Act covered things like “[s]alaries, wages, commissions, fees, [and] bonuses.” But the regulation nowhere suggested that stock was taxable. In light of these textual and structural clues and others, the Court thinks it’s clear enough that the term “money” unambiguously excludes “stock.” Pp. 2–8. 856 F. 3d 490, reversed and remanded. Gorsuch, J., delivered the opinion of the Court, in which Roberts, C. J., and Kennedy, Thomas, and Alito, JJ., joined. Breyer, J., filed a dissenting opinion, in which Ginsburg, Sotomayor, and Kagan, JJ., joined.

As the Great Depression took its toll, struggling railroad pension funds reached the brink of insolvency. During that time before the modern interstate highway system, privately owned railroads employed large numbers of Americans and provided services vital to the nation’s commerce. To address the emergency, Congress adopted the Railroad Retirement Tax Act of 1937. That legislation federalized private railroad pension plans and it remains in force today. Under the law’s terms, private railroads and their employees pay a tax based on employees’ incomes. 26 U. S. C. §§3201(a)–(b), 3221(a)–(b). In return, the federal government provides employees a pension often more generous than the social security system supplies employees in other industries. See Hisquierdo v. Hisquierdo, 439 U. S. 572, 573–575 (1979). Our case arises from a peculiar feature of the statute and its history. At the time of the Act’s adoption, railroads compensated employees not just with money but also with food, lodging, railroad tickets, and the like. Because railroads typically didn’t count these in-kind benefits when calculating an employee’s pension on retirement, neither did Congress in its new statutory pension scheme. Nor did Congress seek to tax these in-kind benefits. Instead, it limited itself to taxing employee “compensation,” and defined that term to capture only “any form of money remuneration.” §3231(e)(1). It’s this limitation that poses today’s question. To encourage employee performance and align employee and corporate goals, some railroads (like employers in many fields) have adopted employee stock option plans. Typical of many, the plan before us permits an employee to exercise stock options in various ways—purchasing stock with her own money and holding it as an investment; purchasing stock but immediately selling a portion to finance the purchase; or purchasing stock at the option price, selling it all immediately at the market price, and taking the profits. App. 41–42. The government argues that stock options like these qualify as a form of taxable “money remuneration” under the Act because stock can be easily converted into money. The railroads reply that stock options aren’t “money” at all and remind us that when Congress passed the Act it sought to mimic existing in- dustry pension practices that generally took no notice of in-kind benefits. Who has the better of it? Courts have divided on the answer, so we agreed to take up the question. We start with the key statutory term: “money remuneration.” As usual, our job is to interpret the words consistent with their “ordinary meaning . . . at the time Congress enacted the statute.” Perrin v. United States, 444 U. S. 37, 42 (1979). And when Congress adopted the Act in 1937, “money” was ordinarily understood to mean currency “issued by [a] recognized authority as a medium of exchange.” Webster’s New International Dictionary 1583 (2d ed. 1942); see also 6 Oxford English Dictionary 603 (1st ed. 1933) (“In mod[ern] use commonly applied indifferently to coin and to such promissory documents representing coin (esp. government and bank notes) as are currently accepted as a medium of exchange”); Black’s Law Dictionary 1200 (3d ed. 1933) (in its “popular sense, ‘money’ means any currency, tokens, bank-notes, or other circulating medium in general use as the representative of value”); Railway Express Agency, Inc. v. Virginia, 347 U. S. 359, 365 (1954) (“[M]oney . . . is a medium of exchange”). Pretty obviously, stock options do not fall within that definition. While stock can be bought or sold for money, few of us buy groceries or pay rent or value goods and services in terms of stock. When was the last time you heard a friend say his new car cost “2,450 shares of Microsoft”? Good luck, too, trying to convince the IRS to treat your stock options as a medium of exchange at tax time. See Rev. Rul. 76–350, 1976–2 Cum. Bull. 396; see also, e.g., In re Boyle’s Estate, 2 Cal. App. 2d 234, 236 (1934) (“[T]he word ‘money’ when taken in its ordinary and grammatical sense does not include corporate stocks”); Helvering v. Credit Alliance Corp., 316 U. S. 107, 112 (1942) (distinguishing between “money and . . . stock”). Nor does adding the word “remuneration” alter the calculus. Of course, “remuneration” can encompass any kind of reward or compensation, not just money. 8 Oxford English Dictionary 439. But in the sentence before us, the adjective “money” modifies the noun “remuneration.” So “money” limits the kinds of remuneration that will qualify for taxation; “remuneration” doesn’t expand what counts as money. When the statute speaks of taxing “any form of money remuneration,” then, it indicates Congress wanted to tax monetary compensation in any of the many forms an employer might choose—coins, paper currency, checks, wire transfers, and the like. It does not prove Congress wanted to tax things, like stock, that aren’t money at all. The broader statutory context points to the same conclusion the immediate text suggests. The 1939 Internal Revenue Code, part of the same title as our statute and adopted just two years later, expressly treated “money” and “stock” as different things. Consider a few examples. The Code described “stock of the corporation” as “property other than money.” §27(d). It explained that a corporate distribution is taxable when distributed “either (A) in [the company’s] stock . . . or (B) in money.” §115(f)(2). And it discussed transfers of “money in addition to . . . stock or securities.” §372(b). While ultimately ruling for the government, even the Court of Appeals in this case conceded that the 1939 Code “treat[ed] ‘money’ and ‘stock’ as different concepts.” 856 F. 3d 490, 492 (CA7 2017). That’s not all. The same Congress that enacted the Railroad Retirement Tax Act enacted a companion statute, the Federal Insurance Contributions Act (FICA), to fund social security pensions for employees in other industries. And while the Railroad Retirement Tax Act taxes only “money remuneration,” FICA taxes “all remuneration”— including benefits “paid in any medium other than cash.” §3121(a) (emphasis added). We usually “presume differences in language like this convey differences in meaning.” Henson v. Santander Consumer USA Inc., 582 U. S. ___, ___ (2017) (slip op., at 6). And that presumption must bear particular strength when the same Congress passed both statutes to handle much the same task. See INS v. Cardoza-Fonseca, 480 U. S. 421, 432 (1987). The Congress that enacted both of these pension schemes knew well the difference between “money” and “all” forms of remuneration. Its choice to use the narrower term in the context of railroad pensions alone requires respect, not disregard. Even the IRS (then the Bureau of Internal Revenue) seems to have understood all this back in 1938. Shortly after the Railroad Retirement Tax Act’s enactment, the IRS issued a regulation explaining that the Act taxes “all remuneration in money, or in something which may be used in lieu of money.” 26 CFR §410.5 (1938). By way of example, the regulation said the Act taxed things like “[s]alaries, wages, commissions, fees, [and] bonuses.” §410.6(a). But it nowhere suggested that stock was tax- able. Nor was the possibility lost on the IRS. The IRS said the Act did tax money payments related to stock—“[p]ayments made by an employer into a stock bonus . . . fund.” §410.6(f). But the agency did not seek to extend the same treatment to stock itself. So even assuming the validity of the regulation, it seems only to confirm our understanding. To be sure, the regulation also lists “scrip and merchandise orders” as examples of qualifying mediums of exchange. §410.5. For argument’s sake, too, we will accept that the word “scrip” can sometimes embrace stock. But even if “scrip” is capable of bearing this meaning, at the time the IRS promulgated the regulation in 1938 that was not its ordinary meaning. As even the government acknowledged before the Court of Appeals, “scrip” ordinarily meant “company-issued certificates” that employees could use in lieu of cash “to purchase merchandise at a company store.” Brief for United States in Nos. 16–3300 etc. (CA7 2017), p. 37. This understanding fits perfectly as well with the whole phrase in which the term appears; both “scrip and merchandise orders” were frequently used at the time to purchase goods at company stores. See, e.g., Webster’s New International Dictionary 2249 (defining “scrip” as a “certificate . . . issued to circulate in lieu of government currency” or “by a corporation that pays wages partly in orders on a company store”); Keokee Consol. Coke Co. v. Taylor, 234 U. S. 224, 226 (1914) (company gave its employees “scrip . . . as an advance of monthly wages in payment for labor performed” that could be used to purchase merchandise at the company store); Gatch, Local Money in the United States During the Great Depression, 26 Essays in Economic & Bus. History 47–48 (2008). What does the government have to say about all this? It concedes that money remuneration often means remuneration in a commonly used medium of exchange. But, it submits, the term can carry a much more expansive meaning too. At least sometimes, the government says, “money” means any “property or possessions of any kind viewed as convertible into money or having value expressible in terms of money.” 6 Oxford English Dictionary 603. The dissent takes the same view. See post, at 3 (opinion of Breyer, J.). But while the term “money” sometimes might be used in this much more expansive sense, that isn’t how the term was ordinarily used at the time of the Act’s adoption (or is even today). Baseball cards, vinyl records, snow globes, and fidget spinners all have “value expressible in terms of money.” Even that “priceless” Picasso has a price. Really, almost anything can be reduced to a “value expressible in terms of money.” But in ordinary usage does “money” mean almost everything? The government and the dissent supply no persuasive proof that Congress sought to invoke their idiosyncratic definition. If Congress really thought everything is money, why did it take such pains to differentiate between money and stock in the Internal Revenue Code of 1939? Why did it so carefully distinguish “money remuneration” in the Act and “all remuneration” in FICA? Why did it include the word “money” to qualify “remuneration” if all remuneration counts as money? And wouldn’t the everything-is-money interpretation encompass railroad tickets, food, and lodging—exactly the sort of in-kind benefits we know the Act was written to exclude? These questions they cannot answer. To be sure, the government and dissent do seek to offer a different structural argument of their own. They point to certain of the Act’s tax exemptions, most notably the exemption for qualified stock options. See 26 U. S. C. §3231(e)(12); post, at 6 (Breyer, J., dissenting). Because the Act excludes qualified stock options from taxation, the argument goes, to avoid superfluity it must include other sorts of stock options like the nonqualified stock options the railroads issued here. The problem, though, is that the exemption covers “any remuneration on account of” qualified stock options. §3231(e)(12) (emphasis added). And, as the government concedes, companies sometimes include money payments when qualified stock options are exercised (often to compensate for fractional shares due an employee). Brief for United States 30. As a result, the exemption does work under anyone’s reading. The government replies that Congress would not have bothered to write an exemption that does only this modest work. To have been worth the candle, Congress must have assumed that stock options would qualify as money remuneration without a specific exemption. But we will not join this guessing game. It is not our function “to rewrite a constitutionally valid statutory text under the banner of speculation about what Congress might have” intended. Henson, 582 U. S., at ___ (slip op., at 9). Besides, even if the railroads’ interpretation of the statute threatens to leave one of many exemptions with little to do, that’s hardly a reason to abandon it, for the government’s and dissent’s alternative promises a graver surplusage problem of its own. As it did in 1939, the Internal Revenue Code today repeatedly distinguishes between “stock” and “money.” See, e.g., §306(c)(2) (referring to a situation where “money had been distributed in lieu of . . . stock”). All these distinctions the government and dissent would simply obliterate. Reaching further afield, the government and dissent point to a 1938 agency interpretation of another companion statute, the Railroad Retirement Act of 1937. See post, at 8–9 (Breyer, J., dissenting). Here, the Railroad Retirement Board suggested that the term “money remuneration” in the Railroad Retirement Act could sometimes include in-kind benefits. Again we may assume the valid- ity of the regulation because, even taken on its own terms, it only ends up confirming our interpretation. The Board indicated that in-kind benefits could count as money remuneration only if the employer and employee agreed to this treatment and to the dollar value of the benefit. 20 CFR §222.2 (1938). That same year, the Board made clear that stock was treated just like any other in-kind benefit under this rule: “stock cannot be considered as a ‘form of money remuneration earned by an individual for services rendered’ ” unless part of an employee’s “agreed compensation” and awarded “at a definite agreed value.” Railroad Retirement Bd. Gen. Counsel Memorandum No. L–38–440, pp. 1–2 (1938). Later, the Board provided fuller explanation for its longstanding view, stating that these conditions are necessary because, unlike FICA, the Act does not cover “ ‘remuneration . . . paid in any medium.’ ” Railroad Retirement Bd. Gen. Counsel Memorandum No. L–1986–82, p. 6 (1986). For decades, then, the Board has taken the view that nonmonetary remuneration is “not . . . included in compensation under the [Act] unless the employer and employee first agree to [its] dollar value . . . and then agree that this dollar value shall be part of the employee’s compensation package.” Ibid. None of these preconditions would be needed, of course, if the Act automatically taxed in-kind benefits as the government and dissent insist. Finally, the government seeks Chevron deference for a more recent IRS interpretation treating “compensation” under the Act as having “the same meaning as the term wages in” FICA “except as specifically limited by the Railroad Retirement Tax Act.” 26 CFR §31.3231(e)–1 (2017). But in light of all the textual and structural clues before us, we think it’s clear enough that the term “money” excludes “stock,” leaving no ambiguity for the agency to fill. See Chevron U. S. A. Inc. v. Natural Resources Defense Council, Inc., 467 U. S. 837, 843, n. 9 (1984). Nor does the regulation help the government even on its own terms. FICA’s definition of wages—“all remuneration”—is “specifically limited by the Railroad Retirement Tax Act,” which applies only to “money remuneration.” So in the end all the regulation winds up saying is that everyone should look carefully at the relevant statutory texts. We agree, and that is what we have done. The Court of Appeals in this case tried a different tack still, if over a dissent. The majority all but admitted that stock isn’t money, but suggested it would make “good practical sense” for our statute to cover stock as well as money. 856 F. 3d, at 492. Meanwhile, Judge Manion dissented, countering that it’s a judge’s job only to apply, not revise or update, the terms of statutes. See id., at 493. The Eighth Circuit made much the same point when it addressed the question. See Union Pacific R. Co. v. United States, 865 F. 3d 1045, 1048–1049 (2017). Judge Man- ion and the Eighth Circuit were right. Written laws are meant to be understood and lived by. If a fog of uncertainty surrounded them, if their meaning could shift with the latest judicial whim, the point of reducing them to writing would be lost. That is why it’s a “fundamental canon of statutory construction” that words generally should be “interpreted as taking their ordinary, contemporary, common meaning . . . at the time Congress enacted the statute.” Perrin, 444 U. S., at 42. Congress alone has the institutional competence, democratic legitimacy, and (most importantly) constitutional authority to revise statutes in light of new social problems and preferences. Until it exercises that power, the people may rely on the original meaning of the written law. This hardly leaves us, as the dissent worries, “trapped in a monetary time warp, forever limited to those forms of money commonly used in the 1930’s.” Post, at 3 (opinion of Breyer, J.). While every statute’s meaning is fixed at the time of enactment, new applications may arise in light of changes in the world. So “money,” as used in this statute, must always mean a “medium of exchange.” But what qualifies as a “medium of exchange” may depend on the facts of the day. Take electronic transfers of paychecks. Maybe they weren’t common in 1937, but we do not doubt they would qualify today as “money remuneration” under the statute’s original public meaning. The problem with the government’s and the dissent’s position today is not that stock and stock options weren’t common in 1937, but that they were not then—and are not now—recognized as mediums of exchange. The judgment of the Seventh Circuit is reversed, and the case is remanded for further proceedings consistent with this opinion. It is so ordered.

586.US.2018_17-1104

Petitioners produced equipment for three Navy ships. The equipment required asbestos insulation or asbestos parts to function as intended, but the manufacturers did not always incorporate the asbestos into their products. Instead, the manufacturers delivered much of the equipment to the Navy without asbestos, and the Navy later added the asbestos to the equipment. Two Navy veterans, Kenneth McAfee and John DeVries, were exposed to asbestos on the ships and developed cancer. They and their wives sued the manufacturers, alleging that the asbestos exposure caused the cancer and contending that the manufacturers were negligent in failing to warn about the dangers of asbestos in the integrated products. Raising the “bare-metal defense,” the manufacturers argued that they should not be liable for harms caused by later-added third-party parts. The District Court granted summary judgment to the manufacturers, but the Third Circuit, adopting a foreseeability approach, vacated and remanded. Held: In the maritime tort context, a product manufacturer has a duty to warn when its product requires incorporation of a part, the manufacturer knows or has reason to know that the integrated product is likely to be dangerous for its intended uses, and the manufacturer has no reason to believe that the product’s users will realize that danger. Pp. 4–11. (a) Tort law imposes a duty to exercise reasonable care on those whose conduct presents a risk of harm to others. That includes a duty to warn when the manufacturer “knows or has reason to know” that its product “is or is likely to be dangerous for the use for which it is supplied” and “has no reason to believe” that the product’s users will realize that danger. 2 Restatement (Second) of Torts §388. Three approaches have emerged on how to apply that “duty to warn” principle when a manufacturer’s product requires later incorporation of a dangerous part in order for the integrated product to function as intended. The first—the foreseeability rule—provides that a manufacturer may be liable when it was foreseeable that its product would be used with another product or part, even if the manufacturer’s product did not require use or incorporation of that other product or part. The second—the bare-metal defense—provides that if a manufacturer did not itself make, sell, or distribute the part or incorporate the part into the product, the manufacturer is not liable for harm caused by the integrated product—even if the product required incorporation of the part and the manufacturer knew that the integrated product was likely to be dangerous for its intended uses. A third approach, falling between those two, imposes on the manufacturer a duty to warn when its product requires incorporation of a part and the manufacturer knows or has reason to know that the integrated product is likely to be dangerous for its intended uses. The third approach is most appropriate for this maritime context. The foreseeability rule would sweep too broadly, imposing a difficult and costly burden on manufacturers, while simultaneously overwarning users. The bare-metal defense ultimately goes too far in the other direction. After all, a manufacturer that supplies a product that is dangerous in and of itself and a manufacturer that supplies a product that requires incorporation of a part that the manufacturer knows or has reason to know is likely to make the integrated product dangerous for its intended uses both “kno[w] or ha[ve] reason to know” that the product “is or is likely to be dangerous for the use for which it is supplied.” And in the latter case, the product manufacturer will often be in a better position than the parts manufacturer to warn of the danger, because the product manufacturer knows the nature of the ultimate integrated product. Requiring a warning in these circumstances will not impose a significant burden on manufacturers, who already have a duty to warn of the dangers of their own products. Nor will it result in substantial uncertainty about when product manufacturers must provide warnings, because the rule requires a manufacturer to warn only when its product requires a part in order for the integrated product to function as intended. And this Court is unaware of any substantial overwarning problems in those jurisdictions that have adopted the approach taken here. Requiring the product manufacturer to warn when its product requires incorporation of a part that makes the integrated product dangerous for its intended uses is especially appropriate in the context of maritime law, which has always recognized a “ ‘special solicitude for the welfare’ ” of sailors. American Export Lines, Inc. v. Alvez, 446 U.S. 274, 285. Pp. 4–10. (b) The maritime tort rule adopted here encompasses all of the following circumstances, so long as the manufacturer knows or has reason to know that the integrated product is likely to be dangerous for its intended uses, and the manufacturer has no reason to believe that the product’s users will realize that danger: (i) a manufacturer directs that the part be incorporated; (ii) a manufacturer itself makes the product with a part that the manufacturer knows will require replacement with a similar part; or (iii) a product would be useless without the part. P. 10. 873 F.3d 232, affirmed. Kavanaugh, J., delivered the opinion of the Court, in which Roberts, C. J., and Ginsburg, Breyer, Sotomayor, and Kagan, JJ., joined. Gorsuch, J., filed a dissenting opinion, in which Thomas and Alito, JJ., joined.

In maritime tort cases, we act as a common-law court, subject to any controlling statutes enacted by Congress. See Exxon Shipping Co. v. Baker, 554 U.S. 471, 507–508 (2008). This maritime tort case raises a question about the scope of a manufacturer’s duty to warn. The manufacturers here produced equipment such as pumps, blowers, and turbines for three Navy ships. The equipment required asbestos insulation or asbestos parts in order to function as intended. When used on the ships, the equipment released asbestos fibers into the air. Two Navy veterans who were exposed to asbestos on the ships developed cancer and later died. The veterans’ families sued the equipment manufacturers, claiming that the manufacturers were negligent in failing to warn of the dangers of asbestos. The plaintiffs contend that a manufacturer has a duty to warn when the manufacturer’s product requires incorporation of a part (here, asbestos) that the manufacturer knows is likely to make the integrated product dangerous for its intended uses. The manufacturers respond that they had no duty to warn because they did not themselves incorporate the asbestos into their equipment; rather, the Navy added the asbestos to the equipment after the equipment was already on board the ships. We agree with the plaintiffs. In the maritime tort context, a product manufacturer has a duty to warn when (i) its product requires incorporation of a part, (ii) the manufacturer knows or has reason to know that the integrated product is likely to be dangerous for its intended uses, and (iii) the manufacturer has no reason to believe that the product’s users will realize that danger. The District Court did not apply that test when granting summary judgment to the defendant manufacturers. Although we do not agree with all of the reasoning of the U. S. Court of Appeals for the Third Circuit, we affirm its judgment requiring the District Court to reconsider its prior grants of summary judgment to the defendant manufacturers. I Kenneth McAfee served in the U. S. Navy for more than 20 years. As relevant here, McAfee worked on the U. S. S. Wanamassa from 1977 to 1980 and then on the U. S. S. Commodore from 1982 to 1986. John DeVries served in the U. S. Navy from 1957 to 1960. He worked on the U. S. S. Turner. Those ships were outfitted with equipment such as pumps, blowers, and turbines. That equipment required asbestos insulation or asbestos parts in order to function as intended. When used as intended, that equipment can cause the release of asbestos fibers into the air. If inhaled or ingested, those fibers may cause various illnesses. Five businesses—Air and Liquid Systems, CBS, Foster Wheeler, Ingersoll Rand, and General Electric—produced some of the equipment that was used on the ships. Al-though the equipment required asbestos insulation or asbestos parts in order to function as intended, those businesses did not always incorporate the asbestos into their products. Instead, the businesses delivered much of the equipment to the Navy without asbestos. The equipment was delivered in a condition known as “bare-metal.” In those situations, the Navy later added the asbestos to the equipment.[1] McAfee and DeVries allege that their exposure to the asbestos caused them to develop cancer. They and their wives sued the equipment manufacturers in Pennsylvania state court. (McAfee and DeVries later died during the course of the ongoing litigation.) The plaintiffs did not sue the Navy because they apparently believed the Navy was immune. See Feres v. United States, 340 U.S. 135 (1950). The plaintiffs also could not recover much from the manufacturers of the asbestos insulation and asbestos parts because those manufacturers had gone bankrupt. As to the manufacturers of the equipment—such as the pumps, blowers, and turbines—the plaintiffs claimed that those manufacturers negligently failed to warn them of the dangers of asbestos in the integrated products. If the manufacturers had provided warnings, the workers on the ships presumably could have worn respiratory masks and thereby avoided the danger. Invoking federal maritime jurisdiction, the manufacturers removed the cases to federal court. The manufacturers then moved for summary judgment on the ground that manufacturers should not be liable for harms caused by later-added third-party parts. That defense is known as the “bare-metal defense.” The District Court granted the manufacturers’ motions for summary judgment. The U. S. Court of Appeals for the Third Circuit vacated and remanded. In re Asbestos Prods. Liability Litigation, 873 F.3d 232, 241 (2017). The Third Circuit held that “a manufacturer of a bare-metal product may be held liable for a plaintiff’s injuries suffered from later-added asbestos-containing materials” if the manufacturer could foresee that the product would be used with the later-added asbestos-containing materials. Id., at 240. We granted certiorari to resolve a disagreement among the Courts of Appeals about the validity of the bare-metal defense under maritime law. 584 U. S. ___ (2018). Compare 873 F.3d 232 (case below), with Lindstrom v. A-C Prod. Liability Trust, 424 F.3d 488 (CA6 2005). II Article III of the Constitution grants the federal courts jurisdiction over maritime cases. Under 28 U. S. C. §1333, the federal courts have “original jurisdiction, exclusive of the courts of the States, of . . . [a]ny civil case of admiralty or maritime jurisdiction, saving to suitors in all cases all other remedies to which they are otherwise entitled.” When a federal court decides a maritime case, it acts as a federal “common law court,” much as state courts do in state common-law cases. Exxon Shipping Co., 554 U. S., at 507. Subject to direction from Congress, the federal courts fashion federal maritime law. See id., at 508, n. 21; Miles v. Apex Marine Corp., 498 U.S. 19, 27 (1990); United States v. Reliable Transfer Co., 421 U.S. 397, 409 (1975); Detroit Trust Co. v. The Thomas Barlum, 293 U.S. 21, 42–44 (1934). In formulating federal maritime law, the federal courts may examine, among other sources, judicial opinions, legislation, treatises, and scholarly writings. See Exxon Co., U. S. A. v. Sofec, Inc., 517 U.S. 830, 839 (1996); East River S. S. Corp. v. Transamerica Delaval Inc., 476 U.S. 858, 864 (1986). This is a maritime tort case. The plaintiffs allege that the defendant equipment manufacturers were negligent in failing to warn about the dangers of asbestos. “The general maritime law has recognized the tort of negligence for more than a century . . . .” Norfolk Shipbuilding & Drydock Corp. v. Garris, 532 U.S. 811, 820 (2001); see also Kermarec v. Compagnie Generale Transatlantique, 358 U.S. 625, 631–632 (1959). Maritime law has likewise recognized common-law principles of products liability for decades. See East River S. S. Corp., 476 U. S., at 865. In this negligence case, we must decide whether a manufacturer has a duty to warn when the manufacturer’s product requires later incorporation of a dangerous part—here, asbestos—in order for the integrated product to function as intended. We start with basic tort-law principles. Tort law im- poses “a duty to exercise reasonable care” on those whose conduct presents a risk of harm to others. 1 Restatement (Third) of Torts: Liability for Physical and Emotional Harm §7, p. 77 (2005). For the manufacturer of a product, the general duty of care includes a duty to warn when the manufacturer “knows or has reason to know” that its product “is or is likely to be dangerous for the use for which it is supplied” and the manufacturer “has no reason to believe” that the product’s users will realize that danger. 2 Restatement (Second) of Torts §388, p. 301 (1963–1964). In tort cases, the federal and state courts have not reached consensus on how to apply that general tort-law “duty to warn” principle when the manufacturer’s product requires later incorporation of a dangerous part in order for the integrated product to function as intended. Three approaches have emerged. The first approach is the more plaintiff-friendly foreseeability rule that the Third Circuit adopted in this case: A manufacturer may be liable when it was foreseeable that the manufacturer’s product would be used with another product or part, even if the manufacturer’s product did not require use or incorporation of that other product or part. See, e.g., 873 F. 3d, at 240; Kochera v. Foster Wheeler, LLC, 2015 WL 5584749, *4 (SD Ill., Sept. 23, 2015); Chicano v. General Elec. Co., 2004 WL 2250990, *9 (ED Pa., Oct. 5, 2004); McKenzie v. A. W. Chesterson Co., 277 Ore. App. 728, 749–750, 373 P.3d 150, 162 (2016). The second approach is the more defendant-friendly bare-metal defense that the manufacturers urge here: If a manufacturer did not itself make, sell, or distribute the part or incorporate the part into the product, the manufacturer is not liable for harm caused by the integrated product—even if the product required incorporation of the part and the manufacturer knew that the integrated product was likely to be dangerous for its intended uses. See, e.g., Lindstrom, 424 F. 3d, at 492, 495–497; Evans v. CBS Corp., 230 F. Supp. 3d 397, 403–405 (Del. 2017); Cabasug v. Crane Co., 989 F. Supp. 2d 1027, 1041 (Haw. 2013). The third approach falls between those two approaches. Under the third approach, foreseeability that the product may be used with another product or part that is likely to be dangerous is not enough to trigger a duty to warn. But a manufacturer does have a duty to warn when its product requires incorporation of a part and the manufacturer knows or has reason to know that the integrated product is likely to be dangerous for its intended uses. Under that approach, the manufacturer may be liable even when the manufacturer does not itself incorporate the required part into the product. See, e.g., Quirin v. Lorillard Tobacco Co., 17 F. Supp. 3d 760, 769–770 (ND Ill. 2014); In re New York City Asbestos Litigation, 27 N.Y.3d 765, 793–794, 59 N. E. 3d 458, 474 (2016); May v. Air & Liquid Systems Corp., 446 Md. 1, 29, 129 A.3d 984, 1000 (2015). We conclude that the third approach is the most appropriate for this maritime tort context. To begin, we agree with the manufacturers that a rule of mere foreseeability would sweep too broadly. See gener- ally 1 Restatement (Third) of Torts: Liability for Physical and Emotional Harm §7, Comment j, at 82; 2 Restatement (Second) of Torts §395, Comment j, at 330. Many products can foreseeably be used in numerous ways with numerous other products and parts. Requiring a product manufacturer to imagine and warn about all of those possible uses—with massive liability looming for failure to correctly predict how its product might be used with other prod- ucts or parts—would impose a difficult and costly burden on manufacturers, while simultaneously overwarning users. In light of that uncertainty and unfairness, we reject the foreseeability approach for this maritime context. That said, we agree with the plaintiffs that the bare-metal defense ultimately goes too far in the other direction. In urging the bare-metal defense, the manufacturers contend that a business generally has “no duty” to “control the conduct of a third person as to prevent him from causing physical harm to another.” Id., §315, at 122. That is true, but it is also beside the point here. After all, when a manufacturer’s product is dangerous in and of itself, the manufacturer “knows or has reason to know” that the product “is or is likely to be dangerous for the use for which it is supplied.” Id., §388, at 301. The same holds true, we conclude, when the manufacturer’s product requires incorporation of a part that the manufacturer knows or has reason to know is likely to make the integrated product dangerous for its intended uses. As a matter of maritime tort law, we find no persuasive reason to distinguish those two similar situations for purposes of a manufacturer’s duty to warn. See Restatement (Third) of Torts: Products Liability §2, Comment i, p. 30 (1997) (“[W]arnings also may be needed to inform users and consumers of nonobvious and not generally known risks that unavoidably inhere in using or consuming the product”). Importantly, the product manufacturer will often be in a better position than the parts manufacturer to warn of the danger from the integrated product. See generally G. Calabresi, The Costs of Accidents 311–318 (1970). The product manufacturer knows the nature of the ultimate integrated product and is typically more aware of the risks associated with that integrated product. By contrast, a parts manufacturer may be aware only that its part could conceivably be used in any number of ways in any number of products. A parts manufacturer may not always be aware that its part will be used in a way that poses a risk of danger.[2] To be sure, as the manufacturers correctly point out, issuing a warning costs time and money. But the burden usually is not significant. Manufacturers already have a duty to warn of the dangers of their own products. That duty typically imposes a light burden on manufacturers. See, e.g., Davis v. Wyeth Labs., Inc., 399 F.2d 121, 131 (CA9 1968); Butler v. L. Sonneborn Sons, Inc., 296 F.2d 623, 625–626 (CA2 1961); Ross Labs. v. Thies, 725 P.2d 1076, 1079 (Alaska 1986); Moran v. Faberge, Inc., 273 Md. 538, 543–544, 332 A.2d 11, 15 (1975). Requiring a manufacturer to also warn when the manufacturer knows or has reason to know that a required later-added part is likely to make the integrated product dangerous for its intended uses should not meaningfully add to that burden. The manufacturers also contend that requiring a warning even when they have not themselves incorporated the part into the product will lead to uncertainty about when product manufacturers must provide warnings. But the manufacturers have not pointed to any substantial confusion in those jurisdictions that have adopted this approach. And the rule that we adopt here is tightly cabined. The rule does not require that manufacturers warn in cases of mere foreseeability. The rule requires that manufacturers warn only when their product requires a part in order for the integrated product to function as intended. The manufacturers further assert that requiring a warning in these circumstances will lead to excessive warning of consumers. Again, however, we are not aware of substantial overwarning problems in those jurisdictions that have adopted this approach. And because the rule we adopt here applies only in certain narrow circumstances, it will not require a plethora of new warnings. Requiring the product manufacturer to warn when its product requires incorporation of a part that makes the integrated product dangerous for its intended uses—and not just when the manufacturer itself incorporates the part into the product—is especially appropriate in the maritime context. Maritime law has always recognized a “special solicitude for the welfare” of those who undertake to “venture upon hazardous and unpredictable sea voy- ages.” American Export Lines, Inc. v. Alvez, 446 U.S. 274, 285 (1980) (internal quotation marks omitted). The plaintiffs in this case are the families of veterans who served in the U. S. Navy. Maritime law’s longstanding solicitude for sailors reinforces our decision to require a warning in these circumstances. See Yamaha Motor Corp., U. S. A. v. Calhoun, 516 U.S. 199, 213 (1996); Miles, 498 U. S., at 36; Moragne v. States Marine Lines, Inc., 398 U.S. 375, 387 (1970). For those reasons, we conclude as follows: In the maritime tort context, a product manufacturer has a duty to warn when (i) its product requires incorporation of a part, (ii) the manufacturer knows or has reason to know that the integrated product is likely to be dangerous for its intended uses, and (iii) the manufacturer has no reason to believe that the product’s users will realize that danger. We do not purport to define the proper tort rule outside of the maritime context. One final point for clarity: Courts have determined that this rule applies in certain related situations, including when: (i) a manufacturer directs that the part be incorporated, see, e.g., Bell v. Foster Wheeler Energy Corp., 2016 WL 5780104, *6–*7 (ED La., Oct. 4, 2016); (ii) a manufacturer itself makes the product with a part that the manufacturer knows will require replacement with a similar part, see, e.g., Chesher v. 3M Co., 234 F. Supp. 3d 693, 713–714 (S. C. 2017); Quirin, 17 F. Supp. 3d, at 769–770; May, 446 Md., at 29, 129 A. 3d, at 1000; or (iii) a product would be useless without the part, see, e.g., In re New York City Asbestos Litigation, 27 N. Y. 3d, at 793–794, 59 N. E. 3d, at 474. In all of those situations, courts have said that the product in effect requires the part in order for the integrated product to function as intended. We agree. The maritime tort rule we adopt today therefore encompasses those situations, so long as the manufacturer knows or has reason to know that the integrated product is likely to be dangerous for its intended uses, and the manufacturer has no reason to believe that the product’s users will realize that danger. * * * In the maritime tort context, we hold that a product manufacturer has a duty to warn when (i) its product requires incorporation of a part, (ii) the manufacturer knows or has reason to know that the integrated product is likely to be dangerous for its intended uses, and (iii) the manufacturer has no reason to believe that the product’s users will realize that danger. The District Court should evaluate the evidence under that rule. Although we do not agree with all of the reasoning of the Third Circuit, we affirm its judgment requiring the District Court to reconsider its prior grants of summary judgment to the defendant manufacturers. It is so ordered. Notes 1 Sometimes, the equipment manufacturers themselves added the asbestos to the equipment. Even in those situations, however, the Navy later replaced the asbestos parts with third-party asbestos parts. 2 We do not rule out the possibility that, in certain circumstances, the parts manufacturer may also have a duty to warn.

587.US.2018_17-204

Apple Inc. sells iPhone applications, or apps, directly to iPhone owners through its App Store—the only place where iPhone owners may lawfully buy apps. Most of those apps are created by independent developers under contracts with Apple. Apple charges the developers a $99 annual membership fee, allows them to set the retail price of the apps, and charges a 30% commission on every app sale. Respondents, four iPhone owners, sued Apple, alleging that the company has unlawfully monopolized the aftermarket for iPhone apps. Apple moved to dismiss, arguing that the iPhone owners could not sue because they were not direct purchasers from Apple under Illinois Brick Co. v. Illinois, 431 U.S. 720. The District Court agreed, but the Ninth Circuit reversed, concluding that the iPhone owners were direct purchasers because they purchased apps directly from Apple. Held: Under Illinois Brick, the iPhone owners were direct purchasers who may sue Apple for alleged monopolization. Pp. 4–14. (a) This straightforward conclusion follows from the text of the antitrust laws and from this Court’s precedent. Section 4 of the Clayton Act provides that “any person who shall be injured in his business or property by reason of anything forbidden in the antitrust laws may sue.” 15 U. S. C. §15(a). That broad text readily covers consumers who purchase goods or services at higher-than-competitive prices from an allegedly monopolistic retailer. Applying §4, this Court has consistently stated that “the immediate buyers from the alleged antitrust violators” may maintain a suit against the antitrust violators, Kansas v. UtiliCorp United Inc., 497 U.S. 199, 207, but has ruled that indirect purchasers who are two or more steps removed from the violator in a distribution chain may not sue. Unlike the consumer in Illinois Brick, the iPhone owners here are not consumers at the bottom of a vertical distribution chain who are attempting to sue manufacturers at the top of the chain. The absence of an intermediary in the distribution chain between Apple and the consumer is dispositive. Pp. 4–7. (b) Apple argues that Illinois Brick allows consumers to sue only the party who sets the retail price, whether or not the party sells the good or service directly to the complaining party. But that theory suffers from three main problems. First, it contradicts statutory text and precedent by requiring the Court to rewrite the rationale of Illinois Brick and to gut its longstanding bright-line rule. Any ambiguity in Illinois Brick should be resolved in the direction of the statutory text, which states that “any person” injured by an antitrust violation may sue to recover damages. Second, Apple’s theory is not persuasive economically or legally. It would draw an arbitrary and unprincipled line among retailers based on their financial arrangements with their manufacturers or suppliers. And it would permit a consumer to sue a monopolistic retailer when the retailer set the retail price by marking up the price it had paid the manufacturer or supplier for the good or service but not when the manufacturer or supplier set the retail price and the retailer took a commission on each sale. Third, Apple’s theory would provide a roadmap for monopolistic retailers to structure transactions with manufacturers or suppliers so as to evade antitrust claims by consumers and thereby thwart effective antitrust enforcement. Pp. 7–11. (c) Contrary to Apple’s argument, the three Illinois Brick rationales for adopting the direct-purchaser rule cut strongly in respondents’ favor. First, Apple posits that allowing only the upstream app developers—and not the downstream consumers—to sue Apple would mean more effective antitrust enforcement. But that makes little sense, and it would directly contradict the longstanding goal of effective private enforcement and consumer protection in antitrust cases. Second, Apple warns that calculating the damages in successful consumer antitrust suits against monopolistic retailers might be complicated. But Illinois Brick is not a get-out-of-court-free card for monopolistic retailers to play any time that a damages calculation might be complicated. Third, Apple claims that allowing consumers to sue will result in “conflicting claims to a common fund—the amount of the alleged overcharge.” Illinois Brick, 431 U. S., at 737. But this is not a case where multiple parties at different levels of a distribution chain are trying to recover the same passed-through overcharge initially levied by the manufacturer at the top of the chain, cf. id., at 726–727. Pp. 11–14. 846 F.3d 313, affirmed. Kavanaugh, J., delivered the opinion of the Court, in which Ginsburg, Breyer, Sotomayor, and Kagan, JJ., joined. Gorsuch, J., filed a dissenting opinion, in which Roberts, C. J., and Thomas and Alito, JJ., joined.

In 2007, Apple started selling iPhones. The next year, Apple launched the retail App Store, an electronic store where iPhone owners can purchase iPhone applications from Apple. Those “apps” enable iPhone owners to send messages, take photos, watch videos, buy clothes, order food, arrange transportation, purchase concert tickets, donate to charities, and the list goes on. “There’s an app for that” has become part of the 21st-century American lexicon. In this case, however, several consumers contend that Apple charges too much for apps. The consumers argue, in particular, that Apple has monopolized the retail market for the sale of apps and has unlawfully used its monopolistic power to charge consumers higher-than-competitive prices. A claim that a monopolistic retailer (here, Apple) has used its monopoly to overcharge consumers is a classic antitrust claim. But Apple asserts that the consumer-plaintiffs in this case may not sue Apple because they supposedly were not “direct purchasers” from Apple under our decision in Illinois Brick Co. v. Illinois, 431 U.S. 720, 745–746 (1977). We disagree. The plaintiffs purchased apps directly from Apple and therefore are direct purchasers under Illinois Brick. At this early pleadings stage of the litigation, we do not assess the merits of the plaintiffs’ antitrust claims against Apple, nor do we consider any other defenses Apple might have. We merely hold that the Illinois Brick direct-purchaser rule does not bar these plaintiffs from suing Apple under the antitrust laws. We affirm the judgment of the U. S. Court of Appeals for the Ninth Circuit. I In 2007, Apple began selling iPhones. In July 2008, Apple started the App Store. The App Store now contains about 2 million apps that iPhone owners can download. By contract and through technological limitations, the App Store is the only place where iPhone owners may lawfully buy apps. For the most part, Apple does not itself create apps. Rather, independent app developers create apps. Those independent app developers then contract with Apple to make the apps available to iPhone owners in the App Store. Through the App Store, Apple sells the apps directly to iPhone owners. To sell an app in the App Store, app developers must pay Apple a $99 annual membership fee. Apple requires that the retail sales price end in $0.99, but otherwise allows the app developers to set the retail price. Apple keeps 30 percent of the sales price, no matter what the sales price might be. In other words, Apple pockets a 30 percent commission on every app sale. In 2011, four iPhone owners sued Apple. They allege that Apple has unlawfully monopolized “the iPhone apps aftermarket.” App. to Pet. for Cert. 53a. The plaintiffs allege that, via the App Store, Apple locks iPhone owners “into buying apps only from Apple and paying Apple’s 30% fee, even if” the iPhone owners wish “to buy apps elsewhere or pay less.” Id., at 45a. According to the complaint, that 30 percent commission is “pure profit” for Apple and, in a competitive environment with other retailers, “Apple would be under considerable pressure to substantially lower its 30% profit margin.” Id., at 54a–55a. The plaintiffs allege that in a competitive market, they would be able to “choose between Apple’s high-priced App Store and less costly alternatives.” Id., at 55a. And they allege that they have “paid more for their iPhone apps than they would have paid in a competitive market.” Id., at 53a. Apple moved to dismiss the complaint, arguing that the iPhone owners were not direct purchasers from Apple and therefore may not sue. In Illinois Brick, this Court held that direct purchasers may sue antitrust violators, but also ruled that indirect purchasers may not sue. The District Court agreed with Apple and dismissed the complaint. According to the District Court, the iPhone owners were not direct purchasers from Apple because the app developers, not Apple, set the consumers’ purchase price. The Ninth Circuit reversed. The Ninth Circuit concluded that the iPhone owners were direct purchasers under Illinois Brick because the iPhone owners purchased apps directly from Apple. According to the Ninth Circuit, Illinois Brick means that a consumer may not sue an alleged monopolist who is two or more steps removed from the consumer in a vertical distribution chain. See In re Apple iPhone Antitrust Litig., 846 F.3d 313, 323 (2017). Here, however, the consumers purchased directly from Apple, the alleged monopolist. Therefore, the Ninth Circuit held that the iPhone owners could sue Apple for allegedly monopolizing the sale of iPhone apps and charging higher-than-competitive prices. Id., at 324. We granted certiorari. 585 U. S. ___ (2018). II A The plaintiffs’ allegations boil down to one straightforward claim: that Apple exercises monopoly power in the retail market for the sale of apps and has unlawfully used its monopoly power to force iPhone owners to pay Apple higher-than-competitive prices for apps. According to the plaintiffs, when iPhone owners want to purchase an app, they have only two options: (1) buy the app from Apple’s App Store at a higher-than-competitive price or (2) do not buy the app at all. Any iPhone owners who are dissatisfied with the selection of apps available in the App Store or with the price of the apps available in the App Store are out of luck, or so the plaintiffs allege. The sole question presented at this early stage of the case is whether these consumers are proper plaintiffs for this kind of antitrust suit—in particular, our precedents ask, whether the consumers were “direct purchasers” from Apple. Illinois Brick, 431 U. S., at 745–746. It is undisputed that the iPhone owners bought the apps directly from Apple. Therefore, under Illinois Brick, the iPhone owners were direct purchasers who may sue Apple for alleged monopolization. That straightforward conclusion follows from the text of the antitrust laws and from our precedents. First is text: Section 2 of the Sherman Act makes it unlawful for any person to “monopolize, or attempt to monopolize, or combine or conspire with any other person or persons, to monopolize any part of the trade or commerce among the several States, or with foreign nations.” 26Stat. 209, 15 U. S. C. §2. Section 4 of the Clayton Act in turn provides that “any person who shall be injured in his business or property by reason of anything forbidden in the antitrust laws may sue . . . the defendant . . . and shall recover threefold the damages by him sustained, and the cost of suit, including a reasonable attorney’s fee.” 38 Stat. 731, 15 U. S. C. §15(a) (emphasis added). The broad text of §4—“any person” who has been “injured” by an antitrust violator may sue—readily covers consumers who purchase goods or services at higher-than-competitive prices from an allegedly monopolistic retailer. Second is precedent: Applying §4, we have consistently stated that “the immediate buyers from the alleged antitrust violators” may maintain a suit against the antitrust violators. Kansas v. UtiliCorp United Inc., 497 U.S. 199, 207 (1990); see also Illinois Brick, 431 U. S., at 745–746. At the same time, incorporating principles of proximate cause into §4, we have ruled that indirect purchasers who are two or more steps removed from the violator in a distribution chain may not sue. Our decision in Illinois Brick established a bright-line rule that authorizes suits by direct purchasers but bars suits by indirect purchasers. Id., at 746.[1] The facts of Illinois Brick illustrate the rule. Illinois Brick Company manufactured and distributed concrete blocks. Illinois Brick sold the blocks primarily to masonry contractors, and those contractors in turn sold masonry structures to general contractors. Those general contractors in turn sold their services for larger construction projects to the State of Illinois, the ultimate consumer of the blocks. The consumer State of Illinois sued the manufacturer Illinois Brick. The State alleged that Illinois Brick had engaged in a conspiracy to fix the price of concrete blocks. According to the complaint, the State paid more for the concrete blocks than it would have paid absent the price-fixing conspiracy. The monopoly overcharge allegedly flowed all the way down the distribution chain to the ultimate consumer, who was the State of Illinois. This Court ruled that the State could not bring an antitrust action against Illinois Brick, the alleged violator, because the State had not purchased concrete blocks directly from Illinois Brick. The proper plaintiff to bring that claim against Illinois Brick, the Court stated, would be an entity that had purchased directly from Illinois Brick. Ibid. The bright-line rule of Illinois Brick, as articulated in that case and as we reiterated in UtiliCorp, means that indirect purchasers who are two or more steps removed from the antitrust violator in a distribution chain may not sue. By contrast, direct purchasers—that is, those who are “the immediate buyers from the alleged antitrust violators”—may sue. UtiliCorp, 497 U. S., at 207. For example, if manufacturer A sells to retailer B, and retailer B sells to consumer C, then C may not sue A. But B may sue A if A is an antitrust violator. And C may sue B if B is an antitrust violator. That is the straightforward rule of Illinois Brick. See Loeb Industries, Inc. v. Sumi- tomo Corp., 306 F.3d 469, 481–482 (CA7 2002) (Wood, J.).[2] In this case, unlike in Illinois Brick, the iPhone owners are not consumers at the bottom of a vertical distribution chain who are attempting to sue manufacturers at the top of the chain. There is no intermediary in the distribution chain between Apple and the consumer. The iPhone owners purchase apps directly from the retailer Apple, who is the alleged antitrust violator. The iPhone owners pay the alleged overcharge directly to Apple. The absence of an intermediary is dispositive. Under Illinois Brick, the iPhone owners are direct purchasers from Apple and are proper plaintiffs to maintain this antitrust suit. B All of that seems simple enough. But Apple argues strenuously against that seemingly simple conclusion, and we address its arguments carefully. For this kind of retailer case, Apple’s theory is that Illinois Brick allows consumers to sue only the party who sets the retail price, whether or not that party sells the good or service directly to the complaining party. Apple says that its theory accords with the economics of the transaction. Here, Apple argues that the app developers, not Apple, set the retail price charged to consumers, which according to Apple means that the consumers may not sue Apple. We see three main problems with Apple’s “who sets the price” theory. First, Apple’s theory contradicts statutory text and precedent. As we explained above, the text of §4 broadly affords injured parties a right to sue under the antitrust laws. And our precedent in Illinois Brick established a bright-line rule where direct purchasers such as the consumers here may sue antitrust violators from whom they purchased a good or service. Illinois Brick, as we read the opinion, was not based on an economic theory about who set the price. Rather, Illinois Brick sought to ensure an effective and efficient litigation scheme in antitrust cases. To do so, the Court drew a bright line that allowed direct purchasers to sue but barred indirect purchasers from suing. When there is no intermediary between the purchaser and the antitrust violator, the purchaser may sue. The Illinois Brick bright-line rule is grounded on the “belief that simplified administration improves antitrust enforcement.” 2A P. Areeda, H. Hovenkamp, R. Blair, & C. Durrance, Antitrust Law ¶346e, p. 194 (4th ed. 2014) (Areeda & Hovenkamp). Apple’s theory would require us to rewrite the rationale of Illinois Brick and to gut the longstanding bright-line rule. To the extent that Illinois Brick leaves any ambiguity about whether a direct purchaser may sue an antitrust violator, we should resolve that ambiguity in the direction of the statutory text. And under the text, direct purchasers from monopolistic retailers are proper plaintiffs to sue those retailers. Second, in addition to deviating from statutory text and precedent, Apple’s proposed rule is not persuasive economically or legally. Apple’s effort to transform Illinois Brick from a direct-purchaser rule to a “who sets the price” rule would draw an arbitrary and unprincipled line among retailers based on retailers’ financial arrangements with their manufacturers or suppliers. In the retail context, the price charged by a retailer to a consumer is often a result (at least in part) of the price charged by the manufacturer or supplier to the retailer, or of negotiations between the manufacturer or supplier and the retailer. Those agreements between manufacturer or supplier and retailer may take myriad forms, including for example a markup pricing model or a commission pricing model. In a traditional markup pricing model, a hypothetical monopolistic retailer might pay $6 to the manufacturer and then sell the product for $10, keeping $4 for itself. In a commission pricing model, the retailer might pay nothing to the manufacturer; agree with the manufacturer that the retailer will sell the product for $10 and keep 40 percent of the sales price; and then sell the product for $10, send $6 back to the manufacturer, and keep $4. In those two different pricing scenarios, everything turns out to be economically the same for the manufacturer, retailer, and consumer. Yet Apple’s proposed rule would allow a consumer to sue the monopolistic retailer in the former situation but not the latter. In other words, under Apple’s rule a consumer could sue a monopolistic retailer when the retailer set the retail price by marking up the price it had paid the manufacturer or supplier for the good or service. But a consumer could not sue a monopolistic retailer when the manufacturer or supplier set the retail price and the retailer took a commission on each sale. Apple’s line-drawing does not make a lot of sense, other than as a way to gerrymander Apple out of this and similar lawsuits. In particular, we fail to see why the form of the upstream arrangement between the manufacturer or supplier and the retailer should determine whether a monopolistic retailer can be sued by a downstream consumer who has purchased a good or service directly from the retailer and has paid a higher-than-competitive price because of the retailer’s unlawful monopolistic conduct. As the Court of Appeals aptly stated, “the distinction between a markup and a commission is immaterial.” 846 F. 3d, at 324. A leading antitrust treatise likewise states: “Denying standing because ‘title’ never passes to a broker is an overly lawyered approach that ignores the reality that a distribution system that relies on brokerage is economically indistinguishable from one that relies on purchaser-resellers.” 2A Areeda & Hovenkamp ¶345, at 183. If a retailer has engaged in unlawful monopolistic conduct that has caused consumers to pay higher-than-competitive prices, it does not matter how the retailer structured its relationship with an upstream manufacturer or supplier—whether, for example, the retailer employed a markup or kept a commission. To be sure, if the monopolistic retailer’s conduct has not caused the consumer to pay a higher-than-competitive price, then the plaintiff’s damages will be zero. Here, for example, if the competitive commission rate were 10 percent rather than 30 percent but Apple could prove that app developers in a 10 percent commission system would always set a higher price such that consumers would pay the same retail price regardless of whether Apple’s commission was 10 percent or 30 percent, then the consumers’ damages would presumably be zero. But we cannot assume in all cases—as Apple would necessarily have us do—that a monopolistic retailer who keeps a commission does not ever cause the consumer to pay a higher-than-competitive price. We find no persuasive legal or economic basis for such a blanket assertion. In short, we do not understand the relevance of the upstream market structure in deciding whether a downstream consumer may sue a monopolistic retailer. Apple’s rule would elevate form (what is the precise arrangement between manufacturers or suppliers and retailers?) over substance (is the consumer paying a higher price because of the monopolistic retailer’s actions?). If the retailer’s unlawful monopolistic conduct caused a consumer to pay the retailer a higher-than-competitive price, the consumer is entitled to sue the retailer under the antitrust laws. Third, if accepted, Apple’s theory would provide a roadmap for monopolistic retailers to structure transactions with manufacturers or suppliers so as to evade antitrust claims by consumers and thereby thwart effective antitrust enforcement. Consider a traditional supplier-retailer relationship, in which the retailer purchases a product from the supplier and sells the product with a markup to consumers. Under Apple’s proposed rule, a retailer, instead of buying the product from the supplier, could arrange to sell the product for the supplier without purchasing it from the sup- plier. In other words, rather than paying the supplier a certain price for the product and then marking up the price to sell the product to consumers, the retailer could collect the price of the product from consumers and remit only a fraction of that price to the supplier. That restructuring would allow a monopolistic retailer to insulate itself from antitrust suits by consumers, even in situations where a monopolistic retailer is using its monopoly to charge higher-than-competitive prices to consumers. We decline to green-light monopolistic retailers to exploit their market position in that way. We refuse to rubber-stamp such a blatant evasion of statutory text and judicial precedent. In sum, Apple’s theory would disregard statutory text and precedent, create an unprincipled and economically senseless distinction among monopolistic retailers, and furnish monopolistic retailers with a how-to guide for evasion of the antitrust laws. C In arguing that the Court should transform the direct-purchaser rule into a “who sets the price” rule, Apple insists that the three reasons that the Court identified in Illinois Brick for adopting the direct-purchaser rule apply to this case—even though the consumers here (unlike in Illinois Brick) were direct purchasers from the alleged monopolist. The Illinois Brick Court listed three reasons for barring indirect-purchaser suits: (1) facilitating more effective enforcement of antitrust laws; (2) avoiding complicated damages calculations; and (3) eliminating duplicative damages against antitrust defendants. As we said in UtiliCorp, however, the bright-line rule of Illinois Brick means that there is no reason to ask whether the rationales of Illinois Brick “apply with equal force” in every individual case. 497 U. S., at 216. We should not engage in “an unwarranted and counterproductive exercise to litigate a series of exceptions.” Id., at 217. But even if we engage with this argument, we conclude that the three Illinois Brick rationales—whether considered individually or together—cut strongly in the plaintiffs’ favor here, not Apple’s. First, Apple argues that barring the iPhone owners from suing Apple will better promote effective enforcement of the antitrust laws. Apple posits that allowing only the upstream app developers—and not the downstream consumers—to sue Apple would mean more effective enforcement of the antitrust laws. We do not agree. Leaving consumers at the mercy of monopolistic retailers simply because upstream suppliers could also sue the retailers makes little sense and would directly contradict the longstanding goal of effective private enforcement and consumer protection in antitrust cases. Second, Apple warns that calculating the damages in successful consumer antitrust suits against monopolistic retailers might be complicated. It is true that it may be hard to determine what the retailer would have charged in a competitive market. Expert testimony will often be necessary. But that is hardly unusual in antitrust cases. Illinois Brick is not a get-out-of-court-free card for monopolistic retailers to play any time that a damages calculation might be complicated. Illinois Brick surely did not wipe out consumer antitrust suits against monopolistic retailers from whom the consumers purchased goods or services at higher-than-competitive prices. Moreover, the damages calculation may be just as complicated in a retailer markup case as it is in a retailer commission case. Yet Apple apparently accepts consumers suing monopolistic retailers in a retailer markup case. If Apple accepts that kind of suit, then Apple should also accept consumers suing monopolistic retailers in a retailer commission case. Third, Apple claims that allowing consumers to sue will result in “conflicting claims to a common fund—the amount of the alleged overcharge.” Illinois Brick, 431 U. S., at 737. Apple is incorrect. This is not a case where multiple parties at different levels of a distribution chain are trying to all recover the same passed-through overcharge initially levied by the manufacturer at the top of the chain. Cf. id., at 726–727; Hanover Shoe, Inc. v. United Shoe Machinery Corp., 392 U.S. 481, 483–484 (1968). If the iPhone owners prevail, they will be entitled to the full amount of the unlawful overcharge that they paid to Apple. The overcharge has not been passed on by anyone to anyone. Unlike in Illinois Brick, there will be no need to “trace the effect of the overcharge through each step in the distribution chain.” 431 U. S., at 741. It is true that Apple’s alleged anticompetitive conduct may leave Apple subject to multiple suits by different plaintiffs. But Illinois Brick did not purport to bar multiple liability that is unrelated to passing an overcharge down a chain of distribution. Basic antitrust law tells us that the “mere fact that an antitrust violation produces two different classes of victims hardly entails that their injuries are duplicative of one another.” 2A Areeda & Hovenkamp ¶339d, at 136. Multiple suits are not atypical when the intermediary in a distribution chain is a bottleneck monopolist or monopsonist (or both) between the manufacturer on the one end and the consumer on the other end. A retailer who is both a monopolist and a monopsonist may be liable to different classes of plaintiffs—both to downstream consumers and to upstream suppliers—when the retailer’s unlawful conduct affects both the downstream and upstream markets. Here, some downstream iPhone consumers have sued Apple on a monopoly theory. And it could be that some upstream app developers will also sue Apple on a monopsony theory. In this instance, the two suits would rely on fundamentally different theories of harm and would not assert dueling claims to a “common fund,” as that term was used in Illinois Brick. The consumers seek damages based on the difference between the price they paid and the competitive price. The app developers would seek lost profits that they could have earned in a competitive retail market. Illinois Brick does not bar either category of suit. In short, the three Illinois Brick rationales do not persuade us to remake Illinois Brick and to bar direct-purchaser suits against monopolistic retailers who employ commissions rather than markups. The plaintiffs seek to hold retailers to account if the retailers engage in unlawful anticompetitive conduct that harms consumers who purchase from those retailers. That is why we have antitrust law. * * * Ever since Congress overwhelmingly passed and President Benjamin Harrison signed the Sherman Act in 1890, “protecting consumers from monopoly prices” has been “the central concern of antitrust.” 2A Areeda & Hovenkamp ¶345, at 179. The consumers here purchased apps directly from Apple, and they allege that Apple used its monopoly power over the retail apps market to charge higher-than-competitive prices. Our decision in Illinois Brick does not bar the consumers from suing Apple for Apple’s allegedly monopolistic conduct. We affirm the judgment of the U. S. Court of Appeals for the Ninth Circuit. It is so ordered. Notes 1 Illinois Brick held that the direct-purchaser requirement applies to claims for damages. Illinois Brick did not address injunctive relief, and we likewise do not address injunctive relief in this case. 2 Thirty States and the District of Columbia filed an amicus brief supporting the plaintiffs, and they argue that C should be able to sue A in that hypothetical. They ask us to overrule Illinois Brick to allow such suits. In light of our ruling in favor of the plaintiffs in this case, we have no occasion to consider that argument for overruling Illinois Brick.

587.US.2018_17-1484

The Medicare program offers additional payments to institutions that serve a “disproportionate number” of low-income patients. 42 U. S. C. §§1395ww(d)(5)(F)(i)(I). These payments are calculated in part using what is called a hospital’s “Medicare fraction.” The fraction’s denominator is the time the hospital spent caring for patients who were “entitled to benefits under” Medicare Part A, while the numerator is the time the hospital spent caring for Part-A-entitled patients who were also entitled to income support payments under the Social Security Act. §1395ww(d)(5)(F)(vi)(I). Congress created Medicare Part C in 1997, leading to the question whether Part C enrollees should be counted as “entitled to benefits under” Part A when calculating a hospital’s Medicare fraction. Respondents claim that, because Part C enrollees tend to be wealthier than Part A enrollees, counting them makes the fraction smaller and reduces hospitals’ payments considerably. In 2004, the agency overseeing Medicare issued a final rule declaring that it would count Part C patients, but that rule was later vacated after hospitals filed legal challenges. In 2013, it issued a new rule prospectively readopting the policy of counting Part C patients. In 2014, unable to rely on either the vacated 2004 rule or the prospective 2013 rule, the agency posted on its website the Medicare fractions for fiscal year 2012, noting that they included Part C patients. A group of hospitals, respondents here, sued. They claimed, among other things, that the government had violated the Medicare Act’s requirement to provide public notice and a 60-day comment period for any “rule, requirement, or other statement of policy . . . that establishes or changes a substantive legal standard governing . . . the payment for services,” §1395hh(a)(2). The court of appeals ultimately sided with the hospitals. Held: Because the government has not identified a lawful excuse for neglecting its statutory notice-and-comment obligations, its policy must be vacated. Pp. 5–17. (a) This case turns on whether the government’s 2014 announcement established or changed a “substantive legal standard.” The government suggests the statute means to distinguish a substantive from an interpretive legal standard and thus tracks the Administrative Procedure Act (APA), under which “substantive rules” have the “force and effect of law,” while “interpretive rules” merely “advise the public of the agency’s construction of the statutes and rules which it administers,” Perez v. Mortgage Bankers Assn., 575 U.S. 92, ___. Because the policy of counting Part C patients in the Medicare fractions would be treated as interpretive rather than substantive under the APA, the government submits, it had no statutory obligation to provide notice and comment before adopting the policy. The government’s interpretation is incorrect because the Medicare Act and the APA do not use the word “substantive” in the same way. First, the Medicare Act contemplates that “statements of policy” can establish or change a “substantive legal standard,” §1395hh(a)(2), while APA statements of policy are not substantive by definition but are grouped with and treated as interpretive rules, 5 U. S. C. §553(b)(A). Second, §1395hh(e)(1)—which gives the government limited authority to make retroactive “substantive change[s]” in, among other things, “interpretative rules” and “statements of policy”—would make no sense if the Medicare Act used the term “substantive” as the APA does, because interpretive rules and statements of policy—and any changes to them—are not substantive under the APA by definition. Third, had Congress wanted to follow the APA in the Medicare Act and exempt interpretive rules and policy statements from notice and comment, it could have simply cross-referenced the exemption in §553(b)(A) of the APA. And the fact that Congress did cross-reference the APA’s neighboring good cause exemption found in §553(b)(B), see §1395hh(b)(2)(C), strongly suggests that it “act[ed] intentionally and purposefully in the disparate” decisions, Russello v. United States, 464 U.S. 16, 23. Pp. 5–12. (b) The Medicare Act’s text and structure foreclose the government’s position in this case, and the legislative history presented by the government is ambiguous at best. The government also advances a policy argument: Requiring notice and comment for Medicare interpretive rules would be excessively burdensome. But courts are not free to rewrite clear statutes under the banner of their own policy concerns, and the government’s argument carries little force even on its own terms. Pp. 13–16. (c) Because this Court affirms the court of appeals’ judgment under §1395hh(a)(2), there is no need to address that court’s alternative holding that §1395hh(a)(4) independently required notice and comment. Nor does this Court consider the argument, not pursued by the government here, that the policy did not “establis[h] or chang[e]” a substantive legal standard—and so did not require notice and comment under §1395hh(a)(2)—because the statute itself required the government to count Part C patients in the Medicare fraction. Pp. 16–17. 863 F.3d 937, affirmed. Gorsuch, J., delivered the opinion of the Court, in which Roberts, C. J., and Thomas, Ginsburg, Alito, Sotomayor, and Kagan, JJ., joined. Breyer, J., filed a dissenting opinion. Kavanaugh, J., took no part in the consideration or decision of the case.

One way or another, Medicare touches the lives of nearly all Americans. Recognizing this reality, Congress has told the government that, when it wishes to establish or change a “substantive legal standard” affecting Medicare benefits, it must first afford the public notice and a chance to comment. 42 U. S. C. §1395hh(a)(2). In 2014, the government revealed a new policy on its website that dramatically—and retroactively—reduced payments to hospitals serving low-income patients. Because affected members of the public received no advance warning and no chance to comment first, and because the government has not identified a lawful excuse for neglecting its statutory notice-and-comment obligations, we agree with the court of appeals that the new policy cannot stand. I Today, Medicare stands as the largest federal program after Social Security. It spends about $700 billion annually to provide health insurance for nearly 60 million aged or disabled Americans, nearly one-fifth of the Nation’s population. Needless to say, even seemingly modest modifications to the program can affect the lives of millions. As Medicare has grown, so has Congress’s interest in ensuring that the public has a chance to be heard before changes are made to its administration. As originally enacted in 1965, the Medicare Act didn’t address the possibility of public input. Nor did the notice-and-comment procedures of the Administrative Procedure Act apply. While the APA requires many other agencies to offer public notice and a comment period before adopting new regulations, it does not apply to public benefit programs like Medicare. 5 U. S. C. §553(a)(2). Soon enough, though, the government volunteered to follow the informal notice-and-comment rulemaking procedures found in the APA when proceeding under the Medicare Act. See Clarian Health West, LLC v. Hargan, 878 F.3d 346, 356–357 (CADC 2017). This solution came under stress in the 1980s. By then, Medicare had grown exponentially and the burdens and benefits of public comment had come under new scrutiny. The government now took the view that following the APA’s procedures had become too troublesome and proposed to relax its commitment to them. See 47 Fed. Reg. 26860–26861 (1982). But Congress formed a different judgment. It decided that, with the growing scope of Medicare, notice and comment should become a matter not merely of administrative grace, but of statutory duty. See §9321(e)(1), 100Stat. 2017; §4035(b), 101Stat. 1330–78. Notably, Congress didn’t just adopt the APA’s notice-and-comment regime for the Medicare program. That, of course, it could have easily accomplished in just a few words. Instead, Congress chose to write a new, Medicare-specific statute. The new statute required the government to provide public notice and a 60-day comment period (twice the APA minimum of 30 days) for any “rule, requirement, or other statement of policy (other than a national coverage determination) that establishes or changes a substantive legal standard governing the scope of benefits, the payment for services, or the eligibility of individuals, entities, or organizations to furnish or receive services or benefits under [Medicare].” 42 U. S. C. §1395hh(a)(2). Our case involves a dispute over this language. Since Medicare’s creation and under what’s called “Medicare Part A,” the federal government has paid hospitals directly for providing covered patient care. To ensure hospitals have the resources and incentive to serve low-income patients, the government has also long offered additional payments to institutions that serve a “disproportionate number” of such persons. §1395ww(d)(5)(F)(i)(I). These payments are calculated in part using a hospital’s so-called “Medicare fraction,” which asks how much of the care the hospital provided to Medicare patients in a given year was provided to low-income Medicare patients. The fraction’s denominator is the time the hospital spent caring for patients who were “entitled to benefits under” Medicare Part A. The numerator is the time the hospital spent caring for Part-A-entitled patients who were also entitled to income support payments under the Social Security Act. §1395ww(d)(5)(F)(vi)(I). The bigger the fraction, the bigger the payment. Calculating Medicare fractions got more complicated in 1997. That year, Congress created “Medicare Part C,” sometimes referred to as Medicare Advantage. Under Part C, beneficiaries may choose to have the government pay their private insurance premiums rather than pay for their hospital care directly. This development led to the question whether Part C patients should be counted as “entitled to benefits under” Part A when calculating a hospital’s Medicare fraction. The question is important as a practical matter because Part C enrollees, we’re told, tend to be wealthier than patients who opt for traditional Part A coverage. Allina Health Services v. Price, 863 F.3d 937, 939 (CADC 2017). So counting them makes the fraction smaller and reduces hospitals’ payments considerably—by between $3 and $4 billion over a 9-year period, according to the government. Pet. for Cert. 23. The agency overseeing Medicare has gone back and forth on whether to count Part C participants in the Medicare fraction. At first, it did not include them. See Northeast Hospital Corp. v. Sebelius, 657 F.3d 1, 15–16 (CADC 2011). In 2003, the agency even proposed codifying that practice in a formal rule. 68 Fed. Reg. 27208. But after the public comment period, the agency reversed field and issued a final rule in 2004 declaring that it would begin counting Part C patients. 69 Fed. Reg. 49099. This abrupt change prompted various legal challenges from hospitals. In one case, a court held that the agency couldn’t apply the 2004 rule retroactively. Northeast Hospital, 657 F. 3d, at 14. In another case, a court vacated the 2004 rule because the agency had “ ‘pull[ed] a surprise switcheroo’ ” by doing the opposite of what it had proposed. Allina Health Services v. Sebelius, 746 F.3d 1102, 1108 (CADC 2014). Eventually, and in response to these developments, the agency in 2013 issued a new rule that prospectively “readopt[ed] the policy” of counting Part C patients. 78 Fed. Reg. 50620. Challenges to the 2013 rule are pending. The case before us arose in 2014. That’s when the agency got around to calculating hospitals’ Medicare fractions for fiscal year 2012. When it did so, the agency still wanted to count Part C patients. But it couldn’t rely on the 2004 rule, which had been vacated. And it couldn’t rely on the 2013 rule, which bore only prospective effect. The agency’s solution? It posted on a website a spreadsheet announcing the 2012 Medicare fractions for 3,500 hospitals nationwide and noting that the fractions included Part C patients. That Internet posting led to this lawsuit. A group of hospitals who provided care to low-income Medicare patients in 2012 argued (among other things) that the government had violated the Medicare Act by skipping its statutory notice-and-comment obligations. In reply, the government admitted that it hadn’t provided notice and comment but argued it wasn’t required to do so in these circumstances. Ultimately, the court of appeals sided with the hospitals. 863 F. 3d, at 938. But in doing so the court created a conflict with other circuits that had suggested, if only in passing, that notice and comment wasn’t needed in cases like this. See, e.g., Via Christi Regional Medical Center, Inc. v. Leavitt, 509 F.3d 1259, 1271, n. 11 (CA10 2007); Baptist Health v. Thompson, 458 F.3d 768, 776, n. 8 (CA8 2006). We granted the government’s petition for certiorari to resolve the conflict. 585 U. S. ___ (2018). II This case hinges on the meaning of a single phrase in the notice-and-comment statute Congress drafted specially for Medicare in 1987. Recall that the law requires the government to provide the public with advance notice and a chance to comment on any “rule, requirement, or other statement of policy” that “establishes or changes a substantive legal standard governing . . . the payment for services.” §1395hh(a)(2). Before us, everyone agrees that the government’s 2014 announcement of the 2012 Medicare fractions governed “payment for services.” It’s clear, too, that the government’s announcement was at least a “statement of policy” because it “le[t] the public know [the agency’s] current . . . adjudicatory approach” to a critical question involved in calculating payments for thousands of hospitals nationwide. Syncor Int’l Corp. v. Shalala, 127 F.3d 90, 94 (CADC 1997). So whether the government had an obligation to provide notice and comment winds up turning on whether its 2014 announcement established or changed a “substantive legal standard.” That phrase doesn’t seem to appear anywhere else in the entire United States Code, and the parties offer at least two ways to read it. The hospitals suggest the statute means to distinguish a substantive from a procedural legal standard. On this account, a substantive standard is one that “creates duties, rights and obligations,” while a procedural standard specifies how those duties, rights, and obligations should be enforced. Black’s Law Dictionary 1281 (5th ed. 1979) (defining “substantive law”). And everyone agrees that a policy of counting Part C patients in the Medicare fraction is substantive in this sense, because it affects a hospital’s right to payment. From this it follows that the public had a right to notice and comment before the government could adopt the policy at hand. 863 F. 3d, at 943. Very differently, the government suggests the statute means to distinguish a substantive from an interpretive legal standard. Under the APA, “substantive rules” are those that have the “force and effect of law,” while “interpretive rules” are those that merely “ ‘advise the public of the agency’s construction of the statutes and rules which it administers.’ ” Perez v. Mortgage Bankers Assn., 575 U.S. 92, ___–___ (2015) (slip op., at 2–3). On the government’s view, the 1987 Medicare notice-and-comment statute meant to track the APA’s usage in this respect. And the government submits that, because the policy of counting Part C patients in the Medicare fractions would be treated as interpretive rather than substantive under the APA, it had no statutory obligation to provide notice and comment before adopting its new policy. Who has the better reading? Several statutory clues persuade us of at least one thing: The government’s interpretation can’t be right. Pretty clearly, the Medicare Act doesn’t use the word “substantive” in the same way the APA does—to identify only those legal standards that have the “force and effect of law.” First, the Medicare Act contemplates that “statements of policy” like the one at issue here can establish or change a “substantive legal standard.” 42 U. S. C. §1395hh(a)(2) (emphasis added). Yet, by definition under the APA, statements of policy are not substantive; instead they are grouped with and treated as interpretive rules. 5 U. S. C. §553(b)(A). This strongly suggests the Medicare Act just isn’t using the word “substantive” in the same way as the APA. Even the government acknowledges that its contrary reading leaves the Medicare Act’s treatment of policy statements “incoherent.” Tr. of Oral Arg. 19. To be sure, the government suggests that the statutory incoherence produced by its reading turns out to serve a rational purpose: It clarifies that the agency overseeing Medicare can’t evade its notice-and-comment obligations for new rules that bear the “force and effect” of law by the simple expedient of “call[ing]” them mere “statements of policy.” Id., at 19–20. The dissent echoes this argument, suggesting that Congress included “statements of policy” in §1395hh(a)(2) in order to capture “substantive rules in disguise.” Post, at 5 (opinion of Breyer, J.). But the statute doesn’t refer to things that are labeled or disguised as statements of policy; it just refers to “statements of policy.” Everyone agrees that when Congress used that phrase in the APA and in other provisions of §1395hh, it referred to things that really are statements of policy. See, e.g., Pacific Gas & Elec. Co. v. Federal Power Comm’n, 506 F.2d 33, 38 (CADC 1974); post, at 4–5 (discussing §1395hh(e)(1)). Yet, to accept the government’s view, we’d have to hold that when Congress used the very same phrase in §1395hh(a)(2), it sought to refer to things an agency calls statements of policy but that in fact are nothing of the sort. The dissent admits this “may seem odd at first blush,” post, at 5, but further blushes don’t bring much improvement. This Court does not lightly assume that Congress silently attaches different meanings to the same term in the same or related statutes. See Law v. Siegel, 571 U.S. 415, 422 (2014). Besides, even if the statute’s reference to “statements of policy” could bear such an odd construction, the government and the dissent fail to explain why Congress would have thought it necessary or appropriate. Agencies have never been able to avoid notice and comment simply by mislabeling their substantive pronouncements. On the contrary, courts have long looked to the contents of the agency’s action, not the agency’s self-serving label, when deciding whether statutory notice-and-comment demands apply. See, e.g., General Motors Corp. v. Ruckelshaus, 742 F.2d 1561, 1565 (CADC 1984) (en banc) (“[T]he agency’s own label, while relevant, is not dispositive”); Guardian Fed. Sav. & Loan Assn. v. Federal Sav. & Loan Ins. Corp., 589 F.2d 658, 666–667 (CADC 1978) (if “a so-called policy statement is in purpose or likely effect . . . a binding rule of substantive law,” it “will be taken for what it is”). Nor is there any evidence before us suggesting that Congress thought it important to underscore this prosaic point in the Medicare Act (and yet not in the APA)—let alone any reason to think Congress would have sought to make the point in such an admittedly incoherent way. Second, the government’s reading would introduce another incoherence into the Medicare statute. Subsection (e)(1) of §1395hh gives the government limited authority to make retroactive “substantive change[s]” in, among other things, “interpretative rules” and “statements of policy.” But this statutory authority would make no sense if the Medicare Act used the term “substantive” as the APA does. It wouldn’t because, again, interpretive rules and statements of policy—and any changes to them—are not substantive under the APA by definition. Here, too, the government offers no satisfactory reply. It concedes, as it must, that the term “substantive” in subsection (e)(1) can’t carry the meaning it wishes to ascribe to the same word in subsection (a)(2). Tr. of Oral Arg. 16–18. So that leaves the government to suggest (again) that the same word should mean two different things in the same statute. In (e)(1), the government says, it may bear the meaning the hospitals propose, but in (a)(2) it means the same thing it does in the APA. But, once more, the government fails to offer any good reason or evidence to unseat our normal presumption that, when Congress uses a term in multiple places within a single statute, the term bears a consistent meaning throughout. See Law, 571 U. S., at 422. Third, the government suggests Congress used the phrase “substantive legal standard” in the Medicare Act as a way to exempt interpretive rules and policy statements from notice and comment. But Congress had before it—and rejected—a much more direct path to that destination. In a single sentence the APA sets forth two exemp- tions from the government’s usual notice-and-comment obligations: “Except when notice or hearing is required by statute, this subsection [requiring notice and comment] does not apply— “(A) to interpretative rules, general statements of policy, or rules of agency organization, procedure, or practice; or “(B) when the agency for good cause finds . . . that notice and public procedure thereon are impracticable, unnecessary, or contrary to the public interest.” 5 U. S. C. §553(b). In the Medicare Act, Congress expressly borrowed one of the APA’s exemptions, the good cause exemption, by cross-referencing it in §1395hh(b)(2)(C). If, as the government supposes, Congress had also wanted to borrow the other APA exemption, for interpretive rules and policy statements, it could have easily cross-referenced that exemption in exactly the same way. Congress had recently done just that, cross-referencing both of the APA’s exceptions in the Clean Air Act. See §305(a), 91Stat. 772, 42 U. S. C. §7607(d)(1). Yet it didn’t do the same thing in the Medicare Act, and Congress’s choice to include a cross-reference to one but not the other of the APA’s neighboring exemptions strongly suggests it acted “ ‘intentionally and purposefully in the disparate’ ” decisions. Russello v. United States, 464 U.S. 16, 23 (1983). The government’s response asks us to favor a most unlikely reading over this obvious one. The government submits that Congress simply preferred to mimic the APA’s interpretive-rule exemption in the Medicare Act by using the novel and enigmatic phrase “substantive legal standard” instead of a simple cross-reference. But the government supplies no persuasive account why Congress would have thought it necessary or wise to proceed in this convoluted way. The dissent suggests that a cross-reference could not have taken the place of other language in §1395hh(a)(2) limiting the notice-and-comment requirement to rules governing benefits, payment, or eligibility, post, at 17; but we can’t see why this would have made a cross-reference less desirable than the phrase “substantive legal standard” as a means of incorporating the APA’s interpretive-rule exemption. So we’re left with nothing but the doubtful proposition that Congress sought to accomplish in a “surpassingly strange manner” what it could have accomplished in a much more straightforward way. RadLAX Gateway Hotel, LLC v. Amalgamated Bank, 566 U.S. 639, 647 (2012); see Advocate Health Care Network v. Stapleton, 581 U. S. ___, ___ (2017) (slip op., at 8) (“When legislators did not adopt ‘obvious alternative’ language, ‘the natural implication is that they did not intend’ the alternative”). The dissent would have us disregard all of the textual clues we’ve found significant because the word “substantive” carried “a special meaning in the context of administrative law” in the 1980s, making it “almost a certainty” that Congress had that meaning in mind when it used the word “substantive” in §1395hh(a)(2). Post, at 3, 8. But it was the phrase “substantive rule” that was a term of art in administrative law, and Congress chose not to use that term in the Medicare Act. Instead, it introduced a seemingly new phrase to the statute books when it spoke of “substantive legal standards.” And, for all the reasons we have already explored, the term “substantive legal standard” in the Medicare Act appears to carry a more expansive scope than that borne by the term “substantive rule” under the APA. In reply, the dissent stresses that §1395hh refers to agency actions requiring notice and comment as “regulations.” This is significant, the dissent says, because “courts had sometimes treated [the term ‘regulations’] as interchangeable with the term ‘substantive rules’ ” around the time of the 1987 Medicare Act amendments. Post, at 4. So if only “regulations” must proceed through notice and comment, the dissent reasons, that necessarily encompasses only things that qualify as substantive rules under the APA. In fact, however, by 1987 courts had commonly referred to both substantive and interpretive rules as “regulations,” so the dissent’s logical syllogism fails on its own terms. To see this, one need look no further than Chrysler Corp. v. Brown, 441 U.S. 281 (1979), which described the substantive-interpretive divide as “the central distinction among agency regulations found in the APA.” Id., at 301 (emphasis added); see also, e.g., Batterton v. Francis, 432 U.S. 416, 425, n. 9 (1977) (distinguishing between “[l]egislative, or substantive, regulations” and “interpretative regulation[s]”); United Technologies Corp. v. EPA, 821 F.2d 714, 719 (CADC 1987) (“most of the regulations at issue are . . . interpretative”).[1] In the end, all of the available evidence persuades us that the phrase “substantive legal standard,” which appears in §13955hh(a)(2) and apparently nowhere else in the U. S. Code, cannot bear the same construction as the term “substantive rule” in the APA. We need not, however, go so far as to say that the hospitals’ interpretation, adopted by the court of appeals, is correct in every particular. To affirm the judgment before us, it is enough to say the government’s arguments for reversal fail to withstand scrutiny. Other questions about the statute’s meaning can await other cases. The dissent would like us to provide more guidance, post, at 13–14, but the briefing before us focused on the issue whether the Medicare Act borrows the APA’s interpretive-rule exception, and we limit our holding accordingly. In doing so, we follow the well-worn path of declining “to issue a sweeping ruling when a narrow one will do.” McWilliams v. Dunn, 582 U. S. ___, ___ (2017) (slip op., at 14).[2] III Unable to muster support for its position in the statu- tory text or structure, the government encourages us to look elsewhere. It begins by inviting us to follow it into the legislative history lurking behind the Medicare Act. “But legislative history is not the law.” Epic Systems Corp. v. Lewis, 584 U. S. ___, ___ (2018) (slip op., at 23). And even those of us who believe that clear legislative history can “illuminate ambiguous text” won’t allow “ambiguous legislative history to muddy clear statutory language.” Milner v. Department of Navy, 562 U.S. 562, 572 (2011). Yet the text before us clearly forecloses the government’s position in this case, and the legislative history presented to us is ambiguous at best. The government points us first to a conference report on the 1986 bill that adopted §1395hh(b). The 1986 report opined that the bill adopted at that time wouldn’t require notice and comment for interpretive rules. See H. R. Conf. Rep. No. 99–1012, p. 311 (1986). But the 1986 bill didn’t include the statutory language at issue here. Congress added that language only the following year, when it enacted §1395hh(a)(2). Nor does the government try to explain how a report on a 1986 bill sheds light on the meaning of statutory terms first introduced in 1987. If anything, the fact that Congress revisited the statute in 1987 may suggest it wasn’t satisfied with the 1986 notice-and-comment requirements and wished to enhance them. Some legislative history even says as much. See H. R. Rep. No. 100–391(I), p. 430 (1987) (expressing concern that, despite the 1986 legislation, the agency was still announcing “important policies” without notice and comment). The conference report on the 1987 bill that did adopt the statutory language before us today doesn’t offer much help to the government either. The House version of the bill would have required notice and comment for rules with a “significant effect” on payments, a condition no doubt present here. H. R. 3545, 100th Cong., 1st Sess., reprinted in 133 Cong. Rec. 30019. Later, the conference committee replaced the House’s language with the current language of subsection (a)(2), which the report said “reflect[ed] recent court rulings.” H. R. Conf. Rep. No. 100–495, p. 566 (1987). The government contends that this was an oblique reference to a then-recent decision discussing the APA’s interpretive-rule exception and an implicit suggestion that interpretive rules shouldn’t be subject to notice and comment. See American Hospital Assn. v. Bowen, 834 F.2d 1037, 1045–1046 (CADC 1987). But, as the hospitals point out, Bowen was mostly about the APA’s treatment of procedural rules. See id., at 1047–1057. So it seems at least equally plausible that the conference committee revised the House’s language because it feared that language would have subjected procedural rules to notice-and-comment obligations. The hospitals call our attention to other indications, too, that Members of Congress didn’t understand the conference’s language to track the APA. For example, the relevant provision in the final bill was titled “Publication as Regulations of Significant Policies.” §4035(b), 101Stat. 1330–78 (emphasis added). And, as we’ve seen, “significant policies” don’t always amount to substantive rules under the APA. The House Ways and Means Committee likewise described the final bill as requiring notice and comment for “[s]ignificant policy changes,” not just substantive rules. Summary of Conference Agreement on Reconciliation Provisions Within the Jurisdiction of the Committee on Ways and Means, 100th Cong., 1st Sess., 12–13 (Comm. Print 1987). So in the end and at most, we are left with exactly the kind of murky legislative history that we all agree can’t overcome a statute’s clear text and structure. That leads us to the government’s final redoubt: a policy argument. But as the government knows well, courts aren’t free to rewrite clear statutes under the banner of our own policy concerns. If the government doesn’t like Congress’s notice-and-comment policy choices, it must take its complaints there. See, e.g., Henson v. Santander Consumer USA Inc., 582 U. S. ___, ___–___ (2017) (slip op., at 9–10); Sebelius v. Cloer, 569 U.S. 369, 381 (2013). Besides, the government’s policy arguments don’t carry much force even on their own terms. The government warns that providing the public with notice and a chance to comment on all Medicare interpretive rules, like those in its roughly 6,000-page “Provider Reimbursement Manual,” would take “ ‘many years’ ” to complete. Brief for Petitioner 18, 42. But the dissent points to only eight manual provisions that courts have deemed interpretive over the last four decades, see post, at 10–12, and the government hasn’t suggested that providing notice and comment for these or any other specific manual provisions would prove excessively burdensome. Nor has the government identified any court decision invalidating a manual provision under §1395hh(a)(2) in the nearly two years since the court of appeals issued its opinion in this case. For their part, the hospitals claim that only a few dozen pages of the Provider Reimbursement Manual might even arguably require notice and comment. Tr. of Oral Arg. 49–51. And they tell us that the agency regularly and without much difficulty undertakes notice-and-comment rulemaking for many other decisions affecting the Medicare program. See Brief for Respondents 58; App. to Brief in Opposition 1a–3a. The government hasn’t rebutted any of these points. Not only has the government failed to document any draconian costs associated with notice and comment, it also has neglected to acknowledge the potential countervailing benefits. Notice and comment gives affected parties fair warning of potential changes in the law and an opportunity to be heard on those changes—and it affords the agency a chance to avoid errors and make a more informed decision. See 1 K. Hickman & R. Pierce, Administrative Law §4.8 (6th ed. 2019). Surely a rational Congress could have thought those benefits especially valuable when it comes to a program where even minor changes to the agency’s approach can impact millions of people and billions of dollars in ways that are not always easy for regulators to anticipate. None of this is to say Congress had to proceed as it did. It is only to say that Congress reasonably could have believed that the policy decision reflected in the statute would yield benefits sufficient to outweigh the speculative burdens the government has suggested. And if notice and comment really does threaten to “become a major roadblock to the implementation of” Medicare, post, at 10, the agency can seek relief from Congress, which—unlike the courts—is both qualified and constitutionally entitled to weigh the costs and benefits of different approaches and make the necessary policy judgment. IV There are two more lines of argument that deserve brief acknowledgment. One concerns §1395hh(a)(4), which provides that a Medicare regulation struck down for not being a logical outgrowth of the government’s proposal can’t “take effect” until the agency provides a “further opportunity for public comment.” The hospitals claim, and the court of appeals held, that subsection (a)(4) also and independently required notice and comment here. But given our holding affirming the court of appeals’ judgment under §1395hh(a)(2), we have no need to reach this question. Separately, we can imagine that the government might have sought to argue that the policy at issue here didn’t “establis[h] or chang[e]” a substantive legal standard—and so didn’t require notice and comment under §1395hh(a)(2)—because the statute itself required it to count Part C patients in the Medicare fraction. But we need not consider this argument either, this time because the government hasn’t pursued it and we normally have no obligation to entertain grounds for reversal that a party hasn’t presented. Far from suggesting that the Medicare Act supplies the controlling legal standard for determining whether to count Part C patients, the government has insisted that the statute “does not speak directly to the issue,” Brief for Appellant in Northeast Hospital Corp. v. Sebelius, No. 10–5163 (CADC), p. 22, and thus leaves a “ ‘gap’ ” for the agency to fill, Brief for Appellee in Allina v. Price, No. 16–5255 (CADC), p. 50 (quoting Northeast Hospital Corp., 657 F. 3d, at 13). The courts below ac- cepted the government’s submission, and the government hasn’t sought to take a different position in this Court. So we express no opinion on whether the statute in fact contains such a “gap.” We hold simply that, when the government establishes or changes an avowedly “gap”-filling policy, it can’t evade its notice-and-comment obligations under §1395hh(a)(2) on the strength of the arguments it has advanced in this case. * The judgment of the court of appeals is Affirmed. Justice Kavanaugh took no part in the consideration or decision of this case. Notes 1 Nor does §1395hh(e)(1) imply that the statute is using “regulations” and “interpretative rules” to mean different things. Post, at 4–5. True, that provision refers to “regulations, manual instructions, interpretative rules, statements of policy, or guidelines of general applicability.” But contrary to the dissent’s suggestion that each item in the list “refers to something different,” post, at 5, the items appear to have substantial overlap. For example, many manual instructions surely qualify as guidelines of general applicability; and, as explained above, the statute explicitly requires some statements of policy to be issued as regulations. 2 Nor is it obvious that the dissent’s approach would provide significantly clearer guidance. Lower courts have often observed “that it is quite difficult to distinguish between substantive and interpretative rules,” Syncor Int’l Corp. v. Shalala, 127 F.3d 90, 93 (CADC 1997), and precisely where to draw the boundary has been a subject “of much scholarly and judicial debate,” Perez v. Mortgage Bankers Assn., 575 U.S. 92, ___ (2015) (slip op., at 3).

587.US.2018_17-1184

Petitioner Michael Biestek, a former construction worker, applied for social security disability benefits, claiming he could no longer work due to physical and mental disabilities. The Social Security Administration (SSA) assigned an Administrative Law Judge (ALJ) to conduct a hearing, at which the ALJ had to determine whether Biestek could successfully transition to less physically demanding work. For guidance on that issue, the ALJ heard testimony from a vocational expert regarding the types of jobs Biestek could still perform and the number of such jobs that existed in the national economy. See 20 CFR §§404.1560(c)(1), 416.960(c)(1). On cross-examination, Biestek’s attorney asked the expert “where [she was] getting [her numbers] from,” and the expert explained they were from her own individual labor market surveys. Biestek’s attorney then requested that the expert turn over the surveys. The expert declined. The ALJ ultimately denied Biestek benefits, basing his conclusion on the expert’s testimony about the number of jobs available to him. Biestek sought review in federal court, where an ALJ’s factual findings are “conclusive” if supported by “substantial evidence,” 42 U. S. C. §405(g). The District Court rejected Biestek’s argument that the expert’s testimony could not possibly constitute substantial evidence because she had declined to produce her supporting data. The Sixth Circuit affirmed. Held: A vocational expert’s refusal to provide private market-survey data upon the applicant’s request does not categorically preclude the testimony from counting as “substantial evidence.” Substantial evidence is “more than a mere scintilla,” and means only “such relevant evidence as a reasonable mind might accept as adequate to support a conclusion.” Consolidated Edison Co. v. NLRB, 305 U.S. 197, 229. Biestek proposes a categorical rule that the testimony of a vocational expert who refuses a request for supporting data about job availability can never clear that bar. To assess that proposal, the Court begins with the parties’ common ground: Assuming no demand, a vocational expert’s testimony may count as substantial evidence even when unaccompanied by supporting data. If that is true, is it not obvious why one additional fact—a refusal to a request for that data—should make an expert’s testimony categorically inadequate. In some cases, the refusal to disclose data, considered along with other shortcomings, will undercut an expert’s credibility and prevent a court from finding that “a reasonable mind” could accept the expert’s testimony. But in other cases, the refusal will have no such consequence. Similarly, the refusal will sometimes interfere with effective cross-examination, which a reviewing court may consider in deciding how much to credit an expert’s opinion. But other times, even without supporting data, an applicant will be able to probe the strength of the expert’s testimony on cross-examination. Ultimately, Biestek’s error lies in his pressing for a categorical rule, applying to every case in which a vocational expert refuses a request for underlying data. The inquiry, as is usually true in determining the substantiality of evidence, is case-by-case. It takes into account all features of the vocational expert’s testimony, as well as the rest of the administrative record, and defers to the presiding ALJ, who has seen the hearing up close. Pp. 5–11. 880 F.3d 778, affirmed. Kagan, J., delivered the opinion of the Court, in which Roberts, C. J., and Thomas, Breyer, Alito, and Kavanaugh, JJ., joined. Sotomayor, J., filed a dissenting opinion. Gorsuch, J., filed a dissenting opinion, in which Ginsburg, J., joined.

The Social Security Administration (SSA) provides benefits to individuals who cannot obtain work because of a physical or mental disability. To determine whether an applicant is entitled to benefits, the agency may hold an informal hearing examining (among other things) the kind and number of jobs available for someone with the applicant’s disability and other characteristics. The agency’s factual findings on that score are “conclusive” in judicial review of the benefits decision so long as they are sup- ported by “substantial evidence.” 42 U. S. C. §405(g). This case arises from the SSA’s reliance on an expert’s testimony about the availability of certain jobs in the economy. The expert largely based her opinion on private market-survey data. The question presented is whether her refusal to provide that data upon the applicant’s request categorically precludes her testimony from counting as “substantial evidence.” We hold it does not. I Petitioner Michael Biestek once worked as a carpenter and general laborer on construction sites. But he stopped working after he developed degenerative disc disease, Hepatitis C, and depression. He then applied for social security disability benefits, claiming eligibility as of October 2009. After some preliminary proceedings, the SSA assigned an Administrative Law Judge (ALJ) to hold a hearing on Biestek’s application. Those hearings, as described in the Social Security Act, 49Stat. 620, as amended, 42 U. S. C. §301 et seq., are recognizably adjudicative in nature. The ALJ may “receive evidence” and “examine witnesses” about the contested issues in a case. §§405(b)(1), 1383(c) (1)(A). But many of the rules governing such hear- ings are less rigid than those a court would follow. See Richardson v. Perales, 402 U.S. 389, 400–401 (1971). An ALJ is to conduct a disability hearing in “an informal, non-adversarial manner.” 20 CFR §404.900(b) (2018); §416.1400(b). Most notably, an ALJ may receive evidence in a disability hearing that “would not be admissible in court.” §§404.950(c), 416.1450(c); see 42 U. S. C. §§405(b) (1), 1383(c)(1)(A). To rule on Biestek’s application, the ALJ had to determine whether the former construction laborer could successfully transition to less physically demanding work. That required exploring two issues. The ALJ needed to identify the types of jobs Biestek could perform notwithstanding his disabilities. See 20 CFR §§404.1560(c)(1), 416.960(c)(1). And the ALJ needed to ascertain whether those kinds of jobs “exist[ed] in significant numbers in the national economy.” §§404.1560(c)(1), 416.960(c)(1); see §§404.1566, 416.966. For guidance on such questions, ALJs often seek the views of “vocational experts.” See §§404.1566(e), 416.966(e); SSA, Hearings, Appeals, and Litigation Law Manual I–2–5–50 (Aug. 29, 2014). Those experts are professionals under contract with SSA to provide impartial testimony in agency proceedings. See id., at I–2–1–31.B.1 (June 16, 2016); id., at I–2–5–48. They must have “expertise” and “current knowledge” of “[w]orking conditions and physical demands of various” jobs; “[k]nowledge of the existence and numbers of [those jobs] in the national economy”; and “[i]nvolvement in or knowledge of placing adult workers[ ] with disabilities[ ] into jobs.” Id., at I–2–1–31.B.1. Many vocational experts simultaneously work in the private sector locating employment for persons with disabilities. See C. Kubitschek & J. Dubin, Social Security Disability Law & Procedure in Federal Court §3:89 (2019). When offering testimony, the experts may invoke not only publicly available sources but also “information obtained directly from employers” and data otherwise developed from their own “experience in job placement or career counseling.” Social Security Ruling, SSR 00–4p, 65 Fed. Reg. 75760 (2000). At Biestek’s hearing, the ALJ asked a vocational expert named Erin O’Callaghan to identify a sampling of “sedentary” jobs that a person with Biestek’s disabilities, education, and job history could perform. Tr. 59 (July 21, 2015); see 20 CFR §§404.1567(a), 416.967(a) (defining a “sedentary” job as one that “involves sitting” and requires “lifting no more than 10 pounds”). O’Callaghan had served as a vocational expert in SSA proceedings for five years; she also had more than ten years’ experience counseling people with disabilities about employment opportunities. See Stachowiak v. Commissioner of Social Security, 2013 WL 593825, *1 (ED Mich., Jan. 11, 2013); Record in No. 16–10422 (ED Mich.), Doc. 17–13, p. 1274 (resume). In response to the ALJ’s query, O’Callaghan listed sedentary jobs “such as a bench assembler [or] sorter” that did not require many skills. Tr. 58–59. And she further testified that 240,000 bench assembler jobs and 120,000 sorter jobs existed in the national economy. See ibid. On cross-examination, Biestek’s attorney asked O’Callaghan “where [she was] getting those [numbers] from.” Id., at 71. O’Callaghan replied that they came from the Bureau of Labor Statistics and her “own individual labor market surveys.” Ibid. The lawyer then requested that O’Callaghan turn over the private surveys so he could review them. Ibid. O’Callaghan responded that she wished to keep the surveys confidential because they were “part of [her] client files.” Id., at 72. The lawyer suggested that O’Callaghan could “take the clients’ names out.” Ibid. But at that point the ALJ interjected that he “would not require” O’Callaghan to produce the files in any form. Ibid. Biestek’s counsel asked no further questions about the basis for O’Callaghan’s assembler and sorter numbers. After the hearing concluded, the ALJ issued a decision granting Biestek’s application in part and denying it in part. According to the ALJ, Biestek was entitled to benefits beginning in May 2013, when his advancing age (he turned fifty that month) adversely affected his ability to find employment. See App. to Pet. for Cert. 19a, 112a–113a. But before that time, the ALJ held, Biestek’s dis- abilities should not have prevented a “successful adjustment to other work.” Id., at 110a–112a. The ALJ based that conclusion on O’Callaghan’s testimony about the availability in the economy of “sedentary unskilled occupations such as bench assembler [or] sorter.” Id., at 111a (emphasis deleted). Biestek sought review in federal court of the ALJ’s denial of benefits for the period between October 2009 and May 2013. On judicial review, an ALJ’s factual findings—such as the determination that Biestek could have found sedentary work—“shall be conclusive” if supported by “substantial evidence.” 42 U. S. C. §405(g); see supra, at 1. Biestek contended that O’Callaghan’s testimony could not possibly constitute such evidence because she had declined, upon request, to produce her supporting data. See Plaintiff’s Motion for Summary Judgment in No. 16–10422 (ED Mich.), Doc. 22, p. 23. But the District Court rejected that argument. See 2017 WL 1173775, *2 (Mar. 30, 2017). And the Court of Appeals for the Sixth Circuit affirmed. See Biestek v. Commissioner of Social Security, 880 F.3d 778 (2018). That court recognized that the Seventh Circuit had adopted the categorical rule Biestek proposed, precluding a vocational expert’s testimony from qualifying as substantial if the expert had declined an applicant’s request to provide supporting data. See id., at 790 (citing McKinnie v. Barnhart, 368 F.3d 907, 910–911 (2004)). But that rule, the Sixth Circuit observed in joining the ranks of unconvinced courts, “ha[d] not been a popular export.” 880 F. 3d, at 790 (internal quotation marks omitted). And no more is it so today. II The phrase “substantial evidence” is a “term of art” used throughout administrative law to describe how courts are to review agency factfinding. T-Mobile South, LLC v. Roswell, 574 U. S. ___, ___ (2015) (slip op., at 7). Under the substantial-evidence standard, a court looks to an existing administrative record and asks whether it contains “sufficien[t] evidence” to support the agency’s factual determinations. Consolidated Edison Co. v. NLRB, 305 U.S. 197, 229 (1938) (emphasis deleted). And whatever the meaning of “substantial” in other contexts, the threshold for such evidentiary sufficiency is not high. Substantial evidence, this Court has said, is “more than a mere scintilla.” Ibid.; see, e.g., Perales, 402 U. S., at 401 (internal quotation marks omitted). It means—and means only—“such relevant evidence as a reasonable mind might accept as adequate to support a conclusion.” Consolidated Edison, 305 U. S., at 229. See Dickinson v. Zurko, 527 U.S. 150, 153 (1999) (comparing the substantial-evidence standard to the deferential clearly-erroneous standard). Today, Biestek argues that the testimony of a vocational expert who (like O’Callaghan) refuses a request for supporting data about job availability can never clear the substantial-evidence bar. See Brief for Petitioner 21–34. As that formulation makes clear, Biestek’s proposed rule is categorical, rendering expert testimony insufficient to sustain an ALJ’s factfinding whenever such a refusal has occurred.[1] But Biestek hastens to add two caveats. The first is to clarify what the rule is not, the second to stress where its limits lie. Biestek initially takes pains—and understandably so—to distinguish his argument from a procedural claim. Reply Brief 12–14. At no stage in this litigation, Biestek says, has he ever espoused “a free-standing procedural rule under which a vocational expert would always have to produce [her underlying data] upon request.” Id., at 2. That kind of rule exists in federal court: There, an expert witness must produce all data she has considered in reaching her conclusions. See Fed. Rule Civ. Proc. 26(a)(2)(B). But as Biestek appreciates, no similar requirement applies in SSA hearings. As explained above, Congress intended those proceedings to be “informal” and provided that the “strict rules of evidence, applicable in the courtroom, are not to” apply. Perales, 402 U. S., at 400; see 42 U. S. C. §405(b)(1); supra, at 2. So Biestek does not press for a “procedural rule” governing “the means through which an evidentiary record [must be] created.” Tr. of Oral Arg. 6; Reply Brief 13. Instead, he urges a “substantive rule” for “assess[ing] the quality and quantity of [record] evidence”—which would find testimony like O’Callaghan’s inadequate, when taken alone, to support an ALJ’s factfinding. Id., at 12. And Biestek also emphasizes a limitation within that proposed rule. For the rule to kick in, the applicant must make a demand for the expert’s supporting data. See Brief for Petitioner i, 5, 18, 40, 55; Tr. of Oral Arg. 25–26. Consider two cases in which vocational experts rely on, but do not produce, nonpublic information. In the first, the applicant asks for the data; in the second, not. According to Biestek, the expert’s testimony in the first case cannot possibly clear the substantial-evidence bar; but in the second case, it may well do so, even though the administrative record is otherwise the same. And Biestek underscores that this difference in outcome has nothing to do with waiver or forfeiture: As he acknowledges, an applicant “cannot waive the substantial evidence standard.” Id., at 27. It is just that the evidentiary problem arises from the expert’s refusal of a demand, not from the data’s absence alone. In his words, the testimony “can constitute substantial evidence if unchallenged, but not if challenged.” Reply Brief 18. To assess Biestek’s proposal, we begin with the parties’ common ground: Assuming no demand, a vocational expert’s testimony may count as substantial evidence even when unaccompanied by supporting data. Take an example. Suppose an expert has top-of-the-line credentials, including professional qualifications and many years’ experience; suppose, too, she has a history of giving sound testimony about job availability in similar cases (perhaps before the same ALJ). Now say that she testifies about the approximate number of various sedentary jobs an applicant for benefits could perform. She explains that she arrived at her figures by surveying a range of representative employers; amassing specific information about their labor needs and employment of people with disabilities; and extrapolating those findings to the national economy by means of a well-accepted methodology. She answers cogently and thoroughly all questions put to her by the ALJ and the applicant’s lawyer. And nothing in the rest of the record conflicts with anything she says. But she never produces her survey data. Still, her testimony would be the kind of evidence—far “more than a mere scintilla”—that “a reasonable mind might accept as adequate to support” a finding about job availability. Consolidated Edison, 305 U. S., at 229. Of course, the testimony would be even better—more reliable and probative—if she had produced supporting data; that would be a best practice for the SSA and its experts.[2] And of course, a different (maybe less qualified) expert failing to produce such data might offer testimony that is so feeble, or contradicted, that it would fail to clear the substantial-evidence bar. The point is only—as, again, Biestek accepts—that expert testimony can sometimes surmount that bar absent underlying data. But if that is true, why should one additional fact—a refusal to a request for that data—make a vocational expert’s testimony categorically inadequate? Assume that an applicant challenges our hypothetical expert to turn over her supporting data; and assume the expert declines because the data reveals private information about her clients and making careful redactions will take a fair bit of time. Nothing in the expert’s refusal changes her testimony (as described above) about job availability. Nor does it alter any other material in the record. So if our expert’s opinion was sufficient—i.e., qualified as substantial evidence—before the refusal, it is hard to see why the opinion has to be insufficient afterward. Biestek suggests two reasons for that non-obvious result. First, he contends that the expert’s rejection of a request for backup data necessarily “cast[s her testimony] into doubt.” Reply Brief 16. And second, he avers that the refusal inevitably “deprives an applicant of the material necessary for an effective cross-examination.” Id., at 2. But Biestek states his arguments too broadly—and the nuggets of truth they contain cannot justify his proposed flat rule. Consider Biestek’s claim about how an expert’s refusal undercuts her credibility. Biestek here invokes the established idea of an “adverse inference”: If an expert declines to back up her testimony with information in her control, then the factfinder has a reason to think she is hiding something. See id., at 16 (citing cases). We do not dispute that possibility—but the inference is far from always required. If an ALJ has no other reason to trust the expert, or finds her testimony iffy on its face, her refusal of the applicant’s demand for supporting data may properly tip the scales against her opinion. (Indeed, more can be said: Even if the applicant makes no demand, such an expert’s withholding of data may count against her.) But if (as in our prior hypothetical example, see supra, at 7–8) the ALJ views the expert and her testimony as otherwise trustworthy, and thinks she has good reason to keep her data private, her rejection of an applicant’s demand need not make a difference. So too when a court reviews the ALJ’s decision under the deferential substantial-evidence standard. In some cases, the refusal to disclose data, considered along with other shortcomings, will prevent a court from finding that “a reasonable mind” could accept the expert’s testimony. Consolidated Edison, 305 U. S., at 229. But in other cases, that refusal will have no such consequence. Even taking it into account, the expert’s opinion will qualify as “more than a mere scintilla” of evidence supporting the ALJ’s conclusion. Which is to say it will count, contra Biestek, as substantial. And much the same is true of Biestek’s claim that an expert’s refusal precludes meaningful cross-examination. We agree with Biestek that an ALJ and reviewing court may properly consider obstacles to such questioning when deciding how much to credit an expert’s opinion. See Perales, 402 U. S., at 402–406. But Biestek goes too far in suggesting that the refusal to provide supporting data always interferes with effective cross-examination, or that the absence of such testing always requires treating an opinion as unreliable. Even without specific data, an applicant may probe the strength of testimony by asking an expert about (for example) her sources and methods—where she got the information at issue and how she analyzed it and derived her conclusions. See, e.g., Chavez v. Berryhill, 895 F.3d 962, 969–970 (CA7 2018). And even without significant testing, a factfinder may conclude that testimony has sufficient indicia of reliability to support a conclusion about whether an applicant could find work. Indeed, Biestek effectively concedes both those points in cases where supporting data is missing, so long as an expert has not refused an applicant’s demand. See supra, at 7. But once that much is acknowledged, Biestek’s argument cannot hold. For with or without an express refusal, the absence of data places the selfsame limits on cross-examination. Where Biestek goes wrong, at bottom, is in pressing for a categorical rule, applying to every case in which a vocational expert refuses a request for underlying data. Sometimes an expert’s withholding of such data, when combined with other aspects of the record, will prevent her testimony from qualifying as substantial evidence. That would be so, for example, if the expert has no good reason to keep the data private and her testimony lacks other markers of reliability. But sometimes the reservation of data will have no such effect. Even though the applicant might wish for the data, the expert’s testimony still will clear (even handily so) the more-than-a-mere-scintilla threshold. The inquiry, as is usually true in determining the substantiality of evidence, is case-by-case. See, e.g., Perales, 402 U. S., at 399, 410 (rejecting a categorical rule pertaining to the substantiality of medical reports in a disability hearing). It takes into account all features of the vocational expert’s testimony, as well as the rest of the administrative record. And in so doing, it defers to the presiding ALJ, who has seen the hearing up close. That much is sufficient to decide this case. Biestek petitioned us only to adopt the categorical rule we have now rejected. He did not ask us to decide whether, in the absence of that rule, substantial evidence supported the ALJ in denying him benefits. Accordingly, we affirm the Court of Appeals’ judgment. It is so ordered. Notes 1 In contrast, the principal dissent cannot decide whether it favors such a categorical rule. At first, Justice Gorsuch endorses the rule Biestek and the Seventh Circuit have proposed. See post, at 2. But in then addressing our opinion, he takes little or no issue with the reasoning we offer to show why that rule is too broad. See post, at 4–7. So the dissent tries to narrow the scope of Biestek’s categorical rule—to only cases that look just like his. See post, at 7–8. And still more, it shelves all the “categorical” talk and concentrates on Biestek’s case alone. See post, at 1, 4–8. There, Justice Gorsuch’s dissent joins Justice Sotomayor’s in concluding that the expert evidence in this case was insubstantial. But as we later explain, see infra, at 11, Biestek did not petition us to resolve that factbound question; nor did his briefing and argument focus on anything other than the Seventh Circuit’s categorical rule. We confine our opinion accordingly. 2 The SSA itself appears to agree. In the handbook given to voca-tional experts, the agency states: “You should have available, atthe hearing, any vocational resource materials that you are likelyto rely upon” because “the ALJ may ask you to provide relevantportions of [those] materials.” SSA, Vocational Expert Handbook 37(Aug. 2017), https://www.ssa.gov/appeals/public_experts/Vocational_Experts_(VE)_Handbook-508.pdf (as last visited Mar. 28, 2019).

586.US.2018_17-1042

Respondent Michael Loos sued petitioner BNSF Railway Company under the Federal Employers’ Liability Act (FELA) for injuries he received while working at BNSF’s railyard. A jury awarded him $126,212.78, ascribing $30,000 of that amount to wages lost during the time Loos was unable to work. BNSF asserted that the lost wages constituted “compensation” taxable under the Railroad Retirement Tax Act (RRTA) and asked to withhold $3,765 of the $30,000 to cover Loos’s share of the RRTA taxes. The District Court and the Eighth Circuit rejected the requested offset, holding that an award of damages compensating an injured railroad worker for lost wages is not taxable under the RRTA. Held: A railroad’s payment to an employee for working time lost due to an on-the-job injury is taxable “compensation” under the RRTA. Pp. 2–14. (a) In 1937, Congress created a self-sustaining retirement benefits system for railroad workers. The RRTA funds the program by imposing a payroll tax on both railroads and their employees, referring to the railroad’s contribution as an “excise” tax, 26 U. S. C. §3221, and the employee’s share as an “income” tax, §3201. The Railroad Retirement Act (RRA) entitles railroad workers to various benefits. Taxes under the RRTA and benefits under the RRA are measured by the employee’s “compensation,” which both statutes define as “any form of money remuneration paid to an individual for services rendered as an employee.” §3231(e)(1); 45 U. S. C. §231(h)(1). The statutory foundation of the railroad retirement system mirrors that of the Social Security system. The Federal Insurance Contributions Act (FICA) taxes employers and employees to fund benefits distributed pursuant to the Social Security Act (SSA). Tax and benefit amounts are determined by the worker’s “wages,” the Social Security equivalent to “compensation.” Both the FICA and the SSA define “wages” employing language resembling the RRTA and the RRA definitions of “compensation.” The term “wages” means “all remuneration” for “any service, of whatever nature, performed . . . by an employee.” 26 U. S. C. §3121(a)–(b) (FICA); see 42 U. S. C. §§409(a), 410(a) (SSA). Pp. 2–4. (b) Given the textual similarity between the definitions of “compensation” and “wages,” the decisions on the meaning of “wages” in Social Security Bd. v. Nierotko, 327 U.S. 358, and United States v. Quality Stores, Inc., 572 U.S. 141, inform this Court’s comprehension of the RRTA term “compensation.” In Nierotko, the Court held that “wages” embraced pay for active service as well as pay received for periods of absence from active service, 327 U. S., at 366, and concluded that backpay for time lost due to “the employer’s wrong” counted as “wages,” id., at 364. In Quality Stores, the Court held that severance payments qualified as “wages” taxable under the FICA. 572 U. S., at 146–147. In line with these decisions, the Court holds that “compensation” under the RRTA encompasses not simply pay for active service but also pay for periods of absence from active service—provided that the remuneration in question stems from the “employer-employee relationship.” Nierotko, 327 U. S., at 366. Damages awarded under the FELA for lost wages fit comfortably within this definition. See BNSF R. Co. v. Tyrrell, 581 U. S. ___, ___. If a railroad negligently fails to maintain a safe railyard and a worker is injured as a result, the FELA requires the railroad to compensate the injured worker for working time lost due to the employer’s wrongdoing. FELA damages for lost wages, like backpay, are “compensation” taxable under the RRTA. Pp. 4–7. (c) The Eighth Circuit construed “compensation” for RRTA purposes to mean only pay for active service, but this reading cannot be reconciled with Nierotko and Quality Stores. In addition, the RRTA’s pinpointed exclusions for certain types of payments for time lost signal that nonexcluded pay for time lost remains RRTA-taxable “compensation.” Pp. 7–10. (d) Loos contends that “compensation” does not include payments made to compensate for an injury. This reading, however, is at odds with Nierotko, which held that “wages” included backpay awarded to redress “the loss of wages” occasioned by “the employer’s wrong.” 327 U. S., at 364. Loos also argues that the exclusion of personal injury damages from “gross income” for federal income tax purposes, see 26 U. S. C. §104(a)(2), should carry over to the RRTA’s tax on the “income” of railroad workers. The RRTA, however, uses the term “income” merely to distinguish the “income” tax on an employee from the matching “excise” tax on a railroad. Further, Congress specified not “gross income” but employee “compensation” as the tax base for RRTA taxes. Congress did not exclude personal injury damages from “compensation.” Pp. 10–14. 865 F.3d 1106, reversed and remanded. Ginsburg, J., delivered the opinion of the Court, in which Roberts, C. J., and Breyer, Alito, Sotomayor, Kagan, and Kavanaugh, JJ., joined. Gorsuch, J., filed a dissenting opinion, in which Thomas, J., joined.

Respondent Michael Loos was injured while working at petitioner BNSF Railway Company’s railyard. Loos sued BNSF under the Federal Employers’ Liability Act (FELA), 35Stat. 65, as amended, 45 U. S. C. §51 et seq., and gained a $126,212.78 jury verdict. Of that amount the jury ascribed $30,000 to wages lost during the time Loos was unable to work. BNSF moved for an offset against the judgment. The lost wages awarded Loos, BNSF asserted, constituted “compensation” taxable under the Railroad Retirement Tax Act (RRTA), 26 U. S. C. §3201 et seq. Therefore, BNSF urged, the railway was required to withhold a portion of the $30,000 attributable to lost wages to cover Loos’s share of RRTA taxes, which came to $3,765. The District Court and the Court of Appeals for the Eighth Circuit rejected the requested offset, holding that an award of damages compensating an injured railroad worker for lost wages is not taxable under the RRTA. The question presented: Is a railroad’s payment to an employee for working time lost due to an on-the-job injury taxable “compensation” under the RRTA, 26 U. S. C. §3231(e)(1)? We granted review to resolve a division of opinion on the answer to that question. 584 U. S. ___ (2018). Compare Hance v. Norfolk S. R. Co., 571 F.3d 511, 523 (CA6 2009) (“compensation” includes pay for time lost); Phillips v. Chicago Central & Pacific R. Co., 853 N.W.2d 636, 650–651 (Iowa 2014) (agency reasonably interpreted “compensation” as including pay for time lost); Heckman v. Burlington N. Santa Fe R. Co., 286 Neb. 453, 463, 837 N.W.2d 532, 540 (2013) (“compensation” includes pay for time lost), with 865 F.3d 1106, 1117–1118 (CA8 2017) (case below) (“compensation” does not include pay for time lost); Mickey v. BNSF R. Co., 437 S.W.3d 207, 218 (Mo. 2014) (“compensation” does not include FELA damages for lost wages). We now hold that an award compensating for lost wages is subject to taxation under the RRTA. I In 1937, Congress created a self-sustaining retirement benefits system for railroad workers. The system provides generous pensions as well as benefits “correspon[ding] . . . to those an employee would expect to receive were he covered by the Social Security Act.” Hisquierdo v. Hisquierdo, 439 U.S. 572, 575 (1979). Two statutes operate in concert to ensure that retired railroad workers receive their allotted pensions and benefits. The first, the RRTA, funds the program by imposing a payroll tax on both railroads and their employees. The RRTA refers to the railroad’s contribution as an “excise” tax, 26 U. S. C. §3221, and describes the employee’s share as an “income” tax, §3201. Congress assigned to the Internal Revenue Service (IRS) responsibility for collecting both taxes. §§3501, 7801.[1] The second statute, the Railroad Retirement Act (RRA), 50Stat. 307, as restated and amended, 45 U. S. C. §231 et seq., entitles railroad workers to various benefits and prescribes eligibility requirements. The RRA is administered by the Railroad Retirement Board. See §231f(a). Taxes under the RRTA and benefits under the RRA are measured by the employee’s “compensation.” 26 U. S. C. §§3201, 3221; 45 U. S. C. §231b. The RRTA and RRA separately define “compensation,” but both statutes state that the term means “any form of money remuneration paid to an individual for services rendered as an employee.” 26 U. S. C. §3231(e)(1); 45 U. S. C. §231(h)(1). This language has remained basically unchanged since the RRTA’s enactment in 1937. See Carriers Taxing Act of 1937 (1937 RRTA), §1(e), 50Stat. 436 (defining “compensation” as “any form of money remuneration earned by an individual for services rendered as an employee”). The RRTA excludes from “compensation” certain types of sick pay and disability pay. See 26 U. S. C. §3231(e)(1), (4)(A). The IRS’s reading of the word “compensation” as it appears in the RRTA has remained constant. One year after the RRTA’s adoption, the IRS stated that “compensation” is not limited to pay for active service but reaches, as well, pay for periods of absence. See 26 CFR §410.5 (1938). This understanding has governed for more than eight decades. As restated in the current IRS regulations, “[t]he term compensation is not confined to amounts paid for active service, but includes amounts paid for an identifiable period during which the employee is absent from the active service of the employer.” §31.3231(e)–1(a)(3) (2017). In 1994, the IRS added, specifically, that “compensation” includes “pay for time lost.” §31.3231(e)–1(a)(4); see 59 Fed. Reg. 66188 (1994). Congress created both the railroad retirement system and the Social Security system during the Great Depression primarily to ensure the financial security of members of the workforce when they reach old age. See Wisconsin Central Ltd. v. United States, 585 U. S. ___, ___ (2018) (slip op., at 1); Helvering v. Davis, 301 U.S. 619, 641 (1937). Given the similarities in timing and purpose of the two programs, it is hardly surprising that their statutory foundations mirror each other. Regarding Social Security, the Federal Insurance Contributions Act (FICA), 26 U. S. C. §3101 et seq., taxes employers and employees to fund benefits, which are distributed pursuant to the Social Security Act (SSA), 49Stat. 620, as amended, 42 U. S. C. §301 et seq. Tax and benefit amounts are determined by the worker’s “wages,” the Social Security equivalent to “compensation.” See Davis, 301 U. S., at 635–636. Both the FICA and the SSA define “wages” employing language resembling the RRTA and the RRA definitions of “compensation.” “Wages” under the FICA and the SSA mean “all remuneration for employment,” and “employment,” in turn, means “any service, of whatever nature, performed . . . by an employee.” 26 U. S. C. §3121(a)–(b) (FICA); see 42 U. S. C. §§409(a), 410(a) (SSA). Reading these prescriptions together, the term “wages” encompasses “all remuneration” for “any service, of whatever nature, performed . . . by an employee.” Ibid. II A To determine whether RRTA-qualifying “compensation” includes an award of damages for lost wages, we begin with the statutory text.[2] The RRTA defines “compensation” as “remuneration paid to an individual for services rendered as an employee.” 26 U. S. C. §3231(e)(1). This definition, as just noted, is materially indistinguishable from the FICA’s definition of “wages” to include “remuneration” for “any service, of whatever nature, performed . . . by an employee.” §3121. Given the textual similarity between the definitions of “compensation” for railroad retirement purposes and “wages” for Social Security purposes, our decisions on the meaning of “wages” in Social Security Bd. v. Nierotko, 327 U.S. 358 (1946), and United States v. Quality Stores, Inc., 572 U.S. 141 (2014), inform our comprehension of the RRTA term “compensation.” In Nierotko, the National Labor Relations Board found that an employee had been “wrongfully discharged for union activity” and awarded him backpay. 327 U. S., at 359. The Social Security Board refused to credit the backpay award in calculating the employee’s benefits. Id., at 365–366. In the Board’s view, “wages” covered only pay for active service. Ibid. We disagreed. Emphasizing that the phrase “any service . . . performed” denotes “breadth of coverage,” we held that “wages” means remuneration for “the entire employer-employee relationship”; in other words, “wages” embraced pay for active service plus pay received for periods of absence from active service. Id., at 366. Backpay, we reasoned, counts as “wages” because it compensates for “the loss of wages which the employee suffered from the employer’s wrong.” Id., at 364. In Quality Stores, we again trained on the meaning of “wages,” reiterating that “Congress chose to define wages . . . broadly.” 572 U. S., at 146 (internal quotation marks omitted). Guided by Nierotko, Quality Stores held that severance payments qualified as “wages” taxable under the FICA. “[C]ommon sense,” we observed, “dictates that employees receive th[ose] payments ‘for employment.’ ” 572 U. S., at 146. Severance payments, the Court spelled out, “are made to employees only,” “are made in consideration for employment,” and are calculated “according to the function and seniority of the [terminated] employee.” Id., at 146–147. In line with Nierotko, Quality Stores, and the IRS’s long held construction, we hold that “compensation” under the RRTA encompasses not simply pay for active service but, in addition, pay for periods of absence from active service—provided that the remuneration in question stems from the “employer-employee relationship.” Nierotko, 327 U. S., at 366. B Damages awarded under the FELA for lost wages fit comfortably within this definition. The FELA “makes railroads liable in money damages to their employees for on-the-job injuries.” BNSF R. Co. v. Tyrrell, 581 U. S. ___, ___ (2017) (slip op., at 1); see 45 U. S. C. §51. If a railroad negligently fails to maintain a safe railyard and a worker is injured as a result, the FELA requires the railroad to compensate the injured worker for, inter alia, working time lost due to the employer’s wrongdoing. FELA damages for lost wages, then, are functionally equivalent to an award of backpay, which compensates an employee “for a period of time during which” the employee is “wrongfully separated from his job.” Nierotko, 327 U. S., at 364. Just as Nierotko held that backpay falls within the definition of “wages,” ibid., we conclude that FELA damages for lost wages qualify as “compensation” and are therefore taxable under the RRTA. III A The Eighth Circuit construed “compensation” for RRTA purposes to mean only pay for “services that an employee actually renders,” in other words, pay for active service. Consequently, the court held that “compensation” within the RRTA’s compass did not reach pay for periods of absence. 865 F. 3d, at 1117. In so ruling, the Court of Appeals attempted to distinguish Nierotko and Quality Stores. The Social Security decisions, the court said, were inapposite because the FICA “taxes payment for ‘employment,’ ” whereas the RRTA “tax[es] payment for ‘services.’ ” 865 F. 3d, at 1117. As noted, however, supra, at 3–4, the FICA defines “employment” in language resembling the RRTA in all relevant respects. Compare 26 U. S. C. §3121(b) (FICA) (“any service, of whatever nature, performed . . . by an employee”) with §3231(e)(1) (RRTA) (“services rendered as an employee”). Construing RRTA “compensation” as less embracive than “wages” covered by the FICA would introduce an unwarranted disparity between terms Congress appeared to regard as equivalents. The reasoning of Nierotko and Quality Stores, as we see it, resists the Eighth Circuit’s swift writeoff.[3] Nierotko and Quality Stores apart, we would in any event conclude that the RRTA term “compensation” covers pay for time lost. Restricting “compensation” to pay for active service, the Court of Appeals relied on statutory history and, in particular, the eventual deletion of two references to pay for time lost contained in early renditions of the RRTA. See also post, at 6–7 (presenting the Eighth Circuit’s statutory history argument). To understand the Eighth Circuit’s position, and why, in our judgment, that position does not withstand scrutiny, some context is in order. On enactment of the RRTA in 1937, Congress made “compensation” taxable at the time it was earned and provided specific guidance on when pay for time lost should be “deemed earned.” Congress instructed: “The term ‘compensation’ means any form of money remuneration earned by an individual for services rendered as an employee . . . , including remuneration paid for time lost as an employee, but [such] remuneration . . . shall be deemed earned in the month in which such time is lost.” 1937 RRTA, §1(e), 50Stat. 436 (emphasis added). In 1946, Congress clarified that the phrase “pa[y] for time lost” meant payment for “an identifiable period of absence from the active service of the employer, including absence on account of personal injury.” Act of July 31, 1946 (1946 Act), §2, 60Stat. 722. Thus, originally, the RRTA stated that “compensation” included pay for time lost, and the language added in 1946 presupposed the same. In subsequent amendments, however, Congress removed the references to pay for time lost. First, in 1975, Congress made “compensation” taxable when paid rather than when earned. Congress simultaneously removed the 1937 language that both referred to pay for time lost and specified when such pay should be “deemed earned.” So amended, the definitional sentence, in its current form, reads: “The term ‘compensation’ means any form of money remuneration paid to an individual for services rendered as an employee . . . .” Act of Aug. 9, 1975 (1975 Act), §204, 89Stat. 466 (emphasis added). Second, in 1983, Congress shifted the wage base for RRTA taxes from monthly “compensation” to annual “compensation.” See Railroad Retirement Solvency Act of 1983 (1983 Act), §225, 97Stat. 424–425. Because the “monthly wage bases for railroad retirement taxes [were being] changed to annual amounts,” the House Report explained, the RRTA required “[s]everal technical and conforming amendments.” H. R. Rep. No. 98–30, pt. 2, p. 29 (1983). In a section of the 1983 Act titled “Technical Amendments,” Congress struck the subsection containing, among other provisions, the 1946 Act’s clarification of pay for time lost. 1983 Act, §225, 97Stat. 424–425. In lieu of the deleted subsection, Congress inserted detailed instructions concerning the new annual wage base. As the Court of Appeals and the dissent see it, the 1975 and 1983 deletions show that “compensation” no longer includes pay for time lost. 865 F. 3d, at 1119; see post, at 6–7. We are not so sure. The 1975 Act left unaltered the language at issue here, “remuneration . . . for services rendered as an employee.” That Act also left intact the 1946 Act’s description of pay for time lost. Continuing after the 1975 Act, then, such pay remained RRTA-taxable “compensation.” The 1983 Act, as billed by Congress, effected only “[t]echnical [a]mendments” relating to the change from monthly to annual computation of “compensation.” Concerning the 1975 and 1983 alterations, the IRS concluded that Congress revealed no “inten[tion] to exclude payments for time lost from compensation.” 59 Fed. Reg. 66188 (1994). We credit the IRS reading. It would be passing strange for Congress to restrict substantially what counts as “compensation” in a manner so oblique. Moreover, the text of the RRTA continues to indicate that “compensation” encompasses pay for time lost. The RRTA excludes from “compensation” a limited subset of payments for time lost, notably certain types of sick pay and disability pay. See 26 U. S. C. §3231(e)(1), (4). These enumerated exclusions would be entirely superfluous if, as the Court of Appeals held, the RRTA broadly excludes from “compensation” any and all pay received for time lost. In justification of its confinement of RRTA-taxable receipts to pay for active service, the Court of Appeals also referred to the RRA. The RRA, like the RRTA as enacted in 1937, states that “compensation” “includ[es] remuneration paid for time lost as an employee” and specifies that such pay “shall be deemed earned in the month in which such time is lost.” 45 U. S. C. §231(h)(1). Pointing to the discrepancy between the RRA and the amended RRTA, which no longer contains the above-quoted language, the Court of Appeals concluded that Congress intended the RRA, but not the RRTA, to include pay for time lost. Accord post, at 7. Although “ ‘[w]e usually presume differences in language . . . convey differences in meaning,’ ” Wisconsin Central, 585 U. S., at ___ (slip op., at 4), Congress’ failure to reconcile the RRA and the amended RRTA is inconsequential. As just explained, the RRTA’s pinpointed exclusions from RRTA taxation signal that nonexcluded pay for time lost remains RRTA-taxable “compensation.” B Instead of adopting lockstep the Court of Appeals’ interpretation, Loos takes a different approach. In his view, echoed by the dissent, “remuneration . . . for services rendered” means the “package of benefits” an employer pays “to retain the employee.” Brief for Respondent 37; post, at 3–4. He therefore agrees with BNSF that benefits like sick pay and vacation pay are taxable “compensation.” He contends, however, that FELA damages for lost wages are of a different order. They are not part of an employee’s “package of benefits,” he observes, and therefore should not count as “compensation.” Such damages, Loos urges, “compensate for an injury” rather than for services rendered. Brief for Respondent 20; post, at 3–4. Loos argues in the alternative that even if voluntary settlements qualify as “compensation,” “involuntary payment[s]” in the form of damages do not. Brief for Respondent 33. Our decision in Nierotko undermines Loos’s argument that, unlike sick pay and vacation pay, payments “compensat[ing] for an injury,” Brief for Respondent 20, are not taxable under the RRTA. We held in Nierotko that an award of backpay compensating an employee for his wrongful discharge ranked as “wages” under the SSA. That was so, we explained, because the backpay there awarded to the employee redressed “the loss of wages” occasioned by “the employer’s wrong.” 327 U. S., at 364; see supra, at 5. Applying that reasoning here, there should be no dispositive difference between a payment voluntarily made and one required by law.[4] Nor does United States v. Cleveland Indians Baseball Co., 532 U.S. 200 (2001), aid Loos’s argument, repeated by the dissent. See post, at 8. Indeed, Cleveland Indians reasserted Nierotko’s holding that “backpay for a time in which the employee was not on the job” counts as pay for services, and therefore ranks as wages. 532 U. S., at 210. Cleveland Indians then took up a discrete, “secondary issue” Nierotko presented, one not in contention here, i.e., whether for taxation purposes backpay is allocable to the tax period when paid rather than an earlier time-earned period. 532 U. S., at 211, 213–214, 219–220. Moreover, Quality Stores, which postdated Cleveland Indians, left no doubt that what qualifies under Nierotko as “wages” for benefit purposes also qualifies as such for taxation purposes. 572 U. S., at 146–147. C Loos presses a final reason why he should not owe RRTA taxes on his lost wages award. Loos argues, and the District Court held, that the RRTA’s tax on employees does not apply to personal injury damages. He observes that the RRTA taxes “the income of each employee.” 26 U. S. C. §3201(a)–(b) (emphasis added). He then cites a provision of the Internal Revenue Code, 26 U. S. C. §104(a)(2). This provision exempts “damages . . . received . . . on account of personal physical injuries” from federal income taxation by excluding such damages from “gross income.” Loos urges that the exclusion of personal injury damages from “gross income” should carry over to the RRTA’s tax on the “income” of railroad workers, §3201(a)–(b). The argument is unconvincing. As the Government points out, the District Court, echoed by Loos, conflated “the distinct concepts of ‘gross income,’ [a prime component of] the tax base on which income tax is collected, and ‘compensation,’ the separately defined category of payments that are taxable under the RRTA.” Brief for United States as Amicus Curiae 15. Blending tax bases that Congress kept discrete, the District Court and Loos proffer a scheme in which employees pay no tax on damages compensating for personal injuries; railroads pay the full excise tax on such compensation; and employees receive full credit for the compensation in determining their retirement benefits. That scheme, however, is not plausibly attributable to Congress. For federal income tax purposes, “gross income” means “all income” “[e]xcept as otherwise provided.” 26 U. S. C. §61; see §§1, 63 (imposing a tax on “taxable income,” defined as “gross income minus . . . deductions”). Congress provided detailed prescriptions on the scope of “gross income,” excluding from its reach numerous items, among them, personal injury damages. See §§101–140. Conspicuously absent from the RRTA, however, is any reference to “gross income.” As employed in the RRTA, the word “income” merely distinguishes the tax on the employee, an “income . . . tax,” §3201, from the matching tax on the railroad, called an “excise tax.” §§3201, 3221. See also 1937 RRTA, §§2–3 (establishing an “income tax on employees” and an “excise tax on employers”); S. Rep. No. 818, 75th Cong., 1st Sess., 5 (1937) (stating that the RRTA imposes an “income tax on employees” and an “excise tax on employers”); H. R. Rep. No. 1071, 75th Cong., 1st Sess., 6 (1937) (same). Congress, we reiterate, specified not “gross income” but employee “compensation” as the tax base for the RRTA’s income and excise taxes. §§3201, 3221. Congress then excepted certain payments from the calculation of “compensation.” See §3231(e); supra, at 9. Congress adopted by cross-reference particular Internal Revenue Code exclusions from “gross income,” thereby carving out those specified items from RRTA coverage. See §3231(e)(5)–(6), (9)–(11). Tellingly, Congress did not adopt for RRTA purposes the exclusion of personal injury damages from federal income taxation set out in §104(a)(2). We note, furthermore, that if RRTA taxes were based on “income” or “gross income” rather than “compensation,” the RRTA tax base would sweep in nonrailroad income, including, for example, dividends, interest accruals, even lottery winnings. Shifting from “compensation” to “income” as the RRTA tax base would thus saddle railroad workers with more RRTA taxes. Given the multiple flaws in Loos’s last ditch argument, we conclude that §104(a)(2) does not exempt FELA damages from the RRTA’s income and excise taxes. * * * In harmony with this Court’s decisions in Nierotko and Quality Stores, we hold that “compensation” for RRTA purposes includes an employer’s payments to an employee for active service and for periods of absence from active service. It is immaterial whether the employer chooses to make the payment or is legally required to do so. Either way, the payment is remitted to the recipient because of his status as a service-rendering employee. See 26 U. S. C. §3231(e)(1); 45 U. S. C. §231(h)(1). For the reasons stated, FELA damages for lost wages qualify as RRTA-taxable “compensation.” The judgment of the Court of Appeals for the Eighth Circuit is accordingly reversed, and the case is remanded for proceedings consistent with this opinion. It is so ordered. Notes 1 The railroad remits both taxes to the IRS. As to the income tax, the railroad deducts the amount owed by the employee from her earnings and then forwards that amount to the IRS. See Tr. of Oral Arg. 22–23. See also 26 U. S. C. §3402(a)(1) (employers must “deduct and withhold” income taxes from earnings). 2 Before turning to the language of the RRTA, the dissent endeavors to unearth the reason why BNSF has pursued this case. The railroad’s “gambit,” the dissent surmises, is to increase pressure on injured workers to settle their claims. Post, at 3. Contrast with the dissent’s conjecture, BNSF’s entirely plausible account of a railroad’s stake in this dispute. Because the RRA credits lost wages toward an employee’s benefits, see 45 U. S. C. §231(h)(1), BNSF posits that immunizing those payments from RRTA taxes would expose the system to “a long-term risk of insolvency.” Tr. of Oral Arg. 4; see Reply Brief for Petitioner 14. 3 The dissent’s reduction of Nierotko’s significance fares no better. Nierotko, the dissent urges, is distinguishable because it involved “a different factual context.” Post, at 7. But as just explained, supra, at 6–7, the facts in Nierotko resemble those here in all material respects. 4 The dissent, building on Loos’s argument, tenders an inapt analogy between passengers and employees. If BNSF were ordered to pay damages for lost wages to an injured passenger, the dissent asserts, one would not say the passenger had been compensated “for services rendered.” There is no reason, the dissent concludes, to “reach a different result here simply because the victim of BNSF’s negligence happened to be one of its own workers.” Post, at 5. Under the RRTA, however, this distinction is of course critical. The passenger’s damages for lost wages are not taxable under the RRTA, for she has no employment relationship with the railroad. In contrast, FELA damages for lost wages are taxable because they are paid only if the injured person previously “rendered [services] as an employee,” 26 U. S. C. §3231(e)(1), and, indeed, was working for the railroad when the injury occurred, see 45 U. S. C. §51.

587.US.2018_17-8151

In Baze v. Rees, 553 U.S. 35, a plurality of this Court concluded that a State’s refusal to alter its execution protocol could violate the Eighth Amendment only if an inmate first identified a “feasible, readily implemented” alternative procedure that would “significantly reduce a substantial risk of severe pain.” Id., at 52. A majority of the Court subsequently held Baze’s plurality opinion to be controlling. See Glossip v. Gross, 576 U. S. ___. Petitioner Russell Bucklew was convicted of murder and sentenced to death. The State of Missouri plans to execute him by lethal injection using a single drug, pentobarbital. Mr. Bucklew presented an as-applied Eighth Amendment challenge to the State’s lethal injection protocol, alleging that, regardless whether it would cause excruciating pain for all prisoners, it would cause him severe pain because of his particular medical condition. The District Court dismissed his challenge. The Eighth Circuit, applying the Baze-Glossip test, remanded the case to allow Mr. Bucklew to identify a feasible, readily implemented alternative procedure that would significantly reduce his alleged risk of pain. Eventually, Mr. Bucklew identified nitrogen hypoxia, but the District Court found the proposal lacking and granted the State’s motion for summary judgment. The Eighth Circuit affirmed. Held: 1. Baze and Glossip govern all Eighth Amendment challenges, whether facial or as-applied, alleging that a method of execution inflicts unconstitutionally cruel pain. Pp. 8–20. (a) The Eighth Amendment forbids “cruel and unusual” methods of capital punishment but does not guarantee a prisoner a painless >death. See Glossip, 576 U. S., at ___. As originally understood, the Eighth Amendment tolerated methods of execution, like hanging, that involved a significant risk of pain, while forbidding as cruel only those methods that intensified the death sentence by “superadding” terror, pain, or disgrace. To establish that a State’s chosen method cruelly “superadds” pain to the death sentence, a prisoner must show a feasible and readily implemented alternative method that would significantly reduce a substantial risk of severe pain and that the State has refused to adopt without a legitimate penological reason. Baze, 553 U. S., at 52; Glossip, 576 U. S., at ___. And Glossip left no doubt that this standard governs “all Eighth Amendment method-of-execution claims.” Id., at ___. Baze and Glossip recognized that the Constitution affords a “measure of deference to a State’s choice of execution procedures” and does not authorize courts to serve as “boards of inquiry charged with determining ‘best practices’ for executions.” Baze, 553 U. S., at 51–52. Nor do they suggest that traditionally accepted methods of execution are necessarily rendered unconstitutional as soon as an arguably more humane method becomes available. Pp. 8–14. (b) Precedent forecloses Mr. Bucklew’s argument that methods posing a “substantial and particular risk of grave suffering” when applied to a particular inmate due to his “unique medical condition” should be considered “categorically” cruel. Because distinguishing between constitutionally permissible and impermissible degrees of pain is a necessarily comparative exercise, the Court held in Glossip, identifying an available alternative is “a requirement of all Eighth Amendment method-of-execution claims” alleging cruel pain. 576 U. S., at ___. Mr. Bucklew’s argument is also inconsistent with the original and historical understanding of the Eighth Amendment on which Baze and Glossip rest: When it comes to determining whether a punishment is unconstitutionally cruel because of the pain involved, the law has always asked whether the punishment superadds pain well beyond what’s needed to effectuate a death sentence. And answering that question has always involved a comparison with available alternatives, not an abstract exercise in “categorical” classification. The substantive meaning of the Eighth Amendment does not change depending on how broad a remedy the plaintiff chooses to seek. Mr. Bucklew’s solution also invites pleading games, and there is little likelihood that an inmate facing a serious risk of pain will be unable to identify an available alternative. Pp. 14–20. 2. Mr. Bucklew has failed to satisfy the Baze-Glossip test. Pp. 20–28. (a) He fails for two independent reasons to present a triable question on the viability of nitrogen hypoxia as an alternative to the State’s lethal injection protocol. First, an inmate must show that his proposed alternative method is not just theoretically “feasible” but also “ ‘readily implemented,’ ” Glossip, 576 U. S., at ___–___. This means the inmate’s proposal must be sufficiently detailed to permit a finding that the State could carry it out relatively easily and reasonably quickly. Mr. Bucklew’s proposal falls well short of that standard. He presented no evidence on numerous questions essential to implementing his preferred method; instead, he merely pointed to reports from correctional authorities in other States indicating the need for additional study to develop a nitrogen hypoxia protocol. Second, the State had a “legitimate” reason for declining to switch from its current method of execution as a matter of law, Baze, 553 U. S., at 52, namely, choosing not to be the first to experiment with a new, “untried and untested” method of execution. Id., at 41. Pp. 20–22. (b) Even if nitrogen hypoxia were a viable alternative, neither of Mr. Bucklew’s theories shows that nitrogen hypoxia would significantly reduce a substantial risk of severe pain. First, his contention that the State may use painful procedures to administer the lethal injection, including forcing him to lie flat on his back (which he claims could impair his breathing even before the pentobarbital is administered), rests on speculation unsupported, if not affirmatively contradicted, by the record. And to the extent the record is unclear, he had ample opportunity to conduct discovery and develop a factual record concerning the State’s planned procedures. Second, Mr. Bucklew contends that while either method will cause him to experience feelings of suffocation for some period of time before he is rendered fully unconscious, the duration of that period will be shorter with nitrogen than with pentobarbital. But nothing in the record suggests that he will be capable of experiencing pain for significantly more time after receiving pentobarbital than he would after receiving nitrogen. His claim to the contrary rested on his expert’s testimony regarding a study of euthanasia in horses that everyone now agrees the expert misunderstood or misremembered. Pp. 23–28. 883 F.3d 1087, affirmed. Gorsuch, J., delivered the opinion of the Court, in which Roberts, C. J., and Thomas, Alito, and Kavanaugh, JJ., joined. Thomas, J., and Kavanaugh, J., filed concurring opinions. Breyer, J., filed a dissenting opinion, in which Ginsburg, Sotomayor, and Kagan, JJ., joined as to all but Part III. Sotomayor, J., filed a dissenting opinion.

Russell Bucklew concedes that the State of Missouri lawfully convicted him of murder and a variety of other crimes. He acknowledges that the U. S. Constitution permits a sentence of execution for his crimes. He accepts, too, that the State’s lethal injection protocol is constitutional in most applications. But because of his unusual medical condition, he contends the protocol is unconstitutional as applied to him. Mr. Bucklew raised this claim for the first time less than two weeks before his scheduled execution. He received a stay of execution and five years to pursue the argument, but in the end neither the district court nor the Eighth Circuit found it supported by the law or evidence. Now, Mr. Bucklew asks us to overturn those judgments. We can discern no lawful basis for doing so. I A In 1996, when Stephanie Ray announced that she wanted to end their relationship, Mr. Bucklew grew violent. He cut her jaw, punched her in the face, and threatened her with a knife. Frightened to remain in the home they had shared, Ms. Ray sought refuge with her children in Michael Sanders’ nearby residence. But then one night Mr. Bucklew invaded that home. Bearing a pistol in each hand, he shot Mr. Sanders in the chest; fired at Mr. Sanders’ 6-year-old son (thankfully, he missed); and pistol-whipped Ms. Ray, this time breaking her jaw. Then Mr. Bucklew handcuffed Ms. Ray, drove her to a secluded spot, and raped her at gunpoint. After a trooper spotted Mr. Bucklew, a shootout followed and he was finally arrested. While all this played out, Mr. Sanders bled to death. As a coda, Mr. Bucklew escaped from jail while awaiting trial and attacked Ms. Ray’s mother with a hammer before he could be recaptured. After a decade of litigation, Mr. Bucklew was seemingly out of legal options. A jury had convicted him of murder and other crimes and recommended a death sentence, which the court had imposed. His direct appeal had proved unsuccessful. State v. Bucklew, 973 S.W.2d 83 (Mo. 1998), cert. denied, 525 U.S. 1082 (1999). Separate rounds of state and federal post-conviction proceedings also had failed to yield relief. Bucklew v. State, 38 S.W.3d 395 (Mo.), cert. denied, 534 U.S. 964 (2001); Bucklew v. Luebbers, 436 F.3d 1010 (CA8), cert. denied, 549 U.S. 1079 (2006). B As it turned out, though, Mr. Bucklew’s case soon became caught up in a wave of litigation over lethal injection procedures. Like many States, Missouri has periodically sought to improve its administration of the death penalty. Early in the 20th century, the State replaced hanging with the gas chamber. Later in the century, it authorized the use of lethal injection as an alternative to lethal gas. By the time Mr. Bucklew’s post-conviction proceedings ended, Missouri’s protocol called for lethal injections to be carried out using three drugs: sodium thiopental, pancuronium bromide, and potassium chloride. And by that time, too, various inmates were in the process of challenging the constitutionality of the State’s protocol and others like it around the country. See Taylor v. Crawford, 457 F.3d 902 (CA8 2006); Note, A New Test for Evaluating Eighth Amendment Challenges to Lethal Injections, 120 Harv. L. Rev. 1301, 1304 (2007) (describing flood of lethal injection lawsuits around 2006 that “severely constrained states’ ability to carry out executions”); Denno, The Lethal Injection Quandary: How Medicine Has Dismantled the Death Penalty, 76 Ford. L. Rev. 49, 102–116 (2007). Ultimately, this Court answered these legal challenges in Baze v. Rees, 553 U.S. 35 (2008). Addressing Kentucky’s similar three-drug protocol, The Chief Justice, joined by Justice Alito and Justice Kennedy, concluded that a State’s refusal to alter its lethal injection protocol could violate the Eighth Amendment only if an inmate first identified a “feasible, readily implemented” alternative procedure that would “significantly reduce a substantial risk of severe pain.” Id., at 52. Justice Thomas, joined by Justice Scalia, thought the protocol passed muster because it was not intended “to add elements of terror, pain, or disgrace to the death penalty.” Id., at 107. Justice Breyer reached the same result because he saw no evidence that the protocol created “a significant risk of unnecessary suffering.” Id., at 113. And though Justice Stevens objected to the continued use of the death penalty, he agreed that petitioners’ evidence was insufficient. Id., at 87. After this Court decided Baze, it denied review in a case seeking to challenge Missouri’s similar lethal injection protocol. Taylor v. Crawford, 487 F.3d 1072 (2007), cert. denied, 553 U.S. 1004 (2008). But that still was not the end of it. Next, Mr. Bucklew and other inmates unsuccessfully challenged Missouri’s protocol in state court, alleging that it had been adopted in contravention of Missouri’s Administrative Procedure Act. Middleton v. Missouri Dept. of Corrections, 278 S.W.3d 193 (Mo.), cert. denied, 556 U.S. 1255 (2009). They also unsuccessfully challenged the protocol in federal court, this time alleging it was pre-empted by various federal statutes. Ringo v. Lombardi, 677 F.3d 793 (CA8 2012). And Mr. Bucklew sought to intervene in yet another lawsuit alleging that Missouri’s protocol violated the Eighth Amendment because unqualified personnel might botch its administration. That lawsuit failed too. Clemons v. Crawford, 585 F.3d 1119 (CA8 2009), cert. denied, 561 U.S. 1026 (2010). While all this played out, pressure from anti-death-penalty advocates induced the company that manufactured sodium thiopental to stop supplying it for use in executions. As a result, the State was unable to proceed with executions until it could change its lethal injection protocol again. This it did in 2012, prescribing the use of a single drug, the sedative propofol. Soon after that, Mr. Bucklew and other inmates sued to invalidate this new protocol as well, alleging that it would produce excruciating pain and violate the Eighth Amendment on its face. After the State revised the protocol in 2013 to use the sedative pentobarbital instead of propofol, the inmates amended their complaint to allege that pentobarbital would likewise violate the Constitution. C Things came to a head in 2014. With its new protocol in place and the necessary drugs now available, the State scheduled Mr. Bucklew’s execution for May 21. But 12 days before the execution Mr. Bucklew filed yet another lawsuit, the one now before us. In this case, he presented an as-applied Eighth Amendment challenge to the State’s new protocol. Whether or not it would cause excruciating pain for all prisoners, as his previous lawsuit alleged, Mr. Bucklew now contended that the State’s protocol would cause him severe pain because of his particular medical condition. Mr. Bucklew suffers from a disease called cavernous hemangioma, which causes vascular tumors—clumps of blood vessels—to grow in his head, neck, and throat. His complaint alleged that this condition could prevent the pentobarbital from circulating properly in his body; that the use of a chemical dye to flush the intravenous line could cause his blood pressure to spike and his tumors to rupture; and that pentobarbital could interact adversely with his other medications. These latest protocol challenges yielded mixed results. The district court dismissed both the inmates’ facial challenge and Mr. Bucklew’s as-applied challenge. But, at Mr. Bucklew’s request, this Court agreed to stay his execution until the Eighth Circuit could hear his appeal. Bucklew v. Lombardi, 572 U.S. 1131 (2014). Ultimately, the Eighth Circuit affirmed the dismissal of the facial challenge. Zink v. Lombardi, 783 F.3d 1089 (en banc) (per curiam), cert. denied, 576 U. S. ___ (2015). Then, turning to the as-applied challenge and seeking to apply the test set forth by the Baze plurality, the court held that Mr. Bucklew’s complaint failed as a matter of law to identify an alternative procedure that would significantly reduce the risks he alleged would flow from the State’s lethal injection protocol. Yet, despite this dispositive shortcoming, the court of appeals decided to give Mr. Bucklew another chance to plead his case. The court stressed that, on remand before the district court, Mr. Bucklew had to identify “at the earliest possible time” a feasible, readily implemented alternative procedure that would address those risks. Bucklew v. Lombardi, 783 F.3d 1120, 1127–1128 (2015) (en banc). Shortly after the Eighth Circuit issued its judgment, this Court decided Glossip v. Gross, 576 U. S. ___ (2015), rejecting a challenge to Oklahoma’s lethal injection protocol. There, the Court clarified that The Chief Justice’s plurality opinion in Baze was controlling under Marks v. United States, 430 U.S. 188 (1977). In doing so, it reaffirmed that an inmate cannot successfully challenge a method of execution under the Eighth Amendment unless he identifies “an alternative that is ‘feasible, readily implemented, and in fact significantly reduces a substantial risk of severe pain.’ ” 576 U. S., at ___–___ (slip op., at 12–13). Justice Thomas, joined by Justice Scalia, reiterated his view that the Eighth Amendment “prohibits only those methods of execution that are deliberately designed to inflict pain,” but he joined the Court’s opinion because it correctly explained why petitioners’ claim failed even under the controlling opinion in Baze. Glossip, 576 U. S., at ___ (concurring opinion) (slip op., at 1) (internal quotation marks and alterations omitted). D Despite the Eighth Circuit’s express instructions, when Mr. Bucklew returned to the district court in 2015 he still refused to identify an alternative procedure that would significantly reduce his alleged risk of pain. Instead, he insisted that inmates should have to carry this burden only in facial, not as-applied, challenges. Finally, after the district court gave him “one last opportunity,” App. 30, Mr. Bucklew filed a fourth amended complaint in which he claimed that execution by “lethal gas” was a feasible and available alternative method that would significantly reduce his risk of pain. Id., at 42. Mr. Bucklew later clarified that the lethal gas he had in mind was nitrogen, which neither Missouri nor any other State had ever used to carry out an execution. The district court allowed Mr. Bucklew “extensive discovery” on his new proposal. 883 F.3d 1087, 1094 (CA8 2018). But even at the close of discovery in 2017, the district court still found the proposal lacking and granted the State’s motion for summary judgment. By this point in the proceedings, Mr. Bucklew’s contentions about the pain he might suffer had evolved considerably. He no longer complained about circulation of the drug, the use of dye, or adverse drug interactions. Instead, his main claim now was that he would experience pain during the period after the pentobarbital started to take effect but before it rendered him fully unconscious. According to his expert, Dr. Joel Zivot, while in this semiconscious “twilight stage” Mr. Bucklew would be unable to prevent his tumors from obstructing his breathing, which would make him feel like he was suffocating. Dr. Zivot declined to say how long this twilight stage would last. When pressed, however, he referenced a study on euthanasia in horses. He claimed that the horses in the study had displayed some amount of brain activity, as measured with an electroencephalogram (or EEG), for up to four minutes after they were given a large dose of pentobarbital. Based on Dr. Zivot’s testi- mony, the district court found a triable issue as to whether there was a “substantial risk” that Mr. Bucklew would “experience choking and an inability to breathe for up to four minutes” if he were executed by lethal injection. App. 827. Even so, the court held, Mr. Bucklew’s claim failed because he had produced no evidence that his proposed alternative, execution by nitrogen hypoxia, would significantly reduce that risk. This time, a panel of the Eighth Circuit affirmed. The panel held that Mr. Bucklew had produced no evidence that the risk of pain he alleged “would be substantially reduced by use of nitrogen hypoxia instead of lethal injection as the method of execution.” 883 F. 3d, at 1096. Judge Colloton dissented, arguing that the evidence raised a triable issue as to whether nitrogen gas would “render Bucklew insensate more quickly than pentobarbital.” Id., at 1099. The full court denied rehearing en banc over a dissent by Judge Kelly, who maintained that, while prisoners pursuing facial challenges to a state execution protocol must plead and prove an alternative method of execution under Baze and Glossip, prisoners like Mr. Bucklew who pursue as-applied challenges should not have to bear that burden. 885 F.3d 527, 528 (2018). On the same day Mr. Bucklew was scheduled to be executed, this Court granted him a second stay of execution. 583 U. S. ___ (2018). We then agreed to hear his case to clarify the legal standards that govern an as-applied Eighth Amendment challenge to a State’s method of carrying out a death sentence. 584 U. S. ___ (2018). II We begin with Mr. Bucklew’s suggestion that the test for lethal injection protocol challenges announced in Baze and Glossip should govern only facial challenges, not as-applied challenges like his. In evaluating this argument, we first examine the original and historical understanding of the Eighth Amendment and our precedent in Baze and Glossip. We then address whether, in light of those authorities, it would be appropriate to adopt a different constitutional test for as-applied claims. A The Constitution allows capital punishment. See Glossip, 576 U. S., at ___–___ (slip op., at 2–4); Baze, 553 U. S., at 47. In fact, death was “the standard penalty for all serious crimes” at the time of the founding. S. Banner, The Death Penalty: An American History 23 (2002) (Banner). Nor did the later addition of the Eighth Amendment outlaw the practice. On the contrary—the Fifth Amendment, added to the Constitution at the same time as the Eighth, expressly contemplates that a defendant may be tried for a “capital” crime and “deprived of life” as a pen- alty, so long as proper procedures are followed. And the First Congress, which proposed both Amendments, made a number of crimes punishable by death. See Act of Apr. 30, 1790, 1Stat. 112. Of course, that doesn’t mean the American people must continue to use the death penalty. The same Constitution that permits States to authorize capital punishment also allows them to outlaw it. But it does mean that the judiciary bears no license to end a debate reserved for the people and their representatives. While the Eighth Amendment doesn’t forbid capital punishment, it does speak to how States may carry out that punishment, prohibiting methods that are “cruel and unusual.” What does this term mean? At the time of the framing, English law still formally tolerated certain punishments even though they had largely fallen into disuse—punishments in which “terror, pain, or disgrace [were] superadded” to the penalty of death. 4 W. Blackstone, Commentaries on the Laws of England 370 (1769). These included such “[d]isgusting” practices as dragging the prisoner to the place of execution, disemboweling, quartering, public dissection, and burning alive, all of which Blackstone observed “savor[ed] of torture or cruelty.” Ibid. Methods of execution like these readily qualified as “cruel and unusual,” as a reader at the time of the Eighth Amendment’s adoption would have understood those words. They were undoubtedly “cruel,” a term often defined to mean “[p]leased with hurting others; inhuman; hard-hearted; void of pity; wanting compassion; savage; barbarous; unrelenting,” 1 S. Johnson, A Dictionary of the English Language (4th ed. 1773), or “[d]isposed to give pain to others, in body or mind; willing or pleased to torment, vex or afflict; inhuman; destitute of pity, compassion or kindness,” 1 N. Webster, An American Dictionary of the English Language (1828). And by the time of the founding, these methods had long fallen out of use and so had become “unusual.” 4 Blackstone, supra, at 370; Banner 76; Baze, 553 U. S., at 97 (Thomas, J., concurring in judgment); see also Stinneford, The Original Meaning of “Unusual”: The Eighth Amendment as a Bar to Cruel Innovation, 102 Nw. U. L. Rev. 1739, 1770–1771, 1814 (2008) (observing that Americans in the late 18th and early 19th centuries described as “unusual” governmental actions that had “fall[en] completely out of usage for a long period of time”). Contemporary evidence confirms that the people who ratified the Eighth Amendment would have understood it in just this way. Patrick Henry, for one, warned that unless the Constitution was amended to prohibit “cruel and unusual punishments,” Congress would be free to inflict “tortures” and “barbarous” punishments. 3 Debates on the Federal Constitution 447–448 (J. Elliot 2d ed. 1891). Many early commentators likewise described the Eighth Amendment as ruling out “the use of the rack or the stake, or any of those horrid modes of torture devised by human ingenuity for the gratification of fiendish passion.” J. Bayard, A Brief Exposition of the Constitution of the United States 140 (1833); see B. Oliver, The Rights of an American Citizen 186 (1832) (the Eighth Amendment prohibits such “barbarous and cruel punishments” as “[b]reaking on the wheel, flaying alive, rending asunder with horses, . . . maiming, mutilating and scourging to death”). Justice Story even remarked that he thought the prohibition of cruel and unusual punishments likely “unnecessary” because no “free government” would ever authorize “atrocious” methods of execution like these. 3 J. Story, Commentaries on the Constitution of the United States §1896, p. 750 (1833). Consistent with the Constitution’s original understanding, this Court in Wilkerson v. Utah, 99 U.S. 130 (1879), permitted an execution by firing squad while observing that the Eighth Amendment forbade the gruesome methods of execution described by Blackstone “and all others in the same line of unnecessary cruelty.” Id., at 135–136. A few years later, the Court upheld a sentence of death by electrocution while observing that, though electrocution was a new mode of punishment and therefore perhaps could be considered “unusual,” it was not “cruel” in the constitutional sense: “[T]he punishment of death is not cruel, within the meaning of that word as used in the Constitution. [Cruelty] implies . . . something inhuman and barbarous, something more than the mere extinguishment of life.” In re Kemmler, 136 U.S. 436, 447 (1890). It’s instructive, too, to contrast the modes of execution the Eighth Amendment was understood to forbid with those it was understood to permit. At the time of the Amendment’s adoption, the predominant method of execution in this country was hanging. Glossip, 576 U. S., at ___ (slip op., at 2). While hanging was considered more humane than some of the punishments of the Old World, it was no guarantee of a quick and painless death. “Many and perhaps most hangings were evidently painful for the condemned person because they caused death slowly,” and “[w]hether a hanging was painless or painful seems to have been largely a matter of chance.” Banner 48, 170. The force of the drop could break the neck and sever the spinal cord, making death almost instantaneous. But that was hardly assured given the techniques that prevailed at the time. More often it seems the prisoner would die from loss of blood flow to the brain, which could produce unconsciousness usually within seconds, or suffocation, which could take several minutes. Id., at 46–47; J. Laurence, The History of Capital Punishment 44–46 (1960); Gardner, Executions and Indignities: An Eighth Amendment Assessment of Methods of Inflicting Capital Punishment, 39 Ohio St. L. J. 96, 120 (1978). But while hanging could and often did result in significant pain, its use “was virtually never questioned.” Banner 170. Presumably that was because, in contrast to punishments like burning and disemboweling, hanging wasn’t “intended to be painful” and the risk of pain involved was considered “unfortunate but inevitable.” Ibid.; see also id., at 48. What does all this tell us about how the Eighth Amendment applies to methods of execution? For one thing, it tells us that the Eighth Amendment does not guarantee a prisoner a painless death—something that, of course, isn’t guaranteed to many people, including most victims of capital crimes. Glossip, 576 U. S., at ___ (slip op., at 4). Instead, what unites the punishments the Eighth Amendment was understood to forbid, and distinguishes them from those it was understood to allow, is that the former were long disused (unusual) forms of punishment that intensified the sentence of death with a (cruel) “ ‘superadd[ition]’ ” of “ ‘terror, pain, or disgrace.’ ” Baze, 553 U. S., at 48; accord, id., at 96 (Thomas, J., concurring in judgment). This Court has yet to hold that a State’s method of execution qualifies as cruel and unusual, and perhaps understandably so. Far from seeking to superadd terror, pain, or disgrace to their executions, the States have often sought more nearly the opposite, exactly as Justice Story predicted. Through much of the 19th century, States experimented with technological innovations aimed at making hanging less painful. See Banner 170–177. In the 1880s, following the recommendation of a commission tasked with finding “ ‘the most humane and practical method known to modern science of carrying into effect the sentence of death,’ ” the State of New York replaced hanging with electrocution. Glossip, 576 U. S., at ___ (slip op., at 2). Several States followed suit in the “ ‘ “belief that electrocution is less painful and more humane than hanging.” ’ ” Ibid. Other States adopted lethal gas after concluding it was “ ‘the most humane [method of execution] known to modern science.’ ” Ibid. And beginning in the 1970s, the search for less painful modes of execution led many States to switch to lethal injection. Id., at ___ (slip op., at 3); Baze, 553 U. S., at 42, 62; see also Banner 178–181, 196–197, 297. Notably, all of these innovations occurred not through this Court’s intervention, but through the initiative of the people and their representatives. Still, accepting the possibility that a State might try to carry out an execution in an impermissibly cruel and unusual manner, how can a court determine when a State has crossed the line? The Chief Justice’s opinion in Baze, which a majority of the Court held to be controlling in Glossip, supplies critical guidance. It teaches that where (as here) the question in dispute is whether the State’s chosen method of execution cruelly superadds pain to the death sentence, a prisoner must show a feasible and readily implemented alternative method of execution that would significantly reduce a substantial risk of severe pain and that the State has refused to adopt without a legitimate penological reason. See Glossip, 576 U. S., at ___–___ (slip op., at 12–13); Baze, 553 U. S., at 52. Glossip left no doubt that this standard governs “all Eighth Amendment method-of-execution claims.” 576 U. S., at ___ (slip op., at 1). In reaching this conclusion, Baze and Glossip recognized that the Eighth Amendment “does not demand the avoidance of all risk of pain in carrying out executions.” Baze, 553 U. S., at 47. To the contrary, the Constitution affords a “measure of deference to a State’s choice of execution procedures” and does not authorize courts to serve as “boards of inquiry charged with determining ‘best practices’ for executions.” Id., at 51–52, and nn. 2–3. The Eighth Amendment does not come into play unless the risk of pain associated with the State’s method is “substantial when compared to a known and available alternative.” Glossip, 576 U. S., at ___ (slip op., at 13); see Baze, 553 U. S., at 61. Nor do Baze and Glossip suggest that traditionally accepted methods of execution—such as hanging, the firing squad, electrocution, and lethal injection—are necessarily rendered unconstitutional as soon as an arguably more humane method like lethal injection becomes available. There are, the Court recognized, many legitimate reasons why a State might choose, consistent with the Eighth Amendment, not to adopt a prisoner’s preferred method of execution. See, e.g., Glossip, 576 U. S., at ___–___ (slip op., at 13–14) (a State can’t be faulted for failing to use lethal injection drugs that it’s unable to procure through good-faith efforts); Baze, 553 U. S., at 57 (a State has a legitimate interest in selecting a method it regards as “preserving the dignity of the procedure”); id., at 66 (Alito, J., concurring) (a State isn’t required to modify its protocol in ways that would require the involvement of “persons whose professional ethics rules or traditions impede their participation”). As we’ve seen, two Members of the Court whose votes were essential to the judgment in Glossip argued that establishing cruelty consistent with the Eighth Amendment’s original meaning demands slightly more than the majority opinion there (or the Baze plurality opinion it followed) suggested. Instead of requiring an inmate to establish that a State has unreasonably refused to alter its method of execution to avoid a risk of unnecessary pain, Justice Thomas and Justice Scalia contended that an inmate must show that the State intended its method to inflict such pain. See Glossip, 576 U. S., at ___ (Thomas, J., concurring) (slip op., at 1); Baze, 553 U. S., at 94–107 (Thomas, J., concurring in judgment). But revisiting that debate isn’t necessary here because, as we’ll see, the State was entitled to summary judgment in this case even under the more forgiving Baze-Glossip test. See Part III, infra. B Before turning to the application of Baze and Glossip, however, we must confront Mr. Bucklew’s argument that a different standard entirely should govern as-applied challenges like his. He admits that Baze and Glossip supply the controlling test in facial challenges to a State’s chosen method of execution. But he suggests that he should not have to prove an alternative method of execution in his as-applied challenge because “certain categories” of punishment are “manifestly cruel . . . without reference to any alternative methods.” Brief for Petitioner 41–42 (internal quotation marks omitted). He points to “ ‘burning at the stake, crucifixion, [and] breaking on the wheel’ ” as examples of “categorically” cruel methods. Ibid. And, he says, we should use this case to add to the list of “categorically” cruel methods any method that, as applied to a particular inmate, will pose a “substantial and particular risk of grave suffering” due to the inmate’s “unique medical condition.” Id., at 44. The first problem with this argument is that it’s foreclosed by precedent. Glossip expressly held that identifying an available alternative is “a requirement of all Eighth Amendment method-of-execution claims” alleging cruel pain. 576 U. S., at ___ (slip op., at 1) (emphasis added). And just as binding as this holding is the reasoning underlying it. Distinguishing between constitutionally permissible and impermissible degrees of pain, Baze and Glossip explained, is a necessarily comparative exercise. To decide whether the State has cruelly “superadded” pain to the punishment of death isn’t something that can be accomplished by examining the State’s proposed method in a vacuum, but only by “compar[ing]” that method with a viable alternative. Glossip, 576 U. S., at ___ (slip op., at 13); see Baze, 553 U. S., at 61. As Mr. Bucklew acknowledges when speaking of facial challenges, this comparison “provides the needed metric” to measure whether the State is lawfully carrying out an execution or inflicting “gratuitous” pain. Brief for Petitioner 42–43. Yet it is that very comparison and needed metric Mr. Bucklew would now have us discard. Nor does he offer some persuasive reason for overturning our precedent. To the contrary, Mr. Bucklew simply repeats the same argument the principal dissent offered and the Court expressly and thoughtfully rejected in Glossip. Just as Mr. Bucklew argues here, the dissent there argued that “certain methods of execution” like “burning at the stake” should be declared “categorically off-limits.” And just as Mr. Bucklew submits here, the dissent there argued that any other “intolerably painful” method of execution should be added to this list. 576 U. S., at ___–___ (Sotomayor, J., dissenting) (slip op., at 23–24). Mr. Bucklew’s submission, thus, amounts to no more than a headlong attack on precedent. Mr. Bucklew’s argument fails for another independent reason: It is inconsistent with the original and historical understanding of the Eighth Amendment on which Baze and Glossip rest. As we’ve seen, when it comes to determining whether a punishment is unconstitutionally cruel because of the pain involved, the law has always asked whether the punishment “superadds” pain well beyond what’s needed to effectuate a death sentence. And answering that question has always involved a comparison with available alternatives, not some abstract exercise in “categorical” classification. At common law, the ancient and barbaric methods of execution Mr. Bucklew cites were understood to be cruel precisely because—by comparison to other available methods—they went so far beyond what was needed to carry out a death sentence that they could only be explained as reflecting the infliction of pain for pain’s sake. Meanwhile, hanging carried with it an acknowledged and substantial risk of pain but was not considered cruel because that risk was thought—by comparison to other known methods—to involve no more pain than was reasonably necessary to impose a lawful death sentence. See supra, at 9–12. What does the principal dissent have to say about all this? It acknowledges that Glossip’s comparative requirement helps prevent facial method-of-execution claims from becoming a “backdoor means to abolish” the death penalty. Post, at 8 (opinion of Breyer, J.). But, the dissent assures us, there’s no reason to worry that as-applied method-of-execution challenges might be used that way. This assurance misses the point. As we’ve explained, the alternative-method requirement is compelled by our understanding of the Constitution, not by mere policy concerns. With that, the dissent is left only to rehash the same argument that Mr. Bucklew offers. The dissent insists that some forms of execution are just categorically cruel. Post, at 10–11. At first and like others who have made this argument, the dissent offers little more than intuition to support its conclusion. Ultimately, though, even it bows to the necessity of something firmer. If a “comparator is needed” to assess whether an execution is cruel, the dissent tells us, we should compare the pain likely to follow from the use of a lethal injection in this case with the pain-free use of lethal injections in mine-run cases. Post, at 10. But that’s just another way of saying executions must always be carried out painlessly because they can be carried out painlessly most of the time, a standard the Constitution has never required and this Court has re- jected time and time again. Supra, at 12. To determine whether the State is cruelly superadding pain, our precedents and history require asking whether the State had some other feasible and readily available method to carry out its lawful sentence that would have significantly reduced a substantial risk of pain. That Mr. Bucklew and the dissent fail to respect the force of our precedents—or to grapple with the understanding of the Constitution on which our precedents rest—is more than enough reason to reject their view that as-applied and facial challenges should be treated differently. But it turns out their position on this score suffers from further problems too—problems that neither Mr. Bucklew nor the dissent even attempts to address. Take this one. A facial challenge is really just a claim that the law or policy at issue is unconstitutional in all its applications. So classifying a lawsuit as facial or as-applied affects the extent to which the invalidity of the challenged law must be demonstrated and the corresponding “breadth of the remedy,” but it does not speak at all to the substantive rule of law necessary to establish a constitutional violation. Citizens United v. Federal Election Comm’n, 558 U.S. 310, 331 (2010). Surely it would be strange for the same words of the Constitution to bear entirely different meanings depending only on how broad a remedy the plaintiff chooses to seek. See Gross v. United States, 771 F.3d 10, 14–15 (CADC 2014) (“ ‘[T]he substantive rule of law is the same for both [facial and as-applied] challenges’ ”); Brooklyn Legal Servs. Corp. v. Legal Servs. Corp., 462 F.3d 219, 228 (CA2 2006) (the facial/as-applied distinction affects “the extent to which the invalidity of a statute need be demonstrated,” not “the substantive rule of law to be used”). And surely, too, it must count for something that we have found not a single court decision in over 200 years suggesting that the Eighth Amendment’s meaning shifts in this way. To the contrary, our precedent suggests just the opposite. In the related context of an Eighth Amendment challenge to conditions of confinement, we have seen “no basis whatever” for applying a different legal standard to “deprivations inflicted upon all prisoners” and those “inflicted upon particular prisoners.” Wilson v. Seiter, 501 U.S. 294, 299, n. 1 (1991). Here’s yet another problem with Mr. Bucklew’s argument: It invites pleading games. The line between facial and as-applied challenges can sometimes prove “amorphous,” Elgin v. Department of Treasury, 567 U.S. 1, 15 (2012), and “not so well defined,” Citizens United, 558 U. S., at 331. Consider an example. Suppose an inmate claims that the State’s lethal injection protocol violates the Eighth Amendment when used to execute anyone with a very common but not quite universal health condition. Should such a claim be regarded as facial or as-applied? In another context, we sidestepped a debate over how to categorize a comparable claim—one that neither sought “to strike [the challenged law] in all its applications” nor was “limited to plaintiff’s particular case”—by concluding that “[t]he label is not what matters.” Doe v. Reed, 561 U.S. 186, 194 (2010). To hold now, for the first time, that choosing a label changes the meaning of the Constitution would only guarantee a good deal of litigation over labels, with lawyers on each side seeking to classify cases to maximize their tactical advantage. Unless increasing the delay and cost involved in carrying out executions is the point of the exercise, it’s hard to see the benefit in placing so much weight on what can be an abstruse exercise. Finally, the burden Mr. Bucklew must shoulder under the Baze-Glossip test can be overstated. An inmate seeking to identify an alternative method of execution is not limited to choosing among those presently authorized by a particular State’s law. Missouri itself seemed to acknowledge as much at oral argument. Tr. of Oral Arg. 65. So, for example, a prisoner may point to a well-established protocol in another State as a potentially viable option. Of course, in a case like that a court would have to inquire into the possibility that one State possessed a legitimate reason for declining to adopt the protocol of another. See supra, at 13–14. And existing state law might be relevant to determining the proper procedural vehicle for the inmate’s claim. See Hill v. McDonough, 547 U.S. 573, 582–583 (2006) (if the relief sought in a 42 U. S. C. §1983 action would “foreclose the State from implementing the [inmate’s] sentence under present law,” then “recharacterizing a complaint as an action for habeas corpus might be proper”). But the Eighth Amendment is the supreme law of the land, and the comparative assessment it requires can’t be controlled by the State’s choice of which methods to authorize in its statutes. In light of this, we see little likelihood that an inmate facing a serious risk of pain will be unable to identify an available alternative—assuming, of course, that the inmate is more interested in avoiding unnecessary pain than in delaying his execution. III Having (re)confirmed that anyone bringing a method of execution claim alleging the infliction of unconstitution- ally cruel pain must meet the Baze-Glossip test, we can now turn to the question whether Mr. Bucklew is able to sat- isfy that test. Has he identified a feasible and readily im- plemented alternative method of execution the State refused to adopt without a legitimate reason, even though it would significantly reduce a substantial risk of severe pain? Because the case comes to us after the entry of summary judgment, this appeal turns on whether Mr. Bucklew has shown a genuine issue of material fact warranting a trial. A We begin with the question of a proposed alternative method. Through much of this case and despite many opportunities, Mr. Bucklew refused to identify any alternative method of execution, choosing instead to stand on his argument that Baze and Glossip’s legal standard doesn’t govern as-applied challenges like his (even after the Eighth Circuit rejected that argument). Only when the district court warned that his continued refusal to abide this Court’s precedents would result in immediate dismissal did Mr. Bucklew finally point to nitrogen hy- poxia. The district court then afforded Mr. Bucklew “exten- sive discovery” to explore the viability of that alternative. 883 F. 3d, at 1094. But even after all that, we conclude Mr. Bucklew has failed for two independent reasons to present a triable question on the viability of nitrogen hypoxia as an alternative to the State’s lethal injection protocol. First, an inmate must show that his proposed alternative method is not just theoretically “ ‘feasible’ ” but also “ ‘readily implemented.’ ” Glossip, 576 U. S., at ___–___ (slip op., at 12–13). This means the inmate’s proposal must be sufficiently detailed to permit a finding that the State could carry it out “relatively easily and reasonably quickly.” McGehee v. Hutchinson, 854 F.3d 488, 493 (CA8 2017); Arthur v. Commissioner, Ala. Dept. of Corrections, 840 F.3d 1268, 1300 (CA11 2016). Mr. Bucklew’s bare-bones proposal falls well short of that standard. He has presented no evidence on essential questions like how nitrogen gas should be administered (using a gas chamber, a tent, a hood, a mask, or some other delivery device); in what concentration (pure nitrogen or some mixture of gases); how quickly and for how long it should be introduced; or how the State might ensure the safety of the execution team, including protecting them against the risk of gas leaks. Instead of presenting the State with a read- ily implemented alternative method, Mr. Bucklew (and the principal dissent) point to reports from correctional authorities in other States indicating that additional study is needed to develop a protocol for execution by nitrogen hypoxia. See App. 697 (Oklahoma grand jury report recommending that the State “retain experts” and conduct “further research” to “determine how to carry out the sentence of death by this method”); id., at 736 (report of Louisiana Dept. of Public Safety & Corrections stating that “[r]esearch . . . is ongoing” to develop a nitrogen hypoxia protocol). That is a proposal for more research, not the readily implemented alternative that Baze and Glossip require. Second, and relatedly, the State had a “legitimate” reason for declining to switch from its current method of execution as a matter of law. Baze, 553 U. S., at 52. Rather than point to a proven alternative method, Mr. Bucklew sought the adoption of an entirely new method—one that had “never been used to carry out an execution” and had “no track record of successful use.” McGehee, 854 F. 3d, at 493. But choosing not to be the first to experiment with a new method of execution is a legitimate reason to reject it. In Baze we observed that “no other State ha[d] adopted” the one-drug protocol the inmates sought and they had “proffered no study showing” their one-drug protocol would be as effective and humane as the State’s existing three-drug protocol. 553 U. S., at 57. Under those circumstances, we held as a matter of law that Kentucky’s refusal to adopt the inmates’ proffered protocol could not “constitute a violation of the Eighth Amendment.” Ibid. The Eighth Amendment prohibits States from dredging up archaic cruel punishments or perhaps inventing new ones, but it does not compel a State to adopt “untried and untested” (and thus unusual in the constitutional sense) methods of execution. Id., at 41.[1] B Even if a prisoner can carry his burden of showing a readily available alternative, he must still show that it would significantly reduce a substantial risk of severe pain. Glossip, 576 U. S., at ___ (slip op., at 13); Baze, 553 U. S., at 52. A minor reduction in risk is insufficient; the difference must be clear and considerable. Over the course of this litigation, Mr. Bucklew’s explanation why nitrogen hypoxia meets this standard has evolved significantly. But neither of the two theories he has advanced in this Court turns out to be supported by record evidence. First, Mr. Bucklew points to several risks that he alleges could result from use of the State’s lethal injection protocol that would not be present if the State used nitrogen gas. For example, he says the execution team might try to insert an IV into one of his peripheral veins, which could cause the vein to rupture; or the team might instead use an allegedly painful “cut-down” procedure to access his femoral vein. He also says that he might be forced to lie flat on his back during the execution, which could impair his breathing even before the pentobarbital is administered. And he says the stress from all this could cause his tumors to bleed, further impairing his breathing. These risks, we may assume, would not exist if Mr. Bucklew were exe- cuted by his preferred method of nitrogen hypoxia. The problem with all of these contentions is that they rest on speculation unsupported, if not affirmatively contradicted, by the evidence in this case. Nor does the principal dissent contend otherwise. So, for example, uncontroverted record evidence indicates that the execution team will have discretion to adjust the gurney to whatever position is in Mr. Bucklew’s best medical interests. 883 F. 3d, at 1092, n. 3; App. 531. Moreover, the State agreed in the district court that it would not try to place an IV in Mr. Bucklew’s compromised peripheral veins. Id., at 820; see Brief for Appellant in No. 17–3052 (CA8), p. 7. And, assuming without granting that using a cut-down would raise issues under the Eighth Amendment—but see Nooner v. Norris, 594 F.3d 592, 604 (CA8 2010) (holding otherwise)—the State’s expert, Dr. Michael Antognini, testified without contradiction that it should be possible to place an IV in Mr. Bucklew’s femoral vein without using a cut-down procedure, App. 350. Mr. Bucklew responds by pointing to the warden’s testimony that he once saw medical staff perform a cut-down as part of an execution; but there’s no evidence that what the warden saw was an attempt to access a femoral vein, as opposed to some other vein. Moreover, to the extent the record is unclear on any of these issues, Mr. Bucklew had ample opportunity to conduct discovery and develop a factual record concerning exactly what procedures the State planned to use. He failed to do so—presumably because the thrust of his constitutional claim was that any attempt to execute him via lethal injection would be unconstitutional, regardless of the specific procedures the State might use. As the court of appeals explained: “Having taken the position that any lethal injection procedure would violate the Eighth Amendment,” Mr. Bucklew “made no effort to determine what changes, if any, the [State] would make in applying its lethal injection protocol” to him, and he “never urged the district court to establish a suitable fact-finding procedure . . . to define the as-applied lethal injection protocol [the State] intends to use.” 883 F. 3d, at 1095–1096.[2] Second, Mr. Bucklew contends that the lethal injection itself will expose him to a substantial risk of severe pain that could be eliminated by adopting his preferred method. He claims that once the sedative pentobarbital is injected he will “lose the ability to manage” the tumors in his airway and, as a result, will experience a “sense of suffocation” for some period of time before the State’s sedative renders him fully unconscious. Brief for Petitioner 12–13. “It is during this in-between twilight stage,” according to his expert, Dr. Zivot, “that Mr. Bucklew is likely to experience prolonged feelings of suffocation and excruciating pain.” App. 234. Mr. Bucklew admits that similar feelings of suffocation could occur with nitrogen, the only difference being the potential duration of the so-called “twilight stage.” He contends that with nitrogen the stage would last at most 20 to 30 seconds, while with pentobarbital it could last up to several minutes. But here again the record contains insufficient evidence to permit Mr. Bucklew to avoid summary judgment. For starters, in the courts below Mr. Bucklew maintained he would have trouble managing his airway only if he were forced to lie supine, which (as we’ve explained) the evidence shows he won’t be. (The dissenters don’t address this point.) But even indulging his new claim that he will have this difficulty regardless of position, he still has failed to present colorable evidence that nitrogen would significantly reduce his risk of pain. We can assume for argument’s sake that Mr. Bucklew is correct that with nitrogen the twilight stage would last 20 to 30 seconds. The critical question, then, is how long that period might last with pentobarbital. The State’s expert, Dr. Antognini, testified that pentobarbital, too, would render Mr. Bucklew fully unconscious and incapable of experiencing pain within 20 to 30 seconds. Id., at 299–301, 432–433. Dr. Zivot disagreed; but when he was asked how long he thought the twilight stage would last with pentobarbital, his testimony was evasive. Eventually, he said his “number would be longer than” 20 to 30 seconds, but he declined to say how much longer. Id., at 195. Instead, he referenced a 2015 study on euthanasia in horses. He said the study found that when horses were given a large dose of pentobarbital (along with other drugs), they exhibited “isoelectric EEG”—a complete absence of detectable brain activity—after 52 to 240 seconds. Id., at 194–196. The district court assumed Dr. Zivot meant that “pain might be felt until measurable brain activity ceases” and that, extrapolating from the horse study, it might take up to four minutes for pentobarbital to “induc[e] a state in which [Mr. Bucklew] could no longer sense that he is choking or unable to breathe.” The district court acknowledged, however, that this might be “a generous interpretation of Dr. Zivot’s testimony.” Id., at 822, and n. 5. In fact, there’s nothing in the record to suggest that Mr. Bucklew will be capable of experiencing pain for significantly more than 20 to 30 seconds after being injected with pentobarbital. For one thing, Mr. Bucklew’s lawyer now admits that Dr. Zivot “crossed up the numbers” from the horse study. Tr. of Oral Arg. 7–8, 11–12. The study actually reported that the horses displayed isoelectric EEG between 2 and 52 seconds after infusion of pentobarbital was completed, with an average time of less than 24 seconds. App. 267. So if anything, the horse study appears to bolster Dr. Antognini’s time estimate. For another thing, everyone now also seems to acknowledge that isoelectric EEG is the wrong measure. Dr. Zivot never claimed the horses were capable of experiencing pain until they reached isoelectric EEG. And Mr. Bucklew’s lawyer now concedes that doctors perform major surgery on human patients with measurable EEG readings, which strongly suggests that Mr. Bucklew will be insensible to pain before reaching isoelectric EEG. Tr. of Oral Arg. 9. Finally, the record evidence even allows the possibility that nitrogen could increase the risk of pain. Because Dr. Zivot declined to testify about the likely effects of nitrogen gas, Mr. Bucklew must rely on Dr. Antognini’s testimony. And while Dr. Antognini did say he thought nitrogen’s “onset of action” could be “relatively fast,” App. 458, he added that the effects of nitrogen could vary depending on exactly how it would be administered—information Mr. Bucklew hadn’t provided. Indeed, he stated that “depending on . . . how it’s used, you might get more suffering from nitrogen gas than you would have” from the State’s current protocol. Id., at 460–461. Of course, the principal dissent maintains that Dr. Zivot’s testimony supports an inference that pentobarbital might cause Mr. Bucklew to suffer for a prolonged period. But its argument rests on a number of mistakes about the record. For example, the dissent points to Dr. Zivot’s remark that, with pentobarbital, “ ‘the period of time between receiving the injection and death could range over a few minutes to many minutes.’ ” Post, at 4, 6 (quoting App. 222). From this, the dissent concludes that Mr. Bucklew may suffer for “up to several minutes.” Post, at 1, 6, 9. But everyone agrees that the relevant question isn’t how long it will take for Mr. Bucklew to die, but how long he will be capable of feeling pain. Seeking to address the problem, the dissent next points to another part of Dr. Zivot’s testimony and says it means Mr. Bucklew could experience pain during the entire time between injection and death. Post, at 6, 13 (quoting App. 222). But the dissent clips the relevant quotation. As the full quotation makes clear, Dr. Zivot claimed that Mr. Bucklew might be unable to “maintain the integrity of his airway” until he died—but he carefully avoided claiming that Mr. Bucklew would be capable of feeling pain until he died.[3] To avoid this problem, the dissent quotes Dr. Zivot’s assertions that pentobarbital might not produce “ ‘rapid unconsciousness’ ” and that Mr. Bucklew’s suffering with pentobarbital could be “ ‘prolonged.’ ” Post, at 4–6, 13 (quoting App. 233–234). But Dr. Zivot’s statements here, too, fail to specify how long Mr. Bucklew is likely to be able to feel pain. The hard fact is that, when Dr. Zivot was finally compelled to offer a view on this question, his only response was to refer to the horse study. Id., at 195–196. The dissent’s effort to suggest that Dr. Zivot “did not rely exclusively or even heavily on that study,” post, at 7, is belied by (among other things) Mr. Bucklew’s own brief in this Court, which asserted that the twilight stage during which he might feel pain could last “between 52 and 240 seconds,” based entirely on a citation of Dr. Zivot’s incorrect testimony about the horse study. Brief for Petitioner 13. In sum, even if execution by nitrogen hypoxia were a feasible and readily implemented alternative to the State’s chosen method, Mr. Bucklew has still failed to present any evidence suggesting that it would significantly reduce his risk of pain. For that reason as well, the State was entitled to summary judgment on Mr. Bucklew’s Eighth Amendment claim.[4] IV “Both the State and the victims of crime have an important interest in the timely enforcement of a sentence.” Hill, 547 U. S., at 584. Those interests have been frustrated in this case. Mr. Bucklew committed his crimes more than two decades ago. He exhausted his appeal and separate state and federal habeas challenges more than a decade ago. Yet since then he has managed to secure delay through lawsuit after lawsuit. He filed his current challenge just days before his scheduled execution. That suit has now carried on for five years and yielded two appeals to the Eighth Circuit, two 11th-hour stays of execution, and plenary consideration in this Court. And despite all this, his suit in the end amounts to little more than an attack on settled precedent, lacking enough evidence even to survive summary judgment—and on not just one but many essential legal elements set forth in our case law and required by the Constitution’s original meaning. The people of Missouri, the surviving victims of Mr. Bucklew’s crimes, and others like them deserve better. Even the principal dissent acknowledges that “the long delays that now typically occur between the time an offender is sentenced to death and his execution” are “excessive.” Post, at 16. The answer is not, as the dissent incongruously suggests, to reward those who interpose delay with a decree ending capital punishment by judicial fiat. Post, at 18. Under our Constitution, the question of capital punishment belongs to the people and their representatives, not the courts, to resolve. The proper role of courts is to ensure that method-of-execution challenges to law- fully issued sentences are resolved fairly and expeditiously. Courts should police carefully against attempts to use such challenges as tools to interpose unjustified delay. Last-minute stays should be the extreme exception, not the norm, and “the last-minute nature of an application” that “could have been brought” earlier, or “an applicant’s attempt at manipulation,” “may be grounds for denial of a stay.” Hill, 547 U. S., at 584 (internal quotation marks omitted). So, for example, we have vacated a stay entered by a lower court as an abuse of discretion where the inmate waited to bring an available claim until just 10 days before his scheduled execution for a murder he had committed 24 years earlier. See Dunn v. Ray, 586 U. S. ___ (2019).[5] If litigation is allowed to proceed, federal courts “can and should” protect settled state judgments from “undue interference” by invoking their “equitable powers” to dismiss or curtail suits that are pursued in a “dilatory” fashion or based on “speculative” theories. Id., at 584–585. * The judgment of the court of appeals is Affirmed. Notes 1 While this case has been pending, a few States have authorized nitrogen hypoxia as a method of execution. See 2018 Ala. Acts no. 2018–353 (allowing condemned inmates to elect execution by nitrogen hypoxia); 2017 Miss. Laws ch. 406, p. 905 (authorizing execution by nitrogen hypoxia only if lethal injection is held unconstitutional or is otherwise unavailable); 2015 Okla. Sess. Laws ch. 75, p. 244 (same). In March 2018, officials in Oklahoma announced that, due to the unavailability of lethal injection drugs, the State would use nitrogen gas for its executions going forward. See Williams, Oklahoma Proposes To Use Nitrogen Gas for Executions by Asphyxiation, N. Y. Times, Mar. 15, 2018, p. A22. But Oklahoma has so far been unable to find a manufacturer willing to sell it a gas delivery device for use in executions. See Clay, State Not Ready for Executions, The Oklahoman, Jan. 27, 2019, p. A1. To date, no one in this case has pointed us to an execution in this country using nitrogen gas. 2 While the district court allowed discovery on many other matters, Mr. Bucklew protests that it did not permit him to learn the identities of the lethal injection execution team members, to depose them, or to inquire into their qualifications, training, and experience. Like the Eighth Circuit, we see no abuse of discretion in the district court’s discovery rulings. As the district court explained, Mr. Bucklew argues that there is no way he may be constitutionally executed by lethal injection, even with modifications to the State’s lethal injection protocol. And in a case like that, discovery into such granular matters as who administers the protocol simply is not relevant. 3 Here’s the full quotation, with the portion quoted by the dissent underlined: “As a result of his inability to maintain the integrity of his airway for the period of time beginning with the injection of the Pentobarbital solution and ending with Mr. Bucklew’s death several minutes to as long as many minutes later, Mr. Bucklew would be highly likely to experience feelings of ‘air hunger’ and the excruciating pain of prolonged suffocation resulting from the complete obstruction of his airway by the large vascular tumor.” App. 222. 4 The State contends that Mr. Bucklew’s claim should fail for yet another reason: because, in the State’s view, the evidence does not show that he is very likely to suffer “ ‘severe pain’ ” cognizable under the Eighth Amendment. Glossip v. Gross, 576 U. S. ___, ___ (2015) (slip op., at 13) (quoting Baze v. Rees, 553 U.S. 35, 52 (2008); emphasis added). We have no need, however, to address that argument because (as explained above) Mr. Bucklew fails even to show that a feasible and readily available alternative could significantly reduce the pain he alleges. 5 Seeking to relitigate Dunn v. Ray, the principal dissent asserts that that case involved no undue delay because the inmate “brought his claim only five days after he was notified” that the State would not allow his spiritual adviser to be present with him in the execution chamber itself, although it would allow the adviser to be present on the other side of a glass partition. Post, at 17. But a state statute listed “[t]he spiritual adviser of the condemned” as one of numerous individ-uals who would be allowed to “be present at an execution,” many of whom—such as “newspaper reporters,” “relatives or friends of the condemned person,” and “the victim’s immediate family members”—obviously would not be allowed into the chamber itself. Ala. Code §15–18–83 (2018). The inmate thus had long been on notice that there was a question whether his adviser would be allowed into the chamber or required to remain on the other side of the glass. Yet although he had been on death row since 1999, and the State had set a date for his execution on November 6, 2018, he waited until January 23, 2019—just 15 days before the execution—to ask for clarification. He then brought a claim 10 days before the execution and sought an indefinite stay. This delay implicated the “strong equitable presumption” that no stay should be granted “where a claim could have been brought at such a time as to allow consideration of the merits without requiring entry of a stay.” Hill v. McDonough, 547 U.S. 573, 584 (2006).

587.US.2018_18-315

The False Claims Act permits a private person, known as a relator, to bring a qui tam civil action “in the name of the [Federal] Government,” 31 U. S. C. §3730(b), against “any person” who “knowingly presents . . . a false or fraudulent claim for payment” to the Government or to certain third parties acting on the Government’s behalf, §§3729(a), (b)(2). The Government may choose to intervene in the action. See §§3730(b)(2), (4). Two limitations periods apply to a “civil action under section 3730.” §3731(b). An action must be brought within either 6 years after the statutory violation occurred, §3731(b)(1), or 3 years after the “the official of the United States charged with responsibility to act in the circumstances” knew or should have known the relevant facts, but not more than 10 years after the violation, §3731(b)(2). The period providing the later date serves as the limitations period. In November 2013, respondent Hunt filed a complaint alleging that petitioners—two defense contractors (collectively, Cochise)—defrauded the Government by submitting false payment claims for providing security services in Iraq up until early 2007. Hunt claims that he revealed Cochise’s allegedly fraudulent scheme during a November 30, 2010, interview with federal officials about his role in an unrelated contracting fraud in Iraq. The United States declined to intervene in the action, and Cochise moved to dismiss the complaint as barred by the statute of limitations. Hunt countered that his complaint was timely under §3731(b)(2). In dismissing the action, the District Court considered three potential interpretations: that §3731(b)(2) does not apply to a relator-initiated action in which the Government elects not to intervene; that §3731(b)(2) applies in nonintervened actions, and the limitations period begins when the relator knew or should have known the relevant facts; or that §3731(b)(2) applies in nonintervened actions, and the limitations period begins when the Government official responsible for acting knew or should have known the relevant facts. The court rejected the third interpretation and found that Hunt’s complaint would be untimely under either of the first two. The Eleventh Circuit reversed and remanded, adopting the third interpretation. Held: 1. The limitations period in §3731(b)(2) applies in a relator-initiated suit in which the Government has declined to intervene. Both Government-initiated suits under §3730(a) and relator-initiated suits under §3730(b) are “civil action[s] under section 3730.” Thus, the plain text of the statute makes the two limitations periods applicable in both types of suits. Cochise claims that starting a limitations period when the party entitled to bring a claim learns the relevant facts is a default rule of tolling provisions, so subsection (b)(2) should apply only when the Government is a party. But treating a relator-initiated, nonintervened suit as a “civil action under section 3730” for purposes of subsection (b)(1) but not subsection (b)(2) is at odds with fundamental rules of statutory interpretation. Because a single use of a statutory phrase generally must have a fixed meaning, see Ratzlaf v. United States, 510 U.S. 135, 143, interpretations that would “attribute different meanings to the same phrase” should be avoided, Reno v. Bossier Parish School Bd., 528 U.S. 320, 329. Here, the clear text of the statute controls. Cochise’s reliance on Graham County Soil & Water Conservation Dist. v. United States ex rel. Wilson, 545 U.S. 409, is misplaced. Nothing in Graham County supports giving the phrase “civil action under section 3730” in §3731(b) two different meanings depending on whether the Government intervenes. While the Graham County Court sought “a construction that avoids . . . counterintuitive results,” there the text “admit[ted] of two plausible interpretations.” Id., at 421, 419, n. 2. Here, Cochise points to no other plausible interpretation of the text, so the “ ‘judicial inquiry is complete.’ ” Barnhart v. Sigmon Coal Co., 534 U.S. 438, 462. Pp. 4–8. 2. The relator in a nonintervened suit is not “the official of the United States” whose knowledge triggers §3731(b)(2)’s 3-year limitations period. The statute provides no support for such a reading. First, a private relator is neither appointed as an officer of the United States nor employed by the United States. Second, the provision authorizing qui tam suits is entitled “Actions by Private Persons.” §3730(b). Third, the statute refers to “the” official “charged with responsibility to act in the circumstances.” Regardless of precisely which official or officials the statute is referring to, §3731(b)(2)’s use of the definite article “the” suggests that Congress did not intend for private relators to be considered “the official of the United States.” See Rumsfeld v. Padilla, 542 U.S. 426, 434. Nor are private relators “charged with responsibility to act” in the sense contemplated by §3731(b), as they are not required to investigate or prosecute a False Claims Act action. Pp. 8–9. 887 F.3d 1081, affirmed. Thomas, J., delivered the opinion for a unanimous Court.

The False Claims Act contains two limitations periods that apply to a “civil action under section 3730”—that is, an action asserting that a person presented false claims to the United States Government. 31 U. S. C. §3731(b). The first period requires that the action be brought within 6 years after the statutory violation occurred. The second period requires that the action be brought within 3 years after the United States official charged with the responsibility to act knew or should have known the relevant facts, but not more than 10 years after the violation. Whichever period provides the later date serves as the limitations period. This case requires us to decide how to calculate the limitations period for qui tam suits in which the United States does not intervene. The Court of Appeals held that these suits are “civil action[s] under section 3730” and that the limitations periods in §3731(b) apply in accordance with their terms, regardless of whether the United States intervenes. It further held that, for purposes of the second period, the private person who initiates the qui tam suit cannot be deemed the official of the United States. We agree, and therefore affirm. I As relevant, the False Claims Act imposes civil liability on “any person” who “knowingly presents, or causes to be presented, a false or fraudulent claim for payment or approval” to the Government or to certain third parties acting on the Government’s behalf. 31 U. S. C. §§3729(a), (b)(2). Section 3730 authorizes two types of actions: First, the Attorney General, who “diligently shall investigate a violation under section 3729,” may bring a civil action against the alleged false claimant. §3730(a). Second, a private person, known as a relator, may bring a qui tam civil action “for the person and for the United States Government” against the alleged false claimant, “in the name of the Government.” §3730(b). If a relator initiates the action, he must deliver a copy of the complaint and supporting evidence to the Government, which then has 60 days to intervene in the action. §§3730(b)(2), (4). During this time, the complaint remains sealed. §3730(b)(2). If the Government intervenes, it assumes primary responsibility for prosecuting the action, though the relator may continue to participate. §3730(c). Otherwise, the relator has the right to pursue the action. §§3730(b)(4), (c)(3). Even if it does not intervene, the Government is entitled to be served with all pleadings upon request and may intervene at any time with good cause. §3730(c)(3). The relator receives a share of any proceeds from the action—generally 15 to 25 percent if the Government intervenes, and 25 to 30 percent if it does not—plus attorney’s fees and costs. §§3730(d)(1)–(2). See Vermont Agency of Natural Resources v. United States ex rel. Stevens, 529 U.S. 765, 769–770 (2000). At issue here is the Act’s statute of limitations, which provides: “(b) A civil action under section 3730 may not be brought— “(1) more than 6 years after the date on which the violation of section 3729 is committed, or “(2) more than 3 years after the date when facts material to the right of action are known or reasonably should have been known by the official of the United States charged with responsibility to act in the circumstances, but in no event more than 10 years after the date on which the violation is committed, “whichever occurs last.” §3731(b). On November 27, 2013, respondent Billy Joe Hunt filed a complaint alleging that petitioners—two defense contractors (collectively, Cochise)—defrauded the Government by submitting false claims for payment under a subcontract to provide security services in Iraq “from some time prior to January 2006 until early 2007.” App. 43a. A little less than three years before bringing his complaint, Hunt was interviewed by federal agents about his role in an unrelated contracting fraud in Iraq. Hunt claims to have revealed Cochise’s allegedly fraudulent scheme during this November 30, 2010, interview. The United States declined to intervene in Hunt’s action, and Cochise moved to dismiss the complaint as barred by the statute of limitations. Hunt conceded that the 6-year limitations period in §3731(b)(1) had elapsed before he filed suit on November 27, 2013. But Hunt argued that his complaint was timely under §3731(b)(2) because it was filed within 3 years of the interview in which he informed federal agents about the alleged fraud (and within 10 years after the violation occurred). The District Court dismissed the action. It considered three potential interpretations of §3731(b). Under the first interpretation, §3731(b)(2) does not apply to a relator-initiated action in which the Government elects not to intervene, so any such action must be filed within six years after the violation. Under the second interpretation, §3731(b)(2) applies in nonintervened actions, and the limitations period begins when the relator knew or should have known the relevant facts. Under the third interpretation, §3731(b)(2) applies in nonintervened actions, and the limitations period begins when “the official of the United States charged with responsibility to act in the circumstances” knew or should have known the relevant facts. The District Court rejected the third interpretation and declined to choose between the first two because it found that Hunt’s complaint would be untimely under either. The Court of Appeals reversed and remanded, adopting the third interpretation. 887 F.3d 1081 (CA11 2018). Given a conflict between the Courts of Appeals,[1]* we granted certiorari. 586 U. S. ___ (2018). II The first question before us is whether the limitations period in §3731(b)(2) is available in a relator-initiated suit in which the Government has declined to intervene. If so, the second question is whether the relator in such a case should be considered “the official of the United States” whose knowledge triggers §3731(b)(2)’s 3-year limitations period. A Section 3731(b) sets forth two limitations periods that apply to “civil action[s] under section 3730.” Both Government-initiated suits under §3730(a) and relator-initiated suits under §3730(b) are “civil action[s] under section 3730.” Thus, the plain text of the statute makes the two limitations periods applicable in both types of suits. Cochise agrees with that view as to the limitations period in §3731(b)(1), but argues that the period in §3731(b)(2) is available in a relator-initiated suit only if the Government intervenes. According to Cochise, starting a limitations period when the party entitled to bring a claim learns the relevant facts is a default rule of tolling provisions, so subsection (b)(2) should be read to apply only when the Government is a party. In short, under Cochise’s reading, a relator-initiated, nonintervened suit is a “civil action under section 3730” for purposes of subsection (b)(1) but not subsection (b)(2). This reading is at odds with fundamental rules of statutory interpretation. In all but the most unusual situations, a single use of a statutory phrase must have a fixed meaning. See Ratzlaf v. United States, 510 U.S. 135, 143 (1994). We therefore avoid interpretations that would “attribute different meanings to the same phrase.” Reno v. Bossier Parish School Bd., 528 U.S. 320, 329 (2000). Here, either a relator-initiated, nonintervened suit is a “civil action under section 3730”—and thus subject to the limitations periods in subsections (b)(1) and (b)(2)—or it is not. It is such an action. Whatever the default tolling rule might be, the clear text of the statute controls this case. Under Cochise’s reading, a relator-initiated civil action would convert to “[a] civil action under section 3730” for purposes of subsection (b)(2) if and when the Government intervenes. That reading cannot be correct. If the Government intervenes, the civil action remains the same—it simply has one additional party. There is no textual basis to base the meaning of “[a] civil action under section 3730” on whether the Government has intervened. Cochise relies on our decision in Graham County Soil & Water Conservation Dist. v. United States ex rel. Wilson, 545 U.S. 409 (2005), which addressed the question whether §3731(b)(1) or federal common law provided the limitations period for §3730(h) retaliation actions. Section 3730(h) creates a cause of action for an employee who suffers retaliation for, among other things, assisting with the prosecution of a False Claims Act action. At the time, §3730(h) did not specify a time limit for bringing a retaliation action, so the question before us was whether the phrase “civil action under section 3730” in §3731(b) encompassed actions under §3730(h). We considered the statute “ambiguous because its text, literally read, admits of two plausible interpretations.” Id., at 419, n. 2. One reading was that a “civil action under section 3730” includes §3730(h) actions because such actions arise under §3730. Id., at 415. “Another reasonable reading” was that a “civil action under section 3730” “applies only to actions arising under §§3730(a) and (b)” because “§3731(b)(1) t[ies] the start of the time limit to ‘the date on which the violation of section 3729 is committed.’ ” Ibid. That reading had force because retaliation claims need not involve an actual violation of §3729. Ibid. Looking to statutory context, we explained that the phrase “ ‘civil action under section 3730’ means only those civil actions under §3730 that have as an element a ‘violation of section 3729,’ that is, §§3730(a) and (b) actions”—not §3730(h) retaliation actions. Id., at 421–422. A relator-initiated, nonintervened suit arises under §3730(b) and has as an element a violation of §3729. Graham County supports our reading. Nonetheless, Cochise points out that in considering the statutory context, we discussed a similar phrase contained in §3731(c) (now §3731(d)), which stated: “In any action brought under section 3730, the United States shall be required to prove all essential elements of the cause of action, including damages, by a preponderance of the evidence.” (Emphasis added.) We explained that §3731(c) “use[d] the similarly unqualified phrase ‘action brought under section 3730’ to refer only to §§3730(a) and (b) actions.” Id., at 417–418. We then stated: “As [respondent] and the United States concede, the context of this provision implies that the phrase ‘any action brought under section 3730’ is limited to §3730(a) actions brought by the United States and §3730(b) actions in which the United States intervenes as a party, as those are the types of §3730 actions in which the United States necessarily participates.” Id., at 418. Cochise contends that we should adopt a similar construction of the phrase “civil action under section 3730” in §3731(b). We disagree. Our discussion of §3731(c) was focused on “the context of th[at] provision” and on whether it could be read to impose the burden of proof on the Government even in cases where the Government did not participate. Id., at 418. Those considerations do not apply here; there is nothing illogical about reading §3731(b) to apply in accordance with its plain terms. Moreover, if a “civil action under section 3730” included only an action in which the Government participates for purposes of §3731(b)(2), then we would be obligated to give it a like meaning for purposes of §3731(b)(1). This would mean that a relator-initiated, nonintervened suit would be subject to neither §3731(b)(1) nor §3731(b)(2)—a reading Cochise expressly disclaims. See Brief for Petitioners 20, n. 3. Nothing in Graham County supports giving the same phrase in §3731(b) two different meanings depending on whether the Government intervenes. Again pointing to Graham County, Cochise next contends that our reading would lead to “ ‘counterintuitive results.’ ” Brief for Petitioners 26. For instance, if the Government discovers the fraud on the day it occurred, it would have 6 years to bring suit, but if a relator instead discovers the fraud on the day it occurred and the Government does not discover it, the relator could have as many as 10 years to bring suit. That discrepancy arises because §3731(b)(2) begins its limitations period on the date that “the official of the United States charged with responsibility to act” obtained knowledge of the relevant facts. But we see nothing unusual about extending the limitations period when the Government official did not know and should not reasonably have known the relevant facts, given that the Government is the party harmed by the false claim and will receive the bulk of any recovery. See §3730(d). In any event, a result that “may seem odd . . . is not absurd.” Exxon Mobil Corp. v. Allapattah Services, Inc., 545 U.S. 546, 565 (2005). Although in Graham County we sought “a construction that avoids . . . counterintuitive results,” there the text “admit[ted] of two plausible interpretations.” 545 U. S., at 421, 419, n. 2. Here, Cochise points to no other plausible interpretation of the text, so the “ ‘judicial inquiry is complete.’ ” Barnhart v. Sigmon Coal Co., 534 U.S. 438, 462 (2002). B Cochise’s fallback argument is that the relator in a nonintervened suit should be considered “the official of the United States charged with responsibility to act in the circumstances,” meaning that §3731(b)(2)’s 3-year limitations period would start when the relator knew or should have known about the fraud. But the statute provides no support for reading “the official of the United States” to encompass a private relator. First, a private relator is not an “official of the United States” in the ordinary sense of that phrase. A relator is neither appointed as an officer of the United States, see U. S. Const., Art. II, §2, cl. 2, nor employed by the United States. Indeed, the provision that authorizes qui tam suits is entitled “Actions by Private Persons.” §3730(b). Although that provision explains that the action is brought “for the person and for the United States Government” and “in the name of the Government,” ibid., it does not make the relator anything other than a private person, much less “the official of the United States” referenced by the statute. Cf. Stevens, 529 U. S., at 773, n. 4 (“[A] qui tam relator is, in effect, suing as a partial as- signee of the United States” (emphasis deleted)). Second, the statute refers to “the” official “charged with responsibility to act in the circumstances.” The Government argues that, in context, “the” official refers to the Attorney General (or his delegate), who by statute “shall investigate a violation under section 3729.” §3730(a). Regardless of precisely which official or officials the statute is referring to, §3731(b)(2)’s use of the definite article “the” suggests that Congress did not intend for any and all private relators to be considered “the official of the United States.” See Rumsfeld v. Padilla, 542 U.S. 426, 434 (2004) (explaining that the “use of the definite article . . . indicates that there is generally only one” person covered). More fundamentally, private relators are not “charged with responsibility to act” in the sense contemplated by §3731(b), as they are not required to investigate or prosecute a False Claims Act action. * * * For the foregoing reasons, the judgment of the Court of Appeals is Affirmed. Notes 1 *Compare 887 F.3d 1081, 1089–1097 (CA11 2018) (adopting the third interpretation), with United States ex rel. Hyatt v. Northrop Corp., 91 F.3d 1211, 1216–1218 (CA9 1996) (adopting the second interpretation); United States ex rel. Sanders v. North Am. Bus Industries, Inc., 546 F.3d 288, 293–294 (CA4 2008) (adopting the first interpretation); and United States ex rel. Sikkenga v. Regence Bluecross Blueshield of Utah, 472 F.3d 702, 725–726 (CA10 2006) (same).

586.US.2018_17-773

The Social Security Act regulates the fees that attorneys may charge claimants seeking Title II benefits for representation both before the Social Security Administration and in federal court. For representation in administrative proceedings, the Act provides two ways to determine fees. If a fee agreement exists, fees are capped at the lesser of 25% of past-due benefits or a set dollar amount—currently $6,000. 42 U. S. C. §406(a)(2)(A). Absent an agreement, the agency may set any “reasonable” fee. §406(a)(1). In either case, the agency is required to withhold up to 25% of past-due benefits for direct payment of any fee. §406(a)(4). For representation in court proceedings, fees are capped at 25% of past-due benefits, and the agency has authority to withhold such benefits to pay these fees. §406(b)(1)(A). Petitioner Culbertson represented Katrina Wood in Social Security disability benefit proceedings before the agency and in District Court. The agency ultimately awarded Wood past-due benefits, withheld 25% of those benefits to pay any attorney’s fees, and awarded Culbertson fees under §406(a) for representation before the agency. Culbertson then moved for a separate fee award under §406(b) for the court proceedings, requesting a full 25% of past-due benefits. The District Court granted the request, but only in part, because Culbertson did not subtract the amount he had already received under §406(a) for his agency-level representation. The Eleventh Circuit affirmed, holding that the 25% limit under §406(b) applies to the total fees awarded under both §§406(a) and (b). Held: Section 406(b)(1)(A)’s 25% cap applies only to fees for court representation and not to the aggregate fees awarded under §§406(a) and (b). Pp. 5–9. (a) Section 406(b) provides that a court rendering a favorable judgment to a claimant “represented before the court by an attorney” may award “a reasonable fee for such representation, not in excess of 25 percent” of past-due benefits. Here, the adjective “such,” which means “[o]f the kind or degree already described or implied,” refers to the only form of representation “already described” in §406(b)—i.e., “represent[ation] before the court.” Thus, the 25% cap applies only to fees for representation before the court, not the agency. Subsections (a) and (b) address different stages of the representation and use different methods for calculating fees. Given this statutory structure, applying §406(b)’s 25% cap on court-stage fees to §406(a) agency-stage fees, or the aggregate of §§406(a) and (b) fees, would make little sense. For example, such a reading would subject §406(a)(1)’s reasonableness limitation to §406(b)’s 25% cap—a limitation not included in the relevant provision of the statute. Had Congress wanted agency-stage fees to be capped at 25%, it presumably would have said so directly in subsection (a). Pp. 5–7. (b) The fact that the agency presently withholds a single pool of 25% of past-due benefits for direct payment of agency and court fees does not support an aggregate reading. The statutory text provides for two pools of money for direct payment of fees. See §§406(a)(4), (b)(1)(A). The agency’s choice to withhold only one pool of 25% of past-due benefits does not alter this text. More fundamentally, the amount of past-due benefits that the agency can withhold for direct payment does not delimit the amount of fees that can be approved for representation before the agency or the court. Pp. 7–9. 861 F. 3d 1197, reversed and remanded. Thomas, J., delivered the opinion for a unanimous Court.

Federal law regulates the fees that attorneys may charge Social Security claimants for representation before the Social Security Administration and a reviewing court. See 42 U. S. C. §§406(a)–(b). The question in this case is whether the statutory scheme limits the aggregate amount of fees for both stages of representation to 25% of the claimant’s past-due benefits. Because §406(b) by its terms imposes a 25% cap on fees only for representation before a court, and §406(a) has separate caps on fees for representation before the agency, we hold that the statute does not impose a 25% cap on aggregate fees. I A Title II of the Social Security Act, 49Stat. 622, as amended, 42 U. S. C. §401 et seq., “is an insurance program” that “provides old-age, survivor, and disability benefits to insured individuals irrespective of financial need.” Bowen v. Galbreath, 485 U. S. 74, 75 (1988). A claimant’s application for Title II benefits can result in payments of past-due benefits—i.e., benefits that accrued before a favorable decision, 20 CFR §404.1703 (2018)—as well as ongoing monthly benefits, see 42 U. S. C. §423(a). A claimant who has been denied benefits “in whole or in part” by the Social Security Administration may seek administrative review of the initial agency determination, §405(b), and may then seek judicial review of the resulting final agency decision, §405(g). As presently written, the Social Security Act “discretely” addresses attorney’s fees for the administrative and judicial-review stages: “§406(a) governs fees for representa-tion in administrative proceedings; §406(b) controls fees for representation in court.” Gisbrecht v. Barnhart, 535 U. S. 789, 794 (2002). The original Social Security Act made no such provision for attorney’s fees in either proceeding. Id., at 793, n. 2. But in 1939, “Congress amended the Act to permit the Social Security Board to prescribe maximum fees attorneys could charge for representation of claimants before the agency.” Ibid. In 1965, Congress added a new subsection (b) to §406 that explicitly prescribed fees for representation before a court and “allow[ed] withholding of past-due benefits to pay” these fees directly to the attorney. Social Security Amendments of 1965, §332, 79Stat. 403; Bowen, 485 U. S., at 76. In 1968, Congress amended subsection (a) to give the agency similar withholding authority to pay attorney’s fees incurred in administrative proceedings. Id., at 76. Section 406(a) is titled “Recognition of representatives; fees for representation before Commissioner” of Social Security. It includes two ways to determine fees for representation before the agency, depending on whether a prior fee agreement exists. If the claimant has a fee agreement, subsection (a)(2) caps fees at the lesser of 25% of past-due benefits or a set dollar amount—currently $6,000. §406(a)(2)(A); Maximum Dollar Limit in the Fee Agreement Process, 74 Fed. Reg. 6080 (2009). Absent a fee agreement, subsection (a)(1) gives the agency authority to “prescribe the maximum fees which may be charged for services performed in connection with any claim” before the agency. If the claimant obtains a favorable agency determination, the agency may allot “a reasonable fee to compensate such attorney for the services performed by him.” Subsection (a)(4) requires the agency to withhold up to 25% of past-due benefits for direct payment of any fee for representation before the agency: “[I]f the claimant is determined to be entitled to past-due benefits under this subchapter and the person representing the claimant is an attorney, the Commissioner of Social Security shall . . . certify for payment out of such past-due benefits . . . to such attorney an amount equal to so much of the maximum fee as does not exceed 25 percent of such past-due benefits . . . .” Section 406(b) is titled “Fees for representation before court.” Subsection (b)(1)(A) both limits these fees to no more than 25% of past-due benefits and allows the agency to withhold past-due benefits to pay these fees: “Whenever a court renders a judgment favorable to a claimant under this subchapter who was represented before the court by an attorney, the court may determine and allow as part of its judgment a reasonable fee for such representation, not in excess of 25 percent of the total of the past-due benefits to which the claimant is entitled by reason of such judgment, and the Commissioner of Social Security may . . . certify the amount of such fee for payment to such attorney out of, and not in addition to, the amount of such past-due benefits.” At issue is whether §406(b)’s 25% cap limits the aggregate fees awarded for representation before both the agency under §406(a) and the court under §406(b), or instead limits only the fee awarded for court representation under §406(b). B Petitioner Richard Culbertson represented claimant Katrina Wood in proceedings seeking Social Security disability benefits. After the agency denied Wood benefits, she brought an action in district court. For the court action, Wood signed a contingency-fee agreement “to pay a fee of 25 percent of the total of the past-due benefits to which [she] is entitled” in consideration for Culbertson’s “representation of [her] in Federal Court.” App. 8–9. The agreement excludes fees for “any representation before” the agency. Id., at 9. The District Court reversed the agency’s denial of benefits and remanded for further proceedings. The court granted Wood attorney’s fees under the Equal Access to Justice Act (EAJA), which authorizes an award against the Government for reasonable fees in “civil action[s].” 28 U. S. C. §§2412(d)(1)(A) and (2)(A). On remand, the agency awarded Wood past-due disability benefits and withheld 25% of those benefits to pay any attorney’s fees that might ultimately be awarded. The agency also awarded Culbertson §406(a) fees for representing Wood before the agency. Culbertson then moved the District Court for a separate fee award under §406(b) for representing Wood there. After accounting for the EAJA award, see Gisbrecht, supra, at 796; App. 9, this request amounted to a full 25% of past-due benefits. The court granted Culbertson’s request only in part because he did not subtract the amount he had already received under §406(a) for his agency-level representation. The Eleventh Circuit affirmed, relying on Circuit precedent to hold that “the 25% limit from §406(b) applies to total fees awarded under both §406(a) and (b), ‘preclud[ing] the aggregate allowance of attorney’s fees greater than twenty-five percent of the past due benefits received by the claimant.’ ” Wood v. Commissioner of Social Security, 861 F. 3d 1197, 1205 (2017) (quoting Dawson v. Finch, 425 F. 2d 1192, 1195 (CA5 1970); emphasis deleted).[1]* Given a conflict between the Circuits on this question, see 861 F. 3d, at 1205–1206, we granted certiorari. 584 U. S. ___ (2018). Because no party defends the judgment, we appointed Amy Weil to brief and argue this case as amicus curiae in support of the judgment below. 584 U. S. ___ (2018). Amicus Weil has ably discharged her assigned responsibilities. II A We “begi[n] with the language of the statute itself, and that is also where the inquiry should end, for the statute’s language is plain.” Puerto Rico v. Franklin Cal. Tax-Free Trust, 579 U. S. ___, ___ (2016) (slip op., at 9) (internal quotation marks omitted). Under §406(b), when a court “renders a judgment favorable to a claimant . . . who was represented before the court by an attorney,” the court may award “a reasonable fee for such representation, not in excess of 25 percent of the total of the past-due benefits to which the claimant is entitled by reason of such judgment.” 42 U. S. C. §406(b)(1)(A) (emphasis added). Both at the time of enactment and today, the adjective “such” means “[o]f the kind or degree already described or implied.” H. Fowler & F. Fowler, Concise Oxford Dictionary of Current English 1289 (5th ed. 1964); Black’s Law Dictionary 1661 (10th ed. 2014) (“[t]hat or those; having just been mentioned”). Here, the only form of representation “already described” in §406(b) is “represent[ation] before the court by an attorney.” Accordingly, the 25% cap applies only to fees for representation before the court, not the agency. This interpretation is supported by “the structure of the statute and its other provisions.” Maracich v. Spears, 570 U. S. 48, 60 (2013). As an initial matter, subsections (a) and (b) address different stages of the representation. Section 406(a) addresses fees for representation “before the Commissioner,” whereas §406(b) addresses fees for representation in court. Because some claimants will prevail before the agency and have no need to bring a court action, it is unsurprising that the statute contemplates separate fees for each stage of representation. These subsections also calculate fees differently. Section 406(b) applies a flat 25% cap on fees for court representation. By contrast, §406(a) provides two ways to determine fees for agency proceedings. Subsection (a)(2) caps fees based on a fee agreement at the lesser of 25% of past-due benefits or $6,000. Supra, at 2. If there is no fee agreement, the agency may set any fee, including a fee greater than 25% of past-due benefits, so long as the fee is “reasonable.” §406(a)(1). Given this statutory structure, applying §406(b)’s 25% cap on court-stage fees to §406(a) agency-stage fees, or the aggregate of §§406(a) and (b) fees, would make little sense. Many claimants will never litigate in court, yet under the aggregate reading, agency fees would be capped at 25% based on a provision related exclusively to representation in court. Absent a fee agreement, §406(a)(1) subjects agency fees only to a reasonableness limitation, so applying §406(b)’s cap to such fees would add a limitation that Congress did not include in the relevant provision of the statute. If Congress had wanted these fees to be capped at 25%, it presumably would have said so directly in subsection (a), instead of providing for a “reasonable fee” in that subsection and adding a 25% cap in §406(b) without even referencing subsection (a). Thus, the structure of the statute confirms that §406(b) caps only court representation fees. B Amicus Amy Weil agrees that “§406(a) and §406(b) provide separate avenues for an award of attorney’s fees for representation of a Social Security claimant,” but emphasizes that “these fees are certified for payment out of a single source: the 25% of past-due benefits withheld by the Commissioner.” Brief for Court-Appointed Amicus Curiae 10. According to Amicus, “[b]ecause the Commissioner withholds only one pool of 25% of past-due benefits from which to pay attorney’s fees for both agency and court representation, for an attorney to collect a fee that exceeds the 25% pool of withheld disability benefits,” the attorney may “need to file a lawsuit against his disabled client” to collect the difference. Id., at 23–24. Therefore, Amicus urges, “[w]hen the statute is read as a whole,” “it is evident that Congress placed a cumulative 25% cap on attorney’s fees payable for successful representation of a Social Security claimant before both the agency and the court.” Id., at 10. Amicus is quite right that presently the agency withholds a single pool of 25% of past-due benefits for direct payment of agency and court fees. See Social Security Administration, Program Operations Manual System (POMS), GN 03920.035(A), online at https://policy.ssa.gov/ poms.nsf/lnx/0203920035 (as last visited Jan. 2, 2019); see also 20 CFR §§404.1730(a) and (b)(1)(i). And Amicus sensibly argues that if there is only a single 25% pool for direct payment of fees, Congress might not have intended aggregate fees higher than 25%. This argument is plausible, but the statutory text in fact provides for two pools of money for direct payment of fees. Any shortage of withheld benefits for direct payment of fees is thus due to agency policy. Under §406(a)(4), the agency “shall” certify for direct payment of agency representation fees “an amount equal to so much of the maximum fee as does not exceed 25 percent of” past-due benefits. In other words, this subsection requires that the agency withhold the approved fees for work performed in agency proceedings, up to 25% of the amount of the claimant’s past-due benefits. But this is not the only subsection that enables the agency to withhold past-due benefits for direct payment of fees. Section 406(b)(1)(A) provides that the agency “may” certify past-due benefits for direct payment of court representation fees. As the Government explains, the agency has nevertheless “exercised its discretion . . . to withhold a total of 25% of past-due benefits for direct payment of the approved agency and court fees.” Reply Brief for Respondent 8 (emphasis added). The agency’s choice to withhold only one pool of 25% of past-due benefits does not alter the statutory text, which differentiates between agency representation in §406(a) and court representation in §406(b), contains separate caps on fees for each type of representation, and authorizes two pools of withheld benefits. More fundamentally, the amount of past-due benefits that the agency can withhold for direct payment does not delimit the amount of fees that can be approved for representation before the agency or the court. The attorney might receive a direct payment out of past-due benefits, but that payment could be less than the fees to which the attorney is entitled. Indeed, prior to 1968, the statute allowed fees for agency representation but lacked a provision for direct payment of such fees from past-due benefits. See supra, at 2. And under the current §§406(a)(1) and (4), the agency can award a “reasonable fee” that exceeds the 25% of past-due benefits it can withhold for direct payment. In short, despite the force of Amicus’ arguments, the statute does not bear her reading. Any concerns about a shortage of withheld benefits for direct payment and the consequences of such a shortage are best addressed to the agency, Congress, or the attorney’s good judgment. * * * Because the 25% cap in §406(b)(1)(A) applies only to fees for court representation, and not to the aggregate fees awarded under §§406(a) and (b), the judgment of the United States Court of Appeals for the Eleventh Circuit is reversed, and the case is remanded for further proceedings consistent with this opinion. It is so ordered. Notes 1 * See Bonner v. Prichard, 661 F. 2d 1206, 1209 (CA11 1981) (en banc) (adopting all decisions of the former Fifth Circuit announced prior to October 1, 1981, as binding precedent in the Eleventh Circuit).

586.US.2018_17-419

After petitioner James Dawson retired from the U. S. Marshals Service, his home State of West Virginia taxed his federal pension benefits as it does all former federal employees. The pension benefits of certain former state and local law enforcement employees, however, are exempt from state taxation. See W. Va. Code Ann. §11–21–12(c)(6). Mr. Dawson sued, alleging that the state statute violates the intergovernmental tax immunity doctrine as codified at 4 U. S. C. §111. Under that statute, the United States consents to state taxation of the pay or compensation of federal employees, but only if the state tax does not discriminate on the basis of the source of the pay or compensation. A West Virginia trial court found no significant differences between Mr. Dawson’s job duties as a federal marshal and those of the state and local law enforcement officers exempted from taxation and held that the state statute violates §111’s antidiscrimination provision. Reversing, the West Virginia Supreme Court of Appeals emphasized that the state tax exemption applies only to a narrow class of state retirees and was never intended to discriminate against former federal marshals. Held: The West Virginia statute unlawfully discriminates against Mr. Dawson as §111 forbids. A State violates §111 when it treats retired state employees more favorably than retired federal employees and no “significant differences between the two classes” justify the differential treatment. Davis v. Michigan Dept. of Treasury, 489 U.S. 803, 814–816. Here, West Virginia expressly affords state law enforcement retirees a tax benefit that federal retirees cannot receive, and there are no “significant differences” between Mr. Dawson’s former job responsibilities and those of the tax-exempt state law enforcement retirees. The narrow preference should be permitted, the State argues, because it affects too few people to meaningfully interfere with federal government operations. Section 111, however, disallows any state tax that discriminates against a federal officer or employee—not just those that seem especially cumbersome. And in Davis the Court refused a similar invitation to add unwritten qualifications to §111. That is not to say that the narrowness of a state tax exemption is irrelevant. If a State exempts only a narrow subset of state retirees, it can comply with §111 by exempting only the comparable class of federal retirees. The State also argues that the statute is not intended to harm federal retirees but to help certain state retirees. The “State’s interest in adopting the discriminatory tax,” however, “is simply irrelevant.” Davis, 489 U. S., at 816. For reasons other than job responsibilities, the State insists, retired U. S. Marshals and tax-exempt state law enforcement retirees are not “similarly situated.” But the State’s statute does not draw any such lines. It singles out for preferential treatment retirement plans associated with particular state law enforcement officers. The distinguishing characteristic of the retirement plans is the nature of the jobs previously held by retirees who may participate in them. The state trial court found no “significant differences” between Mr. Dawson’s former job responsibilities as a U. S. Marshal and those of the state law enforcement retirees who qualify for the tax exemption, and the West Virginia Supreme Court of Appeals did not upset that finding. By submitting that Mr. Dawson’s former job responsibilities are also similar to those of other state law enforcement retirees who do not qualify for a tax exemption, the State mistakes the nature of the inquiry. The relevant question under §111 is not whether federal retirees are similarly situated to state retirees who do not receive a tax break; it is whether they are similarly situated to those who do. Finally, the State says that the real distinction may not be based on job duties at all but on the relative generosity of pension benefits. The statute as enacted, however, does not classify persons or groups on that basis. And an implicit but lawful distinction cannot save an express and unlawful one. See, e.g., id., at 817. Pp. 3–8. Reversed and remanded. Gorsuch, J., delivered the opinion for a unanimous Court.

If you spent your career as a state law enforcement officer in West Virginia, you’re likely to be eligible for a generous tax exemption when you retire. But if you served in federal law enforcement, West Virginia will deny you the same benefit. The question we face is whether a State may discriminate against federal retirees in that way. For most of his career, James Dawson worked in the U. S. Marshals Service. After he retired, he began looking into the tax treatment of his pension. It turns out that his home State, West Virginia, doesn’t tax the pension benefits of certain former state law enforcement employees. But it does tax the benefits of all former federal employees. So Mr. Dawson brought this lawsuit alleging that West Virginia violated 4 U. S. C. §111. In that statute, the United States has consented to state taxation of the “pay or compensation” of “officer[s] or employee[s] of the United States,” but only if the “taxation does not discriminate against the officer or employee because of the source of the pay or compensation.” §111(a). Section 111 codifies a legal doctrine almost as old as the Nation. In McCulloch v. Maryland, 4 Wheat. 316 (1819), this Court invoked the Constitution’s Supremacy Clause to invalidate Maryland’s effort to levy a tax on the Bank of the United States. Chief Justice Marshall explained that “the power to tax involves the power to destroy,” and he reasoned that if States could tax the Bank they could “defeat” the federal legislative policy establishing it. Id., at 431–432. For the next few decades, this Court interpreted McCulloch “to bar most taxation by one sovereign of the employees of another.” Davis v. Michigan Dept. of Treasury, 489 U.S. 803, 810 (1989). In time, though, the Court softened its stance and upheld neutral income taxes—those that treated federal and state employees with an even hand. See Helvering v. Gerhardt, 304 U.S. 405 (1938); Graves v. New York ex rel. O’Keefe, 306 U.S. 466 (1939). So eventually the intergovernmental tax immu- nity doctrine came to be understood to bar only discrimina- tory taxes. It was this understanding that Congress “consciously . . . drew upon” when adopting §111 in 1939. Davis, 489 U. S., at 813. It is this understanding, too, that has animated our application of §111. Since the statute’s adoption, we have upheld an Alabama income tax that did not discriminate on the basis of the source of the employees’ compensation. Jefferson County v. Acker, 527 U.S. 423 (1999). But we have invalidated a Michigan tax that discriminated “in favor of retired state employees and against retired federal employees.” Davis, 489 U. S., at 814. We have struck down a Kansas law that taxed the retirement benefits of federal military personnel at a higher rate than state and local government retirement benefits. Barker v. Kansas, 503 U.S. 594, 599 (1992). And we have rejected a Texas scheme that imposed a property tax on a private company operating on land leased from the federal government, but a “less burdensome” tax on property leased from the State. Phillips Chemical Co. v. Dumas Independent School Dist., 361 U.S. 376, 378, 380 (1960). Mr. Dawson’s own attempt to invoke §111 met with mixed success. A West Virginia trial court found it “undisputed” that “there are no significant differences between Mr. Dawson’s powers and duties as a US Marshal and the powers and duties of the state and local law enforcement officers” that West Virginia exempts from income tax. App. to Pet. for Cert. 22a. In the trial court’s judgment, the State’s statute thus represented “precisely the type of favoritism” §111 prohibits. Id., at 23a. But the West Virginia Supreme Court of Appeals saw it differ- ently. In reversing, the court emphasized that relatively few state employees receive the tax break denied Mr. Dawson. The court stressed, too, that the statute’s “intent . . . was to give a benefit to a narrow class of state retirees,” not to harm federal retirees. Id., at 15a. Because cases in this field have yielded inconsistent results, much as this one has, we granted certiorari to afford additional guidance. 585 U. S. ___ (2018). We believe the state trial court had it right. A State violates §111 when it treats retired state employees more favorably than retired federal employees and no “significant differences between the two classes” justify the differential treatment. Davis, 489 U. S., at 814–816 (1989) (internal quotation marks omitted); Phillips Chemical Co., 361 U. S., at 383. Here, West Virginia expressly affords state law enforcement retirees a tax benefit that federal retirees cannot receive. And before us everyone accepts the trial court’s factual finding that there aren’t any “significant differences” between Mr. Dawson’s former job responsibilities and those of the tax-exempt state law enforcement retirees. Given all this, we have little difficulty concluding that West Virginia’s law unlawfully “discriminate[s]” against Mr. Dawson “because of the source of [his] pay or compensation,” just as §111 forbids. The State offers this ambitious rejoinder. Even if its statute favors some state law enforcement retirees, the favored class is very small. Most state retirees are treated no better than Mr. Dawson. And this narrow preference, the State suggests, should be permitted because it affects so few people that it couldn’t meaningfully interfere with the operations of the federal government. We are unpersuaded. Section 111 disallows any state tax that discriminates against a federal officer or employee—not just those that seem to us especially cumbersome. Nor are we inclined to accept West Virginia’s invitation to adorn §111 with a new and judicially manufactured qualification that cannot be found in its text. In fact, we have already refused an almost identical request. In Davis, we rejected Michigan’s suggestion that a discriminatory state income tax should be allowed to stand so long as it treats federal employees or retirees the same as “the vast major- ity of voters in the State.” 489 U. S., at 815, n. 4. We rejected, too, any suggestion that a discriminatory tax is permissible so long as it “does not interfere with the Federal Government’s ability to perform its governmental functions.” Id., at 814. In fact, as long ago as McCulloch, Chief Justice Marshall warned against enmeshing courts in the “perplexing” business, “so unfit for the judicial department,” of attempting to delineate “what degree of taxation is the legitimate use, and what degree may amount to the abuse of power.” 4 Wheat., at 430. That’s not to say the breadth or narrowness of a state tax exemption is irrelevant. Under §111, the scope of a State’s tax exemption may affect the scope of its resulting duties. So if a State exempts from taxation all state employees, it must likewise exempt all federal employees. Conversely, if the State decides to exempt only a narrow subset of state retirees, the State can comply with §111 by exempting only the comparable class of federal retirees. But the narrowness of a discriminatory state tax law has never been enough to render it necessarily lawful. With its primary argument lost, the State now proceeds more modestly. Echoing the West Virginia Supreme Court of Appeals, the State argues that we should uphold its statute because it isn’t intended to harm federal retirees, only to help certain state retirees. But under the terms of §111, the “State’s interest in adopting the discriminatory tax, no matter how substantial, is simply irrelevant.” Davis, 489 U. S., at 816. We can safely assume that discriminatory laws like West Virginia’s are almost always enacted with the purpose of benefiting state employees rather than harming their federal counterparts. Yet that wasn’t enough to save the state statutes in Davis, Barker, or Phillips, and it can’t be enough here. Under §111 what matters isn’t the intent lurking behind the law but whether the letter of the law “treat[s] those who deal with” the federal government “as well as it treats those with whom [the State] deals itself.” Phillips Chemical Co., 361 U. S., at 385. If treatment rather than intent is what matters, the State suggests that it should still prevail for other reasons. Section 111 prohibits “discriminat[ion],” something we’ve often described as treating similarly situated persons differently. See Davis, 489 U. S., at 815–816; Phillips Chemical Co., 361 U. S., at 383. And before us West Virginia insists that even if retired U. S. Marshals and tax-exempt state law enforcement retirees had similar job responsibilities, they aren’t “similarly situated” for other reasons. Put another way, the State contends that the difference in treatment its law commands doesn’t qualify as unlawful discrimination because it is “directly related to, and justified by,” a lawful and “significant difference” between the two classes. Davis, 489 U. S., at 816 (internal quotation marks and alteration omitted). In approaching this argument, everyone before us agrees on at least one thing. Whether a State treats similarly situated state and federal employees differently depends on how the State has defined the favored class. See id., at 817. So if the State defines the favored class by reference to job responsibilities, a similarly situated fed- eral worker will be one who performs comparable duties. But if the State defines the class by reference to some other criteria, our attention should naturally turn there. If a State gives a tax benefit to all retirees over a certain age, for example, the comparable federal retiree would be someone who is also over that age. So how has West Virginia chosen to define the favored class in this case? The state statute singles out for preferential treatment retirement plans associated with West Virginia police, firefighters, and deputy sheriffs. See W. Va. Code Ann. §11–21–12(c)(6) (Lexis 2017). The distinguishing characteristic of these plans is the nature of the jobs previously held by retirees who may participate in them; thus, a similarly situated federal retiree is someone who had similar job responsibilities to a state police officer, firefighter, or deputy sheriff. The state trial court correctly focused on this point of comparison and found no “significant differences” between Mr. Dawson’s former job responsibilities as a U. S. Marshal and those of the state law enforcement retirees who qualify for the tax exemption. App. to Pet. for Cert. 22a. Nor did the West Virginia Supreme Court of Appeals upset this factual finding. So looking to how the State has chosen to define its favored class only seems to confirm that it has treated similarly situated persons differently because of the source of their compensation. Of course, West Virginia sees it otherwise. It accepts (for now) that its statute distinguishes between persons based on their former job duties. It accepts, too, the trial court’s finding that Mr. Dawson’s former job responsibilities are materially identical to those of state retirees who qualify for its tax exemption. But, the State submits, Mr. Dawson’s former job responsibilities are also similar to those of other state law enforcement retirees who don’t qualify for its tax exemption. And, the State insists, the fact that it treats federal retirees no worse than (some) similarly situated state employees should be enough to save its statute. But this again mistakes the nature of our inquiry. Under §111, the relevant question isn’t whether federal retirees are similarly situated to state retirees who don’t receive a tax benefit; the relevant question is whether they are similarly situated to those who do. So, for example, in Phillips we compared the class of federal lessees with the favored class of state lessees, even though the State urged us to focus instead on the disfavored class of private lessees. 361 U. S., at 381–382. In Davis, we likewise rejected the State’s effort to compare the class of federal retirees with state residents who did not benefit from the tax exemption rather than those who did. See 489 U. S., at 815, n. 4. At this point the State is left to play its final card. Now, it says, maybe the real distinction its statute draws isn’t based on former job duties at all. Maybe its statute actually favors certain state law enforcement retirees only because their pensions are less generous than those of their federal law enforcement counterparts. At the least, the State suggests, we should remand the case to the West Virginia courts to explore this possibility. The problem here is fundamental. While the State was free to draw whatever classifications it wished, the statute it enacted does not classify persons or groups based on the relative generosity of their pension benefits. Instead, it extends a special tax benefit to retirees who served as West Virginia police officers, firefighters, or deputy sheriffs—and it categorically denies that same benefit to retirees who served in similar federal law enforcement positions. Even if Mr. Dawson’s pension turned out to be identical to a state law enforcement officer’s pension, the law as written would deny him a tax exemption. West Virginia’s law thus discriminates “because of the source of . . . compensation or pay” in violation of §111. Whether the unlawful classification found in the text of a statute might serve as some sort of proxy for a lawful classification hidden behind it is neither here nor there. No more than a beneficent legislative intent, an implicit but lawful distinction cannot save an express and unlawful one. Our precedent confirms this too. In Davis, Michigan argued that a state law expressly discriminating between federal and state retirees was really just distinguishing between those with more and less generous pensions. Id., at 816. We rejected this attempt to rerationalize the statute, explaining that “[a] tax exemption truly intended to account for differences in retirement benefits would not discriminate on the basis of the source of those benefits” but “would discriminate on the basis of the amount of benefits received by individual retirees.” Id., at 817. The fact is, when States seek to tax the use of a fellow sovereign’s property, the Constitution and Congress have always carefully constrained their authority. Id., at 810–814. And in this sensitive field it is not too much to ask that, if a State wants to draw a distinction based on the generosity of pension benefits, it enact a law that actually does that. Because West Virginia’s statute unlawfully discriminates against Mr. Dawson, we reverse the judgment of the West Virginia Supreme Court of Appeals and remand the case for further proceedings not inconsistent with this opinion, including the determination of an appropriate remedy. It is so ordered.

588.US.2018_18-966

In order to apportion congressional representatives among the States, the Constitution requires an “Enumeration” of the population every 10 years, to be made “in such Manner” as Congress “shall by Law direct,” Art. I, §2, cl. 3; Amdt. 14, §2. In the Census Act, Congress delegated to the Secretary of Commerce the task of conducting the decennial census “in such form and content as he may determine.” 13 U. S. C. §141(a). The Secretary is aided by the Census Bureau, a statistical agency in the Department of Commerce. The population count is also used to allocate federal funds to the States and to draw electoral districts. The census additionally serves as a means of collecting demographic information used for a variety of purposes. There have been 23 decennial censuses since 1790. All but one between 1820 and 2000 asked at least some of the population about their citizenship or place of birth. The question was asked of all households until 1950, and was asked of a fraction of the population on an alternative long-form questionnaire between 1960 and 2000. In 2010, the citizenship question was moved from the census to the American Community Survey, which is sent each year to a small sample of households. In March 2018, Secretary of Commerce Wilbur Ross announced in a memo that he had decided to reinstate a citizenship question on the 2020 census questionnaire at the request of the Department of Justice (DOJ), which sought census block level citizenship data to use in enforcing the Voting Rights Act (VRA). The Secretary’s memo explained that the Census Bureau initially analyzed, and the Secretary considered, three possible courses of action before he chose a fourth option that combined two of the proposed options: reinstate a citizenship question on the decennial census, and use administrative records from other agencies, e.g., the Social Security Administration, to provide additional citizenship data. The Secretary “carefully considered” the possibility that reinstating a citizenship question would depress the response rate, the long history of the citizenship question on the census, and several other factors before concluding that “the need for accurate citizenship data and the limited burden of the question” outweighed fears about a lower response rate. Here, two separate suits filed in Federal District Court in New York were consolidated: one filed by a group States, counties, cities, and others, alleging that the Secretary’s decision violated the Enumeration Clause and the requirements of the Administrative Procedure Act; the other filed by non-governmental organizations, adding an equal protection claim. The District Court dismissed the Enumeration Clause claim but allowed the other claims to proceed. In June 2018, the Government submitted the Commerce Department’s “administrative record”—materials that Secretary Ross considered in making his decision—including DOJ’s letter requesting reinstatement of the citizenship question. Shortly thereafter, at DOJ’s urging, the Government supplemented the record with a new memo from the Secretary, which stated that he had begun considering the addition of a citizenship question in early 2017 and had asked whether DOJ would formally request its inclusion. Arguing that the supplemental memo indicated that the record was incomplete, respondents asked the District Court to compel the Government to complete the administrative record. The court granted that request, and the parties jointly stipulated to the inclusion of additional materials that confirmed that the Secretary and his staff began exploring reinstatement of a citizenship question shortly after his 2017 confirmation, attempted to elicit requests for citizenship data from other agencies, and eventually persuaded DOJ to make the request. The court also authorized discovery outside the administrative record, including compelling a deposition of Secretary Ross, which this Court stayed pending further review. After a bench trial, the District Court determined that respondents had standing to sue. On the merits, it ruled that the Secretary’s action was arbitrary and capricious, based on a pretextual rationale, and violated the Census Act, and held that respondents had failed to show an equal protection violation. Held: 1. At least some respondents have Article III standing. For a legal dispute to qualify as a genuine case or controversy, at least one plaintiff must “present an injury that is concrete, particularized, and actual or imminent; fairly traceable to the defendant’s challenged behavior; and likely to be redressed by a favorable ruling.” Davis v. Federal Election Comm’n, 554 U.S. 724, 733. The District Court concluded that the evidence at trial established a sufficient likelihood that reinstating a citizenship question would result in noncitizen households responding to the census at lower rates than other groups, which would cause them to be undercounted and lead to many of the injuries respondents asserted—diminishment of political representation, loss of federal funds, degradation of census data, and diversion of resources. For purposes of standing, these findings of fact were not so suspect as to be clearly erroneous. Several state respondents have shown that if noncitizen households are undercounted by as little as 2%, they will lose out on federal funds that are distributed on the basis of state population. That is a sufficiently concrete and imminent injury to satisfy Article III, and there is no dispute that a ruling in favor of respondents would redress that harm. Pp. 8–11. 2. The Enumeration Clause permits Congress, and by extension the Secretary, to inquire about citizenship on the census questionnaire. That conclusion follows from Congress’s broad authority over the census, as informed by long and consistent historical practice that “has been open, widespread, and unchallenged since the early days of the Republic.” NLRB v. Noel Canning, 573 U.S. 513, 572 (Scalia, J., concurring in judgment). Pp. 11–13. 3. The Secretary’s decision is reviewable under the Administrative Procedure Act. The APA instructs reviewing courts to set aside agency action that is “arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law,” 5 U. S. C. §706(2)(A), but it makes review unavailable “to the extent that” the agency action is “committed to agency discretion by law,” §701(a)(2). The Census Act confers broad authority on the Secretary, but it does not leave his discretion unbounded. The §701(a)(2) exception is generally limited to “certain categories of administrative decisions that courts traditionally have regarded as ‘committed to agency discretion,’ ” Lincoln v. Vigil, 508 U.S. 182, 191. The taking of the census is not one of those areas. Nor is the statute drawn so that it furnishes no meaningful standard by which to judge the Secretary’s action, which is amenable to review for compliance with several Census Act provisions according to the general requirements of reasoned agency decisionmaking. Because this is not a case in which there is “no law to apply,” Citizens to Preserve Overton Park, Inc. v. Volpe, 401 U.S. 402, 410, the Secretary’s decision is subject to judicial review. Pp. 13–16. 4. The Secretary’s decision was supported by the evidence before him. He examined the Bureau’s analysis of various ways to collect improved citizenship data and explained why he thought the best course was to both reinstate a citizenship question and use citizenship data from administrative records to fill in the gaps. He then weighed the value of obtaining more complete and accurate citizenship data against the uncertain risk that reinstating a citizenship question would result in a materially lower response rate, and explained why he thought the benefits of his approach outweighed the risk. That decision was reasonable and reasonably explained, particularly in light of the long history of the citizenship question on the census. Pp. 16–20. 5. The District Court also erred in ruling that the Secretary violated two particular provisions of the Census Act, §6(c) and §141(f). Section 6’s first two subsections authorize the Secretary to acquire administrative records from other federal agencies and state and local governments, while subsection (c) requires the Secretary, to the maximum extent possible, to use that information “instead of conducting direct inquiries.” Assuming that §6(c) applies, the Secretary complied with it for essentially the same reasons that his decision was not arbitrary and capricious: Administrative records would not, in his judgment, provide the more complete and accurate data that DOJ sought. The Secretary also complied with §141(f), which requires him to make a series of reports to Congress about his plans for the census. And even if he had violated that provision, the error would be harmless because he fully informed Congress of, and explained, his decision. Pp. 20–23. 6. In order to permit meaningful judicial review, an agency must “ ‘disclose the basis’ ” of its action. Burlington Truck Lines, Inc. v. United States, 371 U.S. 156, 167–169. A court is ordinarily limited to evaluating the agency’s contemporaneous explanation in light of the existing administrative record, Vermont Yankee Nuclear Power Corp. v. Natural Resources Defense Council, Inc., 435 U.S. 519, but it may inquire into “the mental processes of administrative decisionmakers” upon a “strong showing of bad faith or improper behavior,” Overton Park, 401 U. S., at 420. While the District Court pre- maturely invoked that exception in ordering extra-record discovery here, it was ultimately justified in light of the expanded administrative record. Accordingly, the District Court’s ruling on pretext will be reviewed in light of all the evidence in the record, including the extra-record discovery. It is hardly improper for an agency head to come into office with policy preferences and ideas, discuss them with affected parties, sound out other agencies for support, and work with staff attorneys to substantiate the legal basis for a preferred policy. Yet viewing the evidence as a whole, this Court shares the District Court’s conviction that the decision to reinstate a citizenship question cannot adequately be explained in terms of DOJ’s request for improved citizenship data to better enforce the VRA. Several points, taken together, reveal a significant mismatch between the Secretary’s decision and the rationale he provided. The record shows that he began taking steps to reinstate the question a week into his tenure, but gives no hint that he was considering VRA enforcement. His director of policy attempted to elicit requests for citizenship data from the Department of Homeland Security and DOJ’s Office of Immigration Review before turning to the VRA rationale and DOJ’s Civil Rights Division. For its part, DOJ’s actions suggest that it was more interested in helping the Commerce Department than in securing the data. Altogether, the evidence tells a story that does not match the Secretary’s explanation for his decision. Unlike a typical case in which an agency may have both stated and unstated reasons for a decision, here the VRA enforcement rationale—the sole stated reason—seems to have been contrived. The reasoned explanation requirement of administrative law is meant to ensure that agencies offer genuine justifications for important decisions, reasons that can be scrutinized by courts and the interested public. The explanation provided here was more of a distraction. In these unusual circumstances, the District Court was warranted in remanding to the agency. See Florida Power & Light Co. v. Lorion, 470 U.S. 729, 744. Pp. 23–28. 351 F. Supp. 3d 502, affirmed in part, reversed in part, and remanded. Roberts, C. J., delivered the opinion for a unanimous Court with respect to Parts I and II, and the opinion of the Court with respect to Parts III, IV–B, and IV–C, in which Thomas, Alito, Gorsuch, and Kav- anaugh, JJ., joined; with respect to Part IV–A, in which Thomas, Ginsburg, Breyer, Sotomayor, Kagan, and Kavanaugh, JJ., joined; and with respect to Part V, in which Ginsburg, Breyer, Sotomayor, and Kagan, JJ., joined. Thomas, J., filed an opinion concurring in part and dissenting in part, in which Gorsuch and Kavanaugh, JJ., joined. Breyer, J., filed an opinion concurring in part and dissenting in part, in which Ginsburg, Sotomayor, and Kagan, JJ., joined. Alito, J., filed an opinion concurring in part and dissenting in part.

The Secretary of Commerce decided to reinstate a question about citizenship on the 2020 census questionnaire. A group of plaintiffs challenged that decision on constitutional and statutory grounds. We now decide whether the Secretary violated the Enumeration Clause of the Constitution, the Census Act, or otherwise abused his discretion. I A In order to apportion Members of the House of Representatives among the States, the Constitution requires an “Enumeration” of the population every 10 years, to be made “in such Manner” as Congress “shall by Law direct.” Art. I, §2, cl. 3; Amdt. 14, §2. In the Census Act, Congress delegated to the Secretary of Commerce the task of conducting the decennial census “in such form and content as he may determine.” 13 U. S. C. §141(a). The Secretary is aided in that task by the Census Bureau, a statistical agency housed within the Department of Commerce. See §§2, 21. The population count derived from the census is used not only to apportion representatives but also to allocate federal funds to the States and to draw electoral districts. Wisconsin v. City of New York, 517 U.S. 1, 5–6 (1996). The census additionally serves as a means of collecting demographic information, which “is used for such varied purposes as computing federal grant-in-aid benefits, drafting of legislation, urban and regional planning, business planning, and academic and social studies.” Baldrige v. Shapiro, 455 U.S. 345, 353–354, n. 9 (1982). Over the years, the census has asked questions about (for example) race, sex, age, health, education, occupation, housing, and military service. It has also asked about radio ownership, age at first marriage, and native tongue. The Census Act obliges everyone to answer census questions truthfully and requires the Secretary to keep individual answers confidential, including from other Government agencies. §§221, 8(b), 9(a). There have been 23 decennial censuses from the first census in 1790 to the most recent in 2010. Every census between 1820 and 2000 (with the exception of 1840) asked at least some of the population about their citizenship or place of birth. Between 1820 and 1950, the question was asked of all households. Between 1960 and 2000, it was asked of about one-fourth to one-sixth of the population. That change was part of a larger effort to simplify the census by asking most people a few basic demographic questions (such as sex, age, race, and marital status) on a short-form questionnaire, while asking a sample of the population more detailed demographic questions on a long-form questionnaire. In explaining the decision to move the citizenship question to the long-form questionnaire, the Census Bureau opined that “general census information on citizenship had become of less importance compared with other possible questions to be included in the census, particularly in view of the recent statutory requirement for annual alien registration which could provide the Immigration and Naturalization Service, the principal user of such data, with the information it needed.” Dept. of Commerce, Bureau of Census, 1960 Censuses of Population and Housing 194 (1966).[1] In 2010, the year of the latest census, the format changed again. All households received the same questionnaire, which asked about sex, age, race, Hispanic origin, and living arrangements. The more detailed demographic questions previously asked on the long-form questionnaire, including the question about citizenship, were instead asked in the American Community Survey (or ACS), which is sent each year to a rotating sample of about 2.6% of households. The Census Bureau and former Bureau officials have resisted occasional proposals to resume asking a citizenship question of everyone, on the ground that doing so would discourage noncitizens from responding to the census and lead to a less accurate count of the total population. See, e.g., Federation of Am. Immigration Reform v. Klutznick, 486 F. Supp. 564, 568 (DC 1980) (“[A]ccording to the Bureau[,] any effort to ascertain citizenship will inevitably jeopardize the overall accuracy of the population count”); Brief for Former Directors of the U. S. Census Bureau as Amici Curiae in Evenwel v. Abbott, O. T. 2014, No. 14–940, p. 25 (inquiring about citizenship would “invariably lead to a lower response rate”). B In March 2018, Secretary of Commerce Wilbur Ross announced in a memo that he had decided to reinstate a question about citizenship on the 2020 decennial census questionnaire. The Secretary stated that he was acting at the request of the Department of Justice (DOJ), which sought improved data about citizen voting-age population for purposes of enforcing the Voting Rights Act (or VRA)—specifically the Act’s ban on diluting the influence of minority voters by depriving them of single-member districts in which they can elect their preferred candidates. App. to Pet. for Cert. 548a. DOJ explained that federal courts determine whether a minority group could constitute a majority in a particular district by looking to the citizen voting-age population of the group. According to DOJ, the existing citizenship data from the American Community Survey was not ideal: It was not reported at the level of the census block, the basic component of legislative districting plans; it had substantial margins of error; and it did not align in time with the census-based population counts used to draw legislative districts. DOJ therefore formally requested reinstatement of the citizenship question on the census questionnaire. Id., at 565a–569a. The Secretary’s memo explained that the Census Bureau initially analyzed, and the Secretary considered, three possible courses of action. The first was to continue to collect citizenship information in the American Community Survey and attempt to develop a data model that would more accurately estimate citizenship at the census block level. The Secretary rejected that option because the Bureau “did not assert and could not confirm” that such ACS-based data modeling was possible “with a sufficient degree of accuracy.” Id., at 551a. The second option was to reinstate a citizenship question on the decennial census. The Bureau predicted that doing so would discourage some noncitizens from responding to the census. That would necessitate increased “non-response follow up” operations—procedures the Bureau uses to attempt to count people who have not responded to the census—and potentially lead to a less accurate count of the total population. Option three was to use administrative records from other agencies, such as the Social Security Administration and Citizenship and Immigration Services, to provide DOJ with citizenship data. The Census Bureau recommended this option, and the Secretary found it a “potentially appealing solution” because the Bureau has long used administrative records to supplement and improve census data. Id., at 554a. But the Secretary concluded that administrative records alone were inadequate because they were missing for more than 10% of the population. The Secretary ultimately asked the Census Bureau to develop a fourth option that would combine options two and three: reinstate a citizenship question on the census questionnaire, and also use the time remaining until the 2020 census to “further enhance” the Bureau’s “administrative record data sets, protocols, and statistical models.” Id., at 555a. The memo explained that, in the Secretary’s judgment, the fourth option would provide DOJ with the “most complete and accurate” citizen voting-age population data in response to its request. Id., at 556a. The Secretary “carefully considered” the possibility that reinstating a citizenship question would depress the response rate. Ibid. But after evaluating the Bureau’s “limited empirical evidence” on the question— evidence drawn from estimated non-response rates to previous American Community Surveys and census questionnaires—the Secretary concluded that it was not possible to “determine definitively” whether inquiring about citizenship in the census would materially affect response rates. Id., at 557a, 562a. He also noted the long history of the citizenship question on the census, as well as the facts that the United Nations recommends collecting census-based citizenship information, and other major democracies such as Australia, Canada, France, Indonesia, Ireland, Germany, Mexico, Spain, and the United Kingdom inquire about citizenship in their censuses. Altogether, the Secretary determined that “the need for accurate citizenship data and the limited burden that the reinstatement of the citizenship question would impose outweigh fears about a potentially lower response rate.” Id., at 557a. C Shortly after the Secretary announced his decision, two groups of plaintiffs filed suit in Federal District Court in New York, challenging the decision on several grounds. The first group of plaintiffs included 18 States, the District of Columbia, various counties and cities, and the United States Conference of Mayors. They alleged that the Secretary’s decision violated the Enumeration Clause of the Constitution and the requirements of the Administrative Procedure Act. The second group of plaintiffs consisted of several non-governmental organizations that work with immigrant and minority communities. They added an equal protection claim. The District Court consolidated the two cases. Both groups of plaintiffs are respondents here. The Government moved to dismiss the lawsuits, arguing that the Secretary’s decision was unreviewable and that respondents had failed to state cognizable claims under the Enumeration Clause and the Equal Protection Clause. The District Court dismissed the Enumeration Clause claim but allowed the other claims to proceed. 315 F. Supp. 3d 766 (SDNY 2018). In June 2018, the Government submitted to the District Court the Commerce Department’s “administrative record”: the materials that Secretary Ross considered in making his decision. That record included DOJ’s December 2017 letter requesting reinstatement of the citizenship question, as well as several memos from the Census Bureau analyzing the predicted effects of reinstating the question. Shortly thereafter, at DOJ’s urging, the Government supplemented the record with a new memo from the Secretary, “intended to provide further background and context regarding” his March 2018 memo. App. to Pet. for Cert. 546a. The supplemental memo stated that the Secretary had begun considering whether to add the citizenship question in early 2017, and had inquired whether DOJ “would support, and if so would request, inclusion of a citizenship question as consistent with and useful for enforcement of the Voting Rights Act.” Ibid. According to the Secretary, DOJ “formally” requested reinstatement of the citizenship question after that inquiry. Ibid. Respondents argued that the supplemental memo indicated that the Government had submitted an incomplete record of the materials considered by the Secretary. They asked the District Court to compel the Government to complete the administrative record. The court granted that request, and the parties jointly stipulated to the inclusion of more than 12,000 pages of additional materials in the administrative record. Among those materials were emails and other records confirming that the Secretary and his staff began exploring the possibility of reinstating a citizenship question shortly after he was confirmed in early 2017, attempted to elicit requests for citizenship data from other agencies, and eventually persuaded DOJ to request reinstatement of the question for VRA enforcement purposes. In addition, respondents asked the court to authorize discovery outside the administrative record. They claimed that such an unusual step was warranted because they had made a strong preliminary showing that the Secretary had acted in bad faith. See Citizens to Preserve Overton Park, Inc. v. Volpe, 401 U.S. 402, 420 (1971). The court also granted that request, authorizing expert discovery and depositions of certain DOJ and Commerce Department officials. In August and September 2018, the District Court issued orders compelling depositions of Secretary Ross and of the Acting Assistant Attorney General for DOJ’s Civil Rights Division. We granted the Government’s request to stay the Secretary’s deposition pending further review, but we declined to stay the Acting AAG’s deposition or the other extra-record discovery that the District Court had authorized. The District Court held a bench trial and issued findings of fact and conclusions of law on respondents’ statutory and equal protection claims. After determining that respondents had standing to sue, the District Court ruled that the Secretary’s action was arbitrary and capricious, based on a pretextual rationale, and violated certain provisions of the Census Act. On the equal protection claim, however, the District Court concluded that respondents had not met their burden of showing that the Secretary was motivated by discriminatory animus. The court granted judgment to respondents on their statutory claims, vacated the Secretary’s decision, and enjoined him from reinstating the citizenship question until he cured the legal errors the court had identified. 351 F. Supp. 3d 502 (SDNY 2019). The Government appealed to the Second Circuit, but also filed a petition for writ of certiorari before judgment, asking this Court to review the District Court’s decision directly because the case involved an issue of imperative public importance, and the census questionnaire needed to be finalized for printing by the end of June 2019. We granted the petition. 586 U. S. ___ (2019). At the Government’s request, we later ordered the parties to address whether the Enumeration Clause provided an alternative basis to affirm. 586 U. S. ___ (2019). II We begin with jurisdiction. Article III of the Constitution limits federal courts to deciding “Cases” and “Controversies.” For a legal dispute to qualify as a genuine case or controversy, at least one plaintiff must have standing to sue. The doctrine of standing “limits the category of litigants empowered to maintain a lawsuit in federal court to seek redress for a legal wrong” and “confines the federal courts to a properly judicial role.” Spokeo, Inc. v. Robins, 578 U. S. ___, ___ (2016) (slip op., at 6). To have standing, a plaintiff must “present an injury that is concrete, particularized, and actual or imminent; fairly traceable to the defendant’s challenged behavior; and likely to be redressed by a favorable ruling.” Davis v. Federal Election Comm’n, 554 U.S. 724, 733 (2008). Respondents assert a number of injuries—diminishment of political representation, loss of federal funds, degradation of census data, and diversion of resources—all of which turn on their expectation that reinstating a citizenship question will depress the census response rate and lead to an inaccurate population count. Several States with a disproportionate share of noncitizens, for example, anticipate losing a seat in Congress or qualifying for less federal funding if their populations are undercounted. These are primarily future injuries, which “may suffice if the threatened injury is certainly impending, or there is a substantial risk that the harm will occur.” Susan B. Anthony List v. Driehaus, 573 U.S. 149, 158 (2014) (internal quotation marks omitted). The District Court concluded that the evidence at trial established a sufficient likelihood that the reinstatement of a citizenship question would result in noncitizen households responding to the census at lower rates than other groups, which in turn would cause them to be undercounted and lead to many of respondents’ asserted injuries. For purposes of standing, these findings of fact were not so suspect as to be clearly erroneous. We therefore agree that at least some respondents have Article III standing. Several state respondents here have shown that if noncitizen households are undercounted by as little as 2%—lower than the District Court’s 5.8% prediction—they will lose out on federal funds that are distributed on the basis of state population. That is a sufficiently concrete and imminent injury to satisfy Article III, and there is no dispute that a ruling in favor of respondents would redress that harm. The Government contends, however, that any harm to respondents is not fairly traceable to the Secretary’s decision, because such harm depends on the independent action of third parties choosing to violate their legal duty to respond to the census. The chain of causation is made even more tenuous, the Government argues, by the fact that such intervening, unlawful third-party action would be motivated by unfounded fears that the Federal Government will itself break the law by using noncitizens’ answers against them for law enforcement purposes. The Government invokes our steady refusal to “endorse standing theories that rest on speculation about the decisions of independent actors,” Clapper v. Amnesty Int’l USA, 568 U.S. 398, 414 (2013), particularly speculation about future unlawful conduct, Los Angeles v. Lyons, 461 U.S. 95, 105 (1983). But we are satisfied that, in these circumstances, respondents have met their burden of showing that third parties will likely react in predictable ways to the citizenship question, even if they do so unlawfully and despite the requirement that the Government keep individual answers confidential. The evidence at trial established that noncitizen households have historically responded to the census at lower rates than other groups, and the District Court did not clearly err in crediting the Census Bureau’s theory that the discrepancy is likely attributable at least in part to noncitizens’ reluctance to answer a citizenship question. Respondents’ theory of standing thus does not rest on mere speculation about the decisions of third parties; it relies instead on the predictable effect of Government action on the decisions of third parties. Cf. Bennett v. Spear, 520 U.S. 154, 169–170 (1997); Davis, 554 U. S., at 734–735. Because Article III “requires no more than de facto causality,” Block v. Meese, 793 F.2d 1303, 1309 (CADC 1986) (Scalia, J.), traceability is satisfied here. We may therefore consider the merits of respondents’ claims, at least as far as the Constitution is concerned. III The Enumeration Clause of the Constitution does not provide a basis to set aside the Secretary’s decision. The text of that clause “vests Congress with virtually unlimited discretion in conducting the decennial ‘actual Enumeration,’ ” and Congress “has delegated its broad authority over the census to the Secretary.” Wisconsin, 517 U. S., at 19. Given that expansive grant of authority, we have rejected challenges to the conduct of the census where the Secretary’s decisions bore a “reasonable relationship to the accomplishment of an actual enumeration.” Id., at 20. Respondents ask us to evaluate the Secretary’s decision to reinstate a citizenship question under that “reasonable relationship” standard, but we agree with the District Court that a different analysis is needed here. Our cases applying that standard concerned decisions about the population count itself—such as a postcensus decision not to use a particular method to adjust an undercount, id., at 4, and a decision to allocate overseas military personnel to their home States, Franklin v. Massachusetts, 505 U.S. 788, 790–791 (1992). We have never applied the standard to decisions about what kinds of demographic information to collect in the course of taking the census. Indeed, as the District Court recognized, applying the “reasonable relationship” standard to every census-related decision “would lead to the conclusion that it is unconstitutional to ask any demographic question on the census” because “asking such questions bears no relationship whatsoever to the goal of an accurate headcount.” 315 F. Supp. 3d, at 804–805. Yet demographic questions have been asked in every census since 1790, and questions about citizenship in particular have been asked for nearly as long. Like the District Court, we decline respondents’ invitation to measure the constitutionality of the citizenship question by a stand- ard that would seem to render every census since 1790 unconstitutional. We look instead to Congress’s broad authority over the census, as informed by long and consistent historical practice. All three branches of Government have understood the Constitution to allow Congress, and by extension the Secretary, to use the census for more than simply counting the population. Since 1790, Congress has sought, or permitted the Secretary to seek, information about matters as varied as age, sex, marital status, health, trade, profession, literacy, and value of real estate owned. See id., at 801. Since 1820, it has sought, or permitted the Secretary to seek, information about citizenship in particular. Federal courts have approved the practice of collecting demographic data in the census. See, e.g., United States v. Moriarity, 106 F. 886, 891 (CC SDNY 1901) (duty to take a census of population “does not prohibit the gathering of other statistics, if ‘necessary and proper,’ for the intelligent exercise of other powers enumerated in the constitution”). While we have never faced the question directly, we have assumed that Congress has the power to use the census for information-gathering purposes, see Legal Tender Cases, 12 Wall. 457, 536 (1871), and we have recognized the role of the census as a “linchpin of the federal statistical system by collecting data on the characteristics of individuals, households, and housing units throughout the country,” Department of Commerce v. United States House of Representatives, 525 U.S. 316, 341 (1999) (internal quotation marks omitted). That history matters. Here, as in other areas, our interpretation of the Constitution is guided by a Government practice that “has been open, widespread, and unchallenged since the early days of the Republic.” NLRB v. Noel Canning, 573 U.S. 513, 572 (2014) (Scalia, J., concurring in judgment); see also Wisconsin, 517 U. S., at 21 (noting “importance of historical practice” in census context). In light of the early understanding of and long practice under the Enumeration Clause, we conclude that it permits Congress, and by extension the Secretary, to inquire about citizenship on the census questionnaire. We need not, and do not, decide the constitutionality of any other question that Congress or the Secretary might decide to include in the census. IV The District Court set aside the Secretary’s decision to reinstate a citizenship question on the grounds that the Secretary acted arbitrarily and violated certain provisions of the Census Act. The Government contests those rulings, but also argues that the Secretary’s decision was not judicially reviewable under the Administrative Procedure Act in the first place. We begin with that contention. A The Administrative Procedure Act embodies a “basic presumption of judicial review,” Abbott Laboratories v. Gardner, 387 U.S. 136, 140 (1967), and instructs reviewing courts to set aside agency action that is “arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law,” 5 U. S. C. §706(2)(A). Review is not available, however, “to the extent that” a relevant statute precludes it, §701(a)(1), or the agency action is “committed to agency discretion by law,” §701(a)(2). The Government argues that the Census Act commits to the Secretary’s unreviewable discretion decisions about what questions to include on the decennial census questionnaire. We disagree. To be sure, the Act confers broad authority on the Secretary. Section 141(a) instructs him to take “a decennial census of population” in “such form and content as he may determine, including the use of sampling procedures and special surveys.” 13 U. S. C. §141. The Act defines “census of population” to mean “a census of population, housing, and matters relating to population and housing,” §141(g), and it authorizes the Secretary, in “connection with any such census,” to “obtain such other census information as necessary,” §141(a). It also states that the “Secretary shall prepare questionnaires, and shall determine the inquiries, and the number, form, and subdivisions thereof, for the statistics, surveys, and censuses provided for in this title.” §5. And it authorizes him to acquire materials, such as administrative records, from other federal, state, and local agencies in aid of conducting the census. §6. Those provisions leave much to the Secretary’s discretion. See Wisconsin, 517 U. S., at 19 (“Through the Census Act, Congress has delegated its broad authority over the census to the Secretary.”). But they do not leave his discretion unbounded. In order to give effect to the command that courts set aside agency action that is an abuse of discretion, and to honor the presumption of judicial review, we have read the §701(a)(2) exception for action committed to agency discretion “quite narrowly, restricting it to ‘those rare circumstances where the relevant statute is drawn so that a court would have no meaningful standard against which to judge the agency’s exercise of discretion.’ ” Weyerhaeuser Co. v. United States Fish and Wildlife Serv., 586 U. S. ___, ___ (2018) (slip op., at 12) (quoting Lincoln v. Vigil, 508 U.S. 182, 191 (1993)). And we have generally limited the exception to “certain categories of administrative decisions that courts traditionally have regarded as ‘committed to agency discretion,’ ” id., at 191, such as a decision not to institute enforcement proceedings, Heckler v. Chaney, 470 U.S. 821, 831–832 (1985), or a decision by an intelligence agency to terminate an employee in the interest of national security, Webster v. Doe, 486 U.S. 592, 600–601 (1988). The taking of the census is not one of those areas traditionally committed to agency discretion. We and other courts have entertained both constitutional and statutory challenges to census-related decisionmaking. See, e.g., Department of Commerce, 525 U.S. 316; Wisconsin, 517 U.S. 1; Carey v. Klutznick, 637 F.2d 834 (CA2 1980). Nor is the statute here drawn so that it furnishes no meaningful standard by which to judge the Secretary’s action. In contrast to the National Security Act in Webster, which gave the Director of Central Intelligence discretion to terminate employees whenever he “deem[ed]” it “advisable,” 486 U. S., at 594, the Census Act constrains the Secretary’s authority to determine the form and content of the census in a number of ways. Section 195, for example, governs the extent to which he can use statistical sampling. Section 6(c), which will be considered in more detail below, circumscribes his power in certain circumstances to collect information through direct inquiries when administrative records are available. More generally, by mandating a population count that will be used to apportion representatives, see §141(b), 2 U. S. C. §2a, the Act imposes “a duty to conduct a census that is accurate and that fairly accounts for the crucial representational rights that depend on the census and the apportionment.” Franklin, 505 U. S., at 819–820 (Stevens, J., concurring in part and concurring in judgment). The Secretary’s decision to reinstate a citizenship question is amenable to review for compliance with those and other provisions of the Census Act, according to the general requirements of reasoned agency decisionmaking. Because this is not a case in which there is “no law to apply,” Overton Park, 401 U. S., at 410, the Secretary’s decision is subject to judicial review. B At the heart of this suit is respondents’ claim that the Secretary abused his discretion in deciding to reinstate a citizenship question. We review the Secretary’s exercise of discretion under the deferential “arbitrary and capricious” standard. See 5 U. S. C. §706(2)(A). Our scope of review is “narrow”: we determine only whether the Secretary examined “the relevant data” and articulated “a satisfactory explanation” for his decision, “including a rational connection between the facts found and the choice made.” Motor Vehicle Mfrs. Assn. of United States, Inc. v. State Farm Mut. Automobile Ins. Co., 463 U.S. 29, 43 (1983) (internal quotation marks omitted). We may not substitute our judgment for that of the Secretary, ibid., but instead must confine ourselves to ensuring that he remained “within the bounds of reasoned decisionmaking,” Baltimore Gas & Elec. Co. v. Natural Resources Defense Council, Inc., 462 U.S. 87, 105 (1983). The District Court set aside the Secretary’s decision for two independent reasons: His course of action was not supported by the evidence before him, and his stated rationale was pretextual. We focus on the first point here and take up the question of pretext later. The Secretary examined the Bureau’s analysis of various ways to collect improved citizenship data and explained why he thought the best course was to both reinstate a citizenship question and use citizenship data from administrative records to fill in the gaps. He considered but rejected the Bureau’s recommendation to use administrative records alone. As he explained, records are lacking for about 10% of the population, so the Bureau would still need to estimate citizenship for millions of voting-age people. Asking a citizenship question of everyone, the Secretary reasoned, would eliminate the need to estimate citizenship for many of those people. And supplementing census responses with administrative record data would help complete the picture and allow the Bureau to better estimate citizenship for the smaller set of cases where it was still necessary to do so. The evidence before the Secretary supported that decision. As the Bureau acknowledged, each approach—using administrative records alone, or asking about citizenship and using records to fill in the gaps—entailed tradeoffs between accuracy and completeness. Without a citizenship question, the Bureau would need to estimate the citizenship of about 35 million people; with a citizenship question, it would need to estimate the citizenship of only 13.8 million. Under either approach, there would be some errors in both the administrative records and the Bureau’s estimates. With a citizenship question, there would also be some erroneous self-responses (about 500,000) and some conflicts between responses and administrative record data (about 9.5 million). The Bureau explained that the “relative quality” of the citizenship data generated by each approach would depend on the “relative importance of the errors” in each, but it was not able to “quantify the relative magnitude of the errors across the alternatives.” App. 148. The Bureau nonetheless recommended using administrative records alone because it had “high confidence” that it could develop an accurate model for estimating the citizenship of the 35 million people for whom administrative records were not available, and it thought the resulting citizenship data would be of superior quality. Id., at 146, 158–159. But when the time came for the Secretary to make a decision, the model did not yet exist, and even if it had, there was no way to gauge its relative accuracy. As the Bureau put it, “we will most likely never possess a fully adequate truth deck to benchmark” the model—which appears to be bureaucratese for “maybe, maybe not.” Id., at 146. The Secretary opted instead for the approach that would yield a more complete set of data at an acceptable rate of accuracy, and would require estimating the citizenship of fewer people. The District Court overruled that choice, agreeing with the Bureau’s assessment that its recommended approach would yield higher quality citizenship data on the whole. But the choice between reasonable policy alternatives in the face of uncertainty was the Secretary’s to make. He considered the relevant factors, weighed risks and benefits, and articulated a satisfactory explanation for his decision. In overriding that reasonable exercise of discretion, the court improperly substituted its judgment for that of the agency. The Secretary then weighed the benefit of collecting more complete and accurate citizenship data against the risk that inquiring about citizenship would depress census response rates, particularly among noncitizen households. In the Secretary’s view, that risk was difficult to assess. The Bureau predicted a 5.1% decline in response rates among noncitizen households if the citizenship question were reinstated.[2] It relied for that prediction primarily on studies showing that, while noncitizens had responded at lower rates than citizens to the 2000 short-form and 2010 censuses, which did not ask about citizenship, they responded at even lower rates than citizens to the 2000 long-form census and the 2010 American Community Survey, which did ask about citizenship. The Bureau thought it was reasonable to infer that the citizenship question accounted for the differential decline in noncitizen responses. But, the Secretary explained, the Bureau was unable to rule out other causes. For one thing, the evidence before the Secretary suggested that noncitizen households tend to be more distrustful of, and less likely to respond to, any government effort to collect information. For another, both the 2000 long-form census and 2010 ACS asked over 45 questions on a range of topics, including employment, income, and housing characteristics. Noncitizen households might disproportionately fail to respond to a lengthy and intrusive Government questionnaire for a number of reasons besides reluctance to answer a citizenship question—reasons relating to education level, socioeconomic status, and less exposure to Government outreach efforts. See App. to Pet. for Cert. 553a–554a, 557a–558a. The Secretary justifiably found the Bureau’s analysis inconclusive. Weighing that uncertainty against the value of obtaining more complete and accurate citizenship data, he determined that reinstating a citizenship question was worth the risk of a potentially lower response rate. That decision was reasonable and reasonably explained, particularly in light of the long history of the citizenship question on the census. Justice Breyer would conclude otherwise, but only by subordinating the Secretary’s policymaking discretion to the Bureau’s technocratic expertise. Justice Breyer’s analysis treats the Bureau’s (pessimistic) prediction about response rates and (optimistic) assumptions about its data modeling abilities as touchstones of substantive reason- ableness rather than simply evidence for the Secretary to consider. He suggests that the Secretary should have deferred to the Bureau or at least offered some special justification for drawing his own inferences and adopting his own assumptions. But the Census Act authorizes the Secretary, not the Bureau, to make policy choices within the range of reasonable options. And the evidence before the Secretary hardly led ineluctably to just one reasonable course of action. It called for value-laden decisionmaking and the weighing of incommensurables under conditions of uncertainty. The Secretary was required to consider the evidence and give reasons for his chosen course of action. He did so. It is not for us to ask whether his decision was “the best one possible” or even whether it was “better than the alternatives.” FERC v. Electric Power Supply Assn., 577 U. S. ___, ___ (2016) (slip op., at 30). By second-guessing the Secretary’s weighing of risks and benefits and penalizing him for departing from the Bureau’s inferences and assumptions, Justice Breyer—like the District Court—substitutes his judgment for that of the agency. C The District Court also ruled that the Secretary violated two particular provisions of the Census Act, §6(c) and §141(f). Section 6 has three subsections. Subsections (a) and (b) authorize the Secretary to acquire administrative records from other federal agencies and from state and local governments.[3] Subsection (c) states: “To the maximum extent possible and consistent with the kind, timeliness, quality and scope of the statistics required, the Secretary shall acquire and use information available from any source referred to in subsection (a) or (b) of this section instead of conducting direct inquiries.” 13 U. S. C. §6(c). The District Court held, and respondents argue, that the Secretary failed to comply with §6(c) because he opted to collect citizenship data using direct inquiries when it was possible to provide DOJ with data from administrative records alone. At the outset, §6(c) may not even apply here. It governs the Secretary’s choices with respect to “statistics required.” The parties have assumed that phrase refers to census-related data that the Secretary wishes to acquire, but it may instead refer to particular kinds of statistics that other provisions of the Census Act actually do require the Secretary to collect and publish. See, e.g., §41 (“The Secretary shall collect and publish statistics concerning [cotton and cotton production].”); §61 (“The Secretary shall collect, collate, and publish monthly statistics concerning [vegetable and animal oils and the like].”); §91 (“The Secretary shall collect and publish quarterly financial statistics of business operations, organization, practices, management, and relation to other businesses.”). If so, §6(c) would seem to have nothing to say about the Secretary’s collection of census-related citizenship data, which is not a “statistic” he is “required” to collect. Regardless, assuming the provision applies, the Secretary complied with it, for essentially the same reasons that his decision was not arbitrary and capricious. As he explained, administrative records would not, in his judgment, provide the more complete and accurate data that DOJ sought. He thus could not, “consistent with” the kind and quality of the “statistics required,” use administrative records instead of asking about citizenship directly. Respondents’ arguments to the contrary rehash their dis- agreement with the Secretary’s policy judgment about which approach would yield the most complete and accurate citizenship data. For the reasons already discussed, we may not substitute our judgment for that of the Secretary here. We turn now to §141(f), which requires the Secretary to report to Congress about his plans for the census. Paragraph (1) instructs him to submit, at least three years before the census date, a report containing his “determination of the subjects proposed to be included, and the types of information to be compiled,” in the census. Paragraph (2) then tells him to submit, at least two years before the census date, a report containing his “determination of the questions proposed to be included” in the census. Paragraph (3) provides: “[A]fter submission of a report under paragraph (1) or (2) of this subsection and before the appropriate census date, if the Secretary finds new circumstances exist which necessitate that the subjects, types of information, or questions contained in reports so submitted be modified, [he shall submit] areport containing the Secretary’s determination of the subjects, types of information, or questions as proposed to be modified.” The Secretary timely submitted his paragraph (1) report in March 2017. It did not mention citizenship. In December 2017, he received DOJ’s formal request. Three months later, in March 2018, he timely submitted his para- graph (2) report. It did propose asking a question about citizenship. The District Court held that the Secretary’s failure to mention citizenship in his March 2017 report violated §141(f)(1) and provided an independent basis to set aside his action. Assuming without deciding that the Secretary’s compliance with the reporting requirement is for courts—rather than Congress—to police, we disagree. The Secretary’s March 2018 report satisfied the requirements of paragraph (3): By informing Congress that he proposed to include a citizenship question, the Secretary necessarily also informed Congress that he proposed to modify the original list of subjects that he submitted in the March 2017 report. Nothing in §141(f) suggests that the same report cannot simultaneously fulfill the requirements of paragraphs (2) and (3). And to the extent paragraph (3) requires the Secretary to explain his finding of new circumstances, he did so in his March 2018 memo, which described DOJ’s intervening request. In any event, even if we agreed with the District Court that the Secretary technically violated §141(f) by submitting a paragraph (2) report that doubled as a paragraph (3) report, the error would surely be harmless in these circumstances, where the Secretary nonetheless fully informed Congress of, and explained, his decision. See 5 U. S. C. §706 (in reviewing agency action, “due account shall be taken of the rule of prejudicial error”). V We now consider the District Court’s determination that the Secretary’s decision must be set aside because it rested on a pretextual basis, which the Government conceded below would warrant a remand to the agency. We start with settled propositions. First, in order to permit meaningful judicial review, an agency must “disclose the basis” of its action. Burlington Truck Lines, Inc. v. United States, 371 U.S. 156, 167–169 (1962) (internal quotation marks omitted); see also SEC v. Chenery Corp., 318 U.S. 80, 94 (1943) (“[T]he orderly functioning of the process of review requires that the grounds upon which the administrative agency acted be clearly disclosed and adequately sustained.”). Second, in reviewing agency action, a court is ordinarily limited to evaluating the agency’s contemporaneous explanation in light of the existing administrative record. Vermont Yankee Nuclear Power Corp. v. Natural Resources Defense Council, Inc., 435 U.S. 519, 549 (1978); Camp v. Pitts, 411 U.S. 138, 142–143 (1973) (per curiam). That principle reflects the recognition that further judicial inquiry into “executive motivation” represents “a substantial intrusion” into the workings of another branch of Government and should normally be avoided. Arlington Heights v. Metropolitan Housing Development Corp., 429 U.S. 252, 268, n. 18 (1977); see Overton Park, 401 U. S., at 420. Third, a court may not reject an agency’s stated reasons for acting simply because the agency might also have had other unstated reasons. See Jagers v. Federal Crop Ins. Corp., 758 F.3d 1179, 1185–1186 (CA10 2014) (rejecting argument that “the agency’s subjective desire to reach a particular result must necessarily invalidate the result, regardless of the objective evidence supporting the agency’s conclusion”). Relatedly, a court may not set aside an agency’s policymaking decision solely because it might have been influenced by political considerations or prompted by an Administration’s priorities. Agency policymaking is not a “rarified technocratic process, unaffected by political considerations or the presence of Presidential power.” Sierra Club v. Costle, 657 F.2d 298, 408 (CADC 1981). Such decisions are routinely informed by unstated considerations of politics, the legislative process, public relations, interest group relations, foreign relations, and national security concerns (among others). Finally, we have recognized a narrow exception to the general rule against inquiring into “the mental processes of administrative decisionmakers.” Overton Park, 401 U. S., at 420. On a “strong showing of bad faith or improper behavior,” such an inquiry may be warranted and may justify extra-record discovery. Ibid. The District Court invoked that exception in ordering extra-record discovery here. Although that order was premature, we think it was ultimately justified in light of the expanded administrative record. Recall that shortly after this litigation began, the Secretary, prodded by DOJ, filed a supplemental memo that added new, pertinent information to the administrative record. The memo disclosed that the Secretary had been considering the citizenship question for some time and that Commerce had inquired whether DOJ would formally request reinstatement of the question. That supplemental memo prompted respondents to move for both completion of the administrative record and extra-record discovery. The District Court granted both requests at the same hearing, agreeing with respondents that the Government had submitted an incomplete administrative record and that the existing evidence supported a prima facie showing that the VRA rationale was pretextual. The Government did not challenge the court’s conclusion that the administrative record was incomplete, and the parties stipulated to the inclusion of more than 12,000 pages of internal deliberative materials as part of the administrative record, materials that the court later held were sufficient on their own to demonstrate pretext. The Government did, however, challenge the District Court’s order authorizing extra-record discovery, as well as the court’s later orders compelling depositions of the Secretary and of the Acting Assistant Attorney General for DOJ’s Civil Rights Division. We agree with the Government that the District Court should not have ordered extra-record discovery when it did. At that time, the most that was warranted was the order to complete the administrative record. But the new material that the parties stipulated should have been part of the administrative record—which showed, among other things, that the VRA played an insignificant role in the decisionmaking process—largely justified such extra-record discovery as occurred (which did not include the deposition of the Secretary himself). We accordingly review the District Court’s ruling on pretext in light of all the evidence in the record before the court, including the extra-record discovery. That evidence showed that the Secretary was determined to reinstate a citizenship question from the time he entered office; instructed his staff to make it happen; waited while Commerce officials explored whether another agency would request census-based citizenship data; subsequently contacted the Attorney General himself to ask if DOJ would make the request; and adopted the Voting Rights Act rationale late in the process. In the District Court’s view, this evidence established that the Secretary had made up his mind to reinstate a citizenship question “well before” receiving DOJ’s request, and did so for reasons unknown but unrelated to the VRA. 351 F. Supp. 3d, at 660. The Government, on the other hand, contends that there was nothing objectionable or even surprising in this. And we agree—to a point. It is hardly improper for an agency head to come into office with policy preferences and ideas, discuss them with affected parties, sound out other agencies for support, and work with staff attorneys to substantiate the legal basis for a preferred policy. The record here reflects the sometimes involved nature of Executive Branch decisionmaking, but no particular step in the process stands out as inappropriate or defective. And yet, viewing the evidence as a whole, we share the District Court’s conviction that the decision to reinstate a citizenship question cannot be adequately explained in terms of DOJ’s request for improved citizenship data to better enforce the VRA. Several points, considered together, reveal a significant mismatch between the decision the Secretary made and the rationale he provided. The record shows that the Secretary began taking steps to reinstate a citizenship question about a week into his tenure, but it contains no hint that he was considering VRA enforcement in connection with that project. The Secretary’s Director of Policy did not know why the Secretary wished to reinstate the question, but saw it as his task to “find the best rationale.” Id., at 551. The Director initially attempted to elicit requests for citizenship data from the Department of Homeland Security and DOJ’s Executive Office for Immigration Review, neither of which is responsible for enforcing the VRA. After those attempts failed, he asked Commerce staff to look into whether the Secretary could reinstate the question without receiving a request from another agency. The possibility that DOJ’s Civil Rights Division might be willing to request citizenship data for VRA enforcement purposes was proposed by Commerce staff along the way and eventually pursued. Even so, it was not until the Secretary contacted the Attorney General directly that DOJ’s Civil Rights Division expressed interest in acquiring census-based citizenship data to better enforce the VRA. And even then, the record suggests that DOJ’s interest was directed more to helping the Commerce Department than to securing the data. The December 2017 letter from DOJ drew heavily on contributions from Commerce staff and advisors. Their influence may explain why the letter went beyond a simple entreaty for better citizenship data—what one might expect of a typical request from another agency—to a specific request that Commerce collect the data by means of reinstating a citizenship question on the census. Finally, after sending the letter, DOJ declined the Census Bureau’s offer to discuss alternative ways to meet DOJ’s stated need for improved citizenship data, further suggesting a lack of interest on DOJ’s part. Altogether, the evidence tells a story that does not match the explanation the Secretary gave for his decision. In the Secretary’s telling, Commerce was simply acting on a routine data request from another agency. Yet the materials before us indicate that Commerce went to great lengths to elicit the request from DOJ (or any other willing agency). And unlike a typical case in which an agency may have both stated and unstated reasons for a decision, here the VRA enforcement rationale—the sole stated reason—seems to have been contrived. We are presented, in other words, with an explanation for agency action that is incongruent with what the record reveals about the agency’s priorities and decisionmaking process. It is rare to review a record as extensive as the one before us when evaluating informal agency action—and it should be. But having done so for the sufficient reasons we have explained, we cannot ignore the disconnect between the decision made and the explanation given. Our review is deferential, but we are “not required to exhibit a naiveté from which ordinary citizens are free.” United States v. Stanchich, 550 F.2d 1294, 1300 (CA2 1977) (Friendly, J.). The reasoned explanation requirement of administrative law, after all, is meant to ensure that agencies offer genuine justifications for important decisions, reasons that can be scrutinized by courts and the interested public. Accepting contrived reasons would defeat the purpose of the enterprise. If judicial review is to be more than an empty ritual, it must demand something better than the explanation offered for the action taken in this case. In these unusual circumstances, the District Court was warranted in remanding to the agency, and we affirm that disposition. See Florida Power & Light Co. v. Lorion, 470 U.S. 729, 744 (1985). We do not hold that the agency decision here was substantively invalid. But agencies must pursue their goals reasonably. Reasoned decisionmaking under the Administrative Procedure Act calls for an explanation for agency action. What was provided here was more of a distraction. * * * The judgment of the United States District Court for the Southern District of New York is affirmed in part and reversed in part, and the case is remanded for further proceedings consistent with this opinion. It is so ordered. Notes 1 The annual alien registration requirement was repealed in 1981. See §11, 95Stat. 1617 (1981). 2 Several months after the Secretary made his decision, the Bureau updated its prediction to 5.8%, the figure the District Court later relied on in its standing analysis. See 351 F. Supp. 3d 502, 579 (SDNY 2019). 3 The full text of subsections (a) and (b) provides: “(a) The Secretary, whenever he considers it advisable, may call upon any other department, agency, or establishment of the FederalGovernment, or of the government of the District of Columbia, for information pertinent to the work provided for in this title. “(b) The Secretary may acquire, by purchase or otherwise, from States, counties, cities, or other units of government, or their instrumentalities, or from private persons and agencies, such copies of records, reports, and other material as may be required for the efficient and economical conduct of the censuses and surveys provided for in this title.” 13 U. S. C. §6.

588.US.2018_17-9572

Petitioner Curtis Flowers has been tried six separate times for the murder of four employees of a Mississippi furniture store. Flowers is black; three of the four victims were white. At the first two trials, the State used its peremptory strikes on all of the qualified black prospective jurors. In each case, the jury convicted Flowers and sentenced him to death, but the convictions were later reversed by the Mississippi Supreme Court based on prosecutorial misconduct. At the third trial, the State used all of its 15 peremptory strikes against black prospective jurors, and the jury convicted Flowers and sentenced him to death. The Mississippi Supreme Court reversed again, this time concluding that the State exercised its peremptory strikes on the basis of race in violation of Batson v. Kentucky, 476 U.S. 79. Flowers’ fourth and fifth trials ended in mistrials. At the fourth, the State exercised 11 peremptory strikes—all against black prospective jurors. No available racial information exists about the prospective jurors in the fifth trial. At the sixth trial, the State exercised six peremptory strikes—five against black prospective jurors, allowing one black juror to be seated. Flowers again raised a Batson claim, but the trial court concluded that the State had offered race-neutral reasons for each of the five peremptory strikes. The jury convicted Flowers and sentenced him to death. The Mississippi Supreme Court affirmed. After this Court vacated that judgment and remanded in light of Foster v. Chatman, 578 U. S. ___, the Mississippi Supreme Court again upheld Flowers’ conviction in a divided 5-to-4 decision. Justice King dissented on the Batson issue and was joined by two other Justices. Held: All of the relevant facts and circumstances taken together establish that the trial court at Flowers’ sixth trial committed clear error in concluding that the State’s peremptory strike of black prospective juror Carolyn Wright was not motivated in substantial part by discriminatory intent. Pp. 7–31. (a) Under Batson, once a prima facie case of discrimination has been shown by a defendant, the State must provide race-neutral reasons for its peremptory strikes. The trial judge then must determine whether the prosecutor’s stated reasons were the actual reasons or instead were a pretext for discrimination. The Batson Court rejected four arguments. First, the Batson Court rejected the idea that a defendant must demonstrate a history of racially discriminatory strikes in order to make out a claim of race discrimination. Second, the Batson Court rejected the argument that a prosecutor could strike a black juror based on an assumption or belief that the black juror would favor a black defendant. Third, the Batson Court rejected the argument that race-based peremptories should be permissible because black, white, Asian, and Hispanic defendants and jurors were all “equally” subject to race-based discrimination. Fourth, the Batson Court rejected the argument that race-based peremptories are permissible because both the prosecution and defense could employ them in any individual case and in essence balance things out. Pp. 7–15. (b) Four categories of evidence loom large in assessing the Batson issue here, where the State had a persistent pattern of striking black prospective jurors from Flowers’ first through his sixth trial. Pp. 15–30. (1) A review of the history of the State’s peremptory strikes in Flowers’ first four trials strongly supports the conclusion that the State’s use of peremptory strikes in Flowers’ sixth trial was motivated in substantial part by discriminatory intent. The State tried to strike all 36 black prospective jurors over the course of the first four trials. And the state courts themselves concluded that the State had violated Batson on two separate occasions. The State’s relentless, determined effort to rid the jury of black individuals strongly suggests that the State wanted to try Flowers before a jury with as few black jurors as possible, and ideally before an all-white jury. Pp. 19–22. (2) The State’s use of peremptory strikes in Flowers’ sixth trial followed the same pattern as the first four trials. Pp. 22–23. (3) Disparate questioning can be probative of discriminatory intent. Miller-El v. Cockrell, 537 U.S. 322, 331–332, 344–345. Here, the State spent far more time questioning the black prospective jurors than the accepted white jurors—145 questions asked of 5 black prospective jurors and 12 questions asked of 11 white seated jurors. The record refutes the State’s explanation that it questioned black and white prospective jurors differently only because of differences in the jurors’ characteristics. Along with the historical evidence from the earlier trials, as well as the State’s striking of five of six black prospective jurors at the sixth trial, the dramatically disparate questioning and investigation of black prospective jurors and white prospective jurors at the sixth trial strongly suggest that the State was motivated in substantial part by a discriminatory intent. Pp. 23–26. (4) Comparing prospective jurors who were struck and not struck can be an important step in determining whether a Batson violation occurred. See Snyder v. Louisiana, 552 U.S. 472, 483–484. Here, Carolyn Wright, a black prospective juror, was struck, the State says, in part because she knew several defense witnesses and had worked at Wal-Mart where Flowers’ father also worked. But three white prospective jurors also knew many individuals involved in the case, and the State asked them no individual questions about their connections to witnesses. White prospective jurors also had relationships with members of Flowers’ family, but the State did not ask them follow-up questions in order to explore the depth of those relationships. The State also incorrectly explained that it exercised a peremptory strike against Wright because she had worked with one of Flowers’ sisters and made apparently incorrect statements to justify the strikes of other black prospective jurors. When considered with other evidence, a series of factually inaccurate explanations for striking black prospective jurors can be another clue showing discriminatory intent. The overall context here requires skepticism of the State’s strike of Carolyn Wright. The trial court at Flowers’ sixth trial committed clear error in concluding that the State’s peremptory strike of black prospective juror Carolyn Wright was not motivated in substantial part by discriminatory intent. Pp. 26–30. 240 So. 3d 1082, reversed and remanded. Kavanaugh, J., delivered the opinion of the Court, in which Roberts, C. J., and Ginsburg, Breyer, Alito, Sotomayor, and Kagan, JJ., joined. Alito, J., filed a concurring opinion. Thomas, J., filed a dissenting opinion, in which Gorsuch, J., joined as to Parts I, II, and III.

In Batson v. Kentucky, 476 U.S. 79 (1986), this Court ruled that a State may not discriminate on the basis of race when exercising peremptory challenges against prospective jurors in a criminal trial. In 1996, Curtis Flowers allegedly murdered four people in Winona, Mississippi. Flowers is black. He has been tried six separate times before a jury for murder. The same lead prosecutor represented the State in all six trials. In the initial three trials, Flowers was convicted, but the Mississippi Supreme Court reversed each conviction. In the first trial, Flowers was convicted, but the Mississippi Supreme Court reversed the conviction due to “numerous instances of prosecutorial misconduct.” Flowers v. State, 773 So. 2d 309, 327 (2000). In the second trial, the trial court found that the prosecutor discriminated on the basis of race in the peremptory challenge of a black juror. The trial court seated the black juror. Flowers was then convicted, but the Mississippi Supreme Court again reversed the conviction because of prosecutorial misconduct at trial. In the third trial, Flowers was convicted, but the Mississippi Supreme Court yet again reversed the conviction, this time because the court concluded that the prosecutor had again discriminated against black prospective jurors in the jury selection process. The court’s lead opinion stated: “The instant case presents us with as strong a prima facie case of racial discrimination as we have ever seen in the context of a Batson challenge.” Flowers v. State, 947 So. 2d 910, 935 (2007). The opinion further stated that the “State engaged in racially discriminatory practices during the jury selection process” and that the “case evinces an effort by the State to exclude African-Americans from jury service.” Id., at 937, 939. The fourth and fifth trials of Flowers ended in mistrials due to hung juries. In his sixth trial, which is the one at issue here, Flowers was convicted. The State struck five of the six black prospective jurors. On appeal, Flowers argued that the State again violated Batson in exercising peremptory strikes against black prospective jurors. In a divided 5-to-4 decision, the Mississippi Supreme Court affirmed the conviction. We granted certiorari on the Batson question and now reverse. See 586 U. S. ___ (2018). Four critical facts, taken together, require reversal. First, in the six trials combined, the State employed its peremptory challenges to strike 41 of the 42 black prospective jurors that it could have struck—a statistic that the State acknowledged at oral argument in this Court. Tr. of Oral Arg. 32. Second, in the most recent trial, the sixth trial, the State exercised peremptory strikes against five of the six black prospective jurors. Third, at the sixth trial, in an apparent effort to find pretextual reasons to strike black prospective jurors, the State engaged in dramatically disparate questioning of black and white prospective jurors. Fourth, the State then struck at least one black prospective juror, Carolyn Wright, who was similarly situated to white prospective jurors who were not struck by the State. We need not and do not decide that any one of those four facts alone would require reversal. All that we need to decide, and all that we do decide, is that all of the relevant facts and circumstances taken together establish that the trial court committed clear error in concluding that the State’s peremptory strike of black prospective juror Carolyn Wright was not “motivated in substantial part by discriminatory intent.” Foster v. Chatman, 578 U. S. ___, ___ (2016) (slip op., at 23) (internal quotation marks omitted). In reaching that conclusion, we break no new legal ground. We simply enforce and reinforce Batson by applying it to the extraordinary facts of this case. We reverse the judgment of the Supreme Court of Mississippi, and we remand the case for further proceedings not inconsistent with this opinion. I The underlying events that gave rise to this case took place in Winona, Mississippi. Winona is a small town in northern Mississippi, just off I–55 almost halfway between Jackson and Memphis. The total population of Winona is about 5,000. The town is about 53 percent black and about 46 percent white. In 1996, Bertha Tardy, Robert Golden, Derrick Stewart, and Carmen Rigby were murdered at the Tardy Furniture store in Winona. All four victims worked at the Tardy Furniture store. Three of the four victims were white; one was black. In 1997, the State charged Curtis Flowers with murder. Flowers is black. Since then, Flowers has been tried six separate times for the murders. In each of the first two trials, Flowers was tried for one individual murder. In each subsequent trial, Flowers was tried for all four of the murders together. The same state prosecutor tried Flowers each time. The prosecutor is white. At Flowers’ first trial, 36 prospective jurors—5 black and 31 white—were presented to potentially serve on the jury. The State exercised a total of 12 peremptory strikes, and it used 5 of them to strike the five qualified black prospective jurors. Flowers objected, arguing under Batson that the State had exercised its peremptory strikes in a racially discriminatory manner. The trial court rejected the Batson challenge. Because the trial court allowed the State’s peremptory strikes, Flowers was tried in front of an all-white jury. The jury convicted Flowers and sentenced him to death. On appeal, the Mississippi Supreme Court reversed the conviction, concluding that the State had committed prosecutorial misconduct in front of the jury by, among other things, expressing baseless grounds for doubting the credibility of witnesses and mentioning facts that had not been allowed into evidence by the trial judge. Flowers, 773 So. 2d, at 317, 334. In its opinion, the Mississippi Supreme Court described “numerous instances of prosecutorial misconduct” at the trial. Id., at 327. Because the Mississippi Supreme Court reversed based on prosecutorial misconduct at trial, the court did not reach Flowers’ Batson argument. See Flowers, 773 So. 2d, at 327. At the second trial, 30 prospective jurors—5 black and 25 white—were presented to potentially serve on the jury. As in Flowers’ first trial, the State again used its strikes against all five black prospective jurors. But this time, the trial court determined that the State’s asserted reason for one of the strikes was a pretext for discrimination. Specifically, the trial court determined that one of the State’s proffered reasons—that the juror had been inattentive and was nodding off during jury selection—for striking that juror was false, and the trial court therefore sustained Flowers’ Batson challenge. The trial court disallowed the strike and sat that black juror on the jury. The jury at Flowers’ second trial consisted of 11 white jurors and 1 black juror. The jury convicted Flowers and sentenced him to death. On appeal, the Mississippi Supreme Court again reversed. The court ruled that the prosecutor had again engaged in prosecutorial misconduct in front of the jury by, among other things, impermissibly referencing evidence and attempting to undermine witness credibility without a factual basis. See Flowers v. State, 842 So. 2d 531, 538, 553 (2003). At Flowers’ third trial, 45 prospective jurors—17 black and 28 white—were presented to potentially serve on the jury. One of the black prospective jurors was struck for cause, leaving 16. The State exercised a total of 15 peremptory strikes, and it used all 15 against black prospective jurors. Flowers again argued that the State had used its peremptory strikes in a racially discriminatory manner. The trial court found that the State had not discriminated on the basis of race. See Flowers, 947 So. 2d, at 916. The jury in Flowers’ third trial consisted of 11 white jurors and 1 black juror. The lone black juror who served on the jury was seated after the State ran out of peremptory strikes. The jury convicted Flowers and sentenced him to death. On appeal, the Mississippi Supreme Court yet again reversed, concluding that the State had again violated Batson by discriminating on the basis of race in exercising all 15 of its peremptory strikes against 15 black prospective jurors. See Flowers, 947 So. 2d, at 939. The court’s lead opinion stated: “The instant case presents us with as strong a prima facie case of racial discrimination as we have ever seen in the context of a Batson challenge.” Id., at 935. The opinion explained that although “each individual strike may have justifiably appeared to the trial court to be sufficiently race neutral, the trial court also has a duty to look at the State’s use of peremptory challenges in toto.” Id., at 937. The opinion emphasized that “trial judges should not blindly accept any and every reason put forth by the State, especially” when “the State continues to exercise challenge after challenge only upon members of a particular race.” Ibid. The opinion added that the “State engaged in racially discriminatory practices” and that the “case evinces an effort by the State to exclude African-Americans from jury service.” Id., at 937, 939. At Flowers’ fourth trial, 36 prospective jurors—16 black and 20 white—were presented to potentially serve on the jury. The State exercised a total of 11 peremptory strikes, and it used all 11 against black prospective jurors. But because of the relatively large number of prospective jurors who were black, the State did not have enough peremptory challenges to eliminate all of the black prospective jurors. The seated jury consisted of seven white jurors and five black jurors. That jury could not reach a verdict, and the proceeding ended in a mistrial. As to the fifth trial, there is no available racial information about the prospective jurors, as distinct from the jurors who ultimately sat on the jury. The jury was composed of nine white jurors and three black jurors. The jury could not reach a verdict, and the trial again ended in a mistrial. At the sixth trial, which we consider here, 26 prospective jurors—6 black and 20 white—were presented to potentially serve on the jury. The State exercised a total of six peremptory strikes, and it used five of the six against black prospective jurors, leaving one black juror to sit on the jury. Flowers again argued that the State had exercised its peremptory strikes in a racially discrimina-tory manner. The trial court concluded that the State had offered race-neutral reasons for each of the five peremp-tory strikes against the five black prospective jurors. The jury at Flowers’ sixth trial consisted of 11 white jurors and 1 black juror. That jury convicted Flowers of murder and sentenced him to death. In a divided decision, the Mississippi Supreme Court agreed with the trial court on the Batson issue and stated that the State’s “race-neutral reasons were valid and not merely pretextual.” Flowers v. State, 158 So. 3d 1009, 1058 (2014). Flowers then sought review in this Court. This Court granted Flowers’ petition for a writ of certio-rari, vacated the judgment of the Mississippi Supreme Court, and remanded for further consideration in light of the decision in Foster, 578 U. S. ___. Flowers v. Mississippi, 579 U. S. ___ (2016). In Foster, this Court held that the defendant Foster had established a Batson violation. 578 U. S., at ___ (slip op., at 25). On remand, the Mississippi Supreme Court by a 5-to-4 vote again upheld Flowers’ conviction. See 240 So. 3d 1082 (2017). Justice King wrote a dissent for three justices. He stated: “I cannot conclude that Flowers received a fair trial, nor can I conclude that prospective jurors were not subjected to impermissible discrimination.” Id., at 1172. According to Justice King, both the trial court and the Mississippi Supreme Court “completely disregard[ed] the constitutional right of prospective jurors to be free from a racially discriminatory selection process.” Id., at 1171. We granted certiorari. See 586 U. S. ___. II A Other than voting, serving on a jury is the most substantial opportunity that most citizens have to participate in the democratic process. See Powers v. Ohio, 499 U.S. 400, 407 (1991). Jury selection in criminal cases varies significantly based on state and local rules and practices, but ordinarily consists of three phases, which we describe here in general terms. First, a group of citizens in the community is randomly summoned to the courthouse on a particular day for potential jury service. Second, a subgroup of those prospective jurors is called into a particular courtroom for a specific case. The prospective jurors are often questioned by the judge, as well as by the prosecutor and defense attorney. During that second phase, the judge may excuse certain prospective jurors based on their answers. Third, the prosecutor and defense attorney may challenge certain prospective jurors. The attorneys may challenge prospective jurors for cause, which usually stems from a potential juror’s conflicts of interest or inability to be impartial. In addition to challenges for cause, each side is typically afforded a set number of peremptory challenges or strikes. Peremptory strikes have very old credentials and can be traced back to the common law. Those peremptory strikes traditionally may be used to remove any potential juror for any reason—no questions asked. That blanket discretion to peremptorily strike prospective jurors for any reason can clash with the dictates of the Equal Protection Clause of the Fourteenth Amendment to the United States Constitution. This case arises at the intersection of the peremptory challenge and the Equal Protection Clause. And to understand how equal protection law applies to peremptory challenges, it helps to begin at the beginning. Ratified in 1868 in the wake of the Civil War, the Equal Protection Clause of the Fourteenth Amendment provides that no State shall “deny to any person within its jurisdiction the equal protection of the laws.” A primary objective of the Equal Protection Clause, this Court stated just five years after ratification, was “the freedom of the slave race, the security and firm establishment of that freedom, and the protection of the newly-made freeman and citizen from the oppressions of those who had formerly exercised unlimited dominion over him.” Slaughter-House Cases, 16 Wall. 36, 71 (1873). In 1875, to help enforce the Fourteenth Amendment, Congress passed and President Ulysses S. Grant signed the Civil Rights Act of 1875. Ch. 114, 18Stat. 335. Among other things, that law made it a criminal offense for state officials to exclude individuals from jury service on account of their race. 18 U. S. C. §243. The Act provides: “No citizen possessing all other qualifications which are or may be prescribed by law shall be disqualified for service as grand or petit juror in any court of the United States, or of any State on account of race, color, or previous condition of servitude.” In 1880, just 12 years after ratification of the Fourteenth Amendment, the Court decided Strauder v. West Virginia, 100 U.S. 303. That case concerned a West Virginia statute that allowed whites only to serve as jurors. The Court held the law unconstitutional. In reaching its conclusion, the Court explained that the Fourteenth Amendment required “that the law in the States shall be the same for the black as for the white; that all persons, whether colored or white, shall stand equal before the laws of the States, and, in regard to the colored race, for whose protection the amendment was primarily designed, that no discrimination shall be made against them by law because of their color.” Id., at 307. In the words of the Strauder Court: “The very fact that colored people are singled out and expressly denied by a statute all right to participate in the administration of the law, as jurors, because of their color, though they are citizens, and may be in other respects fully qualified, is practically a brand upon them, affixed by the law, an assertion of their inferiority, and a stimulant to that race prejudice which is an impediment to securing to individuals of the race that equal justice which the law aims to secure to all others.” Id., at 308. For those reasons, the Court ruled that the West Virginia statute exclud- ing blacks from jury service violated the Fourteenth Amendment. As the Court later explained in Brown v. Board of Education, 347 U.S. 483 (1954), the Court’s decisions in the Slaughter-House Cases and Strauder interpreted the Fourteenth Amendment “as proscribing all state-imposed discriminations against the Negro race,” including in jury service. Brown, 347 U. S., at 490. In the decades after Strauder, the Court reiterated that States may not discriminate on the basis of race in jury selection. See, e.g., Neal v. Delaware, 103 U.S. 370, 397 (1881); Carter v. Texas, 177 U.S. 442, 447 (1900); Norris v. Alabama, 294 U.S. 587, 597–599 (1935); Hale v. Kentucky, 303 U.S. 613, 616 (1938) (per curiam); Pierre v. Louisiana, 306 U.S. 354, 362 (1939); Smith v. Texas, 311 U.S. 128, 130–131 (1940); Avery v. Georgia, 345 U.S. 559, 562 (1953); Hernandez v. Texas, 347 U.S. 475, 477–478, 482 (1954); Coleman v. Alabama, 377 U.S. 129, 133 (1964). But critical problems persisted. Even though laws barring blacks from serving on juries were unconstitutional after Strauder, many jurisdictions employed various discriminatory tools to prevent black persons from being called for jury service. And when those tactics failed, or were invalidated, prosecutors could still exercise peremptory strikes in individual cases to remove most or all black prospective jurors. In the century after Strauder, the freedom to exercise peremptory strikes for any reason meant that “the problem of racial exclusion from jury service” remained “widespread” and “deeply entrenched.” 5 U. S. Commission on Civil Rights Report 90 (1961). Simple math shows how that happened. Given that blacks were a minority of the population, in many jurisdictions the number of peremp-tory strikes available to the prosecutor exceeded the number of black prospective jurors. So prosecutors could routinely exercise peremptories to strike all the black prospective jurors and thereby ensure all-white juries. The exclusion of black prospective jurors was almost total in certain jurisdictions, especially in cases involving black defendants. Similarly, defense counsel could use—and routinely did use—peremptory challenges to strike all the black prospective jurors in cases involving white defendants and black victims. In the aftermath of Strauder, the exclusion of black jurors became more covert and less overt—often accomplished through peremptory challenges in individual courtrooms rather than by blanket operation of law. But as this Court later noted, the results were the same for black jurors and black defendants, as well as for the black community’s confidence in the fairness of the American criminal justice system. See Batson, 476 U. S., at 98–99. Eighty-five years after Strauder, the Court decided Swain v. Alabama, 380 U.S. 202 (1965). The defendant Swain was black. Swain was convicted of a capital offense in Talladega County, Alabama, and sentenced to death. Swain presented evidence that no black juror had served on a jury in Talladega County in more than a decade. See id., at 226. And in Swain’s case, the prosecutor struck all six qualified black prospective jurors, ensuring that Swain was tried before an all-white jury. Swain invoked Strauder to argue that the prosecutor in his case had impermis-sibly discriminated on the basis of race by using peremptory challenges to strike the six black prospective jurors. See 380 U. S., at 203, 210. This Court ruled that Swain had not established unconstitutional discrimination. Most importantly, the Court held that a defendant could not object to the State’s use of peremptory strikes in an individual case. In the Court’s words: “[W]e cannot hold that the striking of Negroes in a particular case is a denial of equal protection of the laws.” Id., at 221. The Swain Court reasoned that prosecutors do not always judge prospective jurors individually when exercising peremptory strikes. Instead, prosecutors choose which prospective jurors to strike “in light of the limited knowledge counsel has of them, which may include their group affiliations, in the context of the case to be tried.” Ibid. In the Court’s view, the prosecutor could strike prospective jurors on the basis of their group affiliations, including race. In other words, a prosecutor could permissibly strike a prospective juror for any reason, including the assumption or belief that a black prospective juror, because of race, would be favorable to a black defendant or unfavorable to the State. See id., at 220–221. To be sure, the Swain Court held that a defendant could make out a case of racial discrimination by showing that the State “in case after case, whatever the circumstances, whatever the crime and whoever the defendant or the victim may be,” had been responsible for the removal of qualified black prospective jurors so that no black jurors “ever serve on petit juries.” Id., at 223. But Swain’s high bar for establishing a constitutional violation was almost impossible for any defendant to surmount, as the aftermath of Swain amply demonstrated. Twenty-one years later, in its 1986 decision in Batson, the Court revisited several critical aspects of Swain and in essence overruled them. In so doing, the Batson Court emphasized that “the central concern” of the Fourteenth Amendment “was to put an end to governmental discrimination on account of race.” 476 U. S., at 85. The Batson Court noted that Swain had left prosecutors’ peremptory challenges “largely immune from constitutional scrutiny.” 476 U. S., at 92–93. In his concurrence in Batson, Justice Byron White (the author of Swain) agreed that Swain should be overruled. He stated: “[T]he practice of peremptorily eliminating blacks from petit juries in cases with black defendants remains widespread, so much so” that “I agree with the Court that the time has come to rule as it has.” 476 U. S., at 101–102. Under Batson, once a prima facie case of discrimination has been shown by a defendant, the State must provide race-neutral reasons for its peremptory strikes. The trial judge must determine whether the prosecutor’s stated reasons were the actual reasons or instead were a pretext for discrimination. Id., at 97–98. Four parts of Batson warrant particular emphasis here. First, the Batson Court rejected Swain’s insistence that a defendant demonstrate a history of racially discriminatory strikes in order to make out a claim of race discrimination. See 476 U. S., at 95. According to the Batson Court, defendants had run into “practical difficulties” in trying to prove that a State had systematically “exercised peremptory challenges to exclude blacks from the jury on account of race.” Id., at 92, n. 17. The Batson Court explained that, in some jurisdictions, requiring a defendant to “investigate, over a number of cases, the race of persons tried in the particular jurisdiction, the racial composition of the venire and petit jury, and the manner in which both parties exercised their peremptory challenges” posed an “insurmountable” burden. Ibid. In addition to that practical point, the Court stressed a basic equal protection point: In the eyes of the Constitution, one racially discriminatory peremptory strike is one too many. For those reasons, the Batson Court held that a criminal defendant could show “purposeful discrimination in selection of the petit jury solely on evidence concerning the prosecutor’s exercise of peremptory challenges at the defendant’s trial.” Id., at 96 (emphasis added). Second, the Batson Court rejected Swain’s statement that a prosecutor could strike a black juror based on an assumption or belief that the black juror would favor a black defendant. In some of the most critical sentences in the Batson opinion, the Court emphasized that a prosecutor may not rebut a claim of discrimination “by stating merely that he challenged jurors of the defendant’s race on the assumption—or his intuitive judgment—that they would be partial to the defendant because of their shared race.” 476 U. S., at 97. The Court elaborated: The Equal Protection Clause “forbids the States to strike black veniremen on the assumption that they will be biased in a particular case simply because the defendant is black. The core guarantee of equal protection, ensuring citizens that their State will not discriminate on account of race, would be meaningless were we to approve the exclusion of jurors on the basis of such assumptions, which arise solely from the jurors’ race.” Id., at 97–98. In his concurrence, Justice Thurgood Marshall drove the point home: “Exclusion of blacks from a jury, solely because of race, can no more be justified by a belief that blacks are less likely than whites to consider fairly or sympathetically the State’s case against a black defendant than it can be justified by the notion that blacks lack the intelligence, experience, or moral integrity to be entrusted with that role.” Id., at 104–105 (internal quotation marks and citations omitted). Third, the Batson Court did not accept the argument that race-based peremptories should be permissible because black, white, Asian, and Hispanic defendants and jurors were all “equally” subject to race-based discrimination. The Court stated that each removal of an individual juror because of his or her race is a constitutional violation. Discrimination against one defendant or juror on account of race is not remedied or cured by discrimination against other defendants or jurors on account of race. As the Court later explained: Some say that there is no equal protection violation if individuals “of all races are subject to like treatment, which is to say that white jurors are subject to the same risk of peremptory challenges based on race as are all other jurors. The suggestion that racial classifications may survive when visited upon all persons is no more authoritative today than the case which advanced the theorem, Plessy v. Ferguson, 163 U.S. 537 (1896). This idea has no place in our modern equal protection jurisprudence. It is axiomatic that racial classifications do not become legitimate on the assumption that all persons suffer them in equal degree.” Powers, 499 U. S., at 410 (citing Loving v. Virginia, 388 U.S. 1 (1967)). Fourth, the Batson Court did not accept the argument that race-based peremptories are permissible because both the prosecution and defense could employ them in any individual case and in essence balance things out. Under the Equal Protection Clause, the Court stressed, even a single instance of race discrimination against a prospective juror is impermissible. Moreover, in criminal cases involving black defendants, the both-sides-can-do-it argument overlooks the percentage of the United States population that is black (about 12 percent) and the cold reality of jury selection in most jurisdictions. Because blacks are a minority in most jurisdictions, prosecutors often have more peremptory strikes than there are black prospective jurors on a particular panel. In the pre-Batson era, therefore, allowing each side in a case involving a black defendant to strike prospective jurors on the basis of race meant that a prosecutor could eliminate all of the black jurors, but a black defendant could not eliminate all of the white jurors. So in the real world of criminal trials against black defendants, both history and math tell us that a system of race-based peremptories does not treat black defendants and black prospective jurors equally with prosecutors and white prospective jurors. Cf. Batson, 476 U. S., at 99. B Equal justice under law requires a criminal trial free of racial discrimination in the jury selection process. Enforcing that constitutional principle, Batson ended the widespread practice in which prosecutors could (and often would) routinely strike all black prospective jurors in cases involving black defendants. By taking steps to eradicate racial discrimination from the jury selection process, Batson sought to protect the rights of defendants and jurors, and to enhance public confidence in the fairness of the criminal justice system. Batson immediately revolutionized the jury selection process that takes place every day in federal and state criminal courtrooms throughout the United States. In the decades since Batson, this Court’s cases have vigorously enforced and reinforced the decision, and guarded against any backsliding. See Foster, 578 U. S. ___; Snyder v. Louisiana, 552 U.S. 472 (2008); Miller-El v. Dretke, 545 U.S. 231 (2005) (Miller-El II). Moreover, the Court has extended Batson in certain ways. A defendant of any race may raise a Batson claim, and a defendant may raise a Batson claim even if the defendant and the excluded juror are of different races. See Hernandez, 347 U. S., at 477–478; Powers, 499 U. S., at 406. Moreover, Batson now applies to gender discrimination, to a criminal defendant’s peremptory strikes, and to civil cases. See J. E. B. v. Alabama ex rel. T. B., 511 U.S. 127, 129 (1994); Georgia v. McCollum, 505 U.S. 42, 59 (1992); Edmonson v. Leesville Concrete Co., 500 U.S. 614, 616 (1991). Of particular relevance here, Batson’s holding raised several important evidentiary and procedural issues, three of which we underscore. First, what factors does the trial judge consider in evaluating whether racial discrimination occurred? Our precedents allow criminal defendants raising Batson challenges to present a variety of evidence to support a claim that a prosecutor’s peremptory strikes were made on the basis of race. For example, defendants may present: statistical evidence about the prosecutor’s use of peremptory strikes against black prospective jurors as compared to white prospective jurors in the case; evidence of a prosecutor’s disparate questioning and investigation of black and white prospective jurors in the case; side-by-side comparisons of black prospective jurors who were struck and white prospective jurors who were not struck in the case; a prosecutor’s misrepresentations of the record when defending the strikes during the Batson hearing; relevant history of the State’s peremptory strikes in past cases; or other relevant circumstances that bear upon the issue of racial discrimination. See Foster, 578 U. S. ___; Snyder, 552 U.S. 472; Miller-El II, 545 U.S. 231; Batson, 476 U.S. 79. Second, who enforces Batson? As the Batson Court itself recognized, the job of enforcing Batson rests first and foremost with trial judges. See id., at 97, 99, n. 22. America’s trial judges operate at the front lines of American justice. In criminal trials, trial judges possess the primary responsibility to enforce Batson and prevent racial discrimination from seeping into the jury selection process. As the Batson Court explained and as the Court later reiterated, once a prima facie case of racial discrimination has been established, the prosecutor must provide race-neutral reasons for the strikes. The trial court must consider the prosecutor’s race-neutral explanations in light of all of the relevant facts and circumstances, and in light of the arguments of the parties. The trial judge’s assessment of the prosecutor’s credibility is often important. The Court has explained that “the best evidence of discriminatory intent often will be the demeanor of the attorney who exercises the challenge.” Snyder, 552 U. S., at 477 (quotation altered). “We have recognized that these determinations of credibility and demeanor lie peculiarly within a trial judge’s province.” Ibid. (internal quotation marks omitted). The trial judge must determine whether the prosecutor’s proffered reasons are the actual reasons, or whether the proffered reasons are pretextual and the prosecutor instead exercised peremptory strikes on the basis of race. The ultimate inquiry is whether the State was “motivated in substantial part by discriminatory intent.” Foster, 578 U. S., at ___ (slip op., at 23) (internal quotation marks omitted). Third, what is the role of appellate review? An appeals court looks at the same factors as the trial judge, but is necessarily doing so on a paper record. “Since the trial judge’s findings in the context under consideration here largely will turn on evaluation of credibility, a reviewing court ordinarily should give those findings great deference.” Batson, 476 U. S., at 98, n. 21. The Court has described the appellate standard of review of the trial court’s factual determinations in a Batson hearing as “highly deferential.” Snyder, 552 U. S., at 479. “On appeal, a trial court’s ruling on the issue of discriminatory intent must be sustained unless it is clearly erroneous.” Id., at 477. III In accord with the principles set forth in Batson, we now address Flowers’ case. The Constitution forbids striking even a single prospective juror for a discriminatory purpose. See Foster, 578 U. S., at ___ (slip op., at 9). The question for this Court is whether the Mississippi trial court clearly erred in concluding that the State was not “motivated in substantial part by discriminatory intent” when exercising peremp-tory strikes at Flowers’ sixth trial. Id., at ___ (slip op., at 23) (internal quotation marks omitted); see also Snyder, 552 U. S., at 477. Because this case arises on direct review, we owe no deference to the Mississippi Supreme Court, as distinct from deference to the Mississippi trial court. Four categories of evidence loom large in assessing the Batson issue in Flowers’ case: (1) the history from Flowers’ six trials, (2) the prosecutor’s striking of five of six black prospective jurors at the sixth trial, (3) the prosecutor’s dramatically disparate questioning of black and white prospective jurors at the sixth trial, and (4) the prosecutor’s proffered reasons for striking one black juror (Carolyn Wright) while allowing other similarly situated white jurors to serve on the jury at the sixth trial. We address each in turn. A First, we consider the relevant history of the case. Recall that in Swain, the Court held that a defendant may prove racial discrimination by establishing a historical pattern of racial exclusion of jurors in the jurisdiction in question. Indeed, under Swain, that was the only way that a defendant could make out a claim that the State discriminated on the basis of race in the use of peremptory challenges. In Batson, the Court ruled that Swain had imposed too heavy a burden on defendants seeking to prove that a prosecutor had used peremptory strikes in a racially discriminatory manner. Batson lowered the evidentiary burden for defendants to contest prosecutors’ use of peremptory strikes and made clear that demonstrating a history of discriminatory strikes in past cases was not necessary. In doing so, however, Batson did not preclude defendants from still using the same kinds of historical evidence that Swain had allowed defendants to use to support a claim of racial discrimination. Most importantly for present purposes, after Batson, the trial judge may still consider historical evidence of the State’s discriminatory peremptory strikes from past trials in the jurisdiction, just as Swain had allowed. After Batson, the defendant may still cast Swain’s “wide net” to gather “ ‘relevant’ ” evidence. Miller-El II, 545 U. S., at 239–240. A defendant may rely on “all relevant circumstances.” Batson, 476 U. S., at 96–97. Here, our review of the history of the prosecutor’s peremptory strikes in Flowers’ first four trials strongly supports the conclusion that his use of peremptory strikes in Flowers’ sixth trial was motivated in substantial part by discriminatory intent. (Recall that there is no record evidence from the fifth trial regarding the race of the prospective jurors.) The numbers speak loudly. Over the course of the first four trials, there were 36 black prospective jurors against whom the State could have exercised a peremptory strike. The State tried to strike all 36. The State used its avail-able peremptory strikes to attempt to strike every single black prospective juror that it could have struck. (At oral argument in this Court, the State acknowledged that statistic. Tr. of Oral Arg. 32.) Not only did the State’s use of peremptory strikes in Flowers’ first four trials reveal a blatant pattern of striking black prospective jurors, the Mississippi courts themselves concluded on two separate occasions that the State violated Batson. In Flowers’ second trial, the trial court concluded that the State discriminated against a black juror. Specifically, the trial court determined that one of the State’s proffered reasons—that the juror had been inattentive and was nodding off during jury selection—for striking that juror was false, and the trial court therefore sustained Flowers’ Batson challenge. In Flowers’ next trial—his third trial—the prosecutor used all 15 of its peremptories to strike 15 black prospective jurors. The lead opinion of the Mississippi Supreme Court stated: “The instant case presents us with as strong a prima facie case of racial discrimination as we have ever seen in the context of a Batson challenge.” Flowers, 947 So. 2d, at 935. The opinion further stated that “the State engaged in racially discriminatory practices during the jury selection process” and that the “case evinces an effort by the State to exclude African-Americans from jury service.” Id., at 937, 939. To summarize the most relevant history: In Flowers’ first trial, the prosecutor successfully used peremptory strikes against all of the black prospective jurors. Flowers faced an all-white jury. In Flowers’ second trial, the prosecutor tried again to strike all of the black prospective jurors, but the trial court decided that the State could not strike one of those jurors. The jury consisted of 11 white jurors and 1 black juror. In Flowers’ third trial, there were 17 black prospective jurors. The prosecutor used 15 out of 15 peremptory strikes against black prospective jurors. After one black juror was struck for cause and the prosecutor ran out of strikes, one black juror remained. The jury again consisted of 11 white jurors and 1 black juror. In Flowers’ fourth trial, the prosecutor again used 11 out of 11 peremptory strikes against black prospective jurors. Because of the large number of black prospective jurors at the trial, the prosecutor ran out of peremptory strikes before it could strike all of the black prospective jurors. The jury for that trial consisted of seven white jurors and five black jurors, and the jury was unable to reach a verdict. To reiterate, there is no available information about the race of prospective jurors in the fifth trial. The jury for that trial consisted of nine white jurors and three black jurors, and the jury was unable to reach a verdict. Stretching across Flowers’ first four trials, the State employed its peremptory strikes to remove as many black prospective jurors as possible. The State appeared to proceed as if Batson had never been decided. The State’s relentless, determined effort to rid the jury of black individuals strongly suggests that the State wanted to try Flowers before a jury with as few black jurors as possible, and ideally before an all-white jury. The trial judge was aware of the history. But the judge did not sufficiently account for the history when considering Flowers’ Batson claim. The State’s actions in the first four trials necessarily inform our assessment of the State’s intent going into Flowers’ sixth trial. We cannot ignore that history. We cannot take that history out of the case. B We turn now to the State’s strikes of five of the six black prospective jurors at Flowers’ sixth trial, the trial at issue here. As Batson noted, a “ ‘pattern’ of strikes against black jurors included in the particular venire might give rise to an inference of discrimination.” 476 U. S., at 97. Flowers’ sixth trial occurred in June 2010. At trial, 26 prospective jurors were presented to potentially serve on the jury. Six of the prospective jurors were black. The State accepted one black prospective juror—Alexander Robinson. The State struck the other five black prospective jurors—Carolyn Wright, Tashia Cunningham, Edith Burnside, Flancie Jones, and Dianne Copper. The resulting jury consisted of 11 white jurors and 1 black juror. The State’s use of peremptory strikes in Flowers’ sixth trial followed the same pattern as the first four trials, with one modest exception: It is true that the State accepted one black juror for Flowers’ sixth trial. But especially given the history of the case, that fact alone cannot insulate the State from a Batson challenge. In Miller-El II, this Court skeptically viewed the State’s decision to accept one black juror, explaining that a prosecutor might do so in an attempt “to obscure the otherwise consistent pattern of opposition to” seating black jurors. 545 U. S., at 250. The overall record of this case suggests that the same tactic may have been employed here. In light of all of the circumstances here, the State’s decision to strike five of the six black prospective jurors is further evidence suggesting that the State was motivated in substantial part by discriminatory intent. C We next consider the State’s dramatically disparate questioning of black and white prospective jurors in the jury selection process for Flowers’ sixth trial. As Batson explained, “the prosecutor’s questions and statements during voir dire examination and in exercising his challenges may support or refute an inference of discrimina-tory purpose.” 476 U. S., at 97. The questioning process occurred through an initial group voir dire and then more in-depth follow-up questioning by the prosecutor and defense counsel of individual prospective jurors. The State asked the five black prospective jurors who were struck a total of 145 questions. By contrast, the State asked the 11 seated white jurors a total of 12 questions. On average, therefore, the State asked 29 questions to each struck black prospective juror. The State asked an average of one question to each seated white juror. One can slice and dice the statistics and come up with all sorts of ways to compare the State’s questioning of excluded black jurors with the State’s questioning of the accepted white jurors. But any meaningful comparison yields the same basic assessment: The State spent far more time questioning the black prospective jurors than the accepted white jurors. The State acknowledges, as it must under our precedents, that disparate questioning can be probative of discriminatory intent. See Miller-El v. Cockrell, 537 U.S. 322, 331–332, 344–345 (2003) (Miller-El I). As Miller-El I stated, “if the use of disparate questioning is determined by race at the outset, it is likely [that] a justification for a strike based on the resulting divergent views would be pretextual. In this context the differences in the questions posed by the prosecutors are some evidence of purposeful discrimination.” Id., at 344. But the State here argues that it questioned black and white prospective jurors differently only because of differences in the jurors’ characteristics. The record refutes that explanation. For example, Dianne Copper was a black prospective juror who was struck. The State asked her 18 follow-up questions about her relationships with Flowers’ family and with witnesses in the case. App. 188–190. Pamela Chesteen was a white juror whom the State accepted for the jury. Although the State asked questions of Chesteen during group voir dire, the State asked her no individual follow-up questions about her relationships with Flowers’ family, even though the State was aware that Chesteen knew several members of Flowers’ family. Compare id., at 83, with id., at 111. Similarly, the State asked no individual follow-up questions to four other white prospective jurors who, like Dianne Copper, had relationships with defense witnesses, even though the State was aware of those relationships. Those white prospective jurors were Larry Blaylock, Harold Waller, Marcus Fielder, and Bobby Lester. Likewise, the State conducted disparate investigations of certain prospective jurors. Tashia Cunningham, who is black, stated that she worked with Flowers’ sister, but that the two did not work closely together. To try to disprove that statement, the State summoned a witness to challenge Cunningham’s testimony. Id., at 148–150. The State apparently did not conduct similar investigations of white prospective jurors. It is certainly reasonable for the State to ask follow-up questions or to investigate the relationships of jurors to the victims, potential witnesses, and the like. But white prospective jurors who were acquainted with the Flowers’ family or defense witnesses were not questioned extensively by the State or investigated. White prospective jurors who admitted that they or a relative had been convicted of a crime were accepted without apparent further inquiry by the State. The difference in the State’s approaches to black and white prospective jurors was stark. Why did the State ask so many more questions—and conduct more vigorous inquiry—of black prospective jurors than it did of white prospective jurors? No one can know for certain. But this Court’s cases explain that disparate questioning and investigation of prospective jurors on the basis of race can arm a prosecutor with seemingly race-neutral reasons to strike the prospective jurors of a particular race. See Miller-El I, 537 U. S., at 331–332, 344–345. In other words, by asking a lot of questions of the black prospective jurors or conducting additional inquiry into their backgrounds, a prosecutor can try to find some pretextual reason—any reason—that the prosecutor can later articulate to justify what is in reality a racially motivated strike. And by not doing the same for white prospective jurors, by not asking white prospective jurors those same questions, the prosecutor can try to distort the record so as to thereby avoid being accused of treating black and white jurors differently. Disparity in questioning and investigation can produce a record that says little about white prospective jurors and is therefore resistant to characteristic-by-characteristic comparisons of struck black prospective jurors and seated white jurors. Prosecutors can decline to seek what they do not want to find about white prospective jurors. A court confronting that kind of pattern cannot ignore it. The lopsidedness of the prosecutor’s questioning and inquiry can itself be evidence of the prosecutor’s objective as much as it is of the actual qualifications of the black and white prospective jurors who are struck or seated. The prosecutor’s dramatically disparate questioning of black and white prospective jurors—at least if it rises to a certain level of disparity—can supply a clue that the prosecutor may have been seeking to paper the record and disguise a discriminatory intent. See ibid. To be clear, disparate questioning or investigation alone does not constitute a Batson violation. The disparate questioning or investigation of black and white prospective jurors may reflect ordinary race-neutral considerations. But the disparate questioning or investigation can also, along with other evidence, inform the trial court’s evaluation of whether discrimination occurred. Here, along with the historical evidence we described above from the earlier trials, as well as the State’s striking of five of six black prospective jurors at the sixth trial, the dramatically disparate questioning and investigation of black prospective jurors and white prospective jurors at the sixth trial strongly suggests that the State was motivated in substantial part by a discriminatory intent. We agree with the observation of the dissenting justices of the Mississippi Supreme Court: The “numbers described above are too disparate to be explained away or categorized as mere happenstance.” 240 So. 3d, at 1161 (opinion of King, J.). D Finally, in combination with the other facts and circumstances in this case, the record of jury selection at the sixth trial shows that the peremptory strike of at least one of the black prospective jurors (Carolyn Wright) was motivated in substantial part by discriminatory intent. As this Court has stated, the Constitution forbids striking even a single prospective juror for a discriminatory purpose. See Foster, 578 U. S., at ___ (slip op., at 9). Comparing prospective jurors who were struck and not struck can be an important step in determining whether a Batson violation occurred. See Snyder, 552 U. S., at 483–484; Miller-El II, 545 U. S., at 241. The comparison can suggest that the prosecutor’s proffered explanations for striking black prospective jurors were a pretext for discrimination. When a prosecutor’s “proffered reason for striking a black panelist applies just as well to an otherwise-similar nonblack panelist who is permitted to serve, that is evidence tending to prove purposeful discrimination.” Foster, 578 U. S., at ___ (slip op., at 23) (quotation altered). Although a defendant ordinarily will try to identify a similar white prospective juror whom the State did not strike, a defendant is not required to identify an identical white juror for the side-by-side comparison to be suggestive of discriminatory intent. Miller-El II, 545 U. S., at 247, n. 6. In this case, Carolyn Wright was a black prospective juror who said she was strongly in favor of the death penalty as a general matter. And she had a family member who was a prison security guard. Yet the State exercised a peremptory strike against Wright. The State said it struck Wright in part because she knew several defense witnesses and had worked at Wal-Mart where Flowers’ father also worked. Winona is a small town. Wright had some sort of connection to 34 people involved in Flowers’ case, both on the prosecution witness side and the defense witness side. See, 240 So. 3d, at 1126. But three white prospective jurors—Pamela Chesteen, Harold Waller, and Bobby Lester—also knew many individuals involved in the case. Chesteen knew 31 people, Waller knew 18 people, and Lester knew 27 people. See ibid. Yet as we explained above, the State did not ask Chesteen, Waller, and Lester individual follow-up questions about their connections to witnesses. That is a telling statistic. If the State were concerned about prospective jurors’ connections to witnesses in the case, the State presumably would have used individual questioning to ask those potential white jurors whether they could remain impartial despite their relationships. A “State’s failure to engage in any meaningful voir dire examination on a subject the State alleges it is concerned about is evidence suggesting that the explanation is a sham and a pretext for discrimination.” Miller-El II, 545 U. S., at 246 (internal quotation marks omitted). Both Carolyn Wright and Archie Flowers, who is the defendant’s father, had worked at the local Wal-Mart. But there was no evidence that they worked together or were close in any way. Importantly, the State did not ask individual follow-up questions to determine the nature of their relationship. And during group questioning, Wright said she did not know whether Flowers’ father still worked at Wal-Mart, which “supports an inference that Wright and Flowers did not have a close working relationship.” 240 So. 3d, at 1163 (King, J., dissenting). And white prospective jurors also had relationships with members of Flowers’ family. Indeed, white prospective juror Pamela Chesteen stated that she had provided service to Flowers’ family members at the bank and that she knew several members of the Flowers family. App. 83. Likewise, white prospective juror Bobby Lester worked at the same bank and also encountered Flowers’ family members. Id., at 86. Although Chesteen and Lester were questioned during group voir dire, the State did not ask Chesteen or Lester individual follow-up questions in order to explore the depth of their relationships with Flowers’ family. And instead of striking those jurors, the State accepted them for the jury. To be sure, both Chesteen and Lester were later struck by the defense. But the State’s acceptance of Chesteen and Lester necessarily informs our assessment of the State’s intent in striking similarly situated black prospective jurors such as Wright. The State also noted that Wright had once been sued by Tardy Furniture for collection of a debt 13 years earlier. Id., at 209. Wright said that the debt was paid off and that it would not affect her evaluation of the case. Id., at 71, 90–91. The victims in this case worked at Tardy Furniture. But the State did not explain how Wright’s 13-year-old, paid-off debt to Tardy Furniture could affect her ability to serve impartially as a juror in this quadruple murder case. The “State’s unsupported characterization of the lawsuit is problematic.” 240 So. 3d, at 1163 (King, J., dissenting). In any event, the State did not purport to rely on that reason alone as the basis for the Wright strike, and the State in this Court does not rely on that reason alone in defending the Wright strike. The State also explained that it exercised a peremptory strike against Wright because she had worked with one of Flowers’ sisters. App. 209. That was incorrect. The trial judge immediately stated as much. Id., at 218–219. But incorrect statements of that sort may show the State’s intent: When a prosecutor misstates the record in explaining a strike, that misstatement can be another clue showing discriminatory intent. That incorrect statement was not the only one made by the prosecutor. The State made apparently incorrect statements to justify the strikes of black prospective jurors Tashia Cunningham, Edith Burnside, and Flancie Jones. The State contradicted Cunningham’s earlier statement that she had only a working relationship with Flowers’ sister by inaccurately asserting that Cunningham and Flowers’ sister were close friends. See id., at 84, 220. The State asserted that Burnside had tried to cover up a Tardy Furniture suit. See id., at 226. She had not. See id., 70–71. And the State explained that it struck Jones in part because Jones was Flowers’ aunt. See id., at 229. That, too, was not true. See id., at 86–88. The State’s pattern of factually inaccurate statements about black prospective jurors suggests that the State intended to keep black prospective jurors off the jury. See Foster, 578 U. S., at ___ (slip op., at 23); Miller-El II, 545 U. S., at 240, 245. To be sure, the back and forth of a Batson hearing can be hurried, and prosecutors can make mistakes when providing explanations. That is entirely understandable, and mistaken explanations should not be confused with racial discrimination. But when considered with other evidence of discrimination, a series of factually inaccurate explanations for striking black prospective jurors can be telling. So it is here. The side-by-side comparison of Wright to white prospective jurors whom the State accepted for the jury cannot be considered in isolation in this case. In a different context, the Wright strike might be deemed permissible. But we must examine the whole picture. Our disagreement with the Mississippi courts (and our agreement with Justice King’s dissent in the Mississippi Supreme Court) largely comes down to whether we look at the Wright strike in isolation or instead look at the Wright strike in the context of all the facts and circumstances. Our precedents require that we do the latter. As Justice King explained in his dissent in the Mississippi Supreme Court, the Mississippi courts appeared to do the former. 240 So. 3d, at 1163–1164. As we see it, the overall context here requires skepticism of the State’s strike of Carolyn Wright. We must examine the Wright strike in light of the history of the State’s use of peremptory strikes in the prior trials, the State’s decision to strike five out of six black prospective jurors at Flowers’ sixth trial, and the State’s vastly disparate questioning of black and white prospective jurors during jury selection at the sixth trial. We cannot just look away. Nor can we focus on the Wright strike in isolation. In light of all the facts and circumstances, we conclude that the trial court clearly erred in ruling that the State’s peremptory strike of Wright was not motivated in substantial part by discriminatory intent. * * * In sum, the State’s pattern of striking black prospective jurors persisted from Flowers’ first trial through Flowers’ sixth trial. In the six trials combined, the State struck 41 of the 42 black prospective jurors it could have struck. At the sixth trial, the State struck five of six. At the sixth trial, moreover, the State engaged in dramatically disparate questioning of black and white prospective jurors. And it engaged in disparate treatment of black and white prospective jurors, in particular by striking black prospective juror Carolyn Wright. To reiterate, we need not and do not decide that any one of those four facts alone would require reversal. All that we need to decide, and all that we do decide, is that all of the relevant facts and circumstances taken together establish that the trial court at Flowers’ sixth trial committed clear error in concluding that the State’s peremptory strike of black prospective juror Carolyn Wright was not motivated in substantial part by discriminatory intent. In reaching that conclusion, we break no new legal ground. We simply enforce and reinforce Batson by applying it to the extraordinary facts of this case. We reverse the judgment of the Supreme Court of Mississippi, and we remand the case for further proceedings not inconsistent with this opinion. It is so ordered.

588.US.2018_18-481

Respondent Argus Leader Media filed a Freedom of Information Act (FOIA) request with the United States Department of Agriculture (USDA), seeking the names and addresses of all retail stores that participate in the national food-stamp program—known as the Supplemental Nutrition Assistance Program (SNAP)—and each store’s annual SNAP redemption data from fiscal years 2005 to 2010. The USDA declined to disclose the store-level SNAP data, invoking FOIA’s Exemption 4, which shields from disclosure “trade secrets and commercial or financial information obtained from a person and privileged or confidential,” 5 U. S. C. §552(b)(4). Argus Leader sued the USDA. Following circuit precedent, the District Court employed the “competitive harm” test, under which commercial information cannot be deemed “confidential” unless disclosure is “likely . . . to cause substantial harm to the competitive position of the person from whom the information was obtained.” The court agreed that revealing store-level SNAP data could work some competitive harm, but it could not say that disclosure would cause “substantial competitive harm,” and thus ordered disclosure. Petitioner Food Marketing Institute, a trade association representing grocery retailers, intervened and filed an appeal. The Eighth Circuit affirmed, rejecting the Institute’s argument that the court should discard the “substantive competitive harm” test in favor of the ordinary public meaning of the statutory term “confidential.” Held: 1. The Institute has standing to appeal. Disclosure of the contested data would cause its members some financial injury in the highly competitive grocery industry; this concrete injury is directly traceable to the judgment ordering disclosure; and a favorable ruling from this Court would redress the retailers’ injury by reversing that judgment. Pp. 4–5. 2. Where commercial or financial information is both customarily and actually treated as private by its owner and provided to the government under an assurance of privacy, the information is “confidential” within Exemption 4’s meaning. Pp. 5–12. (a) At the time of FOIA’s enactment, the term “confidential” meant “private” or “secret.” Contemporary dictionaries suggest two conditions that might be required for information communicated to another to be considered confidential: when the information is customarily kept private, or at least closely held, by the person imparting it; and when the party receiving the information provides some assurance that it will remain secret. At least the first of these conditions must be met; it is hard to see how information could be deemed confidential if its owner shares it freely. But the Court need not resolve whether both conditions are necessary because both conditions are clearly met here. Uncontested testimony established that the Institute’s retailers customarily do not disclose store-level SNAP data or make it publicly available. And to induce retailers to participate in SNAP and provide store-level information, the government has long promised retailers that it will keep their information private. Early courts of appeals confronting Exemption 4 interpreted its terms in ways consistent with these understandings. Pp. 5–7. (b) Argus Leader pins its hopes on the “substantial competitive harm” requirement from the D. C. Circuit’s decision in National Parks & Conservation Assn. v. Morton, 498 F.2d 765. There, the court inappropriately resorted to legislative history before consulting the statute’s text and structure and relied heavily on statements from witnesses in congressional hearings years earlier on a different bill that was never enacted into law. Unsurprisingly, National Parks has drawn considerable criticism over the years, and even the D. C. Circuit has distanced itself from the decision. Pp. 7–10. (c) Argus Leader’s attempt to salvage National Parks is unpersuasive. First, it rearranges the text of Exemption 4 to create a phrase that does not appear in the statute: “confidential commercial information.” It suggests that this synthetic term mirrors a preexisting common law term of art that covers only information whose release would lead to substantial competitive harm, but points to no treatise or case decided before Exemption 4’s adoption that assigned any such meaning to the terms actually before the Court. Nor will this Court ordinarily imbue statutory terms with a specialized common law meaning when Congress has not itself invoked the common law terms of art associated with that meaning. See, e.g., Bruesewitz v. Wyeth LLC, 562 U.S. 223, 233–235. Alternatively, the company suggests that Congress effectively ratified its understanding of the term “confidential” by enacting similar phrases in other statutes in the years since National Parks was decided. But the ratification canon applies when Congress re-enacts the same statute using the same language, and Congress has never re-enacted Exemption 4. Finally, Argus Leader urges the Court to adopt a “substantial competitive harm” requirement as a matter of policy because it believes FOIA exemptions should be narrowly construed. But the Court cannot arbitrarily constrict Exemption 4 by adding limitations found nowhere in its terms. Pp. 10–12. 889 F.3d 914, reversed and remanded. Gorsuch, J., delivered the opinion of the Court, in which Roberts, C. J., and Thomas, Alito, Kagan, and Kavanaugh, JJ., joined. Breyer, J., filed an opinion concurring in part and dissenting in part, in which Ginsburg and Sotomayor, JJ., joined.

Congress has instructed that the disclosure re- quirements of the Freedom of Information Act do “not apply” to “confidential” private-sector “commercial or financial information” in the government’s possession. But when does information provided to a federal agency qualify as “confidential”? The Food Marketing Institute says it’s enough if the owner keeps the information private rather than releasing it publicly. The government suggests that an agency’s promise to keep information from disclosure may also suffice to render it confidential. But the courts below imposed a different requirement yet, holding that information can never be deemed confidential unless disclosing it is likely to result in “substantial competitive harm” to the business that provided it. Finding at least this “competitive harm” requirement inconsistent with the terms of the statute, we reverse. I This case began when Argus Leader, a South Dakota newspaper, filed a FOIA request for data collected by the United States Department of Agriculture. The USDA administers the national food-stamp program, known as the Supplemental Nutrition Assistance Program. Argus Leader asked the USDA for the names and addresses of all retail stores that participate in SNAP and each store’s annual SNAP redemption data from fiscal years 2005 to 2010, which we refer to as “store-level SNAP data.” The USDA tried to meet the paper halfway. It released the names and addresses of the participating stores but declined to disclose the requested store-level SNAP data. As relevant here, the USDA invoked FOIA’s Exemption 4, which shields from disclosure “trade secrets and commercial or financial information obtained from a person and privileged or confidential.” 5 U. S. C. §552(b)(4). Unsatisfied by the agency’s disclosure, Argus sued the USDA in federal court to compel release of the store-level SNAP data. Like several other courts of appeals, the Eighth Circuit has engrafted onto Exemption 4 a so-called “competitive harm” test, under which commercial information cannot be deemed “confidential” unless disclosure is “likely . . . to cause substantial harm to the competitive position of the person from whom the information was obtained.” Argus Leader Media v. United States Dept. of Agriculture, 889 F.3d 914, 915 (2018) (internal quotation marks omitted). So the district court held a two-day bench trial to determine whether disclosure of the store-level SNAP data would cause substantial competitive harm to participating retailers. At trial, witnesses for the USDA testified that retailers closely guard store-level SNAP data and that disclosure would threaten stores’ competitive positions. They explained that retailers use models of consumer behavior to help choose new store locations and to plan sales strategies. Competitors’ estimated sales volumes represent an important component of these models and can be time consuming and expensive to generate. And a model’s accuracy and utility increase significantly if it includes a rival’s actual sales data rather than mere estimates. So disclosure of store-level SNAP data could create a windfall for competitors: Stores with high SNAP redemptions could see increased competition for SNAP customers from existing competitors, new market entrants could use SNAP data to determine where to build their stores, and SNAP-redemption data could be used to discern a rival retailer’s overall sales and develop strategies to win some of that business too. For its part, Argus Leader offered no fact witnesses and did not dispute that retailers customarily keep this data private or that it bears competitive significance. Instead, the company contended that any competitive harm associated with disclosure would not be substantial. In the end, the district court agreed; while “[c]ompetition in the grocery business is fierce,” and while the record supported the conclusion that revealing store-level SNAP data could work some competitive harm, the court could not say that disclosure would rise to the level of causing “substantial competitive harm,” and thus ordered disclosure. Argus Leader Media v. United States Dept. of Agriculture, 224 F. Supp. 3d 827, 833–835 (SD 2016) (emphasis added). The USDA declined to appeal, but it alerted the retailers who had provided the data so that they could consider intervening to pursue the case further. The Food Marketing Institute, a trade association representing grocery retailers, answered the call. It successfully moved to intervene under Federal Rule of Civil Procedure 24(a) and then filed its own appeal. Meanwhile, the USDA assured the district court that it would not disclose the retailers’ data pending appeal. Before the Eighth Circuit, the Institute argued that the court should discard the “substantial competitive harm” test and apply instead the ordinary public meaning of the statutory term “confidential.” The court rejected that argument and affirmed. We granted the Institute a stay of the Eighth Circuit’s mandate and, later, its petition for certiorari. 585 U. S. ___ (2018); 586 U. S. ___ (2019). II Before turning to the merits, we confront a threshold challenge to our jurisdiction: Argus Leader questions whether the Institute has standing to pursue this appeal. To show standing under Article III, an appealing litigant must demonstrate that it has suffered an actual or imminent injury that is “fairly traceable” to the judgment below and that could be “redress[ed] by a favorable ruling.” Monsanto Co. v. Geertson Seed Farms, 561 U.S. 139, 149–150 (2010). The Institute satisfies each of these criteria. Whether or not disclosure of the contested data would cause its member retailers “substantial competitive harm,” the record before us reveals (and Argus Leader does not meaningfully dispute) that disclosure likely would cause them some financial injury. As the Eighth Circuit observed, the grocery industry is “highly competitive,” and disclosure of store-level SNAP data likely would help competitors win business from the Institute’s members. 889 F. 3d, at 916. This concrete injury is, as well, directly traceable to the judgment ordering disclosure. And a favorable ruling from this Court would redress the retailers’ injury by reversing that judgment. Argus Leader insists that the Institute’s injury is not redressable because a favorable ruling would merely restore the government’s discretion to withhold the requested data under Exemption 4, and it might just as easily choose to provide the data anyway. But the government has represented unequivocally that, consistent with its longstanding policy and past assurances of confidentiality to retailers, it “will not disclose” the contested data unless compelled to do so by the district court’s order. Brief for United States as Amicus Curiae 35; accord, Tr. of Oral Arg. 18–22. A reversal here thus would ensure exactly the relief the Institute requests. That is enough to satisfy Article III. Monsanto, 561 U. S., at 152–153. III A As we’ve seen, Exemption 4 shields from mandatory disclosure “commercial or financial information obtained from a person and privileged or confidential.” 5 U. S. C. §552(b)(4). But FOIA nowhere defines the term “confidential.” So, as usual, we ask what that term’s “ordinary, contemporary, common meaning” was when Congress enacted FOIA in 1966. Perrin v. United States, 444 U.S. 37, 42 (1979). We’ve done the same with other undefined terms in FOIA. See, e.g., Milner v. Department of Navy, 562 U.S. 562, 569 (2011); United States v. Weber Aircraft Corp., 465 U.S. 792, 804 (1984). The term “confidential” meant then, as it does now, “private” or “secret.” Webster’s Seventh New Collegiate Dictionary 174 (1963). Contemporary dictionaries suggest two conditions that might be required for information communicated to another to be considered confidential. In one sense, information communicated to another remains confidential whenever it is customarily kept private, or at least closely held, by the person imparting it. See, e.g., Webster’s Third New International Dictionary 476 (1961) (“known only to a limited few” or “not publicly disseminated”); Black’s Law Dictionary 370 (rev. 4th ed. 1968) (“intended to be held in confidence or kept secret”). In another sense, information might be considered confidential only if the party receiving it provides some assurance that it will remain secret. See, e.g., 1 Oxford Universal Dictionary Illustrated 367 (3d ed. 1961) (“spoken or written in confidence”); Webster’s New World Dictionary 158 (1960) (“told in confidence”). Must both of these conditions be met for information to be considered confidential under Exemption 4? At least the first condition has to be; it is hard to see how information could be deemed confidential if its owner shares it freely. And there’s no question that the Institute’s members satisfy this condition; uncontested testimony established that the Institute’s retailers customarily do not disclose store-level SNAP data or make it publicly available “in any way.” See, e.g., App. 93–94. Even within a company, witnesses testified, only small groups of employees usually have access to it. But what about the second condition: Can privately held information lose its confidential character for purposes of Exemption 4 if it’s communicated to the government without assurances that the government will keep it private? As it turns out, there’s no need to resolve that question in this case because the retailers before us clearly satisfy this condition too. Presumably to induce retailers to participate in SNAP and provide store-level information it finds useful to its adminstration of the program, the government has long promised them that it will keep their information private. See, e.g., 43 Fed. Reg. 43275 (1978); see also Brief for United States as Amicus Curiae 27–30. Early courts of appeals confronting Exemption 4 interpreted its terms in ways consistent with these understandings. In GSA v. Benson, 415 F.2d 878, 881 (1969), for example, the Ninth Circuit concluded that Exemption 4 would “ ‘protect information that a private individual wishes to keep confidential for his own purposes, but reveals to the government under the express or implied promise’ ” of confidentiality. The D. C. Circuit similarly held that Exemption 4 covered sales documents “ ‘which would customarily not be released to the public’ ” and which the government “agreed to treat . . . as confidential.” Sterling Drug Inc. v. FTC, 450 F.2d 698, 709 (1971); see also Grumman Aircraft Eng. Corp. v. Renegotiation Bd., 425 F.2d 578, 580, 582 (1970) (information a private party “submitted ‘in confidence’ ” or “would not reveal to the public [is] exempt from disclosure”). B Notably lacking from dictionary definitions, early case law, or any other usual source that might shed light on the statute’s ordinary meaning is any mention of the “substantial competive harm” requirement that the courts below found unsatisfied and on which Argus Leader pins its hopes. Indeed, when called on some years ago to interpret the similar phrase “information furnished by a confidential source” in FOIA Exemption 7(D), §552(b)(7)(D), this Court looked, as we do now, to “common usage” and never suggested that the government must prove that the disclosure of a source’s information would result in substantial harm. Department of Justice v. Landano, 508 U.S. 165, 173–174 (1993). So where did the “substantial competitive harm” requirement come from? In 1974, the D. C. Circuit declared that, in addition to the requirements actually set forth in Exemption 4, a “court must also be satisfied that non-disclosure is justified by the legislative purpose which underlies the exemption.” National Parks & Conservation Assn. v. Morton, 498 F.2d 765, 767. Then, after a selective tour through the legislative history, the court concluded that “commercial or financial matter is ‘confidential’ [only] if disclosure of the information is likely . . . (1) to impair the Government’s ability to obtain necessary information in the future; or (2) to cause substantial harm to the competitive position of the person from whom the information was obtained.” Id., at 770. Without much independent analysis, a number of courts of appeals eventually fell in line and adopted variants of the National Parks test. See Contract Freighters, Inc. v. Secretary of U. S. Dept. of Transp., 260 F.3d 858, 861 (CA8 2001) (collecting cases). We cannot approve such a casual disregard of the rules of statutory interpretation. In statutory interpretation disputes, a court’s proper starting point lies in a careful examination of the ordinary meaning and structure of the law itself. Schindler Elevator Corp. v. United States ex rel. Kirk, 563 U.S. 401, 407 (2011). Where, as here, that examination yields a clear answer, judges must stop. Hughes Aircraft Co. v. Jacobson, 525 U.S. 432, 438 (1999). Even those of us who sometimes consult legislative history will never allow it to be used to “muddy” the meaning of “clear statutory language.” Milner, 562 U. S., at 572. Indeed, this Court has repeatedly refused to alter FOIA’s plain terms on the strength only of arguments from legislative history. See, e.g., Landano, 508 U. S., at 178 (refusing to expand the plain meaning of Exemption 7(D) based on legislative history); Weber Aircraft, 465 U. S., at 800–803 (refusing to restrict Exemption 5 based on legislative history). National Parks’ contrary approach is a relic from a “bygone era of statutory construction.” Brief for United States as Amicus Curiae 19. Not only did National Parks inappropriately resort to legislative history before consulting the statute’s text and structure, once it did so it went even further astray. The court relied heavily on statements from witnesses in congressional hearings years earlier on a different bill that was never enacted into law. 498 F. 2d, at 767–769. Yet we can all agree that “excerpts from committee hearings” are “ ‘among the least illuminating forms of legislative history.’ ” Advocate Health Care Network v. Stapleton, 581 U. S. ___, ___ (2017) (slip op., at 12); see also Kelly v. Robinson, 479 U.S. 36, 51, n. 13 (1986) (declining to “accord any significance” to “comments in [legislative] hearings”). Perhaps especially so in cases like this one, where the witness statements do not comport with official committee reports that are consistent with the plain and ordinary meaning of the statute’s terms. See S. Rep. No. 813, 89th Cong., 1st Sess., 9 (1965) (Exemption 4 protects information “which would customarily not be released to the public by the person from whom it was obtained” such as “business sales statistics” and “customer lists”); H. R. Rep. No. 1497, 89th Cong., 2d Sess., 10 (1966) (Exemption 4 exempts material “if it would not customarily be made public by the person from whom it was obtained by the Government” and “information which is given to an agency in confidence” such as “business sales statistics”). Unsurprisingly, National Parks has drawn considerable criticism over the years. See, e.g., Critical Mass Energy Project v. NRC, 931 F.2d 939, 947 (CADC 1991) (Randolph, J., concurring) (National Parks was “ ‘fabricated . . . out of whole cloth’ ”); New Hampshire Right to Life v. Department of Health and Human Servs., 577 U. S. ___ (2015) (Thomas, J., joined by Scalia, J., dissenting from denial of certiorari). Even the D. C. Circuit has distanced itself from the decision. While retaining National Parks principally as a matter of stare decisis in the context of information a private entity is required to provide to the government, the court has pointedly declined to extend the National Parks test to information provided voluntarily to the government under Exemption 4. There, the court has adhered to a much more traditional understanding of the statutory term “confidential,” holding that information qualifies as confidential “if it is of a kind that would customarily not be released to the public by the person from whom it was obtained.” Critical Mass Energy Project v. NRC, 975 F.2d 871, 879–880 (CADC 1992) (en banc); see also id., at 880–882 (Randolph, J., concurring). Nor, unbound by D. C. Circuit precedent, can we discern a persuasive reason to afford the same statutory term two such radically different constructions. Ratzlaf v. United States, 510 U.S. 135, 143 (1994). C That leaves Argus Leader to try to salvage the result, if not the reasoning, of National Parks. But here its arguments prove no more persuasive. The company begins by rearranging the text of Exemption 4 to create a phrase that does not appear in the statute: “confidential commercial information.” Then, it suggests this synthetic term mirrors a preexisting common law term of art. And finally it asserts that the common law term covers only information whose release would lead to substantial competitive harm. But Argus Leader points to no treatise or case decided before Exemption 4’s adoption that assigned any such meaning to the terms actually before us: “commercial or financial information [that is] privileged or confidential.” So even accepting (without granting) that other phrases may carry the specialized common law meaning Argus Leader supposes, the parties have mustered no evidence that the terms of Exemption 4 did at the time of their adoption. Nor will this Court ordinarily imbue statutory terms with a specialized common law meaning when Congress hasn’t itself invoked the common law terms of art associated with that meaning. See, e.g., Bruesewitz v. Wyeth LLC, 562 U.S. 223, 233–235 (2011). Alternatively, the company suggests that, whatever the merits of National Parks as an initial matter, Congress effectively ratified its understanding of the term “confidential” by enacting similar phrases in other statutes in the years since that case was decided. To be sure, the ratification canon can sometimes prove a useful interpretive tool. But it derives from the notion that Congress is aware of a definitive judicial interpretation of a statute when it reenacts the same statute using the same language. Helsinn Healthcare S. A. v. Teva Pharmaceuticals USA, Inc., 586 U. S. ___, ___ (2019) (slip op., at 7). And Congress has never reenacted Exemption 4. So whether Congress’s use of similar language in other statutes after National Parks might (or might not) tell us what later Congresses understood those other statutes to mean, it tells us nothing about Congress’s understanding of the language it enacted in Exemption 4 in 1966. Finally, Argus urges us to adopt a “substantial competitive harm” requirement as a matter of policy because it believes FOIA exemptions should be narrowly construed. But as we have explained in connection with another federal statute, we normally “have no license to give [statutory] exemption[s] anything but a fair reading.” Encino Motorcars, LLC v. Navarro, 584 U. S. ___, ___ (2018) (slip op., at 9). Nor do we discern a reason to depart from that rule here: FOIA expressly recognizes that “important interests [are] served by [its] exemptions,” FBI v. Abramson, 456 U.S. 615, 630–631 (1982), and “[t]hose exemptions are as much a part of [FOIA’s] purpose[s and policies] as the [statute’s disclosure] requirement,” Encino Motorcars, 584 U. S., at ___ (slip op., at 9). So, just as we cannot properly expand Exemption 4 beyond what its terms permit, see, e.g., Milner, 562 U. S., at 570–571, we cannot arbitrarily constrict it either by adding limitations found nowhere in its terms. Our dissenting colleagues appear to endorse something like this final argument. They seem to agree that the law doesn’t demand proof of “substantial” or “competitive” harm, but they think it would be a good idea to require a showing of some harm. Neither side, however, has advocated for such an understanding of the statute’s terms. And our colleagues’ brief brush with the statutory text doesn’t help; they cite exclusively from specialized dictionary definitions lifted from the national security classification context that have no bearing on Exemption 4. Really, our colleagues’ submission boils down to a policy argument about the benefits of broad disclosure. But as Justice Breyer has noted, when Congress enacted FOIA it sought a “workable balance” between disclosure and other governmental interests—interests that may include providing private parties with sufficient assurances about the treatment of their proprietary information so they will cooperate in federal programs and supply the government with information vital to its work. See Milner, 562 U. S., at 589 (dissenting opinion) (arguing for a broad exemption from FOIA disclosure obligations to honor a “workable balance” between disclosure and privacy). * At least where commercial or financial information is both customarily and actually treated as private by its owner and provided to the government under an assurance of privacy, the information is “confidential” within the meaning of Exemption 4. Because the store-level SNAP data at issue here is confidential under that construction, the judgment of the court of appeals is reversed and the case is remanded for further proceedings consistent with this opinion. It is so ordered.

587.US.2018_18-525

Title VII of the Civil Rights Act of 1964 prohibits discrimination in employment on the basis of race, color, religion, sex, or national origin. 42 U. S. C. §2000e–2(a)(1). The Act instructs a complainant, before commencing a Title VII action in court, to file a charge with the Equal Employment Opportunity Commission (EEOC or Commission). §2000e‒5(e)(1), (f)(1). On receipt of a charge, the EEOC is to notify the employer and investigate the allegations. §2000e‒5(b). The Commission may “endeavor to eliminate [the] alleged unlawful employment practice by informal methods of . . . conciliation.” Ibid. The EEOC also has first option to “bring a civil action” against the employer in court. §2000e‒5(f)(1). But the Commission has no authority itself to adjudicate discrimination complaints. If the EEOC chooses not to sue, and whether or not the EEOC otherwise acts on the charge, a complainant is entitled to a “right-to-sue” notice 180 days after the charge is filed. Ibid.; 29 CFR §1601.28. On receipt of the right-to-sue notice, the complainant may commence a civil action against her employer. §2000e‒5(f)(1). Respondent Lois M. Davis filed a charge against her employer, petitioner Fort Bend County. Davis alleged sexual harassment and retaliation for reporting the harassment. While her EEOC charge was pending, Fort Bend fired Davis because she failed to show up for work on a Sunday and went to a church event instead. Davis attempted to supplement her EEOC charge by handwriting “religion” on a form called an “intake questionnaire,” but she did not amend the formal charge document. Upon receiving a right-to-sue letter, Davis commenced suit in Federal District Court, alleging discrimination on account of religion and retaliation for reporting sexual harassment. After years of litigation, only the religion-based discrimination claim remained in the case. Fort Bend then asserted for the first time that the District Court lacked jurisdiction to adjudicate Davis’ case because her EEOC charge did not state a religion-based discrimination claim. The District Court agreed and granted Fort Bend’s motion to dismiss Davis’ suit. On appeal from the dismissal, the Court of Appeals for the Fifth Circuit reversed. Title VII’s charge-filing requirement, the Court of Appeals held, is not jurisdictional; instead, the requirement is a prudential prerequisite to suit, forfeited in Davis’ case because Fort Bend had waited too long to raise the objection. Held: Title VII’s charge-filing requirement is not jurisdictional. Pp. 5–11. (a) The word “jurisdictional” is generally reserved for prescriptions delineating the classes of cases a court may entertain (subject-matter jurisdiction) and the persons over whom the court may exercise adjudicatory authority (personal jurisdiction). Kontrick v. Ryan, 540 U.S. 443, 455. A claim-processing rule requiring parties to take certain procedural steps in, or prior to, litigation, may be mandatory in the sense that a court must enforce the rule if timely raised. Eberhart v. United States, 546 U.S. 12, 19. But a mandatory rule of that sort, unlike a prescription limiting the kinds of cases a court may adjudicate, is ordinarily forfeited if not timely asserted. Id., at 15. Pp. 5‒9. (b) Title VII’s charge-filing requirement is a nonjurisdictional claim-processing rule. The requirement is stated in provisions of Title VII discrete from the statutory provisions empowering federal courts to exercise jurisdiction over Title VII actions. The charge-filing instruction is kin to prescriptions the Court has ranked as nonjurisdictional—for example, directions to raise objections in an agency rulemaking before asserting them in court, EPA v. EME Homer City Generation, L. P., 572 U.S. 489, 511‒512, or to follow procedures governing copyright registration before suing for infringement, Reed Elsevier, Inc. v. Muchnick, 559 U.S. 154, 157. Pp. 9‒11. 893 F.3d 300, affirmed. Ginsburg, J., delivered the opinion for a unanimous Court.

Title VII of the Civil Rights Act of 1964 proscribes discrimination in employment on the basis of race, color, religion, sex, or national origin. 78Stat. 255, 42 U. S. C. §2000e–2(a)(1). The Act also prohibits retaliation against persons who assert rights under the statute. §2000e–3(a). As a precondition to the commencement of a Title VII action in court, a complainant must first file a charge with the Equal Employment Opportunity Commission (EEOC or Commission). §2000e–5(e)(1), (f)(1). The question this case presents: Is Title VII’s charge-filing precondition to suit a “jurisdictional” requirement that can be raised at any stage of a proceeding; or is it a procedural prescription mandatory if timely raised, but subject to forfeiture if tardily asserted? We hold that Title VII’s charge-filing instruction is not jurisdictional, a term generally reserved to describe the classes of cases a court may entertain (subject-matter jurisdiction) or the persons over whom a court may exercise adjudicatory authority (personal jurisdiction). Kontrick v. Ryan, 540 U.S. 443, 455 (2004). Prerequisites to suit like Title VII’s charge-filing instruction are not of that character; they are properly ranked among the array of claim-processing rules that must be timely raised to come into play. I Title VII directs that a “charge . . . shall be filed” with the EEOC “by or on behalf of a person claiming to be aggrieved” within 180 days “after the alleged unlawful employment practice occur[s].” 42 U. S. C. §2000e–5(b), (e)(1). For complaints concerning a practice occurring in a State or political subdivision that has a fair employment agency of its own empowered “to grant or seek relief,” Title VII instructs the complainant to file her charge first with the state or local agency. §2000e–5(c). The complainant then has 300 days following the challenged practice, or 30 days after receiving notice that state or local proceedings have ended, “whichever is earlier,” to file a charge with the EEOC. §2000e–5(e)(1). If the state or local agency has a “worksharing” agreement with the EEOC, a complainant ordinarily need not file separately with federal and state agencies. She may file her charge with one agency, and that agency will then relay the charge to the other. See 29 CFR §1601.13 (2018); Brief for United States as Amicus Curiae 3. When the EEOC receives a charge, in contrast to agencies like the National Labor Relations Board, 29 U. S. C. §160, and the Merit Systems Protection Board, 5 U. S. C. §1204, it does not “adjudicate [the] clai[m],” Alexander v. Gardner-Denver Co., 415 U.S. 36, 44 (1974). Instead, Title VII calls for the following course. Upon receiving a charge, the EEOC notifies the employer and investigates the allegations. 42 U. S. C. §2000e–5(b). If the Commission finds “reasonable cause” to believe the charge is true, the Act instructs the Commission to “endeavor to eliminate [the] alleged unlawful employment practice by informal methods of conference, conciliation, and persuasion.” Ibid. When informal methods do not resolve the charge, the EEOC has first option to “bring a civil action” against the employer in court. §2000e–5(f)(1). Where the discrimination charge is lodged against state or local government employers, the Attorney General is the federal authority empowered to commence suit. Ibid.[1] In the event that the EEOC determines there is “n[o] reasonable cause to believe that the charge is true,” the Commission is to dismiss the charge and notify the complainant of his or her right to sue in court. 42 U. S. C. §2000e–5(b), f(1); 29 CFR §1601.28. Whether or not the EEOC acts on the charge, a complainant is entitled to a “right-to-sue” notice 180 days after the charge is filed. §2000e–5(f)(1); 29 CFR §1601.28. And within 90 days following such notice, the complainant may commence a civil action against the allegedly offending employer. §2000e–5(f)(1). II Respondent Lois M. Davis worked in information technology for petitioner Fort Bend County. In 2010, she informed Fort Bend’s human resources department that the director of information technology, Charles Cook, was sexually harassing her. Following an investigation by Fort Bend, Cook resigned. Davis’ supervisor at Fort Bend, Kenneth Ford, was well acquainted with Cook. After Cook resigned, Davis alleges, Ford began retaliating against her for reporting Cook’s sexual harassment. Ford did so, according to Davis, by, inter alia, curtailing her work responsibilities. Seeking redress for the asserted harassment and retaliation, Davis submitted an “intake questionnaire” in February 2011, followed by a charge in March 2011.[2] While her EEOC charge was pending, Davis was told to report to work on an upcoming Sunday. Davis informed her supervisor Ford that she had a commitment at church that Sunday, and she offered to arrange for another employee to replace her at work. Ford responded that if Davis did not show up for the Sunday work, she would be subject to termination. Davis went to church, not work, that Sunday. Fort Bend thereupon fired her. Attempting to supplement the allegations in her charge, Davis handwrote “religion” on the “Employment Harms or Actions” part of her intake questionnaire, and she checked boxes for “discharge” and “reasonable accommodation” on that form. She made no change, however, in the formal charge document. A few months later, the Department of Justice notified Davis of her right to sue. In January 2012, Davis commenced a civil action in the United States District Court for the Southern District of Texas, alleging discrimination on account of religion and retaliation for reporting sexual harassment.[3] The District Court granted Fort Bend’s motion for summary judgment. Davis v. Fort Bend County, 2013 WL 5157191 (SD Tex., Sept. 11, 2013). On appeal, the Court of Appeals for the Fifth Circuit affirmed as to Davis’ retaliation claim, but reversed as to her religion-based discrimination claim. Davis v. Fort Bend County, 765 F.3d 480 (2014). Fort Bend filed a petition for certiorari, which this Court denied. 576 U. S. ___ (2015). When the case returned to the District Court on Davis’ claim of discrimination on account of religion, Fort Bend moved to dismiss the complaint. Years into the litigation, Fort Bend asserted for the first time that the District Court lacked jurisdiction to adjudicate Davis’ religion-based discrimination claim because she had not stated such a claim in her EEOC charge. Granting the motion, the District Court held that Davis had not satisfied the charge-filing requirement with respect to her claim of religion-based discrimination, and that the requirement qualified as “jurisdictional,” which made it nonforfeitable. 2016 WL 4479527 (SD Tex., Aug. 24, 2016). The Fifth Circuit reversed. 893 F.3d 300 (2018). Title VII’s charge-filing requirement, the Court of Appeals held, is not jurisdictional; instead, the requirement is a prudential prerequisite to suit, forfeited in Davis’ case because Fort Bend did not raise it until after “an entire round of appeals all the way to the Supreme Court.” Id., at 307–308. We granted Fort Bend’s petition for certiorari, 586 U. S. ___ (2019), to resolve a conflict among the Courts of Appeals over whether Title VII’s charge-filing requirement is jurisdictional. Compare, e.g., 893 F. 3d, at 306 (case below) (charge-filing requirement is nonjurisdictional), with, e.g., Jones v. Calvert Group, Ltd., 551 F.3d 297, 300 (CA4 2009) (federal courts lack subject-matter jurisdiction when the charge-filing requirement is not satisfied). III “Jurisdiction,” the Court has observed, “is a word of many, too many, meanings.” Kontrick, 540 U. S., at 454 (quoting Steel Co. v. Citizens for Better Environment, 523 U.S. 83, 90 (1998)).[4] In recent years, the Court has undertaken “[t]o ward off profligate use of the term.” Sebelius v. Auburn Regional Medical Center, 568 U.S. 145, 153 (2013). As earlier noted, see supra, at 1, the word “jurisdictional” is generally reserved for prescriptions delineating the classes of cases a court may entertain (subject-matter jurisdiction) and the persons over whom the court may exercise adjudicatory authority (personal jurisdiction). Kontrick, 540 U. S., at 455. Congress may make other prescriptions jurisdictional by incorporating them into a jurisdictional provision, as Congress has done with the amount-in-controversy requirement for federal-court diversity jurisdiction. See 28 U. S. C. §1332(a) (“The district courts shall have original jurisdiction of all civil actions where the matter in controversy exceeds the sum or value of $75,000 . . . and is between (1) citizens of different States . . . .”). In addition, the Court has stated it would treat a requirement as “jurisdictional” when “a long line of [Supreme] Cour[t] decisions left undisturbed by Congress” attached a jurisdictional label to the prescription. Union Pacific R. Co. v. Locomotive Engineers, 558 U.S. 67, 82 (2009) (citing Bowles v. Russell, 551 U.S. 205, 209–211 (2007)). See also John R. Sand & Gravel Co. v. United States, 552 U.S. 130, 132 (2008). Characterizing a rule as a limit on subject-matter jurisdiction “renders it unique in our adversarial system.” Auburn, 568 U. S., at 153. Unlike most arguments, challenges to subject-matter jurisdiction may be raised by the defendant “at any point in the litigation,” and courts must consider them sua sponte. Gonzalez v. Thaler, 565 U.S. 134, 141 (2012). “[H]arsh consequences” attend the jurisdictional brand. United States v. Kwai Fun Wong, 575 U.S. 402, ___ (2015) (slip op., at 6). “Tardy jurisdictional objections” occasion wasted court resources and “disturbingly disarm litigants.” Auburn, 568 U. S., at 153. The Court has therefore stressed the distinction between jurisdictional prescriptions and nonjurisdictional claim-processing rules, which “seek to promote the orderly progress of litigation by requiring that the parties take certain procedural steps at certain specified times.” Henderson v. Shinseki, 562 U.S. 428, 435 (2011). A claim-processing rule may be “mandatory” in the sense that a court must enforce the rule if a party “properly raise[s]” it. Eberhart v. United States, 546 U.S. 12, 19 (2005) (per curiam). But an objection based on a mandatory claim-processing rule may be forfeited “if the party asserting the rule waits too long to raise the point.” Id., at 15 (quoting Kontrick, 540 U. S., at 456).[5] The Court has characterized as nonjurisdictional an array of mandatory claim-processing rules and other preconditions to relief. These include: the Copyright Act’s requirement that parties register their copyrights (or receive a denial of registration from the Copyright Register) before commencing an infringement action, Reed Elsevier, Inc. v. Muchnick, 559 U.S. 154, 157, 163–164 (2010); the Railway Labor Act’s direction that, before arbitrating, parties to certain railroad labor disputes “attempt settlement ‘in conference,’ ” Union Pacific, 558 U. S., at 82 (quoting 45 U. S. C. §152); the Clean Air Act’s instruction that, to maintain an objection in court on certain issues, one must first raise the objection “with reasonable specificity” during agency rulemaking, EPA v. EME Homer City Generation, L. P., 572 U.S. 489, 511–512 (2014) (quoting 42 U. S. C. §7607(d)(7)(B)); the Antiterrorism and Effective Death Penalty Act’s requirement that a certificate of appealability “indicate [the] specific issue” warranting issuance of the certificate, Gonzalez, 565 U. S., at 137 (quoting 28 U. S. C. §2253(c)(3)); Title VII’s limitation of covered “employer[s]” to those with 15 or more employees, Arbaugh v. Y & H Corp., 546 U.S. 500, 503–504 (2006) (quoting 42 U. S. C. §2000e(b)); Title VII’s time limit for filing a charge with the EEOC, Zipes v. Trans World Airlines, Inc., 455 U.S. 385, 393 (1982); and several other time prescriptions for procedural steps in judicial or agency forums. See, e.g., Hamer v. Neighborhood Housing Servs. of Chicago, 583 U. S. ___, ___ (2017) (slip op., at 1); Musacchio v. United States, 577 U. S. ___, ___ (2016) (slip op., at 8); Kwai Fun Wong, 575 U. S., at ___ (slip op., at 9); Auburn, 568 U. S., at 149; Henderson, 562 U. S., at 431; Eberhart, 546 U. S., at 13; Scarborough v. Principi, 541 U.S. 401, 414 (2004); Kontrick, 540 U. S., at 447.[6] While not demanding that Congress “incant magic words” to render a prescription jurisdictional, Auburn, 568 U. S., at 153, the Court has clarified that it would “leave the ball in Congress’ court”: “If the Legislature clearly states that a [prescription] count[s] as jurisdictional, then courts and litigants will be duly instructed and will not be left to wrestle with the issue[;] [b]ut when Congress does not rank a [prescription] as jurisdictional, courts should treat the restriction as nonjurisdictional in character.” Arbaugh, 546 U. S., at 515–516 (footnote and citation omitted). IV Title VII’s charge-filing requirement is not of jurisdictional cast. Federal courts exercise jurisdiction over Title VII actions pursuant to 28 U. S. C. §1331’s grant of general federal-question jurisdiction, and Title VII’s own jurisdictional provision, 42 U. S. C. §2000e–5(f)(3) (giving federal courts “jurisdiction [over] actions brought under this subchapter”).[7] Separate provisions of Title VII, §2000e–5(e)(1) and (f)(1), contain the Act’s charge-filing requirement. Those provisions “d[o] not speak to a court’s authority,” EME Homer, 572 U. S., at 512, or “refer in any way to the jurisdiction of the district courts,” Arbaugh, 546 U. S., at 515 (quoting Zipes, 455 U. S., at 394). Instead, Title VII’s charge-filing provisions “speak to . . . a party’s procedural obligations.” EME Homer, 572 U. S., at 512. They require complainants to submit information to the EEOC and to wait a specified period before commencing a civil action. Like kindred provisions directing parties to raise objections in agency rulemaking, id., at 511–512; follow procedures governing copyright registration, Reed Elsevier, 559 U. S., at 157; or attempt settlement, Union Pacific, 558 U. S., at 82, Title VII’s charge-filing requirement is a processing rule, albeit a mandatory one, not a jurisdictional prescription delineating the adjudicatory authority of courts.[8] Resisting this conclusion, Fort Bend points to statutory schemes that channel certain claims to administrative agency adjudication first, followed by judicial review in a federal court. In Elgin v. Department of Treasury, 567 U.S. 1 (2012), for example, the Court held that claims earmarked for initial adjudication by the Merit Systems Protection Board, then review in the Court of Appeals for the Federal Circuit, may not proceed instead in federal district court. Id., at 5–6, 8. See also Thunder Basin Coal Co. v. Reich, 510 U.S. 200, 202–204 (1994) (no district court jurisdiction over claims assigned in the first instance to a mine safety commission, whose decisions are review- able in a court of appeals). Nowhere do these cases, or others cited by Fort Bend, address the issue here presented: whether a precondition to suit is a mandatory claim-processing rule subject to forfeiture, or a jurisdictional prescription. Fort Bend further maintains that “[t]he congressional purposes embodied in the Title VII scheme,” notably, encouraging conciliation and affording the EEOC first option to bring suit, support jurisdictional characterization of the charge-filing requirement. Brief for Petitioner 27. But a prescription does not become jurisdictional whenever it “promotes important congressional objectives.” Reed Elsevier, 559 U. S., at 169, n. 9. And recognizing that the charge-filing requirement is nonjurisdictional gives plaintiffs scant incentive to skirt the instruction. Defendants, after all, have good reason promptly to raise an objection that may rid them of the lawsuit filed against them. A Title VII complainant would be foolhardy consciously to take the risk that the employer would forgo a potentially dispositive defense. In sum, a rule may be mandatory without being jurisdictional, and Title VII’s charge-filing requirement fits that bill. * * * For the reasons stated, the judgment of the Court of Appeals for the Fifth Circuit is Affirmed. Notes 1 A different provision of Title VII, 42 U. S. C. §2000e‒16, prohibits employment discrimination by the Federal Government and sets out procedures applicable to claims by federal employees. 2 Davis submitted these documents to the Texas Workforce Commission. Complaints lodged with that commission are relayed to the EEOC, under a “worksharing” agreement between the two agencies. See How To Submit an Employment Discrimination Complaint, Texas Workforce Commission, https://twc.texas.gov/jobseekers/how-submit-employment-discrimination-complaint (as last visited May 30, 2019). 3 Davis also alleged intentional infliction of emotional distress, but she did not appeal the District Court’s grant of summary judgment to Fort Bend on that claim. 4 “Courts, including this Court, . . . have more than occasionally [mis]used the term ‘jurisdictional’ ” to refer to nonjurisdictional prescriptions. Scarborough v. Principi, 541 U.S. 401, 413 (2004) (quoting Kontrick v. Ryan, 540 U.S. 443, 454 (2004) (alterations in original)). Passing references to Title VII’s charge-filing requirement as “jurisdictional” in prior Court opinions, see, e.g., McDonnell Douglas Corp. v. Green, 411 U.S. 792, 798 (1973), display the terminology employed when the Court’s use of “jurisdictional” was “less than meticulous,” Kontrick, 540 U. S., at 454. 5 The Court has “reserved whether mandatory claim-processing rules may [ever] be subject to equitable exceptions.” Hamer v. Neighborhood Housing Servs. of Chicago, 583 U. S. ___, ___, n. 3 (2017) (slip op., at 3, n. 3). 6 “If a time prescription governing the transfer of adjudicatory authority from one Article III court to another appears in a statute, the limitation [will rank as] jurisdictional; otherwise, the time specification fits within the claim-processing category.” Hamer, 583 U. S., at ___ (slip op., at 8) (citation omitted). 7 When Title VII was passed in 1964, 28 U. S. C. §1331’s grant of general federal-question jurisdiction included an amount-in-controversy requirement. See §1331(a) (1964 ed.). To ensure that this “limitation would not impede an employment-discrimination complainant’s access to a federal forum,” Arbaugh v. Y & H Corp., 546 U.S. 500, 505 (2006), Congress enacted Title VII’s jurisdiction-conferring provision, 42 U. S. C. §2000e‒5(f )(3). See Arbaugh, 546 U. S., at 505‒506. In 1980, Congress eliminated §1331’s amount-in-controversy requirement. See Federal Question Jurisdictional Amendments Act of 1980, §2, 94Stat. 2369. Since then, “Title VII’s own jurisdictional provision, 42 U. S. C. §2000e‒5(f )(3), has served simply to underscore Congress’ intention to provide a federal forum for the adjudication of Title VII claims.” Arbaugh, 546 U. S., at 506. Title VII also contains a separate jurisdictional provision, §2000e‒6(b), giving federal courts jurisdiction over actions by the Federal Government to enjoin “pattern or practice” discrimination. 8 Fort Bend argues that Title VII’s charge-filing requirement is jurisdictional because it is “textually linked” to Title VII’s jurisdictional provision. Brief for Petitioner 50. Title VII states in 42 U. S. C. §2000e‒5(f )(1) that “a civil action may be brought” after the charge-filing procedures are followed. Section 2000e‒5(f )(3) gives federal courts jurisdiction over “actions brought under this subchapter,” a subchapter that includes §2000e‒5(f )(1). Therefore, Fort Bend insists, federal jurisdiction lies under §2000e‒5(f )(3) only when a proper EEOC charge is filed. But as just observed, see supra, at 9, the charge-filing requirement is stated in provisions discrete from Title VII’s conferral of jurisdiction on federal courts. See Sebelius v. Auburn Regional Medical Center, 568 U.S. 145, 155 (2013) (a requirement “does not become jurisdictional simply because it is placed in a section of a statute that also contains jurisdictional provisions”); Gonzalez v. Thaler, 565 U.S. 134, 145 (2012) (a nonjurisdictional provision does not metamorphose into a jurisdictional limitation by cross-referencing a jurisdictional provision).

586.US.2018_17-571

Petitioner Fourth Estate Public Benefit Corporation (Fourth Estate), a news organization, licensed works to respondent Wall-Street.com, LLC (Wall-Street), a news website. Fourth Estate sued Wall-Street and its owner for copyright infringement of news articles that Wall-Street failed to remove from its website after canceling the parties’ license agreement. Fourth Estate had filed applications to register the articles with the Copyright Office, but the Register of Copyrights had not acted on those applications. Title 17 U. S. C. §411(a) states that “no civil action for infringement of the copyright in any United States work shall be instituted until . . . registration of the copyright claim has been made in accordance with this title.” The District Court dismissed the complaint, and the Eleventh Circuit affirmed, holding that “registration . . . has [not] been made” under §411(a) until the Copyright Office registers a copyright. Held: Registration occurs, and a copyright claimant may commence an infringement suit, when the Copyright Office registers a copyright. Upon registration of the copyright, however, a copyright owner can recover for infringement that occurred both before and after registration. Pp. 3–12. (a) Under the Copyright Act of 1976, as amended, a copyright author gains “exclusive rights” in her work immediately upon the work’s creation. 17 U. S. C. §106. A copyright owner may institute a civil action for infringement of those exclusive rights, §501(b), but generally only after complying with §411(a)’s requirement that “registration . . . has been made.” Registration is thus akin to an administrative exhaustion requirement that the owner must satisfy before suing to enforce ownership rights. P. 3. (b) In limited circumstances, copyright owners may file an infringement suit before undertaking registration. For example, a copy- right owner who is preparing to distribute a work of a type vulnerable to predistribution infringement—e.g., a movie or musical composition—may apply to the Copyright Office for preregistration. §408(f)(2). A copyright owner may also sue for infringement of a live broadcast before “registration . . . has been made.” §411(c). Outside of statutory exceptions not applicable here, however, §411(a) bars a copyright owner from suing for infringement until “registration . . . has been made.” Fourth Estate advances the “application approach” to this provision, arguing that registration occurs when a copyright owner submits a proper application for registration. Wall-Street advocates the “registration approach,” urging that registration occurs only when the Copyright Office grants registration of a copyright. The registration approach reflects the only satisfactory reading of §411(a)’s text. Pp. 3–12. (1) Read together, §411(a)’s first two sentences focus on action by the Copyright Office—namely, its registration or refusal to register a copyright claim. If application alone sufficed to “ma[ke]” registration, §411(a)’s second sentence—which permits a copyright claimant to file suit when the Register has refused her application—would be superfluous. Similarly, §411(a)’s third sentence—which allows the Register to “become a party to the action with respect to the issue of registrability of the copyright claim”—would be negated if an infringement suit could be filed and resolved before the Register acted on an application. The registration approach reading of §411(a) is supported by other provisions of the Copyright Act. In particular, §410 confirms that application is discrete from, and precedes, registration, while §408(f)’s preregistration option would have little utility if a completed application sufficed to make registration. Pp. 4–7. (2) Fourth Estate primarily contends that the Copyright Act uses the phrases “make registration” and “registration has been made” to describe submissions by the copyright owner. Fourth Estate therefore insists that §411(a)’s requirement that “registration . . . has been made in accordance with this title” most likely refers to a copyright owner’s compliance with statutory requirements for registration applications. Fourth Estate points to other Copyright Act provisions that appear to use the phrase “make registration” or one of its variants to describe what a copyright claimant does. Fourth Estate acknowledges, however, that determining how the Copyright Act uses the word “registration” in a particular provision requires examining the “specific context” in which the term is used. The “specific context” of §411(a) permits only one sensible reading: The phrase “registration . . . has been made” refers to the Copyright Office’s act granting registration, not to the copyright claimant’s request for registration. Fourth Estate’s contrary reading stems in part from its misapprehension of the significance of certain 1976 revisions to the Copyright Act. But in enacting §411(a), Congress both reaffirmed the general rule that registration must precede an infringement suit and added an exception in that provision’s second sentence to cover instances in which registration is refused. That exception would have no work to do if Congress intended the 1976 revisions to clarify that a copyright claimant may sue immediately upon applying for registration. Noteworthy, too, in years following the 1976 revisions, Congress resisted efforts to eliminate §411(a), which contains the registration requirement. Fourth Estate also argues that, because “registration is not a condition of copyright protection,” §408(a), §411(a) should not bar a copyright claimant from enforcing that protection in court once she has applied for registration. But the Copyright Act safeguards copyright owners by vesting them with exclusive rights upon creation of their works and prohibiting infringement from that point forward. To recover for such infringement, copyright owners must simply apply for registration and await the Register’s decision. Further, Congress has authorized preregistration infringement suits with respect to works vulnerable to predistribution infringement, and Fourth Estate’s fear that a copyright owner might lose the ability to enforce her rights entirely is overstated. True, registration processing times have increased from one to two weeks in 1956 to many months today. Delays, in large part, are the result of Copyright Office staffing and budgetary shortages that Congress can alleviate, but courts cannot cure. Unfortunate as the current administrative lag may be, that factor does not allow this Court to revise §411(a)’s congressionally composed text. Pp. 7–12. 856 F.3d 1338, affirmed. Ginsburg, J., delivered the opinion for a unanimous Court.

Impelling prompt registration of copyright claims, 17 U. S. C. §411(a) states that “no civil action for infringement of the copyright in any United States work shall be instituted until . . . registration of the copyright claim has been made in accordance with this title.” The question this case presents: Has “registration . . . been made in accordance with [Title 17]” as soon as the claimant delivers the required application, copies of the work, and fee to the Copyright Office; or has “registration . . . been made” only after the Copyright Office reviews and registers the copyright? We hold, in accord with the United States Court of Appeals for the Eleventh Circuit, that registration occurs, and a copyright claimant may commence an infringement suit, when the Copyright Office registers a copyright. Upon registration of the copyright, however, a copyright owner can recover for infringement that occurred both before and after registration. Petitioner Fourth Estate Public Benefit Corporation (Fourth Estate) is a news organization producing online journalism. Fourth Estate licensed journalism works to respondent Wall-Street.com, LLC (Wall-Street), a news website. The license agreement required Wall-Street to remove from its website all content produced by Fourth Estate before canceling the agreement. Wall-Street canceled, but continued to display articles produced by Fourth Estate. Fourth Estate sued Wall-Street and its owner, Jerrold Burden, for copyright infringement. The complaint alleged that Fourth Estate had filed “applications to register [the] articles [licensed to Wall-Street] with the Register of Copyrights.” App. to Pet. for Cert. 18a.[1] Because the Register had not yet acted on Fourth Estate’s applications,[2] the District Court, on Wall-Street and Burden’s motion, dismissed the complaint, and the Eleventh Circuit affirmed. 856 F.3d 1338 (2017). Thereafter, the Register of Copyrights refused registration of the articles Wall-Street had allegedly infringed.[3] We granted Fourth Estate’s petition for certiorari to resolve a division among U. S. Courts of Appeals on when registration occurs in accordance with §411(a). 585 U. S. ___ (2018). Compare, e.g., 856 F. 3d, at 1341 (case below) (registration has been made under §411(a) when the Register of Copyrights registers a copyright), with, e.g., Cosmetic Ideas, Inc. v. IAC/Interactivecorp, 606 F.3d 612, 621 (CA9 2010) (registration has been made under §411(a) when the copyright claimant’s “complete application” for registration is received by the Copyright Office). I Under the Copyright Act of 1976, as amended, copyright protection attaches to “original works of authorship”—prominent among them, literary, musical, and dramatic works—“fixed in any tangible medium of expression.” 17 U. S. C. §102(a). An author gains “exclusive rights” in her work immediately upon the work’s creation, including rights of reproduction, distribution, and display. See §106; Eldred v. Ashcroft, 537 U.S. 186, 195 (2003) (“[F]ederal copyright protection . . . run[s] from the work’s creation.”). The Copyright Act entitles a copyright owner to institute a civil action for infringement of those exclusive rights. §501(b). Before pursuing an infringement claim in court, how- ever, a copyright claimant generally must comply with §411(a)’s requirement that “registration of the copyright claim has been made.” §411(a). Therefore, although an owner’s rights exist apart from registration, see §408(a), registration is akin to an administrative exhaustion requirement that the owner must satisfy before suing to enforce ownership rights, see Tr. of Oral Arg. 35. In limited circumstances, copyright owners may file an infringement suit before undertaking registration. If a copyright owner is preparing to distribute a work of a type vulnerable to predistribution infringement—notably, a movie or musical composition—the owner may apply for preregistration. §408(f)(2); 37 CFR §202.16(b)(1) (2018). The Copyright Office will “conduct a limited review” of the application and notify the claimant “[u]pon completion of the preregistration.” §202.16(c)(7), (c)(10). Once “preregistration . . . has been made,” the copyright claimant may institute a suit for infringement. 17 U. S. C. §411(a). Preregistration, however, serves only as “a preliminary step prior to a full registration.” Preregistration of Certain Unpublished Copyright Claims, 70 Fed. Reg. 42286 (2005). An infringement suit brought in reliance on preregistration risks dismissal unless the copyright owner applies for registration promptly after the preregistered work’s publication or infringement. §408(f)(3)–(4). A copyright owner may also sue for infringement of a live broadcast before “registration . . . has been made,” but faces dismissal of her suit if she fails to “make registration for the work” within three months of its first transmission. §411(c). Even in these exceptional scenarios, then, the copyright owner must eventually pursue registration in order to maintain a suit for infringement. II All parties agree that, outside of statutory exceptions not applicable here, §411(a) bars a copyright owner from suing for infringement until “registration . . . has been made.” Fourth Estate and Wall-Street dispute, however, whether “registration . . . has been made” under §411(a) when a copyright owner submits the application, materials, and fee required for registration, or only when the Copyright Office grants registration. Fourth Estate advances the former view—the “application approach”—while Wall-Street urges the latter reading—the “registration approach.” The registration approach, we conclude, reflects the only satisfactory reading of §411(a)’s text. We therefore reject Fourth Estate’s application approach. A Under §411(a), “registration . . . has been made,” and a copyright owner may sue for infringement, when the Copyright Office registers a copyright.[4] Section 411(a)’s first sentence provides that no civil infringement action “shall be instituted until preregistration or registration of the copyright claim has been made.” The section’s next sentence sets out an exception to this rule: When the required “deposit, application, and fee . . . have been delivered to the Copyright Office in proper form and registration has been refused,” the claimant “[may] institute a civil action, if notice thereof . . . is served on the Register.” Read together, §411(a)’s opening sentences focus not on the claimant’s act of applying for registration, but on action by the Copyright Office—namely, its registration or refusal to register a copyright claim. If application alone sufficed to “ma[ke]” registration, §411(a)’s second sentence—allowing suit upon refusal of registration—would be superfluous. What utility would that allowance have if a copyright claimant could sue for infringement immediately after applying for registration without awaiting the Register’s decision on her application? Proponents of the application approach urge that §411(a)’s second sentence serves merely to require a copyright claimant to serve “notice [of an infringement suit] . . . on the Register.” See Brief for Petitioner 29–32. This reading, however, requires the implausible assumption that Congress gave “registration” different meanings in consecutive, related sentences within a single statutory provision. In §411(a)’s first sentence, “registration” would mean the claimant’s act of filing an application, while in the section’s second sentence, “registration” would entail the Register’s review of an application. We resist this improbable construction. See, e.g., Mid-Con Freight Systems, Inc. v. Michigan Pub. Serv. Comm’n, 545 U.S. 440, 448 (2005) (declining to read “the same words” in con- secutive sentences as “refer[ring] to something totally different”). The third and final sentence of §411(a) further persuades us that the provision requires action by the Register before a copyright claimant may sue for infringement. The sentence allows the Register to “become a party to the action with respect to the issue of registrability of the copyright claim.” This allowance would be negated, and the court conducting an infringement suit would lack the benefit of the Register’s assessment, if an infringement suit could be filed and resolved before the Register acted on an application. Other provisions of the Copyright Act support our reading of “registration,” as used in §411(a), to mean action by the Register. Section 410 states that, “after examination,” if the Register determines that “the material deposited constitutes copyrightable subject matter” and “other legal and formal requirements . . . [are] met, the Register shall register the claim and issue to the applicant a certificate of registration.” §410(a). But if the Register determines that the deposited material “does not constitute copyrightable subject matter or that the claim is invalid for any other reason, the Register shall refuse registration.” §410(b). Section 410 thus confirms that application is discrete from, and precedes, registration. Section 410(d), furthermore, provides that if the Copyright Office registers a claim, or if a court later determines that a refused claim was registrable, the “effective date of [the work’s] copyright registration is the day on which” the copyright owner made a proper submission to the Copyright Office. There would be no need thus to specify the “effective date of a copyright registration” if submission of the required materials qualified as “registration.” Section 408(f)’s preregistration option, too, would have little utility if a completed application constituted registration. Preregistration, as noted supra, at 3–4, allows the author of a work vulnerable to predistribution infringement to enforce her exclusive rights in court before obtaining registration or refusal thereof. A copyright owner who fears prepublication infringement would have no reason to apply for preregistration, however, if she could instead simply complete an application for registration and immediately commence an infringement suit. Cf. TRW Inc. v. Andrews, 534 U.S. 19, 29 (2001) (rejecting an interpretation that “would in practical effect render [a provision] superfluous in all but the most unusual circumstances”). B Challenging the Eleventh Circuit’s judgment, Fourth Estate primarily contends that the Copyright Act uses “the phrase ‘make registration’ and its passive-voice counterpart ‘registration has been made’ ” to describe submissions by the copyright owner, rather than Copyright Office responses to those submissions. Brief for Petitioner 21. Section 411(a)’s requirement that “registration . . . has been made in accordance with this title,” Fourth Estate insists, most likely refers to a copyright owner’s compliance with the statutory specifications for registration applications. In support, Fourth Estate points to Copyright Act provisions that appear to use the phrase “make registration” or one of its variants to describe what a copyright claimant does. See id., at 22–26 (citing 17 U. S. C. §§110, 205(c), 408(c)(3), 411(c), 412(2)). Furthermore, Fourth Estate urges that its reading reflects the reality that, eventually, the vast majority of applications are granted. See Brief for Petitioner 41. Fourth Estate acknowledges, however, that the Copyright Act sometimes uses “registration” to refer to activity by the Copyright Office, not activity undertaken by a copyright claimant. See id., at 27–28 (citing 17 U. S. C. §708(a)). Fourth Estate thus agrees that, to determine how the statute uses the word “registration” in a particular prescription, one must “look to the specific context” in which the term is used. Brief for Petitioner 29. As explained supra, at 4–7, the “specific context” of §411(a) permits only one sensible reading: The phrase “registration . . . has been made” refers to the Copyright Office’s act granting registration, not to the copyright claimant’s request for registration. Fourth Estate’s contrary reading of §411(a) stems in part from its misapprehension of the significance of certain 1976 revisions to the Copyright Act. Before that year, §411(a)’s precursor provided that “[n]o action or proceeding shall be maintained for infringement of copyright in any work until the provisions of this title with respect to the deposit of copies and registration of such work shall have been complied with.” 17 U. S. C. §13 (1970 ed.). Fourth Estate urges that this provision posed the very question we resolve today—namely, whether a claimant’s application alone effects registration. The Second Circuit addressed that question, Fourth Estate observes, in Vacheron & Constantin-Le Coultre Watches, Inc. v. Benrus Watch Co., 260 F.2d 637 (1958). Brief for Petitioner 32–34. In that case, in an opinion by Judge Learned Hand, the court held that a copyright owner who completed an application could not sue for infringement immediately upon the Copyright Office’s refusal to register. Vacheron, 260 F. 3d, at 640–641. Instead, the owner first had to obtain a registration certificate by bringing a mandamus action against the Register. The Second Circuit dissenter would have treated the owner’s application as sufficient to permit commencement of an action for infringement. Id., at 645. Fourth Estate sees Congress’ 1976 revision of the registration requirement as an endorsement of the Vacheron dissenter’s position. Brief for Petitioner 34–36. We dis- agree. The changes made in 1976 instead indicate Congress’ agreement with Judge Hand that it is the Register’s action that triggers a copyright owner’s entitlement to sue. In enacting 17 U. S. C. §411(a), Congress both reaffirmed the general rule that registration must precede an infringement suit, and added an exception in that provision’s second sentence to cover instances in which registration is refused. See H. R. Rep. No. 94‒1476, p. 157 (1976). That exception would have no work to do if, as Fourth Estate urges, Congress intended the 1976 revisions to clarify that a copyright claimant may sue immediately upon applying for registration. A copyright claimant would need no statutory authorization to sue after refusal of her application if she could institute suit as soon as she has filed the application. Noteworthy, too, in years following the 1976 revisions, Congress resisted efforts to eliminate §411(a) and the registration requirement embedded in it. In 1988, Congress removed foreign works from §411(a)’s dominion in order to comply with the Berne Convention for the Protection of Literary and Artistic Works’ bar on copyright formalities for such works. See §9(b)(1), 102Stat. 2859. Despite proposals to repeal §411(a)’s registration requirement entirely, however, see S. Rep. No. 100‒352, p. 36 (1988), Congress maintained the requirement for domestic works, see §411(a). Subsequently, in 1993, Congress considered, but declined to adopt, a proposal to allow suit immediately upon submission of a registration application. See H. R. Rep. No. 103–338, p. 4 (1993). And in 2005, Congress made a preregistration option available for works vulnerable to predistribution infringement. See Artists’ Rights and Theft Prevention Act of 2005, §104, 119Stat. 221. See also supra, at 3–4. Congress chose that course in face of calls to eliminate registration in cases of predistribution infringement. 70 Fed. Reg. 42286. Time and again, then, Congress has maintained registration as prerequisite to suit, and rejected proposals that would have eliminated registration or tied it to the copyright claimant’s application instead of the Register’s action.[5] Fourth Estate additionally argues that, as “registration is not a condition of copyright protection,” 17 U. S. C. §408(a), §411(a) should not be read to bar a copyright claimant from enforcing that protection in court once she has submitted a proper application for registration. Brief for Petitioner 37. But as explained supra, at 3, the Copyright Act safeguards copyright owners, irrespective of registration, by vesting them with exclusive rights upon creation of their works and prohibiting infringement from that point forward. If infringement occurs before a copyright owner applies for registration, that owner may eventually recover damages for the past infringement, as well as the infringer’s profits. §504. She must simply apply for registration and receive the Copyright Office’s decision on her application before instituting suit. Once the Register grants or refuses registration, the copyright owner may also seek an injunction barring the infringer from continued violation of her exclusive rights and an order requiring the infringer to destroy infringing materials. §§502, 503(b). Fourth Estate maintains, however, that if infringement occurs while the Copyright Office is reviewing a registration application, the registration approach will deprive the owner of her rights during the waiting period. Brief for Petitioner 41. See also 1 P. Goldstein, Copyright §3.15, p. 3:154.2 (3d ed. 2018 Supp.) (finding application approach “the better rule”); 2 M. Nimmer & D. Nimmer, Copyright §7.16[B][3][a], [b][ii] (2018) (infringement suit is conditioned on application, while prima facie presumption of validity depends on certificate of registration). The Copyright Act’s explicit carveouts from §411(a)’s general registration rule, however, show that Congress adverted to this concern. In the preregistration option, §408(f ), Congress provided that owners of works especially susceptible to prepublication infringement should be allowed to institute suit before the Register has granted or refused registration. See §411(a). Congress made the same determination as to live broadcasts. §411(c); see supra, at 4.[6] As to all other works, however, §411(a)’s general rule requires owners to await action by the Register before filing suit for infringement. Fourth Estate raises the specter that a copyright owner may lose the ability to enforce her rights if the Copyright Act’s three-year statute of limitations runs out before the Copyright Office acts on her application for registration. Brief for Petitioner 41. Fourth Estate’s fear is overstated, as the average processing time for registration applications is currently seven months, leaving ample time to sue after the Register’s decision, even for infringement that began before submission of an application. See U. S. Copyright Office, Registration Processing Times (Oct. 2, 2018) (Registration Processing Times), https://www.copyright.gov/ registration/docs/processing-times-faqs.pdf (as last visited Mar. 1, 2019). True, the statutory scheme has not worked as Congress likely envisioned. Registration processing times have increased from one or two weeks in 1956 to many months today. See GAO, Improving Productivity in Copyright Registration 3 (GAO–AFMD–83–13 1982); Registration Processing Times. Delays in Copyright Office processing of applications, it appears, are attributable, in large measure, to staffing and budgetary shortages that Congress can alleviate, but courts cannot cure. See 5 W. Patry, Copyright §17:83 (2019). Unfortunate as the current administrative lag may be, that factor does not allow us to revise §411(a)’s congressionally composed text. * * * For the reasons stated, we conclude that “registration . . . has been made” within the meaning of 17 U. S. C. §411(a) not when an application for registration is filed, but when the Register has registered a copyright after examining a properly filed application. The judgment of the Court of Appeals for the Eleventh Circuit is accordingly Affirmed. Notes 1 The Register of Copyrights is the “director of the Copyright Office of the Library of Congress” and is appointed by the Librarian of Congress. 17 U. S. C. §701(a). The Copyright Act delegates to the Register “[a]ll administrative functions and duties under [Title 17].” Ibid. 2 Consideration of Fourth Estate’s filings was initially delayed because the check Fourth Estate sent in payment of the filing fee was rejected by Fourth Estate’s bank as uncollectible. App. to Brief for United States as Amicus Curiae 1a. 3 The merits of the Copyright Office’s decision refusing registration are not at issue in this Court. 4 Section 411(a) provides, in principal part: “[N]o civil action for infringement of the copyright in any United States work shall be insti-tuted until preregistration or registration of the copyright claim has been made in accordance with this title. In any case, however, where the deposit, application, and fee required for registration have been delivered to the Copyright Office in proper form and registration has been refused, the applicant is entitled to institute a civil action for infringement if notice thereof, with a copy of the complaint, is served on the Register of Copyrights. The Register may, at his or her option, become a party to the action with respect to the issue of registrability of the copyright claim . . . .” 5 Fourth Estate asserts that, if a copyright owner encounters a lengthy delay in the Copyright Office, she may be forced to file a mandamus action to compel the Register to rule on her application, the very problem exposed in Vacheron & Constantin-Le Coultre Watches, Inc. v. Benrus Watch Co., 260 F.2d 637 (CA2 1958), see supra, at 8. But Congress’ answer to Vacheron, codified in §411(a)’s second sentence, was to permit an infringement suit upon refusal of registration, not to eliminate Copyright Office action as the trigger for an infringement suit. 6 Further, in addition to the Act’s provisions for preregistration suit, the Copyright Office allows copyright claimants to seek expedited processing of a claim for an additional $800 fee. See U. S. Copyright Office, Special Handling: Circular No. 10, pp. 1–2 (2017). The Copyright Office grants requests for special handling in situations involving, inter alia, “[p]ending or prospective litigation,” and “make[s] every attempt to examine the application . . . within five working days.” Compendium of U. S. Copyright Practices §623.2, 623.4 (3d ed. 2017).

587.US.2018_17-1299

Respondent Hyatt sued petitioner Franchise Tax Board of California (Board) in Nevada state court for alleged torts committed during a tax audit. The Nevada Supreme Court rejected the Board’s argument that the Full Faith and Credit Clause required Nevada courts to apply California law and immunize the Board from liability. The court held instead that general principles of comity entitled the Board only to the same immunity that Nevada law afforded Nevada agencies. This Court affirmed, holding that the Full Faith and Credit Clause did not prohibit Nevada from applying its own immunity law. On remand, the Nevada Supreme Court declined to apply a cap on tort liability applicable to Nevada state agencies. This Court reversed, holding that the Full Faith and Credit Clause required Nevada courts to grant the Board the same immunity that Nevada agencies enjoy. The Court was equally divided, however, on whether to overrule Nevada v. Hall, 440 U.S. 410, which held that the Constitution does not bar suits brought by an individual against a State in the courts of another State. On remand, the Nevada Supreme Court instructed the trial court to enter damages in accordance with Nevada’s statutory cap. The Board sought certiorari a third time, raising only the question whether Nevada v. Hall should be overruled. Held: Nevada v. Hall is overruled; States retain their sovereign immunity from private suits brought in courts of other States. Pp. 4–18. (a) The Hall majority held that nothing “implicit in the Constitution” requires States to adhere to the sovereign immunity that prevailed at the time of the founding. 440 U. S., at 417–418, 424–427. The Court concluded that the Founders assumed that “prevailing notions of comity would provide adequate protection against the unlikely prospect of an attempt by the courts of one State to assert jurisdiction over another.” Id., at 419. The Court’s view rested primarily on the idea that the States maintained sovereign immunity vis-à-vis each other in the same way that foreign nations do. Pp. 4–5. (b) Hall’s determination misreads the historical record and misapprehends the constitutional design created by the Framers. Although the Constitution assumes that the States retain their sovereign immunity except as otherwise provided, it also fundamentally adjusts the States’ relationship with each other and curtails the States’ ability, as sovereigns, to decline to recognize each other’s immunity in their own courts. Pp. 5–16. (1) At the time of the founding, it was well settled that States were immune from suit both under the common law and under the law of nations. The States retained these aspects of sovereignty, “except as altered by the plan of the Convention or certain constitutional Amendments.” Alden v. Maine, 527 U.S. 706, 713. Pp. 6–9. (2) Article III abrogated certain aspects of the States’ traditional immunity by providing a neutral federal forum in which the States agreed to be amenable to suits brought by other States. And in ratifying the Constitution, the States similarly surrendered a portion of their immunity by consenting to suits brought against them by the United States in federal courts. When this Court held in Chisholm v. Georgia, 2 Dall. 419, that Article III extended the federal judicial power over controversies between a State and citizens of another State, Congress and the States acted swiftly to draft and ratify the Eleventh Amendment, which confirms that the Constitution was not meant to “rais[e] up” any suits against the States that were “anomalous and unheard of when the Constitution was adopted,” Hans v. Louisiana, 134 U.S. 1, 18. The “natural inference” from the Amendment’s speedy adoption is that “the Constitution was understood, in light of its history and structure, to preserve the States’ traditional immunity from private suits.” Alden, supra, at 723–724. This view of the States’ sovereign immunity accorded with the understanding of the Constitution by its leading advocates, including Hamilton, Madison, and Marshall, when it was ratified. Pp. 9–12. (3) State sovereign immunity in another State’s courts is integral to the structure of the Constitution. The problem with Hyatt’s argument—that interstate sovereign immunity exists only as a matter of comity and can be disregarded by the forum State—is that the Constitution affirmatively altered the relationships between the States so that they no longer relate to each other as true foreign sovereigns. Numerous provisions reflect this reality. Article I divests the States of the traditional diplomatic and military tools that foreign sovereigns possess. And Article IV imposes duties on the States not required by international law. The Constitution also reflects alterations to the States’ relationships with each other, confirming that they are no longer fully independent nations free to disregard each other’s sovereignty. See New Hampshire v. Louisiana, 108 U.S. 76, 90. Hyatt’s argument is precisely the type of “ahistorical literalism” this Court has rejected when “interpreting the scope of the States’ sovereign immunity since the discredited decision in Chisholm.” Alden, supra, at 730. Moreover, his argument proves too much. Many constitutional doctrines not spelled out in the Constitution are nevertheless implicit in its structure and supported by historical practice, e.g., judicial review, Marbury v. Madison, 1 Cranch 137, 176–180. Pp. 12–16. (c) Stare decisis is “ ‘not an inexorable command,’ ” Pearson v. Callahan, 555 U.S. 223, 233, and is “at its weakest” when interpreting the Constitution, Agostini v. Felton, 521 U.S. 203, 235. The Court’s precedents identify, as relevant here, four factors to consider: the quality of the decision’s reasoning, its consistency with related decisions, legal developments since the decision, and reliance on the decision. See Janus v. State, County, and Municipal Employees, 585 U. S. ___, ___–___. The first three factors support overruling Hall. As to the fourth, case-specific reliance interests are not sufficient to persuade this Court to adhere to an incorrect resolution of an important constitutional question. Pp. 16–17. 133 Nev. ___, 407 P.3d 717, reversed and remanded. Thomas, J., delivered the opinion of the Court, in which Roberts, C. J., and Alito, Gorsuch, and Kavanaugh, JJ., joined. Breyer, J., filed a dissenting opinion, in which Ginsburg, Sotomayor, and Kagan, JJ., joined.

This case, now before us for the third time, requires us to decide whether the Constitution permits a State to be sued by a private party without its consent in the courts of a different State. We hold that it does not and overrule our decision to the contrary in Nevada v. Hall, 440 U.S. 410 (1979). I In the early 1990s, respondent Gilbert Hyatt earned substantial income from a technology patent for a com- puter formed on a single integrated circuit chip. Although Hyatt’s claim was later canceled, see Hyatt v. Boone, 146 F.3d 1348 (CA Fed. 1998), his royalties in the interim totaled millions of dollars. Prior to receiving the patent, Hyatt had been a long-time resident of California. But in 1991, Hyatt sold his house in California and rented an apartment, registered to vote, obtained insurance, opened a bank account, and acquired a driver’s license in Nevada. When he filed his 1991 and 1992 tax returns, he claimed Nevada—which collects no personal income tax, see Nev. Const., Art. 10, §1(9)—as his primary place of residence. Petitioner Franchise Tax Board of California (Board), the state agency responsible for assessing personal income tax, suspected that Hyatt’s move was a sham. Thus, in 1993, the Board launched an audit to determine whether Hyatt underpaid his 1991 and 1992 state income taxes by misrepresenting his residency. In the course of the audit, employees of the Board traveled to Nevada to conduct interviews with Hyatt’s estranged family members and shared his personal information with business contacts. In total, the Board sent more than 100 letters and demands for information to third parties. The Board ultimately concluded that Hyatt had not moved to Nevada until April 1992 and owed California more than $10 million in back taxes, interest, and penalties. Hyatt protested the audit before the Board, which upheld the audit after an 11-year administrative proceeding. The appeal of that decision remains pending before the California Office of Tax Appeals. In 1998, Hyatt sued the Board in Nevada state court for torts he alleged the agency committed during the audit. After the trial court denied in part the Board’s motion for summary judgment, the Board petitioned the Nevada Supreme Court for a writ of mandamus ordering dismissal on the ground that the State of California was immune from suit. The Board argued that, under the Full Faith and Credit Clause, Nevada courts must apply California’s statute immunizing the Board from liability for all injuries caused by its tax collection. See U. S. Const., Art. IV, §1; Cal. Govt. Code Ann. §860.2 (West 1995). The Nevada Supreme Court rejected that argument and held that, under general principles of comity, the Board was entitled to the same immunity that Nevada law afforded Nevada agencies—that is, immunity for negligent but not intentional torts. We granted certiorari and unanimously affirmed, holding that the Full Faith and Credit Clause did not prohibit Nevada from applying its own immunity law to the case. Franchise Tax Bd. of Cal. v. Hyatt, 538 U.S. 488, 498–499 (2003) (Hyatt I). Because the Board did not ask us to overrule Nevada v. Hall, supra, we did not revisit that decision. Hyatt I, supra, at 497. On remand, the trial court conducted a 4-month jury trial that culminated in a verdict for Hyatt that, with prejudgment interest and costs, exceeded $490 million. On appeal, the Nevada Supreme Court rejected most of the damages awarded by the lower court, upholding only a $1 million judgment on one of Hyatt’s claims and remanding for a new damages trial on another. Although the court recognized that tort liability for Nevada state agencies was capped at $50,000 under state law, it nonetheless held that Nevada public policy precluded it from applying that limitation to the California agency in this case. We again granted certiorari and this time reversed, holding that the Full Faith and Credit Clause required Nevada courts to grant the Board the same immunity that Nevada agencies enjoy. Franchise Tax Bd. of Cal. v. Hyatt, 578 U. S. ___, ___–___ (2016) (slip op., at 4–9) (Hyatt II ). Although the question was briefed and argued, the Court was equally divided on whether to overrule Hall and thus affirmed the jurisdiction of the Nevada Supreme Court. Hyatt II, supra, at ___ (slip op., at 1). On remand, the Nevada Supreme Court instructed the trial court to enter damages in accordance with the statutory cap for Nevada agencies. 133 Nev. ___, 407 P.3d 717 (2017). We granted, for a third time, the Board’s petition for certiorari, 585 U. S. ___ (2018). The sole question presented is whether Nevada v. Hall should be overruled.[1] II Nevada v. Hall is contrary to our constitutional design and the understanding of sovereign immunity shared by the States that ratified the Constitution. Stare decisis does not compel continued adherence to this erroneous precedent. We therefore overrule Hall and hold that States retain their sovereign immunity from private suits brought in the courts of other States. A Hall held that the Constitution does not bar private suits against a State in the courts of another State. 440 U. S., at 416–421. The opinion conceded that States were immune from such actions at the time of the founding, but it nonetheless concluded that nothing “implicit in the Constitution” requires States “to adhere to the sovereign-immunity doctrine as it prevailed when the Constitution was adopted.” Id., at 417–418, 424–427. Instead, the Court concluded that the Founders assumed that “prevailing notions of comity would provide adequate protection against the unlikely prospect of an attempt by the courts of one State to assert jurisdiction over another.” Id., at 419. The Court’s view rested primarily on the idea that the States maintained sovereign immunity vis-à-vis each other in the same way that foreign nations do, meaning that immunity is available only if the forum State “voluntar[ily]” decides “to respect the dignity of the [defendant State] as a matter of comity.” Id., at 416; see also id., at 424–427. The Hall majority was unpersuaded that the Constitution implicitly altered the relationship between the States. In the Court’s view, the ratification debates, the Eleventh Amendment, and our sovereign-immunity precedents did not bear on the question because they “concerned questions of federal-court jurisdiction.” Id., at 420. The Court also found unpersuasive the fact that the Constitution delineates several limitations on States’ authority, such as Article I powers granted exclusively to Congress and Article IV requirements imposed on States. Id., at 425. Despite acknowledging “that ours is not a union of 50 wholly independent sovereigns,” Hall inferred from the lack of an express sovereign immunity granted to the States and from the Tenth Amendment that the States retained the power in their own courts to deny immunity to other States. Ibid. Chief Justice Burger, Justice Blackmun, and Justice Rehnquist dissented. B Hall’s determination that the Constitution does not contemplate sovereign immunity for each State in a sister State’s courts misreads the historical record and misapprehends the “implicit ordering of relationships within the federal system necessary to make the Constitution a workable governing charter and to give each provision within that document the full effect intended by the Framers.” Id., at 433 (Rehnquist, J., dissenting). As Chief Justice Marshall explained, the Founders did not state every postulate on which they formed our Republic—“we must never forget, that it is a constitution we are expounding.” McCulloch v. Maryland, 4 Wheat. 316, 407 (1819). And although the Constitution assumes that the States retain their sovereign immunity except as otherwise provided, it also fundamentally adjusts the States’ relationship with each other and curtails their ability, as sovereigns, to decline to recognize each other’s immunity. 1 After independence, the States considered themselves fully sovereign nations. As the Colonies proclaimed in 1776, they were “Free and Independent States” with “full Power to levy War, conclude Peace, contract Alliances, establish Commerce, and to do all other Acts and Things which Independent States may of right do.” Declaration of Independence ¶4. Under international law, then, independence “entitled” the Colonies “to all the rights and powers of sovereign states.” McIlvaine v. Coxe’s Lessee, 4 Cranch 209, 212 (1808). “An integral component” of the States’ sovereignty was “their immunity from private suits.” Federal Maritime Comm’n v. South Carolina Ports Authority, 535 U.S. 743, 751–752 (2002); see Alden v. Maine, 527 U.S. 706, 713 (1999) (“[A]s the Constitution’s structure, its history, and the authoritative interpretations by this Court make clear, the States’ immunity from suit is a fundamental aspect of the sovereignty which the States enjoyed before the ratification of the Constitution, and which they retain today . . . ”). This fundamental aspect of the States’ “inviolable sovereignty” was well established and widely accepted at the founding. The Federalist No. 39, p. 245 (C. Rossiter ed. 1961) (J. Madison); see Alden, supra, at 715–716 (“[T]he doctrine that a sovereign could not be sued without its consent was universal in the States when the Constitution was drafted and ratified”). As Alexander Hamilton explained: “It is inherent in the nature of sovereignty not to be amenable to the suit of an individual without its consent. This is the general sense and the general practice of mankind; and the exemption, as one of the attributes of sovereignty, is now enjoyed by the government of every State in the Union.” The Federalist No. 81, at 487 (emphasis deleted). The Founders believed that both “common law sovereign immunity” and “law-of-nations sovereign immunity” prevented States from being amenable to process in any court without their consent. See Pfander, Rethinking the Supreme Court’s Original Jurisdiction in State-Party Cases, 82 Cal. L. Rev. 555, 581–588 (1994); see also Nelson, Sovereign Immunity as a Doctrine of Personal Jurisdiction, 115 Harv. L. Rev. 1559, 1574–1579 (2002). The common-law rule was that “no suit or action can be brought against the king, even in civil matters, because no court can have jurisdiction over him.” 1 W. Blackstone, Commentaries on the Laws of England 235 (1765) (Blackstone). The law-of-nations rule followed from the “perfect equality and absolute independence of sovereigns” under that body of international law. Schooner Exchange v. McFaddon, 7 Cranch 116, 137 (1812); see C. Phillipson, Wheaton’s Elements of International Law 261 (5th ed. 1916) (recognizing that sovereigns “enjoy equality before international law”); 1 J. Kent, Commentaries on American Law 20 (G. Comstock ed. 1867). According to the founding era’s foremost expert on the law of nations, “[i]t does not . . . belong to any foreign power to take cognisance of the administration of [another] sovereign, to set himself up for a judge of his conduct, and to oblige him to alter it.” 2 E. de Vattel, The Law of Nations §55, p. 155 (J. Chitty ed. 1883). The sovereign is “exemp[t] . . . from all [foreign] jurisdiction.” 4 id., §108, at 486. The founding generation thus took as given that States could not be haled involuntarily before each other’s courts. See Woolhandler, Interstate Sovereign Immunity, 2006 S. Ct. Rev. 249, 254–259. This understanding is perhaps best illustrated by preratification examples. In 1781, a creditor named Simon Nathan tried to recover a debt that Virginia allegedly owed him by attaching some of its property in Philadelphia. James Madison and other Virginia delegates to the Confederation Congress responded by sending a communique to Pennsylvania requesting that its executive branch have the action dismissed. See Letter from Virginia Delegates to Supreme Executive Council of Pennsylvania (July 9, 1781), in 3 The Papers of James Madison, 184–185 (W. Hutchinson & W. Rachal eds. 1963). As Madison framed it, the Commonwealth’s property could not be attached by process issuing from a court of “any other State in the Union.” Id., at 184. To permit otherwise would require Virginia to “abandon its Sovereignty by descending to answer before the Tribunal of another Power.” Ibid. Pennsylvania Attorney General William Bradford intervened, urging the Court of Common Pleas to dismiss the action. See Nathan v. Virginia, 1 Dall. 77, 78 (C. P. Phila. Cty. 1781). According to Bradford, the suit violated international law because “all sovereigns are in a state of equality and independence, exempt from each other’s jurisdiction.” Ibid. “[A]ll jurisdiction implies superiority over the party,” Bradford argued, “but there could be no superiority” between the States, and thus no jurisdiction, because the States were “perfect[ ly] equa[l]” and “entire[ly] independen[t].” Ibid. The court agreed and refused to grant Nathan the writ of attachment. Id., at 80. Similarly, a Pennsylvania Admiralty Court that very same year dismissed a libel action against a South Caro- lina warship, brought by its crew to recover unpaid wages. The court reasoned that the vessel was owned by a “sovereign independent state.” Moitez v. The South Carolina, 17 F. Cas. 574 (No. 9697) (1781). The Founders were well aware of the international-law immunity principles behind these cases. Federalists and Antifederalists alike agreed in their preratification debates that States could not be sued in the courts of other States. One Federalist, who argued that Article III would waive the States’ immunity in federal court, admitted that the waiver was desirable because of the “impossibility of calling a sovereign state before the jurisdiction of another sovereign state.” 3 Debates on the Constitution 549 (J. Elliot ed. 1876) (Pendleton) (Elliot’s Debates). Two of the most prominent Antifederalists—Federal Farmer and Brutus—disagreed with the Federalists about the desir- ability of a federal forum in which States could be sued, but did so for the very reason that the States had previously been “subject to no such actions” in any court and were not “oblige[d]” “to answer to an individual in a court of law.” Federal Farmer No. 3 (Oct. 10, 1787), in 4 The Founders’ Constitution 227 (P. Kurland & R. Lerner eds. 1987). They found it “humiliating and degrading” that a State might have to answer “the suit of an individual.” Brutus No. 13 (Feb. 21, 1788), in id., at 238. In short, at the time of the founding, it was well settled that States were immune under both the common law and the law of nations. The Constitution’s use of the term “States” reflects both of these kinds of traditional immu- nity. And the States retained these aspects of sovereignty, “except as altered by the plan of the Convention or certain constitutional Amendments.” Alden, 527 U. S., at 713. 2 One constitutional provision that abrogated certain aspects of this traditional immunity was Article III, which provided a neutral federal forum in which the States agreed to be amenable to suits brought by other States. Art. III, §2; see Alden, supra, at 755. “The establishment of a permanent tribunal with adequate authority to determine controversies between the States, in place of an inadequate scheme of arbitration, was essential to the peace of the Union.” Principality of Monaco v. Mississippi, 292 U.S. 313, 328 (1934). As James Madison explained during the Convention debates, “there can be no impropriety in referring such disputes” between coequal sovereigns to a superior tribunal. Elliot’s Debates 532. The States, in ratifying the Constitution, similarly surrendered a portion of their immunity by consenting to suits brought against them by the United States in federal courts. See Monaco, supra, at 328; Federal Maritime Comm’n, 535 U. S., at 752. “While that jurisdiction is not conferred by the Constitution in express words, it is inherent in the constitutional plan.” Monaco, supra, at 329. Given that “all jurisdiction implies superiority of power,” Blackstone 235, the only forums in which the States have consented to suits by one another and by the Federal Government are Article III courts. See Federal Maritime Comm’n, supra, at 752. The Antifederalists worried that Article III went even further by extending the federal judicial power over controversies “between a State and Citizens of another State.” They suggested that this provision implicitly waived the States’ sovereign immunity against private suits in federal courts. But “[t]he leading advocates of the Constitution assured the people in no uncertain terms” that this reading was incorrect. Alden, 527 U. S., at 716; see id., at 716–718 (citing arguments by Hamilton, Madison, and John Marshall). According to Madison: “[A federal court’s] jurisdiction in controversies between a state and citizens of another state is much objected to, and perhaps without reason. It is not in the power of individuals to call any state into court. The only operation it can have, is that, if a state should wish to bring a suit against a citizen, it must be brought before the federal court. This will give satisfaction to individuals, as it will prevent citizens, on whom a state may have a claim, being dissatisfied with the state courts.” Elliot’s Debates 533. John Marshall echoed these sentiments: “With respect to disputes between a state and the citizens of another state, its jurisdiction has been decried with unusual vehemence. I hope no gentleman will think that a state will be called at the bar of the federal court. . . . The intent is, to enable states to re- cover claims of individuals residing in other states. I contend this construction is warranted by the words.” Id., at 555 (emphasis in original). Not long after the founding, however, the Antifederalists’ fears were realized. In Chisholm v. Georgia, 2 Dall. 419 (1793), the Court held that Article III allowed the very suits that the “Madison-Marshall-Hamilton triumvirate” insisted it did not. Hall, 440 U. S., at 437 (Rehnquist, J., dissenting). That decision precipitated an immediate “furor” and “uproar” across the country. 1 J. Goebel, Antecedents and Beginnings to 1801, History of the Supreme Court of the United States 734, 737 (1971); see id., at 734–741. Congress and the States accordingly acted swiftly to remedy the Court’s blunder by drafting and ratifying the Eleventh Amendment.[2] See Edelman v. Jordan, 415 U.S. 651, 660–662 (1974); see also Federal Maritime Comm’n, supra, at 753 (acknowledging that Chisholm was incorrect); Alden, supra, at 721–722 (same). The Eleventh Amendment confirmed that the Constitution was not meant to “rais[e] up” any suits against the States that were “anomalous and unheard of when the Constitution was adopted.” Hans v. Louisiana, 134 U.S. 1, 18 (1890). Although the terms of that Amendment address only “the specific provisions of the Constitution that had raised concerns during the ratification debates and formed the basis of the Chisholm decision,” the “natural inference” from its speedy adoption is that “the Constitution was understood, in light of its history and structure, to preserve the States’ traditional immunity from private suits.” Alden, supra, at 723–724. We have often emphasized that “[t]he Amendment is rooted in a recognition that the States, although a union, maintain certain attributes of sovereignty, including sovereign immunity.” Puerto Rico Aqueduct and Sewer Authority v. Metcalf & Eddy, Inc., 506 U.S. 139, 146 (1993). In proposing the Amendment, “Congress acted not to change but to restore the original constitutional design.” Alden, 527 U. S., at 722. The “sovereign immunity of the States,” we have said, “neither derives from, nor is limited by, the terms of the Eleventh Amendment.” Id., at 713. Consistent with this understanding of state sovereign immunity, this Court has held that the Constitution bars suits against nonconsenting States in a wide range of cases. See, e.g., Federal Maritime Comm’n, supra (actions by private parties before federal administrative agencies); Alden, supra (suits by private parties against a State in its own courts); Blatchford v. Native Village of Noatak, 501 U.S. 775 (1991) (suits by Indian tribes in federal court); Monaco, 292 U.S. 313 (suits by foreign states in federal court); Ex parte New York, 256 U.S. 490 (1921) (admiralty suits by private parties in federal court); Smith v. Reeves, 178 U.S. 436 (1900) (suits by federal corporations in federal court). 3 Despite this historical evidence that interstate sovereign immunity is preserved in the constitutional design, Hyatt insists that such immunity exists only as a “matter of comity” and can be disregarded by the forum State. Hall, supra, at 416. He reasons that, before the Constitution was ratified, the States had the power of fully independent nations to deny immunity to fellow sovereigns; thus, the States must retain that power today with respect to each other because “nothing in the Constitution or formation of the Union altered that balance among the still-sovereign states.” Brief for Respondent 14. Like the majority in Hall, he relies primarily on our early foreign immunity decisions. For instance, he cites Schooner Exchange v. McFaddon, in which the Court dismissed a libel action against a French warship docked in Philadelphia because, under the law of nations, a sovereign’s warships entering the ports of a friendly nation are exempt from the jurisdiction of its courts. 7 Cranch, at 145–146. But whether the host nation respects that sovereign immunity, Chief Justice Marshall noted, is for the host nation to decide, for “[t]he jurisdiction of [a] nation within its own territory is necessarily exclusive and absolute” and “is susceptible of no limitation not imposed by itself.” Id., at 136. Similar reasoning is found in The Santissima Trinidad, 7 Wheat. 283, 353 (1822), where Justice Story noted that the host nation’s consent to provide immunity “may be withdrawn upon notice at any time, without just offence.” The problem with Hyatt’s argument is that the Constitution affirmatively altered the relationships between the States, so that they no longer relate to each other solely as foreign sovereigns. Each State’s equal dignity and sovereignty under the Constitution implies certain constitutional “limitation[s] on the sovereignty of all of its sister States.” World-Wide Volkswagen Corp. v. Woodson, 444 U.S. 286, 293 (1980). One such limitation is the inability of one State to hale another into its courts without the latter’s consent. The Constitution does not merely allow States to afford each other immunity as a matter of com- ity; it embeds interstate sovereign immunity within the constitutional design. Numerous provisions reflect this reality. To begin, Article I divests the States of the traditional diplomatic and military tools that foreign sovereigns possess. Specifically, the States can no longer prevent or remedy departures from customary international law because the Constitution deprives them of the independent power to lay imposts or duties on imports and exports, to enter into treaties or compacts, and to wage war. Compare Art. I, §10, with Declaration of Independence ¶4 (asserting the power to “levy War, conclude Peace, contract Alliances, [and] establish Commerce”); see Kansas v. Colorado, 185 U.S. 125, 143 (1902). Article IV also imposes duties on the States not required by international law. The Court’s Full Faith and Credit Clause precedents, for example, demand that state-court judgments be accorded full effect in other States and preclude States from “adopt[ing] any policy of hostility to the public Acts” of other States. Hyatt II, 578 U. S., at ___ (slip op., at 5) (internal quotation marks omitted); see Art. IV, §1. States must also afford citizens of each State “all Privileges and Immunities of Citizens in the several States” and honor extradition requests upon “Demand of the executive Authority of the State” from which the fugitive fled. Art. IV, §2. Foreign sovereigns cannot demand these kinds of reciprocal responsibilities absent consent or compact. But the Constitution imposes them as part of its transformation of the States from a loose league of friendship into a perpetual Union based on the “fundamental principle of equal sovereignty among the States.” Shelby County v. Holder, 570 U.S. 529, 544 (2013) (emphasis in original and internal quotation marks omitted). The Constitution also reflects implicit alterations to the States’ relationships with each other, confirming that they are no longer fully independent nations. See New Hampshire v. Louisiana, 108 U.S. 76, 90 (1883). For example, States may not supply rules of decision governing “disputes implicating the[ir] conflicting rights.” Texas Industries, Inc. v. Radcliff Materials, Inc., 451 U.S. 630, 641 (1981). Thus, no State can apply its own law to interstate disputes over borders, Cissna v. Tennessee, 246 U.S. 289, 295 (1918), water rights, Hinderlider v. La Plata River & Cherry Creek Ditch Co., 304 U.S. 92, 110 (1938), or the interpretation of interstate compacts, Petty v. Tennessee-Missouri Bridge Comm’n, 359 U.S. 275, 278–279 (1959). The States would have had the raw power to apply their own law to such matters before they entered the Union, but the Constitution implicitly forbids that exercise of power because the “interstate . . . nature of the contro- versy makes it inappropriate for state law to control.” Texas Industries, supra, at 641. Some subjects that were decided by pure “political power” before ratification now turn on federal “rules of law.” Rhode Island v. Massachusetts, 12 Pet. 657, 737 (1838). See Clark, Federal Common Law: A Structural Reinterpretation, 144 U. Pa. L. Rev. 1245, 1322–1331 (1996). Interstate sovereign immunity is similarly integral to the structure of the Constitution. Like a dispute over borders or water rights, a State’s assertion of compulsory judicial process over another State involves a direct conflict between sovereigns. The Constitution implicitly strips States of any power they once had to refuse each other sovereign immunity, just as it denies them the power to resolve border disputes by political means. Interstate immunity, in other words, is “implied as an essential component of federalism.” Hall, 440 U. S., at 430–431 (Blackmun, J., dissenting). Hyatt argues that we should find no right to sovereign immunity in another State’s courts because no constitutional provision explicitly grants that immunity. But this is precisely the type of “ahistorical literalism” that we have rejected when “interpreting the scope of the States’ sovereign immunity since the discredited decision in Chisholm.” Alden, 527 U. S., at 730; see id., at 736 (“[T]he bare text of the Amendment is not an exhaustive description of the States’ constitutional immunity from suit”). In light of our constitutional structure, the historical understanding of state immunity, and the swift enactment of the Eleventh Amendment after the Court departed from this understanding in Chisholm, “[i]t is not rational to suppose that the sovereign power should be dragged before a court.” Elliot’s Debates 555 (Marshall). Indeed, the spirited historical debate over Article III courts and the immediate reaction to Chisholm make little sense if the Eleventh Amendment were the only source of sovereign immunity and private suits against the States could already be brought in “partial, local tribunals.” Elliot’s Debates 532 (Madison). Nor would the Founders have objected so strenuously to a neutral federal forum for private suits against States if they were open to a State being sued in a different State’s courts. Hyatt’s view thus inverts the Founders’ concerns about state-court parochialism. Hall, supra, at 439 (Rehnquist, J., dissenting). Moreover, Hyatt’s ahistorical literalism proves too much. There are many other constitutional doctrines that are not spelled out in the Constitution but are nevertheless implicit in its structure and supported by historical practice—including, for example, judicial review, Marbury v. Madison, 1 Cranch 137, 176–180 (1803); intergovernmental tax immunity, McCulloch, 4 Wheat., at 435–436; executive privilege, United States v. Nixon, 418 U.S. 683, 705–706 (1974); executive immunity, Nixon v. Fitzgerald, 457 U.S. 731, 755–758 (1982); and the President’s re- moval power, Myers v. United States, 272 U.S. 52, 163–164 (1926). Like these doctrines, the States’ sovereign immunity is a historically rooted principle embedded in the text and structure of the Constitution. C With the historical record and precedent against him, Hyatt defends Hall on the basis of stare decisis. But stare decisis is “ ‘not an inexorable command,’ ” Pearson v. Callahan, 555 U.S. 223, 233 (2009), and we have held that it is “at its weakest when we interpret the Constitution because our interpretation can be altered only by constitutional amendment,” Agostini v. Felton, 521 U.S. 203, 235 (1997). The Court’s precedents identify a number of factors to consider, four of which warrant mention here: the quality of the decision’s reasoning; its consistency with related decisions; legal developments since the decision; and reliance on the decision. See Janus v. State, County, and Municipal Employees, 585 U. S. ___, ___–___ (2018) (slip op., at 34–35); United States v. Gaudin, 515 U.S. 506, 521 (1995). The first three factors support our decision to overrule Hall. We have already explained that Hall failed to account for the historical understanding of state sovereign immunity and that it failed to consider how the deprivation of traditional diplomatic tools reordered the States’ relationships with one another. We have also demon- strated that Hall stands as an outlier in our sovereign-immunity jurisprudence, particularly when compared to more recent decisions. As to the fourth factor, we acknowledge that some plaintiffs, such as Hyatt, have relied on Hall by suing sovereign States. Because of our decision to overrule Hall, Hyatt unfortunately will suffer the loss of two decades of litigation expenses and a final judgment against the Board for its egregious conduct. But in virtually every case that overrules a controlling precedent, the party relying on that precedent will incur the loss of litigation expenses and a favorable decision below. Those case-specific costs are not among the reliance interests that would persuade us to adhere to an incorrect resolution of an important constitutional question. * * * Nevada v. Hall is irreconcilable with our constitutional structure and with the historical evidence showing a widespread preratification understanding that States retained immunity from private suits, both in their own courts and in other courts. We therefore overrule that decision. Because the Board is thus immune from Hyatt’s suit in Nevada’s courts, the judgment of the Nevada Supreme Court is reversed, and the case is remanded for proceedings not inconsistent with this opinion. It is so ordered. Notes 1 Hyatt argues that the law-of-the-case doctrine precludes our review of this question, but he failed to raise that nonjurisdictional issue in his brief in opposition. We therefore deem this argument waived. See this Court’s Rule 15.2; Arizona v. California, 460 U.S. 605, 618 (1983) (“Law of the case directs a court’s discretion, it does not limit the tribunal’s power”). We also reject Hyatt’s argument that the Board waived its immunity. The Board has raised an immunity-based argument from this suit’s inception, though it was initially based on the Full Faith and Credit Clause. 2 The Eleventh Amendment provides: “The Judicial power of the United States shall not be construed to extend to any suit in law or equity, commenced or prosecuted against one of the United States by Citizens of another State, or by Citizens or Subjects of any Foreign State.”

587.US.2018_17-646

Petitioner Gamble pleaded guilty to a charge of violating Alabama’s felon-in-possession-of-a-firearm statute. Federal prosecutors then indicted him for the same instance of possession under federal law. Gamble moved to dismiss, arguing that the federal indictment was for “the same offence” as the one at issue in his state conviction, thus exposing him to double jeopardy under the Fifth Amendment. The District Court denied this motion, invoking the dual-sovereignty doctrine, according to which two offenses “are not the ‘same offence’ ” for double jeopardy purposes if “prosecuted by different sovereigns,” Heath v. Alabama, 474 U.S. 82, 92. Gamble pleaded guilty to the federal offense but appealed on double jeopardy grounds. The Eleventh Circuit affirmed. Held: This Court declines to overturn the longstanding dual-sovereignty doctrine. Pp. 3–31. (a) The dual-sovereignty doctrine is not an exception to the double jeopardy right but follows from the Fifth Amendment’s text. The Double Jeopardy Clause protects individuals from being “twice put in jeopardy” “for the same offence.” As originally understood, an “offence” is defined by a law, and each law is defined by a sovereign. Thus, where there are two sovereigns, there are two laws and two “offences.” Gamble attempts to show from the Clause’s drafting history that Congress must have intended to bar successive prosecutions regardless of the sovereign bringing the charge. But even if conjectures about subjective goals were allowed to inform this Court’s reading of the text, the Government’s contrary arguments on that score would prevail. Pp. 3–5. (b) This Court’s cases reflect the sovereign-specific reading of the phrase “same offence.” Three antebellum cases—Fox v. Ohio, 5 How. 410; United States v. Marigold, 9 How. 560; and Moore v. Illinois, 14 How. 13—laid the foundation that a crime against two sovereigns constitutes two offenses because each sovereign has an interest to vindicate. Seventy years later, that foundation was cemented in United States v. Lanza, 260 U.S. 377, which upheld a federal prosecution that followed one by a State. This Court applied that precedent for decades until 1959, when it refused two requests to reverse course, see Bartkus v. Illinois, 359 U.S. 121; Abbate v. United States, 359 U.S. 187, and it has reinforced that precedent over the following six decades, see, e.g., Puerto Rico v. Sanchez Valle, 579 U. S. ___. Pp. 5–10. (c) Gamble claims that this Court’s precedent contradicts the common-law rights that the Double Jeopardy Clause was originally understood to engraft onto the Constitution, pointing to English and American cases and treatises. A departure from precedent, however, “demands special justification,” Arizona v. Rumsey, 467 U.S. 203, 212, and Gamble’s historical evidence is too feeble to break the chain of precedent linking dozens of cases over 170 years. This Court has previously concluded that the probative value of early English decisions on which Gamble relies was “dubious” due to “confused and inadequate reporting.” Bartkus, 359 U. S., at 128, n. 9. On closer inspection, that assessment has proven accurate; the passing years have not made those early cases any clearer or more valuable. Nor do the treatises cited by Gamble come close to settling the historical question with enough force to meet his particular burden. His position is also not supported by state court cases, which are equivocal at best. Less useful still are the two federal cases cited by Gamble—Houston v. Moore, 5 Wheat. 1, which squares with the dual-sovereignty doctrine, and United States v. Furlong, 5 Wheat. 184, which actually supports it. Pp. 11–28. (d) Gamble’s attempts to blunt the force of stare decisis here do not succeed. He contends that the recognition of the Double Jeopardy Clause’s incorporation against the States washed away any theoretical foundation for the dual-sovereignty rule. But this rule rests on the fact that only same-sovereign prosecutions can involve the “same offence,” and that is just as true after incorporation as before. Gamble also argues that the proliferation of federal criminal laws has raised the risk of successive prosecutions under state and federal law for the same criminal conduct, thus compounding the harm inflicted by precedent. But this objection obviously assumes that precedent was erroneous from the start, so it is only as strong as the historical arguments found wanting. In any case, eliminating the dual-sovereignty rule would do little to trim the reach of federal criminal law or prevent many successive state and federal prosecutions for the same criminal conduct, see Blockburger v. United States, 284 U.S. 299. Pp. 28–31. 694 Fed. Appx. 750, affirmed. Alito, J., delivered the opinion of the Court, in which Roberts, C. J., and Thomas, Breyer, Sotomayor, Kagan, and Kavanaugh, JJ., joined. Thomas, J., filed a concurring opinion. Ginsburg, J., and Gorsuch, J., filed dissenting opinions.

We consider in this case whether to overrule a longstanding interpretation of the Double Jeopardy Clause of the Fifth Amendment. That Clause provides that no person may be “twice put in jeopardy” “for the same offence.” Our double jeopardy case law is complex, but at its core, the Clause means that those acquitted or convicted of a particular “offence” cannot be tried a second time for the same “offence.” But what does the Clause mean by an “offence”? We have long held that a crime under one sovereign’s laws is not “the same offence” as a crime under the laws of another sovereign. Under this “dual-sovereignty” doctrine, a State may prosecute a defendant under state law even if the Federal Government has prosecuted him for the same conduct under a federal statute. Or the reverse may happen, as it did here. Terance Gamble, convicted by Alabama for possessing a firearm as a felon, now faces prosecution by the United States under its own felon-in-possession law. Attacking this second prosecution on double jeopardy grounds, Gamble asks us to overrule the dual-sovereignty doctrine. He contends that it departs from the founding-era understanding of the right enshrined by the Double Jeopardy Clause. But the historical evidence assembled by Gamble is feeble; pointing the other way are the Clause’s text, other historical evidence, and 170 years of precedent. Today we affirm that precedent, and with it the decision below. I In November 2015, a local police officer in Mobile, Alabama, pulled Gamble over for a damaged headlight. Smelling marijuana, the officer searched Gamble’s car, where he found a loaded 9-mm handgun. Since Gamble had been convicted of second-degree robbery, his possession of the handgun violated an Alabama law providing that no one convicted of “a crime of violence” “shall own a firearm or have one in his or her possession.” Ala. Code §13A–11–72(a) (2015); see §13A–11–70(2) (defining “crime of violence” to include robbery). After Gamble pleaded guilty to this state offense, federal prosecutors indicted him for the same instance of possession under a federal law—one forbidding those convicted of “a crime punishable by imprisonment for a term exceeding one year . . . to ship or transport in interstate or foreign commerce, or possess in or affecting commerce, any firearm or ammunition.” 18 U. S. C. §922(g)(1). Gamble moved to dismiss on one ground: The federal indictment was for “the same offence” as the one at issue in his state conviction and thus exposed him to double jeopardy. But because this Court has long held that two offenses “are not the ‘same offence’ ” for double jeopardy purposes if “prosecuted by different sovereigns,” Heath v. Alabama, 474 U.S. 82, 92 (1985), the District Court denied Gamble’s motion to dismiss. Gamble then pleaded guilty to the federal offense while preserving his right to challenge the denial of his motion to dismiss on double jeopardy grounds. But on appeal the Eleventh Circuit affirmed, citing the dual-sovereignty doctrine. 694 Fed. Appx. 750 (2017). We granted certiorari to determine whether to overturn that doctrine.[1] 585 U. S. ___ (2018). II Gamble contends that the Double Jeopardy Clause must forbid successive prosecutions by different sovereigns because that is what the founding-era common law did. But before turning to that historical claim, see Part III infra, we review the Clause’s text and some of the cases Gamble asks us to overturn. A We start with the text of the Fifth Amendment. Al- though the dual-sovereignty rule is often dubbed an “exception” to the double jeopardy right, it is not an exception at all. On the contrary, it follows from the text that defines that right in the first place. “[T]he language of the Clause . . . protects individuals from being twice put in jeopardy ‘for the same offence,’ not for the same conduct or actions,” Grady v. Corbin, 495 U.S. 508, 529 (1990), as Justice Scalia wrote in a soon-vindicated dissent, see United States v. Dixon, 509 U.S. 688 (1993) (overruling Grady). And the term “ ‘[o]ffence’ was commonly understood in 1791 to mean ‘transgression,’ that is, ‘the Violation or Breaking of a Law.’ ” Grady, 495 U. S., at 529 (Scalia, J., dissenting) (quoting Dictionarium Britannicum (Bailey ed. 1730)). See also 2 R. Burn & J. Burn, A New Law Dictionary 167 (1792) (“OFFENCE, is an act committed against law, or omitted where the law requires it”). As originally understood, then, an “offence” is defined by a law, and each law is defined by a sovereign. So where there are two sovereigns, there are two laws, and two “offences.” See Grady, 495 U. S., at 529 (Scalia, J., dissenting) (“If the same conduct violates two (or more) laws, then each offense may be separately prosecuted”); Moore v. Illinois, 14 How. 13, 17 (1852) (“The constitutional provision is not, that no person shall be subject, for the same act, to be twice put in jeopardy of life or limb; but for the same offence, the same violation of law, no person’s life or limb shall be twice put in jeopardy” (emphasis added)). Faced with this reading, Gamble falls back on an episode from the Double Jeopardy Clause’s drafting history.[2] The first Congress, working on an earlier draft that would have banned “ ‘more than one trial or one punishment for the same offence,’ ” voted down a proposal to add “ ‘by any law of the United States.’ ” 1 Annals of Cong. 753 (1789). In rejecting this addition, Gamble surmises, Congress must have intended to bar successive prosecutions regardless of the sovereign bringing the charge. Even if that inference were justified—something that the Government disputes—it would count for little. The private intent behind a drafter’s rejection of one version of a text is shoddy evidence of the public meaning of an altogether different text. Cf. United States v. Craft, 535 U.S. 274, 287 (2002) (“[F]ailed legislative proposals are a particularly dangerous ground on which to rest an interpretation of a prior statute” (internal quotation marks omitted)). Besides, if we allowed conjectures about purpose to inform our reading of the text, the Government’s conjecture would prevail. The Government notes that the Declaration of Independence denounced King George III for “protecting [British troops] by a mock Trial, from punishment for any Murders which they should commit on the Inhabitants of these States.” ¶ 17. The Declaration was alluding to “the so-called Murderers’ Act, passed by Parliament after the Boston Massacre,” Amar, Sixth Amendment First Principles, 84 Geo. L. J. 641, 687, n. 181 (1996), a law that allowed British officials indicted for murder in America to be “ ‘tried in England, beyond the control of local juries.’ ” Ibid. (quoting J. Blum et al., The National Experience 95 (3d ed. 1973)). “During the late colonial period, Americans strongly objected to . . . [t]his circumvention of the judgment of the victimized community.” Amar, 84 Geo. L. Rev., at 687, n. 181. Yet on Gamble’s reading, the same Founders who quite literally revolted against the use of acquittals abroad to bar criminal prosecutions here would soon give us an Amendment allow- ing foreign acquittals to spare domestic criminals. We doubt it. We see no reason to abandon the sovereign-specific reading of the phrase “same offence,” from which the dual-sovereignty rule immediately follows. B Our cases reflect the same reading. A close look at them reveals how fidelity to the Double Jeopardy Clause’s text does more than honor the formal difference between two distinct criminal codes. It honors the substantive differences between the interests that two sovereigns can have in punishing the same act. The question of successive federal and state prosecutions arose in three antebellum cases implying and then spelling out the dual-sovereignty doctrine. The first, Fox v. Ohio, 5 How. 410 (1847), involved an Ohio prosecution for the passing of counterfeit coins. The defendant argued that since Congress can punish counterfeiting, the States must be barred from doing so, or else a person could face two trials for the same offense, contrary to the Fifth Amendment. We rejected the defendant’s premise that under the Double Jeopardy Clause “offences falling within the competency of different authorities to restrain or punish them would not properly be subjected to the consequences which those authorities might ordain and affix to their perpetration.” Id., at 435. Indeed, we observed, the nature of the crime or its effects on “public safety” might well “deman[d]” separate prosecutions. Ibid. Generalizing from this point, we declared in a second case that “the same act might, as to its character and tendencies, and the consequences it involved, constitute an offence against both the State and Federal governments, and might draw to its commission the penalties denounced by either, as appropriate to its character in reference to each.” United States v. Marigold, 9 How. 560, 569 (1850). A third antebellum case, Moore v. Illinois, 14 How. 13, expanded on this concern for the different interests of separate sovereigns, after tracing it to the text in the manner set forth above. Recalling that the Fifth Amendment prohibits double jeopardy not “for the same ac[t]” but “for the same offence,” and that “[a]n offence, in its legal signification, means the transgression of a law,” id., at 19, we drew the now-familiar inference: A single act “may be an offence or transgression of the laws of” two sovereigns, and hence punishable by both, id., at 20. Then we gave color to this abstract principle—and to the diverse interests it might vindicate—with an example. An assault on a United States marshal, we said, would offend against the Nation and a State: the first by “hindering” the “execution of legal process,” and the second by “breach[ing]” the “peace of the State.” Ibid. That duality of harm explains how “one act” could constitute “two offences, for each of which [the offender] is justly punishable.” Ibid. This principle comes into still sharper relief when we consider a prosecution in this country for crimes committed abroad. If, as Gamble suggests, only one sovereign may prosecute for a single act, no American court—state or federal—could prosecute conduct already tried in a foreign court. Imagine, for example, that a U. S. national has been murdered in another country. That country could rightfully seek to punish the killer for committing an act of violence within its territory. The foreign country’s interest lies in protecting the peace in that territory rather than protecting the American specifically. But the United States looks at the same conduct and sees an act of violence against one of its nationals, a person under the particular protection of its laws. The murder of a U. S. national is an offense to the United States as much as it is to the country where the murder occurred and to which the victim is a stranger. That is why the killing of an American abroad is a federal offense that can be prose- cuted in our courts, see 18 U. S. C. §2332(a)(1), and why customary international law allows this exercise of jurisdiction. There are other reasons not to offload all prosecutions for crimes involving Americans abroad. We may lack confidence in the competence or honesty of the other country’s legal system. Less cynically, we may think that special protection for U. S. nationals serves key national interests related to security, trade, commerce, or scholarship. Such interests might also give us a stake in punishing crimes committed by U. S. nationals abroad—especially crimes that might do harm to our national security or foreign relations. See, e.g., §2332a(b) (bombings). These examples reinforce the foundation laid in our antebellum cases: that a crime against two sovereigns constitutes two offenses because each sovereign has an interest to vindicate. We cemented that foundation 70 years after the last of those antebellum cases, in a decision upholding a federal prosecution that followed one by a State. See United States v. Lanza, 260 U.S. 377, 382 (1922) (“[A]n act denounced as a crime by both national and state sovereignties is an offense against the peace and dignity of both and may be punished by each”). And for decades more, we applied our precedent without qualm or quibble. See, e.g., Screws v. United States, 325 U.S. 91 (1945); Jerome v. United States, 318 U.S. 101 (1943); Puerto Rico v. Shell Co. (P. R.), Ltd., 302 U.S. 253 (1937); Westfall v. United States, 274 U.S. 256 (1927); Hebert v. Louisiana, 272 U.S. 312 (1926). When petitioners in 1959 asked us twice to reverse course, we twice refused, finding “[n]o consideration or persuasive reason not presented to the Court in the prior cases” for disturbing our “firmly established” doctrine. Abbate v. United States, 359 U.S. 187, 195; see also Bartkus v. Illinois, 359 U.S. 121. And then we went on enforcing it, adding another six decades of cases to the doctrine’s history. See, e.g., Puerto Rico v. Sánchez Valle, 579 U. S. ___ (2016); Heath v. Alabama, 474 U.S. 82 (1985); United States v. Wheeler, 435 U.S. 313 (1978); Rinaldi v. United States, 434 U.S. 22 (1977) ( per curiam). C We briefly address two objections to this analysis. First, the dissents contend that our dual-sovereignty rule errs in treating the Federal and State Governments as two separate sovereigns when in fact sovereignty belongs to the people. See post, at 3 (opinion of Ginsburg, J.); post, at 7 (opinion of Gorsuch, J.). This argument is based on a non sequitur. Yes, our Constitution rests on the principle that the people are sovereign, but that does not mean that they have conferred all the attributes of sovereignty on a single government. Instead, the people, by adopting the Constitution, “ ‘split the atom of sovereignty.’ ” Alden v. Maine, 527 U.S. 706, 751 (1999) (alteration omitted) (internal quotation marks and citation omitted). As we explained last Term: “When the original States declared their independence, they claimed the powers inherent in sovereignty . . . . The Constitution limited but did not abolish the sovereign powers of the States, which retained ‘a residuary and inviolable sovereignty.’ The Federalist No. 39, p. 245 (C. Rossiter ed. 1961). Thus, both the Federal Government and the States wield sovereign powers, and that is why our system of government is said to be one of ‘dual sovereignty.’ Gregory v. Ashcroft, 501 U.S. 452, 457 (1991).” Murphy v. National Collegiate Athletic Assn., 584 U. S. ___, ___ (2018) (slip op., at 14). It is true that the Republic is “ ‘ONE WHOLE,’ ” post, at 3 (opinion of Ginsburg, J.) (quoting The Federalist No. 82, p. 493 (C. Rossiter ed. 1961) (A. Hamilton)); accord, post, at 7 (opinion of Gorsuch, J.). But there is a difference between the whole and a single part, and that difference underlies decisions as foundational to our legal system as McCulloch v. Maryland, 4 Wheat. 316 (1819). There, in terms so directly relevant as to seem presciently tailored to answer this very objection, Chief Justice Marshall distinguished precisely between “the people of a State” and “[t]he people of all the States,” id., at 428, 435; between the “sovereignty which the people of a single state possess” and the sovereign powers “conferred by the people of the United States on the government of the Union,” id., at 429–430; and thus between “the action of a part” and “the action of the whole,” id., at 435–436. In short, McCulloch’s famous holding that a State may not tax the national bank rested on a recognition that the States and the Nation have different “interests” and “right[s].” Id., 431, 436. One strains to imagine a clearer statement of the premises of our dual-sovereignty rule, or a more authoritative source. The United States is a federal republic; it is not, contrary to Justice Gorsuch’s suggestion, post, at 10–11, a unitary state like the United Kingdom. Gamble and the dissents lodge a second objection to this line of reasoning. They suggest that because the division of federal and state power was meant to promote liberty, it cannot support a rule that exposes Gamble to a second sentence. See post, at 3–4 (opinion of Ginsburg, J.); post, at 8–9 (opinion of Gorsuch, J.). This argument fundamentally misunderstands the governmental structure established by our Constitution. Our federal system advances individual liberty in many ways. Among other things, it limits the powers of the Federal Government and protects certain basic liberties from infringement. But because the powers of the Federal Government and the States often overlap, allowing both to regulate often results in two layers of regulation. Taxation is an example that comes immediately to mind. It is also not at all uncommon for the Federal Government to permit activities that a State chooses to forbid or heavily restrict—for example, gambling and the sale of alcohol. And a State may choose to legalize an activity that federal law prohibits, such as the sale of marijuana. So while our system of federalism is fundamental to the protection of liberty, it does not always maximize individual liberty at the expense of other interests. And it is thus quite extraordinary to say that the venerable dual-sovereignty doctrine represents a “ ‘desecrat[ion]’ ” of federalism. Post, at 9 (opinion of Gorsuch, J.). III Gamble claims that our precedent contradicts the common-law rights that the Double Jeopardy Clause was originally understood to engraft onto the Constitution—rights stemming from the “common-law pleas of auterfoits acquit [former acquittal] and auterfoits convict [former conviction].” Grady, 495 U. S., at 530 (Scalia, J., dissenting). These pleas were treated as “reason[s] why the prisoner ought not to answer [an indictment] at all, nor put himself upon his trial for the crime alleged.” 4 W. Blackstone, Commentaries on the Laws of England 335 (1773) (Blackstone). Gamble argues that those who ratified the Fifth Amendment understood these common-law principles (which the Amendment constitutionalized) to bar a domestic prosecution following one by a foreign nation. For support, he appeals to early English and American cases and treatises. We have highlighted one hurdle to Gamble’s reading: the sovereign-specific original meaning of “offence.” But the doctrine of stare decisis is another obstacle. Stare decisis “promotes the evenhanded, predictable, and consistent development of legal principles, fosters reliance on judicial decisions, and contributes to the actual and perceived integrity of the judicial process.” Payne v. Tennessee, 501 U.S. 808, 827 (1991). Of course, it is also important to be right, especially on constitutional matters, where Congress cannot override our errors by ordinary legislation. But even in constitutional cases, a departure from precedent “demands special justification.” Arizona v. Rumsey, 467 U.S. 203, 212 (1984). This means that something more than “ambiguous historical evidence” is required before we will “flatly overrule a number of major decisions of this Court.” Welch v. Texas Dept. of Highways and Public Transp., 483 U.S. 468, 479 (1987). And the strength of the case for adhering to such decisions grows in proportion to their “antiquity.” Montejo v. Louisiana, 556 U.S. 778, 792 (2009). Here, as noted, Gamble’s historical arguments must overcome numerous “major decisions of this Court” spanning 170 years. In light of these factors, Gamble’s historical evidence must, at a minimum, be better than middling. And it is not. The English cases are a muddle. Trea- tises offer spotty support. And early state and federal cases are by turns equivocal and downright harmful to Gamble’s position. All told, this evidence does not establish that those who ratified the Fifth Amendment took it to bar successive prosecutions under different sovereigns’ laws—much less do so with enough force to break a chain of precedent linking dozens of cases over 170 years. A Gamble’s core claim is that early English cases reflect an established common-law rule barring domestic prosecution following a prosecution for the same act under a different sovereign’s laws. But from the very dawn of the common law in medieval England until the adoption of the Fifth Amendment in 1791, there is not one reported decision barring a prosecution based on a prior trial under foreign law. We repeat: Gamble has not cited and we have not found a single pre- Fifth Amendment case in which a foreign acquittal or conviction barred a second trial in a British or American court. Given this void, Gamble faces a considerable challenge in convincing us that the Fifth Amendment was originally understood to establish such a bar. Attempting to show that such a bar was available, Gamble points to five early English decisions for which we have case reports. We will examine these in some detail, but we note at the outset that they play only a secondary role for Gamble. The foundation of his argument is a decision for which we have no case report: the prosecution in England in 1677 of a man named Hutchinson. (We have a report of a decision denying Hutchinson bail but no report of his trial.) As told by Gamble, Hutchinson, having been tried and acquitted in a foreign court for a murder committed abroad, was accused of the same homicide in an English tribunal, but the English court held that the foreign prosecution barred retrial. Everything for Gamble stems from this one unreported decision. To the extent that the cases he cites provide any support for his argument—and for the most part, they do not—those cases purport to take their cue from the Hutchinson episode; the same is true of the treatises on which Gamble relies. So what evidence do we have about what actually happened to Hutchinson? The most direct evidence is a report of his application for bail before the Court of King’s Bench. The report spans all of one sentence: “On Habeas Corpus it appeared the Defendant was committed to Newgate on suspicion of Murder in Portugal, which by Mr. Attorny being a Fact out of the Kings Dominions, is not triable by Commission, upon 35 H. 8. Cap. 2. §. I. N. 2. but by a Constable and Marshal, and the Court refused to Bail him, & c.” Rex v. Hutchinson, 3 Keb. 785, 84 Eng. Rep. 1011 (1677). From this report, all that we can tell about the court’s thinking is that it found no convincing reason to grant bail, as was typical in murder cases.[3] The rest of the report concerns claims by an attorney. We are told that he contested the jurisdiction of the commission before which Hutchinson was to be tried, apparently a special commission that would have issued pursuant to a statute enacted under Henry VIII.[4] The commission lacked jurisdiction, the attorney seemed to suggest, because the crime had occurred in Portugal and thus “out of the Kings Dominions.” The attorney claimed that jurisdiction lay instead with “a Constable and Marshal”—an apparent reference to the High Court of Chivalry, which dealt with treason and murder committed abroad.[5] But what, if anything, did the King’s Bench make of the attorney’s jurisdictional claims? And more to the point, what happened after bail was denied? The bail report does not say. If Hutchinson did ultimately appear before the Court of Chivalry—and if that court accepted a plea of prior acquittal in Portugal—this would be paltry evidence of any common-law principle, which is what Gamble cites Hutchinson to establish. After all, the High Court of Chivalry was a civil-law court prohibited from proceeding under the common law (unlike every other English court of the time save Admiralty). 8 Ric. 2 ch. 5; see also Squibb 162; id., at xxv–xxvi (“The essential distinction between the Court of Chivalry and other courts is . . . that it administers justice in relation to those military matters which are not governed by the common law”). Nor would it be any surprise that we have no report of the proceeding; in fact, “[t]here is no report of a case in which a judge of the Court [of Chivalry] has set out the reasons for his decision earlier than the [20th] century.” Id., at 162. In the end, we have only two early accounts from judges of what finally became of Hutchinson, and both are indirect and shaky. First, they appear in the reports of cases decided in the Court of Chancery more than a half century after Hutchinson. Second, both judges cite only one source, and it is of lower authority than their own: namely, an account of Hutchinson given by an interested party (a defendant) in a previous, non-criminal case—an account on which the court in that case did not rely or even comment.[6] Insofar as our two judges seem to add their own details to the Hutchinson saga, we are not told where they obtained this information or whether it reflects mere guesses as to how gaps in the story should be filled in, decades after the fact. Finally, the two judges’ accounts are not entirely consistent. Still, they are the only early judicial glosses on Hutchinson that we have, so we will work with them. The more extensive account appears in the case of Gage v. Bulkeley, Ridg. T. H. 263, 27 Eng. Rep. 824 (Ch. 1744), and what the court said there—far from supporting Gamble’s argument—cuts against it. Gage involved a bill in chancery for an account of money deposited with a banker in Paris. The defendants pleaded, as a bar to this lawsuit, “a sentence” “given upon” the same demand in a French court. Ibid. In addressing this plea, Lord Chancellor Hardwicke first determined that foreign judgments are not binding in an English court of law. Here his reasoning was very similar to that found in our dual-sovereignty decisions. Because each judgment rests on the authority of a particular sovereign, the Chancellor thought, it cannot bind foreign courts, which operate by the power of a different sovereign. Id., at 263–264, 27 Eng. Rep., at 824. Turning next to courts of equity, the Lord Chancellor saw no reason that the rule should be any different; there too, he thought, a foreign judgment is not binding. Id., at 273, 27 Eng. Rep., at 827. But he did allow that in equity a foreign judgment could serve as “evidence, which may affect the right of [a plaintiff] when the cause comes to be heard.” Ibid. Elaborating on why foreign judgments did not bind English courts, whether of law or equity, the Lord Chancellor explained why Hutchinson was “no proof” to the contrary. In the Chancellor’s telling, Hutchinson was not indicted by the Court of King’s Bench, which could have tried a murder committed in England,[7] because that court had no jurisdiction over a homicide committed in Portugal. Gage, Ridge. T. H., at 271, 27 Eng. Rep., at 826–827. Instead, Hutchinson was (as the bail decision indicates) before that court on a writ of habeas corpus, and his case “was referred to the judges to know whether a commission should issue” under a statute similar to the one mentioned in the bail decision. Ibid., 27 Eng. Rep., at 827; see 33 Hen. 8 ch. 28 (1541–1542).[8] “And,” he explained, “the judges very rightly and mercifully thought not, because he had undergone one trial already.” Gage, Ridg. T. H., at 271–272, 27 Eng. Rep., at 827 (emphasis added). This suggests that Hutchinson was spared retrial as a matter of discretion (“merc[y]”)—which must be true if the Chancellor was right that foreign judgments were not binding. Indeed, at least one modern scholar agrees (on other grounds as well) that the result in Hutchinson may have been based on “expediency rather than law.” M. Friedland, Double Jeopardy 362–363 (1969). In the end, then, Gage is doubly damaging to Gamble. First, it squarely rejects the proposition that a litigant in an English court—even a civil litigant in equity—had a right to the benefit of a foreign judgment, a right that the Fifth Amendment might have codified. And second, Gage undermines Gamble’s chief historical example, Hutchinson, by giving a contrary reading of that case—and doing so, no less, in one of the only two judicial accounts of Hutchinson that we have from before the Fifth Amendment. The other account appears in Burrows v. Jemino, 2 Str. 733, 93 Eng. Rep. 815 (K. B. 1726).[9] In Burrows, a party that was sued in England on a bill of exchange sought an injunction against this suit in the Court of Chancery, contending that the suit was barred by the judgment of a court in Italy. In explaining why he would grant the injunction, Lord Chancellor King cited Hutchinson, which he thought had involved an acquittal in Spanish court that was “allowed to be a good bar to any proceedings here.” 2 Str., at 733, 93 Eng. Rep., at 815. This remark, showing that at least one English judge before the founding saw Hutchinson as Gamble does, provides a modicum of support for Gamble’s argument. But that support soft- ens just a few lines down in the report, where the Chan- cellor discusses the status of foreign judgments in courts of law in particular (as distinct from courts of equity like his own)—i.e., the courts that actually applied the common-law rules later codified by the Fifth Amendment. Here the Chancellor explained that while he personally would have accepted an Italian judgment as barring any suit at law, “other Judges might be of a different opinion.” Ibid. As a whole, then, the Chancellor’s comments in Burrows can hardly be cited to prove that the common law had made up its mind on this matter; just the opposite. Gamble’s other cases have even less force. The “most instructive” case, he claims, see Brief for Petitioner 13, is the 1775 case of King v. Roche, 1 Leach 134,[10] 168 Eng. Rep. 169 (K.B.), but that is a curious choice since the Roche court does not so much as mention Hutchinson or even tacitly affirm its supposed holding. The defendant in Roche entered two pleas: prior acquittal abroad and not guilty of the charged crime. All that the Roche court held was that, as a procedural matter, it made no sense to charge the jury with both pleas at once, because a finding for Roche on the first (prior acquittal) would, if successful, bar consideration of the second (not guilty). Roche, 1 Leach, at 135, 168 Eng. Rep., at 169. But on our key question—whether a plea based on a foreign acquittal could be successful—the Roche court said absolutely nothing; it had no occasion to do so. Before the prosecution could reply to Roche’s plea of prior acquittal, he withdrew it, opting for a full trial. The name Hutchinson does not appear even in the marginalia of the 1789 edition of Roche, which existed in 1791. See Captain Roche’s Case, 1 Leach at 138–139. Hutchinson is mentioned in connection with Roche only after the Fifth Amendment’s ratification, and only in a compiler’s annotation to the 1800 edition of the Roche case report. See 168 Eng. Rep., at 169, n. (a). That annotation in turn cites one case as support for its reading of Hutchinson: Beak v. Thyrwhit, 3 Mod. 194, 87 Eng. Rep. 124 (K. B. 1688). But Beak did not involve a foreign prosecution; indeed, it did not involve a prosecution at all. It was an admiralty case for trover and conversion of a ship, and—more to the point—Hutchinson is discussed only in the defendant’s argument in that case, not the court’s response. A report relaying the actual decision in Beak shows that the court ultimately said nothing about the defendant’s Hutchinson argument one way or another. See Beake v. Tyrrell, 1 Show. K. B. 6, 89 Eng. Rep. 411 (1688).[11] This same defendant’s argument was the only source of information about Hutchinson on which the Chancellors in Gage and Burrows explicitly relied, as we noted above. All later accounts of Hutchinson seem to stem from this one shallow root. The last of Gamble’s five pre- Fifth Amendment cases, Rex v. Thomas, 1 Lev. 118, 83 Eng. Rep. 326 (K. B. 1664), did not even involve a foreign prosecution. The defendant was indicted for murder in England, and he pleaded a prior acquittal by a Welsh court. But Wales was then part of the “kingdom of England”; its laws were “the laws of England and no other.” 1 Blackstone 94–95; see Thomas, 1 Lev., at 118, 83 Eng. Rep., at 326–327. So the prior trial in Thomas was not under another sovereign’s laws, making it totally irrelevant for present purposes. Summing up the import of the preratification cases on which Gamble’s argument rests, we have the following: (1) not a single reported case in which a foreign acquittal or conviction barred a later prosecution for the same act in either Britain or America; (2) not a single reported decision in which a foreign judgment was held to be binding in a civil case in a court of law; (3) fragmentary and not entirely consistent evidence about a 17th-century case in which a defendant named Hutchinson, having been tried and acquitted for murder someplace in the Iberian Peninsula, is said to have been spared a second trial for this crime on some ground, perhaps out of “merc[y],” not as a matter of right; (4) two cases (one criminal, one in admiralty) in which a party invoked a prior foreign judgment, but the court did not endorse or rest anything on the party’s reliance on that judgment; and (5) two Court of Chancery cases actually holding that foreign judgments were not (or not generally) treated as barring trial at common law. This is the flimsy foundation in case law for Gamble’s argument that when the Fifth Amendment was ratified, it was well understood that a foreign criminal judgment would bar retrial for the same act. Surveying the pre- Fifth Amendment cases in 1959, we concluded that their probative value was “dubious” due to “confused and inadequate reporting.” Bartkus, 359 U. S., at 128, n. 9. Our assessment was accurate then, and the passing years have not made those early cases any clearer or more valuable. B Not to worry, Gamble responds: Whatever the English courts actually did prior to adoption of the Fifth Amendment, by that time the early English cases were widely thought to support his view. This is a curious argument indeed. It would have us hold that the Fifth Amendment codified a common-law right that existed in legend, not case law. In any event, the evidence that this right was thought to be settled is very thin. Gamble’s argument is based on treatises, but they are not nearly as helpful as he claims. Alone they do not come close to settling the historical question with enough force to meet Gamble’s particular burden under stare decisis. Gamble begins with Blackstone, but he reads volumes into a flyspeck. In the body of his Commentaries, all that Blackstone stated was that successive prosecutions could be barred by prior acquittals by “any court having competent jurisdiction of the offence.” 4 Blackstone 335. This is simply a statement of the general double-jeopardy rule, without a word on separate sovereigns. So Gamble directs our attention to a footnote that appears after the phrase “any court having competent jurisdiction.” The footnote refers to the report of Beak v. Thyrwhit, which, as noted, merely rehearses the argument of the defendant in that case, who in turn mentioned Hutchinson—but not in a criminal prosecution, much less one preceded by a foreign trial. This thread tying Blackstone to Hutchinson—a thread woven through footnotes and reports of reports but not a single statement by a court (or even by a party to an actual prosecution)—is tenuous evidence that Blackstone endorsed Gamble’s reading of Hutchinson. When Gamble’s attorney was asked at argument which other treatises he found most likely to have informed those who ratified the Fifth Amendment, he offered four. See Tr. of Oral Arg. 30–31. But two of the four treatises did not exist when the Fifth Amendment was ratified. See 1 J. Chitty, Criminal Law 458 (1816); 1 T. Starkie, Criminal Pleading 300–301, n. h (1814). And a third discusses not a single case involving a prior prosecution under foreign law. See 2 W. Hawkins, Pleas of the Crown 372 (1739). That leaves one treatise cited by Gamble that spoke to this issue before ratification, F. Buller, An Introduction to the Law Relative to Trials at Nisi Prius (5th ed. 1788). That treatise concerned the trial of civil cases, id., at 2, and its discussion of prior judgments appeared under the heading “Of Evidence in general,” id., at 221. After considering the evidentiary value of such documents as acts of Parliament, deeds, and depositions, Buller addressed what we would later call issue preclusion. Lifting language from an earlier publication, H. Bathurst, The Theory of Evidence 39 (1761), Buller wrote that a final judgment was “conclusive Evidence” “against all the World” of the factual determinations underlying the judgment. Buller, Nisi Prius, at 245. And it is on this basis that Buller (again lifting from Bathurst) said that even someone acquitted of a crime in Spain “might,” upon indictment in England, “plead the Acquittal in Spain in Bar.” Ibid. This endorsement of the preclusive effect of a foreign judgment in civil litigation (which even today is not uniformly accepted in this country[12]) provides no direct support for Gamble since his prior judgment was one of conviction, not acquittal. (There is, after all, a major difference between the preclusive effect of a prior acquittal and that of a prior conviction: Only the first would make a subsequent prosecution pointless, by requiring later courts to assume a defendant’s innocence from the start.) And in any case, the fleeting references in the Buller and Bat- hurst treatises are hardly sufficient to show that the Members of the First Congress and the state legislators who ratified the Fifth Amendment understood the Double Jeopardy Clause to bar a prosecution in this country after acquittal abroad for the same criminal conduct. Gamble attempts to augment his support by citing treatises published after the Fifth Amendment was adopted.[13] And he notes that the Court in District of Columbia v. Heller, 554 U.S. 570, 605–610 (2008), took treatises of a similar vintage to shed light on the public understanding in 1791 of the right codified by the Second Amendment. But the Heller Court turned to these later treatises only after surveying what it regarded as a wealth of authority for its reading—including the text of the Second Amendment and state constitutions. The 19th-century treatises were treated as mere confirmation of what the Court thought had already been established. Here Gamble’s evidence as to the understanding in 1791 of the double jeopardy right is not at all comparable. C When we turn from 19th-century treatises to 19th-century state cases, Gamble’s argument appears no stronger. The last time we looked, we found these state cases to be “inconclusive.” Bartkus, 359 U. S., at 131. They seemed to be evenly split and to “manifest conflict[s] in conscience” rather than confident conclusions about the common law. Ibid. Indeed, two of those cases manifested nothing more than a misreading of a then-recent decision of ours. Id., at 130. We see things no differently today. The distinction between believing successive prosecutions by separate sovereigns unjust and holding them unlawful appears right on the face of the first state case that Gamble discusses. In State v. Brown, 2 N. C. 100, 101 (1794), the court opined that it would be “against natural justice” for a man who stole a horse in the Ohio Territory to be punished for theft in North Carolina just for having brought the horse to that State. To avoid this result, the Brown court simply construed North Carolina’s theft law not to reach the defendant’s conduct. But it did so precisely because the defendant otherwise could face two prosecutions for the same act of theft—despite the common-law rule against double jeopardy for the same “offence”—since “the offence against the laws of this State, and the offence against the laws of [the Ohio Territory] are distinct; and satisfaction made for the offence committed against this State, is no satisfaction for the offence committed against the laws there.” Ibid. Far from undermining the dual-sovereignty rule, Brown expressly affirms it, rejecting outright the idea that a judgment in one sovereign’s court could “be pleadable in bar to an indictment” in another’s. Ibid. Other state courts were divided. Massachusetts and Michigan courts thought that at least some trials in either federal or state court could bar prosecution in the other, see Commonwealth v. Fuller, 49 Mass. 313, 318 (1844); Harlan v. People, 1 Doug. 207, 212 (Mich. 1843), but those antebellum cases are poor images of the founding-era common law, resting as they do on what we have explained, see Bartkus, 359 U. S., at 130, was a misreading of our then-recent decision in Houston v. Moore, 5 Wheat. 1 (1820), which we discuss below. A Vermont court did take the same view based on its own analysis of the question, State v. Randall, 2 Aik. 89, 100–101 (1827), but just a few years later a Virginia court declared the opposite, Hendrick v. Commonwealth, 32 Va. 707, 713 (1834) (punishment for forgery under both federal and Virginia law is not double punishment for the “same offence” since “the law of Virginia punishes the forgery, not because it is an offence against the U. States, but because it is an offence against this commonwealth”). And South Carolina—a perfect emblem of the time—produced cases cutting both ways. See State v. Antonio, 2 Tread. 776, 781 (1816); State v. Tutt, 2 Bail. 44, 47–48 (1831). This is not the quantum of support for Gamble’s claim about early American common law that might withstand his burden under stare decisis. And once we look beyond the Nation’s earliest years, the body of state-court decisions appears even less helpful to Gamble’s position. We aptly summarized those cases in Bartkus, 359 U. S., at 134–136, and need not add to that discussion here.[14] D Less useful still, for Gamble’s purposes, are the two early Supreme Court cases on which he relies. In the first, a member of the Pennsylvania militia was tried by a state court-martial for the federal offense of deserting the militia. See Houston v. Moore, 5 Wheat. 1 (1820). The accused objected that the state court-martial lacked jurisdiction to try this federal offense. Since the offense could be tried in federal court, the defendant argued, allowing the state court-martial to try him for this crime could expose him to successive federal and state prosecutions for the same offense. Justice Washington answered that a ruling in either federal or state court would bar a second trial in the other. See id., at 31. But as we later explained, “that language by Mr. Justice Washington reflected his belief that the state statute imposed state sanctions for violation of a federal criminal law. As he viewed the matter, the two trials would not be of similar crimes arising out of the same conduct; they would be of the same crime. Mr. Justice Johnson agreed that if the state courts had become empowered to try the defendant for the federal offense, then such a state trial would bar a federal prosecution. Thus Houston v. Moore can be cited only for the presence of a bar in a case in which the second trial is for a violation of the very statute whose violation by the same conduct has already been tried in the courts of an- other government empowered to try that question.” Bartkus, 359 U. S., at 130 (citations omitted). In other words, Justice Washington taught only that the law prohibits two sovereigns (in that case, Pennsylvania and the United States) from both trying an offense against one of them (the United States). That is consistent with our doctrine allowing successive prosecutions for offenses against separate sovereigns. In light of this reading of Houston, the case does not undercut our dual-sovereignty doctrine. It may seem strange to think of state courts as prosecuting crimes against the United States, but that is just what state courts and commentators writing within a decade of Houston thought it involved. See, e.g., Tutt, 2 Bail., at 47 (“In [Houston], the act punished by the law of the State, was certainly and exclusively an offence against the general Government . . . [whereas h]ere, certainly there is an offence against the State, and a very different one from that committed against the United States” (emphasis added)); 1 J. Kent, Commentaries on American Law 373–374 (1826) (“[M]any . . . acts of [C]ongress . . . permit jurisdiction, over the offences therein described, to be exercised by state magistrates and courts,” and what Houston bars are successive prosecutions for the same “crime against the United States”). Even the scholar Gamble cites for his cause finds Houston not “[o]n point” because it “was discussing the jurisdiction of the state court to try a crime against the nation and impose a fine payable to the latter government.” Grant, Successive Prosecutions by State and Nation: Common Law and British Empire Comparisons, 4 UCLA L. Rev. 1, 7, and n. 27 (1956) (citing Warren, Federal Criminal Laws and the State Courts, 38 Harv. L. Rev. 545 (1925)). Perhaps feeling Houston wobble, Gamble says pre-emptively that if it is “inconclusive,” Brief for Petitioner 26, other cases are clear. But the other federal case on which he leans is worse for his argument. In United States v. Furlong, 5 Wheat. 184, 197 (1820), we said that an acquittal of piracy in the court of any “civilized State” would bar prosecution in any other nation because piracy, as an “offence within the criminal jurisdiction of all nations,” is “punished by all.”[15] Ending his quotation from Furlong at this point, Gamble gives the impression that Furlong rejects any dual-sovereignty rule. But that impression is shattered by the next sentence: “Not so with the crime of murder.” Ibid. As to that crime, the Furlong Court was “inclined to think that an acquittal” in the United States “would not have been a good plea in a Court of Great Britain.” Ibid. (emphasis added). And that was precisely because murder is “punishable under the laws of each State” rather than falling under some “universal jurisdiction.” Ibid. (emphasis added). When it came to crimes that were understood to offend against more than one sovereign, Furlong treated them as separate offenses—just as we have a dozen times since, and just as we do today. Thus, of the two federal cases that Gamble cites against the dual-sovereignty rule, Houston squares with it and Furlong supports it. Together with the muddle in the early state cases, this undermines Gamble’s claim that the early American bench and bar took the Fifth Amendment to proscribe successive prosecutions by different sovereigns. And without making a splash in the legal practice of the time, a few early treatises by themselves cannot unsettle almost two centuries of precedent. IV Besides appealing to the remote past, Gamble contends that recent changes—one doctrinal, one practical—blunt the force of stare decisis here. They do not. A If historical claims form the chorus of Gamble’s argument, his refrain is “incorporation.” In Gamble’s telling, the recognition of the Double Jeopardy Clause’s incorporation against the States, see Benton v. Maryland, 395 U.S. 784, 794 (1969), washed away any theoretical foundation for the dual-sovereignty rule, see United States v. Gaudin, 515 U.S. 506, 521 (1995) (abrogating precedent when “subsequent decisions of this Court” have “eroded” its foundations). But this incorporation-changes-everything argument trades on a false analogy. The analogy Gamble draws is to the evolution of our doctrine on the Fourth Amendment right against unreasonable searches and seizures.[16] We have long enforced this right by barring courts from relying on evidence gathered in an illegal search. Thus, in Weeks v. United States, 232 U.S. 383, 391–393 (1914), the Court held that federal prosecutors could not rely on the fruits of an unreasonable search undertaken by federal agents. But what if state or local police conducted a search that would have violated the Fourth Amendment if conducted by federal agents? Before incorporation, the state search would not have violated the Federal Constitution, so federal law would not have barred admission of the resulting evidence in a state prosecution. But by the very same token, under what was termed “the silver-platter doctrine,” state authorities could hand such evidence over to federal prosecutors for use in a federal case. See id., at 398. Once the Fourth Amendment was held to apply to the States as well as the Federal Government, however, the silver-platter doctrine was scuttled. See Elkins v. United States, 364 U.S. 206 (1960); Wolf v. Colorado, 338 U.S. 25 (1949). Now the fruits of unreasonable state searches are inadmissible in federal and state courts alike. Gamble contends that the incorporation of the Double Jeopardy Clause should likewise end the dual-sovereignty rule, but his analogy fails. The silver-platter doctrine was based on the fact that the state searches to which it applied did not at that time violate federal law. Once the Fourth Amendment was incorporated against the States, the status of those state searches changed. Now they did violate federal law, so the basis for the silver-platter doctrine was gone. See Elkins, 364 U. S., at 213 (“The foundation upon which the admissibility of state-seized evidence in a federal trial originally rested—that unreasonable state searches did not violate the Federal Constitution—thus disappeared [with incorporation]”). By contrast, the premises of the dual-sovereignty doctrine have survived incorporation intact. Incorporation meant that the States were now required to abide by this Court’s interpretation of the Double Jeopardy Clause. But that interpretation has long included the dual-sovereignty doctrine, and there is no logical reason why incorporation should change it. After all, the doctrine rests on the fact that only same-sovereign successive prosecutions are prosecutions for the “same offense,” see Part II, supra—and that is just as true after incorporation as before. B If incorporation is the doctrinal shift that Gamble invokes to justify a departure from precedent, the practical change he cites is the proliferation of federal criminal law. Gamble says that the resulting overlap of federal and criminal codes heightens the risk of successive prosecutions under state and federal law for the same criminal conduct. Thus, Gamble contends, our precedent should yield to “ ‘far-reaching systemic and structural changes’ ” that make our “earlier error all the more egregious and harmful.” South Dakota v. Wayfair, Inc., 585 U. S. ___, ___ (2018) (slip op., at 18). But unlike Gamble’s appeal to incorporation, this argument obviously assumes that the dual-sovereignty doctrine was legal error from the start. So the argument is only as strong as Gamble’s argument about the original understanding of double jeopardy rights, an argument that we have found wanting. Insofar as the expansion of the reach of federal criminal law has been questioned on constitutional rather than policy grounds, the argument has focused on whether Congress has overstepped its legislative powers under the Constitution. See, e.g., Gonzales v. Raich, 545 U.S. 1, 57–74 (2005) (Thomas, J., dissenting). Eliminating the dual-sovereignty rule would do little to trim the reach of federal criminal law, and it would not even prevent many successive state and federal prosecutions for the same criminal conduct unless we also overruled the long-settled rule that an “offence” for double jeopardy purposes is defined by statutory elements, not by what might be described in a looser sense as a unit of criminal conduct. See Blockburger v. United States, 284 U.S. 299 (1932). Perhaps believing that two revolutionary assaults in the same case would be too much, Gamble has not asked us to overrule Blockburger along with the dual-sovereignty rule. * * * The judgment of the Court of Appeals for the Eleventh Circuit is affirmed. It is so ordered. Notes 1 In addressing that question, we follow the parties’ lead and assume, without deciding, that the state and federal offenses at issue here satisfy the other criteria for being the “same offence” under our double jeopardy precedent. See Blockburger v. United States, 284 U.S. 299, 304 (1932). 2 Gamble also cites founding-era uses of the word “offence” that are not tied to violations of a sovereign’s laws, but the examples are not very telling. Some, for instance, play on the unremarkable fact that at the founding, “offence” could take on a different sense in nonlegal settings, much as “offense” does today. In this vein, Gamble cites a 19th-century dictionary defining “offense” broadly as “any transgression of law, divine or human; a crime; sin; act of wickedness or omission of duty.” 2 N. Webster, An American Dictionary of the English Language (1828). But the question is what “offence” meant in legal contexts. See Moore v. Illinois, 14 How. 13, 19 (1852) (“An offence, in its legal signification, means the transgression of a law. . .” (emphasis added)). 3 See J. Beattie, Crime and the Courts in England: 1660–1800, pp. 281–282 (1986). 4 Although this Act reached conduct committed “out of the King Majesties Realme of Englande and other his Graces [Dominions],” Acte concerning the triall of Treasons 1543–1544, 35 Hen. 8 ch. 2 (1543–1544), it applied only to treasons and misprisions of treason—not to homicide, of which Hutchinson was accused. 5 See G. Squibb, The High Court of Chivalry 54, 147–148 (1959) (Squibb); 4 Blackstone 267. 6 See Gage v. Bulkeley, Ridg. T. H. 263, 271, 27 Eng. Rep. 824, 826–827 (Ch. 1744) (citing Beake v. Tyrrell, 1 Show. K. B. 6); Burrows v. Jemino, 2 Str. 733, 25 Eng. Rep. 235 (K. B. 1726) (same). As noted, the report cited by both judges—which also appears at 89 Eng. Rep. 411 (K. B. 1688)—mentions Hutchinson only in summarizing a defendant’s argument. So does the only other source cited by either judge. See Gage, Ridg. T. H., at 271, 27 Eng. Rep., at 826–827 (citing Beak v. Thyrwhit, 3 Mod. 194, 87 Eng. Rep. 124 (K. B. 1688)). Below we discuss in detail the case that figures in these two reports. See infra, at 19, and n. 11. 7 4 Blackstone 262. 8 This statute authorized commissioners to try certain defendants for acts of treason or murder committed “in whatsoever other Shire or place, within the King’s dominions or without.” But “[d]espite the words ‘or without’, contemporary opinion seems not to have regarded the extra-territorial operation of this Act as clear.” Squibb 149. Indeed, the statute cited in the Hutchinson bail report, dated to just two years later, cited lingering “doubtes and questions” about whether English courts could try treason committed abroad (in the course of clarifying that treason and misprisions of treason abroad could indeed be tried in England). 35 Hen. 8 ch. 2, § I. 9 This case is also reported as Burrows v. Jemineau in Sel. Ca. t. 69, 25 Eng. Rep. 228 (Ch. 1726); as Burroughs v. Jamineau in Mos. 1, 25 Eng. Rep. 235; as Burrows v. Jemineau in 2 Eq. Ca. Abr. 476, 22 Eng. Rep. 405; and as Burrows v. Jemino in 2 Eq. Ca. Abr. 524, 22 Eng. Rep. 443. 10 This case is reported as Captain Roche’s Case in 1 Leach 138 (1789 ed.) and in 2 Leach 125 (1792 ed.). 11 This decision is also reported as Beake v. Tirrell, Com. 120, 90 Eng. Rep. 379. 12 Compare Restatement (Fourth) of Foreign Relations Law of the United States § 481 (2018) (With a few specified exceptions, “a final, conclusive, and enforceable judgment of a court of a foreign state granting or denying recovery of a sum of money, or determining a legal controversy, is entitled to recognition by courts in the United States”) and Restatement (Second) of Conflict of Laws § 98, Comment b. (1969) (“In most respects,” judgments rendered in a foreign nation satisfying specified criteria “will be accorded the same degree of recognition to which sister State judgments are entitled”), with, e.g., Derr v. Swarek, 766 F.3d 430, 437 (CA5 2014) (recognition of foreign judgments is not required but is a matter of comity); Diorinou v. Mezitis, 237 F.3d 133, 142–143 (CA2 2001) (same); id., at 139–140 (“It is well-established that United States courts are not obliged to recognize judgments rendered by a foreign state, but may choose to give res judicata effect to foreign judgments on the basis of comity” (emphasis in original; internal quotation marks omitted)); MacArthur v. San Juan County, 497 F.3d 1057, 1067 (CA10 2007) (“Comity is not an inexorable command . . . and a request for recognition of a foreign judgment may be rebuffed on any number of grounds”); Guinness PLC v. Ward, 955 F.2d 875, 883 (CA4 1992) (“The effect to be given foreign judgments has therefore historically been determined by more flexible principles of comity”). 13 See, e.g., F. Wharton, A Treatise on the Law of Homicide in the United States 283 (1855); F. Wharton, A Treatise on the Criminal Law of the United States 137 (1846); L. MacNally, The Rules of Evidence on Pleas of the Crown 428 (1802). 14 As we put it in Bartkus, 359 U. S., at 134–136: “Of the twenty-eight States which have considered the validity of successive state and federal prosecutions as against a challenge of violation of either a state constitutional double-jeopardy provision or a common-law evidentiary rule of autrefois acquit and autrefois convict, twenty-seven have refused to rule that the second prosecution was or would be barred. These States were not bound to follow this Court and its interpretation of the Fifth Amendment. The rules, constitutional, statutory, or common law which bound them, drew upon the same experience as did the Fifth Amendment, but were and are of separate and independent authority. “Not all of the state cases manifest careful reasoning, for in some of them the language concerning double jeopardy is but offhand dictum. But in an array of state cases there may be found full consideration of the arguments supporting and denying a bar to a second prosecution. These courts interpreted their rules as not proscribing a second prosecution where the first was by a different government and for violation of a different statute.” (Footnote omitted.) 15 Piracy was understood as a violation of the law of nations, which was seen as common to all. That is why any successive prosecution for piracy, being under the same law, would have been for the same offense. See United States v. Smith, 5 Wheat. 153, 163, n. a (1820) (quoting definitions of piracy by several ancient and more recent authorities). See also 4 Blackstone 71 (“[T]he crime of piracy, or robbery and depredation upon the high seas, is an offence against the universal law of society; a pirate being, according to Sir Edward Coke, hostis humani generis [enemies of mankind]. As therefore he has renounced all the benefits of society and government, and has reduced himself afresh to the savage state of nature, by declaring war against all mankind, all mankind must declare war against him: so that every community has a right, by the rule of self-defence, to inflict that punishment upon him, which every individual would in a state of nature have been otherwise entitled to do, for any invasion of his person or personal property” (footnote omitted)). 16 He draws a similar analogy to the Fifth Amendment right against self-incrimination, but our response to his Fourth Amendment analogy would answer that argument as well.

586.US.2018_17-1026

Petitioner Gilberto Garza, Jr., signed two plea agreements, each arising from state criminal charges and each containing a clause stating that Garza waived his right to appeal. Shortly after sentencing, Garza told his trial counsel that he wished to appeal. Instead of filing a notice of appeal, counsel informed Garza that an appeal would be “problematic” given Garza’s appeal waiver. After the time period for Garza to preserve an appeal lapsed, he sought state postconviction relief, alleging that his trial counsel had rendered ineffective assistance by failing to file a notice of appeal despite his repeated requests. The Idaho trial court denied relief, and the Idaho Court of Appeals affirmed. Also affirming, the Idaho Supreme Court held that Garza could not show the requisite deficient performance by counsel and resulting prejudice. In doing so, the court concluded that the presumption of prejudice recognized in Roe v. Flores-Ortega, 528 U.S. 470, when trial counsel fails to file an appeal as instructed does not apply when the defendant has agreed to an appeal waiver. Held: Flores-Ortega’s presumption of prejudice applies regardless of whether a defendant has signed an appeal waiver. Pp. 3–14. (a) Under Strickland v. Washington, 466 U.S. 668, a defendant who claims ineffective assistance of counsel must prove (1) “that counsel’s representation fell below an objective standard of reasonableness,” id., at 687–688, and (2) that any such deficiency was “prejudicial to the defense,” id., at 692. However, “prejudice is presumed” in “certain Sixth Amendment contexts,” ibid., such as “when counsel’s constitutionally deficient performance deprives a defendant of an appeal that he otherwise would have taken,” Flores-Ortega, 528 U. S., at 484. Pp. 3–4. (b) This case hinges on two procedural devices: appeal waivers and notices of appeal. No appeal waiver serves as an absolute bar to all appellate claims. Because a plea agreement is essentially a contract, it does not bar claims outside its scope. And, like any contract, the language of appeal waivers can vary widely, leaving many types of claims unwaived. A waived appellate claim may also proceed if the prosecution forfeits or waives the waiver or if the Government breaches the agreement. Separately, some claims are treated as unwaiveable. Most fundamentally, courts agree that defendants retain the right to challenge whether the waiver itself was knowing and voluntary. The filing of a notice of appeal is “a purely ministerial task that imposes no great burden on counsel.” Flores-Ortega, 528 U. S., at 474. Filing requirements reflect that appellate claims are likely to be ill defined or unknown at the filing stage. And within the division of labor between defendants and their attorneys, the “ultimate authority” to decide whether to “take an appeal” belongs to the accused. Jones v. Barnes, 463 U.S. 745, 751. Pp. 4–7. (c) Garza’s attorney rendered deficient performance by not filing a notice of appeal in light of Garza’s clear requests. Given the possibility that a defendant will end up raising claims beyond an appeal waiver’s scope, simply filing a notice of appeal does not necessarily breach a plea agreement. Thus, counsel’s choice to override Garza’s instructions was not a strategic one. In any event, the bare decision whether to appeal is ultimately the defendant’s to make. Pp. 7–8. (d) Because there is no dispute that Garza wished to appeal, a direct application of Flores-Ortega’s language resolves this case. Flores-Ortega reasoned that because a presumption of prejudice applies whenever “ ‘the accused is denied counsel at a critical stage,’ ” it makes greater sense to presume prejudice when counsel’s deficiency forfeits an “appellate proceeding altogether.” 528 U. S., at 483. Because Garza retained a right to appeal at least some issues despite his waivers, he had a right to a proceeding and was denied that proceeding altogether as a result of counsel’s deficient performance. That he surrendered many claims by signing appeal waivers does not change things. First, the presumption of prejudice does not bend because a particular defendant seems to have had poor prospects. See, e.g., Jae Lee v. United States, 582 U. S. ___, ___. Second, while the defendant in Flores-Ortega did not sign an appeal waiver, he did plead guilty, which “reduces the scope of potentially appealable issues” on its own. 528 U. S., at 480. Pp. 8–10. (e) Contrary to the argument by Idaho and the U. S. Government, as amicus, that Garza never “had a right” to his appeal and thus that any deficient performance by counsel could not have caused the loss of any such appeal, Garza did retain a right to his appeal; he simply had fewer possible claims than some other appellants. The Government also proposes a rule that would require a defendant to show—on a case-by-case basis—that he would have presented claims that would have been considered by the appellate court on the merits. This Court, however, has already rejected attempts to condition the restoration of a defendant’s appellate rights forfeited by ineffective counsel on proof that the defendant’s appeal had merit. See, e.g., Rodriquez v. United States, 395 U.S. 327, 330. Moreover, it is not the defendant’s role to decide what arguments to press, making it especially improper to impose that role upon the defendant simply because his opportunity to appeal was relinquished by deficient counsel. And because there is no right to counsel in postconviction proceedings and, thus, most applicants proceed pro se, the Government’s proposal would be unfair, ill advised, and unworkable. Pp. 10–14. 162 Idaho 791, 405 P.3d 576, reversed and remanded. Sotomayor, J., delivered the opinion of the Court, in which Roberts, C. J., and Ginsburg, Breyer, Kagan, and Kavanaugh, JJ., joined. Thomas, J., filed a dissenting opinion, in which Gorsuch, J., joined, and in which Alito, J., joined as to Parts I and II.

In Roe v. Flores-Ortega, 528 U.S. 470 (2000), this Court held that when an attorney’s deficient performance costs a defendant an appeal that the defendant would have otherwise pursued, prejudice to the defendant should be presumed “with no further showing from the defendant of the merits of his underlying claims.” Id., at 484. This case asks whether that rule applies even when the defendant has, in the course of pleading guilty, signed what is often called an “appeal waiver”—that is, an agreement forgoing certain, but not all, possible appellate claims. We hold that the presumption of prejudice recognized in Flores-Ortega applies regardless of whether the defendant has signed an appeal waiver. I In early 2015, petitioner Gilberto Garza, Jr., signed two plea agreements, each arising from criminal charges brought by the State of Idaho. Each agreement included a clause stating that Garza “waive[d] his right to appeal.” App. to Pet. for Cert. 44a, 49a. The Idaho trial court accepted the agreements and sentenced Garza to terms of prison in accordance with the agreements. Shortly after sentencing, Garza told his trial counsel that he wished to appeal.[1] In the days that followed, he would later attest, Garza “continuously reminded” his attorney of this directive “via phone calls and letters,” Record 210, and Garza’s trial counsel acknowledged in his own affidavit that Garza had “told me he wanted to appeal the sentence(s) of the court,” id., at 151.[2] Garza’s trial counsel, however, did not file a notice of appeal. Instead, counsel “informed Mr. Garza that an appeal was problematic because he waived his right to appeal.” Ibid. The period of time for Garza’s appeal to be preserved came and went with no notice having been filed on Garza’s behalf.Roughly four months after sentencing, Garza sought postconviction relief in Idaho state court. As relevant here, Garza alleged that his trial counsel rendered ineffective assistance by failing to file notices of appeal despite Garza’s requests. The Idaho trial court denied relief, and both the Idaho Court of Appeals and the Idaho Supreme Court affirmed that decision. See 162 Idaho 791, 793, 405 P.3d 576, 578 (2017). The Idaho Supreme Court ruled that Garza, given the appeal waivers, needed to show both deficient performance and resulting prejudice; it concluded that he could not. See id., at 798, 405 P. 3d, at 583. In ruling that Garza needed to show prejudice, the Idaho Supreme Court acknowledged that it was aligning itself with the minority position among courts. For example, 8 of the 10 Federal Courts of Appeals to have considered the question have applied Flores-Ortega’s presumption of prejudice even when a defendant has signed an appeal waiver.[3] 162 Idaho, at 795, 405 P. 3d, at 580. We granted certiorari to resolve the split of authority. 585 U. S. ___ (2018). We now reverse. II A The Sixth Amendment guarantees criminal defendants “the right . . . to have the Assistance of Counsel for [their] defence.” The right to counsel includes “ ‘the right to the effective assistance of counsel.’ ” Strickland v. Washington, 466 U.S. 668, 686 (1984) (quoting McMann v. Richardson, 397 U.S. 759, 771, n. 14 (1970)). Under Strickland, a defendant who claims ineffective assistance of counsel must prove (1) “that counsel’s representation fell below an objective standard of reasonableness,” 466 U. S., at 687–688, and (2) that any such deficiency was “prejudicial to the defense,” id., at 692. “In certain Sixth Amendment contexts,” however, “prejudice is presumed.” Ibid. For example, no showing of prejudice is necessary “if the accused is denied counsel at a critical stage of his trial,” United States v. Cronic, 466 U.S. 648, 659 (1984), or left “entirely without the assistance of counsel on appeal,” Penson v. Ohio, 488 U.S. 75, 88 (1988). Similarly, prejudice is presumed “if counsel entirely fails to subject the prosecution’s case to meaningful adversarial testing.” Cronic, 466 U. S., at 659. And, most relevant here, prejudice is presumed “when counsel’s constitutionally deficient performance deprives a defendant of an appeal that he otherwise would have taken.” Flores-Ortega, 528 U. S., at 484. We hold today that this final presumption applies even when the defendant has signed an appeal waiver. B It is helpful, in analyzing Garza’s case, to first address two procedural devices on which the case hinges: appeal waivers and notices of appeal. 1 We begin with the term “appeal waivers.” While the term is useful shorthand for clauses like those in Garza’s plea agreements, it can misleadingly suggest a monolithic end to all appellate rights.[4] In fact, however, no appeal waiver serves as an absolute bar to all appellate claims. As courts widely agree, “[a] valid and enforceable appeal waiver . . . only precludes challenges that fall within its scope.” United States v. Hardman, 778 F.3d 896, 899 (CA11 2014); see also ibid., n. 2 (collecting cases from the 11 other Federal Courts of Appeals with criminal jurisdiction); State v. Patton, 287 Kan. 200, 228–229, 195 P.3d 753, 771 (2008). That an appeal waiver does not bar claims outside its scope follows from the fact that, “[a]lthough the analogy may not hold in all respects, plea bargains are essentially contracts.” Puckett v. United States, 556 U.S. 129, 137 (2009). As with any type of contract, the language of appeal waivers can vary widely, with some waiver clauses leaving many types of claims unwaived.[5] Additionally, even a waived appellate claim can still go forward if the prosecution forfeits or waives the waiver. E.g., United States v. Story, 439 F.3d 226, 231 (CA5 2006). Accordingly, a defendant who has signed an appeal waiver does not, in directing counsel to file a notice of appeal, necessarily undertake a quixotic or frivolous quest. Separately, all jurisdictions appear to treat at least some claims as unwaiveable. Most fundamentally, courts agree that defendants retain the right to challenge whether the waiver itself is valid and enforceable—for example, on the grounds that it was unknowing or involuntary.[6] Consequently, while signing an appeal waiver means giving up some, many, or even most appellate claims, some claims nevertheless remain. 2 It is also important to consider what it means—and does not mean—for trial counsel to file a notice of appeal. “Filing such a notice is a purely ministerial task that imposes no great burden on counsel.” Flores-Ortega, 528 U. S., at 474. It typically takes place during a compressed window: 42 days in Idaho, for example, and just 14 days in federal court. See Idaho Rule App. Proc. 14(a) (2017); Fed. Rule App. Proc. 4(b)(1)(A). By the time this window has closed, the defendant likely will not yet have important documents from the trial court, such as transcripts of key proceedings, see, e.g., Idaho Rules App. Proc. 19 and 25; Fed. Rule App. Proc. 10(b), and may well be in custody, making communication with counsel difficult, see Peguero v. United States, 526 U.S. 23, 26 (1999). And because some defendants receive new counsel for their appeals, the lawyer responsible for deciding which appellate claims to raise may not yet even be involved in the case. Filing requirements reflect that claims are, accordingly, likely to be ill defined or unknown at this stage. In the federal system, for example, a notice of appeal need only identify who is appealing; what “judgment, order, or part thereof” is being appealed; and “the court to which the appeal is taken.” Fed. Rule App. Proc. 3(c)(1). Generally speaking, state requirements are similarly nonsubstantive.[7] A notice of appeal also fits within a broader division of labor between defendants and their attorneys. While “the accused has the ultimate authority” to decide whether to “take an appeal,” the choice of what specific arguments to make within that appeal belongs to appellate counsel. Jones v. Barnes, 463 U.S. 745, 751 (1983); see also McCoy v. Louisiana, 584 U. S. ___, ___ (2018) (slip op., at 6). In other words, filing a notice of appeal is, generally speaking, a simple, nonsubstantive act that is within the defendant’s prerogative. C With that context in mind, we turn to the precise legal issues here. As an initial matter, we note that Garza’s attorney rendered deficient performance by not filing the notice of appeal in light of Garza’s clear requests. As this Court explained in Flores-Ortega: “We have long held that a lawyer who disregards specific instructions from the defendant to file a notice of appeal acts in a manner that is professionally unreasonable. This is so because a defendant who instructs counsel to initiate an appeal reasonably relies upon counsel to file the necessary notice. Counsel’s failure to do so cannot be considered a strategic decision; filing a notice of appeal is a purely ministerial task, and the failure to file reflects inattention to the defendant’s wishes.” 528 U. S., at 477 (citations omitted); see also id., at 478. Idaho maintains that the risk of breaching the defendant’s plea agreement renders counsel’s choice to override the defendant’s instructions a strategic one. See Strickland, 466 U. S., at 690–691 (“[S]trategic choices made after thorough investigation of law and facts relevant to plausible options are virtually unchallengeable . . . ”). That is not so. While we do not address what constitutes a defendant’s breach of an appeal waiver or any responsibility counsel may have to discuss the potential consequences of such a breach, it should be clear from the foregoing that simply filing a notice of appeal does not necessarily breach a plea agreement, given the possibility that the defendant will end up raising claims beyond the waiver’s scope. And in any event, the bare decision whether to appeal is ultimately the defendant’s, not counsel’s, to make.[8] See McCoy, 584 U. S., at ___ (slip op., at 6); Barnes, 463 U. S., at 751. Where, as here, a defendant has expressly requested an appeal, counsel performs deficiently by disregarding the defendant’s instructions.[9] D We now address the crux of this case: whether Flores-Ortega’s presumption of prejudice applies despite an appeal waiver. The holding, principles, and facts of Flores-Ortega show why that presumption applies equally here. With regard to prejudice, Flores-Ortega held that, to succeed in an ineffective-assistance claim in this context, a defendant need make only one showing: “that, but for counsel’s deficient failure to consult with him about an appeal, he would have timely appealed.” 528 U. S., at 484. So long as a defendant can show that “counsel’s constitutionally deficient performance deprive[d him] of an appeal that he otherwise would have taken,” courts are to “presum[e] prejudice with no further showing from the defendant of the merits of his underlying claims.” Ibid. Because there is no dispute here that Garza wished to appeal, see supra, at 2, a direct application of Flores-Ortega’s language resolves this case. See 528 U. S., at 484. Flores-Ortega’s reasoning shows why an appeal waiver does not complicate this straightforward application. That case, like this one, involves a lawyer who forfeited an appellate proceeding by failing to file a notice of appeal. Id., at 473–475. As the Court explained, given that past precedents call for a presumption of prejudice whenever “ ‘the accused is denied counsel at a critical stage,’ ” it makes even greater sense to presume prejudice when counsel’s deficiency forfeits an “appellate proceeding altogether.” Id., at 483. After all, there is no disciplined way to “accord any ‘presumption of reliability’. . . to judicial proceedings that never took place.” Ibid. (quoting Smith v. Robbins, 528 U.S. 259, 286 (2000)). That rationale applies just as well here because, as discussed supra, at 4–6, Garza retained a right to appeal at least some issues despite the waivers he signed.[10] In other words, Garza had a right to a proceeding, and he was denied that proceeding altogether as a result of counsel’s deficient performance. That Garza surrendered many claims by signing his appeal waivers does not change things. First, this Court has made clear that when deficient counsel causes the loss of an entire proceeding, it will not bend the presumption-of-prejudice rule simply because a particular defendant seems to have had poor prospects. See, e.g., Jae Lee v. United States, 582 U. S. ___, ___ (2017) (slip op., at 9). We hew to that principle again here. Second, while the defendant in Flores-Ortega did not sign an appeal waiver, he did plead guilty, and—as the Court pointed out—“a guilty plea reduces the scope of potentially appealable issues” on its own. See 528 U. S., at 480. In other words, with regard to the defendant’s appellate prospects, Flores-Ortega presented at most a difference of degree, not kind, and prescribed a presumption of prejudice regardless of how many appellate claims were foreclosed. See id., at 484. We do no different today. Instead, we reaffirm that, “when counsel’s constitutionally deficient performance deprives a defendant of an appeal that he otherwise would have taken, the defendant has made out a successful ineffective assistance of counsel claim entitling him to an appeal,” with no need for a “further showing” of his claims’ merit, ibid., regardless of whether the defendant has signed an appeal waiver. III Flores-Ortega states, in one sentence, that the loss of the “entire [appellate] proceeding itself, which a defendant wanted at the time and to which he had a right, . . . demands a presumption of prejudice.” Id., at 483. Idaho and the U. S. Government, participating as an amicus on Idaho’s behalf, seize on this language, asserting that Garza never “had a right” to his appeal and thus that any deficient performance by counsel could not have caused the loss of any such appeal. See Brief for Respondent 11, 23–26; Brief for United States as Amicus Curiae 7, 13, 21–22. These arguments miss the point. Garza did retain a right to his appeal; he simply had fewer possible claims than some other appellants. Especially because so much is unknown at the notice-of-appeal stage, see supra, at 6–7, it is wholly speculative to say that counsel’s deficiency forfeits no proceeding to which a defendant like Garza has a right.[11] The Government also takes its causation argument one step further. Arguing that, in the appeal-waiver context, “a generalized request that an attorney file an appeal . . . is not enough to show that appellate merits review would have followed,” Brief for United States as Amicus Curiae 22, the Government proposes a rule that would require a defendant to show—on a “case-specific” basis, id., at 23—either (1) “that he in fact requested, or at least expressed interest in, an appeal on a non-waived issue,” id., at 21–22, or alternatively (2) “ ‘that there were nonfrivolous grounds for appeal’ despite the waiver,” id., at 22 (quoting Flores-Ortega, 528 U. S., at 485). We decline this suggestion, because it cannot be squared with our precedent and would likely prove both unfair and inefficient in practice. This Court has already rejected attempts to condition the restoration of a defendant’s appellate rights forfeited by ineffective counsel on proof that the defendant’s appeal had merit. In Flores-Ortega, the Court explained that prejudice should be presumed “with no further showing from the defendant of the merits of his underlying claims.” Id., at 484; see also id., at 486. In Rodriquez v. United States, 395 U.S. 327 (1969), similarly, the Court rejected a rule that required a defendant whose appeal had been forfeited by counsel “to specify the points he would raise were his right to appeal reinstated.” Id., at 330. So too here. Moreover, while it is the defendant’s prerogative whether to appeal, it is not the defendant’s role to decide what arguments to press. See Barnes, 463 U. S., at 751, 754. That makes it especially improper to impose that role upon the defendant simply because his opportunity to appeal was relinquished by deficient counsel. “Those whose right to appeal has been frustrated should be treated exactly like any other appellants; they should not be given an additional hurdle to clear just because their rights were violated at some earlier stage in the proceedings.” Rodriquez, 395 U. S., at 330. We accordingly decline to place a pleading barrier between a defendant and an opportunity to appeal that he never should have lost. Meanwhile, the Government’s assumption that unwaived claims can reliably be distinguished from waived claims through case-by-case postconviction review is dubious. There is no right to counsel in postconviction proceedings, see Pennsylvania v. Finley, 481 U.S. 551, 555 (1987), and most applicants proceed pro se.[12] That means that the Government effectively puts its faith in asking “an indigent, perhaps pro se, defendant to demonstrate that his hypothetical appeal might have had merit before any advocate has ever reviewed the record in his case in search of potentially meritorious grounds for appeal,” Flores-Ortega, 528 U. S., at 486. We have already explained why this would be “unfair” and ill advised. See ibid.; see also Rodriquez, 395 U. S., at 330. Compounding the trouble, defendants would be asked to make these showings in the face of the heightened standards and related hurdles that attend many postconviction proceedings. See, e.g., 28 U. S. C. §§2254, 2255; see also Brief for Idaho Association of Criminal Defense Lawyers et al. as Amici Curiae 22–25. The Government’s proposal is also unworkable. For one, it would be difficult and time consuming for a postconviction court to determine—perhaps years later—what appellate claims a defendant was contemplating at the time of conviction.[13] Moreover, because most postconviction petitioners will be pro se, courts would regularly have to parse both (1) what claims a pro se defendant seeks to raise and (2) whether each plausibly invoked claim is subject to the defendant’s appeal waiver (which can be complex, see supra, at 4–6), all without the assistance of counseled briefing. We are not persuaded that this would be a more efficient or trustworthy process than the one we reaffirm today. The more administrable and workable rule, rather, is the one compelled by our precedent: When counsel’s deficient performance forfeits an appeal that a defendant otherwise would have taken, the defendant gets a new opportunity to appeal. That is the rule already in use in 8 of the 10 Federal Circuits to have considered the question, see supra, at 3, and n. 3, and neither Idaho nor its amici have pointed us to any evidence that it has proved unmanageable there.[14] That rule does no more than restore the status quo that existed before counsel’s deficient performance forfeited the appeal, and it allows an appellate court to consider the appeal as that court otherwise would have done—on direct review, and assisted by counsel’s briefing. IV We hold today that the presumption of prejudice recognized in Flores-Ortega applies regardless of whether a defendant has signed an appeal waiver. This ruling follows squarely from Flores-Ortega and from the fact that even the broadest appeal waiver does not deprive a defendant of all appellate claims. Accordingly where, as here, an attorney performed deficiently in failing to file a notice of appeal despite the defendant’s express instructions, prejudice is presumed “with no further showing from the defendant of the merits of his underlying claims.” See Flores-Ortega, 528 U. S., at 484. The judgment of the Supreme Court of Idaho is therefore reversed, and the case is remanded for further proceedings not inconsistent with this opinion. It is so ordered. Notes 1 The record suggests that Garza may have been confused as to whether he had waived his appellate rights in the first place. See Record 97 (answering “No” on a court advisory form asking whether Garza had “waived [his] right to appeal [his] judgment of conviction and sentence as part of [his] plea agreement”); see also id., at 118, 121, 132 (showing that Garza’s sentencing judge and judgments of conviction provided, despite the appeal waiver, generalized notice of a “right to appeal”). Because our ruling does not turn on these facts, we do not address them further. 2 Garza’s affidavit states that he wished to argue, at least in part, that he “was persuaded to plead guilty by [the] prosecuting attorney and [his] counsel which was not voluntarily [sic].” Id., at 210. 3 Compare Campbell v. United States, 686 F.3d 353, 359 (CA6 2012); Watson v. United States, 493 F.3d 960, 964 (CA8 2007); United States v. Poindexter, 492 F.3d 263, 273 (CA4 2007); United States v. Tapp, 491 F.3d 263, 266 (CA5 2007); Campusano v. United States, 442 F.3d 770, 775 (CA2 2006); Gomez-Diaz v. United States, 433 F.3d 788, 791–794 (CA11 2005); United States v. Sandoval-Lopez, 409 F.3d 1193, 1195–1199 (CA9 2005); United States v. Garrett, 402 F. 3d 1262, 1266–1267 (CA10 2005), with Nunez v. United States, 546 F.3d 450, 455 (CA7 2008); United States v. Mabry, 536 F.3d 231, 241 (CA3 2008). At least two state courts have declined to apply Flores-Ortega in the face of appeal waivers. See Buettner v. State, 382 Mont. 410, 363 P.3d 1147 (2015) (Table); Stewart v. United States, 37 A.3d 870, 877 (D. C. 2012); see also Kargus v. State, 284 Kan. 908, 922, 928, 169 P.3d 307, 316, 320 (2007). 4 While this Court has never recognized a “constitutional right to an appeal,” it has “held that if an appeal is open to those who can pay for it, an appeal must be provided for an indigent.” Jones v. Barnes, 463 U.S. 745, 751 (1983); see also Douglas v. California, 372 U.S. 353 (1963); Griffin v. Illinois, 351 U.S. 12, 18 (1956) (plurality opinion). Today, criminal defendants in nearly all States have a right to appeal either by statute or by court rule. See generally Robertson, The Right To Appeal, 91 N. C. L. Rev. 1219, 1222, and n. 8 (2013). Criminal defendants in federal court have appellate rights under 18 U. S. C. §3742(a) and 28 U. S. C. §1291. 5 See generally Brief for Idaho Association of Criminal Defense Lawyers et al. as Amici Curiae 6–10 (collecting examples of appeal waivers that allowed challenges to the defendant’s sentence or conviction or allowed claims based on prosecutorial misconduct or changes in law). 6 See, e.g., United States v. Brown, 892 F.3d 385, 394 (CADC 2018) (“Like all other courts of appeals, our circuit holds that a defendant ‘may waive his right to appeal his sentence as long as his decision is knowing, intelligent, and voluntary’ ”); Spann v. State, 704 N.W.2d 486, 491 (Minn. 2005) (“Jurisdictions allowing a defendant to waive his or her right to appeal a conviction require that the waiver be made ‘intelligently, voluntarily, and with an understanding of the consequences’ ”). Lower courts have also applied exceptions for other kinds of claims, including “claims that a sentence is based on race discrimination, exceeds the statutory maximum authorized, or is the product of ineffective assistance of counsel.” King & O’Neill, Appeal Waivers and the Future of Sentencing Policy, 55 Duke L. J. 209, 224 (2005) (collecting federal cases); see also, e.g., United States v. Puentes-Hurtado, 794 F.3d 1278, 1284 (CA11 2015) (“[A]ppellate review is also permitted when a defendant claims that the government breached the very plea agreement which purports to bar him from appealing or collaterally attacking his conviction and sentence”); State v. Dye, 291 Neb. 989, 999, 870 N.W.2d 628, 634 (2015) (holding that appeal waivers are subject to a “miscarriage of justice” exception). We make no statement today on what particular exceptions may be required. 7 E.g., Miss. Rule Crim. Proc. 29.1(b) (2017); Ohio Rule App. Proc. 3(D) (Lexis 2017). While Idaho requires a notice of appeal to “contain substantially . . . [a] preliminary statement of the issues on appeal which the appellant then intends to assert in the appeal,” the Rule in question also makes clear that “any such list of issues on appeal shall not prevent the appellant from asserting other issues on appeal.” Idaho Rule App. Proc. 17(f ). 8 That does not mean, of course, that appellate counsel must then make unsupportable arguments. After an appeal has been preserved and counsel has reviewed the case, counsel may always, in keeping with longstanding precedent, “advise the court and request permission to withdraw,” while filing “a brief referring to anything in the record that might arguably support the appeal.” Anders v. California, 386 U.S. 738, 744 (1967). The existence of this procedure reinforces that a defendant’s appellate rights should not hinge “on appointed counsel’s bare assertion that he or she is of the opinion that there is no merit to the appeal.” Penson v. Ohio, 488 U.S. 75, 80 (1988). 9 We leave undisturbed today Flores-Ortega’s separate discussion of how to approach situations in which a defendant’s wishes are less clear. See 528 U. S., at 478–481. 10 Or the State might not have invoked the waiver at all. E.g., United States v. Archie, 771 F.3d 217, 223, n. 2 (CA4 2014); State v. Rendon, 2012 WL 9492805, *1, n. 1 (Idaho Ct. App., May 11, 2012). 11 The possibility that an appellate court confronted with a waived claim (and a motion to enforce the waiver) would technically “dismiss the appeal without reaching the merits,” see Brief for United States as Amicus Curiae 17; see also Brief for Respondent 26, does not alter this conclusion. Whatever the label, the defendant loses the opportunity to raise any appellate claims at all—including those that would, or at least could, be heard on the merits. 12 For example, researchers have found that over 90% of noncapital federal habeas petitioners proceed without counsel. See N. King et al., Final Technical Report: Habeas Litigation in U. S. District Courts 23 (2007). 13 To the extent relief would turn on what precisely a defendant said to counsel regarding specific claims, moreover, Garza rightly points out the serious risk of “causing indigent defendants to forfeit their rights simply because they did not know what words to use.” Reply Brief 17. 14 It is, of course, inevitable that some defendants under this rule will seek to raise issues that are within the scope of their appeal waivers. We are confident that courts can continue to deal efficiently with such cases via summary dispositions and the procedures outlined in Anders. See 386 U. S., at 744; n. 9, supra.

588.US.2018_17-6086

Congress has sought, for the past quarter century, to combat sex crimes and crimes against children through sex-offender registration schemes. The Sex Offender Registration and Notification Act (SORNA) makes more “uniform and effective” the prior “patchwork” of reg- istration systems. Reynolds v. United States, 565 U.S. 432, 435. To that end, it requires a broader range of sex offenders to register and backs up those requirements with criminal penalties. Section 20913 elaborates the “[i]nitial registration” requirements for sex offenders. 34 U. S. C. §§20913(b), (d). Subsection (b) sets out the general rule: An offender must register “before completing a sentence of imprisonment with respect to the offense giving rise to the registration requirement.” §20913(b). Subsection (d) addresses the “[i]nitial registration of sex offenders unable to comply with subsection (b).” The provision states that, for individuals convicted of a sex offense before SORNA’s enactment (“pre-Act offenders”), the Attorney General “shall have the authority” to “specify the applicability” of SORNA’s registration requirements and “to prescribe rules for [their] registration.” §20913(d). Under that delegated authority, the Attorney General issued a rule specifying that SORNA’s registration requirements apply in full to pre-Act offenders. Petitioner Herman Gundy, a pre-Act offender, was convicted of failing to register. Both the District Court and the Second Circuit rejected his claim that Congress unconstitutionally delegated legislative power when it authorized the Attorney General to “specify the applicability” of SORNA’s registration requirements to pre-Act offenders. Held: The judgment is affirmed. 695 Fed. Appx. 639, affirmed. Justice Kagan, joined by Justice Ginsburg, Justice Breyer, and Justice Sotomayor, concluded that §20913(d) does not violate the nondelegation doctrine. Pp. 4–18. (a) Article I of the Constitution provides that “[a]ll legislative Powers herein granted shall be vested in a Congress of the United States.” §1. Based on that provision, this Court explained early on that Congress may not transfer to another branch “powers which are strictly and exclusively legislative.” Wayman v. Southard, 10 Wheat. 1, 42–43. But Congress may confer substantial discretion on executive agencies to implement and enforce the laws. Accordingly, the Court has held, time and time again, that a statutory delegation is constitutional as long as Congress “ ‘lay[s] down by legislative act an intelligible principle to which the person or body authorized to [exercise that authority] is directed to conform.’ ” Mistretta v. United States, 488 U.S. 361, 372. Given that standard, a nondelegation inquiry always begins (and often almost ends) with statutory interpretation. Only after a court has determined a challenged statute’s meaning can it decide whether the law sufficiently guides executive discretion to accord with Article I. Pp. 4–6. (b) This Court has already interpreted §20913(d) to require the Attorney General to apply SORNA to all pre-Act offenders as soon as feasible. In Reynolds v. United States, 565 U.S. 432, the Court held that SORNA’s registration requirements did not apply of their own force to pre-Act offenders. But in doing so, it made clear how far SORNA limited the Attorney General’s authority and thereby effectively resolved this case. The Court started from the premise that Congress meant for SORNA’s registration requirements to apply to pre-Act offenders, based on the Act’s statutory purpose, its definition of sex offender, and its history. But the Court found that Congress had conditioned pre-Act offenders’ duty to register on a prior ruling from the Attorney General because “instantaneous registration” of pre-Act offenders “might not prove feasible.” Id., at 440–441. SORNA, the majority explained, created a “practical problem[ ]” because it would require “newly registering or reregistering a large number of pre-Act offenders.” Id., at 440. In addition, many pre-Act offenders were already out of prison and could not comply with the requirement that they register before completing their sentences. Congress therefore “[a]sk[ed] the Department of Justice, charged with responsibility for implementation, to examine [the issues] and to apply the new registration requirements accordingly.” Id., at 441. On that understanding, the Attorney General’s role under §20913(d) was important but limited: It was to apply SORNA to pre-Act offenders as soon as he thought it feasible to do so. Pp. 6–10. (c) Gundy claims that §20913(d) empowers the Attorney General to do whatever he wants as to pre-Act offenders, including exempting them from registration forever. He bases that argument on the first half of §20913(d), isolated from everything else. But this Court has long refused to construe words “in a vacuum,” as Gundy attempts. Davis v. Michigan Dept. of Treasury, 489 U.S. 803, 809. Rather, the Court interprets statutory provisions—including delegations—by reading the text in “context” and in light of the statutory “purpose.” National Broadcasting Co. v. United States, 319 U.S. 190, 214, 216. Applying that approach here, it is clear that §20913(d) requires the Attorney General to register pre-Act offenders as soon as feasible. In SORNA’s statement of purpose, Congress announced that “to protect the public,” it was “establish[ing] a comprehensive national system for the registration” of “sex offenders.” §20901. The term “comprehensive” means “all-encompassing” or “sweeping.” That description could not fit the system SORNA created if the Attorney General could decline, for any reason or no reason at all, to apply SORNA to all pre-Act offenders. The Act’s definition of “sex offender” makes the same point. Under that definition, a “sex offender” is “an individual who was convicted of a sex offense.” §20911(1). Congress’s use of the past tense shows that SORNA was not merely forward-looking and confirms that the delegation allows only temporary exclusions. The Act’s legislative history backs that all up, by showing that the need to register pre-Act offenders was front and center in Congress’s thinking. The text and title of §20913(d) then pinpoint one of the practical problems discussed above: At the moment of SORNA’s enactment, many pre-Act offenders were “unable to comply” with the Act’s initial registration requirements. §20913(d). In identifying that issue, §20913(d) itself reveals the nature of the delegation to the Attorney General. It was to give him the time needed (if any) to address the various implementation issues involved in getting pre-Act offenders into the registration system. Thus, contrary to Gundy, “specify the applicability” does not mean “specify whether to apply SORNA” to pre-Act offenders at all. The phrase instead means “specify how to apply SORNA” to pre-Act offenders if transitional difficulties require some delay. And no Attorney General has used §20913(d) in any more expansive way. Pp. 10–15. (d) Section 20913(d)’s delegation therefore falls well within constitutional bounds. As noted, a delegation is constitutional so long as Congress sets out an intelligible principle to guide the delegee’s exercise of authority. The standards for that principle are not demanding. See Whitman v. American Trucking Assns., Inc., 531 U.S. 457, 474–475. Only twice in this country’s history has the Court found a delegation excessive, in each case because “Congress had failed to articulate any policy or standard” to confine discretion. Mistretta, 488 U. S., at 373, n. 3; see A. L. A. Schechter Poultry Corp. v. United States, 295 U.S. 495; Panama Refining Co. v. Ryan, 293 U.S. 388. By contrast, the Court has over and over upheld even very broad delegations. See, e.g., National Broadcasting Co. v. United States, 319 U.S. 190. In that context, the delegation in SORNA easily passes muster. The authority §20913(d) confers, as compared to the delegations the Court has upheld in the past, is distinctly small bore. Indeed, if SORNA’s delegation is unconstitutional, then most of Government is unconstitutional—dependent as Congress is on the need to give discretion to executive officials to implement its programs. Pp. 15–18. Justice Alito concluded that he cannot say that the statute at issue lacks an adequately discernable standard under the nondelegation approach the Court has taken for the past 84 years, but would reconsider that approach in an appropriate case. P. 1. Kagan, J., announced the judgment of the Court and delivered an opinion, in which Ginsburg, Breyer, and Sotomayor, JJ., joined. Alito, J., filed an opinion concurring in the judgment. Gorsuch, J., filed a dissenting opinion, in which Roberts, C. J., and Thomas, J., joined. Kavanaugh, J., took no part in the consideration or decision of the case.

in which Justice Ginsburg, Justice Breyer, and Justice Sotomayor join. The nondelegation doctrine bars Congress from transferring its legislative power to another branch of Government. This case requires us to decide whether 34 U. S. C. §20913(d), enacted as part of the Sex Offender Registration and Notification Act (SORNA), violates that doctrine. We hold it does not. Under §20913(d), the Attorney General must apply SORNA’s registration requirements as soon as feasible to offenders convicted before the statute’s enactment. That delegation easily passes constitutional muster. I Congress has sought, for the past quarter century, to combat sex crimes and crimes against children through sex-offender registration schemes. In 1994, Congress first conditioned certain federal funds on States’ adoption of registration laws meeting prescribed minimum standards. See Jacob Wetterling Crimes Against Children and Sexually Violent Offender Registration Act, §170101, 108 Stat. 2038, 42 U. S. C. §14071 et seq. (1994 ed.). Two years later, Congress strengthened those standards, most notably by insisting that States inform local communities of registrants’ addresses. See Megan’s Law, §2, 110Stat. 1345, note following 42 U. S. C. §13701 (1994 ed., Supp. II). By that time, every State and the District of Columbia had enacted a sex-offender registration law. But the state statutes varied along many dimensions, and Congress came to realize that their “loopholes and deficiencies” had allowed over 100,000 sex offenders (about 20% of the total) to escape registration. See H. R. Rep. No. 109–218, pt. 1, pp. 20, 23–24, 26 (2005) (referring to those sex offenders as “missing” or “lost”). In 2006, to address those failings, Congress enacted SORNA. See 120Stat. 590, 34 U. S. C. §20901 et seq. SORNA makes “more uniform and effective” the prior “patchwork” of sex-offender registration systems. Reynolds v. United States, 565 U.S. 432, 435 (2012). The Act’s express “purpose” is “to protect the public from sex offenders and offenders against children” by “establish[ing] a comprehensive national system for [their] registration.” §20901. To that end, SORNA covers more sex offenders, and imposes more onerous registration requirements, than most States had before. The Act also backs up those requirements with new criminal penalties. Any person required to register under SORNA who knowingly fails to do so (and who travels in interstate commerce) may be imprisoned for up to ten years. See 18 U. S. C. §2250(a). The basic registration scheme works as follows. A “sex offender” is defined as “an individual who was convicted of” specified criminal offenses: all offenses “involving a sexual act or sexual contact” and additional offenses “against a minor.” 34 U. S. C. §§20911(1), (5)(A), (7). Such an individual must register—provide his name, address, and certain other information—in every State where he resides, works, or studies. See §§20913(a), 20914. And he must keep the registration current, and periodically report in person to a law enforcement office, for a period of between fifteen years and life (depending on the severity of his crime and his history of recidivism). See §§20915, 20918. Section 20913—the disputed provision here—elaborates the “[i]nitial registration” requirements for sex offenders. §§20913(b), (d). Subsection (b) sets out the general rule: An offender must register “before completing a sentence of imprisonment with respect to the offense giving rise to the registration requirement” (or, if the offender is not sentenced to prison, “not later than [three] business days after being sentenced”). Two provisions down, subsection (d) addresses (in its title’s words) the “[i]nitial registration of sex offenders unable to comply with subsection (b).” The provision states: “The Attorney General shall have the authority to specify the applicability of the requirements of this subchapter to sex offenders convicted before the enactment of this chapter . . . and to prescribe rules for the registration of any such sex offenders and for other categories of sex offenders who are unable to comply with subsection (b).” Subsection (d), in other words, focuses on individuals convicted of a sex offense before SORNA’s enactment—a group we will call pre-Act offenders. Many of these individuals were unregistered at the time of SORNA’s enactment, either because pre-existing law did not cover them or because they had successfully evaded that law (so were “lost” to the system). See supra, at 2. And of those potential new registrants, many or most could not comply with subsection (b)’s registration rule because they had already completed their prison sentences. For the entire group of pre-Act offenders, once again, the Attorney General “shall have the authority” to “specify the applicability” of SORNA’s registration requirements and “to prescribe rules for [their] registration.” Under that delegated authority, the Attorney General issued an interim rule in February 2007, specifying that SORNA’s registration requirements apply in full to “sex offenders convicted of the offense for which registration is required prior to the enactment of that Act.” 72 Fed. Reg. 8897. The final rule, issued in December 2010, reiterated that SORNA applies to all pre-Act offenders. 75 Fed. Reg. 81850. That rule has remained the same to this day. Petitioner Herman Gundy is a pre-Act offender. The year before SORNA’s enactment, he pleaded guilty under Maryland law for sexually assaulting a minor. After his release from prison in 2012, Gundy came to live in New York. But he never registered there as a sex offender. A few years later, he was convicted for failing to register, in violation of §2250. He argued below (among other things) that Congress unconstitutionally delegated legislative power when it authorized the Attorney General to “specify the applicability” of SORNA’s registration requirements to pre-Act offenders. §20913(d). The District Court and Court of Appeals for the Second Circuit rejected that claim, see 695 Fed. Appx. 639 (2017), as had every other court (including eleven Courts of Appeals) to consider the issue. We nonetheless granted certiorari. 583 U. S. __ (2018). Today, we join the consensus and affirm. II Article I of the Constitution provides that “[a]ll legislative Powers herein granted shall be vested in a Congress of the United States.” §1. Accompanying that assignment of power to Congress is a bar on its further delegation. Congress, this Court explained early on, may not transfer to another branch “powers which are strictly and exclusively legislative.” Wayman v. Southard, 10 Wheat. 1, 42–43 (1825). But the Constitution does not “deny[ ] to the Congress the necessary resources of flexibility and practicality [that enable it] to perform its function[s].” Yakus v. United States, 321 U.S. 414, 425 (1944) (internal quotation marks omitted). Congress may “obtain[ ] the assistance of its coordinate Branches”—and in particular, may confer substantial discretion on executive agencies to implement and enforce the laws. Mistretta v. United States, 488 U.S. 361, 372 (1989). “[I]n our increasingly complex society, replete with ever changing and more technical problems,” this Court has understood that “Congress simply cannot do its job absent an ability to delegate power under broad general directives.” Ibid. So we have held, time and again, that a statutory delegation is constitutional as long as Congress “lay[s] down by legislative act an intelligible principle to which the person or body authorized to [exercise the delegated authority] is directed to conform.” Ibid. (quoting J. W. Hampton, Jr., & Co. v. United States, 276 U.S. 394, 409 (1928); brackets in original). Given that standard, a nondelegation inquiry always begins (and often almost ends) with statutory interpretation. The constitutional question is whether Congress has supplied an intelligible principle to guide the delegee’s use of discretion. So the answer requires construing the challenged statute to figure out what task it delegates and what instructions it provides. See, e.g., Whitman v. American Trucking Assns., Inc., 531 U.S. 457, 473 (2001) (construing the text of a delegation to place constitutionally adequate “limits on the EPA’s discretion”); American Power & Light Co. v. SEC, 329 U.S. 90, 104–105 (1946) (interpreting a statutory delegation, in light of its “purpose[,] factual background[, and] context,” to provide sufficiently “definite” standards). Only after a court has determined a challenged statute’s meaning can it decide whether the law sufficiently guides executive discretion to accord with Article I. And indeed, once a court interprets the statute, it may find that the constitutional question all but answers itself. That is the case here, because §20913(d) does not give the Attorney General anything like the “unguided” and “unchecked” authority that Gundy says. Brief for Petitioner 37, 45. The provision, in Gundy’s view, “grants the Attorney General plenary power to determine SORNA’s applicability to pre-Act offenders—to require them to register, or not, as she sees fit, and to change her policy for any reason and at any time.” Id., at 42. If that were so, we would face a nondelegation question. But it is not. This Court has already interpreted §20913(d) to say something different—to require the Attorney General to apply SORNA to all pre-Act offenders as soon as feasible. See Reynolds, 565 U. S., at 442–443. And revisiting that issue yet more fully today, we reach the same conclusion. The text, considered alongside its context, purpose, and history, makes clear that the Attorney General’s discretion extends only to considering and addressing feasibility issues. Given that statutory meaning, Gundy’s constitutional claim must fail. Section 20913(d)’s delegation falls well within permissible bounds. A This is not the first time this Court has had to interpret §20913(d). In Reynolds, the Court considered whether SORNA’s registration requirements applied of their own force to pre-Act offenders or instead applied only once the Attorney General said they did. We read the statute as adopting the latter approach. But even as we did so, we made clear how far SORNA limited the Attorney General’s authority. And in that way, we effectively resolved the case now before us. Everything in Reynolds started from the premise that Congress meant for SORNA’s registration requirements to apply to pre-Act offenders. The majority recounted SORNA’s “basic statutory purpose,” found in its text, as follows: “the ‘establish[ment of] a comprehensive national system for the registration of [sex] offenders’ that includes offenders who committed their offenses before the Act became law.” 565 U. S., at 442 (quoting §20901; emphasis and alterations in original; citation omitted). That purpose, the majority further noted, informed SORNA’s “broad[ ]” definition of “sex offender,” which “include[s] any ‘individual who was convicted of a sex offense.’ ” Id., at 442 (quoting §20911(1); emphasis added). And those two provisions were at one with “[t]he Act’s history.” Id., at 442. Quoting statements from both the House and the Senate about the sex offenders then “lost” to the system, Reynolds explained that the Act’s “supporters placed considerable importance upon the registration of pre-Act offenders.” Ibid. In recognizing all this, the majority (temporarily) bonded with the dissenting Justices, who found it obvious that SORNA was “meant to cover pre-Act offenders.” Id., at 448 (Scalia, J., dissenting). And indeed, the dissent emphasized that common ground, remarking that “the Court acknowledges” and “rightly believes” that registration of pre-Act offenders was “what the statute sought to achieve.” Id., at 448–449.[1] But if that was so, why had Congress (as the majority held) conditioned the pre-Act offenders’ duty to register on a prior “ruling from the Attorney General”? Id., at 441. The majority had a simple answer: “[I]nstantaneous registration” of pre-Act offenders “might not prove feasible,” or “[a]t least Congress might well have so thought.” Id., at 440–441, 443. Here, the majority explained that SORNA’s requirements diverged from prior state law. See id., at 440; supra, at 2. Some pre-Act offenders (as defined by SORNA) had never needed to register before; others had once had to register, but had fulfilled their old obligations. And still others (the “lost” or “missing” offenders) should have registered, but had escaped the system. As a result, SORNA created a “practical problem[ ]”: It would require “newly registering or reregistering a large number of pre-Act offenders.” Reynolds, 565 U. S., at 440 (internal quotation marks omitted). And attached to that broad feasibility concern was a more technical one. Recall that under SORNA “a sex offender must initially register before completing his ‘sentence of imprisonment.’ ” Id., at 439 (quoting §20913(b)); see supra, at 3. But many pre-Act offenders were already out of prison, so could not comply with that requirement. That inability raised questions about “how[] the new registration requirements applied to them.” 565 U. S., at 441. “Congress[’s] solution” to both those difficulties was the same: Congress “[a]sk[ed] the Department of Justice, charged with responsibility for implementation, to examine [the issues] and to apply the new registration requirements accordingly.” Ibid. On that understanding, the Attorney General’s role under §20913(d) was important but limited: It was to apply SORNA to pre-Act offenders as soon as he thought it feasible to do so. That statutory delegation, the Court explained, would “involve[] implementation delay.” Id., at 443. But no more than that. Congress had made clear in SORNA’s text that the new registration requirements would apply to pre-Act offenders. See id., at 442–445. So (the Court continued) “there was no need” for Congress to worry about the “unrealistic possibility” that “the Attorney General would refuse to apply” those requirements on some excessively broad view of his authority under §20913(d). Id., at 444–445. Reasonably read, SORNA enabled the Attorney General only to address (as appropriate) the “practical problems” involving pre-Act offenders before requiring them to register. Id., at 440. The delegation was a stopgap, and nothing more.[2] Gundy dismisses Reynolds’s relevance, but his arguments come up short. To begin, he contends that Rey- nolds spoke “tentative[ly]”—with “might[s], may[s], or could[s]”—about Congress’s reasons for enacting §20913(d). Reply Brief 11; see supra, at 7 (quoting such phrases). Gundy concludes from such constructions—which are indeed present—that the Court was “not offering a definitive reading of the statute.” Reply Brief 11. But the Court used those locutions to convey not its own uncertainty but Congress’s. The point of the opinion was that Congress had questions about how best to phase SORNA’s application to pre-Act offenders, so gave the Attorney General flexibility on timing. The “mights, mays, and coulds” were there to describe the legislative mindset responsible for §20913(d), and thus formed part of the Court’s own—yes, “definitive”—view of that provision’s meaning. Anticipating that explanation, Gundy falls back on the claim that the Court’s account of Congress’s motivations “cannot supply the intelligible principle Congress failed to enact into law.” Id., at 12 (citing Whitman, 531 U. S., at 473). But the Court in Reynolds did not invent a standard Congress omitted. Rather, the Court read the statute to contain a standard—again, that the Attorney General should apply SORNA to pre-Act offenders as soon as feasible. And as the next part of this opinion shows, in somewhat greater detail than Reynolds thought necessary, we read the statute in the same way. B Recall again the delegation provision at issue. Congress gave the Attorney General authority to “specify the applicability” of SORNA’s requirements to pre-Act offenders. §20913(d). And in the second half of the same sentence, Congress gave him authority to “prescribe rules for the registration of any such sex offenders . . . who are unable to comply with” subsection (b)’s initial registration requirement. Ibid. What does the delegation in §20913(d) allow the Attorney General to do? The different answers on offer here reflect competing views of statutory interpretation. As noted above, Gundy urges us to read §20913(d) to empower the Attorney General to do whatever he wants as to pre-Act offenders: He may make them all register immediately or he may exempt them from registration forever (or he may do anything in between). See Brief for Petitioner 41–42; supra, at 6. Gundy bases that argument on the first half of §20913(d), isolated from everything else—from the second half of the same section, from surrounding provisions in SORNA, and from any conception of the statute’s history and purpose. Reynolds took a different approach (as does the Government here), understanding statutory interpretation as a “holistic endeavor” which determines meaning by looking not to isolated words, but to text in context, along with purpose and history. United Sav. Assn. of Tex. v. Timbers of Inwood Forest Associates, Ltd., 484 U.S. 365, 371 (1988). This Court has long refused to construe words “in a vacuum,” as Gundy attempts. Davis v. Michigan Dept. of Treasury, 489 U.S. 803, 809 (1989). “It is a fundamental canon of statutory construction that the words of a statute must be read in their context and with a view to their place in the overall statutory scheme.” National Assn. of Home Builders v. Defenders of Wildlife, 551 U.S. 644, 666 (2007) (internal quotation marks omitted); see Utility Air Regulatory Group v. EPA, 573 U.S. 302, 321 (2014) (“[R]easonable statutory interpretation must account for both the specific context in which . . . language is used and the broader context of the statute as a whole” (internal quotation marks omitted)). And beyond context and structure, the Court often looks to “history [and] purpose” to divine the meaning of language. Maracich v. Spears, 570 U.S. 48, 76 (2013) (internal quotation marks omitted). That non-blinkered brand of interpretation holds good for delegations, just as for other statutory provisions. To define the scope of delegated authority, we have looked to the text in “context” and in light of the statutory “purpose.” National Broadcasting Co. v. United States, 319 U.S. 190, 214, 216 (1943) (internal quotation marks omitted); see American Power & Light, 329 U. S., at 104 (stating that the delegation at issue “derive[d] much meaningful content from the purpose of the Act, its factual background and the statutory context”). In keeping with that method, we again do so today. So begin at the beginning, with the “[d]eclaration of purpose” that is SORNA’s first sentence. §20901. There, Congress announced (as Reynolds noted, see supra, at 6–7) that “to protect the public,” it was “establish[ing] a comprehensive national system for the registration” of “sex offenders and offenders against children.” §20901. The term “comprehensive” has a clear meaning—something that is all-encompassing or sweeping. See, e.g., Webster’s Third New International Dictionary 467 (2002) (“covering a matter under consideration completely or nearly completely”); New Oxford American Dictionary 350 (2d ed. 2005) (“complete; including all or nearly all elements or aspects of something”). That description could not fit the system SORNA created if the Attorney General could decline, for any reason or no reason at all, to apply SORNA to all pre-Act offenders. After all, for many years after SORNA’s enactment, the great majority of sex offenders in the country would be pre-Act offenders. If Gundy were right, all of those offenders could be exempt from SORNA’s registration requirements. So the mismatch between SORNA’s statement of purpose and Gundy’s view of §20913(d) is as stark as stark comes. Responding to that patent disparity, Gundy urges us to ignore SORNA’s statement of purpose because it is “located in the Act’s preface” rather than “tied” specifically to §20913(d). Brief for Petitioner 46. But the placement of such a statement within a statute makes no difference. See A. Scalia & B. Garner, Reading Law: The Interpretation of Legal Texts 220 (2012). Wherever it resides, it is “an appropriate guide” to the “meaning of the [statute’s] operative provisions.” Id., at 218. And here it makes clear that SORNA was supposed to apply to all pre-Act offenders—which precludes Gundy’s construction of §20913(d). The Act’s definition of “sex offender” (also noted in Reynolds, see supra, at 7) makes the same point. Under that definition, a “sex offender” is “an individual who was convicted of a sex offense.” §20911(1). Note the tense: “was,” not “is.” This Court has often “looked to Congress’ choice of verb tense to ascertain a statute’s temporal reach,” including when interpreting other SORNA provisions. Carr v. United States, 560 U.S. 438, 447–448 (2010) (holding that because SORNA “sets forth [its] travel requirement in the present tense,” the statute’s criminal penalties do not apply to a person whose interstate travel predated enactment); see, e.g., United States v. Wilson, 503 U.S. 329, 333 (1992); Gwaltney of Smithfield, Ltd. v. Chesapeake Bay Foundation, Inc., 484 U.S. 49, 57 (1987). Here, Congress’s use of the past tense to define the term “sex offender” shows that SORNA was not merely forward-looking. The word “is” would have taken care of all future offenders. The word “was” served to bring in the hundreds of thousands of persons previously found guilty of a sex offense, and thought to pose a current threat to the public. The tense of the “sex offender” definition thus confirms that the delegation allows only temporary exclusions, as necessary to address feasibility issues. Contra Gundy, it does not sweep so wide as to make a laughingstock of the statute’s core definition. The Act’s legislative history backs up everything said above by showing that the need to register pre-Act offenders was front and center in Congress’s thinking. (Once again, the Reynolds majority noted this history, but Justice Scalia’s dissent thought that was gilding the lily. See supra, at 7, and n. 1. He had a point, but we can’t resist.) Recall that Congress designed SORNA to address “loopholes and deficiencies” in existing registration laws. See supra, at 2. And no problem attracted greater attention than the large number of sex offenders who had slipped the system. According to the House Report, “[t]he most significant enforcement issue in the sex offender program is that over 100,000 sex offenders” are “ ‘missing,’ meaning that they have not complied with” then-current requirements. H. R. Rep. No. 109–218, at 26. There is a “strong public interest,” the Report continued, in “having [those offenders] register with current information to mitigate the risks of additional crimes against children.” Id., at 24. Senators struck a similar chord in the debates preceding SORNA’s passage, repeatedly stressing that the new provisions would capture the missing offenders. See, e.g., 152 Cong. Rec. 15338 (2006) (statement of Sen. Kyl) (“The penalties in this bill should be adequate to ensure that [the 100,000 missing offenders] register”); id., at 13050 (statement of Sen. Frist) (“Every day that we don’t have this national sex offender registry, these missing sex predators are out there somewhere”). Imagine how surprising those Members would have found Gundy’s view that they had authorized the Attorney General to exempt the missing “predators” from registering at all. With that context and background established, we may return to §20913(d). As we have noted, Gundy makes his stand there (and there only), insisting that the lonesome phrase “specify the applicability” ends this case. See supra, at 10. But in so doing, Gundy ignores even the rest of the section that phrase is in. Both the title and the remaining text of that section pinpoint one of the “practical problems” discussed above: At the moment of SORNA’s enactment, many pre-Act offenders were “unable to comply” with the Act’s initial registration requirements. §20913(d); Reynolds, 565 U. S., at 440; see supra, at 8. That was because, once again, the requirements assumed that offenders would be in prison, whereas many pre-Act offenders were on the streets. In identifying that issue, §20913(d) itself reveals the nature of the delegation to the Attorney General. It was to give him the time needed (if any) to address the various implementation issues involved in getting pre-Act offenders into the registration system. “Specify the applicability” thus does not mean “specify whether to apply SORNA” to pre-Act offenders at all, even though everything else in the Act commands their coverage. The phrase instead means “specify how to apply SORNA” to pre-Act offenders if transitional difficulties require some delay. In that way, the whole of §20913(d) joins the rest of SORNA in giving the Attorney General only time-limited latitude to excuse pre-Act offenders from the statute’s requirements. Under the law, he had to order their registration as soon as feasible. And no Attorney General has used (or, apparently, thought to use) §20913(d) in any more expansive way. To the contrary. Within a year of SORNA’s enactment (217 days, to be precise), the Attorney General determined that SORNA would apply immediately to pre-Act offenders. See Interim Rule, 72 Fed. Reg. 8897; supra, at 4. That rule has remained in force ever since (save for a technical change to one of the rule’s illustrative examples). See Final Rule, 75 Fed. Reg. 81850.[3] And at oral argument here, the Solicitor General’s office—rarely in a hurry to agree to limits on the Government’s authority—acknowledged that §20913(d) does not allow the Attorney General to excuse a pre-Act offender from registering, except for reasons of “feasibility.” Tr. of Oral Arg. 41–42. We thus end up, on close inspection of the statutory scheme, exactly where Reynolds left us. The Attorney General’s authority goes to transition-period implementation issues, and no further. C Now that we have determined what §20913(d) means, we can consider whether it violates the Constitution. The question becomes: Did Congress make an impermissible delegation when it instructed the Attorney General to apply SORNA’s registration requirements to pre-Act offenders as soon as feasible? Under this Court’s long-established law, that question is easy. Its answer is no. As noted earlier, this Court has held that a delegation is constitutional so long as Congress has set out an “intelligible principle” to guide the delegee’s exercise of authority. J. W. Hampton, Jr., & Co., 276 U. S., at 409; see supra, at 5. Or in a related formulation, the Court has stated that a delegation is permissible if Congress has made clear to the delegee “the general policy” he must pursue and the “boundaries of [his] authority.” American Power & Light, 329 U. S., at 105. Those standards, the Court has made clear, are not demanding. “[W]e have ‘almost never felt qualified to second-guess Congress regarding the permissible degree of policy judgment that can be left to those executing or applying the law.’ ” Whitman, 531 U. S., at 474–475 (quoting Mistretta, 488 U. S., at 416 (Scalia, J., dissenting)). Only twice in this country’s history (and that in a single year) have we found a delegation excessive—in each case because “Congress had failed to articulate any policy or standard” to confine discretion. Mistretta, 488 U. S., at 373, n. 7 (emphasis added); see A. L. A. Schechter Poultry Corp. v. United States, 295 U.S. 495 (1935); Panama Refining Co. v. Ryan, 293 U.S. 388 (1935). By contrast, we have over and over upheld even very broad delegations. Here is a sample: We have approved delegations to various agencies to regulate in the “public interest.” See, e.g., National Broadcasting Co., 319 U. S., at 216; New York Central Securities Corp. v. United States, 287 U.S. 12, 24 (1932). We have sustained authorizations for agencies to set “fair and equitable” prices and “just and reasonable” rates. Yakus, 321 U. S., at 422, 427; FPC v. Hope Natural Gas Co., 320 U.S. 591 (1944). We more recently affirmed a delegation to an agency to issue whatever air quality standards are “requisite to protect the public health.” Whitman, 531 U. S., at 472 (quoting 42 U. S. C. §7409(b)(1)). And so forth. In that context, the delegation in SORNA easily passes muster (as all eleven circuit courts to have considered the question found, see supra, at 4). The statute conveyed Congress’s policy that the Attorney General require pre-Act offenders to register as soon as feasible. Under the law, the feasibility issues he could address were administrative—and, more specifically, transitional—in nature. Those issues arose, as Reynolds explained, from the need to “newly register[ ] or reregister[ ] ‘a large number’ of pre-Act offenders” not then in the system. 565 U. S., at 440; see supra, at 8. And they arose, more technically, from the gap between an initial registration requirement hinged on imprisonment and a set of pre-Act offenders long since released. See 565 U. S., at 441; see supra, at 8. Even for those limited matters, the Act informed the Attorney General that he did not have forever to work things out. By stating its demand for a “comprehensive” registration system and by defining the “sex offenders” required to register to include pre-Act offenders, Congress conveyed that the Attorney General had only temporary authority. Or again, in the words of Reynolds, that he could prevent “instantaneous registration” and impose some “implementation delay.” 565 U. S., at 443. That statutory authority, as compared to the delegations we have upheld in the past, is distinctly small-bore. It falls well within constitutional bounds.[4] Indeed, if SORNA’s delegation is unconstitutional, then most of Government is unconstitutional—dependent as Congress is on the need to give discretion to executive officials to implement its programs. Consider again this Court’s long-time recognition: “Congress simply cannot do its job absent an ability to delegate power under broad general directives.” Mistretta, 488 U. S., at 372; see supra, at 5. Or as the dissent in that case agreed: “[S]ome judgments . . . must be left to the officers executing the law.” 488 U. S., at 415 (opinion of Scalia, J.); see Whitman, 531 U. S., at 475 (“[A] certain degree of discretion[ ] inheres in most executive” action (internal quotation marks omitted)). Among the judgments often left to executive officials are ones involving feasibility. In fact, standards of that kind are ubiquitous in the U. S. Code. See, e.g., 12 U. S. C. §1701z–2(a) (providing that the Secretary of Housing and Urban Development “shall require, to the greatest extent feasible, the employment of new and improved technologies, methods, and materials in housing construction[ ] under [HUD] programs”); 47 U. S. C. §903(d)(1) (providing that “the Secretary of Commerce shall promote efficient and cost-effective use of the spectrum to the maximum extent feasible” in “assigning frequencies for mobile radio services”). In those delegations, Congress gives its delegee the flexibility to deal with real-world constraints in carrying out his charge. So too in SORNA. It is wisdom and humility alike that this Court has always upheld such “necessities of government.” Mistretta, 488 U. S., at 416 (Scalia, J., dissenting) (internal quo- tation marks omitted); see ibid. (“Since Congress is no less endowed with common sense than we are, and better equipped to inform itself of the ‘necessities’ of government; and since the factors bearing upon those necessities are both multifarious and (in the nonpartisan sense) highly political . . . it is small wonder that we have almost never felt qualified to second-guess Congress regarding the permissible degree of policy judgment that can be left to those executing or applying the law”). We therefore affirm the judgment of the Court of Appeals. It is so ordered. Justice Kavanaugh took no part in the consideration or decision of this case. Notes 1 As to that point, the dissent criticized the majority only for basing its view in part on legislative history. 565 U. S., at 448, n. (opinion of Scalia, J.). The dissent found the majority’s excursion into history “quite superfluous” given that the “text of the Act itself makes clear that Congress sought” to ensure the registration of all pre-Act offenders. Ibid. In reaching that conclusion, the dissent relied on the Act’s express statement of purpose and its “sex offender” definition. See infra, at 11–13. 2 Once again, the dissent agreed with the Court that §20913(d) could not sensibly be read to give the Attorney General any greater power. “[I]t is simply implausible,” the dissent concluded, “that the Attorney General was given discretion to determine whether coverage of pre-Act offenders (one of the purposes of the Act) should exist.” 565 U. S., at 450 (opinion of Scalia, J.). The dissent parted ways with the Court only in interpreting §20913(d) to provide the Attorney General with even less authority. 3 Gundy tries to dispute that simple fact, but fails. He points to changes that Attorneys General have made in guidelines to States about how to satisfy SORNA’s funding conditions. See Brief for Petitioner 32–33. But those state-directed rules are independent of the only thing at issue here: the application of registration requirements to pre-Act offenders. Those requirements have been constant since the Attorney General’s initial rule, as the guidelines themselves affirm. See 73 Fed. Reg. 38046 (2008); 76 Fed. Reg. 1639 (2011). Indeed, the guidelines to States are issued not under §20913(d) at all, but under a separate delegation in §20912(b). See 73 Fed. Reg. 38030; 76 Fed. Reg. 1631. 4 Even Gundy conceded at oral argument that if the statute means what we have said, it “likely would be constitutional.” Tr. of Oral Arg. 25. That is why all of his argument is devoted to showing that it means something else.

586.US.2018_17-1229

Petitioner Helsinn Healthcare S. A. makes a treatment for chemotherapy-induced nausea and vomiting using the chemical palonosetron. While Helsinn was developing its palonosetron product, it entered into two agreements with another company granting that company the right to distribute, promote, market, and sell a 0.25 mg dose of palonosetron in the United States. The agreements required that the company keep confidential any proprietary information received under the agreements. Nearly two years later, in January 2003, Helsinn filed a provisional patent application covering a 0.25 mg dose of palonosetron. Over the next 10 years, Helsinn filed four patent applications that claimed priority to the January 2003 date. Relevant here, Helsinn filed its fourth patent application in 2013. That patent (the ’219 patent) covers a fixed dose of 0.25 mg of palonosetron in a 5 ml solution and is covered by the Leahy-Smith America Invents Act (AIA). In 2011, respondents Teva Pharmaceutical Industries, Ltd., and Teva Pharmaceuticals USA, Inc. (collectively Teva), sought approval to market a generic 0.25 mg palonosetron product. Helsinn sued Teva for infringing its patents, including the ’219 patent. Teva countered that the ’219 patent was invalid under the “on sale” provision of the AIA—which precludes a person from obtaining a patent on an invention that was “in public use, on sale, or otherwise available to the public before the effective filing date of the claimed invention,” 35 U. S. C. §102(a)(1)—because the 0.25 mg dose was “on sale” more than one year before Helsinn filed the provisional patent application in 2003. The District Court held that the AIA’s “on sale” provision did not apply because the public disclosure of the agreements did not disclose the 0.25 mg dose. The Federal Circuit reversed, holding that the sale was publicly disclosed, regardless of whether the details of the invention were publicly disclosed in the terms of the sale agreements. Held: A commercial sale to a third party who is required to keep the invention confidential may place the invention “on sale” under §102(a). The patent statute in force immediately before the AIA included an on-sale bar. This Court’s precedent interpreting that provision supports the view that a sale or offer of sale need not make an invention available to the public to constitute invalidating prior art. See, e.g., Pfaff v. Wells Electronics, Inc., 525 U.S. 55, 67. The Federal Circuit had made explicit what was implicit in this Court’s pre-AIA precedent, holding that “secret sales” could invalidate a patent. Special Devices, Inc. v. OEA, Inc., 270 F.3d 1353, 1357. Given this settled pre-AIA precedent, the Court applies the presumption that when Congress reenacted the same “on sale” language in the AIA, it adopted the earlier judicial construction of that phrase. The addition of the catchall phrase “or otherwise available to the public” is not enough of a change for the Court to conclude that Congress intended to alter the meaning of “on sale.” Paroline v. United States, 572 U.S. 434, and Federal Maritime Comm’n v. Seatrain Lines, Inc., 411 U.S. 726, distinguished. Pp. 5–9. 855 F.3d 1356, affirmed. Thomas, J., delivered the opinion for a unanimous Court.

The Leahy-Smith America Invents Act (AIA) bars a person from receiving a patent on an invention that was “in public use, on sale, or otherwise available to the public before the effective filing date of the claimed invention.” 35 U. S. C. §102(a)(1). This case requires us to decide whether the sale of an invention to a third party who is contractually obligated to keep the invention confidential places the invention “on sale” within the meaning of §102(a). More than 20 years ago, this Court determined that an invention was “on sale” within the meaning of an earlier version of §102(a) when it was “the subject of a commercial offer for sale” and “ready for patenting.” Pfaff v. Wells Electronics, Inc., 525 U.S. 55, 67 (1998). We did not further require that the sale make the details of the invention available to the public. In light of this earlier construction, we determine that the reenactment of the phrase “on sale” in the AIA did not alter this meaning. Accordingly, a commercial sale to a third party who is required to keep the invention confidential may place the invention “on sale” under the AIA. I Petitioner Helsinn Healthcare S. A. (Helsinn) is a Swiss pharmaceutical company that makes Aloxi, a drug that treats chemotherapy-induced nausea and vomiting. Helsinn acquired the right to develop palonosetron, the active ingredient in Aloxi, in 1998. In early 2000, it submitted protocols for Phase III clinical trials to the Food and Drug Administration (FDA), proposing to study a 0.25 mg and a 0.75 mg dose of palonosetron. In September 2000, Helsinn announced that it was beginning Phase III clinical trials and was seeking marketing partners for its palonosetron product. Helsinn found its marketing partner in MGI Pharma, Inc. (MGI), a Minnesota pharmaceutical company that markets and distributes drugs in the United States. Helsinn and MGI entered into two agreements: a license agreement and a supply and purchase agreement. The license agreement granted MGI the right to distribute, promote, market, and sell the 0.25 mg and 0.75 mg doses of palonosetron in the United States. In return, MGI agreed to make upfront payments to Helsinn and to pay future royalties on distribution of those doses. Under the supply and purchase agreement, MGI agreed to purchase exclusively from Helsinn any palonosetron product approved by the FDA. Helsinn in turn agreed to supply MGI however much of the approved doses it required. Both agreements included dosage information and required MGI to keep confidential any proprietary information received under the agreements. Helsinn and MGI announced the agreements in a joint press release, and MGI also reported the agreements in its Form 8–K filing with the Securities and Exchange Commission. Although the 8–K filing included redacted copies of the agreements, neither the 8–K filing nor the press releases disclosed the specific dosage formulations covered by the agreements. On January 30, 2003, nearly two years after Helsinn and MGI entered into the agreements, Helsinn filed a provisional patent application covering the 0.25 mg and 0.75 mg doses of palonosetron. Over the next 10 years, Helsinn filed four patent applications that claimed priority to the January 30, 2003, date of the provisional application. Helsinn filed its fourth patent application—the one relevant here—in May 2013, and it issued as U. S. Patent No. 8,598,219 (’219 patent). The ’219 patent covers a fixed dose of 0.25 mg of palonosetron in a 5 ml solution. By virtue of its effective date, the ’219 patent is governed by the AIA. See §101(i). Respondents Teva Pharmaceutical Industries, Ltd., and Teva Pharmaceuticals USA, Inc. (Teva), are, respectively, an Israeli company that manufactures generic drugs and its American affiliate. In 2011, Teva sought approval from the FDA to market a generic 0.25 mg palonosetron product. Helsinn then sued Teva for infringing its patents, including the ’219 patent. In defense, Teva asserted that the ’219 patent was invalid because the 0.25 mg dose was “on sale” more than one year before Helsinn filed the provisional patent application covering that dose in January 2003. The AIA precludes a person from obtaining a patent on an invention that was “on sale” before the effective filing date of the patent application: “A person shall be entitled to a patent unless . . . the claimed invention was patented, described in a printed publication, or in public use, on sale, or otherwise available to the public before the effective filing date of the claimed invention.” 35 U. S. C. §102(a)(1) (emphasis added). See also §102(b)(1) (exception for certain disclosures made within a year before the effective filing date). Disclosures described in §102(a)(1) are often referred to as “prior art.” The patent statute in effect before the passage of the AIA included a similar proscription, known as the “on-sale bar”: “A person shall be entitled to a patent unless— “(a) the invention was known or used by others in this country, or patented or described in a printed publication in this or a foreign country, before the invention thereof by the applicant for patent, or “(b) the invention was patented or described in a printed publication in this or a foreign country or in public use or on sale in this country, more than one year prior to the date of the application for patent in the United States.” 35 U. S. C. §§102(a)–(b) (2006 ed.) (emphasis added). The District Court determined that the “on sale” provision did not apply. It concluded that, under the AIA, an invention is not “on sale” unless the sale or offer in question made the claimed invention available to the public. Helsinn Healthcare S. A. v. Dr. Reddy’s Labs. Ltd., 2016 WL 832089, *45, *51 (D NJ, Mar. 3, 2016). Because the companies’ public disclosure of the agreements between Helsinn and MGI did not disclose the 0.25 mg dose, the court determined that the invention was not “on sale” before the critical date. Id., at *51–*52. The Federal Circuit reversed. 855 F.3d 1356, 1360 (2017). It concluded that “if the existence of the sale is public, the details of the invention need not be publicly disclosed in the terms of sale” to fall within the AIA’s on-sale bar. Id., at 1371. Because the sale between Helsinn and MGI was publicly disclosed, it held that the on-sale bar applied. Id., at 1364, 1371. We granted certiorari to determine whether, under the AIA, an inventor’s sale of an invention to a third party who is obligated to keep the invention confidential qualifies as prior art for purposes of determining the patent- ability of the invention. 585 U. S. ___ (2018). We conclude that such a sale can qualify as prior art. II A The United States Constitution authorizes Congress “[t]o promote the Progress of Science and useful Arts, by securing for limited Times to Authors and Inventors the exclusive Right to their respective Writings and Discoveries.” Art. 1, §8, cl. 8. Under this grant of authority, Congress has crafted a federal patent system that encourages “the creation and disclosure of new, useful, and nonobvious advances in technology and design” by granting inventors “the exclusive right to practice the invention for a period of years.” Bonito Boats, Inc. v. Thunder Craft Boats, Inc., 489 U.S. 141, 151 (1989). To further the goal of “motivating innovation and enlightenment” while also “avoiding monopolies that unnecessarily stifle competition,” Pfaff, 525 U. S., at 63, Congress has imposed several conditions on the “limited opportunity to obtain a property right in an idea,” Bonito Boats, supra, at 149. One such condition is the on-sale bar, which reflects Congress’ “reluctance to allow an inventor to remove existing knowledge from public use” by obtaining a patent covering that knowledge. Pfaff, supra, at 64; see also Pennock v. Dialogue, 2 Pet. 1, 19 (1829) (explaining that “it would materially retard the progress of science and the useful arts” to allow an inventor to “sell his invention publicly” and later “take out a patent” and “exclude the public from any farther use than what should be derived under it”). Every patent statute since 1836 has included an on-sale bar. Pfaff, supra, at 65. The patent statute in force immediately before the AIA prevented a person from receiving a patent if, “more than one year prior to the date of the application for patent in the United States,” “the invention was . . . on sale” in the United States. 35 U. S. C. §102(b) (2006 ed., Supp. IV). The AIA, as relevant here, retained the on-sale bar and added the catchall phrase “or otherwise available to the public.” §102(a)(1) (2012 ed.) (“A person shall be entitled to a patent unless” the “claimed invention was . . . in public use, on sale, or otherwise available to the public . . . ”). We must decide whether these changes altered the meaning of the “on sale” bar. We hold that they did not. B Congress enacted the AIA in 2011 against the backdrop of a substantial body of law interpreting §102’s on-sale bar. In 1998, we determined that the pre-AIA on-sale bar applies “when two conditions are satisfied” more than a year before an inventor files a patent application. Pfaff, 525 U. S., at 67. “First, the product must be the subject of a commercial offer for sale.” Ibid. “Second, the invention must be ready for patenting,” which we explained could be shown by proof of “reduction to practice” or “drawings or other descriptions of the invention that were sufficiently specific to enable a person skilled in the art to practice the invention.” Id., at 67–68. Although this Court has never addressed the precise question presented in this case, our precedents suggest that a sale or offer of sale need not make an invention available to the public. For instance, we held in Pfaff that an offer for sale could cause an inventor to lose the right to patent, without regard to whether the offer discloses each detail of the invention. E.g., id., at 67. Other cases focus on whether the invention had been sold, not whether the details of the invention had been made available to the public or whether the sale itself had been publicly disclosed. E.g., Consolidated Fruit-Jar Co. v. Wright, 94 U.S. 92, 94 (1877) (“[A] single instance of sale or of use by the patentee may, under the circumstances, be fatal to the patent . . . ”); cf. Smith & Griggs Mfg. Co. v. Sprague, 123 U.S. 249, 257 (1887) (“A single sale to another . . . would certainly have defeated his right to a patent . . . ”); Elizabeth v. Pavement Co., 97 U.S. 126, 136 (1878) (“It is not a public knowledge of his invention that precludes the inventor from obtaining a patent for it, but a public use or sale of it”). The Federal Circuit—which has “exclusive jurisdiction” over patent appeals, 28 U. S. C. §1295(a)—has made explicit what was implicit in our precedents. It has long held that “secret sales” can invalidate a patent. E.g., Special Devices, Inc. v. OEA, Inc., 270 F.3d 1353, 1357 (2001) (invalidating patent claims based on “sales for the purpose of the commercial stockpiling of an invention” that “took place in secret”); Woodland Trust v. Flowertree Nursery, Inc., 148 F.3d 1368, 1370 (1998) (“Thus an inventor’s own prior commercial use, albeit kept secret, may constitute a public use or sale under §102(b), barring him from obtaining a patent”). In light of this settled pre-AIA precedent on the meaning of “on sale,” we presume that when Congress reenacted the same language in the AIA, it adopted the earlier judicial construction of that phrase. See Shapiro v. United States, 335 U.S. 1, 16 (1948) (“In adopting the language used in the earlier act, Congress ‘must be considered to have adopted also the construction given by this Court to such language, and made it a part of the enactment’ ”). The new §102 retained the exact language used in its predecessor statute (“on sale”) and, as relevant here, added only a new catchall clause (“or otherwise available to the public”). As amicus United States noted at oral argument, if “on sale” had a settled meaning before the AIA was adopted, then adding the phrase “or otherwise available to the public” to the statute “would be a fairly oblique way of attempting to overturn” that “settled body of law.” Tr. of Oral Arg. 28. The addition of “or otherwise available to the public” is simply not enough of a change for us to conclude that Congress intended to alter the meaning of the reenacted term “on sale.” Cf. Holder v. Martinez Gutierrez, 566 U.S. 583, 593 (2012) (determining that a reenacted provision did not ratify an earlier judicial construction where the provision omitted the word on which the prior judicial constructions were based). Helsinn disagrees, arguing that our construction reads “otherwise” out of the statute. Citing Paroline v. United States, 572 U.S. 434 (2014), and Federal Maritime Comm’n v. Seatrain Lines, Inc., 411 U.S. 726 (1973), Helsinn contends that the associated-words canon requires us to read “otherwise available to the public” to limit the preceding terms in §102 to disclosures that make the claimed invention available to the public. As an initial matter, neither of the cited decisions addresses the reenactment of terms that had acquired a well-settled judicial interpretation. And Helsinn’s argument places too much weight on §102’s catchall phrase. Like other such phrases, “otherwise available to the public” captures material that does not fit neatly into the statute’s enumerated categories but is nevertheless meant to be covered. Given that the phrase “on sale” had acquired a well-settled meaning when the AIA was enacted, we decline to read the addition of a broad catchall phrase to upset that body of precedent. III Helsinn does not ask us to revisit our pre-AIA interpretation of the on-sale bar. Nor does it dispute the Federal Circuit’s determination that the invention claimed in the ’219 patent was “on sale” within the meaning of the pre-AIA statute. Because we determine that Congress did not alter the meaning of “on sale” when it enacted the AIA, we hold that an inventor’s sale of an invention to a third party who is obligated to keep the invention confidential can qualify as prior art under §102(a). We therefore affirm the judgment of the Federal Circuit. It is so ordered.

586.US.2018_17-1272

Respondent Archer & White Sales, Inc., sued petitioner Henry Schein, Inc., alleging violations of federal and state antitrust law and seeking both money damages and injunctive relief. The relevant contract between the parties provided for arbitration of any dispute arising under or related to the agreement, except for, among other things, actions seeking injunctive relief. Invoking the Federal Arbitration Act, Schein asked the District Court to refer the matter to arbitration, but Archer & White argued that the dispute was not subject to arbitration because its complaint sought injunctive relief, at least in part. Schein contended that because the rules governing the contract provide that arbitrators have the power to resolve arbitrability questions, an arbitrator—not the court—should decide whether the arbitration agreement applied. Archer & White countered that Schein’s argument for arbitration was wholly groundless, so the District Court could resolve the threshold arbitrability question. The District Court agreed with Archer & White and denied Schein’s motion to compel arbitration. The Fifth Circuit affirmed. Held: The “wholly groundless” exception to arbitrability is inconsistent with the Federal Arbitration Act and this Court’s precedent. Under the Act, arbitration is a matter of contract, and courts must enforce arbitration contracts according to their terms. Rent-A-Center, West, Inc. v. Jackson, 561 U. S. 63, 67. The parties to such a contract may agree to have an arbitrator decide not only the merits of a particular dispute, but also “ ‘gateway’ questions of ‘arbitrability.’ ” Id., at 68–69. Therefore, when the parties’ contract delegates the arbitrability question to an arbitrator, a court may not override the contract, even if the court thinks that the arbitrability claim is wholly groundless. That conclusion follows also from this Court’s precedent. See AT&T Technologies, Inc. v. Communications Workers, 475 U. S. 643, 649–650. Archer & White’s counterarguments are unpersuasive. First, its argument that §§3 and 4 of the Act should be interpreted to mean that a court must always resolve questions of arbitrability has already been addressed and rejected by this Court. See, e.g., First Options of Chicago, Inc. v. Kaplan, 514 U. S. 938, 944. Second, its argument that §10 of the Act—which provides for back-end judicial review of an arbitrator’s decision if an arbitrator has “exceeded” his or her “powers”—supports the conclusion that the court at the front end should also be able to say that the underlying issue is not arbitrable is inconsistent with the way Congress designed the Act. And it is not this Court’s proper role to redesign the Act. Third, its argument that it would be a waste of the parties’ time and money to send wholly groundless arbitrability questions to an arbitrator ignores the fact that the Act contains no “wholly groundless” exception. This Court may not engraft its own exceptions onto the statutory text. Nor is it likely that the exception would save time and money systemically even if it might do so in some individual cases. Fourth, its argument that the exception is necessary to deter frivolous motions to compel arbitration overstates the potential problem. Arbitrators are already capable of efficiently disposing of frivolous cases and deterring frivolous motions, and such motions do not appear to have caused a substantial problem in those Circuits that have not recognized a “wholly groundless” exception. The Fifth Circuit may address the question whether the contract at issue in fact delegated the arbitrability question to an arbitrator, as well as other properly preserved arguments, on remand. Pp. 4–8. 878 F. 3d 488, vacated and remanded. Kavanaugh, J., delivered the opinion for a unanimous Court.

Under the Federal Arbitration Act, parties to a contract may agree that an arbitrator rather than a court will resolve disputes arising out of the contract. When a dispute arises, the parties sometimes may disagree not only about the merits of the dispute but also about the threshold arbitrability question—that is, whether their arbitration agreement applies to the particular dispute. Who decides that threshold arbitrability question? Under the Act and this Court’s cases, the question of who decides arbitrability is itself a question of contract. The Act allows parties to agree by contract that an arbitrator, rather than a court, will resolve threshold arbitrability questions as well as underlying merits disputes. Rent-A-Center, West, Inc. v. Jackson, 561 U. S. 63, 68−70 (2010); First Options of Chicago, Inc. v. Kaplan, 514 U. S. 938, 943−944 (1995). Even when a contract delegates the arbitrability question to an arbitrator, some federal courts nonetheless will short-circuit the process and decide the arbitrability question themselves if the argument that the arbitration agreement applies to the particular dispute is “wholly groundless.” The question presented in this case is whether the “wholly groundless” exception is consistent with the Federal Arbitration Act. We conclude that it is not. The Act does not contain a “wholly groundless” exception, and we are not at liberty to rewrite the statute passed by Congress and signed by the President. When the parties’ contract delegates the arbitrability question to an arbitrator, the courts must respect the parties’ decision as embodied in the contract. We vacate the contrary judgment of the Court of Appeals. I Archer and White is a small business that distributes dental equipment. Archer and White entered into a contract with Pelton and Crane, a dental equipment manufacturer, to distribute Pelton and Crane’s equipment. The relationship eventually soured. As relevant here, Archer and White sued Pelton and Crane’s successor-in-interest and Henry Schein, Inc. (collectively, Schein) in Federal District Court in Texas. Archer and White’s complaint alleged violations of federal and state antitrust law, and sought both money damages and injunctive relief. The relevant contract between the parties provided: “Disputes. This Agreement shall be governed by the laws of the State of North Carolina. Any dispute arising under or related to this Agreement (except for actions seeking injunctive relief and disputes related to trademarks, trade secrets, or other intellectual property of [Schein]), shall be resolved by binding arbitration in accordance with the arbitration rules of the American Arbitration Association [(AAA)]. The place of arbitration shall be in Charlotte, North Carolina.” App. to Pet. for Cert. 3a. After Archer and White sued, Schein invoked the Federal Arbitration Act and asked the District Court to refer the parties’ antitrust dispute to arbitration. Archer and White objected, arguing that the dispute was not subject to arbitration because Archer and White’s complaint sought injunctive relief, at least in part. According to Archer and White, the parties’ contract barred arbitration of disputes when the plaintiff sought injunctive relief, even if only in part. The question then became: Who decides whether the antitrust dispute is subject to arbitration? The rules of the American Arbitration Association provide that arbitrators have the power to resolve arbitrability questions. Schein contended that the contract’s express incorporation of the American Arbitration Association’s rules meant that an arbitrator—not the court—had to decide whether the arbitration agreement applied to this particular dispute. Archer and White responded that in cases where the defendant’s argument for arbitration is wholly groundless—as Archer and White argued was the case here—the District Court itself may resolve the threshold question of arbitrability. Relying on Fifth Circuit precedent, the District Court agreed with Archer and White about the existence of a “wholly groundless” exception, and ruled that Schein’s argument for arbitration was wholly groundless. The District Court therefore denied Schein’s motion to compel arbitration. The Fifth Circuit affirmed. In light of disagreement in the Courts of Appeals over whether the “wholly groundless” exception is consistent with the Federal Arbitration Act, we granted certiorari, 585 U. S. ___ (2018). Compare 878 F. 3d 488 (CA5 2017) (case below); Simply Wireless, Inc. v. T-Mobile US, Inc., 877 F. 3d 522 (CA4 2017); Douglas v. Regions Bank, 757 F. 3d 460 (CA5 2014); Turi v. Main Street Adoption Servs., LLP, 633 F. 3d 496 (CA6 2011); Qualcomm, Inc. v. Nokia Corp., 466 F. 3d 1366 (CA Fed. 2006), with Belnap v. Iasis Healthcare, 844 F. 3d 1272 (CA10 2017); Jones v. Waffle House, Inc., 866 F. 3d 1257 (CA11 2017); Douglas, 757 F. 3d, at 464 (Dennis, J., dissenting). II In 1925, Congress passed and President Coolidge signed the Federal Arbitration Act. As relevant here, the Act provides: “A written provision in . . . a contract evidencing a transaction involving commerce to settle by arbitration a controversy thereafter arising out of such contract . . . shall be valid, irrevocable, and enforceable, save upon such grounds as exist at law or in equity for the revocation of any contract.” 9 U. S. C. §2. Under the Act, arbitration is a matter of contract, and courts must enforce arbitration contracts according to their terms. Rent-A-Center, 561 U. S., at 67. Applying the Act, we have held that parties may agree to have an arbitrator decide not only the merits of a particular dispute but also “ ‘gateway’ questions of ‘arbitrability,’ such as whether the parties have agreed to arbitrate or whether their agreement covers a particular controversy.” Id., at 68–69; see also First Options, 514 U. S., at 943. We have explained that an “agreement to arbitrate a gateway issue is simply an additional, antecedent agreement the party seeking arbitration asks the federal court to enforce, and the FAA operates on this additional arbitration agree- ment just as it does on any other.” Rent-A-Center, 561 U. S., at 70. Even when the parties’ contract delegates the threshold arbitrability question to an arbitrator, the Fifth Circuit and some other Courts of Appeals have determined that the court rather than an arbitrator should decide the threshold arbitrability question if, under the contract, the argument for arbitration is wholly groundless. Those courts have reasoned that the “wholly groundless” exception enables courts to block frivolous attempts to transfer disputes from the court system to arbitration. We conclude that the “wholly groundless” exception is inconsistent with the text of the Act and with our precedent. We must interpret the Act as written, and the Act in turn requires that we interpret the contract as written. When the parties’ contract delegates the arbitrability question to an arbitrator, a court may not override the contract. In those circumstances, a court possesses no power to decide the arbitrability issue. That is true even if the court thinks that the argument that the arbitration agreement applies to a particular dispute is wholly groundless. That conclusion follows not only from the text of the Act but also from precedent. We have held that a court may not “rule on the potential merits of the underlying” claim that is assigned by contract to an arbitrator, “even if it appears to the court to be frivolous.” AT&T Technologies, Inc. v. Communications Workers, 475 U. S. 643, 649–650 (1986). A court has “ ‘no business weighing the merits of the grievance’ ” because the “ ‘agreement is to submit all grievances to arbitration, not merely those which the court will deem meritorious.’ ” Id., at 650 (quoting Steelworkers v. American Mfg. Co., 363 U. S. 564, 568 (1960)). That AT&T Technologies principle applies with equal force to the threshold issue of arbitrability. Just as a court may not decide a merits question that the parties have delegated to an arbitrator, a court may not decide an arbitrability question that the parties have delegated to an arbitrator. In an attempt to overcome the statutory text and this Court’s cases, Archer and White advances four main arguments. None is persuasive. First, Archer and White points to §§3 and 4 of the Federal Arbitration Act. Section 3 provides that a court must stay litigation “upon being satisfied that the issue” is “referable to arbitration” under the “agreement.” Section 4 says that a court, in response to a motion by an aggrieved party, must compel arbitration “in accordance with the terms of the agreement” when the court is “satisfied that the making of the agreement for arbitration or the failure to comply therewith is not in issue.” Archer and White interprets those provisions to mean, in essence, that a court must always resolve questions of arbitrability and that an arbitrator never may do so. But that ship has sailed. This Court has consistently held that parties may delegate threshold arbitrability questions to the arbitrator, so long as the parties’ agreement does so by “clear and unmistakable” evidence. First Options, 514 U. S., at 944 (alterations omitted); see also Rent-A-Center, 561 U. S., at 69, n. 1. To be sure, before referring a dispute to an arbitrator, the court determines whether a valid arbitration agreement exists. See 9 U. S. C. §2. But if a valid agreement exists, and if the agreement delegates the arbitrability issue to an arbitrator, a court may not decide the arbitrability issue. Second, Archer and White cites §10 of the Act, which provides for back-end judicial review of an arbitrator’s decision if an arbitrator has “exceeded” his or her “powers.” §10(a)(4). According to Archer and White, if a court at the back end can say that the underlying issue was not arbitrable, the court at the front end should also be able to say that the underlying issue is not arbitrable. The dispositive answer to Archer and White’s §10 argument is that Congress designed the Act in a specific way, and it is not our proper role to redesign the statute. Archer and White’s §10 argument would mean, moreover, that courts presumably also should decide frivolous merits questions that have been delegated to an arbitrator. Yet we have already rejected that argument: When the parties’ contract assigns a matter to arbitration, a court may not resolve the merits of the dispute even if the court thinks that a party’s claim on the merits is frivolous. AT&T Technologies, 475 U. S., at 649−650. So, too, with arbitrability. Third, Archer and White says that, as a practical and policy matter, it would be a waste of the parties’ time and money to send the arbitrability question to an arbitrator if the argument for arbitration is wholly groundless. In cases like this, as Archer and White sees it, the arbitrator will inevitably conclude that the dispute is not arbitrable and then send the case back to the district court. So why waste the time and money? The short answer is that the Act contains no “wholly groundless” exception, and we may not engraft our own exceptions onto the statutory text. See Exxon Mobil Corp. v. Allapattah Services, Inc., 545 U. S. 546, 556−557 (2005). In addition, contrary to Archer and White’s claim, it is doubtful that the “wholly groundless” exception would save time and money systemically even if it might do so in some individual cases. Archer and White assumes that it is easy to tell when an argument for arbitration of a particular dispute is wholly groundless. We are dubious. The exception would inevitably spark collateral litigation (with briefing, argument, and opinion writing) over whether a seemingly unmeritorious argument for arbitration is wholly groundless, as opposed to groundless. We see no reason to create such a time-consuming sideshow. Archer and White further assumes that an arbitrator would inevitably reject arbitration in those cases where a judge would conclude that the argument for arbitration is wholly groundless. Not always. After all, an arbitrator might hold a different view of the arbitrability issue than a court does, even if the court finds the answer obvious. It is not unheard-of for one fair-minded adjudicator to think a decision is obvious in one direction but for another fair-minded adjudicator to decide the matter the other way. Fourth, Archer and White asserts another policy argument: that the “wholly groundless” exception is necessary to deter frivolous motions to compel arbitration. Again, we may not rewrite the statute simply to accommodate that policy concern. In any event, Archer and White overstates the potential problem. Arbitrators can efficiently dispose of frivolous cases by quickly ruling that a claim is not in fact arbitrable. And under certain circumstances, arbitrators may be able to respond to frivolous arguments for arbitration by imposing fee-shifting and cost-shifting sanctions, which in turn will help deter and remedy frivolous motions to compel arbitration. We are not aware that frivolous motions to compel arbitration have caused a substantial problem in those Circuits that have not recognized a “wholly groundless” exception. In sum, we reject the “wholly groundless” exception. The exception is inconsistent with the statutory text and with our precedent. It confuses the question of who decides arbitrability with the separate question of who prevails on arbitrability. When the parties’ contract delegates the arbitrability question to an arbitrator, the courts must respect the parties’ decision as embodied in the contract. We express no view about whether the contract at issue in this case in fact delegated the arbitrability question to an arbitrator. The Court of Appeals did not decide that issue. Under our cases, courts “should not assume that the parties agreed to arbitrate arbitrability unless there is clear and unmistakable evidence that they did so.” First Options, 514 U. S., at 944 (alterations omitted). On remand, the Court of Appeals may address that issue in the first instance, as well as other arguments that Archer and White has properly preserved. The judgment of the Court of Appeals is vacated, and the case is remanded for further proceedings consistent with this opinion. It is so ordered.

587.US.2018_17-532

An 1868 treaty between the United States and the Crow Tribe promised that in exchange for most of the Tribe’s territory in modern-day Montana and Wyoming, its members would “have the right to hunt on the unoccupied lands of the United States so long as game may be found thereon . . . and peace subsists . . . on the borders of the hunting districts.” 15Stat. 650. In 2014, Wyoming charged petitioner Clayvin Herrera with off-season hunting in Bighorn National Forest and being an accessory to the same. The state trial court rejected Herrera’s argument that he had a protected right to hunt in the forest pursuant to the 1868 Treaty, and a jury convicted him. On appeal, the state appellate court relied on the reasoning of the Tenth Circuit’s decision in Crow Tribe of Indians v. Repsis, 73 F. 3d 982—which in turn relied upon this Court’s decision in Ward v. Race Horse, 163 U. S. 504—and held that the treaty right expired upon Wyoming’s statehood. The court rejected Herrera’s argument that this Court’s subsequent decision in Minnesota v. Mille Lacs Band of Chippewa Indians, 526 U.S. 172, repudiated Race Horse and therefore undercut the logic of Repsis. In any event, the court concluded, Herrera was precluded from arguing that the treaty right survived Wyoming’s statehood because the Crow Tribe had litigated Repsis on behalf of itself and its members. Even if the 1868 Treaty right survived Wyoming’s statehood, the court added, it did not permit Herrera to hunt in Bighorn National Forest because the treaty right applies only on unoccupied lands and the national forest became categorically occupied when it was created. Held: 1. The Crow Tribe’s hunting rights under the 1868 Treaty did not expire upon Wyoming’s statehood. Pp. 6–17. (a) This case is controlled by Mille Lacs, not Race Horse. Race Horse concerned a hunting right guaranteed in an 1868 treaty with the Shoshone and Bannock Tribes containing language identical to that at issue here. Relying on two lines of reasoning, the Race Horse Court held that Wyoming’s admission to the United States in 1890 extinguished the Shoshone-Bannock Treaty right. First, the doctrine that new States are admitted to the Union on an “equal footing” with existing States led the Court to conclude that affording the Tribes a protected hunting right lasting after statehood would conflict with the power vested in those States—and newly shared by Wyoming—“to regulate the killing of game within their borders.” 163 U. S., at 514. Second, the Court found no evidence in the Shoshone-Bannock Treaty itself that Congress intended the treaty right to continue in “perpetuity.” Id., at 514–515. Mille Lacs undercut both pillars of Race Horse’s reasoning. Mille Lacs established that the crucial inquiry for treaty termination analysis is whether Congress has “clearly express[ed]” an intent to abrogate an Indian treaty right, 526 U. S., at 202, or whether a termination point identified in the treaty itself has been satisfied, id., at 207. Thus, while Race Horse “was not expressly overruled” in Mille Lacs, it “retain[s] no vitality,” Limbach v. Hooven & Allison Co., 466 U.S. 353, 361, and is repudiated to the extent it held that treaty rights can be impliedly extinguished at statehood. Pp. 6–11. (b) Repsis does not preclude Herrera from arguing that the 1868 Treaty right survived Wyoming’s statehood. Even when the elements of issue preclusion are met, an exception may be warranted if there has been an intervening “ ‘change in [the] applicable legal context.’ ” Bobby v. Bies, 556 U.S. 825, 834. Here, Mille Lacs’ repudiation of Race Horse’s reasoning—on which Repsis relied—justifies such an exception. Pp. 11–13. (c) Applying Mille Lacs, Wyoming’s admission into the Union did not abrogate the Crow Tribe’s off-reservation treaty hunting right. First, the Wyoming Statehood Act does not show that Congress “clearly expressed” an intent to end the 1868 Treaty hunting right. See 526 U. S., at 202. There is also no evidence in the treaty itself that Congress intended the hunting right to expire at statehood, or that the Crow Tribe would have understood it to do so. Nor does the historical record support such a reading of the treaty. The State counters that statehood, as a practical matter, rendered all the lands in the State occupied. Even assuming that Wyoming presents an accurate historical picture, the State, by using statehood as a proxy for occupation, subverts this Court’s clear instruction that treaty-protected rights “are not impliedly terminated upon statehood.” Id., at 207. To the extent that the State seeks to rely on historical evidence to establish that all land in Wyoming was functionally “occupied” by 1890, its arguments fall outside the question presented and are unpersuasive in any event. Pp. 13–17. 2. Bighorn National Forest did not become categorically “occupied” within the meaning of the 1868 Treaty when the national forest was created. Construing the treaty’s terms as “ ‘they would naturally be understood by the Indians,’ ” Washington v. Washington State Commercial Passenger Fishing Vessel Assn., 443 U.S. 658, 676, it is clear that the Tribe would have understood the word “unoccupied” to denote an area free of residence or settlement by non-Indians. That interpretation follows from several cues in the treaty’s text. For example, the treaty made the hunting right contingent on peace “among the whites and Indians on the borders of the hunting districts,” 15Stat. 650, thus contrasting the unoccupied hunting districts with areas of white settlement. Historical evidence confirms this reading of “unoccupied.” Wyoming’s counterarguments are unavailing. The Federal Government’s exercise of control and withdrawing of the forest lands from settlement would not categorically transform the territory into an area resided on or settled by non-Indians; quite the opposite. Nor would mining and logging of the forest lands prior to 1897 have caused the Tribe to view the Bighorn Mountains as occupied. Pp. 17–21. 3. This decision is limited in two ways. First, the Court holds that Bighorn National Forest is not categorically occupied, not that all areas within the forest are unoccupied. Second, the state trial court de- cided that Wyoming could regulate the exercise of the 1868 Treaty right “in the interest of conservation,” an issue not reached by the appellate court. The Court also does not address the viability of the State’s arguments on this issue. Pp. 21–22. Vacated and remanded. Sotomayor, J., delivered the opinion of the Court, in which Ginsburg, Breyer, Kagan, and Gorsuch, JJ., joined. Alito, J., filed a dissenting opinion, in which Roberts, C. J., and Thomas and Kavanaugh, JJ., joined.

In 1868, the Crow Tribe ceded most of its territory in modern-day Montana and Wyoming to the United States. In exchange, the United States promised that the Crow Tribe “shall have the right to hunt on the unoccupied lands of the United States so long as game may be found thereon” and “peace subsists . . . on the borders of the hunting districts.” Treaty Between the United States of America and the Crow Tribe of Indians (1868 Treaty), Art. IV, May 7, 1868, 15Stat. 650. Petitioner Clayvin Herrera, a member of the Tribe, invoked this treaty right as a defense against charges of off-season hunting in Bighorn National Forest in Wyoming. The Wyoming courts held that the treaty-protected hunting right expired when Wyoming became a State and, in any event, does not permit hunting in Bighorn National Forest because that land is not “unoccupied.” We disagree. The Crow Tribe’s hunting right survived Wyoming’s statehood, and the lands within Bighorn National Forest did not become categorically “occupied” when set aside as a national reserve. I A The Crow Tribe first inhabited modern-day Montana more than three centuries ago. Montana v. United States, 450 U.S. 544, 547 (1981). The Tribe was nomadic, and its members hunted game for subsistence. J. Medicine Crow, From the Heart of the Crow Country 4–5, 8 (1992). The Bighorn Mountains of southern Montana and northern Wyoming “historically made up both the geographic and the spiritual heart” of the Tribe’s territory. Brief for Crow Tribe of Indians as Amicus Curiae 5. The westward migration of non-Indians began a new chapter in the Tribe’s history. In 1825, the Tribe signed a treaty of friendship with the United States. Treaty With the Crow Tribe, Aug. 4, 1825, 7Stat. 266. In 1851, the Federal Government and tribal representatives entered into the Treaty of Fort Laramie, in which the Crow Tribe and other area tribes demarcated their respective lands. Montana, 450 U. S., at 547–548. The Treaty of Fort Laramie specified that “the tribes did not ‘surrender the privilege of hunting, fishing, or passing over’ any of the lands in dispute” by entering the treaty. Id., at 548. After prospectors struck gold in Idaho and western Montana, a new wave of settlement prompted Congress to initiate further negotiations. See F. Hoxie, Parading Through History 88–90 (1995). Federal negotiators, including Commissioner of Indian Affairs Nathaniel G. Taylor, met with Crow Tribe leaders for this purpose in 1867. Taylor acknowledged that “settlements ha[d] been made” upon the Crow Tribe’s lands and that their “game [was] being driven away.” Institute for the Development of Indian Law, Proceedings of the Great Peace Commission of 1867–1868, p. 86 (1975) (hereinafter Proceedings). He told the assembled tribal leaders that the United States wished to “set apart a tract of [Crow Tribe] country as a home” for the Tribe “forever” and to buy the rest of the Tribe’s land. Ibid. Taylor emphasized that the Tribe would have “the right to hunt upon” the land it ceded to the Federal Government “as long as the game lasts.” Ibid. At the convening, Tribe leaders stressed the vital importance of preserving their hunting traditions. See id., at 88 (Black Foot: “You speak of putting us on a reservation and teaching us to farm. . . . That talk does not please us. We want horses to run after the game, and guns and ammunition to kill it. I would like to live just as I have been raised”); id., at 89 (Wolf Bow: “You want me to go on a reservation and farm. I do not want to do that. I was not raised so”). Although Taylor responded that “[t]he game w[ould] soon entirely disappear,” he also reassured tribal leaders that they would “still be free to hunt” as they did at the time even after the reservation was created. Id., at 90. The following spring, the Crow Tribe and the United States entered into the treaty at issue in this case: the 1868 Treaty. 15Stat. 649. Pursuant to the 1868 Treaty, the Crow Tribe ceded over 30 million acres of territory to the United States. See Montana, 450 U. S., at 547–548; Art. II, 15Stat. 650. The Tribe promised to make its “permanent home” a reservation of about 8 million acres in what is now Montana and to make “no permanent settlement elsewhere.” Art. IV, 15Stat. 650. In exchange, the United States made certain promises to the Tribe, such as agreeing to construct buildings on the reservation, to provide the Tribe members with seeds and implements for farming, and to furnish the Tribe with clothing and other goods. 1868 Treaty, Arts. III–XII, id., at 650–652. Article IV of the 1868 Treaty memorialized Commissioner Taylor’s pledge to preserve the Tribe’s right to hunt off-reservation, stating: “The Indians . . . shall have the right to hunt on the unoccupied lands of the United States so long as game may be found thereon, and as long as peace subsists among the whites and Indians on the borders of the hunting districts.” Id., at 650. A few months after the 1868 Treaty signing, Congress established the Wyoming Territory. Congress provided that the establishment of this new Territory would not “impair the rights of person or property now pertaining to the Indians in said Territory, so long as such rights shall remain unextinguished by treaty.” An Act to Provide a Temporary Government for the Territory of Wyoming (Wyoming Territory Act), July 25, 1868, ch. 235, 15Stat. 178. Around two decades later, the people of the new Territory adopted a constitution and requested admission to the United States. In 1890, Congress formally admitted Wyoming “into the Union on an equal footing with the original States in all respects,” in an Act that did not mention Indian treaty rights. An Act to Provide for the Admission of the State of Wyoming into the Union (Wyoming Statehood Act), July 10, 1890, ch. 664, 26Stat. 222. Finally, in 1897, President Grover Cleveland set apart an area in Wyoming as a public land reservation and declared the land “reserved from entry or settlement.” Presidential Proclamation No. 30, 29Stat. 909. This area, made up of lands ceded by the Crow Tribe in 1868, became known as the Bighorn National Forest. See App. 234; Crow Tribe of Indians v. Repsis, 73 F.3d 982, 985 (CA10 1995). B Petitioner Clayvin Herrera is a member of the Crow Tribe who resides on the Crow Reservation in Montana. In 2014, Herrera and other Tribe members pursued a group of elk past the boundary of the reservation and into the neighboring Bighorn National Forest in Wyoming. They shot several bull elk and returned to Montana with the meat. The State of Wyoming charged Herrera for taking elk off-season or without a state hunting license and with being an accessory to the same. In state trial court, Herrera asserted that he had a protected right to hunt where and when he did pursuant to the 1868 Treaty. The court disagreed and denied Herrera’s pretrial motion to dismiss. See Nos. CT–2015–2687, CT–2015–2688 (4th Jud. Dist. C. C., Sheridan Cty., Wyo., Oct. 16, 2015), App. to Pet. for Cert. 37, 41. Herrera unsuccessfully sought a stay of the trial court’s order from the Wyoming Supreme Court and this Court. He then went to trial, where he was not permitted to advance a treaty-based defense, and a jury convicted him on both counts. The trial court imposed a suspended jail sentence, as well as a fine and a 3-year suspension of Herrera’s hunting privileges. Herrera appealed. The central question facing the state appellate court was whether the Crow Tribe’s off-reservation hunting right was still valid. The U. S. Court of Appeals for the Tenth Circuit, reviewing the same treaty right in 1995 in Crow Tribe of Indians v. Repsis, had ruled that the right had expired when Wyoming became a State. 73 F. 3d, at 992–993. The Tenth Circuit’s decision in Repsis relied heavily on a 19th-century decision of this Court, Ward v. Race Horse, 163 U.S. 504, 516 (1896). Herrera argued in the state court that this Court’s subsequent decision in Minnesota v. Mille Lacs Band of Chippewa Indians, 526 U.S. 172 (1999), repudiated Race Horse, and he urged the Wyoming court to follow Mille Lacs instead of the Repsis and Race Horse decisions that preceded it. The state appellate court saw things differently. Reasoning that Mille Lacs had not overruled Race Horse, the court held that the Crow Tribe’s 1868 Treaty right expired upon Wyoming’s statehood. No. 2016–242 (4th Jud. Dist., Sheridan Cty., Wyo., Apr. 25, 2017), App. to Pet. for Cert. 31–34. Alternatively, the court concluded that the Repsis Court’s judgment merited issue-preclusive effect against Herrera because he is a member of the Crow Tribe, and the Tribe had litigated the Repsis suit on behalf of itself and its members. App. to Pet. for Cert. 15–17, 31; App. 258. Herrera, in other words, was not allowed to relitigate the validity of the treaty right in his own case. The court also held that, even if the 1868 Treaty right survived Wyoming’s entry into the Union, it did not permit Herrera to hunt in Bighorn National Forest. Again following Repsis, the court concluded that the treaty right applies only on “unoccupied” lands and that the national forest became categorically “occupied” when it was created. See App. to Pet. for Cert. 33–34; Repsis, 73 F. 3d, at 994. The state appellate court affirmed the trial court’s judgment and sentence. The Wyoming Supreme Court denied a petition for review, and this Court granted certiorari. 585 U. S. ___ (2018). For the reasons that follow, we now vacate and remand. II We first consider whether the Crow Tribe’s hunting rights under the 1868 Treaty remain valid. Relying on this Court’s decision in Mille Lacs, Herrera and the United States contend that those rights did not expire when Wyoming became a State in 1890. We agree. A Wyoming argues that this Court’s decision in Race Horse establishes that the Crow Tribe’s 1868 Treaty right expired at statehood. But this case is controlled by Mille Lacs, not Race Horse. Race Horse concerned a hunting right guaranteed in a treaty with the Shoshone and Bannock Tribes. The Shoshone-Bannock Treaty and the 1868 Treaty with the Crow Tribe were signed in the same year and contain identical language reserving an off-reservation hunting right. See Treaty Between the United States of America and the Eastern Band of Shoshonees [sic] and the Bannack [sic] Tribe of Indians (Shoshone-Bannock Treaty), July 3, 1868, 15Stat. 674–675 (“[T]hey shall have the right to hunt on the unoccupied lands of the United States so long as game may be found thereon, and so long as peace subsists among the whites and Indians on the borders of the hunting districts”). The Race Horse Court concluded that Wyoming’s admission to the United States extinguished the Shoshone-Bannock Treaty right. 163 U. S., at 505, 514–515. Race Horse relied on two lines of reasoning. The first turned on the doctrine that new States are admitted to the Union on an “equal footing” with existing States. Id., at 511–514 (citing, e.g., Lessee of Pollard v. Hagan, 3 How. 212 (1845)). This doctrine led the Court to conclude that the Wyoming Statehood Act repealed the Shoshone and Bannock Tribes’ hunting rights, because affording the Tribes a protected hunting right lasting after statehood would be “irreconcilably in conflict” with the power—“vested in all other States of the Union” and newly shared by Wyoming—“to regulate the killing of game within their borders.” 163 U. S., at 509, 514. Second, the Court found no evidence in the Shoshone-Bannock Treaty itself that Congress intended the treaty right to continue in “perpetuity.” Id., at 514–515. To the contrary, the Court emphasized that Congress “clearly contemplated the disappearance of the conditions” specified in the treaty. Id., at 509. The Court decided that the rights at issue in the Shoshone-Bannock Treaty were “essentially perishable” and afforded the Tribes only a “temporary and precarious” privilege. Id., at 515. More than a century after Race Horse and four years after Repsis relied on that decision, however, Mille Lacs undercut both pillars of Race Horse’s reasoning. Mille Lacs considered an 1837 Treaty that guaranteed to several bands of Chippewa Indians the privilege of hunting, fishing, and gathering in ceded lands “ ‘during the pleasure of the President.’ ” 526 U. S., at 177 (quoting 1837 Treaty With the Chippewa, 7Stat. 537). In an opinion extensively discussing and distinguishing Race Horse, the Court decided that the treaty rights of the Chippewa bands survived after Minnesota was admitted to the Union. 526 U. S., at 202–208. Mille Lacs approached the question before it in two stages. The Court first asked whether the Act admitting Minnesota to the Union abrogated the treaty right of the Chippewa bands. Next, the Court examined the Chippewa Treaty itself for evidence that the parties intended the treaty right to expire at statehood. These inquires roughly track the two lines of analysis in Race Horse. Despite these parallel analyses, however, the Mille Lacs Court refused Minnesota’s invitation to rely on Race Horse, explaining that the case had “been qualified by later decisions.” 526 U. S., at 203. Although Mille Lacs stopped short of explicitly overruling Race Horse, it methodically repudiated that decision’s logic. To begin with, in addressing the effect of the Minnesota Statehood Act on the Chippewa Treaty right, the Mille Lacs Court entirely rejected the “equal footing” reasoning applied in Race Horse. The earlier case concluded that the Act admitting Wyoming to the Union on an equal footing “repeal[ed]” the Shoshone-Bannock Treaty right because the treaty right was “irreconcilable” with state sovereignty over natural resources. Race Horse, 163 U. S., at 514. But Mille Lacs explained that this conclusion “rested on a false premise.” 526 U. S., at 204. Later decisions showed that States can impose reasonable and nondiscriminatory regulations on an Indian tribe’s treaty-based hunting, fishing, and gathering rights on state land when necessary for conservation. Id., at 204–205 (citing Washington v. Washington State Commercial Passenger Fishing Vessel Assn., 443 U.S. 658, 682 (1979); Antoine v. Washington, 420 U.S. 194, 207–208 (1975); Puyallup Tribe v. Department of Game of Wash., 391 U.S. 392, 398 (1968)). “[B]ecause treaty rights are reconcilable with state sovereignty over natural resources,” the Mille Lacs Court concluded, there is no reason to find statehood itself sufficient “to extinguish Indian treaty rights to hunt, fish, and gather on land within state boundaries.” 526 U. S., at 205. In lieu of adopting the equal-footing analysis, the Court instead drew on numerous decisions issued since Race Horse to explain that Congress “must clearly express” any intent to abrogate Indian treaty rights. 526 U. S., at 202 (citing United States v. Dion, 476 U.S. 734, 738–740 (1986); Fishing Vessel Assn., 443 U. S., at 690; Menominee Tribe v. United States, 391 U.S. 404, 413 (1968)). The Court found no such “ ‘clear evidence’ ” in the Act admitting Minnesota to the Union, which was “silent” with regard to Indian treaty rights. 526 U. S., at 203. The Mille Lacs Court then turned to what it referred to as Race Horse’s “alternative holding” that the rights in the Shoshone-Bannock Treaty “were not intended to survive Wyoming’s statehood.” 526 U. S., at 206. The Court observed that Race Horse could be read to suggest that treaty rights only survive statehood if the rights are “ ‘ “of such a nature as to imply their perpetuity,” ’ ” rather than “ ‘temporary and precarious.’ ” 526 U. S., at 206. The Court rejected such an approach. The Court found the “ ‘temporary and precarious’ ” language “too broad to be useful,” given that almost any treaty rights—which Congress may unilaterally repudiate, see Dion, 476 U. S., at 738—could be described in those terms. 526 U. S., at 206–207. Instead, Mille Lacs framed Race Horse as inquiring into whether the Senate “intended the rights secured by the . . . Treaty to survive statehood.” 526 U. S., at 207. Applying this test, Mille Lacs concluded that statehood did not extinguish the Chippewa bands’ treaty rights. The Chippewa Treaty itself defined the specific “circumstances under which the rights would terminate,” and there was no suggestion that statehood would satisfy those circumstances. Ibid. Maintaining its focus on the treaty’s language, Mille Lacs distinguished the Chippewa Treaty before it from the Shoshone-Bannock Treaty at issue in Race Horse. Specifically, the Court noted that the Shoshone-Bannock Treaty, unlike the Chippewa Treaty, “tie[d] the duration of the rights to the occurrence of some clearly contemplated event[s]”—i.e., to whenever the hunting grounds would cease to “remai[n] unoccupied and owned by the United States.” 526 U. S., at 207. In drawing that distinction, however, the Court took care to emphasize that the treaty termination analysis turns on the events enumerated in the “Treaty itself.” Ibid. Insofar as the Race Horse Court determined that the Shoshone-Bannock Treaty was “impliedly repealed,” Mille Lacs disavowed that earlier holding. 526 U. S., at 207. “Treaty rights,” the Court clarified, “are not impliedly terminated upon statehood.” Ibid. The Court further explained that “[t]he Race Horse Court’s decision to the contrary”—that Wyoming’s statehood did imply repeal of Indian treaty rights—“was informed by” that Court’s erroneous conclusion “that the Indian treaty rights were inconsistent with state sovereignty over natural resources.” Id., at 207–208. In sum, Mille Lacs upended both lines of reasoning in Race Horse. The case established that the crucial inquiry for treaty termination analysis is whether Congress has expressly abrogated an Indian treaty right or whether a termination point identified in the treaty itself has been satisfied. Statehood is irrelevant to this analysis unless a statehood Act otherwise demonstrates Congress’ clear intent to abrogate a treaty, or statehood appears as a termination point in the treaty. See 526 U. S., at 207. “[T]here is nothing inherent in the nature of reserved treaty rights to suggest that they can be extinguished by implication at statehood.” Ibid. Even Wyoming concedes that the Court has rejected the equal-footing reasoning in Race Horse, Brief for Respondent 26, but the State contends that Mille Lacs reaffirmed the alternative holding in Race Horse that the Shoshone-Bannock Treaty right (and thus the identically phrased right in the 1868 Treaty with the Crow Tribe) was in- tended to end at statehood. We are unpersuaded. As explained above, although the decision in Mille Lacs did not explicitly say that it was overruling the alternative ground in Race Horse, it is impossible to harmonize Mille Lacs’ analysis with the Court’s prior reasoning in Race Horse.[1] We thus formalize what is evident in Mille Lacs itself. While Race Horse “was not expressly overruled” in Mille Lacs, “it must be regarded as retaining no vitality” after that decision. Limbach v. Hooven & Allison Co., 466 U.S. 353, 361 (1984). To avoid any future confusion, we make clear today that Race Horse is repudiated to the extent it held that treaty rights can be impliedly extinguished at statehood. B Because this Court’s intervening decision in Mille Lacs repudiated the reasoning on which the Tenth Circuit relied in Repsis, Repsis does not preclude Herrera from arguing that the 1868 Treaty right survived Wyoming’s statehood. Under the doctrine of issue preclusion, “a prior judgment . . . foreclos[es] successive litigation of an issue of fact or law actually litigated and resolved in a valid court determination essential to the prior judgment.” New Hampshire v. Maine, 532 U.S. 742, 748–749 (2001). Even when the elements of issue preclusion are met, however, an exception may be warranted if there has been an intervening “ ‘change in [the] applicable legal context.’ ” Bobby v. Bies, 556 U.S. 825, 834 (2009) (quoting Restatement (Second) of Judgments §28, Comment c (1980)); see Limbach, 466 U. S., at 363 (refusing to find a party bound by “an early decision based upon a now repudiated legal doctrine”); see also Montana v. United States, 440 U.S. 147, 155 (1979) (asking “whether controlling facts or legal principles ha[d] changed significantly” since a judgment before giving it preclusive effect); id., at 157–158 (explaining that a prior judgment was conclusive “[a]bsent significant changes in controlling facts or legal principles” since the judgment); Commissioner v. Sunnen, 333 U.S. 591, 599 (1948) (issue preclusion “is designed to prevent repetitious lawsuits over matters which have once been decided and which have remained substantially static, factually and legally”). The change-in-law exception recognizes that applying issue preclusion in changed circumstances may not “advance the equitable administration of the law.” Bobby, 556 U. S., at 836–837.[2] We conclude that a change in law justifies an exception to preclusion in this case. There is no question that the Tenth Circuit in Repsis relied on this Court’s binding decision in Race Horse to conclude that the 1868 Treaty right terminated upon Wyoming’s statehood. See 73 F. 3d, at 994. When the Tenth Circuit reached its decision in Repsis, it had no authority to disregard this Court’s holding in Race Horse and no ability to predict the analysis this Court would adopt in Mille Lacs. Mille Lacs repudiated Race Horse’s reasoning. Although we recognize that it may be difficult at the margins to discern whether a particular legal shift warrants an exception to issue preclusion, this is not a marginal case. At a minimum, a repudiated decision does not retain preclusive force. See Limbach, 466 U. S., at 363.[3] C We now consider whether, applying Mille Lacs, Wyoming’s admission to the Union abrogated the Crow Tribe’s off-reservation treaty hunting right. It did not. First, the Wyoming Statehood Act does not show that Congress intended to end the 1868 Treaty hunting right. If Congress seeks to abrogate treaty rights, “it must clearly express its intent to do so.” Mille Lacs, 526 U. S., at 202. “There must be ‘clear evidence that Congress actually considered the conflict between its intended action on the one hand and Indian treaty rights on the other, and chose to resolve that conflict by abrogating the treaty.’ ” Id., at 202–203 (quoting Dion, 476 U. S., at 740); see Menominee Tribe, 391 U. S., at 412. Like the Act discussed in Mille Lacs, the Wyoming Statehood Act “makes no mention of Indian treaty rights” and “provides no clue that Congress considered the reserved rights of the [Crow Tribe] and decided to abrogate those rights when it passed the Act.” Cf. Mille Lacs, 526 U. S., at 203; see Wyoming Statehood Act, 26Stat. 222. There simply is no evidence that Congress intended to abrogate the 1868 Treaty right through the Wyoming Statehood Act, much less the “ ‘clear evidence’ ” this Court’s precedent requires. Mille Lacs, 526 U. S., at 203.[4] Nor is there any evidence in the treaty itself that Congress intended the hunting right to expire at statehood, or that the Crow Tribe would have understood it to do so. A treaty is “essentially a contract between two sovereign nations.” Fishing Vessel Assn., 443 U. S., at 675. Indian treaties “must be interpreted in light of the parties’ intentions, with any ambiguities resolved in favor of the Indians,” Mille Lacs, 526 U. S., at 206, and the words of a treaty must be construed “ ‘in the sense in which they would naturally be understood by the Indians,’ ” Fishing Vessel Assn., 443 U. S., at 676. If a treaty “itself defines the circumstances under which the rights would terminate,” it is to those circumstances that the Court must look to determine if the right ends at statehood. Mille Lacs, 526 U. S., at 207. Just as in Mille Lacs, there is no suggestion in the text of the 1868 Treaty with the Crow Tribe that the parties intended the hunting right to expire at statehood. The treaty identifies four situations that would terminate the right: (1) the lands are no longer “unoccupied”; (2) the lands no longer belong to the United States; (3) game can no longer “be found thereon”; and (4) the Tribe and non-Indians are no longer at “peace . . . on the borders of the hunting districts.” Art. IV, 15Stat. 650. Wyoming’s statehood does not appear in this list. Nor is there any hint in the treaty that any of these conditions would necessarily be satisfied at statehood. See Mille Lacs, 526 U. S., at 207. The historical record likewise does not support the State’s position. See Choctaw Nation v. United States, 318 U.S. 423, 431–432 (1943) (explaining that courts “may look beyond the written words to the history of the treaty, the negotiations, and the practical construction adopted by the parties” to determine a treaty’s meaning). Crow Tribe leaders emphasized the importance of the hunting right in the 1867 negotiations, see, e.g., Proceedings 88, and Commissioner Taylor assured them that the Tribe would have “the right to hunt upon [the ceded land] as long as the game lasts,” id., at 86. Yet despite the apparent importance of the hunting right to the negotiations, Wyoming points to no evidence that federal negotiators ever proposed that the right would end at statehood. This silence is especially telling because five States encompassing lands west of the Mississippi River—Nebraska, Nevada, Kansas, Oregon, and Minnesota—had been admitted to the Union in just the preceding decade. See ch. 36, 14Stat. 391 (Nebraska, Feb. 9, 1867); Presidential Proclamation No. 22, 13Stat. 749 (Nevada, Oct. 31, 1864); ch. 20, 12Stat. 126 (Kansas, Jan. 29, 1861); ch. 33, 11Stat. 383 (Oregon, Feb. 14, 1859); ch. 31, 11Stat. 285 (Minnesota, May 11, 1858). Federal negotiators had every reason to bring up statehood if they intended it to extinguish the Tribe’s hunting rights. In the face of this evidence, Wyoming nevertheless contends that the 1868 Treaty expired at statehood pursuant to the Mille Lacs analysis. Wyoming does not argue that the legal act of Wyoming’s statehood abrogated the treaty right, and it cannot contend that statehood is explicitly identified as a treaty expiration point. Instead, Wyoming draws on historical sources to assert that statehood, as a practical matter, marked the arrival of “civilization” in the Wyoming Territory and thus rendered all the lands in the State occupied. Brief for Respondent 48. This claim cannot be squared with Mille Lacs. Wyoming’s arguments boil down to an attempt to read the treaty impliedly to terminate at statehood, precisely as Mille Lacs forbids. The State sets out a potpourri of evidence that it claims shows statehood in 1890 effectively coincided with the disappearance of the wild frontier: for instance, that the buffalo were extinct by the mid-1870s; that by 1880, Indian Department regulations instructed Indian agents to confine tribal members “ ‘wholly within the limits of their respective reservations’ ”; and that the Crow Tribe stopped hunting off-reservation altogether in 1886. Brief for Respondent 47 (quoting §237 Instructions to Indian Agents (1880), as published in Regulations of the Indian Dept. §492 (1884)). Herrera contradicts this account, see Reply Brief for Petitioner 5, n. 3, and the historical record is by no means clear. For instance, game appears to have persisted for longer than Wyoming suggests. See Dept. of Interior, Ann. Rep. of the Comm’r of Indian Affairs 495 (1873) (Black Foot: “On the other side of the river below, there are plenty of buffalo; on the mountains are plenty of elk and black-tail deer; and white-tail deer are plenty at the foot of the mountain”). As for the Indian Department Regulations, there are reports that a group of Crow Tribe members “regularly hunted along the Little Bighorn River” even after the regulation the State cites was in effect. Hoxie, Parading Through History, at 26. In 1889, the Office of Indian Affairs wrote to U. S. Indian Agents in the Northwest that “[f]requent complaints have been made to this Department that Indians are in the habit of leaving their reservations for the purpose of hunting.” 28 Cong. Rec. 6231 (1896). Even assuming that Wyoming presents an accurate historical picture, the State’s mode of analysis is severely flawed. By using statehood as a proxy for occupation, Wyoming subverts this Court’s clear instruction that treaty-protected rights “are not impliedly terminated upon statehood.” Mille Lacs, 526 U. S., at 207. Finally, to the extent that Wyoming seeks to rely on this same evidence to establish that all land in Wyoming was functionally “occupied” by 1890, its arguments fall outside the question presented and are unpersuasive in any event. As explained below, the Crow Tribe would have understood occupation to denote some form of residence or settlement. See infra, at 19–20. Furthermore, Wyoming cannot rely on Race Horse to equate occupation with statehood, because that case’s reasoning rested on the flawed belief that statehood could not coexist with a continuing treaty right. See Race Horse, 163 U. S., at 514; Mille Lacs, 526 U. S., at 207–208. Applying Mille Lacs, this is not a hard case. The Wyoming Statehood Act did not abrogate the Crow Tribe’s hunting right, nor did the 1868 Treaty expire of its own accord at that time. The treaty itself defines the circumstances in which the right will expire. Statehood is not one of them. III We turn next to the question whether the 1868 Treaty right, even if still valid after Wyoming’s statehood, does not protect hunting in Bighorn National Forest because the forest lands are “occupied.” We agree with Herrera and the United States that Bighorn National Forest did not become categorically “occupied” within the meaning of the 1868 Treaty when the national forest was created.[5] Treaty analysis begins with the text, and treaty terms are construed as “ ‘they would naturally be understood by the Indians.’ ” Fishing Vessel Assn., 443 U. S., at 676. Here it is clear that the Crow Tribe would have understood the word “unoccupied” to denote an area free of residence or settlement by non-Indians. That interpretation follows first and foremost from several cues in the treaty’s text. For example, Article IV of the 1868 Treaty made the hunting right contingent on peace “among the whites and Indians on the borders of the hunting districts,” thus contrasting the unoccupied hunting districts with areas of white settlement. 15Stat. 650. The treaty elsewhere used the word “occupation” to refer to the Tribe’s residence inside the reservation boundaries, and referred to the Tribe members as “settlers” on the new reservation. Arts. II, VI, id., at 650–651. The treaty also juxtaposed occupation and settlement by stating that the Tribe was to make “no permanent settlement” other than on the new reservation, but could hunt on the “unoccupied lands” of the United States. Art. IV, id., at 650. Contemporaneous definitions further support a link between occupation and settlement. See W. Anderson, A Dictionary of Law 725 (1889) (defining “occupy” as “[t]o hold in possession; to hold or keep for use” and noting that the word “[i]mplies actual use, possession or cultivation by a particular person”); id., at 944 (defining “settle” as “[t]o establish one’s self upon; to occupy, reside upon”). Historical evidence confirms this reading of the word “unoccupied.” At the treaty negotiations, Commissioner Taylor commented that “settlements ha[d] been made upon [Crow Tribe] lands” and that “white people [were] rapidly increasing and . . . occupying all the valuable lands.” Proceedings 86. It was against this backdrop of white settlement that the United States proposed to buy “the right to use and settle” the ceded lands, retaining for the Tribe the right to hunt. Ibid. A few years after the 1868 Treaty signing, a leader of the Board of Indian Commissioners confirmed the connection between occupation and settlement, explaining that the 1868 Treaty permitted the Crow Tribe to hunt in an area “as long as there are any buffalo, and as long as the white men are not [in that area] with farms.” Dept. of Interior, Ann. Rep. of the Comm’r of Indian Affairs 500. Given the tie between the term “unoccupied” and a lack of non-Indian settlement, it is clear that President Cleveland’s proclamation creating Bighorn National Forest did not “occupy” that area within the treaty’s meaning. To the contrary, the President “reserved” the lands “from entry or settlement.” Presidential Proclamation No. 30, 29Stat. 909. The proclamation gave “[w]arning . . . to all persons not to enter or make settlement upon the tract of land reserved by th[e] proclamation.” Id., at 910. If anything, this reservation made Bighorn National Forest more hospitable, not less, to the Crow Tribe’s exercise of the 1868 Treaty right. Wyoming’s counterarguments are unavailing. The State first asserts that the forest became occupied through the Federal Government’s “exercise of dominion and control” over the forest territory, including federal regulation of those lands. Brief for Respondent 56–60. But as explained, the treaty’s text and the historical record suggest that the phrase “unoccupied lands” had a specific meaning to the Crow Tribe: lack of settlement. The proclamation of a forest reserve withdrawing land from settlement would not categorically transform the territory into an area resided on or settled by non-Indians; quite the opposite. Nor would the restrictions on hunting in national forests that Wyoming cites. See Appropriations Act of 1899, ch. 424, 30Stat. 1095; 36 CFR §§241.2, 241.3 (Supp. 1941); §261.10(d)(1) (2018). Wyoming also claims that exploitative mining and logging of the forest lands prior to 1897 would have caused the Crow Tribe to view the Bighorn Mountains as occupied. But the presence of mining and logging operations did not amount to settlement of the sort that the Tribe would have understood as rendering the forest occupied. In fact, the historical source on which Wyoming primarily relies indicates that there was “very little” settlement of Bighorn National Forest around the time the forest was created. Dept. of Interior, Nineteenth Ann. Rep. of the U. S. Geological Survey 167 (1898). Considering the terms of the 1868 Treaty as they would have been understood by the Crow Tribe, we conclude that the creation of Bighorn National Forest did not remove the forest lands, in their entirety, from the scope of the treaty. IV Finally, we note two ways in which our decision is limited. First, we hold that Bighorn National Forest is not categorically occupied, not that all areas within the forest are unoccupied. On remand, the State may argue that the specific site where Herrera hunted elk was used in such a way that it was “occupied” within the meaning of the 1868 Treaty. See State v. Cutler, 109 Idaho 448, 451, 708 P.2d 853, 856 (1985) (stating that the Federal Government may not be foreclosed from using land in such a way that the Indians would have considered it occupied). Second, the state trial court decided that Wyoming could regulate the exercise of the 1868 Treaty right “in the interest of conservation.” Nos. CT–2015–2687, CT–2015–2688, App. to Pet. for Cert. 39–41; see Antoine, 420 U. S., at 207. The appellate court did not reach this issue. No. 2016–242, App. to Pet. for Cert. 14, n. 3. On remand, the State may press its arguments as to why the application of state conservation regulations to Crow Tribe members exercising the 1868 Treaty right is necessary for conservation. We do not pass on the viability of those arguments today. * * * The judgment of the Wyoming District Court of the Fourth Judicial District, Sheridan County, is vacated, and the case is remanded for further proceedings not inconsistent with this opinion. It is so ordered. Notes 1 Notably, the four Justices who dissented in Mille Lacs protested that the Court “effectively overrule[d] Race Horse sub silentio.” 526 U. S., at 219 (Rehnquist, C. J., dissenting). Others have agreed with this assessment. See, e.g., State v. Buchanan, 138 Wash. 2d 186, 211–212, 978 P.2d 1070, 1083 (1999) (“[T]he United States Supreme Court effectively overruled Race Horse in Minnesota v. Mille Lacs”). 2 The dissent does not disagree outright with this conclusion, noting only that “there is a respectable argument on the other side,” post, at 12. The dissent argues that the cases cited above are distinguishable, but we do not read them as narrowly as does the dissent. We note, too, that the lower federal courts have long applied the change-in-law exception in a variety of contexts. See, e.g., Dow Chemical Co. v. Nova Chemicals Corp. (Canada), 803 F.3d 620, 627–630 (CA Fed. 2015), cert. denied, 578 U. S. ___ (2016); Coors Brewing Co. v. Mendez-Torres, 562 F.3d 3, 11 (CA1 2009), abrogated on other grounds by Levin v. Commerce Energy, Inc., 560 U.S. 413 (2010); Ginters v. Frazier, 614 F.3d 822, 826–827 (CA8 2010); Faulkner v. National Geographic Enterprises Inc., 409 F.3d 26, 37–38 (CA2 2005); Chippewa & Flambeau Improvement Co. v. FERC, 325 F.3d 353, 356–357 (CADC 2003); Spradling v. Tulsa, 198 F.3d 1219, 1222–1223 (CA10 2000); Mendelovitz v. Adolph Coors Co., 693 F.2d 570, 579 (CA5 1982). 3 We do not address whether a different outcome would be justified if the State had identified “compelling concerns of repose or reliance.” See 18 C. Wright, A. Miller, & E. Cooper, Federal Practice and Procedure §4425, p. 726 (3d ed. 2016). Wyoming here has not done so. The State suggests that public support for its conservation efforts may be jeopardized if it no longer has “unquestioned” authority over wildlife management in the Bighorn Mountains. Brief for Respondent 54. Wyoming does not explain why its authority to regulate Indians exercising their treaty rights when necessary for conservation is not sufficient to preserve that public support, see infra, at 22. The State’s passing reference to upsetting the settled expectations of private property owners is unconvincing because the 1868 Treaty right applies only to “unoccupied lands of the United States.” 4 Recall also that the Act establishing the Wyoming Territory declared that the creation of the Territory would not “impair the rights of person or property now pertaining to the Indians in said Territory” unless a treaty extinguished those rights. Wyoming Territory Act, 15Stat. 178. 5 Wyoming argues that the judgment below should be affirmed because the Tenth Circuit held in Repsis that the creation of the forest rendered the land “occupied,” see 73 F. 3d, at 994, and thus Herrera is precluded from raising this issue. We did not grant certiorari on the question of how preclusion principles would apply to the alternative judgment in Repsis, and—although our dissenting colleagues disagree, see post, at 13, and n. 6—the decision below did not address that issue. The Wyoming appellate court agreed with the State that “the pri-mary issue in [Herrera’s] case is identical to the primary issue in the Repsis case.” No. 2016–242 (4th Jud. Dist., Sheridan Cty., Wyo., Apr. 25, 2017), App. to Pet. for Cert. 13 (emphasis added). That “primary issue” was the Race Horse ground of decision, not the “occupation” ground, which Repsis referred to as “an alternative basis for affirmance,” Repsis, 73 F. 3d, at 993, and which the Wyoming court itself described as an “alternativ[e]” holding, No. 2016–242, App. to Pet. for Cert. 33. Reading the state court’s decision to give preclusive effect to the occupation ground as well would not fit with the Wyoming court’s preclusion analysis, which, among other things, relied on a decision of the Federal District Court in Repsis that did not address the occupation issue. See No. 2016–242, App. to Pet. for Cert. 14, 18; see also Repsis, 73 F. 3d, at 993 (explaining that “the district court did not reach [the occupation] issue”). Context thus makes clear that the state court gave issue-preclusive effect only to Repsis’ holding that the 1868 Treaty was no longer valid, not to Repsis’ independent, narrower holding that Bighorn National Forest in particular was “occupied” land. The court may not have addressed the issue-preclusive effect of the latter holding because of ambiguity in the State’s briefing. See Appellee’s Supplemental Brief in No. 2016–242, pp. 4, 11–12. While the dissent questions whether forfeiture could have played a part in the state court’s analysis given that the court invited the parties to submit supplemental briefs on preclusion, post, at 13, n. 6, the parties suggest that Wyoming failed adequately to raise the claim even in its supplemental brief. See Brief for Petitioner 49 (“the state made no such argument before” the state court); Brief for United States as Amicus Curiae 31 (noting ambiguity in the State’s supplemental brief). It can be “appropriate in special circumstances” for a court to address a preclusion argument sua sponte. Arizona v. California, 530 U.S. 392, 412 (2000). But because the Wyoming District Court “did not address” this contention, “we decline to address it here.” County of Los Angeles v. Mendez, 581 U. S. ___, ___, n. (2017) (slip op., at 8, n.); see Cutter v. Wilkinson, 544 U.S. 709, 718, n. 7 (2005); Archer v. Warner, 538 U.S. 314, 322–323 (2003). Resolution of this question would require fact-intensive analyses of whether this issue was fully and fairly litigated in Repsis or was forfeited in this litigation, among other matters. These gateway issues should be decided before this Court addresses them, especially given that even the dissent acknowledges that one of the preclusion issues raised by the parties is important and undecided, post, at 14, and some of the parties’ other arguments are equally weighty. Unlike the dissent, we do not address these issues in the first instance.

587.US.2018_17-1471

Citibank, N. A., filed a debt-collection action in state court, alleging that respondent Jackson was liable for charges incurred on a Home Depot credit card. As relevant here, Jackson responded by filing third-party class-action claims against petitioner Home Deport U. S. A., Inc., and Carolina Water Systems, Inc., alleging that they had engaged in unlawful referral sales and deceptive and unfair trade practices under state law. Home Depot filed a notice to remove the case from state to federal court, but Jackson moved to remand, arguing that controlling precedent barred removal by a third-party counterclaim defendant. The District Court granted Jackson’s motion, and the Fourth Circuit affirmed, holding that neither the general removal provision, 28 U. S. C. §1441(a), nor the removal provision in the Class Action Fairness Act of 2005, §1453(b), allowed Home Depot to remove the class-action claims filed against it. Held: 1. Section 1441(a) does not permit removal by a third-party counterclaim defendant. Home Depot emphasizes that it is a “defendant” to a “claim,” but §1441(a) refers to “civil action[s],” not “claims.” And because the action as defined by the plaintiff’s complaint is the “civil action . . . of which the district cour[t]” must have “original jurisdiction,” “the defendant” to that action is the defendant to the complaint, not a party named in a counterclaim. This conclusion is bolstered by the use of the term “defendant” in related contexts. For one, the Federal Rules of Civil Procedure differentiate between third-party defendants, counterclaim defendants, and defendants. See, e.g., Rules 14, 12(a)(1)(A)–(B). And in other removal provisions, Congress has clearly extended removal authority to parties other than the original defendant, see, e.g., §§1452(a), 1454(a), (b), but has not done so here. Finally, if, as this Court has held, a counterclaim defendant who was the original plaintiff is not one of “the defendants,” see Shamrock Oil & Gas Corp. v. Sheets, 313 U.S. 100, 106–109, there is no textual reason to reach a different conclusion for a counterclaim defendant who was not part of the initial lawsuit. This reading, Home Depot asserts, runs counter to the history and purposes of removal by preventing a party involuntarily brought into state-court proceedings from removing the claim against it to federal court. But the limits Congress has imposed on removal show that it did not intend to allow all defendants an unqualified right to remove, see, e.g., §1441(b)(2), and Home Depot’s interpretation makes little sense in the context of other removal provisions, see, e.g., §1446(b)(2)(A). Pp. 5–9. 2. Section 1453(b) does not permit removal by a third-party counterclaim defendant. Home Depot contends that even if §1441(a) does not permit removal here, §1453(b) does because it permits removal by “any defendant” to a “class action.” But the two clauses in §1453(b) that employ the term “any defendant” simply clarify that certain limitations on removal that might otherwise apply do not limit removal under that provision. And neither clause—nor anything else in the statute—alters §1441(a)’s limitation on who can remove, suggesting that Congress intended to leave that limit in place. In addition, §§1453(b) and 1441(a) both rely on the procedures for removal in §1446, which also employs the term “defendant.” Interpreting that term to have different meanings in different sections would render the removal provisions incoherent. Pp. 9–11. 880 F.3d 165, affirmed. Thomas, J., delivered the opinion of the Court, in which Ginsburg, Breyer, Sotomayor, and Kagan, JJ., joined. Alito, J., filed a dissenting opinion, in which Roberts, C. J., and Gorsuch and Kavanaugh, JJ., joined.

The general removal statute, 28 U. S. C. §1441(a), provides that “any civil action” over which a federal court would have original jurisdiction may be removed to federal court by “the defendant or the defendants.” The Class Action Fairness Act of 2005 (CAFA) provides that “[a] class action” may be removed to federal court by “any defendant without the consent of all defendants.” 28 U. S. C. §1453(b). In this case, we address whether either provision allows a third-party counterclaim defendant—that is, a party brought into a lawsuit through a counterclaim filed by the original defendant—to remove the counterclaim filed against it. Because in the context of these removal provisions the term “defendant” refers only to the party sued by the original plaintiff, we conclude that neither provision allows such a third party to remove. I A We have often explained that “[f]ederal courts are courts of limited jurisdiction.” Kokkonen v. Guardian Life Ins. Co. of America, 511 U.S. 375, 377 (1994). Article III, §2, of the Constitution delineates “[t]he character of the controversies over which federal judicial authority may extend.” Insurance Corp. of Ireland v. Compagnie des Bauxites de Guinee, 456 U.S. 694, 701 (1982). And lower federal-court jurisdiction “is further limited to those subjects encompassed within a statutory grant of jurisdiction.” Ibid. Accordingly, “the district courts may not exercise jurisdiction absent a statutory basis.” Exxon Mobil Corp. v. Allapattah Services, Inc., 545 U.S. 546, 552 (2005). In 28 U. S. C. §§1331 and 1332(a), Congress granted federal courts jurisdiction over two general types of cases: cases that “aris[e] under” federal law, §1331, and cases in which the amount in controversy exceeds $75,000 and there is diversity of citizenship among the parties, §1332(a). These jurisdictional grants are known as “federal-question jurisdiction” and “diversity jurisdiction,” respectively. Each serves a distinct purpose: Federal-question jurisdiction affords parties a federal forum in which “to vindicate federal rights,” whereas diversity jurisdiction provides “a neutral forum” for parties from different States. Exxon Mobil Corp., supra, at 552. Congress has modified these general grants of jurisdiction to provide federal courts with jurisdiction in certain other types of cases. As relevant here, CAFA provides district courts with jurisdiction over “class action[s]” in which the matter in controversy exceeds $5,000,000 and at least one class member is a citizen of a State different from the defendant. §1332(d)(2)(A). A “class action” is “any civil action filed under Rule 23 of the Federal Rules of Civil Procedure or similar State statute or rule of judicial procedure.” §1332(d)(1)(B). In addition to granting federal courts jurisdiction over certain types of cases, Congress has enacted provisions that permit parties to remove cases originally filed in state court to federal court. Section 1441(a), the general removal statute, permits “the defendant or the defendants” in a state-court action over which the federal courts would have original jurisdiction to remove that action to federal court. To remove under this provision, a party must meet the requirements for removal detailed in other provisions. For one, a defendant cannot remove unilaterally. Instead, “all defendants who have been properly joined and served must join in or consent to the removal of the action.” §1446(b)(2)(A). Moreover, when federal jurisdiction is based on diversity jurisdiction, the case generally must be removed within “1 year after commencement of the action,” §1446(c)(1), and the case may not be removed if any defendant is “a citizen of the State in which such action is brought,” §1441(b)(2). CAFA also includes a removal provision specific to class actions. That provision permits the removal of a “class action” from state court to federal court “by any defendant without the consent of all defendants” and “without regard to whether any defendant is a citizen of the State in which the action is brought.” §1453(b). At issue here is whether the term “defendant” in either §1441(a) or §1453(b) encompasses a party brought into a lawsuit to defend against a counterclaim filed by the original defendant or whether the provisions limit removal authority to the original defendant. B In June 2016, Citibank, N. A., filed a debt-collection action against respondent George Jackson in North Carolina state court. Citibank alleged that Jackson was liable for charges he incurred on a Home Depot credit card. In August 2016, Jackson answered and filed his own claims: an individual counterclaim against Citibank and third-party class-action claims against Home Depot U. S. A., Inc., and Carolina Water Systems, Inc. Jackson’s claims arose out of an alleged scheme between Home Depot and Carolina Water Systems to induce homeowners to buy water treatment systems at inflated prices. The crux of the claims was that Home Depot and Carolina Water Systems engaged in unlawful referral sales and deceptive and unfair trade practices in violation of North Carolina law, Gen. Stat. Ann. §§25A–37, 75–1.1 (2013). Jackson also asserted that Citibank was jointly and severally liable for the conduct of Home Depot and Carolina Water Systems and that his obligations under the sale were null and void. In September 2016, Citibank dismissed its claims against Jackson. One month later, Home Depot filed a notice of removal, citing 28 U. S. C. §§1332, 1441, 1446, and 1453. Jackson moved to remand, arguing that precedent barred removal by a “third-party/additional counter defendant like Home Depot.” App. 51–52. Shortly thereafter, Jackson amended his third-party class-action claims to remove any reference to Citibank. The District Court granted Jackson’s motion to remand, and the Court of Appeals for the Fourth Circuit granted Home Depot permission to appeal and affirmed. 880 F.3d 165, 167 (2018); see 28 U. S. C. §1453(c)(1). Relying on Circuit precedent, it held that neither the general removal provision, §1441(a), nor CAFA’s removal provision, §1453(b), allowed Home Depot to remove the class-action claims filed against it. 880 F. 3d, at 167–171. We granted Home Depot’s petition for a writ of certio- rari to determine whether a third party named in a class-action counterclaim brought by the original defendant can remove if the claim otherwise satisfies the jurisdictional requirements of CAFA. 585 U. S. ___ (2018). We also directed the parties to address whether the holding in Shamrock Oil & Gas Corp. v. Sheets, 313 U.S. 100 (1941)—that an original plaintiff may not remove a coun- terclaim against it—should extend to third-party counterclaim defendants.[1] 585 U. S. ___. II A We first consider whether 28 U. S. C. §1441(a) permits a third-party counterclaim defendant to remove a claim filed against it.[2] Home Depot contends that because a third-party counterclaim defendant is a “defendant” to the claim against it, it may remove pursuant to §1441(a). The dissent agrees, emphasizing that “a ‘defendant’ is a ‘person sued in a civil proceeding.’ ” Post, at 9 (opinion of Alito, J.). This reading of the statute is plausible, but we do not think it is the best one. Of course the term “defendant,” standing alone, is broad. But the phrase “the defendant or the defendants” “cannot be construed in a vacuum.” Davis v. Michigan Dept. of Treasury, 489 U.S. 803, 809 (1989). “It is a fundamental canon of statutory construction that the words of a statute must be read in their context and with a view to their place in the overall statutory scheme.” Ibid.; see also A. Scalia & B. Garner, Reading Law 167 (2012) (“The text must be construed as a whole”); accord, Bailey v. United States, 516 U.S. 137, 145–146 (1995). Considering the phrase “the defendant or the defendants” in light of the structure of the statute and our precedent, we conclude that §1441(a) does not permit removal by any counterclaim defendant, including parties brought into the lawsuit for the first time by the counterclaim.[3] Home Depot emphasizes that it is a “defendant” to a “claim,” but the statute refers to “civil action[s],” not “claims.” This Court has long held that a district court, when determining whether it has original jurisdiction over a civil action, should evaluate whether that action could have been brought originally in federal court. See Mexican Nat. R. Co. v. Davidson, 157 U.S. 201, 208 (1895); Tennessee v. Union & Planters’ Bank, 152 U.S. 454, 461 (1894). This requires a district court to evaluate whether the plaintiff could have filed its operative complaint in federal court, either because it raises claims arising under federal law or because it falls within the court’s diversity jurisdiction. E.g., Franchise Tax Bd. of Cal. v. Construction Laborers Vacation Trust for Southern Cal., 463 U.S. 1, 10 (1983); cf. Holmes Group, Inc. v. Vornado Air Circulation Systems, Inc., 535 U.S. 826, 831 (2002) (“[A] counterclaim . . . cannot serve as the basis for ‘arising under’ jurisdiction”); §1446(c)(2) (deeming the “sum demanded in good faith in the initial pleading . . . the amount in controversy”). Section 1441(a) thus does not permit removal based on counterclaims at all, as a counterclaim is irrelevant to whether the district court had “original jurisdiction” over the civil action. And because the “civil action . . . of which the district cour[t]” must have “original jurisdiction” is the action as defined by the plaintiff’s complaint, “the defendant” to that action is the defendant to that complaint, not a party named in a counterclaim. It is this statutory context, not “the policy goals behind the [well-pleaded complaint] rule,” post, at 23, that underlies our interpretation of the phrase “the defendant or the defendants.” The use of the term “defendant” in related contexts bolsters our determination that Congress did not intend for the phrase “the defendant or the defendants” in §1441(a) to include third-party counterclaim defendants. For one, the Federal Rules of Civil Procedure differentiate between third-party defendants, counterclaim defendants, and defendants. Rule 14, which governs “Third-Party Practice,” distinguishes between “the plaintiff,” a “defendant” who becomes the “third-party plaintiff,” and “the third-party defendant” sued by the original defendant. Rule 12 likewise distinguishes between defendants and counterclaim defendants by separately specifying when “[a] defendant must serve an answer” and when “[a] party must serve an answer to a counterclaim.” Fed. Rules Civ. Proc. 12(a)(1)(A)–(B). Moreover, in other removal provisions, Congress has clearly extended the reach of the statute to include parties other than the original defendant. For instance, §1452(a) permits “[a] party” in a civil action to “remove any claim or cause of action” over which a federal court would have bankruptcy jurisdiction. And §§1454(a) and (b) allow “any party” to remove “[a] civil action in which any party asserts a claim for relief arising under any Act of Congress relating to patents, plant variety protection, or copyrights.” Section 1441(a), by contrast, limits removal to “the defendant or the defendants” in a “civil action” over which the district courts have original jurisdiction. Finally, our decision in Shamrock Oil suggests that third-party counterclaim defendants are not “the defendant or the defendants” who can remove under §1441(a). Shamrock Oil held that a counterclaim defendant who was also the original plaintiff could not remove under §1441(a)’s predecessor statute. 313 U. S., at 106–109. We agree with Home Depot that Shamrock Oil does not specifically address whether a party who was not the original plaintiff can remove a counterclaim filed against it. And we acknowledge, as Home Depot points out, that a third-party counterclaim defendant, unlike the original plaintiff, has no role in selecting the forum for the suit. But the text of §1441(a) simply refers to “the defendant or the defendants” in the civil action. If a counterclaim defendant who was the original plaintiff is not one of “the defendants,” we see no textual reason to reach a different conclusion for a counterclaim defendant who was not originally part of the lawsuit. In that regard, Shamrock Oil did not view the counterclaim as a separate action with a new plaintiff and a new defendant. Instead, the Court highlighted that the original plaintiff was still “the plaintiff.” Id., at 108 (“We can find no basis for saying that Congress, by omitting from the present statute all reference to ‘plaintiffs,’ intended to save a right of removal to some plaintiffs and not to others”). Similarly here, the filing of counterclaims that included class-action allegations against a third party did not create a new “civil action” with a new “plaintiff” and a new “defendant.” Home Depot asserts that reading “the defendant” in §1441(a) to exclude third-party counterclaim defendants runs counter to the history and purposes of removal by preventing a party involuntarily brought into state-court proceedings from removing the claim against it. But the limits Congress has imposed on removal show that it did not intend to allow all defendants an unqualified right to remove. E.g., §1441(b)(2) (preventing removal based on diversity jurisdiction where any defendant is a citizen of the State in which the action is brought). Moreover, Home Depot’s interpretation makes little sense in the context of other removal provisions. For instance, when removal is based on §1441(a), all defendants must consent to removal. See §1446(b)(2)(A). Under Home Depot’s interpretation, “defendants” in §1446(b)(2)(A) could be read to require consent from the third-party counterclaim defendant, the original plaintiff (as a counterclaim defendant), and the original defendant asserting claims against them. Further, Home Depot’s interpretation would require courts to determine when the original defendant is also a “plaintiff” under other statutory provisions. E.g., §1446(c)(1). Instead of venturing down this path, we hold that a third-party counterclaim defendant is not a “defendant” who can remove under §1441(a). B We next consider whether CAFA’s removal provision, §1453(b), permits a third-party counterclaim defendant to remove.[4] Home Depot contends that even if it could not remove under §1441(a), it could remove under §1453(b) because that statute is worded differently. It argues that although §1441(a) permits removal only by “the defendant or the defendants” in a “civil action,” §1453(b) permits removal by “any defendant” to a “class action.” (Emphasis added.) Jackson responds that this argument ignores the context of §1453(b), which he contends makes clear that Congress intended only to alter certain restrictions on removal, not expand the class of parties who can remove a class action. Although this is a closer question, we agree with Jackson. The two clauses in §1453(b) that employ the term “any defendant” simply clarify that certain limitations on removal that might otherwise apply do not limit removal under §1453(b). Section 1453(b) first states that “[a] class action may be removed . . . without regard to whether any defendant is a citizen of the State in which the action is brought.” There is no indication that this language does anything more than alter the general rule that a civil action may not be removed on the basis of diversity jurisdiction “if any of the . . . defendants is a citizen of the State in which such action is brought.” §1441(b)(2). Section 1453(b) then states that “[a] class action . . . may be removed by any defendant without the consent of all defendants.” This language simply amends the rule that “all defendants who have been properly joined and served must join in or consent to the removal of the action.” §1446(b)(2)(A). Rather than indicate that a counterclaim defendant can remove, “here the word ‘any’ is being employed in connection with the word ‘all’ later in the sentence—‘by any . . . without . . . the consent of all.’ ” Westwood Apex v. Contreras, 644 F.3d 799, 804 (CA9 2011); see Palisades Collections LLC v. Shorts, 552 F.3d 327, 335–336 (CA4 2008). Neither clause—nor anything else in the statute—alters §1441(a)’s limitation on who can remove, which suggests that Congress intended to leave that limit in place. See supra, at 5–8. Thus, although the term “any” ordinarily carries an “ ‘expansive meaning,’ ” post, at 10, the context here demonstrates that Congress did not expand the types of parties eligible to remove a class action under §1453(b) beyond §1441(a)’s limits. If anything, that the language of §1453(b) mirrors the language in the statutory provisions it is amending suggests that the term “defendant” is being used consistently across all provisions. Cf. Mississippi ex rel. Hood v. AU Optronics Corp., 571 U.S. 161, 169–170 (2014) (interpreting CAFA consistently with Rule 20 where Congress used terms in a like manner in both provisions). To the extent Home Depot is arguing that the term “defendant” has a different meaning in §1453(b) than it does in §1441(a), we reject its interpretation. Because §§1453(b) and 1441(a) both rely on the procedures for removal in §1446, which also employs the term “defendant,” interpreting “defendant” to have different meanings in different sections would render the removal provisions incoherent. See First Bank v. DJL Properties, LLC, 598 F.3d 915, 917 (CA7 2010) (Easterbrook, C. J.). Interpreting the removal provisions together, we determine that §1453(b), like §1441(a), does not permit a third-party counterclaim defendant to remove. Finally, the dissent argues that our interpretation allows defendants to use the statute as a “tactic” to prevent removal, post, at 7, but that result is a consequence of the statute Congress wrote. Of course, if Congress shares the dissent’s disapproval of certain litigation “tactics,” it certainly has the authority to amend the statute. But we do not. * * * Because neither §1441(a) nor §1453(b) permits removal by a third-party counterclaim defendant, Home Depot could not remove the class-action claim filed against it. Accordingly, we affirm the judgment of the Fourth Circuit. It is so ordered. Notes 1 In this opinion, we use the term “third-party counterclaim defendant” to refer to a party first brought into the case as an additional defendant to a counterclaim asserted against the original plaintiff. 2 Section 1441(a) provides that “any civil action brought in a State court of which the district courts of the United States have original jurisdiction, may be removed by the defendant or the defendants, to the district court of the United States for the district and division embracing the place where such action is pending.” 3 Even the dissent declines to rely on the dictionary definition of “defendant” alone, as following that approach to its logical conclusion would require overruling Shamrock Oil & Gas Corp. v. Sheets, 313 U.S. 100 (1941). See post, at 10, n. 2. 4 Section 1453(b) provides that “[a] class action may be removed to a district court of the United States in accordance with section 1446 (except that the 1-year limitation under section 1446(c)(1) shall not apply), without regard to whether any defendant is a citizen of the State in which the action is brought, except that such action may be removed by any defendant without the consent of all defendants.”

588.US.2018_18-302

Respondent Erik Brunetti sought federal registration of the trademark FUCT. The Patent and Trademark Office (PTO) denied his application under a provision of the Lanham Act that prohibits registration of trademarks that “[c]onsist[ ] of or comprise[ ] immoral[ ] or scandalous matter,” 15 U. S. C. §1052(a). Brunetti brought a First Amendment challenge to the “immoral or scandalous” bar in the Federal Circuit, which invalidated the provision. Held: The Lanham Act’s prohibition on registration of “immoral[ ] or scandalous” trademarks violates the First Amendment. In Matal v. Tam, 582 U. S. ___, this Court declared unconstitutional the Lanham Act’s ban on registering marks that “disparage” any “person[ ], living or dead.” §1052(a). A divided Court agreed on two propositions. First, if a trademark registration bar is viewpoint based, it is unconstitutional. And second, the disparagement bar was viewpoint based. The “immoral or scandalous” bar similarly discriminates on the basis of viewpoint and so collides with this Court’s First Amendment doctrine. Expressive material is “immoral” when it is “inconsistent with rectitude, purity, or good morals”; “wicked”; or “vicious.” So the Lanham Act permits registration of marks that champion society’s sense of rectitude and morality, but not marks that denigrate those concepts. And material is “scandalous” when it “giv[es] offense to the conscience or moral feelings”; “excite[s] reprobation”; or “call[s] out condemnation.” So the Lanham Act allows registration of marks when their messages accord with, but not when their messages defy, society’s sense of decency or propriety. The statute, on its face, distinguishes between two opposed sets of ideas: those aligned with conventional moral standards and those hostile to them; those inducing societal nods of approval and those provoking offense and condemnation. This facial viewpoint bias in the law results in viewpoint-discriminatory application. The PTO has refused to register marks communicating “immoral” or “scandalous” views about (among other things) drug use, religion, and terrorism. But all the while, it has approved registration of marks expressing more accepted views on the same topics. The Government says the statute is susceptible of a limiting construction that would remove its viewpoint bias. The Government’s idea is to narrow the statutory bar to “marks that are offensive [or] shocking[ ] because of their mode of expression, independent of any views that they may express,” which would mostly restrict the PTO to refusing marks that are lewd, sexually explicit, or profane. But this Court cannot accept the Government’s proposal, because the statute says something markedly different. The “immoral or scandalous” bar does not draw the line at lewd, sexually explicit, or profane marks. Nor does it refer only to marks whose “mode of expression,” independent of viewpoint, is particularly offensive. To cut the statute off where the Government urges is not to interpret the statute Congress enacted, but to fashion a new one. And once the “immoral or scandalous” bar is interpreted fairly, it must be invalidated. Pp. 4–11. 877 F.3d 1330, affirmed. Kagan, J., delivered the opinion of the Court, in which Thomas, Ginsburg, Alito, Gorsuch and Kavanaugh, JJ., joined. Alito, J., filed a concurring opinion. Roberts, C. J., and Breyer, J., filed opinions concurring in part and dissenting in part. Sotomayor, J., filed an opinion concurring in part and dissenting in part, in which Breyer, J., joined.

Two Terms ago, in Matal v. Tam, 582 U. S. ___ (2017), this Court invalidated the Lanham Act’s bar on the registration of “disparag[ing]” trademarks. 15 U. S. C. §1052(a). Although split between two non-majority opinions, all Members of the Court agreed that the provision violated the First Amendment because it discriminated on the basis of viewpoint. Today we consider a First Amendment challenge to a neighboring provision of the Act, prohibiting the registration of “immoral[ ] or scandalous” trademarks. Ibid. We hold that this provision infringes the First Amendment for the same reason: It too disfavors certain ideas. I Respondent Erik Brunetti is an artist and entrepreneur who founded a clothing line that uses the trademark FUCT. According to Brunetti, the mark (which functions as the clothing’s brand name) is pronounced as four letters, one after the other: F-U-C-T. See Brief for Respondent 1. But you might read it differently and, if so, you would hardly be alone. See Tr. of Oral Arg. 5 (describing the brand name as “the equivalent of [the] past participle form of a well-known word of profanity”). That common perception caused difficulties for Brunetti when he tried to register his mark with the U. S. Patent and Trademark Office (PTO). Under the Lanham Act, the PTO administers a federal registration system for trademarks. See 15 U. S. C. §§1051, 1052. Registration of a mark is not mandatory. The owner of an unregistered mark may still use it in commerce and enforce it against infringers. See Tam, 582 U. S., at ___ (slip op., at 4). But registration gives trademark owners valuable benefits. For example, registration constitutes “prima facie evidence” of the mark’s validity. §1115(a). And registration serves as “constructive notice of the registrant’s claim of ownership,” which forecloses some defenses in infringement actions. §1072. Generally, a trademark is eligible for registration, and receipt of such benefits, if it is “used in commerce.” §1051(a)(1). But the Act directs the PTO to “refuse[ ] registration” of certain marks. §1052. For instance, the PTO cannot register a mark that “so resembles” another mark as to create a likelihood of confusion. §1052(d). It cannot register a mark that is “merely descriptive” of the goods on which it is used. §1052(e). It cannot register a mark containing the flag or insignia of any nation or State. See §1052(b). There are five or ten more (depending on how you count). And until we invalidated the criterion two years ago, the PTO could not register a mark that “disparage[d]” a “person[ ], living or dead.” §1052(a); see Tam, 582 U. S. ___. This case involves another of the Lanham Act’s prohibitions on registration—one applying to marks that “[c]onsist[ ] of or comprise[ ] immoral[ ] or scandalous matter.” §1052(a). The PTO applies that bar as a “unitary provision,” rather than treating the two adjectives in it separately. In re Brunetti, 877 F.3d 1330, 1336 (CA Fed. 2017); Brief for Petitioner 6 (stating that the PTO “has long treated the two terms as composing a single category”). To determine whether a mark fits in the category, the PTO asks whether a “substantial composite of the general public” would find the mark “shocking to the sense of truth, decency, or propriety”; “giving offense to the conscience or moral feelings”; “calling out for condemnation”; “disgraceful”; “offensive”; “disreputable”; or “vulgar.” 877 F. 3d, at 1336 (internal quotation marks omitted); see Brief for Petitioner 6 (agreeing that the PTO “generally defines” the category in that way). Both a PTO examining attorney and the PTO’s Trademark Trial and Appeal Board decided that Brunetti’s mark flunked that test. The attorney determined that FUCT was “a total vulgar” and “therefore[ ] unregistrable.” App. 27–28. On review, the Board stated that the mark was “highly offensive” and “vulgar,” and that it had “decidedly negative sexual connotations.” App. to Pet. for Cert. 59a, 64a–65a. As part of its review, the Board also considered evidence of how Brunetti used the mark. It found that Brunetti’s website and products contained imagery, near the mark, of “extreme nihilism” and “anti-social” behavior. Id., at 64a. In that context, the Board thought, the mark communicated “misogyny, depravity, [and] violence.” Ibid. The Board concluded: “Whether one considers [the mark] as a sexual term, or finds that [Brunetti] has used [the mark] in the context of extreme misogyny, nihilism or violence, we have no question but that [the term is] extremely offensive.” Id., at 65a. Brunetti then brought a facial challenge to the “immoral or scandalous” bar in the Court of Appeals for the Federal Circuit. That court found the prohibition to violate the First Amendment. As usual when a lower court has invalidated a federal statute, we granted certiorari. 586 U. S. ___ (2019). II This Court first considered a First Amendment challenge to a trademark registration restriction in Tam, just two Terms ago. There, the Court declared unconstitutional the Lanham Act’s ban on registering marks that “disparage” any “person[ ], living or dead.” §1052(a). The eight-Justice Court divided evenly between two opinions and could not agree on the overall framework for deciding the case. (In particular, no majority emerged to resolve whether a Lanham Act bar is a condition on a government benefit or a simple restriction on speech.) But all the Justices agreed on two propositions. First, if a trademark registration bar is viewpoint-based, it is unconstitutional. See 582 U. S., at ___–___, ___–___ (opinion of Alito, J.) (slip op., at 1–2, 22–23); id., at ___–___, ___ (opinion of Kennedy, J.) (slip op., at 1–2, 5). And second, the disparagement bar was viewpoint-based. See id., at ___–___, ___–___ (opinion of Alito, J.) (slip op., at 1–2, 22–23); id., at ___–___ (opinion of Kennedy, J.) (slip op., at 2–5). The Justices thus found common ground in a core postulate of free speech law: The government may not discriminate against speech based on the ideas or opinions it conveys. See Rosenberger v. Rector and Visitors of Univ. of Va., 515 U.S. 819, 829–830 (1995) (explaining that viewpoint discrimination is an “egregious form of content discrimination” and is “presumptively unconstitutional”). In Justice Kennedy’s explanation, the disparagement bar allowed a trademark owner to register a mark if it was “positive” about a person, but not if it was “derogatory.” Tam, 582 U. S., at ___ (slip op., at 2). That was the “essence of viewpoint discrimination,” he continued, because “[t]he law thus reflects the Government’s disapproval of a subset of messages it finds offensive.” Id., at ___–___ (slip op., at 2–3). Justice Alito emphasized that the statute “denie[d] registration to any mark” whose disparaging message was “offensive to a substantial percentage of the members of any group.” Id., at ___ (slip op., at 22). The bar thus violated the “bedrock First Amendment principle” that the government cannot discriminate against “ideas that offend.” Id., at ___–___ (slip op., at 1–2). Slightly different explanations, then, but a shared conclusion: Viewpoint discrimination doomed the disparagement bar. If the “immoral or scandalous” bar similarly discriminates on the basis of viewpoint, it must also collide with our First Amendment doctrine. The Government does not argue otherwise. In briefs and oral argument, the Government offers a theory for upholding the bar if it is viewpoint-neutral (essentially, that the bar would then be a reasonable condition on a government benefit). See Brief for Petitioner 14–26. But the Government agrees that under Tam it may not “deny registration based on the views expressed” by a mark. Tr. of Oral Arg. 24. “As the Court’s Tam decision establishes,” the Government says, “the criteria for federal trademark registration” must be “viewpoint-neutral to survive Free Speech Clause review.” Pet. for Cert. 19. So the key question becomes: Is the “immoral or scandalous” criterion in the Lanham Act viewpoint-neutral or viewpoint-based? It is viewpoint-based. The meanings of “immoral” and “scandalous” are not mysterious, but resort to some dictionaries still helps to lay bare the problem. When is expressive material “immoral”? According to a standard definition, when it is “inconsistent with rectitude, purity, or good morals”; “wicked”; or “vicious.” Webster’s New International Dictionary 1246 (2d ed. 1949). Or again, when it is “opposed to or violating morality”; or “morally evil.” Shorter Oxford English Dictionary 961 (3d ed. 1947). So the Lanham Act permits registration of marks that champion society’s sense of rectitude and morality, but not marks that denigrate those concepts. And when is such material “scandalous”? Says a typical definition, when it “giv[es] offense to the conscience or moral feelings”; “excite[s] reprobation”; or “call[s] out condemnation.” Webster’s New International Dictionary, at 2229. Or again, when it is “shocking to the sense of truth, decency, or propriety”; “disgraceful”; “offensive”; or “disreputable.” Funk & Wagnalls New Standard Dictionary 2186 (1944). So the Lanham Act allows registration of marks when their messages accord with, but not when their messages defy, society’s sense of decency or propriety. Put the pair of overlapping terms together and the statute, on its face, distinguishes between two opposed sets of ideas: those aligned with conventional moral standards and those hostile to them; those inducing societal nods of approval and those provoking offense and condemnation. The statute favors the former, and disfavors the latter. “Love rules”? “Always be good”? Registration follows. “Hate rules”? “Always be cruel”? Not according to the Lanham Act’s “immoral or scandalous” bar. The facial viewpoint bias in the law results in viewpoint-discriminatory application. Recall that the PTO itself describes the “immoral or scandalous” criterion using much the same language as in the dictionary definitions recited above. See supra, at 3. The PTO, for example, asks whether the public would view the mark as “shocking to the sense of truth, decency, or propriety”; “calling out for condemnation”; “offensive”; or “disrepu- table.” Brief for Petitioner 6 (internal quotation marks omitted). Using those guideposts, the PTO has refused to register marks communicating “immoral” or “scandalous” views about (among other things) drug use, religion, and terrorism. But all the while, it has approved registration of marks expressing more accepted views on the same topics. See generally Gilson & LaLonde, Trademarks Laid Bare, 101 Trademark Reporter 1476, 1510–1513, 1518–1522 (2011); Brief for Barton Beebe et al. as Amici Curiae 28–29. Here are some samples. The PTO rejected marks conveying approval of drug use (YOU CAN’T SPELL HEALTHCARE WITHOUT THC for pain-relief medication, MARIJUANA COLA and KO KANE for beverages) because it is scandalous to “inappropriately glamoriz[e] drug abuse.” PTO, Office Action of Aug. 28, 2010, Serial No. 85038867; see Office Action of Dec. 24, 2009, Serial No. 77833964; Office Action of Nov. 17, 2009, Serial No. 77671304. But at the same time, the PTO registered marks with such sayings as D.A.R.E. TO RESIST DRUGS AND VIOLENCE and SAY NO TO DRUGS—REALITY IS THE BEST TRIP IN LIFE. See PTO, Reg. No. 2975163 (July 26, 2005); Reg. No. 2966019 (July 12, 2005). Similarly, the PTO disapproved registration for the mark BONG HITS 4 JESUS because it “suggests that people should engage in an illegal activity [in connection with] worship” and because “Christians would be morally outraged by a statement that connects Jesus Christ with illegal drug use.” Office Action of Mar. 15, 2008, Serial No. 77305946. And the PTO refused to register trademarks associating religious references with products (AGNUS DEI for safes and MADONNA for wine) because they would be “offensive to most individuals of the Christian faith” and “shocking to the sense of propriety.” Ex parte Summit Brass & Bronze Works, 59 USPQ 22, 23 (Dec. Com. Pat. 1943); In re Riverbank Canning Co., 95 F.2d 327, 329 (CCPA 1938). But once again, the PTO approved marks—PRAISE THE LORD for a game and JESUS DIED FOR YOU on clothing—whose message suggested religious faith rather than blasphemy or irreverence. See Reg. No. 5265121 (Aug. 15, 2017); Reg. No. 3187985 (Dec. 19, 2006). Finally, the PTO rejected marks reflecting support for al-Qaeda (BABY AL QAEDA and AL-QAEDA on t-shirts) “because the bombing of civilians and other terrorist acts are shocking to the sense of decency and call out for condemnation.” Office Action of Nov. 22, 2004, Serial No. 78444968; see Office Action of Feb. 23, 2005, Serial No. 78400213. Yet it approved registration of a mark with the words WAR ON TERROR MEMORIAL. Reg. No. 5495362 (Jun. 19, 2018). Of course, all these decisions are understandable. The rejected marks express opinions that are, at the least, offensive to many Americans. But as the Court made clear in Tam, a law disfavoring “ideas that offend” discriminates based on viewpoint, in violation of the First Amendment. 582 U. S., at ___ (opinion of Alito, J.) (slip op., at 2); see id., at ___–___ (slip op., at 22–23); id., at ___–___ (opinion of Kennedy, J.) (slip op., at 2–3). How, then, can the Government claim that the “immoral or scandalous” bar is viewpoint-neutral? The Government basically asks us to treat decisions like those described above as PTO examiners’ mistakes. See Brief for Petitioner 46. Still more, the Government tells us to ignore how the Lanham Act’s language, on its face, disfavors some ideas. In urging that course, the Government does not dispute that the statutory language—and words used to define it—have just that effect. At oral argument, the Government conceded: “[I]f you just looked at the words like ‘shocking’ and ‘offensive’ on their face and gave them their ordinary meanings[,] they could easily encompass material that was shocking [or offensive] because it expressed an outrageous point of view or a point of view that most members” of society reject. Tr. of Oral Arg. 6. But no matter, says the Government, because the statute is “susceptible of” a limiting construction that would remove this viewpoint bias. Id., at 7 (arguing that the Court should “attempt to construe [the] statute in a way that would render it constitutional”). The Government’s idea, abstractly phrased, is to narrow the statutory bar to “marks that are offensive [or] shocking to a substantial segment of the public because of their mode of expression, independent of any views that they may express.” Id., at 11 (emphasis added); see Brief for Petitioner 27–28. More concretely, the Government explains that this reinterpretation would mostly restrict the PTO to refusing marks that are “vulgar”—meaning “lewd,” “sexually explicit or profane.” Id., at 27, 30. Such a reconfigured bar, the Government says, would not turn on viewpoint, and so we could uphold it. But we cannot accept the Government’s proposal, because the statute says something markedly different. This Court, of course, may interpret “ambiguous statutory language” to “avoid serious constitutional doubts.” FCC v. Fox Television Stations, Inc., 556 U.S. 502, 516 (2009). But that canon of construction applies only when ambiguity exists. “We will not rewrite a law to conform it to constitutional requirements.” United States v. Stevens, 559 U.S. 460, 481 (2010) (internal quotation marks and alteration omitted). So even assuming the Government’s reading would eliminate First Amendment problems, we may adopt it only if we can see it in the statutory language. And we cannot. The “immoral or scandalous” bar stretches far beyond the Government’s proposed construction. The statute as written does not draw the line at lewd, sexually explicit, or profane marks. Nor does it refer only to marks whose “mode of expression,” independent of viewpoint, is particularly offensive. Brief for Petitioner 28 (internal quotation marks omitted). It covers the universe of immoral or scandalous—or (to use some PTO synonyms) offensive or disreputable—material. Whether or not lewd or profane. Whether the scandal and immorality comes from mode or instead from viewpoint. To cut the statute off where the Government urges is not to interpret the statute Congress enacted, but to fashion a new one.[1]* And once the “immoral or scandalous” bar is interpreted fairly, it must be invalidated. The Government just barely argues otherwise. In the last paragraph of its brief, the Government gestures toward the idea that the provision is salvageable by virtue of its constitutionally permissible applications (in the Government’s view, its applications to lewd, sexually explicit, or profane marks). See id., at 47. In other words, the Government invokes our First Amendment overbreadth doctrine, and asks us to uphold the statute against facial attack because its unconstitutional applications are not “substantial” relative to “the statute’s plainly legitimate sweep.” Stevens, 559 U. S., at 473 (internal quotation marks omitted). But to begin with, this Court has never applied that kind of analysis to a viewpoint-discriminatory law. In Tam, for example, we did not pause to consider whether the disparagement clause might admit some permissible applications (say, to certain libelous speech) before striking it down. The Court’s finding of viewpoint bias ended the matter. And similarly, it seems unlikely we would compare permissible and impermissible applications if Congress outright banned “offensive” (or to use some other examples, “divisive” or “subversive”) speech. Once we have found that a law “aim[s] at the suppression of ” views, why would it matter that Congress could have captured some of the same speech through a viewpoint-neutral statute? Tam, 582 U. S., at ___ (opinion of Kennedy, J.) (slip op., at 2). But in any event, the “immoral or scandalous” bar is substantially overbroad. There are a great many immoral and scandalous ideas in the world (even more than there are swearwords), and the Lanham Act covers them all. It therefore violates the First Amendment. We accordingly affirm the judgment of the Court of Appeals. It is so ordered. Notes 1 *We reject the dissent’s statutory surgery for the same reason. Al-though conceding that the term “immoral” cannot be saved, the dissent thinks that the term “scandalous” can be read as the Government proposes. See post, at 1–2 (Sotomayor, J., concurring in part and dissenting in part). But that term is not “ambiguous,” as the dissent argues, post, at 3; it is just broad. Remember that the dictionaries define it to mean offensive, disreputable, exciting reprobation, and so forth. See supra, at 5–6; post, at 3 (accepting those definitions). Even if hived off from “immoral” marks, the category of scandalous marks thus includes both marks that offend by the ideas they convey and marks that offend by their mode of expression. And its coverage of the former means that it discriminates based on viewpoint. We say nothing at all about a statute that covers only the latter—or, in the Government’s more concrete description, a statute limited to lewd, sexually explicit, and profane marks. Nor do we say anything about how to evaluate viewpoint-neutral restrictions on trademark registration, see post, at 14–17—because the “scandalous” bar (whether or not attached to the “immoral” bar) is not one.

586.US.2018_17-1011

In 1945, Congress passed the International Organizations Immunities Act (IOIA), which, among other things, grants international organizations the “same immunity from suit . . . as is enjoyed by foreign governments.” 22 U. S. C. §288a(b). At that time, foreign governments were entitled to virtually absolute immunity as a matter of international grace and comity. In 1952, the State Department adopted a more restrictive theory of foreign sovereign immunity, which Congress subsequently codified in the Foreign Sovereign Immunities Act (FSIA), 28 U. S. C. §1602. The FSIA gives foreign sovereign governments presumptive immunity from suit, §1604, subject to several statutory exceptions, including, as relevant here, an exception for actions based on commercial activity with a sufficient nexus with the United States, §1605(a)(2). Respondent International Finance Corporation (IFC), an IOIA international organization, entered into a loan agreement with Coastal Gujarat Power Limited, a company based in India, to finance the construction of a coal-fired power plant in Gujarat. Petitioners sued the IFC, claiming that pollution from the plant harmed the surrounding air, land, and water. The District Court, however, held that the IFC was immune from suit because it enjoyed the virtually absolute immunity that foreign governments enjoyed when the IOIA was enacted. The D. C. Circuit affirmed in light of its decision in Atkinson v. Inter-American Development Bank, 156 F.3d 1335. Held: The IOIA affords international organizations the same immunity from suit that foreign governments enjoy today under the FSIA. Pp. 6–15. (a) The IOIA “same as” formulation is best understood as making international organization immunity and foreign sovereign immunity continuously equivalent. The IOIA is thus like other statutes that use similar or identical language to place two groups on equal footing. See, e.g., Civil Rights Act of 1866, 42 U. S. C. §§1981(a), 1982; Federal Tort Claims Act, 28 U. S. C. §2674. Whatever the ultimate purpose of international organization immunity may be, the immediate purpose of the IOIA immunity provision is expressed in language that Congress typically uses to make one thing continuously equivalent to another. Pp. 6–9. (b) That reading is confirmed by the “reference canon” of statutory interpretation. When a statute refers to a general subject, the statute adopts the law on that subject as it exists whenever a question under the statute arises. In contrast, when a statute refers to another statute by specific title, the referenced statute is adopted as it existed when the referring statute was enacted, without any subsequent amendments. Federal courts have often relied on the reference canon to harmonize a statute with an external body of law that the statute refers to generally. The IOIA’s reference to the immunity enjoyed by foreign governments is to an external body of potentially evolving law, not to a specific provision of another statute. Nor is it a specific reference to a common law concept with a fixed meaning. The phrase “immunity enjoyed by foreign governments” is not a term of art with substantive content but rather a concept that can be given scope and content only by reference to the rules governing foreign sovereign immunity. Pp. 9–11. (c) The D. C. Circuit relied upon Atkinson’s conclusion that the reference canon’s probative force was outweighed by an IOIA provision authorizing the President to alter the immunity of an international organization. But the fact that the President has power to modify otherwise applicable immunity rules is perfectly compatible with the notion that those rules might themselves change over time in light of developments in the law governing foreign sovereign immunity. The Atkinson court also did not consider the opinion of the State Department, whose views in this area ordinarily receive “special attention,” Bolivarian Republic of Venezuela v. Helmerich & Payne Int’l Drilling Co., 581 U. S. ___, ___, and which took the position that immunity rules of the IOIA and the FSIA were linked following the FSIA’s enactment. Pp. 11–13. (d) The IFC contends that interpreting the IOIA immunity provision to grant only restrictive immunity would defeat the purpose of granting immunity in the first place, by subjecting international organizations to suit under the commercial activity exception of the FSIA for most or all of their core activities. This would be particularly true with respect to international development banks, which use the tools of commerce to achieve their objectives. Those concerns are inflated. The IOIA provides only default rules. An international organization’s charter can always specify a different level of immunity, and many do. Nor is it clear that the lending activity of all development banks qualifies as commercial activity within the meaning of the FSIA. But even if it does qualify as commercial, that does not mean the organization is automatically subject to suit, since other FSIA requirements must also be met, see, e.g., 28 U. S. C. §§1603, 1605(a)(2). Pp. 13–15. 860 F.3d 703, reversed and remanded. Roberts, C. J., delivered the opinion of the Court, in which Thomas, Ginsburg, Alito, Sotomayor, Kagan, and Gorsuch, JJ., joined. Breyer, J., filed a dissenting opinion. Kavanaugh, J., took no part in the consideration or decision of the case.

The International Organizations Immunities Act of 1945 grants international organizations such as the World Bank and the World Health Organization the “same immunity from suit . . . as is enjoyed by foreign governments.” 22 U. S. C. §288a(b). At the time the IOIA was enacted, foreign governments enjoyed virtually absolute immunity from suit. Today that immunity is more limited. Most significantly, foreign governments are not immune from actions based upon certain kinds of commercial activity in which they engage. This case requires us to determine whether the IOIA grants international organizations the virtually absolute immunity foreign governments enjoyed when the IOIA was enacted, or the more limited immunity they enjoy today. Respondent International Finance Corporation is an international organization headquartered in the United States. The IFC finances private-sector development projects in poor and developing countries around the world. About 10 years ago, the IFC financed the construction of a power plant in Gujarat, India. Petitioners are local farmers and fishermen and a small village. They allege that the power plant has polluted the air, land, and water in the surrounding area. Petitioners sued the IFC for damages and injunctive relief in Federal District Court, but the IFC claimed absolute immunity from suit. Petitioners argued that the IFC was entitled under the IOIA only to the limited or “restrictive” immunity that foreign governments currently enjoy. We agree. I A In the wake of World War II, the United States and many of its allies joined together to establish a host of new international organizations. Those organizations, which included the United Nations, the International Monetary Fund, and the World Bank, were designed to allow member countries to collectively pursue goals such as stabilizing the international economy, rebuilding war-torn nations, and maintaining international peace and security. Anticipating that those and other international organizations would locate their headquarters in the United States, Congress passed the International Organizations Immunities Act of 1945, 59Stat. 669. The Act grants international organizations a set of privileges and immunities, such as immunity from search and exemption from property taxes. 22 U. S. C. §§288a(c), 288c. The IOIA defines certain privileges and immunities by reference to comparable privileges and immunities enjoyed by foreign governments. For example, with respect to customs duties and the treatment of official communications, the Act grants international organizations the privileges and immunities that are “accorded under similar circumstances to foreign governments.” §288a(d). The provision at issue in this case provides that international organizations “shall enjoy the same immunity from suit and every form of judicial process as is enjoyed by foreign governments.” §288a(b). The IOIA authorizes the President to withhold, withdraw, condition, or limit the privileges and immunities it grants in light of the functions performed by any given international organization. §288. Those privileges and immunities can also be expanded or restricted by a particular organization’s founding charter. B When the IOIA was enacted in 1945, courts looked to the views of the Department of State in deciding whether a given foreign government should be granted immunity from a particular suit. If the Department submitted a recommendation on immunity, courts deferred to the recommendation. If the Department did not make a recommendation, courts decided for themselves whether to grant immunity, although they did so by reference to State Department policy. Samantar v. Yousuf, 560 U.S. 305, 311–312 (2010). Until 1952, the State Department adhered to the classical theory of foreign sovereign immunity. According to that theory, foreign governments are entitled to “virtually absolute” immunity as a matter of international grace and comity. At the time the IOIA was enacted, therefore, the Department ordinarily requested, and courts ordinarily granted, immunity in suits against foreign governments. Ibid.; Verlinden B. V. v. Central Bank of Nigeria, 461 U.S. 480, 486 (1983).[1] In 1952, however, the State Department announced that it would adopt the newer “restrictive” theory of foreign sovereign immunity. Under that theory, foreign governments are entitled to immunity only with respect to their sovereign acts, not with respect to commercial acts. The State Department explained that it was adopting the restrictive theory because the “widespread and increasing practice on the part of governments of engaging in commercial activities” made it “necessary” to “enable persons doing business with them to have their rights determined in the courts.” Letter from Jack B. Tate, Acting Legal Adviser, Dept. of State, to Acting Attorney General Philip B. Perlman (May 19, 1952), reprinted in 26 Dept. State Bull. 984–985 (1952). In 1976, Congress passed the Foreign Sovereign Immunities Act. The FSIA codified the restrictive theory of foreign sovereign immunity but transferred “primary responsibility for immunity determinations from the Executive to the Judicial Branch.” Republic of Austria v. Altmann, 541 U.S. 677, 691 (2004); see 28 U. S. C. §1602. Under the FSIA, foreign governments are presumptively immune from suit. §1604. But a foreign government may be subject to suit under one of several statutory exceptions. Most pertinent here, a foreign government may be subject to suit in connection with its commercial activity that has a sufficient nexus with the United States. §1605(a)(2). C The International Finance Corporation is an international development bank headquartered in Washington, D. C. The IFC is designated as an international organization under the IOIA. Exec. Order No. 10680, 3 CFR 86 (1957); see 22 U. S. C. §§282, 288. One hundred eighty-four countries, including the United States, are members of the IFC. The IFC is charged with furthering economic development “by encouraging the growth of productive private enterprise in member countries, particularly in the less developed areas, thus supplementing the activities of” the World Bank. Articles of Agreement of the International Finance Corporation, Art. I, Dec. 5, 1955, 7 U. S. T. 2193, T. I. A. S. No. 3620. Whereas the World Bank primarily provides loans and grants to developing countries for public-sector projects, the IFC finances private-sector development projects that cannot otherwise attract capital on reasonable terms. See Art. I(i), ibid. In 2018, the IFC provided some $23 billion in such financing. The IFC expects its loan recipients to adhere to a set of performance standards designed to “avoid, mitigate, and manage risks and impacts” associated with development projects. IFC Performance Standards on Environmental and Social Sustainability, Jan. 1, 2012, p. 2, ¶1. Those standards are usually more stringent than any established by local law. The IFC includes the standards in its loan agreements and enforces them through an internal review process. Brief for Respondent 10. In 2008, the IFC loaned $450 million to Coastal Gujarat Power Limited, a company located in India. The loan helped finance the construction of a coal-fired power plant in the state of Gujarat. Under the terms of the loan agreement, Coastal Gujarat was required to comply with an environmental and social action plan designed to protect areas around the plant from damage. The agreement allowed the IFC to revoke financial support for the project if Coastal Gujarat failed to abide by the terms of the agreement. The project did not go smoothly. According to the IFC’s internal audit, Coastal Gujarat did not comply with the environmental and social action plan in constructing and operating the plant. The audit report criticized the IFC for inadequately supervising the project. In 2015, a group of farmers and fishermen who live near the plant, as well as a local village, sued the IFC in the United States District Court for the District of Columbia. They claimed that pollution from the plant, such as coal dust, ash, and water from the plant’s cooling system, had destroyed or contaminated much of the surrounding air, land, and water. Relying on the audit report, they asserted several causes of action against the IFC, including negligence, nuisance, trespass, and breach of contract. The IFC maintained that it was immune from suit under the IOIA and moved to dismiss for lack of subject matter jurisdiction. The District Court, applying D. C. Circuit precedent, concluded that the IFC was immune from suit because the IOIA grants international organizations the virtually absolute immunity that foreign governments enjoyed when the IOIA was enacted. 172 F. Supp. 3d 104, 108–109 (DC 2016) (citing Atkinson v. Inter-American Development Bank, 156 F.3d 1335 (CADC 1998)). The D. C. Circuit affirmed in light of its precedent. 860 F.3d 703 (2017). Judge Pillard wrote separately to say that she would have decided the question differently were she writing on a clean slate. Id., at 708 (concurring opinion). Judge Pillard explained that she thought the D. C. Circuit “took a wrong turn” when it “read the IOIA to grant international organizations a static, absolute immunity that is, by now, not at all the same ‘as is enjoyed by foreign governments,’ but substantially broader.” Ibid. Judge Pillard also noted that the Third Circuit had expressly declined to follow the D. C. Circuit’s approach. See OSS Nokalva, Inc. v. European Space Agency, 617 F.3d 756 (CA3 2010). We granted certiorari. 584 U. S. ___ (2018). II The IFC contends that the IOIA grants international organizations the “same immunity” from suit that foreign governments enjoyed in 1945. Petitioners argue that it instead grants international organizations the “same immunity” from suit that foreign governments enjoy today. We think petitioners have the better reading of the statute. A The language of the IOIA more naturally lends itself to petitioners’ reading. In granting international organizations the “same immunity” from suit “as is enjoyed by foreign governments,” the Act seems to continuously link the immunity of international organizations to that of foreign governments, so as to ensure ongoing parity between the two. The statute could otherwise have simply stated that international organizations “shall enjoy absolute immunity from suit,” or specified some other fixed level of immunity. Other provisions of the IOIA, such as the one making the property and assets of international organizations “immune from search,” use such noncomparative language to define immunities in a static way. 22 U. S. C. §288a(c). Or the statute could have specified that it was incorporating the law of foreign sovereign immunity as it existed on a particular date. See, e.g., Energy Policy Act of 1992, 30 U. S. C. §242(c)(1) (certain land patents “shall provide for surface use to the same extent as is provided under applicable law prior to October 24, 1992”). Because the IOIA does neither of those things, we think the “same as” formulation is best understood to make international organization immunity and foreign sovereign immunity continuously equivalent. That reading finds support in other statutes that use similar or identical language to place two groups on equal footing. In the Civil Rights Act of 1866, for instance, Congress established a rule of equal treatment for newly freed slaves by giving them the “same right” to make and enforce contracts and to buy and sell property “as is enjoyed by white citizens.” 42 U. S. C. §§1981(a), 1982. That provision is of course understood to guarantee continuous equality between white and nonwhite citizens with respect to the rights in question. See Jones v. Alfred H. Mayer Co., 392 U.S. 409, 427–430 (1968). Similarly, the Federal Tort Claims Act states that the “United States shall be liable” in tort “in the same manner and to the same extent as a private individual under like circumstances.” 28 U. S. C. §2674. That provision is most naturally understood to make the United States liable in the same way as a private individual at any given time. See Richards v. United States, 369 U.S. 1, 6–7 (1962). Such “same as” provisions dot the statute books, and federal and state courts commonly read them to mandate ongoing equal treatment of two groups or objects. See, e.g., Adamson v. Bowen, 855 F.2d 668, 671–672 (CA10 1988) (statute making United States liable for fees and expenses “to the same extent that any other party would be liable under the common law or under the terms of any statute” interpreted to continuously tie liability of United States to that of any other party); Kugler’s Appeal, 55 Pa. 123, 124–125 (1867) (statute making the procedure for dividing election districts “the same as” the procedure for dividing townships interpreted to continuously tie the former procedure to the latter). The IFC objects that the IOIA is different because the purpose of international organization immunity is entirely distinct from the purpose of foreign sovereign immunity. Foreign sovereign immunity, the IFC argues, is grounded in the mutual respect of sovereigns and serves the ends of international comity and reciprocity. The purpose of international organization immunity, on the other hand, is to allow such organizations to freely pursue the collective goals of member countries without undue interference from the courts of any one member country. The IFC therefore urges that the IOIA should not be read to tether international organization immunity to changing foreign sovereign immunity. But that gets the inquiry backward. We ordinarily assume, “absent a clearly expressed legislative intention to the contrary,” that “the legislative purpose is expressed by the ordinary meaning of the words used.” American Tobacco Co. v. Patterson, 456 U.S. 63, 68 (1982) (alterations omitted). Whatever the ultimate purpose of international organization immunity may be—the IOIA does not address that question—the immediate purpose of the immunity provision is expressed in language that Congress typically uses to make one thing continuously equivalent to another. B The more natural reading of the IOIA is confirmed by a canon of statutory interpretation that was well established when the IOIA was drafted. According to the “reference” canon, when a statute refers to a general subject, the statute adopts the law on that subject as it exists whenever a question under the statute arises. 2 J. Sutherland, Statutory Construction §§5207–5208 (3d ed. 1943). For example, a statute allowing a company to “collect the same tolls and enjoy the same privileges” as other companies incorporates the law governing tolls and privileges as it exists at any given moment. Snell v. Chicago, 133 Ill. 413, 437–439, 24 N.E. 532, 537 (1890). In contrast, a statute that refers to another statute by specific title or section number in effect cuts and pastes the referenced statute as it existed when the referring statute was enacted, without any subsequent amendments. See, e.g., Culver v. People ex rel. Kochersperger, 161 Ill. 89, 95–99, 43 N.E. 812, 814–815 (1896) (tax-assessment statute referring to specific article of another statute does not adopt subsequent amendments to that article). Federal courts have often relied on the reference canon, explicitly or implicitly, to harmonize a statute with an external body of law that the statute refers to generally. Thus, for instance, a statute that exempts from disclosure agency documents that “would not be available by law to a party . . . in litigation with the agency” incorporates the general law governing attorney work-product privilege as it exists when the statute is applied. FTC v. Grolier Inc., 462 U.S. 19, 20, 26–27 (1983) (emphasis added); id., at 34, n. 6 (Brennan, J., concurring in part and concurring in judgment). Likewise, a general reference to federal discovery rules incorporates those rules “as they are found on any given day, today included,” El Encanto, Inc. v. Hatch Chile Co., 825 F.3d 1161, 1164 (CA10 2016), and a general reference to “the crime of piracy as defined by the law of nations” incorporates a definition of piracy “that changes with advancements in the law of nations,” United States v. Dire, 680 F.3d 446, 451, 467–469 (CA4 2012). The same logic applies here. The IOIA’s reference to the immunity enjoyed by foreign governments is a general rather than specific reference. The reference is to an external body of potentially evolving law—the law of foreign sovereign immunity—not to a specific provision of another statute. The IOIA should therefore be understood to link the law of international organization immunity to the law of foreign sovereign immunity, so that the one develops in tandem with the other. The IFC contends that the IOIA’s reference to the immunity enjoyed by foreign governments is not a general reference to an external body of law, but is instead a specific reference to a common law concept that had a fixed meaning when the IOIA was enacted in 1945. And because we ordinarily presume that “Congress intends to incorporate the well-settled meaning of the common-law terms it uses,” Neder v. United States, 527 U.S. 1, 23 (1999), the IFC argues that we should read the IOIA to incorporate what the IFC maintains was the then-settled meaning of the “immunity enjoyed by foreign governments”: virtually absolute immunity. But in 1945, the “immunity enjoyed by foreign governments” did not mean “virtually absolute immunity.” The phrase is not a term of art with substantive content, such as “fraud” or “forgery.” See id., at 22; Gilbert v. United States, 370 U.S. 650, 655 (1962). It is rather a concept that can be given scope and content only by reference to the rules governing foreign sovereign immunity. It is true that under the rules applicable in 1945, the extent of immunity from suit was virtually absolute, while under the rules applicable today, it is more limited. But in 1945, as today, the IOIA’s instruction to grant international organizations the immunity “enjoyed by foreign governments” is an instruction to look up the applicable rules of foreign sovereign immunity, wherever those rules may be found—the common law, the law of nations, or a statute. In other words, it is a general reference to an external body of (potentially evolving) law. C In ruling for the IFC, the D. C. Circuit relied upon its prior decision in Atkinson, 156 F.3d 1335. Atkinson acknowledged the reference canon, but concluded that the canon’s probative force was “outweighed” by a structural inference the court derived from the larger context of the IOIA. Id., at 1341. The Atkinson court focused on the provision of the IOIA that gives the President the author- ity to withhold, withdraw, condition, or limit the otherwise applicable privileges and immunities of an international organization, “in the light of the functions performed by any such international organization.” 22 U. S. C. §288. The court understood that provision to “delegate to the President the responsibility for updating the immunities of international organizations in the face of changing circumstances.” Atkinson, 156 F. 3d, at 1341. That delegation, the court reasoned, “undermine[d]” the view that Congress intended the IOIA to in effect update itself by incorporating changes in the law governing foreign sovereign immunity. Ibid. We do not agree. The delegation provision is most naturally read to allow the President to modify, on a case-by-case basis, the immunity rules that would otherwise apply to a particular international organization. The statute authorizes the President to take action with respect to a single organization—“any such organization”—in light of the functions performed by “such organization.” 28 U. S. C. §288. The text suggests retail rather than wholesale action, and that is in fact how authority under §288 has been exercised in the past. See, e.g., Exec. Order No. 12425, 3 CFR 193 (1984) (designating INTERPOL as an international organization under the IOIA but withholding certain privileges and immunities); Exec. Order No. 11718, 3 CFR 177 (1974) (same for INTELSAT). In any event, the fact that the President has power to modify otherwise applicable immunity rules is perfectly compatible with the notion that those rules might themselves change over time in light of developments in the law governing foreign sovereign immunity. The D. C. Circuit in Atkinson also gave no consideration to the opinion of the State Department, whose views in this area ordinarily receive “special attention.” Bolivarian Republic of Venezuela v. Helmerich & Payne Int’l. Drilling Co., 581 U. S. ___, ___ (2017) (slip op., at 9). Shortly after the FSIA was enacted, the State Department took the position that the immunity rules of the IOIA and the FSIA were now “link[ed].” Letter from Detlev F. Vagts, Office of the Legal Adviser, to Robert M. Carswell, Jr., Senior Legal Advisor, OAS, p. 2 (Mar. 24, 1977). The Department reaffirmed that view during subsequent administrations, and it has reaffirmed it again here.[2] That longstanding view further bolsters our understanding of the IOIA’s immunity provision. D The IFC argues that interpreting the IOIA’s immunity provision to grant anything less than absolute immunity would lead to a number of undesirable results. The IFC first contends that affording international organizations only restrictive immunity would defeat the purpose of granting them immunity in the first place. Allowing international organizations to be sued in one member country’s courts would in effect allow that member to second-guess the collective decisions of the others. It would also expose international organizations to money damages, which would in turn make it more difficult and expensive for them to fulfill their missions. The IFC argues that this problem is especially acute for international development banks. Because those banks use the tools of commerce to achieve their objectives, they may be subject to suit under the FSIA’s commercial activity exception for most or all of their core activities, unlike foreign sovereigns. According to the IFC, allowing such suits would bring a flood of foreign-plaintiff litigation into U. S. courts, raising many of the same foreign-relations concerns that we identified when considering similar litigation under the Alien Tort Statute. See Jesner v. Arab Bank, PLC, 584 U. S. ___, ___–___ (2018); Kiobel v. Royal Dutch Petroleum Co., 569 U.S. 108, 116–117 (2013). The IFC’s concerns are inflated. To begin, the privileges and immunities accorded by the IOIA are only default rules. If the work of a given international organization would be impaired by restrictive immunity, the organization’s charter can always specify a different level of immunity. The charters of many international organizations do just that. See, e.g., Convention on Privileges and Immunities of the United Nations, Art. II, §2, Feb. 13, 1946, 21 U. S. T. 1422, T. I. A. S. No. 6900 (“The United Nations . . . shall enjoy immunity from every form of legal process except insofar as in any particular case it has expressly waived its immunity”); Articles of Agreement of the International Monetary Fund, Art. IX, §3, Dec. 27, 1945, 60Stat. 1413, T. I. A. S. No. 1501 (IMF enjoys “immunity from every form of judicial process except to the extent that it expressly waives its immunity”). Notably, the IFC’s own charter does not state that the IFC is absolutely immune from suit. Nor is there good reason to think that restrictive immunity would expose international development banks to excessive liability. As an initial matter, it is not clear that the lending activity of all development banks qualifies as commercial activity within the meaning of the FSIA. To be considered “commercial,” an activity must be “the type” of activity “by which a private party engages in” trade or commerce. Republic of Argentina v. Weltover, Inc., 504 U.S. 607, 614 (1992); see 28 U. S. C. §1603(d). As the Government suggested at oral argument, the lending activity of at least some development banks, such as those that make conditional loans to governments, may not qualify as “commercial” under the FSIA. See Tr. of Oral Arg. 27–30. And even if an international development bank’s lending activity does qualify as commercial, that does not mean the organization is automatically subject to suit. The FSIA includes other requirements that must also be met. For one thing, the commercial activity must have a sufficient nexus to the United States. See 28 U. S. C. §§1603, 1605(a)(2). For another, a lawsuit must be “based upon” either the commercial activity itself or acts performed in connection with the commercial activity. See §1605(a)(2). Thus, if the “gravamen” of a lawsuit is tortious activity abroad, the suit is not “based upon” commercial activity within the meaning of the FSIA’s commercial activity exception. See OBB Personenverkehr AG v. Sachs, 577 U. S. ___, ___–___ (2015); Saudi Arabia v. Nelson, 507 U.S. 349, 356–359 (1993). At oral argument in this case, the Government stated that it has “serious doubts” whether petitioners’ suit, which largely concerns allegedly tortious conduct in India, would satisfy the “based upon” requirement. Tr. of Oral Arg. 25–26. In short, restrictive immunity hardly means unlimited exposure to suit for international organizations. * * * The International Organizations Immunities Act grants international organizations the “same immunity” from suit “as is enjoyed by foreign governments” at any given time. Today, that means that the Foreign Sovereign Immunities Act governs the immunity of international organizations. The International Finance Corporation is therefore not absolutely immune from suit. The judgment of the United States Court of Appeals for the D. C. Circuit is reversed, and the case is remanded for further proceedings consistent with this opinion. It is so ordered. Justice Kavanaugh took no part in the consideration or decision of this case. Notes 1 The immunity was “virtually” absolute because it was subject to occasional exceptions for specific situations. In Republic of Mexico v. Hoffman, 324 U.S. 30 (1945), for example, the State Department declined to recommend, and the Court did not grant, immunity from suit with respect to a ship that Mexico owned but did not possess. 2 See Letter from Roberts B. Owen, Legal Adviser, to Leroy D. Clark, Gen. Counsel, EEOC (June 24, 1980) in Nash, Contemporary Practice of the United States Relating to International Law, 74 Am. J. Int’l. L. 917, 918 (1980) (“By virtue of the FSIA, and unless otherwise specified in their constitutive agreements, international organizations are now subject to the jurisdiction of our courts in respect of their commercial activities, while retaining immunity for their acts of a public character.”); Letter from Arnold Kanter, Acting Secretary of State, to President George H. W. Bush (Sept. 12, 1992) in Digest of United States Practice in International Law 1016–1017 (S. Cummins & D. Stewart eds. 2005) (explaining that the Headquarters Agreement of the Organization of American States affords the OAS “full immunity from judicial process, thus going beyond the usual United States practice of affording restrictive immunity,” in exchange for assurances that OAS would provide for “appropriate modes of settlement of those disputes for which jurisdiction would exist against a foreign government under the” FSIA); Brief for United States as Amicus Curiae 24–29.

588.US.2018_17-647

The Township of Scott, Pennsylvania, passed an ordinance requiring that “[a]ll cemeteries . . . be kept open and accessible to the general public during daylight hours.” Petitioner Rose Mary Knick, whose 90-acre rural property has a small family graveyard, was notified that she was violating the ordinance. Knick sought declaratory and injunctive relief in state court on the ground that the ordinance effected a taking of her property, but she did not bring an inverse condemnation action under state law seeking compensation. The Township responded by withdrawing the violation notice and staying enforcement of the ordinance. Without an ongoing enforcement action, the court held, Knick could not demonstrate the irreparable harm necessary for equitable relief, so it declined to rule on her request. Knick then filed an action in Federal District Court under 42 U. S. C. §1983, alleging that the ordinance violated the Takings Clause of the Fifth Amendment. The District Court dismissed her claim under Williamson County Regional Planning Comm’n v. Hamilton Bank of Johnson City, 473 U.S. 172, which held that property owners must seek just compensation under state law in state court before bringing a federal takings claim under §1983. The Third Circuit affirmed. Held: 1. A government violates the Takings Clause when it takes property without compensation, and a property owner may bring a Fifth Amendment claim under §1983 at that time. Pp. 5–20. (a) In Williamson County, the Court held that, as relevant here, a property developer’s federal takings claim was “premature” because he had not sought compensation through the State’s inverse condemnation procedure. 473 U. S., at 197. The unanticipated consequence of this ruling was that a takings plaintiff who complied with Williamson County and brought a compensation claim in state court would—on proceeding to federal court after the unsuccessful state claim—have the federal claim barred because the full faith and credit statute required the federal court to give preclusive effect to the state court’s decision. San Remo Hotel, L. P. v. City and County of San Francisco, 545 U.S. 323, 347. Pp. 5–6. (b) This Court has long recognized that property owners may bring Fifth Amendment claims for compensation as soon as their property has been taken, regardless of any other post-taking remedies that may be available to the property owner. See Jacobs v. United States, 290 U.S. 13. The Court departed from that understanding in Williamson County and held that a taking gives rise not to a constitutional right to just compensation, but instead gives a right to a state law procedure that will eventually result in just compensation. Just two years after Williamson County, however, the Court returned to its traditional understanding of the Fifth Amendment, holding that the compensation remedy is required by the Constitution in the event of a taking. First English Evangelical Lutheran Church of Glendale v. County of Los Angeles, 482 U.S. 304. A property owner acquires a right to compensation immediately upon an uncompensated taking because the taking itself violates the Fifth Amendment. See San Diego Gas & Elec. Co. v. San Diego, 450 U.S. 621, 654 (Brennan, J., dissenting). The property owner may, therefore, bring a claim under §1983 for the deprivation of a constitutional right at that time. Pp. 6–12. (c) Williamson County’s understanding of the Takings Clause was drawn from Ruckelshaus v. Monsanto Co., 467 U.S. 986, where the plaintiff sought to enjoin a federal statute because it effected a taking, even though the statute set up a mandatory arbitration procedure for obtaining compensation. Id., at 1018. That case does not support Williamson County, however, because Congress—unlike the States—is free to require plaintiffs to exhaust administrative remedies before bringing constitutional claims. Williamson County also analogized its new state-litigation requirement to federal takings practice under the Tucker Act, but a claim for just compensation brought under the Tucker Act is not a prerequisite to a Fifth Amendment takings claim—it is a Fifth Amendment takings claim. Williamson County also looked to Parratt v. Taylor, 451 U.S. 527. But Parratt was not a takings case at all, and the analogy from the due process context to the takings context is strained. The poor reasoning of Williamson County may be partially explained by the circumstances in which the state-litigation issue reached the Court, which may not have permitted the Court to adequately test the logic of the state-litigation requirement or consider its implications. Pp. 12–16. (d) Respondents read too broadly statements in prior opinions that the Takings Clause “does not provide or require that compensation shall be actually paid in advance of the occupancy of the land to be taken. But the owner is entitled to reasonable, certain and adequate provision for obtaining compensation” after a taking. Cherokee Nation v. Southern Kansas R. Co., 135 U.S. 641, 659. Those statements concerned requests for injunctive relief, and the availability of subsequent compensation meant that such an equitable remedy was not available. Simply because the property owner was not entitled to injunctive relief at the time of the taking does not mean there was no violation of the Takings Clause at that time. The history of takings litigation provides valuable context. At the time of the founding, there usually was no compensation remedy available to property owners, who could obtain only retrospective damages, as well as an injunction ejecting the government from the property going forward. But in the 1870s, as state courts began to recognize implied rights of action for damages under the state equivalents of the Takings Clause, they declined to grant injunctions because property owners had an adequate remedy at law. Congress enabled property owners to obtain compensation for takings by the Federal Government when it passed the Tucker Act in 1887, and this Court subsequently joined the state courts in holding that the compensation remedy is required by the Takings Clause itself. Today, because the federal and nearly all state governments provide just compensation remedies to property owners who have suffered a taking, equitable relief is generally unavailable. As long as an adequate provision for obtaining just compensation exists, there is no basis to enjoin government action effecting a taking. Pp. 16–19. 2. The state-litigation requirement of Williamson County is overruled. Several factors counsel in favor of this decision. Williamson County was poorly reasoned and conflicts with much of the Court’s takings jurisprudence. Because of its shaky foundations, the rationale for the state-litigation requirement has been repeatedly recast by this Court and the defenders of Williamson County. The state-litigation requirement also proved to be unworkable in practice because the San Remo preclusion trap prevented takings plaintiffs from ever bringing their claims in federal court, contrary to the expectations of the Williamson County Court. Finally, there are no reliance interests on the state-litigation requirement. As long as post-taking compensation remedies are available, governments need not fear that federal courts will invalidate their regulations as unconstitutional. Pp. 20–23. 862 F.3d 310, vacated and remanded. Roberts, C. J., delivered the opinion of the Court, in which Thomas, Alito, Gorsuch, and Kavanaugh, JJ., joined. Thomas, J., filed a concurring opinion. Kagan, J., filed a dissenting opinion, in which Ginsburg, Breyer, and Sotomayor, JJ., joined.

The Takings Clause of the Fifth Amendment states that “private property [shall not] be taken for public use, without just compensation.” In Williamson County Regional Planning Comm’n v. Hamilton Bank of Johnson City, 473 U.S. 172 (1985), we held that a property owner whose property has been taken by a local government has not suffered a violation of his Fifth Amendment rights—and thus cannot bring a federal takings claim in federal court—until a state court has denied his claim for just compensation under state law. The Williamson County Court anticipated that if the property owner failed to secure just compensation under state law in state court, he would be able to bring a “ripe” federal takings claim in federal court. See id., at 194. But as we later held in San Remo Hotel, L. P. v. City and County of San Francisco, 545 U.S. 323 (2005), a state court’s resolution of a claim for just compensation under state law generally has preclusive effect in any subsequent federal suit. The takings plaintiff thus finds himself in a Catch-22: He cannot go to federal court without going to state court first; but if he goes to state court and loses, his claim will be barred in federal court. The federal claim dies aborning. The San Remo preclusion trap should tip us off that the state-litigation requirement rests on a mistaken view of the Fifth Amendment. The Civil Rights Act of 1871, after all, guarantees “a federal forum for claims of unconstitutional treatment at the hands of state officials,” and the settled rule is that “exhaustion of state remedies ‘is not a prerequisite to an action under [42 U. S. C.] §1983.’ ” Heck v. Humphrey, 512 U.S. 477, 480 (1994) (quoting Patsy v. Board of Regents of Fla., 457 U.S. 496, 501 (1982)). But the guarantee of a federal forum rings hollow for takings plaintiffs, who are forced to litigate their claims in state court. We now conclude that the state-litigation requirement imposes an unjustifiable burden on takings plaintiffs, conflicts with the rest of our takings jurisprudence, and must be overruled. A property owner has an actionable Fifth Amendment takings claim when the government takes his property without paying for it. That does not mean that the government must provide compensation in advance of a taking or risk having its action invalidated: So long as the property owner has some way to obtain compensation after the fact, governments need not fear that courts will enjoin their activities. But it does mean that the property owner has suffered a violation of his Fifth Amendment rights when the government takes his property without just compensation, and therefore may bring his claim in federal court under §1983 at that time. I Petitioner Rose Mary Knick owns 90 acres of land in Scott Township, Pennsylvania, a small community just north of Scranton. Knick lives in a single-family home on the property and uses the rest of the land as a grazing area for horses and other farm animals. The property includes a small graveyard where the ancestors of Knick’s neighbors are allegedly buried. Such family cemeteries are fairly common in Pennsylvania, where “backyard burials” have long been permitted. In December 2012, the Township passed an ordinance requiring that “[a]ll cemeteries . . . be kept open and accessible to the general public during daylight hours.” The ordinance defined a “cemetery” as “[a] place or area of ground, whether contained on private or public property, which has been set apart for or otherwise utilized as a burial place for deceased human beings.” The ordinance also authorized Township “code enforcement” officers to “enter upon any property” to determine the existence and location of a cemetery. App. 21–23. In 2013, a Township officer found several grave markers on Knick’s property and notified her that she was violating the ordinance by failing to open the cemetery to the public during the day. Knick responded by seeking declaratory and injunctive relief in state court on the ground that the ordinance effected a taking of her property. Knick did not seek compensation for the taking by bringing an “inverse condemnation” action under state law. Inverse condemnation is “a cause of action against a governmental defendant to recover the value of property which has been taken in fact by the governmental defendant.” United States v. Clarke, 445 U.S. 253, 257 (1980) (quoting D. Hagman, Urban Planning and Land Development Control Law 328 (1971)). Inverse condemnation stands in contrast to direct condemnation, in which the government initiates proceedings to acquire title under its eminent domain authority. Pennsylvania, like every other State besides Ohio, provides a state inverse condemnation action. 26 Pa. Cons. Stat. §502(c) (2009).[1] In response to Knick’s suit, the Township withdrew the violation notice and agreed to stay enforcement of the ordinance during the state court proceedings. The court, however, declined to rule on Knick’s request for declara- tory and injunctive relief because, without an ongoing en- forcement action, she could not demonstrate the irreparable harm necessary for equitable relief. Knick then filed an action in Federal District Court under 42 U. S. C. §1983, alleging that the ordinance violated the Takings Clause of the Fifth Amendment.[2] The District Court dismissed Knick’s takings claim under Williamson County because she had not pursued an inverse condemnation action in state court. 2016 WL 4701549, *5–*6 (MD Pa., Sept. 8, 2016). On appeal, the Third Circuit noted that the ordinance was “extraordinary and constitutionally suspect,” but affirmed the District Court in light of Williamson County. 862 F.3d 310, 314 (2017). We granted certiorari to reconsider the holding of Williamson County that property owners must seek just compensation under state law in state court before bringing a federal takings claim under §1983. 583 U. S. ___ (2018). II In Williamson County, a property developer brought a takings claim under §1983 against a zoning board that had rejected the developer’s proposal for a new subdivision. Williamson County held that the developer’s Fifth Amendment claim was not “ripe” for two reasons. First, the developer still had an opportunity to seek a variance from the appeals board, so any taking was therefore not yet final. 473 U. S., at 186–194. Knick does not question the validity of this finality requirement, which is not at issue here. The second holding of Williamson County is that the developer had no federal takings claim because he had not sought compensation “through the procedures the State ha[d] provided for doing so.” Id., at 194. That is the holding Knick asks us to overrule. According to the Court, “if a State provides an adequate procedure for seeking just compensation, the property owner cannot claim a violation of the [Takings] Clause until it has used the procedure and been denied just compensation.” Id., at 195. The Court concluded that the developer’s federal takings claim was “premature” because he had not sought compensation through the State’s inverse condemnation procedure. Id., at 197. The unanticipated consequences of this ruling were not clear until 20 years later, when this Court decided San Remo. In that case, the takings plaintiffs complied with Williamson County and brought a claim for compensation in state court. 545 U. S., at 331. The complaint made clear that the plaintiffs sought relief only under the takings clause of the State Constitution, intending to reserve their Fifth Amendment claim for a later federal suit if the state suit proved unsuccessful. Id., at 331–332. When that happened, however, and the plaintiffs proceeded to federal court, they found that their federal claim was barred. This Court held that the full faith and credit statute, 28 U. S. C. §1738, required the federal court to give preclusive effect to the state court’s decision, blocking any subsequent consideration of whether the plaintiff had suffered a taking within the meaning of the Fifth Amendment. 545 U. S., at 347. The adverse state court decision that, according to Williamson County, gave rise to a ripe federal takings claim simultaneously barred that claim, preventing the federal court from ever considering it. The state-litigation requirement relegates the Takings Clause “to the status of a poor relation” among the provisions of the Bill of Rights. Dolan v. City of Tigard, 512 U.S. 374, 392 (1994). Plaintiffs asserting any other constitutional claim are guaranteed a federal forum under §1983, but the state-litigation requirement “hand[s] authority over federal takings claims to state courts.” San Remo, 545 U. S., at 350 (Rehnquist, C. J., concurring in judgment). Fidelity to the Takings Clause and our cases construing it requires overruling Williamson County and restoring takings claims to the full-fledged constitutional status the Framers envisioned when they included the Clause among the other protections in the Bill of Rights. III A Contrary to Williamson County, a property owner has a claim for a violation of the Takings Clause as soon as a government takes his property for public use without paying for it. The Clause provides: “[N]or shall private property be taken for public use, without just compensation.” It does not say: “Nor shall private property be taken for public use, without an available procedure that will result in compensation.” If a local government takes private property without paying for it, that government has violated the Fifth Amendment—just as the Takings Clause says—without regard to subsequent state court proceedings. And the property owner may sue the government at that time in federal court for the “deprivation” of a right “secured by the Constitution.” 42 U. S. C. §1983. We have long recognized that property owners may bring Fifth Amendment claims against the Federal Government as soon as their property has been taken. The Tucker Act, which provides the standard procedure for bringing such claims, gives the Court of Federal Claims jurisdiction to “render judgment upon any claim against the United States founded either upon the Constitution” or any federal law or contract for damages “in cases not sounding in tort.” 28 U. S. C. §1491(a)(1). We have held that “[i]f there is a taking, the claim is ‘founded upon the Constitution’ and within the jurisdiction of the Court of Claims to hear and determine.” United States v. Causby, 328 U.S. 256, 267 (1946). And we have explained that “the act of taking” is the “event which gives rise to the claim for compensation.” United States v. Dow, 357 U.S. 17, 22 (1958). The Fifth Amendment right to full compensation arises at the time of the taking, regardless of post-taking remedies that may be available to the property owner. That principle was confirmed in Jacobs v. United States, 290 U.S. 13 (1933), where we held that a property owner found to have a valid takings claim is entitled to compensation as if it had been “paid contemporaneously with the taking”—that is, the compensation must generally consist of the total value of the property when taken, plus interest from that time. Id., at 17 (quoting Seaboard Air Line R. Co. v. United States, 261 U.S. 299, 306 (1923)). We rejected the view of the lower court that a property owner is entitled to interest only when the government provides a particular remedy—direct condemnation proceedings—and not when the owner brings a takings suit under the Tucker Act. “The form of the remedy d[oes] not qualify the right. It rest[s] upon the Fifth Amendment.” 290 U. S., at 16. Jacobs made clear that, no matter what sort of procedures the government puts in place to remedy a taking, a property owner has a Fifth Amendment entitlement to compensation as soon as the government takes his prop- erty without paying for it. Whether the government does nothing, forcing the owner to bring a takings suit under the Tucker Act, or whether it provides the owner with a statutory compensation remedy by initiating direct condemnation proceedings, the owner’s claim for compensation “rest[s] upon the Fifth Amendment.” Although Jacobs concerned a taking by the Federal Government, the same reasoning applies to takings by the States. The availability of any particular compensation remedy, such as an inverse condemnation claim under state law, cannot infringe or restrict the property owner’s federal constitutional claim—just as the existence of a state action for battery does not bar a Fourth Amendment claim of excessive force. The fact that the State has provided a property owner with a procedure that may subsequently result in just compensation cannot deprive the owner of his Fifth Amendment right to compensation under the Constitution, leaving only the state law right. And that is key because it is the existence of the Fifth Amendment right that allows the owner to proceed directly to federal court under §1983. Williamson County had a different view of how the Takings Clause works. According to Williamson County, a taking does not give rise to a federal constitutional right to just compensation at that time, but instead gives a right to a state law procedure that will eventually result in just compensation. As the Court put it, “if a State provides an adequate procedure for seeking just compensation, the property owner cannot claim a violation of the [Takings] Clause until it has used the procedure and been denied just compensation.” 473 U. S., at 195. In the absence of a state remedy, the Fifth Amendment right to compensation would attach immediately. But, under Williamson County, the presence of a state remedy qualifies the right, preventing it from vesting until exhaustion of the state procedure. That is what Jacobs confirmed could not be done. Just two years after Williamson County, in First English Evangelical Lutheran Church of Glendale v. County of Los Angeles, 482 U.S. 304 (1987), the Court returned to the understanding that the Fifth Amendment right to compensation automatically arises at the time the government takes property without paying for it. Relying heavily on Jacobs and other Fifth Amendment precedents neglected by Williamson County, First English held that a property owner is entitled to compensation for the temporary loss of his property. We explained that “government action that works a taking of property rights necessarily implicates the ‘constitutional obligation to pay just compensation.’ ” 482 U. S., at 315. Because of “the self-executing character” of the Takings Clause “with respect to compensation,” a property owner has a constitutional claim for just compensation at the time of the taking. Ibid. (quoting 6 P. Nichols, Eminent Domain §25.41 (3d rev. ed. 1972)). The government’s post-taking actions (there, repeal of the challenged ordinance) cannot nullify the property owner’s existing Fifth Amendment right: “[W]here the government’s activities have already worked a taking of all use of property, no subsequent action by the government can relieve it of the duty to provide compensation.” 482 U. S., at 321.[3] In holding that a property owner acquires an irrevocable right to just compensation immediately upon a taking, First English adopted a position Justice Brennan had taken in an earlier dissent. See id., at 315, 318 (quoting and citing San Diego Gas & Elec. Co. v. San Diego, 450 U.S. 621, 654, 657 (1981) (Brennan, J., dissenting)).[4] In that opinion, Justice Brennan explained that “once there is a ‘taking,’ compensation must be awarded” because “[a]s soon as private property has been taken, whether through formal condemnation proceedings, occupancy, physical invasion, or regulation, the landowner has already suffered a constitutional violation.” Id., at 654. First English embraced that view, reaffirming that “in the event of a taking, the compensation remedy is required by the Constitution.” 482 U. S., at 316; see ibid., n. 9 (rejecting the view that “the Constitution does not, of its own force, furnish a basis for a court to award money damages against the government” (quoting Brief for United States as Amicus Curiae 14)). Compensation under the Takings Clause is a remedy for the “constitutional violation” that “the landowner has already suffered” at the time of the uncompensated taking. San Diego Gas & Elec. Co., 450 U. S., at 654 (Brennan, J., dissenting); see First English, 482 U. S., at 315. A later payment of compensation may remedy the constitutional violation that occurred at the time of the taking, but that does not mean the violation never took place. The violation is the only reason compensation was owed in the first place. A bank robber might give the loot back, but he still robbed the bank. The availability of a subsequent compensation remedy for a taking without compensation no more means there never was a constitutional violation in the first place than the availability of a damages action renders negligent conduct compliant with the duty of care. In sum, because a taking without compensation violates the self-executing Fifth Amendment at the time of the taking, the property owner can bring a federal suit at that time. Just as someone whose property has been taken by the Federal Government has a claim “founded . . . upon the Constitution” that he may bring under the Tucker Act, someone whose property has been taken by a local government has a claim under §1983 for a “deprivation of [a] right[ ] . . . secured by the Constitution” that he may bring upon the taking in federal court. The “general rule” is that plaintiffs may bring constitutional claims under §1983 “without first bringing any sort of state lawsuit, even when state court actions addressing the underlying behavior are available.” D. Dana & T. Merrill, Property: Takings 262 (2002); see McNeese v. Board of Ed. for Community Unit School Dist. 187, 373 U.S. 668, 672 (1963) (observing that it would defeat the purpose of §1983 “if we held that assertion of a federal claim in a federal court must await an attempt to vindicate the same claim in a state court”); Monroe v. Pape, 365 U.S. 167, 183 (1961) (“The federal remedy is supplementary to the state rem- edy, and the latter need not be first sought and refused before the federal one is invoked.”). This is as true for takings claims as for any other claim grounded in the Bill of Rights. B Williamson County effectively established an exhaustion requirement for §1983 takings claims when it held that a property owner must pursue state procedures for obtaining compensation before bringing a federal suit. But the Court did not phrase its holding in those terms; if it had, its error would have been clear. Instead, Williamson County broke with the Court’s longstanding position that a property owner has a constitutional claim to compensation at the time the government deprives him of his property, and held that there can be no uncompensated taking, and thus no Fifth Amendment claim actionable under §1983, until the property owner has tried and failed to obtain compensation through the available state procedure. “[U]ntil it has used the procedure and been denied just compensation,” the property owner “ ‘has no claim against the Government’ for a taking.” 473 U. S., at 194–195 (quoting Ruckelshaus v. Monsanto Co., 467 U.S. 986, 1018, n. 21 (1984)). Williamson County drew that understanding of the Clause from Ruckelshaus v. Monsanto Co., a decision from the prior Term. Monsanto did not involve a takings claim for just compensation. The plaintiff there sought to enjoin a federal statute because it effected a taking, even though the statute set up a special arbitration procedure for obtaining compensation, and the plaintiff could bring a takings claim pursuant to the Tucker Act if arbitration did not yield sufficient compensation. 467 U. S., at 1018. The Court rejected the plaintiff’s claim because “[e]quitable relief is not available to enjoin an alleged taking of private property for a public use, duly authorized by law, when a suit for compensation can be brought against the sovereign subsequent to the taking.” Id., at 1016 (footnote omitted). That much is consistent with our precedent: Equitable relief was not available because monetary relief was under the Tucker Act. That was enough to decide the case. But Monsanto went on to say that if the plaintiff obtained compensation in arbitration, then “no taking has occurred and the [plaintiff] has no claim against the Government.” Id., at 1018, n. 21. Certainly it is correct that a fully compensated plaintiff has no further claim, but that is because the taking has been remedied by compensation, not because there was no taking in the first place. See First English, 482 U. S., at 316, n. 9. The statute in Monsanto simply required the plaintiff to attempt to vindicate its claim to compensation through arbitration before proceeding under the Tucker Act. The case offers no support to Williamson County in this regard, because Congress—unlike the States—is free to require plaintiffs to exhaust administrative remedies before bringing constitutional claims. See McCarthy v. Madigan, 503 U.S. 140, 144 (1992) (“Where Congress specifically mandates, exhaustion is required.”). Williamson County also relied on Monsanto when it analogized its new state-litigation requirement to federal takings practice, stating that “taking[s] claims against the Federal Government are premature until the property owner has availed itself of the process provided by the Tucker Act.” 473 U. S., at 195. But the Court was simply confused. A claim for just compensation brought under the Tucker Act is not a prerequisite to a Fifth Amendment takings claim—it is a Fifth Amendment takings claim. A party who loses a Tucker Act suit has nowhere else to go to seek compensation for an alleged taking. Other than Monsanto, the principal case to which Williamson County looked was Parratt v. Taylor, 451 U.S. 527 (1981). Like Monsanto, Parratt did not involve a takings claim for just compensation. Indeed, it was not a takings case at all. Parratt held that a prisoner deprived of $23.50 worth of hobby materials by the rogue act of a state employee could not state a due process claim if the State provided adequate post-deprivation process. 451 U. S., at 543–544. But the analogy from the due process context to the takings context is strained, as Williamson County itself recognized. See 473 U. S., at 195, n. 14. It is not even possible for a State to provide pre-deprivation due process for the unauthorized act of a single employee. That is quite different from the taking of property by the government through physical invasion or a regulation that destroys a property’s productive use. The poor reasoning of Williamson County may be partially explained by the circumstances in which the state-litigation issue reached the Court. The Court granted certiorari to decide whether the Fifth Amendment entitles a property owner to just compensation when a regulation temporarily deprives him of the use of his property. (First English later held that the answer was yes.) As amicus curiae in support of the local government, the United States argued in this Court that the developer could not state a Fifth Amendment claim because it had not pursued an inverse condemnation suit in state court. Neither party had raised that argument before.[5] The Court then adopted the reasoning of the Solicitor General in an alternative holding, even though the case could have been resolved solely on the narrower and settled ground that no taking had occurred because the zoning board had not yet come to a final decision regarding the developer’s proposal. In these circumstances, the Court may not have ade- quately tested the logic of the state-litigation requirement or considered its implications, most notably the preclusion trap later sprung by San Remo. That consequence was totally unanticipated in Williamson County. The dissent, doing what respondents do not even dare to attempt, defends the original rationale of Williamson County—that there is no Fifth Amendment violation, and thus no Fifth Amendment claim, until the government denies the property owner compensation in a subsequent proceeding.[6] But although the dissent makes a more thoughtful and considered argument than Williamson County, it cannot reconcile its view with our repeated holdings that a property owner acquires a constitutional right to compensation at the time of the taking. See supra, at 7–11. The only reason that a taking would automatically entitle a property owner to the remedy of compensation is that, as Justice Brennan explained, with the uncompensated taking “the landowner has already suffered a constitutional violation.” San Diego Gas & Elec. Co., 450 U. S., at 654 (dissenting opinion). The dissent here provides no more reason to resist that conclusion than did Williamson County. C The Court in Williamson County relied on statements in our prior opinions that the Clause “does not provide or require that compensation shall be actually paid in advance of the occupancy of the land to be taken. But the owner is entitled to reasonable, certain and adequate provision for obtaining compensation” after a taking. Cherokee Nation v. Southern Kansas R. Co., 135 U.S. 641, 659 (1890). Respondents rely on the same cases in contending that uncompensated takings for which compensation is subsequently available do not violate the Fifth Amendment at the time of the taking. But respondents read those statements too broadly. They concerned requests for injunctive relief, and the availability of subsequent compensation meant that such an equitable remedy was not available. See Regional Rail Reorganization Act Cases, 419 U.S. 102, 107, 149 (1974) (reversing a decision “enjoin[ing]” the enforcement of a federal statute because “the availability of the Tucker Act guarantees an adequate remedy at law for any taking which might occur”); Hurley v. Kincaid, 285 U.S. 95, 99, 105 (1932) (rejecting a request to “enjoin the carrying out of any work” on a flood control project because the Tucker Act provided the plaintiff with “a plain, adequate, and complete remedy at law”). Simply because the property owner was not entitled to injunctive relief at the time of the taking does not mean there was no violation of the Takings Clause at that time. The history of takings litigation provides valuable context. At the time of the founding there usually was no compensation remedy available to property owners. On occasion, when a legislature authorized a particular government action that took private property, it might also create a special owner-initiated procedure for obtaining compensation. But there were no general causes of action through which plaintiffs could obtain compensation for property taken for public use. Brauneis, The First Constitutional Tort: The Remedial Revolution in Nineteenth-Century State Just Compensation Law, 52 Vand. L. Rev. 57, 69–70, and n. 33 (1999). Until the 1870s, the typical recourse of a property owner who had suffered an uncompensated taking was to bring a common law trespass action against the responsible corporation or government official. The official would then raise the defense that his trespass was lawful because authorized by statute or ordinance, and the plaintiff would respond that the law was unconstitutional because it provided for a taking without just compensation. If the plaintiff prevailed, he nonetheless had no way at common law to obtain money damages for a permanent taking—that is, just compensation for the total value of his prop- erty. He could obtain only retrospective damages, as well as an injunction ejecting the government from his property going forward. See id., at 67–69, 97–99. As Chancellor Kent explained when granting a property owner equitable relief, the Takings Clause and its analogs in state constitutions required that “a fair compensation must, in all cases, be previously made to the individuals affected.” Gardner v. Newburgh, 2 Johns. Ch. 162, 166 (N. Y. 1816) (emphasis added). If a government took property without payment, a court would set aside the taking because it violated the Constitution and order the property restored to its owner. The Framers meant to prohibit the Federal Government from taking property without paying for it. Allowing the government to keep the property pending subsequent compensation to the owner, in proceedings that hardly existed in 1787, was not what they envisioned. Antebellum courts, which had no means of compensating a property owner for his loss, had no way to redress the violation of an owner’s Fifth Amendment rights other than ordering the government to give him back his prop- erty. See Callender v. Marsh, 18 Mass. 418, 430–431 (1823) (“[I]f by virtue of any legislative act the land of any citizen should be occupied by the public . . . , without any means provided to indemnify the owner of the property, . . . because such a statute would be directly contrary to the [Massachusetts takings clause]; and as no action can be maintained against the public for damages, the only way to secure the party in his constitutional rights would be to declare void the public appropriation.”). But in the 1870s, as state courts began to recognize implied rights of action for damages under the state equivalents of the Takings Clause, they declined to grant injunctions because prop- erty owners had an adequate remedy at law. See, e.g., Stet- son v. Chicago & Evanston R. Co., 75 Ill. 74, 78 (1874) (“What injury, if any, [the property owner] has sustained, may be compensated by damages recoverable by an action at law.”); see also Brauneis, supra, at 97–99, 110–112. On the federal level, Congress enabled property owners to obtain compensation for takings in federal court when it passed the Tucker Act in 1887, and we subsequently joined the state courts in holding that the compensation remedy is required by the Takings Clause itself. See First English, 482 U. S., at 316 (collecting cases). Today, because the federal and nearly all state governments provide just compensation remedies to property owners who have suffered a taking, equitable relief is generally unavailable. As long as an adequate provision for obtaining just compensation exists, there is no basis to enjoin the government’s action effecting a taking. But that is because, as the Court explained in First English, such a procedure is a remedy for a taking that violated the Constitution, not because the availability of the procedure somehow prevented the violation from occurring in the first place. See supra, at 9–11.[7] The dissent contends that our characterization of Cherokee Nation effectively overrules “a hundred-plus years of legal rulings.” Post, at 6 (opinion of Kagan, J.). But under today’s decision every one of the cases cited by the dissent would come out the same way—the plaintiffs would not be entitled to the relief they requested because they could instead pursue a suit for compensation. The premise of such a suit for compensation is that the prop- erty owner has already suffered a violation of the Fifth Amendment that may be remedied by money damages.[8] * * * We conclude that a government violates the Takings Clause when it takes property without compensation, and that a property owner may bring a Fifth Amendment claim under §1983 at that time. That does not as a practical matter mean that government action or regulation may not proceed in the absence of contemporaneous compensation. Given the availability of post-taking compensation, barring the government from acting will ordinarily not be appropriate. But because the violation is complete at the time of the taking, pursuit of a remedy in federal court need not await any subsequent state action. Takings claims against local governments should be handled the same as other claims under the Bill of Rights. Williamson County erred in holding otherwise. IV The next question is whether we should overrule Williamson County, or whether stare decisis counsels in favor of adhering to the decision, despite its error. The doctrine of stare decisis reflects a judgment “that ‘in most matters it is more important that the applicable rule of law be settled than that it be settled right.’ ” Agostini v. Felton, 521 U.S. 203, 235 (1997) (quoting Burnet v. Coronado Oil & Gas Co., 285 U.S. 393, 406 (1932) (Brandeis, J., dissenting)). The doctrine “is at its weakest when we interpret the Constitution,” as we did in Williamson County, because only this Court or a constitutional amendment can alter our holdings. Agostini, 521 U. S., at 235. We have identified several factors to consider in deciding whether to overrule a past decision, including “the quality of [its] reasoning, the workability of the rule it established, its consistency with other related decisions, . . . and reliance on the decision.” Janus v. State, County, and Municipal Employees, 585 U. S. ___, ___–___ (2018) (slip op., at 34–35). All of these factors counsel in favor of overruling Williamson County. Williamson County was not just wrong. Its reasoning was exceptionally ill founded and conflicted with much of our takings jurisprudence. See supra, at 12–14. Its key conclusion, which it drew from unnecessary language in Monsanto—that a property owner does not have a ripe federal takings claim until he has unsuccessfully pursued an initial state law claim for just compensation—ignored Jacobs and many subsequent decisions holding that a property owner acquires a Fifth Amendment right to compensation at the time of a taking. This contradiction was on stark display just two years later in First English. The decision has come in for repeated criticism over the years from Justices of this Court and many respected commentators. See San Remo, 545 U. S., at 348 (Rehnquist, C. J., joined by O’Connor, Kennedy, and Thomas, JJ., concurring in judgment); Arrigoni Enter- prises, LLC v. Durham, 578 U. S. ___ (2016) (Thomas, J., joined by Kennedy, J., dissenting from denial of certiorari); Merrill, Anticipatory Remedies for Takings, 128 Harv. L. Rev. 1630, 1647–1649 (2015); McConnell, Horne and the Normalization of Takings Litigation: A Response to Professor Echeverria, 43 Env. L. Rep. 10749, 10751 (2013); Friedman, Under the Law of Federal Jurisdiction: Allocating Cases Between Federal and State Courts, 104 Colum. L. Rev. 1211, 1264 (2004); Monaghan, State Law Wrongs, State Law Remedies, and the Fourteenth Amendment, 86 Colum. L. Rev. 979, 989 (1986). Even the academic defenders of the state-litigation requirement base it on federalism concerns (although they do not reconcile those concerns with the settled construction of §1983) rather than the reasoning of the opinion itself. See Echeverria, Horne v. Department of Agriculture: An Invitation To Reexamine “Ripeness” Doctrine in Takings Litigation, 43 Env. L. Rep. 10735, 10744 (2013); Sterk, The Demise of Federal Takings Litigation, 48 Wm. & Mary L. Rev. 251, 288 (2006). Because of its shaky foundations, the state-litigation requirement has been a rule in search of a justification for over 30 years. We eventually abandoned the view that the requirement is an element of a takings claim and recast it as a “prudential” ripeness rule. See Horne v. Department of Agriculture, 569 U.S. 513, 525–526 (2013); Suitum v. Tahoe Regional Planning Agency, 520 U.S. 725, 733–734 (1997). No party defends that approach here. See Brief for Respondents 37; Brief for United States as Amicus Curiae 19–20. Respondents have taken a new tack, adopting a §1983–specific theory at which Williamson County did not even hint. See n. 6, supra. The fact that the justification for the state-litigation requirement continues to evolve is another factor undermining the force of stare decisis. See Janus, 585 U. S., at ___ (slip op., at 23). The state-litigation requirement has also proved to be unworkable in practice. Williamson County envisioned that takings plaintiffs would ripen their federal claims in state court and then, if necessary, bring a federal suit under §1983. But, as we held in San Remo, the state court’s resolution of the plaintiff’s inverse condemnation claim has preclusive effect in any subsequent federal suit. The upshot is that many takings plaintiffs never have the opportunity to litigate in a federal forum that §1983 by its terms seems to provide. That significant consequence was not considered by the Court in Williamson County. The dissent argues that our constitutional holding in Williamson County should enjoy the “enhanced” form of stare decisis we usually reserve for statutory decisions, because Congress could have eliminated the San Remo preclusion trap by amending the full faith and credit statute. Post, at 17 (quoting Kimble v. Marvel Entertainment, LLC, 578 U. S. ___, ___ (slip op., at 8)). But takings plaintiffs, unlike plaintiffs bringing any other constitutional claim, would still have been forced to pursue relief under state law before they could bring suit in federal court. Congress could not have lifted that unjustified exhaustion requirement because, under Williamson County, a property owner had no federal claim until a state court denied him compensation. Finally, there are no reliance interests on the state-litigation requirement. We have recognized that the force of stare decisis is “reduced” when rules that do not “serve as a guide to lawful behavior” are at issue. United States v. Gaudin, 515 U.S. 506, 521 (1995); see Alleyne v. United States, 570 U.S. 99, 119 (2013) (Sotomayor, J., concurring). Our holding that uncompensated takings violate the Fifth Amendment will not expose governments to new liability; it will simply allow into federal court takings claims that otherwise would have been brought as inverse condemnation suits in state court. Governments need not fear that our holding will lead federal courts to invalidate their regulations as unconstitutional. As long as just compensation remedies are available—as they have been for nearly 150 years—injunctive relief will be foreclosed. For the same reason, the Federal Government need not worry that courts will set aside agency actions as unconstitutional under the Administrative Procedure Act. 5 U. S. C. §706(2)(B). Federal courts will not invalidate an otherwise lawful uncompensated taking when the property owner can receive complete relief through a Fifth Amendment claim brought under the Tucker Act. In light of all the foregoing, the dissent cannot, with respect, fairly maintain its extreme assertions regarding our application of the principle of stare decisis. * * * The state-litigation requirement of Williamson County is overruled. A property owner may bring a takings claim under §1983 upon the taking of his property without just compensation by a local government. The judgment of the United States Court of Appeals for the Third Circuit is vacated, and the case is remanded for further proceedings consistent with this opinion. It is so ordered. Notes 1 A property owner in Ohio who has suffered a taking without compensation must seek a writ of mandamus to compel the government to initiate condemnation proceedings. See, e.g., State ex rel. Doner v. Zody, 130 Ohio St. 3d 446, 2011-Ohio-6117, 958 N.E.2d 1235. 2 Section 1983 provides: “Every person who, under color of any statute, ordinance, regulation, custom, or usage, of any State or Territory or the District of Columbia, subjects, or causes to be subjected, any citizen of the United States or other person within the jurisdiction thereof to the deprivation of any rights, privileges, or immunities secured by the Constitution and laws, shall be liable to the party injured in an action at law . . . .” 3 First English distinguished Williamson County in a footnote, explaining that the case addressed only “whether the constitutional claim was ripe for review” before the State denied compensation. 482 U. S., at 320, n. 10. But Williamson County was based on the premise that there was no Fifth Amendment claim at all until the State denies compensation. Having rejected that premise, First English eliminated the rationale for the state-litigation requirement. The author of First English later recognized that it was “not clear . . . that Williamson County was correct in demanding that . . . the claimant must seek compensation in state court before bringing a federal takings claim in federal court.” San Remo Hotel, L. P. v. City and County of San Francisco, 545 U.S. 323, 349 (2005) (Rehnquist, C. J., concurring in judgment). 4 Justice Brennan was joined by Justices Stewart, Marshall, and Powell. The majority did not disagree with Justice Brennan’s analysis of the merits, but concluded that the Court lacked jurisdiction to address the question presented. Justice Rehnquist, concurring on the jurisdictional issue, noted that if he were satisfied that jurisdiction was proper, he “would have little difficulty in agreeing with much of what is said in the dissenting opinion.” 450 U. S., at 633–634. The Court reached the merits of the question presented in San Diego in First English, adopting Justice Brennan’s view in an opinion by Chief Justice Rehnquist. 5 The Solicitor General continues this tradition here, arguing for the first time as amicus curiae that state inverse condemnation claims “aris[e] under” federal law and can be brought in federal court under 28 U. S. C. §1331 through the Grable doctrine. Brief for United States as Amicus Curiae 22–24; see Grable & Sons Metal Products, Inc. v. Darue Engineering & Mfg., 545 U.S. 308 (2005). Because we agree with the Solicitor General’s principal contention that federal takings claims can be brought immediately under §1983, we have no occasion to consider his novel §1331 argument. 6 The dissent thinks that respondents still press this theory. Post, at 6 n. 3. But respondents instead describe Williamson County as resting on an understanding not of the elements of a federal takings claim but of the scope of 42 U. S. C. §1983. They even go so far as to rewrite petitioner’s question presented in such terms. Brief for Respondents i. For respondents, it does not matter whether a property owner has a Fifth Amendment claim at the time of a taking. What matters is that, in respondents’ view, no constitutional violation occurs for purposes of §1983 until the government has subsequently denied compensation. That characterization has no basis in the Williamson County opinion, which did not even quote §1983 and stated that the Court’s reasoning applied with equal force to takings by the Federal Government, not covered by §1983. 473 U. S., at 195. Respondents’ attempt to recast the state-litigation requirement as a §1983-specific rule fails for the same reason as the logic of Williamson County—a property owner has a Fifth Amendment claim for a violation of the Takings Clause as soon as the government takes his property without paying for it. 7 Among the cases invoking the Cherokee Nation language that the parties have raised, only one, Yearsley v. W. A. Ross Constr. Co., 309 U.S. 18 (1940), rejected a demand for compensation. Yearsley concerned a state tort suit alleging a taking by a contractor building dikes for the Federal Government. In ruling for the contractors, we sug-gested that the taking did not violate the Fifth Amendment because the property owner had the opportunity to pursue a claim for just compensation under the Tucker Act. As explained, however, a claim for compensation brought under the Tucker Act is a claim for a violation of the Fifth Amendment; it does not prevent a violation from occurring. Regardless, Yearsley was right to hold that the contractors were immune from suit. Because the Tucker Act provides a complete remedy for any taking by the Federal Government, it “excludes liability of the Government’s representatives lawfully acting on its behalf in relation to the taking,” barring the plaintiffs from seeking any relief from the contractors themselves. Id., at 22. 8 The dissent also asserts that today’s ruling “betrays judicial federalism.” Post, at 15. But since the Civil Rights Act of 1871, part of “judicial federalism” has been the availability of a federal cause of action when a local government violates the Constitution. 42 U. S. C. §1983. Invoking that federal protection in the face of state action violating the Fifth Amendment cannot properly be regarded as a betrayal of federalism.

587.US.2018_17-988

In 2016, a hacker tricked an employee of petitioner Lamps Plus, Inc., into disclosing tax information of about 1,300 company employees. After a fraudulent federal income tax return was filed in the name of respondent Frank Varela, a Lamps Plus employee, Varela filed a putative class action against Lamps Plus in Federal District Court on behalf of employees whose information had been compromised. Relying on the arbitration agreement in Varela’s employment contract, Lamps Plus sought to compel arbitration—on an individual rather than a classwide basis—and to dismiss the suit. The District Court rejected the individual arbitration request, but authorized class arbitration and dismissed Varela’s claims. Lamps Plus appealed, arguing that the District Court erred by compelling class arbitration, but the Ninth Circuit affirmed. This Court had held in Stolt-Nielsen S. A. v. AnimalFeeds Int’l Corp., 559 U.S. 662, that a court may not compel classwide arbitration when an agreement is silent on the availability of such arbitration. The Ninth Circuit ruled that Stolt-Nielsen was not controlling because the agreement in this case was ambiguous rather than silent on the issue of class arbitration. Held: 1. This Court has jurisdiction. An order that both compels arbitration and dismisses the underlying claims qualifies as “a final decision with respect to an arbitration” within the meaning of 9 U. S. C. §16(a)(3), the jurisdictional provision on which Lamps Plus relies. See Green Tree Financial Corp.-Ala. v. Randolph, 531 U.S. 79, 89. Varela attempts to distinguish Randolph on the ground that the appeal here was taken by the party who had already secured the relief it requested, i.e., Lamps Plus had already obtained an order dismissing the claim and compelling arbitration. But Lamps Plus did not secure the relief it requested, since it sought individual rather than class arbitration. The shift from individual to class arbitration is a “fundamental” change, Stolt-Nielsen, 559 U. S., at 686, that “sacrifices the principal advantage of arbitration” and “greatly increases risks to defendants,” AT&T Mobility LLC v. Concepcion, 563 U.S. 333, 348, 350. Avoiding these consequences gives Lamps Plus the “necessary personal stake” to appeal. Camreta v. Greene, 563 U.S. 692, 702. Pp. 3–5. 2. Under the Federal Arbitration Act, an ambiguous agreement cannot provide the necessary contractual basis for concluding that the parties agreed to submit to class arbitration. Pp. 5–12. (a) “Arbitration is strictly a matter of consent,” Granite Rock Co. v. Teamsters, 561 U.S. 287, 299 (internal quotation marks omitted), and the task for courts and arbitrators is “to give effect to the intent of the parties,” Stolt-Nielsen, 559 U. S., 684. In carrying out that responsibility, it is important to recognize the “fundamental” difference between class arbitration and the individualized form of arbitration envisioned by the FAA. Class arbitration “sacrifices the principal advantage of arbitration—its informality—and makes the process slower, more costly, and more likely to generate procedural morass than final judgment.” Concepcion, 563 U. S., at 348. Because of such “crucial differences,” Stolt-Nielsen, 559 U. S., at 687, this Court has held that courts may not infer consent to participate in class arbitration absent an affirmative “contractual basis for concluding that the party agreed to do so,” id., at 684. Silence is not enough. Id., at 687. That reasoning controls here. Like silence, ambiguity does not provide a sufficient basis to conclude that parties to an arbitration agreement agreed to “sacrifice[ ] the principal advantage of arbitration.” Concepcion, 563 U. S., at 348. This conclusion aligns with the Court’s refusal to infer consent when it comes to other fundamental arbitration questions. See, e.g., First Options of Chicago, Inc. v. Kaplan, 514 U.S. 938, 945. Pp. 6–9. (b) The Ninth Circuit’s contrary conclusion was based on the state law contra proferentem doctrine, which counsels that contractual ambiguities should be construed against the drafter. That default rule is based on public policy considerations and seeks ends other than the intent of the parties. Such an approach is flatly inconsistent with “the foundational FAA principle that arbitration is a matter of consent.” Stolt-Nielsen, 559 U. S., at 684. Varela claims that the rule is nondiscriminatory and gives equal treatment to arbitration agreements and other contracts alike, but an equal treatment principle cannot save from preemption general rules “that target arbitration either by name or by more subtle methods, such as by ‘interfer[ing] with fundamental attributes of arbitration,’ ” Epic Systems Corp. v. Lewis, 584 U. S. ___, ___. This conclusion is consistent with the Court’s precedents holding that the FAA provides the default rule for resolving certain ambiguities in arbitration agreements. See, e.g., Mitsubishi Motors Corp. v. Soler Chrysler-Plymouth, Inc., 473 U.S. 614, 626. Pp. 9–12. 701 Fed. Appx. 670, reversed and remanded. Roberts, C. J., delivered the opinion of the Court, in which Thomas, Alito, Gorsuch, and Kavanaugh, JJ., joined. Thomas, J., filed a concurring opinion. Ginsburg, J., filed a dissenting opinion, in which Breyer and Sotomayor, JJ., joined. Breyer, J., and Sotomayor, J., filed dissenting opinions. Kagan, J., filed a dissenting opinion, in which Ginsburg and Breyer, JJ., joined, and in which Sotomayor, J., joined as to Part II.

The Federal Arbitration Act requires courts to enforce covered arbitration agreements according to their terms. See 9 U. S. C. §2. In Stolt-Nielsen S. A. v. AnimalFeeds Int’l Corp., 559 U.S. 662 (2010), we held that a court may not compel arbitration on a classwide basis when an agreement is “silent” on the availability of such arbitration. Because class arbitration fundamentally changes the nature of the “traditional individualized arbitration” envisioned by the FAA, Epic Systems Corp. v. Lewis, 584 U. S. ___, ___ (2018) (slip op., at 8), “a party may not be compelled under the FAA to submit to class arbitration unless there is a contractual basis for concluding that the party agreed to do so,” Stolt-Nielsen, 559 U. S., at 684 (emphasis in original). We now consider whether the FAA similarly bars an order requiring class arbitration when an agreement is not silent, but rather “ambiguous” about the availability of such arbitration. I Petitioner Lamps Plus is a company that sells light fixtures and related products. In 2016, a hacker impersonating a company official tricked a Lamps Plus em- ployee into disclosing the tax information of approximately 1,300 other employees. Soon after, a fraudulent federal income tax return was filed in the name of Frank Varela, a Lamps Plus employee and respondent here. Like most Lamps Plus employees, Varela had signed an arbitration agreement when he started work at the company. But after the data breach, he sued Lamps Plus in Federal District Court in California, bringing state and federal claims on behalf of a putative class of employees whose tax information had been compromised. Lamps Plus moved to compel arbitration on an individual rather than classwide basis, and to dismiss the lawsuit. In a single order, the District Court granted the motion to compel arbitration and dismissed Varela’s claims without prejudice. But the court rejected Lamps Plus’s request for individual arbitration, instead authorizing arbitration on a classwide basis. Lamps Plus appealed the order, arguing that the court erred by compelling class arbitration. The Ninth Circuit affirmed. 701 Fed. Appx. 670 (2017). The court acknowledged that Stolt-Nielsen prohibits forcing a party “to submit to class arbitration unless there is a contractual basis for concluding that the party agreed to do so” and that Varela’s agreement “include[d] no express mention of class proceedings.” 701 Fed. Appx., at 672. But that did not end the inquiry, the court reasoned, because the fact that the agreement “does not expressly refer to class arbitration is not the ‘silence’ contemplated in Stolt-Nielsen.” Ibid. In Stolt-Nielsen, the parties had stipulated that their agreement was silent about class arbitration. Because there was no such stipulation here, the court concluded that Stolt-Nielsen was not controlling. The Ninth Circuit then determined that the agreement was ambiguous on the issue of class arbitration. On the one hand, as Lamps Plus argued, certain phrases in the agreement seemed to contemplate “purely binary claims.” Ibid. At the same time, as Varela asserted, other phrases were capacious enough to include class arbitration, such as one stating that “arbitration shall be in lieu of any and all lawsuits or other civil legal proceedings relating to my employment.” Ibid. The Ninth Circuit followed California law to construe the ambiguity against the drafter, a rule that “applies with peculiar force in the case of a contract of adhesion” such as this. Ibid. (quoting Sandquist v. Lebo Auto., Inc., 1 Cal. 5th 233, 248, 376 P.3d 506, 514 (2016)). Because Lamps Plus had drafted the agreement, the court adopted Varela’s interpretation authorizing class arbitration. Judge Fernandez dissented. In his view, the agreement was not ambiguous, and the majority’s holding was a “palpable evasion of Stolt-Nielsen.” 701 Fed. Appx., at 673. Lamps Plus petitioned for a writ of certiorari, arguing that the Ninth Circuit’s decision contravened Stolt-Nielsen and created a conflict among the Courts of Appeals. In opposition, Varela not only disputed those contentions but also argued for the first time that the Ninth Circuit lacked jurisdiction over the appeal, and that this Court therefore lacked jurisdiction in turn. We granted certiorari. 584 U. S. ___ (2018). II We begin with jurisdiction. Section 16 of the FAA governs appellate review of arbitration orders. 9 U. S. C. §16. Varela contends that the Ninth Circuit lacked statutory jurisdiction because section 16 permits appeal from orders denying motions to compel arbitration, §16(a)(1)(B), but not orders granting such motions, §16(b)(2). Brief for Respondent 9–12; see also post, at 3 (Breyer, J., dissenting). This argument is beside the point, however, because Lamps Plus relies for jurisdiction on a different provision of section 16, section 16(a)(3). Section 16(a)(3) provides that an appeal may be taken from “a final decision with respect to an arbitration that is subject to this title.” We construed that provision in Green Tree Financial Corp.-Ala. v. Randolph, 531 U.S. 79 (2000), a case where, as here, the District Court had issued an order both compelling arbitration and dismissing the underlying claims. We held that such an order directing “the parties to proceed to arbitration, and dismiss[ing] all the claims before [the court], . . . is ‘final’ within the meaning of §16(a)(3), and therefore appealable.” Id., at 89.[1] Varela attempts to distinguish Randolph on the ground that the appeal here was taken by the party who sought an order to dismiss the claim and compel arbitration, Lamps Plus. He claims the company “lacked standing to appeal the dismissal,” because the District Court’s order “provided precisely the relief Lamps Plus sought.” Brief for Respondent 13, 15. But Lamps Plus did not secure the relief it requested. It sought an order compelling individual arbitration. What it got was an order rejecting that relief and instead compelling arbitration on a classwide basis. We have explained—and will elaborate further below—that shifting from individual to class arbitration is a “fundamental” change, Stolt-Nielsen, 559 U. S., at 686, that “sacrifices the principal advantage of arbitration” and “greatly increases risks to defendants,” AT&T Mobility LLC v. Concepcion, 563 U.S. 333, 348, 350 (2011). Lamps Plus’s interest in avoiding those consequences gives it the “necessary personal stake in the appeal” required by our precedent. Camreta v. Greene, 563 U.S. 692, 702 (2011).[2] III The Ninth Circuit applied California contract law to conclude that the parties’ agreement was ambiguous on the availability of class arbitration. In California, an agreement is ambiguous “when it is capable of two or more constructions, both of which are reasonable.” 701 Fed. Appx., at 672 (quoting Powerine Oil Co. v. Superior Ct., 37 Cal. 4th 377, 390, 118 P.3d 589, 598 (2005)). Following our normal practice, we defer to the Ninth Circuit’s interpretation and application of state law and thus accept that the agreement should be regarded as ambiguous. See, e.g., Expressions Hair Design v. Schneiderman, 581 U. S. ___, ___ (2017) (slip op., at 7).[3] We therefore face the question whether, consistent with the FAA, an ambiguous agreement can provide the necessary “contractual basis” for compelling class arbitration. Stolt-Nielsen, 559 U. S., at 684. We hold that it cannot—a conclusion that follows directly from our decision in Stolt-Nielsen. Class arbitration is not only markedly different from the “traditional individualized arbitration” contemplated by the FAA, it also undermines the most important benefits of that familiar form of arbitration. Epic Systems, 584 U. S., at ___ (slip op., at 8); see Stolt-Nielsen, 559 U. S., at 686–687. The statute therefore requires more than ambiguity to ensure that the parties actually agreed to arbitrate on a classwide basis. A The FAA requires courts to “enforce arbitration agreements according to their terms.” Epic Systems, 584 U. S., at ___ (slip op., at 5) (quoting American Express Co. v. Italian Colors Restaurant, 570 U.S. 228, 233 (2013)). Although courts may ordinarily accomplish that end by relying on state contract principles, First Options of Chicago, Inc. v. Kaplan, 514 U.S. 938, 944 (1995), state law is preempted to the extent it “stands as an obstacle to the accomplishment and execution of the full purposes and objectives” of the FAA, Concepcion, 563 U. S., at 352 (internal quotation marks omitted). At issue in this case is the interaction between a state contract principle for addressing ambiguity and a “rule[ ] of fundamental importance” under the FAA, namely, that arbitration “is a matter of consent, not coercion.” Stolt-Nielsen, 559 U. S., at 681 (internal quotation marks omitted). “[T]he first principle that underscores all of our arbitration decisions” is that “[a]rbitration is strictly a matter of consent.” Granite Rock Co. v. Teamsters, 561 U.S. 287, 299 (2010) (internal quotation marks omitted). We have emphasized that “foundational FAA principle” many times. Stolt-Nielsen, 559 U. S., at 684; see also, e.g., Howsam v. Dean Witter Reynolds, Inc., 537 U.S. 79, 83 (2002); First Options, 514 U. S., at 943; Mastrobuono v. Shearson Lehman Hutton, Inc., 514 U.S. 52, 57 (1995); Volt Information Sciences, Inc. v. Board of Trustees of Leland Stanford Junior Univ., 489 U.S. 468, 479 (1989); Mitsubishi Motors Corp. v. Soler Chrysler-Plymouth, Inc., 473 U.S. 614, 626 (1985). Consent is essential under the FAA because arbitrators wield only the authority they are given. That is, they derive their “powers from the parties’ agreement to forgo the legal process and submit their disputes to private dispute resolution.” Stolt-Nielsen, 559 U. S., at 682. Parties may generally shape such agreements to their liking by specifying with whom they will arbitrate, the issues subject to arbitration, the rules by which they will arbitrate, and the arbitrators who will resolve their disputes. Id., at 683–684. Whatever they settle on, the task for courts and arbitrators at bottom remains the same: “to give effect to the intent of the parties.” Id., at 684. In carrying out that responsibility, it is important to recognize the “fundamental” difference between class arbitration and the individualized form of arbitration envisioned by the FAA. Epic Systems, 584 U. S., at ___ (slip op., at 8); see also Concepcion, 563 U. S., at 349, 351; Stolt-Nielsen, 559 U. S., at 686–687. In individual arbitration, “parties forgo the procedural rigor and appellate review of the courts in order to realize the benefits of private dispute resolution: lower costs, greater efficiency and speed, and the ability to choose expert adjudicators to resolve specialized disputes.” Id., at 685. Class arbitration lacks those benefits. It “sacrifices the principal advantage of arbitration—its informality—and makes the process slower, more costly, and more likely to generate procedural morass than final judgment.” Concepcion, 563 U. S., at 348. Indeed, we recognized just last Term that with class arbitration “the virtues Congress originally saw in arbitration, its speed and simplicity and inexpensiveness, would be shorn away and arbitration would wind up looking like the litigation it was meant to displace.” Epic Systems, 584 U. S., at ___ (slip op., at 8). Class arbitration not only “introduce[s] new risks and costs for both sides,” ibid., it also raises serious due process concerns by adjudicating the rights of absent members of the plaintiff class—again, with only limited judicial review. See Concepcion, 563 U. S., 349; see also Stolt-Nielsen, 559 U. S., at 686 (citing Ortiz v. Fibreboard Corp., 527 U.S. 815, 846 (1999)). Because of these “crucial differences” between individ- ual and class arbitration, Stolt-Nielsen explained that there is “reason to doubt the parties’ mutual consent to resolve disputes through classwide arbitration.” 559 U. S., at 687, 685–686. And for that reason, we held that courts may not infer consent to participate in class arbitration absent an affirmative “contractual basis for concluding that the party agreed to do so.” Id., at 684. Silence is not enough; the “FAA requires more.” Id., at 687. Our reasoning in Stolt-Nielsen controls the question we face today. Like silence, ambiguity does not provide a sufficient basis to conclude that parties to an arbitration agreement agreed to “sacrifice[ ] the principal advantage of arbitration.” Concepcion, 563 U. S., at 348. This conclusion aligns with our refusal to infer consent when it comes to other fundamental arbitration questions. For example, we presume that parties have not authorized arbitrators to resolve certain “gateway” questions, such as “whether the parties have a valid arbitration agreement at all or whether a concededly binding arbitration clause applies to a certain type of controversy.” Green Tree Financial Corp. v. Bazzle, 539 U.S. 444, 452 (2003) (plural- ity opinion). Although parties are free to authorize arbitrators to resolve such questions, we will not conclude that they have done so based on “silence or ambiguity” in their agreement, because “doing so might too often force unwilling parties to arbitrate a matter they reasonably would have thought a judge, not an arbitrator, would decide.” First Options, 514 U. S., at 945 (emphasis added); see also Howsam, 537 U. S., at 83–84. We relied on that same reasoning in Stolt-Nielsen, 559 U. S., at 686–687, and it applies with equal force here. Neither silence nor ambiguity provides a sufficient basis for concluding that parties to an arbitration agreement agreed to undermine the central benefits of arbitration itself.[4] B The Ninth Circuit reached a contrary conclusion based on California’s rule that ambiguity in a contract should be construed against the drafter, a doctrine known as contra proferentem. The rule applies “only as a last resort” when the meaning of a provision remains ambiguous after exhausting the ordinary methods of interpretation. 3 A. Corbin, Contracts §559, pp. 268–270 (1960). At that point, contra proferentem resolves the ambiguity against the drafter based on public policy factors, primarily equitable considerations about the parties’ relative bargaining strength. See 2 E. Farnsworth, Contracts §7.11, pp. 300–304 (3d ed. 2004); see also 11 R. Lord, Williston on Contracts §32:12, pp. 788–792 (4th ed. 2012) (stating that application of the rule may vary based on “the degree of sophistication of the contracting parties or the degree to which the contract was negotiated”); Restatement (Second) of Contracts §206, pp. 80–81, 105–107 (1979) (classifying contra proferentem under “Considerations of Fairness and the Public Interest” rather than with rules for interpreting “The Meaning of Agreements”); 3 Corbin, Contracts §559, at 270 (noting that contra proferentem is “chiefly a rule of public policy”). Although the rule enjoys a place in every hornbook and treatise on contracts, we noted in a recent FAA case that “the reach of the canon construing contract language against the drafter must have limits, no matter who the drafter was.” DIRECTV, Inc. v. Imburgia, 577 U. S. ___, ___ (2015) (slip op., at 10). This case brings those limits into focus. Unlike contract rules that help to interpret the meaning of a term, and thereby uncover the intent of the parties, contra proferentem is by definition triggered only after a court determines that it cannot discern the intent of the parties. When a contract is ambiguous, contra proferentem provides a default rule based on public policy considerations; “it can scarcely be said to be designed to ascertain the meanings attached by the parties.” 2 Farnsworth, Contracts §7.11, at 303. Like the contract rule preferring interpretations that favor the public interest, see id., at 304, contra proferentem seeks ends other than the intent of the parties. “[C]lass arbitration, to the extent it is manufactured by [state law] rather than consen[t], is inconsistent with the FAA.” Concepcion, 563 U. S., at 348. We recently reiterated that courts may not rely on state contract principles to “reshape traditional individualized arbitration by mandating classwide arbitration procedures without the parties’ consent.” Epic Systems, 584 U. S., at ___ (slip op., at 8). But that is precisely what the court below did, requiring class arbitration on the basis of a doctrine that “does not help to determine the meaning that the two parties gave to the words, or even the meaning that a reasonable person would have given to the language used.” 3 Corbin, Contracts §559, at 269–270. Such an approach is flatly inconsistent with “the foundational FAA principle that arbitration is a matter of consent.” Stolt-Nielsen, 559 U. S., at 684. Varela and Justice Kagan defend application of the rule on the basis that it is nondiscriminatory. It does not conflict with the FAA, they argue, because it is a neutral rule that gives equal treatment to arbitration agreements and other contracts alike. See Brief for Respondent 18, 25–26; post, at 6–9 (Kagan, J., dissenting). We have explained, however, that such an equal treatment principle cannot save from preemption general rules “that target arbitration either by name or by more subtle methods, such as by ‘interfer[ing] with fundamental attributes of arbitration.’ ” Epic Systems, 584 U. S., at ___ (slip op., at 7) (quoting Concepcion, 563 U. S., at 344). That was the case in Concepcion. There, the Court considered the general contract defense of unconscionability, which had been interpreted by the state court to bar class action waivers in consumer contracts, whether in the litigation or arbitration context. See id., at 341–344. The general applicability of the rule did not save it from preemption under the FAA with respect to arbitration agreements, because it had the consequence of allowing any party to a consumer arbitration agreement to demand class proceedings “without the parties’ consent.” Epic Systems, 584 U. S., at ___ (slip op., at 8) (describing the “essential insight” of Concepcion). That, for the reasons we have explained, “interferes with fundamental attributes of arbitration and thus creates a scheme inconsistent with the FAA.” Concepcion, 563 U. S., at 344; see Epic Systems, 584 U. S., at ___–___ (slip op., at 8–9). The same reasoning applies here: The general contra proferentem rule cannot be applied to impose class arbitration in the absence of the parties’ consent.[5] Our opinion today is far from the watershed Justice Kagan claims it to be. Rather, it is consistent with a long line of cases holding that the FAA provides the default rule for resolving certain ambiguities in arbitration agreements. For example, we have repeatedly held that ambiguities about the scope of an arbitration agreement must be resolved in favor of arbitration. See, e.g., Mitsubishi Motors Corp., 473 U. S., at 626; Moses H. Cone Memorial Hospital v. Mercury Constr. Corp., 460 U.S. 1, 24–25 (1983). In those cases, we did not seek to resolve the ambiguity by asking who drafted the agreement. Instead, we held that the FAA itself provided the rule. As in those cases, the FAA provides the default rule for resolving ambiguity here. * * * Courts may not infer from an ambiguous agreement that parties have consented to arbitrate on a classwide basis. The doctrine of contra proferentem cannot substitute for the requisite affirmative “contractual basis for concluding that the part[ies] agreed to [class arbitration].” Stolt-Nielsen, 559 U. S., at 684. We reverse the judgment of the Court of Appeals for the Ninth Circuit and remand the case for further proceedings consistent with this opinion. It is so ordered. Notes 1 Justice Breyer repeatedly refers to the order in this case as “interlocutory,” post, at 5–7 (dissenting opinion), but—as the language quoted above makes clear—Randolph expressly held that such an order is “final” under the FAA. Justice Breyer also claims that Randolph “explicitly reserved the [jurisdictional] question that we face now,” post, at 7, but Randolph reserved a different question. In that case, the District Court had denied a motion to stay. We noted that, if the District Court had entered a stay instead of dismissing the case, an appeal would have been barred by 9 U. S. C. §16(b)(1). That said, we expressly refrained from addressing whether the District Court should have granted the stay. See 531 U. S., at 87, n. 2. That is the question we reserved. Justice Breyer would have us take up that question today, post, at 3, 7, but there is no basis for doing so. The FAA provides that a district court “shall on application of one of the parties stay” the case pending the arbitration. 9 U. S. C. §3 (emphasis added). Here, no party sought a stay. Thus, Justice Breyer’s jurisdictional analysis is premised on two events that did not happen—a District Court ruling that was never issued denying a stay request that was never made. In short, Justice Breyer has written an opinion for a case other than the one before us. 2 And contrary to Varela’s contention, Brief for Respondent 14–15, and Justice Breyer’s dissent, post, at 6–7, this is hardly a case like Microsoft Corp. v. Baker, 582 U. S. ___ (2017). There, we held that plaintiffs cannot generate a final appealable order by voluntarily dismissing their claim. Here, Lamps Plus was the defendant, and the District Court compelled class arbitration over the company’s vigorous opposition. 3 Justice Kagan offers her own interpretation of the contract, concludes that it unambiguously authorizes class arbitration, post, at 2–4, and criticizes us for “disregard[ing] the actual contract the parties signed,” post, at 14. Justice Sotomayor, on the other hand, concludes that the contract is ambiguous about class arbitration but criticizes us for treating the contract as . . . ambiguous. Post, at 2–3 (dissenting opinion). Again, we simply follow this Court’s ordinary approach, which “accord[s] great deference” to the courts of appeals in their interpretation of state law. Expressions Hair Design, 581 U. S., at ___, (slip op., at 7) (quoting Pembaur v. Cincinnati, 475 U.S. 469, 484 n. 13 (1986) (collecting cases)). 4 This Court has not decided whether the availability of class arbitration is a so-called “question of arbitrability,” which includes these gateway matters. Oxford Health Plans LLC v. Sutter, 569 U.S. 564, 569, n. 2 (2013). We have no occasion to address that question here because the parties agreed that a court, not an arbitrator, should resolve the question about class arbitration. 5 Varela and Justice Kagan contend that our use of contra proferentem in Mastrobuono v. Shearson Lehman Hutton, Inc., 514 U.S. 52, 57 (1995), establishes that the rule is not preempted by the FAA. Brief for Respondent 33–35; post, at 8 (dissenting opinion). In Mastrobuono, however, we had no occasion to consider a conflict between the FAA and contra proferentem because both rules led to the same result. Our holding was primarily based on the FAA policy favoring arbitration, 514 U. S., at 62, and only after establishing that did we apply contra proferentem, noting that the rule was “well suited to the facts of this case,” id., at 63. See also EEOC v. Waffle House, Inc., 534 U.S. 279, 293, n. 9 (2002) (explaining that Mastrobuono resolved an ambiguous provision by “read[ing] the agreement to favor arbitration under the FAA rules”).

587.US.2018_17-1077

Securities and Exchange Commission Rule 10b–5 makes it unlawful to (a) “employ any device, scheme, or artifice to defraud,” (b) “make any untrue statement of a material fact,” or (c) “engage in any act, practice, or course of business” that “operates . . . as a fraud or deceit” in connection with the purchase or sale of securities. In Janus Capital Group, Inc. v. First Derivative Traders, 564 U.S. 135, this Court held that to be a “maker” of a statement under subsection (b) of that Rule, one must have “ultimate authority over the statement, including its content and whether and how to communicate it.” Id., at 142 (emphasis added). On the facts of Janus, this meant that an investment adviser who had merely “participat[ed] in the drafting of a false statement” “made” by another could not be held liable in a private action under subsection (b). Id., at 145. Petitioner Francis Lorenzo, while the director of investment banking at an SEC-registered brokerage firm, sent two e-mails to prospective investors. The content of those e-mails, which Lorenzo’s boss supplied, described a potential investment in a company with “confirmed assets” of $10 million. In fact, Lorenzo knew that the company had recently disclosed that its total assets were worth less than $400,000. In 2015, the Commission found that Lorenzo had violated Rule 10b–5, §10(b) of the Exchange Act, and §17(a)(1) of the Securities Act by sending false and misleading statements to investors with intent to defraud. On appeal, the District of Columbia Circuit held that Lorenzo could not be held liable as a “maker” under subsection (b) of the Rule in light of Janus, but sustained the Commission’s finding with respect to subsections (a) and (c) of the Rule, as well as §10(b) and §17(a)(1). Held: Dissemination of false or misleading statements with intent to defraud can fall within the scope of Rules 10b–5(a) and (c), as well as the relevant statutory provisions, even if the disseminator did not “make” the statements and consequently falls outside Rule 10b–5(b). Pp. 5–13. (a) It would seem obvious that the words in these provisions are, as ordinarily used, sufficiently broad to include within their scope the dissemination of false or misleading information with the intent to defraud. By sending e-mails he understood to contain material untruths, Lorenzo “employ[ed]” a “device,” “scheme,” and “artifice to defraud” within the meaning of subsection (a) of the Rule, §10(b), and §17(a)(1). By the same conduct, he “engage[d] in a[n] act, practice, or course of business” that “operate[d] . . . as a fraud or deceit” under subsection (c) of the Rule. As Lorenzo does not challenge the appeals court’s scienter finding, it is undisputed that he sent the e-mails with “intent to deceive, manipulate, or defraud” the recipients. Aaron v. SEC, 446 U.S. 680, 686, and n. 5. Resort to the expansive dictionary definitions of “device,” “scheme,” and “artifice” in Rule 10b–5(a) and §17(a)(1), and of “act” and “practice” in Rule 10b–5(c), only strengthens this conclusion. Under the circumstances, it is difficult to see how Lorenzo’s actions could escape the reach of these provisions. Pp. 5–7. (b) Lorenzo counters that the only way to be liable for false statements is through those provisions of the securities laws—like Rule 10b–5(b)—that refer specifically to false statements. Holding to the contrary, he and the dissent say, would render subsection (b) “superfluous.” The premise of this argument is that each subsection governs different, mutually exclusive, spheres of conduct. But this Court and the Commission have long recognized considerable overlap among the subsections of the Rule and related provisions of the securities laws. And the idea that each subsection governs a separate type of conduct is difficult to reconcile with the Rule’s language, since at least some conduct that amounts to “employ[ing]” a “device, scheme, or artifice to defraud” under subsection (a) also amounts to “engag[ing] in a[n] act . . . which operates . . . as a fraud” under subsection (c). This Court’s conviction is strengthened by the fact that the plainly fraudulent behavior confronted here might otherwise fall outside the Rule’s scope. Using false representations to induce the purchase of securities would seem a paradigmatic example of securities fraud. Pp. 7–9. (c) Lorenzo and the dissent make a few other important arguments. The dissent contends that applying Rules 10b–5(a) and (c) to conduct like Lorenzo’s would render Janus “a dead letter.” Post, at 9. But Janus concerned subsection (b), and it said nothing about the Rule’s application to the dissemination of false or misleading information. Thus, Janus would remain relevant (and preclude liability) where an individual neither makes nor disseminates false information—provided, of course, that the individual is not involved in some other form of fraud. Lorenzo also claims that imposing primary liability upon his conduct would erase or at least weaken the distinction between primary and secondary liability under the statute’s “aiding and abetting” provision. See 15 U. S. C. §78t(e). But the line the Court adopts today is clear: Those who disseminate false statements with intent to defraud are primarily liable under Rules 10b–5(a) and (c), §10(b), and §17(a)(1), even if they are secondarily liable under Rule 10b–5(b). As for Lorenzo’s suggestion that those like him ought to be held secondarily liable, this offer will, too often, prove illusory. Where a “maker” of a false statement does not violate subsection (b) of the Rule (perhaps because he lacked the necessary intent), a disseminator of those statements, even one knowingly engaged in an egregious fraud, could not be held to have violated the “aiding and abetting” statute. And if, as Lorenzo claims, the disseminator has not primarily violated other parts of Rule 10b–5, then such a fraud, whatever its intent or consequences, might escape liability altogether. That anomalous result is not what Congress intended. Pp. 9–13. 872 F.3d 578, affirmed. Breyer, J., delivered the opinion of the Court, in which Roberts, C. J., and Ginsburg, Alito, Sotomayor, and Kagan, JJ., joined. Thomas, J., filed a dissenting opinion, in which Gorsuch, J., joined. Kav- anaugh, J., took no part in the consideration or decision of the case.

Securities and Exchange Commission Rule 10b–5 makes it unlawful: “(a) To employ any device, scheme, or artifice to defraud, “(b) To make any untrue statement of a material fact . . . , or “(c) To engage in any act, practice, or course of business which operates or would operate as a fraud or deceit . . . in connection with the purchase or sale of any security.” 17 CFR §240.10b–5 (2018). In Janus Capital Group, Inc. v. First Derivative Traders, 564 U.S. 135 (2011), we examined the second of these provisions, Rule 10b–5(b), which forbids the “mak[ing]” of “any untrue statement of a material fact.” We held that the “maker of a statement is the person or entity with ultimate authority over the statement, including its content and whether and how to communicate it.” Id., at 142 (emphasis added). We said that “[w]ithout control, a person or entity can merely suggest what to say, not ‘make’ a statement in its own right.” Ibid. And we illustrated our holding with an analogy: “[W]hen a speechwriter drafts a speech, the content is entirely within the control of the person who delivers it. And it is the speaker who takes credit—or blame—for what is ultimately said.” Id., at 143. On the facts of Janus, this meant that an investment adviser who had merely “participat[ed] in the drafting of a false statement” “made” by another could not be held liable in a private action under subsection (b) of Rule 10b–5. Id., at 145. In this case, we consider whether those who do not “make” statements (as Janus defined “make”), but who disseminate false or misleading statements to potential investors with the intent to defraud, can be found to have violated the other parts of Rule 10b–5, subsections (a) and (c), as well as related provisions of the securities laws, §10(b) of the Securities Exchange Act of 1934, 48Stat. 891, as amended, 15 U. S. C. §78j(b), and §17(a)(1) of the Securities Act of 1933, 48Stat. 84–85, as amended, 15 U. S. C. §77q(a)(1). We believe that they can. I A For our purposes, the relevant facts are not in dispute. Francis Lorenzo, the petitioner, was the director of investment banking at Charles Vista, LLC, a registered broker-dealer in Staten Island, New York. Lorenzo’s only investment banking client at the time was Waste2Energy Holdings, Inc., a company developing technology to convert “solid waste” into “clean renewable energy.” In a June 2009 public filing, Waste2Energy stated that its total assets were worth about $14 million. This figure included intangible assets, namely, intellectual property, valued at more than $10 million. Lorenzo was skeptical of this valuation, later testifying that the intangibles were a “dead asset” because the technology “didn’t really work.” During the summer and early fall of 2009, Waste2Energy hired Lorenzo’s firm, Charles Vista, to sell to investors $15 million worth of debentures, a form of “debt secured only by the debtor’s earning power, not by a lien on any specific asset,” Black’s Law Dictionary 486 (10th ed. 2014). In early October 2009, Waste2Energy publicly disclosed, and Lorenzo was told, that its intellectual property was worthless, that it had “ ‘ “[w]rit[ten] off . . . all [of its] intangible assets,” ’ ” and that its total assets (as of March 31, 2009) amounted to $370,552. Shortly thereafter, on October 14, 2009, Lorenzo sent two e-mails to prospective investors describing the debenture offering. According to later testimony by Lorenzo, he sent the e-mails at the direction of his boss, who supplied the content and “approved” the messages. The e-mails described the investment in Waste2Energy as having “3 layers of protection,” including $10 million in “confirmed assets.” The e-mails nowhere revealed the fact that Waste2Energy had publicly stated that its assets were in fact worth less than $400,000. Lorenzo signed the e-mails with his own name, he identified himself as “Vice President—Investment Banking,” and he invited the recipients to “call with any questions.” B In 2013, the Securities and Exchange Commission instituted proceedings against Lorenzo (along with his boss and Charles Vista). The Commission charged that Lorenzo had violated Rule 10b–5, §10(b) of the Exchange Act, and §17(a)(1) of the Securities Act. Ultimately, the Commission found that Lorenzo had run afoul of these provisions by sending false and misleading statements to investors with intent to defraud. As a sanction, it fined Lorenzo $15,000, ordered him to cease and desist from violating the securities laws, and barred him from working in the securities industry for life. Lorenzo appealed, arguing primarily that in sending the e-mails he lacked the intent required to establish a violation of Rule 10b–5, §10(b), and §17(a)(1), which we have characterized as “ ‘a mental state embracing intent to deceive, manipulate, or defraud.’ ” Aaron v. SEC, 446 U.S. 680, 686, and n. 5 (1980). With one judge dissenting, the Court of Appeals panel rejected Lorenzo’s lack-of-intent argument. 872 F.3d 578, 583 (CADC 2017). Lorenzo does not challenge the panel’s scienter finding. Reply Brief 17. Lorenzo also argued that, in light of Janus, he could not be held liable under subsection (b) of Rule 10b–5. 872 F. 3d, at 586–587. The panel agreed. Because his boss “asked Lorenzo to send the emails, supplied the central content, and approved the messages for distribution,” id., at 588, it was the boss that had “ultimate authority” over the content of the statement “and whether and how to communicate it,” Janus, 563 U. S., at 142. (We took this case on the assumption that Lorenzo was not a “maker” under subsection (b) of Rule 10b–5, and do not revisit the court’s decision on this point.) The Court of Appeals nonetheless sustained (with one judge dissenting) the Commission’s finding that, by knowingly disseminating false information to prospective investors, Lorenzo had violated other parts of Rule 10b–5, subsections (a) and (c), as well as §10(b) and §17(a)(1). Lorenzo then filed a petition for certiorari in this Court. We granted review to resolve disagreement about whether someone who is not a “maker” of a misstatement under Janus can nevertheless be found to have violated the other subsections of Rule 10b–5 and related provisions of the securities laws, when the only conduct involved concerns a misstatement. Compare e.g., 872 F.3d 578, with WPP Luxembourg Gamma Three Sarl v. Spot Runner, Inc., 655 F.3d 1039, 1057–1058 (CA9 2011). II A At the outset, we review the relevant provisions of Rule 10b–5 and of the statutes. See Appendix, infra. As we have said, subsection (a) of the Rule makes it unlawful to “employ any device, scheme, or artifice to defraud.” Subsection (b) makes it unlawful to “make any untrue statement of a material fact.” And subsection (c) makes it unlawful to “engage in any act, practice, or course of business” that “operates . . . as a fraud or deceit.” See 17 CFR §240.10b–5. There are also two statutes at issue. Section 10(b) makes it unlawful to “use or employ . . . any manipulative or deceptive device or contrivance” in contravention of Commission rules and regulations. 15 U. S. C. §78j(b). By its authority under that section, the Commission promulgated Rule 10b–5. The second statutory provision is §17(a), which, like Rule 10b–5, is organized into three subsections. 15 U. S. C. §77q(a). Here, however, we consider only the first subsection, §17(a)(1), for this is the only subsection that the Commission charged Lorenzo with violating. Like Rule 10b–5(a), (a)(1) makes it unlawful to “employ any device, scheme, or artifice to defraud.” B After examining the relevant language, precedent, and purpose, we conclude that (assuming other here-irrelevant legal requirements are met) dissemination of false or misleading statements with intent to defraud can fall within the scope of subsections (a) and (c) of Rule 10b–5, as well as the relevant statutory provisions. In our view, that is so even if the disseminator did not “make” the statements and consequently falls outside subsection (b) of the Rule. It would seem obvious that the words in these provisions are, as ordinarily used, sufficiently broad to include within their scope the dissemination of false or misleading information with the intent to defraud. By sending emails he understood to contain material untruths, Lorenzo “employ[ed]” a “device,” “scheme,” and “artifice to defraud” within the meaning of subsection (a) of the Rule, §10(b), and §17(a)(1). By the same conduct, he “engage[d] in a[n] act, practice, or course of business” that “operate[d] . . . as a fraud or deceit” under subsection (c) of the Rule. Recall that Lorenzo does not challenge the appeals court’s scienter finding, so we take for granted that he sent the emails with “intent to deceive, manipulate, or defraud” the recipients. Aaron, 446 U. S., at 686, n. 5. Under the circumstances, it is difficult to see how his actions could escape the reach of those provisions. Resort to dictionary definitions only strengthens this conclusion. A “ ‘device,’ ” we have observed, is simply “ ‘[t]hat which is devised, or formed by design’ ”; a “ ‘scheme’ ” is a “ ‘project,’ ” “ ‘plan[,] or program of something to be done’ ”; and an “ ‘artifice’ ” is “ ‘an artful stratagem or trick.’ ” Id., at 696, n. 13 (quoting Webster’s International Dictionary 713, 2234, 157 (2d ed. 1934) (Webster’s Second)). By these lights, dissemination of false or misleading material is easily an “artful stratagem” or a “plan,” “devised” to defraud an investor under subsection (a). See Rule 10b–5(a) (making it unlawful to “employ any device, scheme, or artifice to defraud”); §17(a)(1) (same). The words “act” and “practice” in subsection (c) are similarly expansive. Webster’s Second 25 (defining “act” as “a doing” or a “thing done”); id., at 1937 (defining “practice” as an “action” or “deed”); see Rule 10b–5(c) (making it unlawful to “engage in a[n] act, practice, or course of business” that “operates . . . as a fraud or deceit”). These provisions capture a wide range of conduct. Applying them may present difficult problems of scope in borderline cases. Purpose, precedent, and circumstance could lead to narrowing their reach in other contexts. But we see nothing borderline about this case, where the relevant conduct (as found by the Commission) consists of disseminating false or misleading information to prospective investors with the intent to defraud. And while one can readily imagine other actors tangentially involved in dissemination—say, a mailroom clerk—for whom liability would typically be inappropriate, the petitioner in this case sent false statements directly to investors, invited them to follow up with questions, and did so in his capacity as vice president of an investment banking company. C Lorenzo argues that, despite the natural meaning of these provisions, they should not reach his conduct. This is so, he says, because the only way to be liable for false statements is through those provisions that refer specifically to false statements. Other provisions, he says, concern “scheme liability claims” and are violated only when conduct other than misstatements is involved. Brief for Petitioner 4–6, 28–30. Thus, only those who “make” untrue statements under subsection (b) can violate Rule 10b–5 in connection with statements. (Similarly, §17(a)(2) would be the sole route for finding liability for statements under §17(a).) Holding to the contrary, he and the dissent insist, would render subsection (b) of Rule 10b–5 “superfluous.” See post, at 6–7 (opinion of Thomas, J.). The premise of this argument is that each of these provisions should be read as governing different, mutually exclusive, spheres of conduct. But this Court and the Commission have long recognized considerable overlap among the subsections of the Rule and related provisions of the securities laws. See Herman & MacLean v. Huddleston, 459 U.S. 375, 383 (1983) (“[I]t is hardly a novel proposition that” different portions of the securities laws “prohibit some of the same conduct” (internal quotation marks omitted)). As we have explained, these laws marked the “first experiment in federal regulation of the securities industry.” SEC v. Capital Gains Research Bureau, Inc., 375 U.S. 180, 198 (1963). It is “understandable, therefore,” that “in declaring certain practices unlawful,” it was thought prudent “to include both a general proscription against fraudulent and deceptive practices and, out of an abundance of caution, a specific proscription against nondisclosure” even though “a specific proscription against nondisclosure” might in other circumstances be deemed “surplusage.” Id., at 198–199. “Each succeeding prohibition” was thus “meant to cover additional kinds of illegalities—not to narrow the reach of the prior sections.” United States v. Naftalin, 441 U.S. 768, 774 (1979). We have found “ ‘no warrant for narrowing alternative provisions . . . adopted with the purpose of affording added safeguards.’ ” Ibid. (quoting United States v. Gilliland, 312 U.S. 86, 93 (1941)); see Affiliated Ute Citizens of Utah v. United States, 406 U.S. 128, 152–153 (1972) (While “the second subparagraph of [Rule 10b–5] specifies the making of an untrue statement . . . [t]he first and third subparagraphs are not so restricted”). And since its earliest days, the Commission has not viewed these provisions as mutually exclusive. See, e.g., In re R. D. Bayly & Co., 19 S. E. C. 773 (1945) (finding violations of what would become Rules 10b–5(b) and (c) based on the same misrepresentations and omissions); In re Arthur Hays & Co., 5 S. E. C. 271 (1939) (finding violations of both §§17(a)(2) and (a)(3) based on false representations in stock sales). The idea that each subsection of Rule 10b–5 governs a separate type of conduct is also difficult to reconcile with the language of subsections (a) and (c). It should go without saying that at least some conduct amounts to “employ[ing]” a “device, scheme, or artifice to defraud” under subsection (a) as well as “engag[ing] in a[n] act . . . which operates . . . as a fraud” under subsection (c). In Affiliated Ute, for instance, we described the “defendants’ activities” as falling “within the very language of one or the other of those subparagraphs, a ‘course of business’ or a ‘device, scheme, or artifice’ that operated as a fraud.” 406 U. S., at 153. (The dissent, for its part, offers no account of how the superfluity problems that motivate its interpretation can be avoided where subsections (a) and (c) are concerned.) Coupled with the Rule’s expansive language, which readily embraces the conduct before us, this considerable overlap suggests we should not hesitate to hold that Lorenzo’s conduct ran afoul of subsections (a) and (c), as well as the related statutory provisions. Our conviction is strengthened by the fact that we here confront behavior that, though plainly fraudulent, might otherwise fall outside the scope of the Rule. Lorenzo’s view that subsection (b), the making-false-statements provision, exclusively regulates conduct involving false or misleading statements would mean those who disseminate false statements with the intent to cheat investors might escape liability under the Rule altogether. But using false representations to induce the purchase of securities would seem a paradigmatic example of securities fraud. We do not know why Congress or the Commission would have wanted to disarm enforcement in this way. And we cannot easily reconcile Lorenzo’s approach with the basic purpose behind these laws: “to substitute a philosophy of full disclosure for the philosophy of caveat emptor and thus to achieve a high standard of business ethics in the securities industry.” Capital Gains, 375 U. S., at 186. See also, e.g., SEC v. W. J. Howey Co., 328 U.S. 293, 299 (1946) (the securities laws were designed “to meet the countless and variable schemes devised by those who seek the use of the money of others on the promise of profits”). III Lorenzo and the dissent make a few other important arguments. They contend that applying subsections (a) or (c) of Rule 10b–5 to conduct like his would render our decision in Janus (which we described at the outset, supra, at 1–2) “a dead letter,” post, at 9. But we do not see how that is so. In Janus, we considered the language in subsection (b), which prohibits the “mak[ing]” of “any untrue statement of a material fact.” See 564 U. S., at 141–143. We held that the “maker” of a “statement” is the “person or entity with ultimate authority over the statement.” Id., at 142. And we found that subsection (b) did not (under the circumstances) cover an investment adviser who helped draft misstatements issued by a different entity that controlled the statements’ content. Id., at 146–148. We said nothing about the Rule’s application to the dissemination of false or misleading information. And we can assume that Janus would remain relevant (and preclude liability) where an individual neither makes nor disseminates false information—provided, of course, that the individual is not involved in some other form of fraud. Next, Lorenzo points to the statute’s “aiding and abetting” provision. 15 U. S. C. §78t(e). This provision, enforceable only by the Commission (and not by private parties), makes it unlawful to “knowingly or recklessly . . . provid[e] substantial assistance to another person” who violates the Rule. Ibid.; see Janus, 564 U. S., at 143 (citing Central Bank of Denver, N. A. v. First Interstate Bank of Denver, N. A., 511 U.S. 164 (1994)). Lorenzo claims that imposing primary liability upon his conduct would erase or at least weaken what is otherwise a clear distinction between primary and secondary (i.e., aiding and abetting) liability. He emphasizes that, under today’s holding, a disseminator might be a primary offender with respect to subsection (a) of Rule 10b–5 (by employing a “scheme” to “defraud”) and also secondarily liable as an aider and abettor with respect to subsection (b) (by providing substantial assistance to one who “makes” a false statement). And he refers to two cases that, in his view, argue in favor of circumscribing primary liability. See Central Bank, 511 U. S., at 164; Stoneridge Investment Partners, LLC v. Scientific-Atlanta, Inc., 552 U.S. 148 (2008). We do not believe, however, that our decision creates a serious anomaly or otherwise weakens the distinction between primary and secondary liability. For one thing, it is hardly unusual for the same conduct to be a primary violation with respect to one offense and aiding and abetting with respect to another. John, for example, might sell Bill an unregistered firearm in order to help Bill rob a bank, under circumstances that make him primarily li- able for the gun sale and secondarily liable for the bank robbery. For another, the cases to which Lorenzo refers do not help his cause. Take Central Bank, where we held that Rule 10b–5’s private right of action does not permit suits against secondary violators. 511 U. S., at 177. The holding of Central Bank, we have said, suggests the need for a “clean line” between conduct that constitutes a primary violation of Rule 10b–5 and conduct that amounts to a secondary violation. Janus, 564 U. S., at 143, and n. 6. Thus, in Janus, we sought an interpretation of “make” that could neatly divide primary violators and actors too far removed from the ultimate decision to communicate a statement. Ibid. (citing Central Bank, 511 U. S. 164). The line we adopt today is just as administrable: Those who disseminate false statements with intent to defraud are primarily liable under Rules 10b–5(a) and (c), §10(b), and §17(a)(1), even if they are secondarily liable under Rule 10b–5(b). Lorenzo suggests that classifying dissemination as a primary violation would inappropriately subject peripheral players in fraud (including him, naturally) to substantial liability. We suspect the investors who received Lorenzo’s e-mails would not view the deception so favorably. And as Central Bank itself made clear, even a bit participant in the securities markets “may be liable as a primary violator under [Rule] 10b–5” so long as “all of the requirements for primary liability . . . are met.” Id., at 191. Lorenzo’s reliance on Stoneridge is even further afield. There, we held that private plaintiffs could not bring suit against certain securities defendants based on undisclosed deceptions upon which the plaintiffs could not have relied. 552 U. S., at 159. But the Commission, unlike private parties, need not show reliance in its enforcement actions. And even supposing reliance were relevant here, Lorenzo’s conduct involved the direct transmission of false statements to prospective investors intended to induce reliance—far from the kind of concealed fraud at issue in Stoneridge. As for Lorenzo’s suggestion that those like him ought to be held secondarily liable, this offer will, far too often, prove illusory. In instances where a “maker” of a false statement does not violate subsection (b) of the Rule (perhaps because he lacked the necessary intent), a disseminator of those statements, even one knowingly engaged in an egregious fraud, could not be held to have violated the “aiding and abetting” statute. That is because the statute insists that there be a primary violator to whom the secondary violator provided “substantial assistance.” 15 U. S. C. §78t(e). And the latter can be “deemed to be in violation” of the provision only “to the same extent as the person to whom such assistance is provided.” Ibid. In other words, if Acme Corp. could not be held liable under subsection (b) for a statement it made, then a knowing disseminator of those statements could not be held liable for aiding and abetting Acme under subsection (b). And if, as Lorenzo claims, the disseminator has not primarily violated other parts of Rule 10b–5, then such a fraud, whatever its intent or consequences, might escape liability altogether. That is not what Congress intended. Rather, Congress intended to root out all manner of fraud in the securities industry. And it gave to the Commission the tools to accomplish that job. * * * For these reasons, the judgment of the Court of Appeals is affirmed. So ordered. Justice Kavanaugh took no part in the consideration or decision of this case. APPENDIX 17 CFR §240.10b–5 “It shall be unlawful for any person, directly or indirectly, by the use of any means or instrumentality of interstate commerce, or of the mails or of any facility of any national securities exchange, “(a) To employ any device, scheme, or artifice to defraud, “(b) To make any untrue statement of a material fact or to omit to state a material fact necessary in order to make the statements made, in the light of the circumstances under which they were made, not misleading, or “(c) To engage in any act, practice, or course of business which operates or would operate as a fraud or deceit upon any person in connection with the purchase or sale of any security.” 15 U. S. C. §78j “It shall be unlawful for any person, directly or in- directly, by the use of any means or instrumentality of in- terstate commerce or of the mails, or of any facility of any national securities exchange— * * * “(b) To use or employ, in connection with the purchase or sale of any security registered on a national securities ex- change or any security not so registered, or any securities-based swap agreement[,] any manipulative or decep- tive device or contrivance in contravention of such rules and regulations as the Commission may prescribe as necessary or appropriate in the public interest or for the protection of investors.” 15 U. S. C. §77q “(a) Use of interstate commerce for purpose of fraud or deceit “It shall be unlawful for any person in the offer or sale of any securities (including security-based swaps) or any security-based swap agreement . . . by the use of any means or instruments of transportation or communication in interstate commerce or by use of the mails, directly or indirectly— “(1) to employ any device, scheme, or artifice to defraud, or “(2) to obtain money or property by means of any untrue statement of a material fact or any omission to state a material fact necessary in order to make the statements made, in light of the circumstances under which they were made, not misleading; or “(3) to engage in any transaction, practice, or course of business which operates or would operate as a fraud or deceit upon the purchaser.” 15 U. S. C. §78t “(e) Prosecution of persons who aid and abet violations “For purposes of any action brought by the Commission . . . , any person that knowingly or recklessly provides substantial assistance to another person in violation of a provision of this chapter, or of any rule or regulation issued under this chapter, shall be deemed in violation of such provision to the same extent as the person to whom such assistance is provided.

586.US.2018_17-7505

In Ford v. Wainwright, 477 U.S. 399, this Court held that the Eighth Amendment’s ban on cruel and unusual punishments precludes executing a prisoner who has “lost his sanity” after sentencing. Id., at 406. And in Panetti v. Quarterman, 551 U.S. 930, the Court set out the appropriate competency standard: A State may not execute a prisoner whose “mental state is so distorted by a mental illness” that he lacks a “rational understanding” of “the State’s rationale for [his] execution.” Id., at 958–959. Petitioner Vernon Madison was found guilty of capital murder and sentenced to death. While awaiting execution, he suffered a series of strokes and was diagnosed with vascular dementia. In 2016, Madison petitioned the state trial court for a stay of execution on the ground that he was mentally incompetent, stressing that he could not recollect committing the crime for which he had been sentenced to die. Alabama responded that Madison had a rational understanding of the reasons for his execution, even assuming he had no memory of committing his crime. And more broadly, the State claimed that Madison failed to implicate Ford and Panetti because both decisions concerned themselves with gross delusions, which Madison did not have. Following a competency hearing, the trial court found Madison competent to be executed. On federal habeas review, this Court summarily reversed the Eleventh Circuit’s grant of relief, holding that, under the “demanding” and “deferential standard” of the Antiterrorism and Effective Death Penalty Act of 1996 (AEDPA), “[n]either Panetti nor Ford ‘clearly established’ that a prisoner is incompetent to be executed” because of a simple failure to remember his crime. Dunn v. Madison, 583 U. S. ___, ___. But the Court “express[ed] no view” on the question of Madison’s competency outside of the AEDPA context. Id., at ___. When Alabama set a 2018 execution date, Madison returned to state court, arguing once more that his mental condition precluded the State from going forward. The state court again found Madison mentally competent. Held: 1. Under Ford and Panetti, the Eighth Amendment may permit executing a prisoner even if he cannot remember committing his crime. Panetti asks only about a person’s comprehension of the State’s reasons for resorting to punishment, not his memory of the crime itself. And the one may exist without the other. Such memory loss, however, still may factor into the analysis Panetti demands. If that loss combines and interacts with other mental shortfalls to deprive a person of the capacity to comprehend why the State is exacting death as a punishment, then the Panetti standard will be satisfied. Pp. 9–11. 2. Under Ford and Panetti, the Eighth Amendment may prohibit executing a prisoner even though he suffers from dementia or another disorder rather than psychotic delusions. The Panetti standard focuses on whether a mental disorder has had a particular effect; it has no interest in establishing any precise cause. Panetti’s references to “gross delusions,” 551 U. S., at 960, are no more than a predictable byproduct of that case’s facts. Ford and Panetti hinge on the prisoner’s “[in]comprehension of why he has been singled out” to die, 477 U. S., 409, and kick in if and when that failure of understanding is present, irrespective of whether one disease or another is to blame. In evaluating competency, a judge must therefore look beyond any given diagnosis to a downstream consequence. Pp. 12–14. 3. Because this Court is uncertain whether the state court’s decision was tainted by legal error, this case is remanded to that court for renewed consideration of Madison’s competency. The state court’s brief 2018 ruling—which states only that Madison “did not prove a substantial threshold showing of insanity[ ]”—does not provide any assurance that the court knew a person with dementia, and not psychotic delusions, might receive a stay of execution. Nor does that court’s initial 2016 opinion. The sole question on which Madison’s competency depends is whether he can reach a rational understanding of why the State wants to execute him. In answering that question—on which this Court again expresses no view—the state court may not rely on any arguments or evidence tainted with the legal errors addressed by this Court. Pp. 14–18. Vacated and remanded. Kagan, J., delivered the opinion of the Court, in which Roberts, C. J., and Ginsburg, Breyer, and Sotomayor, JJ., joined. Alito, J., filed a dissenting opinion, in which Thomas and Gorsuch, JJ., joined. Kavanaugh, J., took no part in the consideration or decision of the case.

The Eighth Amendment, this Court has held, prohibits the execution of a prisoner whose mental illness prevents him from “rational[ly] understanding” why the State seeks to impose that punishment. Panetti v. Quarterman, 551 U.S. 930, 959 (2007). In this case, Vernon Madison argued that his memory loss and dementia entitled him to a stay of execution, but an Alabama court denied the relief. We now address two questions relating to the Eighth Amendment’s bar, disputed below but not in this Court. First, does the Eighth Amendment forbid execution whenever a prisoner shows that a mental disorder has left him without any memory of committing his crime? We (and, now, the parties) think not, because a person lacking such a memory may still be able to form a rational understanding of the reasons for his death sentence. Second, does the Eighth Amendment apply similarly to a prisoner suffering from dementia as to one experiencing psychotic delusions? We (and, now, the parties) think so, because either condition may—or, then again, may not—impede the requisite comprehension of his punishment. The only issue left, on which the parties still disagree, is what those rulings mean for Madison’s own execution. We direct that issue to the state court for further consideration in light of this opinion. I A This Court decided in Ford v. Wainwright, 477 U.S. 399 (1986), that the Eighth Amendment’s ban on cruel and unusual punishments precludes executing a prisoner who has “lost his sanity” after sentencing. Id., at 406. While on death row, Alvin Ford was beset by “pervasive delusion[s]” associated with “[p]aranoid [s]chizophrenia.” Id., at 402–403. Surveying both the common law and state statutes, the Court found a uniform practice against taking the life of such a prisoner. See id., at 406–409. Among the reasons for that time-honored bar, the Court explained, was a moral “intuition” that “killing one who has no capacity” to understand his crime or punishment “simply offends humanity.” Id., at 407, 409; see id., at 409 (citing the “natural abhorrence civilized societies feel” at performing such an act). Another rationale rested on the lack of “retributive value” in executing a person who has no comprehension of the meaning of the community’s judgment. Ibid.; see id., at 421 (Powell, J., concurring in part and concurring in judgment) (stating that the death penalty’s “retributive force[ ] depends on the defendant’s awareness of the penalty’s existence and purpose”). The resulting rule, now stated as a matter of constitutional law, held “a category of defendants defined by their mental state” incompetent to be executed. Id., at 419. The Court clarified the scope of that category in Panetti v. Quarterman by focusing on whether a prisoner can “reach a rational understanding of the reason for [his] execution.” 551 U. S., at 958. Like Alvin Ford, Scott Panetti suffered from “gross delusions” stemming from “extreme psychosis.” Id., at 936, 960. In reversing a ruling that he could still be executed, the Panetti Court set out the appropriate “standard for competency.” Id., at 957. Ford, the Court now noted, had not provided “specific criteria.” 551 U. S., at 957. But Ford had explored what lay behind the Eighth Amendment’s prohibition, highlighting that the execution of a prisoner who cannot comprehend the reasons for his punishment offends moral values and “serves no retributive purpose.” 551 U. S., at 958. Those principles, the Panetti Court explained, indicate how to identify prisoners whom the State may not execute. The critical question is whether a “prisoner’s mental state is so distorted by a mental illness” that he lacks a “rational understanding” of “the State’s rationale for [his] execution.” Id., at 958–959. Or similarly put, the issue is whether a “prisoner’s concept of reality” is “so impair[ed]” that he cannot grasp the execution’s “meaning and purpose” or the “link between [his] crime and its punishment.” Id., at 958, 960. B Vernon Madison killed a police officer in 1985 during a domestic dispute. An Alabama jury found him guilty of capital murder, and the trial court sentenced him to death. He has spent most of the ensuing decades on the State’s death row. In recent years, Madison’s mental condition has sharply deteriorated. Madison suffered a series of strokes, including major ones in 2015 and 2016. See Tr. 19, 46–47 (Apr. 14, 2016). He was diagnosed as having vascular dementia, with attendant disorientation and confusion, cognitive impairment, and memory loss. See id., at 19–20, 52–54. In particular, Madison claims that he can no longer recollect committing the crime for which he has been sentenced to die. See Tr., Pet. Exh. 2, p. 8. After his 2016 stroke, Madison petitioned the trial court for a stay of execution on the ground that he had become mentally incompetent. Citing Ford and Panetti, he argued that “he no longer understands” the “status of his case” or the “nature of his conviction and sentence.” Pet. for Suspension in No. CC–85–1385.80 (C. C. Mobile Cty., Ala., Feb. 12, 2016), pp. 11, 14. And in a later filing, Madison emphasized that he could not “independently recall the facts of the offense he is convicted of.” Brief Pursuant to Order (Apr. 21, 2016), p. 8. Alabama countered that Madison had “a rational understanding of [the reasons for] his impending execution,” as required by Ford and Panetti, even assuming he had no memory of committing his crime. Brief on Madison’s Competency (April 21, 2016), pp. 4–5, 8. And more broadly, the State claimed that Madison could not possibly qualify as incompetent under those two decisions because both “concerned themselves with ‘[g]ross delusions’ ”—which all agree Madison does not have. Id., at 2; see ibid. (Madison “failed to implicate” Ford and Panetti because he “does not suffer from psychosis or delusions”). Expert reports from two psychologists largely aligned with the parties’ contending positions. Dr. John Goff, Madison’s expert, found that although Madison “underst[ood] the nature of execution” in the abstract, he did not comprehend the “reasoning behind” Alabama’s effort to execute him. Tr., Pet. Exh. 2 (Apr. 14, 2016), p. 8; see id., at 9. Goff stated that Madison had “Major Vascular Neurological Disorder”—also called vascular dementia—which had caused “significant cognitive decline.” Ibid. And Goff underscored that Madison “demonstrate[d] retrograde amnesia” about his crime, meaning that he had no “independent recollection[ ]” of the murder. Id., at 8; see id., at 9. For his part, Dr. Karl Kirkland, the court-appointed expert, reported that Madison “was able to discuss his case” accurately and “appear[ed] to understand his legal situation.” Tr., Ct. Exh. 1, pp. 10–11. Although Kirkland acknowledged that Madison’s strokes had led to cognitive decline, see id., at 10, the psychologist made no men- tion of Madison’s diagnosed vascular dementia. Rather, Kirkland highlighted that “[t]here was no evidence of psychosis, paranoia, or delusion.” Id., at 9; see ibid. (Madison “did not seem delusional at all”). At a competency hearing, Alabama similarly stressed Madison’s absence of psychotic episodes or delusions. The State asked both experts to affirm that Madison was “neither delusional [n]or psychotic.” Tr. 56; see id., at 22. And its closing argument focused on their agreement that he was not. As the State summarized: “He’s not psychotic. He’s not delusional.” Id., at 81. On the State’s view, that fact answered the competency question because “[t]he Supreme Court is looking at whether someone’s delusions or someone’s paranoia or someone’s psychosis is standing in the way of” rationally understanding his punishment. Id., at 82. Madison’s counsel disputed that point. “[T]he State would like to say, well, he’s not delusional, he’s not psychotic,” the attorney recapped. Id., at 83. But, she continued, “[t]hat’s not really the criteria” under Panetti. Tr. 83. Rather, the Court there barred executing a person with any mental illness—“dementia” and “brain injuries” no less than psychosis and delusions—that prevents him from comprehending “why he is being executed.” Ibid. The trial court found Madison competent to be executed. Its order first recounted the evidence given by each expert witness. The summary of Kirkland’s report and testimony began by stating that the psychologist had “found no evidence of paranoia[,] delusion [or] psychosis.” Order (Apr. 29, 2016), p. 5 (2016 Order). The court then noted Kirkland’s view that Madison could “give details of the history of his case” and “appear[ed] to understand his legal situation.” Ibid. Turning to the Goff report, the court noted the expert’s finding that Madison was “amnesic” and could not recollect his crime. Id., at 6; see id., at 7. In a single, final paragraph, the court provided both its ruling and its reasoning. Madison had failed to show, the court wrote, that he did not “rationally understand the punishment he is about to suffer and why he is about to suffer it.” Id., at 10. The court “accept[ed] the testimony of Dr. Kirkland as to the understanding Madison has concerning the situation.” Ibid. “Further,” the court concluded, “the evidence does not support that Mr. Madison is delusional.” Ibid. Madison next sought habeas relief in federal court, where he faced the heavy burden of showing that the state-court ruling “involved an unreasonable application of[ ] clearly established federal law” or rested on an “unreasonable determination of the facts.” Antiterrorism and Effective Death Penalty Act of 1996 (AEDPA), 28 U. S. C. §2254(d). The District Court rejected his petition, but the Court of Appeals for the Eleventh Circuit ruled that Madison had demonstrated both kinds of indisputable error. See Madison v. Commissioner, 851 F.3d 1173 (2017). This Court then summarily reversed the appeals court’s decision. See Dunn v. Madison, 583 U. S. ___ (2017) (per curiam). We explained, contrary to the Eleventh Circuit’s principal holding, that “[n]either Panetti nor Ford ‘clearly established’ that a prisoner is incompetent to be executed” because of a simple failure to remember his crime. Id., at ___ (slip op., at 4). And we found that the state court did not act unreasonably—otherwise put, did not err “beyond any possibility for fairminded disagreement”—when it found that Madison had the necessary understanding to be executed. Ibid. (internal quotation marks omitted). But we made clear that our decision was premised on AEDPA’s “demanding” and “deferential standard.” Id., at ___, ___ (slip op., at 3, 4). “We express[ed] no view” on the question of Madison’s competency “outside of the AEDPA context.” Id., at ___ (slip op., at 4).[1] When Alabama set an execution date in 2018, Madison returned to state court to argue again that his mental condition precluded the State from going forward. In his petition, Madison reiterated the facts and arguments he had previously presented to the state court. But Madison also claimed that since that court’s decision (1) he had suffered further cognitive decline and (2) a state board had suspended Kirkland’s license to practice psychology, thus discrediting his prior testimony. See Pet. to Suspend Execution in No. CC–85–1385.80 (C. C. Mobile Cty., Ala., Dec. 18, 2017), pp. 1–2, 16–19.[2] Alabama responded that nothing material had changed since the court’s first competency hearing. See Motion to Dismiss (Dec. 20, 2017), p. 9. The State also repeated its argument that Panetti permits executing Madison, pointing to the experts’ agreement that he is “not delusional or psychotic” and asserting that neither “memory impairment [n]or dementia [could] suffice to satisfy the Panetti and Ford standards” without “an expansion” of those decisions. Motion to Dismiss 4, 10. A week before the scheduled execution, the state court again found Madison mentally competent. Its brief order stated only that Madison “did not provide a substantial threshold showing of insanity[ ] sufficient to convince this Court to stay the execution.” App. A to Pet. for Cert. Madison then filed in this Court a request to stay his execution and a petition for certiorari. We ordered the stay on the scheduled execution date and granted the petition a few weeks later. See 583 U. S. ___, ___ (2018). Because the case now comes to us on direct review of the state court’s decision (rather than in a habeas proceeding), AEDPA’s deferential standard no longer governs. (And for that reason—contrary to the dissent’s suggestion, post, at 12—our decision on Madison’s habeas petition cannot help resolve the questions raised here.) II Two issues relating to Panetti’s application are before us. Recall that our decision there held the Eighth Amendment to forbid executing a prisoner whose mental illness makes him unable to “reach a rational understanding of the reason for [his] execution.” 551 U. S., at 958; see supra, at 2–3. The first question presented is whether Panetti prohibits executing Madison merely because he cannot remember committing his crime. The second question raised is whether Panetti permits executing Madison merely because he suffers from dementia, rather than psychotic delusions.[3] In prior stages of this case, as we have described, the parties disagreed about those matters. See supra, at 4–8. But at this Court, Madison accepted Alabama’s positon on the first issue and Alabama accepted Madison’s on the second. See, e.g., Tr. of Oral Arg. 11, 36. And rightly so. As the parties now recognize, the standard set out in Panetti supplies the answers to both questions. First, a person lacking memory of his crime may yet rationally understand why the State seeks to execute him; if so, the Eighth Amendment poses no bar to his execution. Second, a person suffering from dementia may be unable to rationally understand the reasons for his sentence; if so, the Eighth Amendment does not allow his execution. What matters is whether a person has the “rational understanding” Panetti requires—not whether he has any particular memory or any particular mental illness. A Consider initially a person who cannot remember his crime because of a mental disorder, but who otherwise has full cognitive function. The memory loss is genuine: Let us say the person has some kind of amnesia, which has produced a black hole where that recollection should be. But the person remains oriented in time and place; he can make logical connections and order his thoughts; and he comprehends familiar concepts of crime and punishment. Can the State execute him for a murder? When we considered this case before, using the deferential standard applicable in habeas, we held that a state court could allow such an execution without committing inarguable error. See Madison, 583 U. S., at ___ (slip op., at 4) (stating that no prior decision had “clearly established” the opposite); supra, at 6. Today, we address the issue straight-up, sans any deference to a state court. Again, is the failure to remember committing a crime alone enough to prevent a State from executing a prisoner? It is not, under Panetti’s own terms. That decision asks about understanding, not memory—more specifically, about a person’s understanding of why the State seeks capital punishment for a crime, not his memory of the crime itself. And the one may exist without the other. Do you have an independent recollection of the Civil War? Obviously not. But you may still be able to reach a rational—indeed, a sophisticated—understanding of that conflict and its consequences. Do you recall your first day of school? Probably not. But if your mother told you years later that you were sent home for hitting a classmate, you would have no trouble grasping the story. And similarly, if you somehow blacked out a crime you committed, but later learned what you had done, you could well appreciate the State’s desire to impose a penalty. Assuming, that is, no other cognitive impairment, loss of memory of a crime does not prevent rational understanding of the State’s reasons for resorting to punishment. And that kind of comprehension is the Panetti standard’s singular focus. The same answer follows from the core justifications Panetti offered for framing its Eighth Amendment test as it did. Echoing Ford, Panetti reasoned that execution has no retributive value when a prisoner cannot appreciate the meaning of a community’s judgment. See 551 U. S., at 958–959 (citing 477 U. S., at 407–408); supra, at 3. But as just explained, a person who can no longer remember a crime may yet recognize the retributive message society intends to convey with a death sentence. Similarly, Ford and Panetti stated that it “offends humanity” to execute a person so wracked by mental illness that he cannot comprehend the “meaning and purpose of the punishment.” 477 U. S., at 407; 551 U. S., at 960; see id., at 958. But that offense to morality must be much less when a person’s mental disorder causes nothing more than an episodic memory loss. Moral values do not exempt the simply forgetful from punishment, whatever the neurological reason for their lack of recall. But such memory loss still may factor into the “rational understanding” analysis that Panetti demands. If that loss combines and interacts with other mental shortfalls to deprive a person of the capacity to comprehend why the State is exacting death as punishment, then the Panetti standard will be satisfied. That may be so when a person has difficulty preserving any memories, so that even newly gained knowledge (about, say, the crime and punishment) will be quickly forgotten. Or it may be so when cognitive deficits prevent the acquisition of such knowledge at all, so that memory gaps go forever uncompensated. As Panetti indicated, neurologists, psychologists, and other experts can contribute to a court’s understanding of issues of that kind. See id., at 962. But the sole inquiry for the court remains whether the prisoner can rationally understand the reasons for his death sentence. B Next consider a prisoner who suffers from dementia or a similar disorder, rather than psychotic delusions. The dementia, as is typical, has compromised this prisoner’s cognitive functions. But it has not resulted in the kind of delusional beliefs that Alvin Ford and Scott Panetti held. May the prisoner nonetheless receive a stay of execution under Ford and Panetti? Or instead, is a delusional disorder a prerequisite to declaring a mentally ill person incompetent to be executed? We did not address that issue when we last considered this case, on habeas review; in that sense, the question is one of first impression. See supra, at 6, n. 1. But here too, Panetti has already answered the question. Its standard focuses on whether a mental disorder has had a particular effect: an inability to rationally understand why the State is seeking execution. See supra, at 2–3. Conversely, that standard has no interest in establishing any precise cause: Psychosis or dementia, delusions or overall cognitive decline are all the same under Panetti, so long as they produce the requisite lack of comprehension. To be sure, Panetti on occasion spoke of “gross delusions” in explaining its holding. 551 U. S., at 960. And similarly, Ford talked about the “insane,” which sometimes refers to persons holding such irrational beliefs. See, e.g., 477 U. S., at 401, 410.[4] But those references are no more than a predictable byproduct of the two cases’ facts. At the same time (and interchangeably), Panetti used more inclusive terms, such as “mental illness,” “mental disorder,” and “psychological dysfunction.” 551 U. S., at 936, 959, 960; see Ford, 477 U. S., at 408–409, n. 2 (referring to prisoners with “mental illness”). And most important, Panetti framed its test, as just described, in a way utterly indifferent to a prisoner’s specific mental illness. The Panetti standard concerns, once again, not the diagnosis of such illness, but a consequence—to wit, the prisoner’s inability to rationally understand his punishment. And here too, the key justifications Ford and Panetti offered for the Eighth Amendment’s bar confirm our conclusion about its reach. As described above, those decisions stated that an execution lacks retributive purpose when a mentally ill prisoner cannot understand the societal judgment underlying his sentence. See Panetti, 551 U. S., at 958–959; Ford, 477 U. S., at 409; supra, at 2–3. And they indicated that an execution offends morality in the same circumstance. See 551 U. S., at 958, 960; 477 U. S., at 409; supra, at 2–3. Both rationales for the constitutional bar thus hinge (just as the Panetti standard deriving from them does) on the prisoner’s “[in]comprehension of why he has been singled out” to die. 477 U. S., at 409; see supra, at 2–3. Or said otherwise, if and when that failure of understanding is present, the rationales kick in—irrespective of whether one disease or another (say, psychotic delusions or dementia) is to blame. In evaluating competency to be executed, a judge must therefore look beyond any given diagnosis to a downstream consequence. As Ford and Panetti recognized, a delusional disorder can be of such severity—can “so impair the prisoner’s concept of reality”—that someone in its thrall will be unable “to come to grips with” the punishment’s meaning. Panetti, 551 U. S., at 958; Ford, 477 U. S., at 409. But delusions come in many shapes and sizes, and not all will interfere with the understanding that the Eighth Amendment requires. See Panetti, 551 U. S., at 962 (remanding the case to consider expert evidence on whether the prisoner’s delusions did so). And much the same is true of dementia. That mental condition can cause such disorientation and cognitive decline as to prevent a person from sustaining a rational understanding of why the State wants to execute him. See supra, at 11–12. But dementia also has milder forms, which allow a person to preserve that understanding. Hence the need—for dementia as for delusions as for any other mental disorder—to attend to the particular circumstances of a case and make the precise judgment Panetti requires. III The only question left—and the only one on which the parties now disagree—is whether Madison’s execution may go forward based on the state court’s decision below. Madison’s counsel says it cannot because that ruling was tainted by legal error—specifically, the idea that only delusions, and not dementia, can support a finding of mental incompetency. See Tr. of Oral Arg. 12, 21, 25, 27. Alabama counters that the state court did not rely on that (concededly) incorrect view of the law. See id., at 37–41. But we come away at the least unsure whether that is so—especially given Alabama’s evidence and arguments in the state court. As noted earlier, the 2018 ruling we review today contains only one sentence of explanation. See supra, at 7–8. It states that Madison “did not provide a substantial threshold showing of insanity[ ] sufficient to convince this Court to stay the execution.” App. A to Pet. for Cert. If the state court used the word “insanity” to refer to a delusional disorder, then error occurred: The court would have denied a stay on the ground that Madison did not have that specific kind of mental illness. And the likelihood that the court made that mistake is heightened by the State’s emphasis, at that stage of the proceedings (as at others), that Madison was “not delusional or psychotic” and that “dementia” could not suffice to bar his execution absent “an expansion of Ford and Panetti.” Motion to Dismiss 4, 10; see supra, at 4–8; but see post, at 9–10, and n. 4 (disregarding those arguments).[5] Alabama argues, however, that the court spoke of “insanity” only because the state statute under which Madison sought relief uses that term. See Tr. of Oral Arg. 37; Ala. Code §15–16–23 (2011) (allowing a stay of execution “on account of the [convict’s] insanity”). But even if so, that does not advance the State’s view that the state court properly understood the Eighth Amendment bar when assessing Madison’s competency. Alabama told this Court in opposing certiorari that its statute covers only those with delusional disorders, and not those with dementia. See Brief in Opposition 12 (“[T]he sole question to be answered under the state statute was whether Madison was insane, not whether he suffered from dementia”). The state court’s (supposed) echoing of statutory language understood in that way cannot provide assurance that the court knew a person with dementia might receive a stay of execution; indeed, it suggests exactly the opposite. The court’s 2018 order thus calls out for a do-over. Alabama further contends, however, that we should look past the state court’s 2018 decision to the court’s initial 2016 determination of competency. (The dissent similarly begins with the 2016 ruling, see post, at 6–7, even though that is not the decision under review here.) According to the State, nothing material changed in the interim period, see supra, at 7; thus, we may find the meaning of the later ruling in the earlier one, see Tr. of Oral Arg. 36–37. And, the State continues, the 2016 opinion gets the law right. Alabama’s proof is that the court, after summarizing the psychologists’ testimony, found that “Madison has a rational[ ] understanding, as required by Panetti,” concerning the “punishment he is about to suffer and why he is about to suffer it.” 2016 Order, at 10; see Tr. of Oral Arg. 39; supra, at 5–6. (The dissent quotes the same passage. See post, at 7.) But the state court’s initial decision does not aid Alabama’s cause. First, we do not know that the court in 2018 meant to incorporate everything in its prior opinion. The order says nothing to that effect; and though it came out the same way as the earlier decision, it need not have rested on all the same reasoning. Second, the 2016 opinion itself does not show that the state court realized that persons suffering from dementia could satisfy the Panetti standard. True enough, as Alabama says, that the court accurately stated that standard in its decision. But as described above, Alabama had repeatedly argued to the court (over Madison’s objection) that only prisoners suffering from delusional disorders could qualify as incompetent under Panetti. See, e.g., Brief on Madison’s Competency 2 (Madison “failed to implicate” Ford and Panetti because he “does not suffer from psychosis or delusions”); Tr. 82 (“The Supreme Court [in Panetti] is looking at whether someone’s delusions or someone’s paranoia or someone’s psychosis is standing in the way of” rationally understanding his punishment); see also supra, at 4–5; but see post, at 9–10, and n. 4 (disregarding those arguments). And Alabama relied on the expert opinion of a psychologist who highlighted Madison’s lack of “psychosis, paranoia, or delusion,” while never mentioning his dementia. Tr., Ct. Exh. 1 (Apr. 14, 2016), p. 9. That too-limited understanding of Panetti’s compass is reflected in the court’s 2016 opinion. In its single paragraph of analysis, the court “accept[ed] the testimony” of the State’s preferred psychologist.[6] And the court further found that “the evidence does not support that Mr. Madison is delusional”—without ever considering his undisputed dementia. 2016 Order, at 10. For those reasons, we must return this case to the state court for renewed consideration of Madison’s competency (assuming Alabama sets a new execution date). See, e.g., Kindred Nursing Centers L. P. v. Clark, 581 U. S. ___, ___ (2017) (slip op., at 9) (remanding when “uncertain” whether “an impermissible taint occurred”); Clemons v. Mississippi, 494 U.S. 738, 751–752 (1990) (similar). In that proceeding, two matters disputed below should now be clear. First, under Ford and Panetti, the Eighth Amendment may permit executing Madison even if he cannot remember committing his crime. Second, under those same decisions, the Eighth Amendment may prohibit executing Madison even though he suffers from dementia, rather than delusions. The sole question on which Madison’s competency depends is whether he can reach a “rational understanding” of why the State wants to execute him. Panetti, 551 U. S., at 958. In answering that question—on which we again express no view, see supra, at 6—the state court may not rely on any arguments or evidence tainted with the legal errors we have addressed. And because that is so, the court should consider whether it needs to supplement the existing record. Some evidence in that record, including portions of the experts’ reports and testimony, expressly reflects an incorrect view of the relevance of delusions or memory; still other evidence might have implicitly rested on those same misjudgments. The state court, we have little doubt, can evaluate such matters better than we. It must do so as the first step in assessing Madison’s competency—and ensuring that if he is to be executed, he understands why. We accordingly vacate the judgment of the state court and remand the case for further proceedings not inconsistent with this opinion. It is so ordered. Justice Kavanaugh took no part in the consideration or decision of this case. Notes 1 Neither did we opine on—or even mention—the subsidiary legal question whether a mental disorder other than delusions may render a person incompetent to be executed. Alabama told the Eleventh Circuit that it could not, thus reprising the claim the State had made in the trial court. See Madison, 851 F. 3d, at 1188 (describing Alabama’s argument that “only a prisoner suffering from gross delusions can show incompetency under Panetti”); Recording of Oral Arg. in No. 16–12279 (CA11, June 23, 2016), at 26:36–26:45 (“In this case, what we have is someone who claims to have a mental illness, dementia,” but does not have “delusions, which is what Panetti requires”); id., at 26:48–27:21 (When asked if someone with “severe dementia” but no delusions could be executed, the State responded “I think so because . . . they don’t have delusions”). (Alabama alternatively argued that the state court’s decision was not based on that view, see Brief for Appellee in No. 16–12279 (CA11), pp. 37–38; the quotations the dissent picks out, see post, at 10, n. 4, come from that additional argument.) The Eleventh Circuit rejected the State’s contention that dementia could not preclude an execution as “inconsistent with the principles underlying” Ford and Panetti. 851 F. 3d, at 1188. But we had no reason to address that holding in light of the errors we saw in other parts of the appeals court’s analysis. 2 As Madison’s petition recounted, the license suspension followed the opening of a criminal investigation into whether Kirkland had committed narcotics offenses. See Pet. to Suspend Execution 17–19. 3 The dissent is in high dudgeon over our taking up the second question, arguing that it was not presented in Madison’s petition for certiorari. See post, at 1–6. But that is incorrect. The petition presented two questions—the same two we address here. The first question asked whether the Eighth Amendment bars executing Madison because he has no “memory of his commission of the capital offense.” Pet. for Cert. iii. The second question asked whether that Amendment bars his execution because his “vascular dementia” and “severe cognitive dysfunction” prevent him from either remembering his crime “or understanding the circumstances of his scheduled execution.” Ibid. So the first question concerned whether memory loss alone could form the basis of a Panetti claim and the second whether the varied consequences of dementia could do so. The body of the petition, to be sure, devoted more space to the first question. But it clearly referenced the second. See Pet. for Cert. 18 (“[T]his Court has never sought to constrain the world of maladies that can give rise to a finding that a prisoner is incompetent to be executed”); id., at 25 (“[C]ourts have recognized dementia and attendant cognitive decline and memory impairment as a basis for a finding of incompetency to be executed”). And in any event, the number of words spent on each is not what matters. Our Rule states that the Court will consider “[o]nly the questions set out in the petition, or fairly included therein.” This Court’s Rule 14.1(a). Here, we consider, in order, the two questions set out in Madison’s petition. 4 Alternatively, however, the term may also be used to encompass persons with other mental conditions, so long as they are “severe enough [to] prevent[ ] a person from having legal capacity and excuse[ ] the person from criminal or civil responsibility.” Black’s Law Dictionary 914 (10th ed. 2014). In that different understanding, “insanity” connotes a general standard of legal competency rather than a more limited description of delusional disorders. 5 The State once again repeated that argument in its Brief in Opposition to Madison’s certiorari petition. See Brief in Opposition 11–12 (“Madison does not argue that he is insane. Instead, he argues that he suffers from dementia” and that his execution should be barred “under a yet-unannounced expansion of Ford and Panetti”). 6 The court well understood that expert’s exclusive focus on whether Madison had psychotic delusions. In summarizing his testimony, the court began as follows: “Dr. Kirkland in his exam found no evidence of paranoia or delusion at the time of his examin[ation], on March 31, 2016. He also found that there was no psychosis present.” 2016 Order, at 5; see supra, at 5.

587.US.2018_17-1702

New York state law requires cable operators to set aside channels on their cable systems for public access. Those channels are operated by the cable operator unless the local government chooses to itself operate the channels or designates a private entity to operate the channels. New York City (the City) has designated a private nonprofit corporation, petitioner Manhattan Neighborhood Network (MNN), to operate the public access channels on Time Warner’s cable system in Manhattan. Respondents DeeDee Halleck and Jesus Papoleto Melendez produced a film critical of MNN to be aired on MNN’s public access channels. MNN televised the film. MNN later suspended Halleck and Melendez from all MNN services and facilities. The producers sued, claiming that MNN violated their First Amendment free-speech rights when it restricted their access to the public access channels because of the content of their film. The District Court dismissed the claim on the ground that MNN is not a state actor and therefore is not subject to First Amendment constraints on its editorial discretion. Reversing in relevant part, the Second Circuit concluded that MNN is a state actor subject to First Amendment constraints. Held: MNN is not a state actor subject to the First Amendment. Pp. 5–16. (a) The Free Speech Clause of the First Amendment prohibits only governmental, not private, abridgment of speech. See, e.g., Denver Area Ed. Telecommunications Consortium, Inc. v. FCC, 518 U.S. 727, 737. This Court’s state-action doctrine distinguishes the government from individuals and private entities. Pp. 5–14. (1) A private entity may qualify as a state actor when, as relevant here, the entity exercises “powers traditionally exclusively reserved to the State.” Jackson v. Metropolitan Edison Co., 419 U.S. 345, 352. The Court has stressed that “very few” functions fall into that category. Flagg Bros., Inc. v. Brooks, 436 U.S. 149, 158. The relevant function in this case—operation of public access channels on a cable system—has not traditionally and exclusively been performed by government. Since the 1970s, a variety of private and public actors have operated public access channels. Early Manhattan public access channels were operated by private cable operators with some help from private nonprofit organizations. That practice continued until the early 1990s, when MNN began to operate the channels. Operating public access channels on a cable system is not a traditional, exclusive public function. Pp. 6–8. (2) The producers contend that the relevant function here is more generally the operation of a public forum for speech, which, they claim, is a traditional, exclusive public function. But that analysis mistakenly ignores the threshold state-action question. Providing some kind of forum for speech is not an activity that only governmental entities have traditionally performed. Therefore, a private entity who provides a forum for speech is not transformed by that fact alone into a state actor. See Hudgens v. NLRB, 424 U.S. 507, 520–521. Pp. 8–10. (3) The producers note that the City has designated MNN to operate the public access channels on Time Warner’s cable system, and that the State heavily regulates MNN with respect to those channels. But the City’s designation is analogous to a government license, a government contract, or a government-granted monopoly, none of which converts a private entity into a state actor—unless the private entity is performing a traditional, exclusive public function. See, e.g., San Francisco Arts & Athletics, Inc. v. United States Olympic Comm., 483 U.S. 522, 543–544. And the fact that MNN is subject to the State’s extensive regulation “does not by itself convert its action into that of the State.” Jackson, 419 U. S., at 350. Pp. 11–14. (b) The producers alternatively contend that the public access channels are actually the City’s property and that MNN is essentially managing government property on the City’s behalf. But the City does not own or lease the public access channels and does not possess any formal easement or other property interest in the channels. It does not matter that a provision in the franchise agreements between the City and Time Warner allowed the City to designate a private entity to operate the public access channels on Time Warner’s cable system. Nothing in the agreements suggests that the City possesses any property interest in the cable system or in the public access channels on that system. Pp. 14–15. 882 F.3d 300, reversed in part and remanded. Kavanaugh, J., delivered the opinion of the Court, in which Roberts, C. J., and Thomas, Alito, and Gorsuch, JJ., joined. Sotomayor, J., filed a dissenting opinion, in which Ginsburg, Breyer, and Kagan, JJ., joined.

The Free Speech Clause of the First Amendment constrains governmental actors and protects private actors. To draw the line between governmental and private, this Court applies what is known as the state-action doctrine. Under that doctrine, as relevant here, a private entity may be considered a state actor when it exercises a function “traditionally exclusively reserved to the State.” Jackson v. Metropolitan Edison Co., 419 U.S. 345, 352 (1974). This state-action case concerns the public access channels on Time Warner’s cable system in Manhattan. Public access channels are available for private citizens to use. The public access channels on Time Warner’s cable system in Manhattan are operated by a private nonprofit corporation known as MNN. The question here is whether MNN—even though it is a private entity—nonetheless is a state actor when it operates the public access channels. In other words, is operation of public access channels on a cable system a traditional, exclusive public function? If so, then the First Amendment would restrict MNN’s exercise of editorial discretion over the speech and speakers on the public access channels. Under the state-action doctrine as it has been articulated and applied by our precedents, we conclude that operation of public access channels on a cable system is not a traditional, exclusive public function. Moreover, a private entity such as MNN who opens its property for speech by others is not transformed by that fact alone into a state actor. In operating the public access channels, MNN is a private actor, not a state actor, and MNN therefore is not subject to First Amendment constraints on its editorial discretion. We reverse in relevant part the judgment of the Second Circuit, and we remand the case for further proceedings consistent with this opinion. I A Since the 1970s, public access channels have been a regular feature on cable television systems throughout the United States. In the 1970s, Federal Communications Commission regulations required certain cable operators to set aside channels on their cable systems for public access. In 1979, however, this Court ruled that the FCC lacked statutory authority to impose that mandate. See FCC v. Midwest Video Corp., 440 U.S. 689 (1979). A few years later, Congress passed and President Reagan signed the Cable Communications Policy Act of 1984. 98Stat. 2779. The Act authorized state and local governments to require cable operators to set aside channels on their cable systems for public access. 47 U. S. C. §531(b). The New York State Public Service Commission regulates cable franchising in New York State and requires cable operators in the State to set aside channels on their cable systems for public access. 16 N. Y. Codes, Rules & Regs. §§895.1(f), 895.4(b) (2018). State law requires that use of the public access channels be free of charge and first-come, first-served. §§895.4(c)(4) and (6). Under state law, the cable operator operates the public access channels unless the local government in the area chooses to itself operate the channels or designates a private entity to operate the channels. §895.4(c)(1). Time Warner (now known as Charter) operates a cable system in Manhattan. Under state law, Time Warner must set aside some channels on its cable system for public access. New York City (the City) has designated a private nonprofit corporation named Manhattan Neighborhood Network, commonly referred to as MNN, to operate Time Warner’s public access channels in Manhattan. This case involves a complaint against MNN regarding its management of the public access channels. B Because this case comes to us on a motion to dismiss, we accept the allegations in the complaint as true. See Ashcroft v. Iqbal, 556 U.S. 662, 678 (2009). DeeDee Halleck and Jesus Papoleto Melendez produced public access programming in Manhattan. They made a film about MNN’s alleged neglect of the East Harlem community. Halleck submitted the film to MNN for airing on MNN’s public access channels, and MNN later televised the film. Afterwards, MNN fielded multiple complaints about the film’s content. In response, MNN temporarily suspended Halleck from using the public access channels. Halleck and Melendez soon became embroiled in another dispute with MNN staff. In the wake of that dispute, MNN ultimately suspended Halleck and Melendez from all MNN services and facilities. Halleck and Melendez then sued MNN, among other parties, in Federal District Court. The two producers claimed that MNN violated their First Amendment free-speech rights when MNN restricted their access to the public access channels because of the content of their film. MNN moved to dismiss the producers’ First Amendment claim on the ground that MNN is not a state actor and therefore is not subject to First Amendment restrictions on its editorial discretion. The District Court agreed with MNN and dismissed the producers’ First Amendment claim. The Second Circuit reversed in relevant part. 882 F.3d 300, 308 (2018). In the majority opinion authored by Judge Newman and joined by Judge Lohier, the court stated that the public access channels in Manhattan are a public forum for purposes of the First Amendment. Reasoning that “public forums are usually operated by governments,” the court concluded that MNN is a state actor subject to First Amendment constraints. Id., at 306–307. Judge Lohier added a concurring opinion, explaining that MNN also qualifies as a state actor for the independent reason that “New York City delegated to MNN the traditionally public function of administering and regulating speech in the public forum of Manhattan’s public access channels.” Id., at 309. Judge Jacobs dissented in relevant part, opining that MNN is not a state actor. He reasoned that a private entity’s operation of an open forum for speakers does not render the host entity a state actor. Judge Jacobs further stated that the operation of public access channels is not a traditional, exclusive public function. We granted certiorari to resolve disagreement among the Courts of Appeals on the question whether private operators of public access cable channels are state actors subject to the First Amendment. 586 U. S. __ (2018). Compare 882 F.3d 300 (case below), with Wilcher v. Akron, 498 F.3d 516 (CA6 2007); and Alliance for Commu- nity Media v. FCC, 56 F.3d 105 (CADC 1995). II Ratified in 1791, the First Amendment provides in relevant part that “Congress shall make no law . . . abridging the freedom of speech.” Ratified in 1868, the Fourteenth Amendment makes the First Amendment’s Free Speech Clause applicable against the States: “No State shall make or enforce any law which shall abridge the privileges or immunities of citizens of the United States; nor shall any State deprive any person of life, liberty, or property, without due process of law . . . .” §1. The text and original meaning of those Amendments, as well as this Court’s longstanding precedents, establish that the Free Speech Clause prohibits only governmental abridgment of speech. The Free Speech Clause does not prohibit private abridgment of speech. See, e.g., Denver Area Ed. Telecommunications Consortium, Inc. v. FCC, 518 U.S. 727, 737 (1996) (plurality opinion); Hurley v. Irish-American Gay, Lesbian and Bisexual Group of Boston, Inc., 515 U.S. 557, 566 (1995); Hudgens v. NLRB, 424 U.S. 507, 513 (1976); cf. Miami Herald Publishing Co. v. Tornillo, 418 U.S. 241, 256 (1974). In accord with the text and structure of the Constitution, this Court’s state-action doctrine distinguishes the government from individuals and private entities. See Brentwood Academy v. Tennessee Secondary School Athletic Assn., 531 U.S. 288, 295–296 (2001). By enforcing that constitutional boundary between the governmental and the private, the state-action doctrine protects a robust sphere of individual liberty. Here, the producers claim that MNN, a private entity, restricted their access to MNN’s public access channels because of the content of the producers’ film. The producers have advanced a First Amendment claim against MNN. The threshold problem with that First Amendment claim is a fundamental one: MNN is a private entity. Relying on this Court’s state-action precedents, the producers assert that MNN is nonetheless a state actor subject to First Amendment constraints on its editorial discretion. Under this Court’s cases, a private entity can qualify as a state actor in a few limited circumstances—including, for example, (i) when the private entity performs a traditional, exclusive public function, see, e.g., Jackson, 419 U. S., at 352–354; (ii) when the government compels the private entity to take a particular action, see, e.g., Blum v. Yaretsky, 457 U.S. 991, 1004–1005 (1982); or (iii) when the government acts jointly with the private entity, see, e.g., Lugar v. Edmondson Oil Co., 457 U.S. 922, 941–942 (1982). The producers’ primary argument here falls into the first category: The producers contend that MNN exercises a traditional, exclusive public function when it operates the public access channels on Time Warner’s cable system in Manhattan. We disagree. A Under the Court’s cases, a private entity may qualify as a state actor when it exercises “powers traditionally exclusively reserved to the State.” Jackson, 419 U. S., at 352. It is not enough that the federal, state, or local government exercised the function in the past, or still does. And it is not enough that the function serves the public good or the public interest in some way. Rather, to qualify as a traditional, exclusive public function within the meaning of our state-action precedents, the government must have traditionally and exclusively performed the function. See Rendell-Baker v. Kohn, 457 U.S. 830, 842 (1982); Jackson, 419 U. S., at 352–353; Evans v. Newton, 382 U.S. 296, 300 (1966). The Court has stressed that “very few” functions fall into that category. Flagg Bros., Inc. v. Brooks, 436 U.S. 149, 158 (1978). Under the Court’s cases, those functions include, for example, running elections and operating a company town. See Terry v. Adams, 345 U.S. 461, 468–470 (1953) (elections); Marsh v. Alabama, 326 U.S. 501, 505–509 (1946) (company town); Smith v. Allwright, 321 U.S. 649, 662–666 (1944) (elections); Nixon v. Condon, 286 U.S. 73, 84–89 (1932) (elections).[1] The Court has ruled that a variety of functions do not fall into that category, including, for example: running sports associations and leagues, administering insurance payments, operating nursing homes, providing special education, representing indigent criminal defendants, resolving private disputes, and supplying electricity. See American Mfrs. Mut. Ins. Co. v. Sullivan, 526 U.S. 40, 55–57 (1999) (insurance payments); National Collegiate Athletic Assn. v. Tarkanian, 488 U.S. 179, 197, n. 18 (1988) (college sports); San Francisco Arts & Athletics, Inc. v. United States Olympic Comm., 483 U.S. 522, 544–545 (1987) (amateur sports); Blum, 457 U. S., at 1011–1012 (nursing home); Rendell-Baker, 457 U. S., at 842 (special education); Polk County v. Dodson, 454 U.S. 312, 318–319 (1981) (public defender); Flagg Bros., 436 U. S., at 157–163 (private dispute resolution); Jackson, 419 U. S., at 352–354 (electric service). The relevant function in this case is operation of public access channels on a cable system. That function has not traditionally and exclusively been performed by government. Since the 1970s, when public access channels became a regular feature on cable systems, a variety of private and public actors have operated public access channels, including: private cable operators; private nonprofit organizations; municipalities; and other public and private community organizations such as churches, schools, and libraries. See Denver Area, 518 U. S., at 761–762 (plurality opinion); R. Oringel & S. Buske, The Access Manager’s Handbook: A Guide for Managing Community Television 14–17 (1987). The history of public access channels in Manhattan further illustrates the point. In 1971, public access channels first started operating in Manhattan. See D. Brenner, M. Price, & M. Meyerson, Cable Television and Other Nonbroadcast Video §6:29, p. 6–47 (2018). Those early Manhattan public access channels were operated in large part by private cable operators, with some help from private nonprofit organizations. See G. Gillespie, Public Access Cable Television in the United States and Canada 37–38 (1975); Janes, History and Structure of Public Access Television, 39 J. Film & Video, No. 3, pp. 15–17 (1987). Those private cable operators continued to operate the public access channels until the early 1990s, when MNN (also a private entity) began to operate the public access channels. In short, operating public access channels on a cable system is not a traditional, exclusive public function within the meaning of this Court’s cases. B To avoid that conclusion, the producers widen the lens and contend that the relevant function here is not simply the operation of public access channels on a cable system, but rather is more generally the operation of a public forum for speech. And according to the producers, operation of a public forum for speech is a traditional, exclusive public function. That analysis mistakenly ignores the threshold state-action question. When the government provides a forum for speech (known as a public forum), the government may be constrained by the First Amendment, meaning that the government ordinarily may not exclude speech or speakers from the forum on the basis of viewpoint, or sometimes even on the basis of content. See, e.g., Southeastern Promotions, Ltd. v. Conrad, 420 U.S. 546, 547, 555 (1975) (private theater leased to the city); Police Dept. of Chicago v. Mosley, 408 U.S. 92, 93, 96 (1972) (sidewalks); Hague v. Committee for Industrial Organization, 307 U.S. 496, 515–516 (1939) (streets and parks). By contrast, when a private entity provides a forum for speech, the private entity is not ordinarily constrained by the First Amendment because the private entity is not a state actor. The private entity may thus exercise editorial discretion over the speech and speakers in the forum. This Court so ruled in its 1976 decision in Hudgens v. NLRB. There, the Court held that a shopping center owner is not a state actor subject to First Amendment requirements such as the public forum doctrine. 424 U. S., at 520–521; see also Lloyd Corp. v. Tanner, 407 U.S. 551, 569–570 (1972); Central Hardware Co. v. NLRB, 407 U.S. 539, 547 (1972); Alliance for Community Media, 56 F. 3d, at 121–123. The Hudgens decision reflects a commonsense principle: Providing some kind of forum for speech is not an activity that only governmental entities have traditionally performed. Therefore, a private entity who provides a forum for speech is not transformed by that fact alone into a state actor. After all, private property owners and private lessees often open their property for speech. Grocery stores put up community bulletin boards. Comedy clubs host open mic nights. As Judge Jacobs persuasively explained, it “is not at all a near-exclusive function of the state to provide the forums for public expression, politics, information, or entertainment.” 882 F. 3d, at 311 (opinion concurring in part and dissenting in part). In short, merely hosting speech by others is not a traditional, exclusive public function and does not alone transform private entities into state actors subject to First Amendment constraints. If the rule were otherwise, all private property owners and private lessees who open their property for speech would be subject to First Amendment constraints and would lose the ability to exercise what they deem to be appropriate editorial discretion within that open forum. Private property owners and private lessees would face the unappetizing choice of allowing all comers or closing the platform altogether. “The Constitution by no means requires such an attenuated doctrine of dedication of private property to public use.” Hudgens, 424 U. S., at 519 (internal quotation marks omitted). Benjamin Franklin did not have to operate his newspaper as “a stagecoach, with seats for everyone.” F. Mott, American Journalism 55 (3d ed. 1962). That principle still holds true. As the Court said in Hudgens, to hold that private property owners providing a forum for speech are constrained by the First Amendment would be “to create a court-made law wholly disregarding the constitutional basis on which private ownership of property rests in this country.” 424 U. S., at 517 (internal quotation marks omitted). The Constitution does not disable private property owners and private lessees from exercising editorial discretion over speech and speakers on their property.[2] The producers here are seeking in effect to circumvent this Court’s case law, including Hudgens. But Hudgens is sound, and we therefore reaffirm our holding in that case.[3] C Next, the producers retort that this case differs from Hudgens because New York City has designated MNN to operate the public access channels on Time Warner’s cable system, and because New York State heavily regulates MNN with respect to the public access channels. Under this Court’s cases, however, those facts do not establish that MNN is a state actor. New York City’s designation of MNN to operate the public access channels is analogous to a government license, a government contract, or a government-granted monopoly. But as the Court has long held, the fact that the government licenses, contracts with, or grants a monopoly to a private entity does not convert the private entity into a state actor—unless the private entity is performing a traditional, exclusive public function. See, e.g., San Francisco Arts & Athletics, 483 U. S., at 543–544 (exclusive-use rights and corporate charters); Blum, 457 U. S., at 1011 (licenses); Rendell-Baker, 457 U. S., at 840–841 (contracts); Polk County, 454 U. S., at 319, n. 9, and 320–322 (law licenses); Jackson, 419 U. S., at 351–352 (electric monopolies); Columbia Broadcasting System, Inc. v. Democratic National Committee, 412 U.S. 94, 120–121 (1973) (broadcast licenses); Moose Lodge No. 107 v. Irvis, 407 U.S. 163, 176–177 (1972) (liquor licenses); cf. Trustees of Dartmouth College v. Woodward, 4 Wheat. 518, 638–639 (1819) (corporate charters). The same principle applies if the government funds or subsidizes a private entity. See Blum, 457 U. S., at 1011; Rendell-Baker, 457 U. S., at 840. Numerous private entities in America obtain government licenses, government contracts, or government-granted monopolies. If those facts sufficed to transform a private entity into a state actor, a large swath of private entities in America would suddenly be turned into state actors and be subject to a variety of constitutional constraints on their activities. As this Court’s many state-action cases amply demonstrate, that is not the law. Here, therefore, the City’s designation of MNN to operate the public access channels on Time Warner’s cable system does not make MNN a state actor. So, too, New York State’s extensive regulation of MNN’s operation of the public access channels does not make MNN a state actor. Under the State’s regulations, air time on the public access channels must be free, and programming must be aired on a first-come, first-served basis. Those regulations restrict MNN’s editorial discretion and in effect require MNN to operate almost like a common carrier. But under this Court’s cases, those restrictions do not render MNN a state actor. In Jackson v. Metropolitan Edison Co., the leading case on point, the Court stated that the “fact that a business is subject to state regulation does not by itself convert its action into that of the State.” 419 U. S., at 350. In that case, the Court held that “a heavily regulated, privately owned utility, enjoying at least a partial monopoly in the providing of electrical service within its territory,” was not a state actor. Id., at 358. The Court explained that the “mere existence” of a “regulatory scheme”—even if “extensive and detailed”—did not render the utility a state actor. Id., at 350, and n. 7. Nor did it matter whether the State had authorized the utility to provide electric service to the community, or whether the utility was the only entity providing electric service to much of that community. This case closely parallels Jackson. Like the electric utility in Jackson, MNN is “a heavily regulated, privately owned” entity. Id., at 358. As in Jackson, the regulations do not transform the regulated private entity into a state actor. Put simply, being regulated by the State does not make one a state actor. See Sullivan, 526 U. S., at 52; Blum, 457 U. S., at 1004; Rendell-Baker, 457 U. S., at 841–842; Jackson, 419 U. S., at 350; Moose Lodge, 407 U. S., at 176–177. As the Court’s cases have explained, the “being heavily regulated makes you a state actor” theory of state action is entirely circular and would significantly endanger individual liberty and private enterprise. The theory would be especially problematic in the speech context, because it could eviscerate certain private entities’ rights to exercise editorial control over speech and speakers on their properties or platforms. Not surprisingly, as Justice Thomas has pointed out, this Court has “never even hinted that regulatory control, and particularly direct regulatory control over a private entity’s First Amendment speech rights,” could justify subjecting the regulated private entity to the constraints of the First Amendment. Denver Area, 518 U. S., at 829 (opinion concurring in judgment in part and dissenting in part). In sum, we conclude that MNN is not subject to First Amendment constraints on how it exercises its editorial discretion with respect to the public access channels. To be sure, MNN is subject to state-law constraints on its editorial discretion (assuming those state laws do not violate a federal statute or the Constitution). If MNN violates those state laws, or violates any applicable contracts, MNN could perhaps face state-law sanctions or liability of some kind. We of course take no position on any potential state-law questions. We simply conclude that MNN, as a private actor, is not subject to First Amendment constraints on how it exercises editorial discretion over the speech and speakers on its public access channels. III Perhaps recognizing the problem with their argument that MNN is a state actor under ordinary state-action principles applicable to private entities and private property, the producers alternatively contend that the public access channels are actually the property of New York City, not the property of Time Warner or MNN. On this theory, the producers say (and the dissent agrees) that MNN is in essence simply managing government property on behalf of New York City. The short answer to that argument is that the public access channels are not the property of New York City. Nothing in the record here suggests that a government (federal, state, or city) owns or leases either the cable system or the public access channels at issue here. Both Time Warner and MNN are private entities. Time Warner is the cable operator, and it owns its cable network, which contains the public access channels. MNN operates those public access channels with its own facilities and equipment. The City does not own or lease the public access channels, and the City does not possess a formal easement or other property interest in those channels. The franchise agreements between the City and Time Warner do not say that the City has any property interest in the public access channels. On the contrary, the franchise agreements expressly place the public access channels “under the jurisdiction” of MNN. App. 22. Moreover, the producers did not allege in their complaint that the City has a property interest in the channels. And the producers have not cited any basis in state law for such a conclusion. Put simply, the City does not have “any formal easement or other property interest in those channels.” Denver Area, 518 U. S., at 828 (opinion of Thomas, J.). It does not matter that a provision in the franchise agreements between the City and Time Warner allowed the City to designate a private entity to operate the public access channels on Time Warner’s cable system. Time Warner still owns the cable system. And MNN still operates the public access channels. To reiterate, nothing in the franchise agreements suggests that the City possesses any property interest in Time Warner’s cable system, or in the public access channels on that system. It is true that the City has allowed the cable operator, Time Warner, to lay cable along public rights-of-way in the City. But Time Warner’s access to public rights-of-way does not alter the state-action analysis. For Time Warner, as for other cable operators, access to public rights-of-way is essential to lay cable and construct a physical cable infrastructure. See Turner Broadcasting System, Inc. v. FCC, 512 U.S. 622, 628 (1994). But the same is true for utility providers, such as the electric utility in Jackson. Put simply, a private entity’s permission from government to use public rights-of-way does not render that private entity a state actor. Having said all that, our point here should not be read too broadly. Under the laws in certain States, including New York, a local government may decide to itself operate the public access channels on a local cable system (as many local governments in New York State and around the country already do), or could take appropriate steps to obtain a property interest in the public access channels. Depending on the circumstances, the First Amendment might then constrain the local government’s operation of the public access channels. We decide only the case before us in light of the record before us. * * * It is sometimes said that the bigger the government, the smaller the individual. Consistent with the text of the Constitution, the state-action doctrine enforces a critical boundary between the government and the individual, and thereby protects a robust sphere of individual liberty. Expanding the state-action doctrine beyond its traditional boundaries would expand governmental control while restricting individual liberty and private enterprise. We decline to do so in this case. MNN is a private entity that operates public access channels on a cable system. Operating public access channels on a cable system is not a traditional, exclusive public function. A private entity such as MNN who opens its property for speech by others is not transformed by that fact alone into a state actor. Under the text of the Constitution and our precedents, MNN is not a state actor subject to the First Amendment. We reverse in relevant part the judgment of the Second Circuit, and we remand the case for further proceedings consistent with this opinion. It is so ordered. Notes 1 Relatedly, this Court has recognized that a private entity may, under certain circumstances, be deemed a state actor when the government has outsourced one of its constitutional obligations to a private entity. In West v. Atkins, for example, the State was constitutionally obligated to provide medical care to prison inmates. 487 U.S. 42, 56 (1988). That scenario is not present here because the government has no such obligation to operate public access channels. 2 A distinct question not raised here is the degree to which the First Amendment protects private entities such as Time Warner or MNN from government legislation or regulation requiring those private entities to open their property for speech by others. Cf. Turner Broadcasting System, Inc. v. FCC, 512 U.S. 622, 636–637 (1994). 3 In Cornelius v. NAACP Legal Defense & Educational Fund, Inc., this Court said in passing dicta that “a speaker must seek access to public property or to private property dedicated to public use to evoke First Amendment concerns.” 473 U.S. 788, 801 (1985). But Cornelius dealt with government-owned property. As Justice Thomas explained in Denver Area Educational Telecommunications Consortium, Inc. v. FCC, the Court’s admittedly imprecise and overbroad phrase in Cornelius is not consistent with this Court’s case law and should not be read to suggest that private property owners or private lessees are subject to First Amendment constraints whenever they dedicate their private property to public use or otherwise open their property for speech. 518 U.S. 727, 827–828 (1996) (opinion concurring in judgment in part and dissenting in part).

588.US.2018_18-485

Petitioner Edward McDonough processed ballots as a commissioner of the county board of elections in a primary election in Troy, New York. Respondent Youel Smith was specially appointed to investigate and to prosecute a case of forged absentee ballots in that election. McDonough became his primary target. McDonough alleges that Smith fabricated evidence against him and used it to secure a grand jury indictment. Smith then brought the case to trial and presented allegedly fabricated testimony. That trial ended in a mistrial. Smith again elicited allegedly fabricated evidence in a second trial, which ended on December 21, 2012, with McDonough’s acquittal on all charges. On December 18, 2015, McDonough sued Smith under 42 U. S. C. §1983, asserting, as relevant here, a claim for fabrication of evidence. The District Court dismissed the claim as untimely, and the Second Circuit affirmed. The court held that the 3-year limitations period began to run “when (1) McDonough learned that the evidence was false and was used against him during the criminal proceedings; and (2) he suffered a loss of liberty as a result of that evidence,” 898 F.3d 259, 265. Thus, the court concluded, McDonough’s claim was untimely, because those events undisputedly had occurred by the time McDonough was arrested and stood trial. Held: The statute of limitations for McDonough’s §1983 fabricated-evidence claim began to run when the criminal proceedings against him terminated in his favor—that is, when he was acquitted at the end of his second trial. Pp. 3–15. (a) The time at which a §1983 claim accrues “is a question of federal law,” “conforming in general to common-law tort principles,” and is presumptively—but not always—“when the plaintiff has ‘a complete and present cause of action.’ ” Wallace v. Kato, 549 U.S. 384, 388. An accrual analysis begins with identifying “the specific constitutional right” alleged to have been infringed. Manuel v. Joliet, 580 U. S. ___, ___. Here, the claimed right is an assumed due process right not to be deprived of liberty as a result of a government official’s fabrication of evidence. Pp. 4–5. (b) Accrual questions are often decided by referring to the common-law principles governing analogous torts. Wallace, 549 U. S., 388. The most analogous common-law tort here is malicious prosecution, which accrues only once the underlying criminal proceedings have resolved in the plaintiff’s favor. Following that analogy where it leads: McDonough could not bring his fabricated-evidence claim under §1983 prior to favorable termination of his prosecution. Malicious prosecution’s favorable-termination requirement is rooted in pragmatic concerns with avoiding parallel criminal and civil litigation over the same subject matter and the related possibility of conflicting civil and criminal judgments, and likewise avoids allowing collateral attacks on criminal judgments through civil litigation. See Heck v. Humphrey, 512 U.S. 477, 484–485. Because a civil claim such as McDonough’s, asserting that fabricated evidence was used to pursue a criminal judgment, implicates the same concerns, it makes sense to adopt the same rule. The principles and reasoning of Heck—which emphasized those concerns with parallel litigation and conflicting judgments—confirm the strength of this analogy. This case differs because the plaintiff in Heck had been convicted and McDonough was acquitted, but McDonough’s claims nevertheless challenge the validity of the criminal proceedings against him in essentially the same manner as the plaintiff in Heck challenged the validity of his conviction. Pp. 5–9. (c) The soundness of this conclusion is reinforced by the consequences that would follow from imposing a ticking limitations clock on criminal defendants as soon as they become aware that fabricated evidence has been used against them. That rule would create practical problems in jurisdictions where prosecutions regularly last nearly as long as—or even longer than—the limitations period. Criminal defendants could face the untenable choice of letting their claims expire or filing a civil suit against the very person who is in the midst of prosecuting them. The parallel civil litigation that would result if plaintiffs chose the second option would run counter to core principles of federalism, comity, consistency, and judicial economy. Smith’s suggested workaround—stays and ad-hoc abstentions—is poorly suited to the type of claim at issue here. Pp. 9–11. (d) Smith’s counterarguments do not sway the result. First, relying on Wallace, Smith argues that Heck is irrelevant to McDonough’s claim. The Court in Wallace rejected the plaintiff’s reliance on Heck, but Wallace involved a false-arrest claim—analogous to common-law false imprisonment—and does not displace the principles in Heck that resolve this case. Second, Smith argues that McDonough theoretically could have been prosecuted without the fabricated evidence, and was not convicted even with it; and thus, because a violation could exist no matter its effect on the outcome, the date of that outcome is irrelevant. Although the argument for adopting a favorable-termination requirement would be weaker in the context of a fabricated-evidence claim that does not allege that the violation’s consequence was a liberty deprivation occasioned by the criminal proceedings themselves, that is not the nature of McDonough’s claim. His claim remains most analogous to a claim of common-law malicious prosecution. Nor does it change the result that McDonough suffered harm prior to his acquittal, because the Court has never suggested that the date on which a constitutional injury first occurs is the only date from which a limitations period may run. Third, Smith argues that the advantages of his rule outweigh its disadvantages as a matter of policy. But his arguments are unconvincing. It is not clear that the Second Circuit’s approach would provide more predictable guidance, and while perverse incentives for prosecutors and risk of foreclosing meritorious claims could be valid considerations in other contexts, they do not overcome other considerations here. Pp. 11–15. 898 F.3d 259, reversed and remanded. Sotomayor, J., delivered the opinion of the Court, in which Roberts, C. J., and Ginsburg, Breyer, Alito, and Kavanaugh, JJ., joined. Thomas, J., filed a dissenting opinion, in which Kagan and Gorsuch, JJ., joined.

Petitioner Edward McDonough alleges that respondent Youel Smith fabricated evidence and used it to pursue criminal charges against him. McDonough was acquitted, then sued Smith under 42 U. S. C. §1983. The courts below, concluding that the limitations period for McDonough’s fabricated-evidence claim began to run when the evidence was used against him, determined that the claim was untimely. We hold that the limitations period did not begin to run until McDonough’s acquittal, and therefore reverse. I This case arises out of an investigation into forged absentee ballots that were submitted in a primary election in Troy, New York, in 2009. McDonough, who processed the ballots in his capacity as a commissioner of the county board of elections, maintains that he was unaware that they had been forged. Smith was specially appointed to investigate and to prosecute the matter. McDonough’s complaint alleges that Smith then set about scapegoating McDonough (against whose family Smith harbored a political grudge), despite evidence that McDonough was innocent. Smith leaked to the press that McDonough was his primary target and pressured him to confess. When McDonough would not, Smith allegedly fabricated evidence in order to inculpate him. Specifically, McDonough alleges that Smith falsified affidavits, coached witnesses to lie, and orchestrated a suspect DNA analysis to link McDonough to relevant ballot envelopes. Relying in part on this allegedly fabricated evidence, Smith secured a grand jury indictment against McDonough. McDonough was arrested, arraigned, and released (with restrictions on his travel) pending trial. Smith brought the case to trial a year later, in January 2012. He again presented the allegedly fabricated testimony during this trial, which lasted more than a month and ended in a mistrial. Smith then reprosecuted McDonough. The second trial also lasted over a month, and again, Smith elicited allegedly fabricated testimony. The second trial ended with McDonough’s acquittal on all charges on December 21, 2012. On December 18, 2015, just under three years after his acquittal, McDonough sued Smith and other defendants under §1983 in the U. S. District Court for the Northern District of New York. Against Smith, McDonough asserted two different constitutional claims: one for fabrication of evidence, and one for malicious prosecution without probable cause. The District Court dismissed the malicious prosecution claim as barred by prosecutorial immunity, though timely. It dismissed the fabricated-evidence claim, however, as untimely. McDonough appealed to the U. S. Court of Appeals for the Second Circuit, which affirmed. 898 F.3d 259 (2018). The Court of Appeals agreed with the District Court’s disposition of the malicious prosecution claim. As for the timeliness of the fabricated-evidence claim, because all agreed that the relevant limitations period is three years, id., at 265, the question was when that limitations period began to run: upon McDonough’s acquittal, or at some point earlier. In essence, given the dates at issue, McDonough’s claim was timely only if the limitations period began running at acquittal. The Court of Appeals held that McDonough’s fabricated-evidence claim accrued, and thus the limitations period began to run, “when (1) McDonough learned that the evidence was false and was used against him during the criminal proceedings; and (2) he suffered a loss of liberty as a result of that evidence.” Ibid. This rule, in the Second Circuit’s view, followed from its conclusion that a plaintiff has a complete fabricated-evidence claim as soon as he can show that the defendant’s knowing use of the fabricated evidence caused him some deprivation of lib- erty. Id., at 266. Those events undisputedly had occurred by the time McDonough was arrested and stood trial. Ibid. As the Second Circuit acknowledged, id., at 267, other Courts of Appeals have held that the statute of limitations for a fabricated-evidence claim does not begin to run until favorable termination of the challenged criminal proceedings.[1] We granted certiorari to resolve the conflict, 586 U. S. ___ (2019), and now reverse. II The statute of limitations for a fabricated-evidence claim like McDonough’s does not begin to run until the criminal proceedings against the defendant (i.e., the §1983 plaintiff) have terminated in his favor. This conclusion follows both from the rule for the most natural common-law analogy (the tort of malicious prosecution) and from the practical considerations that have previously led this Court to defer accrual of claims that would otherwise constitute an untenable collateral attack on a criminal judgment. A The question here is when the statute of limitations began to run. Although courts look to state law for the length of the limitations period, the time at which a §1983 claim accrues “is a question of federal law,” “conforming in general to common-law tort principles.” Wallace v. Kato, 549 U.S. 384, 388 (2007). That time is presumptively “when the plaintiff has ‘a complete and present cause of action,’ ” ibid., though the answer is not always so simple. See, e.g., id., at 388–391, and n. 3; Dodd v. United States, 545 U.S. 353, 360 (2005). Where, for example, a particular claim may not realistically be brought while a violation is ongoing, such a claim may accrue at a later date. See Wallace, 549 U. S., at 389. An accrual analysis begins with identifying “ ‘the specific constitutional right’ ” alleged to have been infringed. Manuel v. Joliet, 580 U. S. ___, ___ (2017) (slip op., at 12) (quoting Albright v. Oliver, 510 U.S. 266, 271 (1994) (plurality opinion)). Though McDonough’s complaint does not ground his fabricated-evidence claim in a particular constitutional provision, the Second Circuit treated his claim as arising under the Due Process Clause. 898 F. 3d, at 266. McDonough’s claim, this theory goes, seeks to vindicate a “ ‘right not to be deprived of liberty as a result of the fabrication of evidence by a government officer.’ ” Ibid. (quoting Zahrey v. Coffey, 221 F.3d 342, 349 (CA2 2000)); see also, e.g., Napue v. Illinois, 360 U.S. 264, 269 (1959). We assume without deciding that the Second Circuit’s articulations of the right at issue and its contours are sound, having not granted certiorari to resolve those separate questions. See Heck v. Humphrey, 512 U.S. 477, 480, n. 2 (1994) (accepting the lower courts’ characterization of the relevant claims).[2] B As noted above, this Court often decides accrual questions by referring to the common-law principles governing analogous torts. See Wallace, 549 U. S., at 388; Heck, 512 U. S., at 483. These “principles are meant to guide rather than to control the definition of §1983 claims,” such that the common law serves “ ‘more as a source of inspired examples than of prefabricated components.’ ” Manuel, 580 U. S., at ___ (slip op., at 12). Relying on our decision in Heck, McDonough analogizes his fabricated-evidence claim to the common-law tort of malicious prosecution, a type of claim that accrues only once the underlying criminal proceedings have resolved in the plaintiff’s favor. 512 U. S., at 484; Prosser & Keeton §119, at 871, 874–875; Restatement (Second) of Torts §§653, 658 (1976); 3 D. Dobbs, P. Hayden, & E. Bublick, Law of Torts §§586, 590, pp. 388–389, 402–404 (2d ed. 2011) (Dobbs). McDonough is correct that malicious prosecution is the most analogous common-law tort here. Common-law malicious prosecution requires showing, in part, that a defendant instigated a criminal proceeding with improper purpose and without probable cause. Restatement (Second) of Torts §653; see also Dobbs §586, at 388–389; Prosser & Keeton §119, at 871.[3] The essentials of McDonough’s claim are similar: His claim requires him to show that the criminal proceedings against him—and consequent deprivations of his liberty[4]—were caused by Smith’s malfeasance in fabricating evidence. At bottom, both claims challenge the integrity of criminal prosecutions undertaken “pursuant to legal process.” See Heck, 512 U. S., at 484.[5] We follow the analogy where it leads: McDonough could not bring his fabricated-evidence claim under §1983 prior to favorable termination of his prosecution. As Heck explains, malicious prosecution’s favorable-termination requirement is rooted in pragmatic concerns with avoiding parallel criminal and civil litigation over the same subject matter and the related possibility of conflicting civil and criminal judgments. See id., at 484–485; see also Prosser & Keeton §119, at 874; Dobbs §589, at 402. The requirement likewise avoids allowing collateral attacks on criminal judgments through civil litigation. Heck, 512 U. S., at 484. These concerns track “similar concerns for finality and consistency” that have motivated this Court to refrain from multiplying avenues for collateral attack on criminal judgments through civil tort vehicles such as §1983. Id., at 485; see also Preiser v. Rodriguez, 411 U.S. 475, 490 (1973) (noting the “strong policy requiring exhaustion of state remedies” in order “to avoid the unnecessary friction between the federal and state court systems”); Younger v. Harris, 401 U.S. 37, 43 (1971) (“Since the beginning of this country’s history Congress has, subject to few exceptions, manifested a desire to permit state courts to try state cases free from interference by federal courts”). Because a civil claim such as McDonough’s, asserting that fabricated evidence was used to pursue a criminal judgment, implicates the same concerns, it makes sense to adopt the same rule.[6] Heck confirms the strength of this analogy. In Heck, a prisoner serving a 15-year sentence for manslaughter sought damages under §1983 against state prosecutors and an investigator for alleged misconduct similar to that alleged here, including knowingly destroying exculpatory evidence and causing an illegal voice identification procedure to be employed at the prisoner’s trial. 512 U. S., at 478–479. The Court took as a given the lower courts’ conclusion that those claims all effectively “challeng[ed] the legality of” the plaintiff’s conviction. Id., at 480, n. 2. Looking first to the common law, the Court observed that malicious prosecution “provide[d] the closest analogy to” such claims because, unlike other potentially analogous common-law claims, malicious prosecution “permits damages for confinement imposed pursuant to legal process.” Id., at 484. Emphasizing the concerns with parallel litigation and conflicting judgments just discussed, see id., at 484–486, the Court in Heck held that “in order to recover damages for allegedly unconstitutional conviction or imprisonment, or for other harm caused by actions whose unlawfulness would render a conviction or sentence invalid,” a plaintiff in a §1983 action first had to prove that his conviction had been invalidated in some way, id., at 486. This favorable-termination requirement, the Court explained, applies whenever “a judgment in favor of the plaintiff would necessarily imply” that his prior conviction or sentence was invalid. Id., at 487. This case differs from Heck because the plaintiff in Heck had been convicted, while McDonough was acquitted. Although some claims do fall outside Heck’s ambit when a conviction is merely “anticipated,” Wallace, 549 U. S., at 393, however, McDonough’s claims are not of that kind, see infra, at 11–12. As articulated by the Court of Appeals, his claims challenge the validity of the criminal proceedings against him in essentially the same manner as the plaintiff in Heck challenged the validity of his conviction. And the pragmatic considerations discussed in Heck apply generally to civil suits within the domain of habeas corpus, not only to those that challenge convictions. See Preiser, 411 U. S., at 490–491. The principles and reasoning of Heck thus point toward a corollary result here: There is not “ ‘a complete and present cause of action,’ ” Wallace, 549 U. S., at 388, to bring a fabricated-evidence challenge to criminal proceedings while those criminal proceedings are ongoing. Only once the criminal proceeding has ended in the defendant’s favor, or a resulting conviction has been invalidated within the meaning of Heck, see 512 U. S., at 486–487, will the statute of limitations begin to run.[7] C The soundness of this conclusion is reinforced by the consequences that would follow from the Second Circuit’s approach, which would impose a ticking limitations clock on criminal defendants as soon as they become aware that fabricated evidence has been used against them. Such a rule would create practical problems in jurisdictions where prosecutions regularly last nearly as long as—or even longer than—the relevant civil limitations period. See Brief for Petitioner 53–55; Brief for Criminal Defense Organizations et al. as Amici Curiae 23–24. A significant number of criminal defendants could face an untenable choice between (1) letting their claims expire and (2) filing a civil suit against the very person who is in the midst of prosecuting them. The first option is obviously undesir- able, but from a criminal defendant’s perspective the latter course, too, is fraught with peril: He risks tipping his hand as to his defense strategy, undermining his privilege against self-incrimination, and taking on discovery obligations not required in the criminal context. See SEC v. Dresser Industries, Inc., 628 F.2d 1368, 1376 (CADC 1980) (en banc). Moreover, as noted above, the parallel civil litigation that would result if plaintiffs chose the second option would run counter to core principles of federalism, comity, consistency, and judicial economy. See supra, at 7–8. Smith suggests that stays and ad hoc abstention are sufficient to avoid the problems of two-track litigation. Such workarounds are indeed available when claims falling outside Heck’s scope nevertheless are initiated while a state criminal proceeding is pending, see Wallace, 549 U. S., at 393–394 (noting the power of district courts to stay civil actions while criminal prosecutions proceed); Heck, 512 U. S., at 487–488, n. 8 (noting possibility of abstention), but Smith’s solution is poorly suited to the type of claim at issue here. When, as here, a plaintiff’s claim “necessarily” questions the validity of a state proceeding, id., at 487, there is no reason to put the onus to safeguard comity on district courts exercising case-by-case discretion—particularly at the foreseeable expense of potentially prejudicing litigants and cluttering dockets with dormant, unripe cases. Cf. Panetti v. Quarterman, 551 U.S. 930, 943 (2007) (noting that a scheme requiring “conscientious defense attorneys” to file unripe suits “would add to the burden imposed on courts, applicants, and the States, with no clear advantage to any”). The accrual rule we adopt today, by contrast, respects the autonomy of state courts and avoids these costs to litigants and federal courts. In deferring rather than inviting such suits, we adhere to familiar principles. The proper approach in our federal system generally is for a criminal defendant who believes that the criminal proceedings against him rest on knowingly fabricated evidence to defend himself at trial and, if necessary, then to attack any resulting conviction through collateral review proceedings. McDonough therefore had a complete and present cause of action for the loss of his liberty only once the criminal proceedings against him terminated in his favor. III Smith’s counterarguments do not sway the result. First, Smith argues that Heck is irrelevant to McDonough’s claim, relying on this Court’s opinion in Wallace. Wallace held that the limitations period begins to run on a §1983 claim alleging an unlawful arrest under the Fourth Amendment as soon as the arrestee “becomes detained pursuant to legal process,” not when he is ultimately released. 549 U. S., at 397. The Court rejected the plaintiff’s reliance on Heck, stating that the Heck rule comes “into play only when there exists ‘a conviction or sentence that has not been . . . invalidated,’ that is to say, an ‘outstanding criminal judgment.’ ” Wallace, 549 U. S., at 393. The Court thus declined to adopt the plaintiff’s theory “that an action which would impugn an anticipated future conviction cannot be brought until that conviction occurs and is set aside,” because doing so in the context of an action for false arrest would require courts and litigants “to speculate about whether a prosecution will be brought, whether it will result in conviction, and whether the pending civil action will impugn that verdict—all this at a time when it can hardly be known what evidence the prosecution has in its possession.” Ibid. (citations omitted).[8] Smith is correct that Heck concerned a plaintiff serving a sentence for a still-valid conviction and that Wallace distinguished Heck on that basis, but Wallace did not displace the principles in Heck that resolve this case. A false-arrest claim, Wallace explained, has a life independent of an ongoing trial or putative future conviction—it attacks the arrest only to the extent it was without legal process, even if legal process later commences. See 549 U. S., at 389–390, 393. That feature made the claim analogous to common-law false imprisonment. Id., at 389. By contrast, a claim like McDonough’s centers on evidence used to secure an indictment and at a criminal trial, so it does not require “speculat[ion] about whether a prosecution will be brought.” Id., at 393. It directly challenges—and thus necessarily threatens to impugn—the prosecution itself. See Heck, 512 U. S., at 486–487. Second, Smith notes (1) that a fabricated-evidence claim in the Second Circuit (unlike a malicious prosecution claim) can exist even if there is probable cause and (2) that McDonough was acquitted. In other words, McDonough theoretically could have been prosecuted without the fabricated evidence, and he was not convicted even with it. Because a violation thus could exist no matter its effect on the outcome, Smith reasons, “the date on which that outcome occurred is irrelevant.” Brief for Respondent 26. Smith is correct in one sense. One could imagine a fabricated-evidence claim that does not allege that the violation’s consequence was a liberty deprivation occasioned by the criminal proceedings themselves. See n. 2, supra. To be sure, the argument for adopting a favorable-termination requirement would be weaker in that context. That is not, however, the nature of McDonough’s claim. As already explained, McDonough’s claim remains most analogous to a claim of common-law malicious prosecution, even if the two are not identical. See supra, at 6–7. Heck explains why favorable termination is both relevant and required for a claim analogous to malicious prosecution that would impugn a conviction, and that rationale extends to an ongoing prosecution as well: The alternative would impermissibly risk parallel litigation and conflicting judgments. See supra, at 7–8. If the date of the favorable termination was relevant in Heck, it is relevant here. It does not change the result, meanwhile, that McDonough suffered harm prior to his acquittal. The Court has never suggested that the date on which a constitutional injury first occurs is the only date from which a limitations period may run. Cf. Wallace, 549 U. S., at 389–391, and n. 3 (explaining that the statute of limitations for false-arrest claims does not begin running when the initial arrest takes place). To the contrary, the injury caused by a classic malicious prosecution likewise first occurs as soon as legal process is brought to bear on a defendant, yet favorable termination remains the accrual date. See Heck, 512 U. S., at 484.[9] Third and finally, Smith argues that the advantages of his rule outweigh its disadvantages as a matter of policy. In his view, the Second Circuit’s approach would provide more predictable guidance, while the favorable-termination approach fosters perverse incentives for prosecutors (who may become reluctant to offer favorable resolutions) and risks foreclosing meritorious claims (for example, where an outcome is not clearly “favorable”). These arguments are unconvincing. We agree that clear accrual rules are valuable but fail to see how assessing when proceedings terminated favorably will be, on balance, more burdensome than assessing when a criminal defendant “learned that the evidence was false and was used against him” and deprived him of liberty as a result. 898 F. 3d, at 265. And while the risk of foreclosing certain claims and the potential incentive effects that Smith identifies could be valid considerations in other contexts,[10] they do not overcome the greater danger that plaintiffs will be deterred under Smith’s theory from suing for redress of egregious misconduct, see supra, at 10—nor do they override the guidance of the common law and precedent. IV The statute of limitations for McDonough’s §1983 claim alleging that he was prosecuted using fabricated evidence began to run when the criminal proceedings against him terminated in his favor—that is, when he was acquitted at the end of his second trial. The judgment of the United States Court of Appeals for the Second Circuit is therefore reversed, and the case is remanded for further proceedings consistent with this opinion. It is so ordered. Notes 1 See Floyd v. Attorney General of Pa., 722 Fed. Appx. 112, 114 (CA3 2018); Mills v. Barnard, 869 F.3d 473, 484 (CA6 2017); Bradford v. Scherschligt, 803 F.3d 382, 388 (CA9 2015); Castellano v. Fragozo, 352 F.3d 939, 959–960 (CA5 2003) (en banc). 2 In accepting the Court of Appeals’ treatment of McDonough’s claim as one sounding in denial of due process, we express no view as to what other constitutional provisions (if any) might provide safeguards against the creation or use of fabricated evidence enforceable through a 42 U. S. C. §1983 action. See Soldal v. Cook County, 506 U.S. 56, 70 (1992) (“Certain wrongs affect more than a single right and, accord-ingly, can implicate more than one of the Constitution’s commands”). Moreover, because the Second Circuit understood McDonough’s due process claim to allege a deprivation of liberty, we have no occasion to consider the proper handling of a fabricated-evidence claim founded on an allegation that the use of fabricated evidence was so egregious as to shock the conscience, see, e.g., County of Sacramento v. Lewis, 523 U.S. 833, 849 (1998), or caused harms exclusively to “interests other than the interest in freedom from physical restraint,” Albright v. Oliver, 510 U.S. 266, 283 (1994) (Kennedy, J., concurring in judgment); see also, e.g., W. Keeton, D. Dobbs, R. Keeton, & D. Owen, Prosser and Keeton on Law of Torts §119, p. 870 (5th ed. 1984) (Prosser & Keeton) (“[O]ne who is wrongfully prosecuted may suffer both in reputation and by confinement”). Accordingly, we do not address what the accrual rule would be for a claim rooted in other types of harm independent of a liberty deprivation, as no such claim is before us. See 898 F.3d 259, 266 (CA2 2018). 3 The Second Circuit borrowed the common-law elements of malicious prosecution to govern McDonough’s distinct constitutional malicious prosecution claim, which is not before us. See 898 F. 3d, at 268, n. 10. This Court has not defined the elements of such a §1983 claim, see Manuel v. Joliet, 580 U. S. ___, ___–___ (2017) (slip op., at 14–15), and this case provides no occasion to opine on what the elements of a constitutional malicious prosecution action under §1983 are or how they may or may not differ from those of a fabricated-evidence claim. Similarly, while noting that only McDonough’s malicious prosecution claim was barred on absolute-immunity grounds below, we make no statement on whether or how the doctrine of absolute immunity would apply to McDonough’s fabricated-evidence claim. Any further consideration of that question is properly addressed by the Second Circuit on remand, subject to ordinary principles of waiver and forfeiture. 4 Though McDonough was not incarcerated pending trial, he was subject to restrictions on his ability to travel and other “ ‘restraints not shared by the public generally,’ ” Justices of Boston Municipal Court v. Lydon, 466 U.S. 294, 301 (1984), and as the case comes to this Court, it is undisputed that McDonough has pleaded a liberty deprivation. See 898 F. 3d, at 266. 5 Smith urges the Court to steer away from the comparison to malicious prosecution, noting that the Second Circuit treats malicious prosecution claims and fabricated-evidence claims as distinct. See id., at 268, and n. 12. But two constitutional claims may differ yet still both resemble malicious prosecution more than any other common-law tort; comparing constitutional and common-law torts is not a one-to-one matching exercise. See, e.g., Heck, 512 U. S., at 479, 484 (analogizing malicious prosecution to several distinct claims). Tellingly, Smith has not suggested an alternative common-law analogy. See Tr. of Oral Arg. 44–46. 6 Such considerations are why Congress has determined that a petition for writ of habeas corpus, not a §1983 action, “is the appropriate remedy for state prisoners attacking the validity of the fact or length of their confinement,” Preiser v. Rodriguez, 411 U.S. 475, 490 (1973), including confinement pending trial before any conviction has occurred, see id., at 491 (citing Braden v. 30th Judicial Circuit Court of Ky., 410 U.S. 484 (1973)). 7 Because McDonough was not free to sue prior to his acquittal, we need not reach his alternative argument that his claim was timely because it alleged a continuing violation. 8 Heck itself suggested that a similar rule might allow at least some Fourth Amendment unlawful-search claims to proceed without a favorable termination. See 512 U. S., at 487, n. 7. 9 As for Smith’s suggestion that the fabricated evidence could not have caused any liberty deprivation where, as here, there could have been probable cause and there was in fact an acquittal, it suffices to reiterate that we assume the contours of the claim as defined by the Second Circuit, see supra, at 5, 6–7, and nn. 2, 4, and thus accept its undisputed conclusion that there was a sufficient liberty deprivation here, see 898 F. 3d, at 266; see also Garnett v. Undercover Officer C0039, 838 F.3d 265, 277 (CA2 2016) (explaining that “a further deprivation of liberty can result from the fabrication of evidence even if the initial arrest is lawful”). 10 Because McDonough’s acquittal was unquestionably a favorable termination, we have no occasion to address the broader range of ways a criminal prosecution (as opposed to a conviction) might end favorably to the accused. Cf. Heck, 512 U. S., at 486–487. To the extent Smith argues that the law in this area should take account of prosecutors’ broad discretion over such matters as the terms on which pleas will be offered or whether charges will be dropped, those arguments more properly bear on the question whether a given resolution should be understood as favorable or not. Such considerations might call for a context-specific and more capacious understanding of what constitutes “favorable” termination for purposes of a §1983 false-evidence claim, but that is not the question before us.

587.US.2018_17-290

Petitioner Merck Sharp & Dohme Corp. manufactures Fosamax, a drug that treats and prevents osteoporosis in postmenopausal women. However, the mechanism through which Fosamax treats and prevents osteoporosis may increase the risk that patients will suffer “atypical femoral fractures,” that is, a rare type of complete, low-energy fracture that affects the thigh bone. When the Food and Drug Administration first approved of the manufacture and sale of Fosamax in 1995, the Fosamax label did not warn of the then-speculative risk of atypical femoral fractures associated with the drug. But stronger evidence connecting Fosamax to atypical femoral fractures developed after 1995. And the FDA ultimately ordered Merck to add a warning about atypical femoral fractures to the Fosamax label in 2011. Respondents are more than 500 individuals who took Fosamax and suffered atypical femoral fractures between 1999 and 2010. Respondents sued Merck seeking tort damages on the ground that state law imposed upon Merck a legal duty to warn respondents and their doctors about the risk of atypical femoral fractures associated with using Fosamax. Merck, in defense, argued that respondents’ state-law failure-to-warn claims should be dismissed as pre-empted by federal law. Merck conceded that the FDA regulations would have permitted Merck to try to change the label to add a warning before 2010, but Merck asserted that the FDA would have rejected that attempt. In particular, Merck claimed that the FDA’s rejection of Merck’s 2008 attempt to warn of a risk of “stress fractures” showed that the FDA would also have rejected any attempt by Merck to warn of the risk of atypical femoral fractures associated with the drug. The District Court agreed with Merck’s pre-emption argument and granted summary judgment to Merck, but the Third Circuit vacated and remanded. The Court of Appeals recognized that its pre-emption analysis was controlled by this Court’s decision in Wyeth v. Levine, 555 U.S. 555, which held that a state-law failure-to-warn claim is pre-empted where there is “clear evidence” that the FDA would not have approved a change to the label. The Court of Appeals, however, suggested that the “clear evidence” standard had led to varying lower court applications and that it would be helpful for this Court to “clarif[y] or buil[d] out the doctrine.” 852 F.3d 268, 284. Held: 1. “Clear evidence” is evidence that shows the court that the drug manufacturer fully informed the FDA of the justifications for the warning required by state law and that the FDA, in turn, informed the drug manufacturer that the FDA would not approve a change to the drug’s label to include that warning. Pp. 9–15. (a) The Wyeth Court undertook a careful review of the history of federal regulation of drugs and drug labeling and found both a reluctance by Congress to displace state laws that would penalize drug manufacturers for failing to warn consumers of the risks associated with their drugs and an insistence by Congress that drug manufacturers bear the responsibility for the content of their drug labels. Accordingly, this Court held in Wyeth that “absent clear evidence that the FDA would not have approved a change” to the label, the Court “will not conclude that it was impossible . . . to comply with both federal and state requirements.” 555 U. S., at 571. Applying that rule to the facts of that case, the Court said that Wyeth’s evidence of pre-emption fell short for two reasons. First, the record did not show that Wyeth “supplied the FDA with an evaluation or analysis concerning the specific dangers” that would have merited the warning. Id., at 572–573. And second, the record did not show that Wyeth “attempted to give the kind of warning required by [state law] but was prohibited from doing so by the FDA.” Ibid., and n. 5. Pp. 10–13. (b) Thus, in a case like Wyeth, showing that federal law prohibited the drug manufacturer from adding a warning that would satisfy state law requires the drug manufacturer to show that it fully informed the FDA of the justifications for the warning required by state law and that the FDA, in turn, informed the drug manufacturer that the FDA would not approve changing the drug’s label to include that warning. These conclusions flow from this Court’s precedents on impossibility pre-emption and the statutory and regulatory scheme that the Court reviewed in Wyeth. See 555 U. S., at 578. In particular, this Court has refused to find clear evidence of impossibility where the laws of one sovereign permit an activity that the laws of the other sovereign restrict or even prohibit. And as explained in Wyeth, FDA regulations permit drug manufacturers to change a label to “reflect newly acquired information” if the changes “add or strengthen a . . . warning” for which there is “evidence of a causal association.” 21 CFR §314.70(c)(6)(iii)(A). Pp. 13–14. (c) The only agency actions that can determine the answer to the pre-emption question are agency actions taken pursuant to the FDA’s congressionally delegated authority. The Supremacy Clause grants “supreme” status only to the “the Laws of the United States.” U. S. Const., Art. VI, cl. 2. And pre-emption takes place “ ‘only when and if [the agency] is acting within the scope of its congressionally delegated authority.’ ” New York v. FERC, 535 U.S. 1, 18 (some alterations omitted). P 15. 2. The question of agency disapproval is primarily one of law for a judge to decide. The question often involves the use of legal skills to determine whether agency disapproval fits facts that are not in dispute. Moreover, judges, rather than lay juries, are better equipped to evaluate the nature and scope of an agency’s determination, and are better suited to understand and to interpret agency decisions in light of the governing statutory and regulatory context. While contested brute facts will sometimes prove relevant to a court’s legal determination about the meaning and effect of an agency decision, such factual questions are subsumed within an already tightly circumscribed legal analysis and do not warrant submission alone or together with the larger pre-emption question to a jury. Pp. 15–17. 852 F.3d 268, vacated and remanded. Breyer, J., delivered the opinion of the Court, in which Thomas, Ginsburg, Sotomayor, Kagan, and Gorsuch, JJ., joined. Thomas, J., filed a concurring opinion. Alito, J., filed an opinion concurring in the judgment, in which Roberts, C. J., and Kavanaugh, J., joined.

When Congress enacted the Federal Food, Drug, and Cosmetic Act, ch. 675, 52Stat. 1040, as amended, 21 U. S. C. §301 et seq., it charged the Food and Drug Administration with ensuring that prescription drugs are “safe for use under the conditions prescribed, recommended, or suggested” in the drug’s “labeling.” §355(d). When the FDA exercises this authority, it makes careful judgments about what warnings should appear on a drug’s label for the safety of consumers. For that reason, we have previously held that “clear evidence” that the FDA would not have approved a change to the drug’s label pre-empts a claim, grounded in state law, that a drug manufacturer failed to warn consumers of the change-related risks associated with using the drug. See Wyeth v. Levine, 555 U.S. 555, 571 (2009). We here determine that this question of pre-emption is one for a judge to decide, not a jury. We also hold that “clear evidence” is evidence that shows the court that the drug manufacturer fully informed the FDA of the justifications for the warning required by state law and that the FDA, in turn, informed the drug manufacturer that the FDA would not approve a change to the drug’s label to include that warning. I The central issue in this case concerns federal pre-emption, which as relevant here, takes place when it is “ ‘impossible for a private party to comply with both state and federal requirements.’ ” Mutual Pharmaceutical Co. v. Bartlett, 570 U.S. 472, 480 (2013). See also U. S. Const., Art. VI, cl. 2. The state law that we consider is state common law or state statutes that require drug manufacturers to warn drug consumers of the risks associated with drugs. The federal law that we consider is the statutory and regulatory scheme through which the FDA regulates the information that appears on brand-name prescription drug labels. The alleged conflict between state and federal law in this case has to do with a drug that was manufactured by petitioner Merck Sharp & Dohme and was administered to respondents without a warning of certain associated risks. A The FDA regulates the safety information that appears on the labels of prescription drugs that are marketed in the United States. 21 U. S. C. §355(b)(1)(F); 21 CFR §201.57(a) (2018). Although we commonly understand a drug’s “label” to refer to the sticker affixed to a prescription bottle, in this context the term refers more broadly to the written material that is sent to the physician who prescribes the drug and the written material that comes with the prescription bottle when the drug is handed to the patient at the pharmacy. 21 U. S. C. §321(m). These (often lengthy) package inserts contain detailed information about the drug’s medical uses and health risks. §355(b)(1)(F); 21 CFR §201.57(a). FDA regulations set out requirements for the content, the format, and the order of the safety information on the drug label. §201.57(c). Those regulations require drug labels to include, among other things: (1) prominent “boxed” warnings about risks that may lead to death or serious injury; (2) contraindications describing any situation in which the drug should not be used because the risk of use outweighs any therapeutic benefit; (3) warnings and precautions about other potential safety hazards; and (4) any adverse reactions for which there is some basis to believe a causal relationship exists between the drug and the occurrence of the adverse event. Ibid. As those requirements make clear, the category in which a particular risk appears on a drug label is an indicator of the likelihood and severity of the risk. The hierarchy of label information is designed to “prevent overwarning” so that less important information does not “overshadow” more important information. 73 Fed. Reg. 49605–49606 (2008). It is also designed to exclude “[e]xaggeration of risk, or inclusion of speculative or hypothetical risks,” that “could discourage appropriate use of a beneficial drug.” Id., at 2851. Prospective drug manufacturers work with the FDA to develop an appropriate label when they apply for FDA approval of a new drug. 21 U. S. C. §§355(a), 355(b), 355(d)(7); 21 CFR §314.125(b)(6). But FDA regulations also acknowledge that information about drug safety may change over time, and that new information may require changes to the drug label. §§314.80(c), 314.81(b)(2)(i). Drug manufacturers generally seek advance permission from the FDA to make substantive changes to their drug labels. However, an FDA regulation called the “changes being effected” or “CBE” regulation permits drug manufacturers to change a label without prior FDA approval if the change is designed to “add or strengthen a . . . warning” where there is “newly acquired information” about the “evidence of a causal association” between the drug and a risk of harm. 21 CFR §314.70(c)(6)(iii)(A). B Petitioner Merck Sharp & Dohme manufactures Fosamax, a drug that treats and prevents osteoporosis in postmenopausal women. App. 192; In re Fosamax (Alendronate Sodium) Products Liability Litigation, 852 F.3d 268, 271, 274–275 (CA3 2017). Fosamax belongs to a class of drugs called “bisphosphonates.” Fosamax and other bisphosphonates work by affecting the “bone remodeling process,” that is, the process through which bones are continuously broken down and built back up again. App. 102, 111. For some postmenopausal women, the two parts of the bone remodeling process fall out of sync; the body removes old bone cells faster than it can replace them. That imbalance can lead to osteoporosis, a disease that is characterized by low bone mass and an increased risk of bone fractures. Fosamax (like other bisphosphonates) slows the breakdown of old bone cells and thereby helps postmenopausal women avoid osteoporotic fractures. Id., at 102. However, the mechanism through which Fosamax decreases the risk of osteoporotic fractures may increase the risk of a different type of fracture. Id., at 400–444, 661–663. That is because all bones—healthy and osteoporotic alike—sometimes develop microscopic cracks that are not due to any trauma, but are instead caused by the mechanical stress of everyday activity. Id., at 102. Those so-called “stress fractures” ordinarily heal on their own through the bone remodeling process. But, by slowing the breakdown of old bone cells, Fosamax and other bisphosphonates may cause stress fractures to progress to complete breaks that cause great pain and require surgical intervention to repair. Id., at 106–109, 139, 144–145. When that rare type of complete, low-energy fracture affects the thigh bone, it is called an “atypical femoral fracture.” Id., at 101. The Fosamax label that the FDA approved in 1995 did not warn of the risk of atypical femoral fractures. 852 F. 3d, at 274–275. At that time, Merck’s scientists were aware of at least a theoretical risk of those fractures. Indeed, as far back as 1990 and 1991, when Fosamax was undergoing preapproval clinical trials, Merck scientists expressed concern in internal discussions that Fosamax could inhibit bone remodeling to such a “ ‘profound’ ” degree that “inadequate repair may take place” and “ ‘micro-fractures would not heal.’ ” App. 111–113. When Merck applied to the FDA for approval of Fosamax, Merck brought those theoretical considerations to the FDA’s attention. 852 F. 3d, at 274–275. But, perhaps because the concerns were only theoretical, the FDA approved Fosamax’s label without requiring any mention of this risk. Ibid. Evidence connecting Fosamax to atypical femoral fractures developed after 1995. Merck began receiving adverse event reports from the medical community indicating that long-term Fosamax users were suffering atypical femoral fractures. App. 122–125. For example, Merck received a report from a doctor who said that hospital staff had begun calling atypical femoral fractures the “ ‘Fosamax Fracture’ ” because “ ‘100% of patients in his practice who have experienced femoral fractures (without being hit by a taxicab), were taking Fosamax . . . for over 5 years. ’ ” Id., at 126. Merck performed a statistical analysis of Fosamax adverse event reports, concluding that these reports revealed a statistically significant incidence of femur fractures. 3 App. in No. 14–1900 (CA3), pp. A1272–A1273, A1443. And about the same time, Merck began to see numerous scholarly articles and case studies documenting possible connections between long-term Fosamax use and atypical femoral fractures. App. 106–110, 116–122. In 2008, Merck applied to the FDA for preapproval to change Fosamax’s label to add language to both the “Adverse Reaction[s]” and the “Precaution[s]” sections of the label. Id., at 670. In particular, Merck proposed adding a reference to “ ‘low-energy femoral shaft fracture’ ” in the Adverse Reactions section, and cross-referencing a longer discussion in the Precautions section that focused on the risk of stress fractures associated with Fosamax. Id., at 728. The FDA approved the addition to the Adverse Reactions section, but rejected Merck’s proposal to warn of a risk of “stress fractures.” Id., at 511–512. The FDA explained that Merck’s “justification” for the proposed change to the Precautions section was “inadequate,” because “[i]dentification of ‘stress fractures’ may not be clearly related to the atypical subtrochanteric fractures that have been reported in the literature.” Id., at 511. The FDA invited Merck to “resubmit” its application and to “fully address all the deficiencies listed.” Id., at 512; see 21 CFR §314.110(b). But Merck instead withdrew its application and decided to make the changes to the Adverse Reactions section through the CBE process. App. 654–660. Merck made no changes to the Precautions section at issue here. Id., at 274. A warning about “atypical femoral fractures” did not appear on the Fosamax label until 2011, when the FDA ordered that change based on its own analyses. Id., at 246–252, 526–534. Merck was initially resistant to the change, proposing revised language that, once again, referred to the risk of “stress fractures.” Id., at 629–634. But the FDA, once again, rejected that language. And this time, the FDA explained that “the term ‘stress fracture’ was considered and was not accepted” because, “for most practitioners, the term ‘stress fracture’ represents a minor fracture and this would contradict the seriousness of the atypical femoral fractures associated with bisphosphonate use.” Id., at 566. In January 2011, Merck and the FDA ultimately agreed upon adding a three-paragraph discussion of atypical femoral fractures to the Warnings and Precautions section of the Fosamax label. Id., at 223–224. The label now refers to the fractures five times as “atypical” without using the term “stress fracture.” Ibid. C The respondents here are more than 500 individuals who took Fosamax and who suffered atypical femoral fractures between 1999 and 2010. Brief for Respondents 7. Respondents, invoking federal diversity jurisdiction, filed separate actions seeking tort damages on the ground that, during the relevant period, state law imposed upon Merck a legal duty to warn them and their doctors about the risk of atypical femoral fractures associated with using Fosamax. Id., at 1. One respondent, for example, filed a complaint alleging that she took Fosamax for roughly 10 years and suffered an atypical femoral fracture. One day in 2009, when the respondent was 70 years old, she turned to unlock the front door of her house, heard a popping sound, and suddenly felt her left leg give out beneath her. She needed surgery, in which doctors repaired her leg with a rod and screws. She explained she would not have used Fosamax for so many years if she had known that she might suffer an atypical femoral fracture as a result. See id., at 18–19. Merck, in defense, argued that respondents’ state-law failure-to-warn claims should be dismissed as pre-empted by federal law. Both Merck and the FDA have long been aware that Fosamax could theoretically increase the risk of atypical femoral fractures. But for some period of time between 1995 (when the FDA first approved a drug label for Fosamax) and 2010 (when the FDA decided to require Merck to add a warning about atypical femoral fractures to Fosamax’s label), both Merck and the FDA were unsure whether the developing evidence of a causal link between Fosamax and atypical femoral fractures was strong enough to require adding a warning to the Fosamax drug label. Merck conceded that the FDA’s CBE regulation would have permitted Merck to try to change the label to add a warning before 2010, but Merck asserted that the FDA would have rejected that attempt. In particular, Merck pointed to the FDA’s rejection of Merck’s 2008 attempt to amend the Fosamax label to warn of the risk of “stress fractures” associated with Fosamax. On that basis, Merck claimed that federal law prevented Merck from complying with any state-law duty to warn the respondents of the risk of atypical femoral fractures associated with Fosamax. The District Court agreed with Merck’s pre-emption argument and granted summary judgment to Merck, In re Fosamax (Alendronate Sodium): Products Liability Litigation, 2014 WL 1266994, *17 (D NJ, Mar. 22, 2017), but the Court of Appeals vacated and remanded, 852 F. 3d, at 302. The Court of Appeals concluded that its pre-emption analysis was controlled by this Court’s decision in Wyeth. Ibid. The Court of Appeals understood that case as making clear that a failure-to-warn claim grounded in state law is pre-empted where there is “ ‘clear evidence that the FDA would not have approved a change to the . . . label.’ ” Id., at 280 (quoting Wyeth, 555 U. S., at 571). The Court of Appeals, however, suggested that this statement had led to varying lower court applications and that it would be helpful for this Court to “clarif[y] or buil[d] out the doctrine.” 852 F. 3d, at 284. In attempting to do so itself, the Court of Appeals held that “the Supreme Court intended to announce a standard of proof when it used the term ‘clear evidence’ in Wyeth.” Ibid. That is, the Court of Appeals believed that “[t]he term ‘clear evidence’ . . . does not refer directly to the type of facts that a manufacturer must show, or to the circumstances in which preemption will be appropriate.” Id., at 285. “Rather, it specifies how difficult it will be for the manufacturer to convince the factfinder that the FDA would have rejected a proposed label change.” Ibid. And in the Court of Appeals’ view, “for a defendant to establish a preemption defense under Wyeth, the factfinder must conclude that it is highly probable that the FDA would not have approved a change to the drug’s label.” Id., at 286. Moreover and importantly, the Court of Appeals also held that “whether the FDA would have rejected a proposed label change is a question of fact that must be answered by a jury.” Ibid. Merck filed a petition for a writ of certiorari. Merck’s petition asked the Court to decide whether Merck’s case and others like it “must . . . go to a jury” to determine whether the FDA, in effect, has disapproved a state-law-required labeling change. In light of differences and uncertainties among the courts of appeals and state supreme courts in respect to the application of Wyeth, we granted certiorari. See, e.g., Mason v. SmithKline Beecham Corp., 596 F.3d 387, 391 (CA7 2010) (“The Supreme Court . . . did not clarify what constitutes ‘clear evidence’ ”); Reckis v. Johnson & Johnson, 471 Mass. 272, 286, 28 N. E. 3d 445, 457 (2015) (“Wyeth did not define ‘clear evidence’ . . . ” (some internal quotation marks omitted)). II We stated in Wyeth v. Levine that state law failure-to-warn claims are pre-empted by the Federal Food, Drug, and Cosmetic Act and related labeling regulations when there is “clear evidence” that the FDA would not have approved the warning that state law requires. 555 U. S., at 571. We here decide that a judge, not the jury, must decide the pre-emption question. And we elaborate Wyeth’s requirements along the way. A We begin by describing Wyeth. In that case, the plaintiff developed gangrene after a physician’s assistant injected her with Phenergan, an antinausea drug. The plaintiff brought a state-law failure-to-warn claim against Wyeth, the drug’s manufacturer, for failing to provide an adequate warning about the risks that accompany various methods of administering the drug. In particular, the plaintiff claimed that directly injecting Phenergan into a patient’s vein (the “IV-push” method of administration) creates a significant risk of catastrophic consequences. And those consequences could be avoided by introducing the drug into a saline solution that slowly descends into a patient’s vein (the “IV-drip” method of administration). A jury concluded that Wyeth’s warning label was inadequate, and that the label’s inadequacy caused the plaintiff’s injury. On appeal, Wyeth argued that the plaintiff’s state-law failure-to-warn claims were pre-empted because it was impossible for Wyeth to comply with both state law duties and federal labeling obligations. The Vermont Supreme Court rejected Wyeth’s pre-emption claim. Id., at 563. We too considered Wyeth’s pre-emption argument, and we too rejected it. In rejecting Wyeth’s argument, we undertook a careful review of the history of federal regulation of drugs and drug labeling. Id., at 566–568. In doing so, we “assum[ed] that the historic police powers of the States were not to be superseded by the Federal Act unless that was the clear and manifest purpose of Congress.” Id., at 565 (internal quotation marks omitted). And we found nothing within that history to indicate that the FDA’s power to approve or to disapprove labeling changes, by itself, pre-empts state law. Rather, we concluded that Congress enacted the FDCA “to bolster consumer protection against harmful products;” that Congress provided no “federal remedy for consumers harmed by unsafe or ineffective drugs”; that Congress was “awar[e] of the prevalence of state tort litigation;” and that, whether Congress’ general purpose was to protect consumers, to provide safety-related incentives to manufacturers, or both, language, history, and purpose all indicate that “Congress did not intend FDA oversight to be the exclusive means of ensuring drug safety and effectiveness.” Id., at 574–575 (emphasis added). See also id., at 574 (“If Congress thought state-law suits posed an obstacle to its objectives, it surely would have enacted an express pre-emption provision at some point during the FDCA’s 70-year history”). We also observed that “through many amendments to the FDCA and to FDA regulations, it has remained a central premise of federal drug regulation that the manufacturer bears responsibility for the content of its label at all times.” Id., at 570–571. A drug manufacturer “is charged both with crafting an adequate label and with ensuring that its warnings remain adequate as long as the drug is on the market.” Id., at 571. Thus, when the risks of a particular drug become apparent, the manufacturer has “a duty to provide a warning that adequately describe[s] that risk.” Ibid. “Indeed,” we noted, “prior to 2007, the FDA lacked the authority to order manufacturers to revise their labels.” Ibid. And even when “Congress granted the FDA this authority,” in the 2007 Amendments to the FDCA, Congress simultaneously “reaffirmed the manufacturer’s obligations and referred specifically to the CBE regulation, which both reflects the manufacturer’s ultimate responsibility for its label and provides a mechanism for adding safety information to the label prior to FDA approval.” Ibid. In light of Congress’ reluctance to displace state laws that would penalize drug manufacturers for failing to warn consumers of the risks associated with their drugs, and Congress’ insistence on requiring drug manufacturers to bear the responsibility for the content of their drug labels, we were unpersuaded by Wyeth’s pre-emption argument. In Wyeth’s case, we concluded, “when the risk of gangrene from IV-push injection of Phenergan became apparent, Wyeth had a duty” under state law “to provide a warning that adequately described that risk, and the CBE regulation permitted it to provide such a warning before receiving the FDA’s approval.” Ibid. At the same time, and more directly relevant here, we pointed out that “the FDA retains authority to reject labeling changes made pursuant to the CBE regulation in its review of the manufacturer’s supplemental application, just as it retains such authority in reviewing all supplemental applications.” Ibid. We then said that, nonetheless, “absent clear evidence that the FDA would not have approved a change to Phenergan’s label, we will not conclude that it was impossible for Wyeth to comply with both federal and state requirements.” Ibid. (emphasis added). We reviewed the record and concluded that “Wyeth has offered no such evidence.” Id., at 572. We said that Wyeth’s evidence of pre-emption fell short for two reasons. First, the record did not show that Wyeth “supplied the FDA with an evaluation or analysis concerning the specific dangers” that would have merited the warning. Id., at 572–573. We could find “no evidence in this record that either the FDA or the manufacturer gave more than passing attention to the issue of IV-push versus IV-drip administration”—the matter at issue in the case. Id., at 572 (internal quotation marks omitted). Second, the record did not show that Wyeth “attempted to give the kind of warning required by [state law] but was prohibited from doing so by the FDA.” Ibid., and n. 5. The “FDA had not made an affirmative decision to preserve” the warning as it was or “to prohibit Wyeth from strengthening its warning.” Id., at 572. For those reasons, we could not “credit Wyeth’s contention that the FDA would have prevented it from adding a stronger warning about the IV-push method of intravenous administration.” And we could not conclude that “it was impossible for Wyeth to comply with both federal and state requirements.” Id., at 573. We acknowledged that meeting the standard we set forth would be difficult, but, we said, “[i]mpossibility pre-emption is a demanding defense.” Ibid. B The underlying question for this type of impossibility pre-emption defense is whether federal law (including appropriate FDA actions) prohibited the drug manufacturer from adding any and all warnings to the drug label that would satisfy state law. And, of course, in order to succeed with that defense the manufacturer must show that the answer to this question is yes. But in Wyeth, we confronted that question in the context of a particular set of circumstances. Accordingly, for purposes of this case, we assume—but do not decide—that, as was true of the warning at issue in Wyeth, there is sufficient evidence to find that Merck violated state law by failing to add a warning about atypical femoral fractures to the Fosamax label. In a case like Wyeth, showing that federal law prohibited the drug manufacturer from adding a warning that would satisfy state law requires the drug manufacturer to show that it fully informed the FDA of the justifications for the warning required by state law and that the FDA, in turn, informed the drug manufacturer that the FDA would not approve changing the drug’s label to include that warning. These conclusions flow from our precedents on impossibility pre-emption and the statutory and regulatory scheme that we reviewed in Wyeth. See 555 U. S., at 578. In particular, “it has long been settled that state laws that conflict with federal law are without effect.” Mutual Pharmaceutical Co., 570 U. S., at 480. But as we have cautioned many times before, the “possibility of impossibility [is] not enough.” PLIVA, Inc. v. Mensing, 564 U.S. 604, 625, n. 8 (2011) (internal quotation marks omitted). Consequently, we have refused to find clear evidence of such impossibility where the laws of one sovereign permit an activity that the laws of the other sovereign restrict or even prohibit. See, e.g., Barnett Bank of Marion Cty., N. A. v. Nelson, 517 U.S. 25, 31 (1996); Michigan Canners & Freezers Assn., Inc. v. Agricultural Marketing and Bargaining Bd., 467 U.S. 461, 478, and n. 21 (1984). And, as we explained in Wyeth, 555 U. S., at 571–573, federal law—the FDA’s CBE regulation—permits drug manufacturers to change a label to “reflect newly acquired information” if the changes “add or strengthen a . . . warning” for which there is “evidence of a causal association,” without prior approval from the FDA. 21 CFR §314.70(c)(6)(iii)(A). Of course, the FDA reviews CBE submissions and can reject label changes even after the manufacturer has made them. See §§314.70(c)(6), (7). And manufacturers cannot propose a change that is not based on reasonable evidence. §314.70(c)(6)(iii)(A). But in the interim, the CBE regulation permits changes, so a drug manufacturer will not ordinarily be able to show that there is an actual conflict between state and federal law such that it was impossible to comply with both. We do not further define Wyeth’s use of the words “clear evidence” in terms of evidentiary standards, such as “preponderance of the evidence” or “clear and convincing evidence” and so forth, because, as we shall discuss, infra, at 15–17, courts should treat the critical question not as a matter of fact for a jury but as a matter of law for the judge to decide. And where that is so, the judge must simply ask himself or herself whether the relevant federal and state laws “irreconcilably conflic[t].” Rice v. Norman Williams Co., 458 U.S. 654, 659 (1982); see ibid. (“The existence of a hypothetical or potential conflict is insufficient to warrant the pre-emption of the state statute”). We do note, however, that the only agency actions that can determine the answer to the pre-emption question, of course, are agency actions taken pursuant to the FDA’s congressionally delegated authority. The Supremacy Clause grants “supreme” status only to the “the Laws of the United States.” U. S. Const., Art. VI, cl. 2. And pre-emption takes place “ ‘only when and if [the agency] is acting within the scope of its congressionally delegated authority, . . . for an agency literally has no power to act, let alone pre-empt the validly enacted legislation of a sovereign State, unless and until Congress confers power upon it.’ ” New York v. FERC, 535 U.S. 1, 18 (2002) (some alterations omitted). Federal law permits the FDA to communicate its disapproval of a warning by means of notice-and-comment rulemaking setting forth labeling standards, see, e.g., 21 U. S. C. §355(d); 21 CFR §§201.57, 314.105; by formally rejecting a warning label that would have been adequate under state law, see, e.g., 21 CFR §§314.110(a), 314.125(b)(6); or with other agency action carrying the force of law, cf., e.g., 21 U. S. C. §355(o)(4)(A). The question of disapproval “method” is not now before us. And we make only the obvious point that, whatever the means the FDA uses to exercise its authority, those means must lie within the scope of the authority Congress has lawfully delegated. III We turn now to what is the determinative question before us: Is the question of agency disapproval primarily one of fact, normally for juries to decide, or is it a question of law, normally for a judge to decide without a jury? The complexity of the preceding discussion of the law helps to illustrate why we answer this question by concluding that the question is a legal one for the judge, not a jury. The question often involves the use of legal skills to determine whether agency disapproval fits facts that are not in dispute. Moreover, judges, rather than lay juries, are better equipped to evaluate the nature and scope of an agency’s determination. Judges are experienced in “[t]he construction of written instruments,” such as those normally produced by a federal agency to memorialize its considered judgments. Markman v. Westview Instruments, Inc., 517 U.S. 370, 388 (1996). And judges are better suited than are juries to understand and to interpret agency decisions in light of the governing statutory and regulatory context. Cf. 5 U. S. C. §706 (specifying that a “reviewing court,” not a jury, “shall . . . determine the meaning or applicability of the terms of an agency action”); see also H. R. Rep. No. 1980, 79th Cong., 2d Sess., 44 (1946) (noting longstanding view that “questions respecting the . . . terms of any agency action” and its “application” are “questions of law”). To understand the question as a legal question for judges makes sense given the fact that judges are normally familiar with principles of administrative law. Doing so should produce greater uniformity among courts; and greater uniformity is normally a virtue when a question requires a determination concerning the scope and effect of federal agency action. Cf. Markman, 517 U. S., at 390–391. We understand that sometimes contested brute facts will prove relevant to a court’s legal determination about the meaning and effect of an agency decision. For example, if the FDA rejected a drug manufacturer’s supplemental application to change a drug label on the ground that the information supporting the application was insufficient to warrant a labeling change, the meaning and scope of that decision might depend on what information the FDA had before it. Yet in litigation between a drug consumer and a drug manufacturer (which will ordinarily lack an official administrative record for an FDA decision), the litigants may dispute whether the drug manufacturer submitted all material information to the FDA. But we consider these factual questions to be subsumed within an already tightly circumscribed legal analysis. And we do not believe that they warrant submission alone or together with the larger pre-emption question to a jury. Rather, in those contexts where we have determined that the question is “for the judge and not the jury,” we have also held that “courts may have to resolve subsidiary factual disputes” that are part and parcel of the broader legal question. Teva Pharmaceuticals USA, Inc. v. Sandoz, Inc., 574 U. S. ___, ___–___ (2015) (slip op., at 6–7). And, as in contexts as diverse as the proper construction of patent claims and the voluntariness of criminal confessions, they create a question that “ ‘falls somewhere between a pristine legal standard and a simple historical fact.’ ” Markman, 517 U. S., at 388 (quoting Miller v. Fenton, 474 U.S. 104, 114 (1985)). In those circum- stances, “ ‘the fact/law distinction at times has turned on a determination that, as a matter of the sound administration of justice, one judicial actor is better positioned than another to decide the issue in question.’ ” Markman, 517 U. S., at 388 (quoting Miller, 474 U. S., at 114). In this context, that “better positioned” decisionmaker is the judge. IV Because the Court of Appeals treated the pre-emption question as one of fact, not law, and because it did not have an opportunity to consider fully the standards we have described in Part II of our opinion, we vacate its judgment and remand the case to that court for further proceedings consistent with this opinion. It is so ordered.

587.US.2018_17-1657

Petitioner Mission Product Holdings, Inc., entered into a contract with Respondent Tempnology, LLC, which gave Mission a license to use Tempnology’s trademarks in connection with the distribution of certain clothing and accessories. Tempnology filed for Chapter 11 bankruptcy and sought to reject its agreement with Mission. Section 365 of the Bankruptcy Code enables a debtor to “reject any executory contract”—meaning a contract that neither party has finished performing. 11 U. S. C. §365(a). It further provides that rejection “constitutes a breach of such contract.” §365(g). The Bankruptcy Court approved Tempnology’s rejection and further held that the rejection terminated Mission’s rights to use Tempnology’s trademarks. The Bankruptcy Appellate Panel reversed, relying on Section 365(g)’s statement that rejection “constitutes a breach” to hold that rejection does not terminate rights that would survive a breach of contract outside bankruptcy. The First Circuit rejected the Panel’s judgment and reinstated the Bankruptcy Court’s decision. Held: 1. This case is not moot. Mission presents a plausible claim for money damages arising from its inability to use Tempnology’s trademarks, which is sufficient to preserve a live controversy. See Chafin v. Chafin, 568 U.S. 165, 172. Tempnology’s various arguments that Mission is not entitled to damages do not so clearly preclude recovery as to render this case moot. Pp. 6–7. 2. A debtor’s rejection of an executory contract under Section 365 of the Bankruptcy Code has the same effect as a breach of that contract outside bankruptcy. Such an act cannot rescind rights that the contract previously granted. Pp. 7–16. (a) Section 365(g) provides that rejection “constitutes a breach.” And “breach” is neither a defined nor a specialized bankruptcy term—it means in the Code what it means in contract law outside bankruptcy. See Field v. Mans, 516 U.S. 59, 69. Outside bankruptcy, a licensor’s breach cannot revoke continuing rights given to a counterparty under a contract (assuming no special contract term or state law). And because rejection “constitutes a breach,” the same result must follow from rejection in bankruptcy. In preserving a counterparty’s rights, Section 365 reflects the general bankruptcy rule that the estate cannot possess anything more than the debtor did outside bankruptcy. See Board of Trade of Chicago v. Johnson, 264 U.S. 1, 15. And conversely, allowing rejection to rescind a counterparty’s rights would circumvent the Code’s stringent limits on “avoidance” actions—the exceptional cases in which debtors may unwind pre-bankruptcy transfers that undermine the bankruptcy process. See, e.g., §548(a). Pp. 8–12. (b) Tempnology’s principal counterargument rests on a negative inference drawn from provisions of Section 365 identifying categories of contracts under which a counterparty may retain specified rights after rejection. See §§365(h), (i), (n). Tempnology argues that these provisions indicate that the ordinary consequence of rejection must be something different—i.e., the termination of contractual rights previously granted. But that argument offers no account of how to read Section 365(g) (rejection “constitutes a breach”) to say essentially its opposite. And the provisions Tempnology treats as a reticulated scheme of exceptions each emerged at a different time and responded to a discrete problem—as often as not, correcting a judicial ruling of just the kind Tempnology urges. Tempnology’s remaining argument turns on how the special features of trademark law may affect the fulfillment of the Code’s goals. Unless rejection terminates a licensee’s right to use a trademark, Tempnology argues, a debtor must choose between monitoring the goods sold under a license or risking the loss of its trademark, either of which would impede a debtor’s ability to reorganize. But the distinctive features of trademarks do not persuade this Court to adopt a construction of Section 365 that will govern much more than trademark licenses. And Tempnology’s plea to facilitate reorganizations cannot overcome what Section 365(a) and (g) direct. In delineating the burdens a debtor may and may not escape, Section 365’s edict that rejection is breach expresses a more complex set of aims than Tempnology acknowledges. Pp. 12–16. 879 F.3d 389, reversed and remanded. Kagan, J., delivered the opinion of the Court, in which Roberts, C. J., and Thomas, Ginsburg, Breyer, Alito, Sotomayor, and Kavanaugh, JJ., joined. Sotomayor, J., filed a concurring opinion. Gorsuch, J., filed a dissenting opinion.

Section 365 of the Bankruptcy Code enables a debtor to “reject any executory contract”—meaning a contract that neither party has finished performing. 11 U. S. C. §365(a). The section further provides that a debtor’s rejection of a contract under that authority “constitutes a breach of such contract.” §365(g). Today we consider the meaning of those provisions in the context of a trademark licensing agreement. The question is whether the debtor-licensor’s rejection of that contract deprives the licensee of its rights to use the trademark. We hold it does not. A rejection breaches a contract but does not rescind it. And that means all the rights that would ordinarily survive a contract breach, including those conveyed here, remain in place. I This case arises from a licensing agreement gone wrong. Respondent Tempnology, LLC, manufactured clothing and accessories designed to stay cool when used in exercise. It marketed those products under the brand name “Coolcore,” using trademarks (e.g., logos and labels) to distinguish the gear from other athletic apparel. In 2012, Tempnology entered into a contract with petitioner Mission Product Holdings, Inc. See App. 203–255. The agreement gave Mission an exclusive license to distribute certain Coolcore products in the United States. And more important here, it granted Mission a non-exclusive license to use the Coolcore trademarks, both in the United States and around the world. The agreement was set to expire in July 2016. But in September 2015, Tempnology filed a petition for Chapter 11 bankruptcy. And it soon afterward asked the Bankruptcy Court to allow it to “reject” the licensing agreement. §365(a). Chapter 11 of the Bankruptcy Code sets out a framework for reorganizing a bankrupt business. See §§1101–1174. The filing of a petition creates a bankruptcy estate consisting of all the debtor’s assets and rights. See §541. The estate is the pot out of which creditors’ claims are paid. It is administered by either a trustee or, as in this case, the debtor itself. See §§1101, 1107. Section 365(a) of the Code provides that a “trustee [or debtor], subject to the court’s approval, may assume or reject any executory contract.” §365(a). A contract is executory if “performance remains due to some extent on both sides.” NLRB v. Bildisco & Bildisco, 465 U.S. 513, 522, n. 6 (1984) (internal quotation marks omitted). Such an agreement represents both an asset (the debtor’s right to the counterparty’s future performance) and a liability (the debtor’s own obligations to perform). Section 365(a) enables the debtor (or its trustee), upon entering bankruptcy, to decide whether the contract is a good deal for the estate going forward. If so, the debtor will want to assume the contract, fulfilling its obligations while benefiting from the counterparty’s performance. But if not, the debtor will want to reject the contract, repudiating any further performance of its duties. The bankruptcy court will generally approve that choice, under the deferential “business judgment” rule. Id., at 523. According to Section 365(g), “the rejection of an execu- tory contract[ ] constitutes a breach of such contract.” As both parties here agree, the counterparty thus has a claim against the estate for damages resulting from the debtor’s nonperformance. See Brief for Petitioner 17, 19; Brief for Respondent 30–31. But such a claim is unlikely to ever be paid in full. That is because the debtor’s breach is deemed to occur “immediately before the date of the filing of the [bankruptcy] petition,” rather than on the actual post-petition rejection date. §365(g)(1). By thus giving the counterparty a pre-petition claim, Section 365(g) places that party in the same boat as the debtor’s unsecured creditors, who in a typical bankruptcy may receive only cents on the dollar. See Bildisco, 465 U. S., at 531–532 (noting the higher priority of post-petition claims). In this case, the Bankruptcy Court (per usual) approved Tempnology’s proposed rejection of its executory licensing agreement with Mission. See App. to Pet. for Cert. 83–84. That meant, as laid out above, two things on which the parties agree. First, Tempnology could stop performing under the contract. And second, Mission could assert (for whatever it might be worth) a pre-petition claim in the bankruptcy proceeding for damages resulting from Tempnology’s nonperformance. But Tempnology thought still another consequence ensued, and it returned to the Bankruptcy Court for a declaratory judgment confirming its view. According to Tempnology, its rejection of the contract also terminated the rights it had granted Mission to use the Coolcore trademarks. Tempnology based its argument on a negative inference. See Motion in No. 15–11400 (Bkrtcy. Ct. NH), pp. 9–14. Several provisions in Section 365 state that a counterparty to specific kinds of agreements may keep exercising contractual rights after a debtor’s rejection. For example, Section 365(h) provides that if a bankrupt landlord rejects a lease, the tenant need not move out; instead, she may stay and pay rent (just as she did before) until the lease term expires. And still closer to home, Section 365(n) sets out a similar rule for some types of intellectual property licenses: If the debtor-licensor rejects the agreement, the licensee can continue to use the property (typically, a patent), so long as it makes whatever payments the contract demands. But Tempnology pointed out that neither Section 365(n) nor any similar provision covers trademark licenses. So, it reasoned, in that sort of contract a different rule must apply: The debtor’s rejection must extinguish the rights that the agreement had conferred on the trademark licensee. The Bankruptcy Court agreed. See In re Tempnology, LLC, 541 B.R. 1 (Bkrtcy. Ct. NH 2015). It held, relying on the same “negative inference,” that Tempnology’s rejection of the licensing agreement revoked Mission’s right to use the Coolcore marks. Id., at 7. The Bankruptcy Appellate Panel reversed, relying heavily on a decision of the Court of Appeals for the Seventh Circuit about the effects of rejection on trademark licensing agreements. See In re Tempnology, LLC, 559 B.R. 809, 820–823 (Bkrtcy. App. Panel CA1 2016); Sunbeam Products, Inc. v. Chicago Am. Mfg., LLC, 686 F.3d 372, 376–377 (CA7 2012). Rather than reason backward from Section 365(n) or similar provisions, the Panel focused on Section 365(g)’s statement that rejection of a contract “constitutes a breach.” Outside bankruptcy, the court explained, the breach of an agreement does not eliminate rights the contract had already conferred on the non-breaching party. See 559 B. R., at 820. So neither could a rejection of an agreement in bankruptcy have that effect. A rejection “convert[s]” a “debtor’s unfulfilled obligations” to a pre-petition damages claim. Id., at 822 (quoting Sunbeam, 686 F. 3d, at 377). But it does not “terminate the contract” or “vaporize[ ]” the counterparty’s rights. 559 B. R., at 820, 822 (quoting Sunbeam, 686 F. 3d, at 377). Mission could thus continue to use the Coolcore trademarks. But the Court of Appeals for the First Circuit rejected the Panel’s and Seventh Circuit’s view, and reinstated the Bankruptcy Court decision terminating Mission’s license. See In re Tempnology, LLC, 879 F.3d 389 (2018). The majority first endorsed that court’s inference from Section 365(n) and similar provisions. It next reasoned that special features of trademark law counsel against allowing a licensee to retain rights to a mark after the licensing agreement’s rejection. Under that body of law, the major- ity stated, the trademark owner’s “[f]ailure to monitor and exercise [quality] control” over goods associated with a trademark “jeopardiz[es] the continued validity of [its] own trademark rights.” Id., at 402. So if (the majority continued) a licensee can keep using a mark after an agreement’s rejection, the licensor will need to carry on its monitoring activities. And according to the majority, that would frustrate “Congress’s principal aim in providing for rejection”: to “release the debtor’s estate from burdensome obligations.” Ibid. (internal quotation marks omitted). Judge Torruella dissented, mainly for the Seventh Circuit’s reasons. See id., at 405–407. We granted certiorari to resolve the division between the First and Seventh Circuits. 586 U. S. ___ (2018). We now affirm the Seventh’s reasoning and reverse the decision below.[1] II Before reaching the merits, we pause to consider Tempnology’s claim that this case is moot. Under settled law, we may dismiss the case for that reason only if “it is impossible for a court to grant any effectual relief whatever” to Mission assuming it prevails. Chafin v. Chafin, 568 U.S. 165, 172 (2013) (internal quotation marks omitted). That demanding standard is not met here. Mission has presented a claim for money damages—essentially lost profits—arising from its inability to use the Coolcore trademarks between the time Tempnology rejected the licensing agreement and its scheduled expiration date. See Reply Brief 22, and n. 8. Such claims, if at all plausible, ensure a live controversy. See Memphis Light, Gas & Water Div. v. Craft, 436 U.S. 1, 8–9 (1978). For better or worse, nothing so shows a continuing stake in a dispute’s outcome as a demand for dollars and cents. See 13C C. Wright, A. Miller & E. Cooper, Federal Practice and Procedure §3533.3, p. 2 (3d ed. 2008) (Wright & Miller) (“[A] case is not moot so long as a claim for monetary relief survives”). Ultimate recovery on that demand may be uncertain or even unlikely for any number of reasons, in this case as in others. But that is of no moment. If there is any chance of money changing hands, Mission’s suit remains live. See Chafin, 568 U. S., at 172. Tempnology makes a flurry of arguments about why Mission is not entitled to damages, but none so clearly precludes recovery as to make this case moot. First, Tempnology contends that Mission suffered no injury because it “never used the trademark[s] during [the post-rejection] period.” Brief for Respondent 24; see Tr. of Oral Arg. 33. But that gets things backward. Mission’s non-use of the marks during that time is precisely what gives rise to its damages claim; had it employed the marks, it would not have lost any profits. So next, Tempnology argues that Mission’s non-use was its own “choice,” for which damages cannot lie. See id., at 26. But recall that the Bankruptcy Court held that Mission could not use the marks after rejection (and its decision remained in effect through the agreement’s expiration). See supra, at 4. And although (as Tempnology counters) the court issued “no injunction,” Brief for Respondent 26, that difference does not matter: Mission need not have flouted a crystal-clear ruling and courted yet more legal trouble to preserve its claim. Cf. 13B Wright & Miller §3533.2.2, at 852 (“[C]ompliance [with a judicial decision] does not moot [a case] if it remains possible to undo the effects of compliance,” as through compensation). So last, Tempnology claims that it bears no blame (and thus should not have to pay) for Mission’s injury because all it did was “ask[ ] the court to make a ruling.” Tr. of Oral Arg. 34–35. But whether Tempnology did anything to Mission amounting to a legal wrong is a prototypical merits question, which no court has addressed and which has no obvious answer. That means it is no reason to find this case moot. And so too for Tempnology’s further argument that Mission will be unable to convert any judgment in its favor to hard cash. Here, Tempnology notes that the bankruptcy estate has recently distributed all of its assets, leaving nothing to satisfy Mission’s judgment. See Brief for Respondent 27. But courts often adjudicate disputes whose “practical impact” is unsure at best, as when “a defendant is insolvent.” Chafin, 568 U. S., at 175. And Mission notes that if it prevails, it can seek the unwinding of prior distributions to get its fair share of the estate. See Reply Brief 23. So although this suit “may not make [Mission] rich,” or even better off, it remains a live controversy—allowing us to proceed. Chafin, 568 U. S., at 176. III What is the effect of a debtor’s (or trustee’s) rejection of a contract under Section 365 of the Bankruptcy Code? The parties and courts of appeals have offered us two starkly different answers. According to one view, a rejection has the same consequence as a contract breach outside bankruptcy: It gives the counterparty a claim for damages, while leaving intact the rights the counterparty has received under the contract. According to the other view, a rejection (except in a few spheres) has more the effect of a contract rescission in the non-bankruptcy world: Though also allowing a damages claim, the rejection terminates the whole agreement along with all rights it conferred. Today, we hold that both Section 365’s text and fundamental principles of bankruptcy law command the first, rejection-as-breach approach. We reject the competing claim that by specifically enabling the counterparties in some contracts to retain rights after rejection, Congress showed that it wanted the counterparties in all other contracts to lose their rights. And we reject an argument for the rescission approach turning on the distinctive features of trademark licenses. Rejection of a contract—any contract—in bankruptcy operates not as a rescission but as a breach. A We start with the text of the Code’s principal provisions on rejection—and find that it does much of the work. As noted earlier, Section 365(a) gives a debtor the option, subject to court approval, to “assume or reject any executory contract.” See supra, at 2. And Section 365(g) describes what rejection means. Rejection “constitutes a breach of [an executory] contract,” deemed to occur “immediately before the date of the filing of the petition.” See supra, at 3. Or said more pithily for current purposes, a rejection is a breach. And “breach” is neither a defined nor a specialized bankruptcy term. It means in the Code what it means in contract law outside bankruptcy. See Field v. Mans, 516 U.S. 59, 69 (1995) (Congress generally meant for the Bankruptcy Code to “incorporate the established meaning” of “terms that have accumulated settled meaning” (internal quotation marks omitted)). So the first place to go in divining the effects of rejection is to non-bankruptcy contract law, which can tell us the effects of breach. Consider a made-up executory contract to see how the law of breach works outside bankruptcy. A dealer leases a photocopier to a law firm, while agreeing to service it every month; in exchange, the firm commits to pay a monthly fee. During the lease term, the dealer decides to stop servicing the machine, thus breaching the agreement in a material way. The law firm now has a choice (assuming no special contract term or state law). The firm can keep up its side of the bargain, continuing to pay for use of the copier, while suing the dealer for damages from the service breach. Or the firm can call the whole deal off, halting its own payments and returning the copier, while suing for any damages incurred. See 13 R. Lord, Williston on Contracts §39:32, pp. 701–702 (4th ed. 2013) (“[W]hen a contract is breached in the course of performance, the injured party may elect to continue the contract or refuse to perform further”). But to repeat: The choice to terminate the agreement and send back the copier is for the law firm. By contrast, the dealer has no ability, based on its own breach, to terminate the agreement. Or otherwise said, the dealer cannot get back the copier just by refusing to show up for a service appointment. The contract gave the law firm continuing rights in the copier, which the dealer cannot unilaterally revoke. And now to return to bankruptcy: If the rejection of the photocopier contract “constitutes a breach,” as the Code says, then the same results should follow (save for one twist as to timing). Assume here that the dealer files a Chapter 11 petition and decides to reject its agreement with the law firm. That means, as above, that the dealer will stop servicing the copier. It means, too, that the law firm has an option about how to respond—continue the contract or walk away, while suing for whatever damages go with its choice. (Here is where the twist comes in: Because the rejection is deemed to occur “immediately before” bankruptcy, the firm’s damages suit is treated as a pre-petition claim on the estate, which will likely receive only cents on the dollar. See supra, at 3.) And most important, it means that assuming the law firm wants to keep using the copier, the dealer cannot take it back. A rejection does not terminate the contract. When it occurs, the debtor and counterparty do not go back to their pre-contract positions. Instead, the counterparty retains the rights it has received under the agreement. As after a breach, so too after a rejection, those rights survive. All of this, it will hardly surprise you to learn, is not just about photocopier leases. Sections 365(a) and (g) speak broadly, to “any executory contract[s].” Many licensing agreements involving trademarks or other property are of that kind (including, all agree, the Tempnology-Mission contract). The licensor not only grants a license, but provides associated goods or services during its term; the licensee pays continuing royalties or fees. If the licensor breaches the agreement outside bankruptcy (again, barring any special contract term or state law), everything said above goes. In particular, the breach does not revoke the license or stop the licensee from doing what it allows. See, e.g., Sunbeam, 686 F. 3d, at 376 (“Outside of bankruptcy, a licensor’s breach does not terminate a licensee’s right to use [the licensed] intellectual property”). And because rejection “constitutes a breach,” §365(g), the same consequences follow in bankruptcy. The debtor can stop performing its remaining obligations under the agreement. But the debtor cannot rescind the license already conveyed. So the licensee can continue to do whatever the license authorizes. In preserving those rights, Section 365 reflects a general bankruptcy rule: The estate cannot possess anything more than the debtor itself did outside bankruptcy. See Board of Trade of Chicago v. Johnson, 264 U.S. 1, 15 (1924) (establishing that principle); §541(a)(1) (defining the estate to include the “interests of the debtor in property” (emphasis added)). As one bankruptcy scholar has put the point: Whatever “limitation[s] on the debtor’s property [apply] outside of bankruptcy[ ] appl[y] inside of bankruptcy as well. A debtor’s property does not shrink by happenstance of bankruptcy, but it does not expand, either.” D. Baird, Elements of Bankruptcy 97 (6th ed. 2014). So if the not-yet debtor was subject to a counterparty’s contractual right (say, to retain a copier or use a trademark), so too is the trustee or debtor once the bankruptcy petition has been filed. The rejection-as-breach rule (but not the rejection-as-rescission rule) ensures that result. By insisting that the same counterparty rights survive rejection as survive breach, the rule prevents a debtor in bankruptcy from recapturing interests it had given up. And conversely, the rejection-as-rescission approach would circumvent the Code’s stringent limits on “avoidance” actions—the exceptional cases in which trustees (or debtors) may indeed unwind pre-bankruptcy transfers that undermine the bankruptcy process. The most not- able example is for fraudulent conveyances—usually, something-for-nothing transfers that deplete the estate (and so cheat creditors) on the eve of bankruptcy. See §548(a). A trustee’s avoidance powers are laid out in a discrete set of sections in the Code, see §§544–553, far away from Section 365. And they can be invoked in only narrow circumstances—unlike the power of rejection, which may be exercised for any plausible economic reason. See, e.g., §548(a) (describing the requirements for avoiding fraudulent transfers); supra, at 2–3. If trustees (or debtors) could use rejection to rescind previously granted interests, then rejection would become functionally equivalent to avoidance. Both, that is, would roll back a prior transfer. And that result would subvert everything the Code does to keep avoidances cabined—so they do not threaten the rule that the estate can take only what the debtor possessed before filing. Again, then, core tenets of bankruptcy law push in the same direction as Section 365’s text: Rejection is breach, and has only its consequences. B Tempnology’s main argument to the contrary, here as in the courts below, rests on a negative inference. See Brief for Respondent 33–41; supra, at 3–4. Several provisions of Section 365, Tempnology notes, “identif[y] categories of contracts under which a counterparty” may retain specified contract rights “notwithstanding rejection.” Brief for Respondent 34. Sections 365(h) and (i) make clear that certain purchasers and lessees of real property and timeshare interests can continue to exercise rights after a debtor has rejected the lease or sales contract. See §365(h)(1) (real-property leases); §365(i) (real-property sales contracts); §§365(h)(2), (i) (timeshare interests). And Section 365(n) similarly provides that licensees of some intellectual property—but not trademarks—retain contractual rights after rejection. See §365(n); §101(35A); supra, at 4. Tempnology argues from those provisions that the ordinary consequence of rejection must be something different—i.e., the termination, rather than survival, of contractual rights previously granted. Otherwise, Tempnology concludes, the statute’s “general rule” would “swallow the exceptions.” Brief for Respondent 19. But that argument pays too little heed to the main provisions governing rejection and too much to subsidiary ones. On the one hand, it offers no account of how to read Section 365(g) (recall, rejection “constitutes a breach”) to say essentially its opposite (i.e., that rejection and breach have divergent consequences). On the other hand, it treats as a neat, reticulated scheme of “narrowly tailored exception[s],” id., at 36 (emphasis deleted), what history reveals to be anything but. Each of the provisions Tempnology highlights emerged at a different time, over a span of half a century. See, e.g., 52Stat. 881 (1938) (real-property leases); §1(b), 102Stat. 2538 (1988) (intellectual property). And each responded to a discrete problem—as often as not, correcting a judicial ruling of just the kind Tempnology urges. See Andrew, Executory Contracts in Bankruptcy, 59 U. Colo. L. Rev. 845, 911–912, 916–919 (1988) (identifying judicial decisions that the provisions overturned); compare, e.g., In re Sombrero Reef Club, Inc., 18 B.R. 612, 618–619 (Bkrtcy. Ct. SD Fla. 1982), with, e.g., §§365(h)(2), (i). Read as generously as possible to Tempnology, this mash-up of legislative interventions says nothing much of anything about the content of Section 365(g)’s general rule. Read less generously, it affirma- tively refutes Tempnology’s rendition. As one bankruptcy scholar noted after an exhaustive review of the history: “What the legislative record [reflects] is that whenever Congress has been confronted with the consequences of the [view that rejection terminates all contractual rights], it has expressed its disapproval.” Andrew, 59 U. Colo. L. Rev., at 928. On that account, Congress enacted the provisions, as and when needed, to reinforce or clarify the general rule that contractual rights survive rejection.[2] Consider more closely, for example, Congress’s enactment of Section 365(n), which addresses certain intellectual property licensing agreements. No one disputes how that provision came about. In Lubrizol Enterprises v. Richmond Metal Finishers, the Fourth Circuit held that a debtor’s rejection of an executory contract worked to revoke its grant of a patent license. See 756 F.2d 1043, 1045–1048 (1985). In other words, Lubrizol adopted the same rule for patent licenses that the First Circuit announced for trademark licenses here. Congress sprang into action, drafting Section 365(n) to reverse Lubrizol and ensure the continuation of patent (and some other intellectual property) licensees’ rights. See 102Stat. 2538 (1988); S. Rep. No. 100–505, pp. 2–4 (1988) (explaining that Section 365(n) “corrects [Lubrizol’s] perception” that “Section 365 was ever intended to be a mechanism for stripping innocent licensee[s] of rights”). As Tempnology highlights, that provision does not cover trademark licensing agreements, which continue to fall, along with most other contracts, within Section 365(g)’s general rule. See Brief for Respondent 38. But what of that? Even put aside the claim that Section 365(n) is part of a pattern—that Congress whacked Tempnology’s view of rejection wherever it raised its head. See supra, at 13. Still, Congress’s repudiation of Lubrizol for patent contracts does not show any intent to ratify that decision’s approach for almost all others. Which is to say that no negative inference arises. Congress did nothing in adding Section 365(n) to alter the natural reading of Section 365(g)—that rejection and breach have the same results. Tempnology’s remaining argument turns on the way special features of trademark law may affect the fulfillment of the Code’s goals. Like the First Circuit below, Tempnology here focuses on a trademark licensor’s duty to monitor and “exercise quality control over the goods and services sold” under a license. Brief for Respondent 20; see supra, at 5. Absent those efforts to keep up quality, the mark will naturally decline in value and may eventually become altogether invalid. See 3 J. McCarthy, Trademarks and Unfair Competition §18:48, pp. 18–129, 18–133 (5th ed. 2018). So (Tempnology argues) unless rejection of a trademark licensing agreement terminates the licensee’s rights to use the mark, the debtor will have to choose between expending scarce resources on quality control and risking the loss of a valuable asset. See Brief for Respondent 59. “Either choice,” Tempnology concludes, “would impede a [debtor’s] ability to reorganize,” thus “undermining a fundamental purpose of the Code.” Id., at 59–60. To begin with, that argument is a mismatch with Tempnology’s reading of Section 365. The argument is trademark-specific. But Tempnology’s reading of Section 365 is not. Remember, Tempnology construes that section to mean that a debtor’s rejection of a contract terminates the counterparty’s rights “unless the contract falls within an express statutory exception.” Id., at 27–28; see supra, at 12. That construction treats trademark agreements identically to most other contracts; the only agreements getting different treatment are those falling within the discrete provisions just discussed. And indeed, Tempnol- ogy could not have discovered, however hard it looked, any trademark-specific rule in Section 365. That section’s special provisions, as all agree, do not mention trademarks; and the general provisions speak, well, generally. So Tempnology is essentially arguing that distinctive features of trademarks should persuade us to adopt a construction of Section 365 that will govern not just trademark agreements, but pretty nearly every executory contract. However serious Tempnology’s trademark-related concerns, that would allow the tail to wag the Doberman. And even putting aside that incongruity, Tempnology’s plea to facilitate trademark licensors’ reorganizations cannot overcome what Sections 365(a) and (g) direct. The Code of course aims to make reorganizations possible. But it does not permit anything and everything that might advance that goal. See, e.g., Florida Dept. of Revenue v. Piccadilly Cafeterias, Inc., 554 U.S. 33, 51 (2008) (observing that in enacting Chapter 11, Congress did not have “a single purpose,” but “str[uck] a balance” among multiple competing interests (internal quotation marks omitted)). Here, Section 365 provides a debtor like Tempnology with a powerful tool: Through rejection, the debtor can escape all of its future contract obligations, without having to pay much of anything in return. See supra, at 3. But in allowing rejection of those contractual duties, Section 365 does not grant the debtor an exemption from all the burdens that generally applicable law—whether involving contracts or trademarks—imposes on property owners. See 28 U. S. C. §959(b) (requiring a trustee to manage the estate in accordance with applicable law). Nor does Section 365 relieve the debtor of the need, against the backdrop of that law, to make economic decisions about preserving the estate’s value—such as whether to invest the resources needed to maintain a trademark. In thus delineating the burdens that a debtor may and may not escape, Congress also weighed (among other things) the legitimate interests and expectations of the debtor’s counterparties. The resulting balance may indeed impede some reorganizations, of trademark licensors and others. But that is only to say that Section 365’s edict that rejection is breach expresses a more complex set of aims than Tempnology acknowledges. IV For the reasons stated above, we hold that under Section 365, a debtor’s rejection of an executory contract in bankruptcy has the same effect as a breach outside bankruptcy. Such an act cannot rescind rights that the contract previously granted. Here, that construction of Section 365 means that the debtor-licensor’s rejection cannot revoke the trademark license. We accordingly reverse the judgment of the Court of Appeals and remand the case for further proceedings consistent with this opinion. It is so ordered. Notes 1 In its briefing before this Court, Mission contends that its exclusive distribution rights survived the licensing agreement’s rejection for the same reason as its trademark rights did. See Brief for Petitioner 40–44; supra, at 2. But the First Circuit held that Mission had waived that argument, see 879 F. 3d, at 401, and we have no reason to doubt that conclusion. Our decision thus affects only Mission’s trademark rights. 2 At the same time, Congress took the opportunity when drafting those provisions to fill in certain details, generally left to state law, about the post-rejection relationship between the debtor and counterparty. See, e.g., Andrew, Executory Contracts in Bankruptcy, 59 U. Colo. L. Rev. 845, 903, n. 200 (1988) (describing Congress’s addition of subsidiary rules for real property leases in Section 365(h)); Brief for United States as Amicus Curiae 29 (noting that Congress similarly set out detailed rules for patent licenses in Section 365(n)). The provisions are therefore not redundant of Section 365(g): Each sets out a remedial scheme embellishing on or tweaking the general rejection-as-breach rule.

588.US.2018_18-6210

Petitioner Gerald Mitchell was arrested for operating a vehicle while intoxicated after a preliminary breath test registered a blood alcohol concentration (BAC) that was triple Wisconsin’s legal limit for driving. As is standard practice, the arresting officer drove Mitchell to a police station for a more reliable breath test using evidence-grade equipment. By the time Mitchell reached the station, he was too lethargic for a breath test, so the officer drove him to a nearby hospital for a blood test. Mitchell was unconscious by the time he arrived at the hospital, but his blood was drawn anyway under a state law that presumes that a person incapable of withdrawing implied consent to BAC testing has not done so. The blood analysis showed Mitchell’s BAC to be above the legal limit, and he was charged with violating two drunk-driving laws. Mitchell moved to suppress the results of the blood test on the ground that it violated his Fourth Amendment right against “unreasonable searches” because it was conducted without a warrant. The trial court denied the motion, and Mitchell was convicted. On certification from the intermediate appellate court, the Wisconsin Supreme Court affirmed the lawfulness of Mitchell’s blood test. Held: The judgment is vacated, and the case is remanded. 2018 WI 84, 383 Wis. 2d 192, 914 N.W.2d 151, vacated and remanded. Justice Alito, joined by The Chief Justice, Justice Breyer, and Justice Kavanaugh, concluded that when a driver is unconscious and cannot be given a breath test, the exigent-circumstances doctrine generally permits a blood test without a warrant. Pp. 5–17. (a) BAC tests are Fourth Amendment searches. See Birchfield v. North Dakota, 579 U. S. ___, ___. A warrant is normally required for a lawful search, but there are well-defined exceptions to this rule, including the “exigent circumstances” exception, which allows warrantless searches “to prevent the imminent destruction of evidence.” Missouri v. McNeely, 569 U.S. 141, 149. In McNeely, this Court held that the fleeting nature of blood-alcohol evidence alone was not enough to bring BAC testing within the exigency exception. Id., at 156. But in Schmerber v. California, 384 U.S. 757, the dissipation of BAC did justify a blood test of a drunk driver whose accident gave police other pressing duties, for then the further delay caused by a warrant application would indeed have threatened the destruction of evidence. Like Schmerber, unconscious-driver cases will involve a heightened degree of urgency for several reasons. And when the driver’s stupor or unconsciousness deprives officials of a reasonable opportunity to administer a breath test using evidence-grade equipment, a blood test will be essential for achieving the goals of BAC testing. Pp. 5–7. (b) Under the exigent circumstances exception, a warrantless search is allowed when “ ‘there is compelling need for official action and no time to secure a warrant.’ ” McNeely, 569 U. S., at 149. Pp. 7–16. (1) There is clearly a “compelling need” for a blood test of drunk-driving suspects whose condition deprives officials of a reasonable opportunity to conduct a breath test. First, highway safety is a vital public interest—a “compelling” and “paramount” interest, Mackey v. Montrym, 443 U.S. 1, 17–18. Second, when it comes to promoting that interest, federal and state lawmakers have long been convinced that legal limits on a driver’s BAC make a big difference. And there is good reason to think that such laws have worked. Birchfield, 579 U. S., at ___. Third, enforcing BAC limits obviously requires a test that is accurate enough to stand up in court. Id., at ___. And such testing must be prompt because it is “a biological certainty” that “[a]lcohol dissipates from the bloodstream,” “literally disappearing by the minute.” McNeely, 569 U. S., at 169 (Roberts, C. J., concurring). Finally, when a breath test is unavailable to promote the interests served by legal BAC limits, “a blood draw becomes necessary.” Id., at 170. Pp. 9–12. (2) Schmerber demonstrates that an exigency exists when (1) BAC evidence is dissipating and (2) some other factor creates pressing health, safety, or law enforcement needs that would take priority over a warrant application. Because both conditions are met when a drunk-driving suspect is unconscious, Schmerber controls. A driver’s unconsciousness does not just create pressing needs; it is itself a medical emergency. In such a case, as in Schmerber, an officer could “reasonably have believed that he was confronted with an emergency.” 384 U. S., at 771. And in many unconscious-driver cases, the exigency will be especially acute. A driver so drunk as to lose consciousness is quite likely to crash, giving officers a slew of urgent tasks beyond that of securing medical care for the suspect—tasks that would require them to put off applying for a warrant. The time needed to secure a warrant may have shrunk over the years, but it has not disappeared; and forcing police to put off other urgent tasks for even a relatively short period of time may have terrible collateral costs. Pp. 12–16. (c) On remand, Mitchell may attempt to show that his was an unusual case, in which his blood would not have been drawn had police not been seeking BAC information and police could not have reasonably judged that a warrant application would interfere with other pressing needs or duties. Pp. 16–17. Justice Thomas would apply a per se rule, under which the natural metabolization of alcohol in the blood stream “creates an exigency once police have probable cause to believe the driver is drunk,” regardless of whether the driver is conscious. Missouri v. McNeely, 569 U.S. 141, 178 (Thomas, J., dissenting). Pp. 1–4. Alito, J., announced the judgment of the Court and delivered an opinion, in which Roberts, C. J., and Breyer and Kavanaugh, JJ., joined. Thomas, J., filed an opinion concurring in the judgment. Sotomayor, J., filed a dissenting opinion, in which Ginsburg and Kagan, JJ., joined. Gorsuch, J., filed a dissenting opinion.

in which The Chief Justice, Justice Breyer, and Justice Kavanaugh join. In this case, we return to a topic that we have addressed twice in recent years: the circumstances under which a police officer may administer a warrantless blood alcohol concentration (BAC) test to a motorist who appears to have been driving under the influence of alcohol. We have previously addressed what officers may do in two broad categories of cases. First, an officer may conduct a BAC test if the facts of a particular case bring it within the exigent-circumstances exception to the Fourth Amendment’s general requirement of a warrant. Second, if an officer has probable cause to arrest a motorist for drunk driving, the officer may conduct a breath test (but not a blood test) under the rule allowing warrantless searches of a person incident to arrest. Today, we consider what police officers may do in a narrow but important category of cases: those in which the driver is unconscious and therefore cannot be given a breath test. In such cases, we hold, the exigent-circumstances rule almost always permits a blood test without a warrant. When a breath test is impossible, enforcement of the drunk-driving laws depends upon the administration of a blood test. And when a police officer encounters an unconscious driver, it is very likely that the driver would be taken to an emergency room and that his blood would be drawn for diagnostic purposes even if the police were not seeking BAC information. In addition, police officers most frequently come upon unconscious drivers when they report to the scene of an accident, and under those circumstances, the officers’ many responsibilities—such as attending to other injured drivers or passengers and preventing further accidents—may be incompatible with the procedures that would be required to obtain a warrant. Thus, when a driver is unconscious, the general rule is that a warrant is not needed. I A In Birchfield v. North Dakota, 579 U. S. ___ (2016), we recounted the country’s efforts over the years to address the terrible problem of drunk driving. Today, “all States have laws that prohibit motorists from driving with a [BAC] that exceeds a specified level.” Id., at ___ (slip op., at 2). And to help enforce BAC limits, every State has passed what are popularly called implied-consent laws. Ibid. As “a condition of the privilege of” using the public roads, these laws require that drivers submit to BAC testing “when there is sufficient reason to believe they are violating the State’s drunk-driving laws.” Id., at ___, ___ (slip op., at 2, 6). Wisconsin’s implied-consent law is much like those of the other 49 States and the District of Columbia. It deems drivers to have consented to breath or blood tests if an officer has reason to believe they have committed one of several drug- or alcohol-related offenses.[1] See Wis. Stat. §§343.305(2), (3). Officers seeking to conduct a BAC test must read aloud a statement declaring their intent to administer the test and advising drivers of their options and the implications of their choice. §343.305(4). If a driver’s BAC level proves too high, his license will be suspended; but if he refuses testing, his license will be revoked and his refusal may be used against him in court. See ibid. No test will be administered if a driver refuses—or, as the State would put it, “withdraws” his statutorily presumed consent. But “[a] person who is unconscious or otherwise not capable of withdrawing consent is presumed not to have” withdrawn it. §343.305(3)(b). See also §§343.305(3)(ar)1–2. More than half the States have provisions like this one regarding unconscious drivers. B The sequence of events that gave rise to this case began when Officer Alexander Jaeger of the Sheboygan Police Department received a report that petitioner Gerald Mitchell, appearing to be very drunk, had climbed into a van and driven off. Jaeger soon found Mitchell wandering near a lake. Stumbling and slurring his words, Mitchell could hardly stand without the support of two officers. Jaeger judged a field sobriety test hopeless, if not dangerous, and gave Mitchell a preliminary breath test. It registered a BAC level of 0.24%, triple the legal limit for driving in Wisconsin. Jaeger arrested Mitchell for operating a vehicle while intoxicated and, as is standard practice, drove him to a police station for a more reliable breath test using better equipment. On the way, Mitchell’s condition continued to deteriorate—so much so that by the time the squad car had reached the station, he was too lethargic even for a breath test. Jaeger therefore drove Mitchell to a nearby hospital for a blood test; Mitchell lost consciousness on the ride over and had to be wheeled in. Even so, Jaeger read aloud to a slumped Mitchell the standard statement giving drivers a chance to refuse BAC testing. Hearing no response, Jaeger asked hospital staff to draw a blood sample. Mitchell remained unconscious while the sample was taken, and analysis of his blood showed that his BAC, about 90 minutes after his arrest, was 0.222%. Mitchell was charged with violating two related drunk-driving provisions. See §§346.63(1)(a), (b). He moved to suppress the results of the blood test on the ground that it violated his Fourth Amendment right against “unreason-able searches” because it was conducted without a warrant. Wisconsin chose to rest its response on the notion that its implied-consent law (together with Mitchell’s free choice to drive on its highways) rendered the blood test a consensual one, thus curing any Fourth Amendment problem. In the end, the trial court denied Mitchell’s motion to suppress, and a jury found him guilty of the charged offenses. The intermediate appellate court certified two questions to the Wisconsin Supreme Court: first, whether compliance with the State’s implied-consent law was sufficient to show that Mitchell’s test was consistent with the Fourth Amendment and, second, whether a warrantless blood draw from an unconscious person violates the Fourth Amendment. See 2018 WI 84, ¶15, 383 Wis. 2d 192, 202–203, 914 N.W.2d 151, 155–156 (2018). The Wisconsin Supreme Court affirmed Mitchell’s convictions, and we granted certiorari, 586 U. S. ___ (2019), to decide “[w]hether a statute author- izing a blood draw from an unconscious motorist provides an exception to the Fourth Amendment warrant requirement,” Pet. for Cert. ii. II In considering Wisconsin’s implied-consent law, we do not write on a blank slate. “Our prior opinions have referred approvingly to the general concept of implied-consent laws that impose civil penalties and evidentiary consequences on motorists who refuse to comply.” Birchfield, 579 U. S., at ___ (slip op., at 36). But our decisions have not rested on the idea that these laws do what their popular name might seem to suggest—that is, create actual consent to all the searches they authorize. Instead, we have based our decisions on the precedent regarding the specific constitutional claims in each case, while keeping in mind the wider regulatory scheme developed over the years to combat drunk driving. That scheme is centered on legally specified BAC limits for drivers—limits enforced by the BAC tests promoted by implied-consent laws. Over the last 50 years, we have approved many of the defining elements of this scheme. We have held that forcing drunk-driving suspects to undergo a blood test does not violate their constitutional right against self-incrimination. See Schmerber v. California, 384 U.S. 757, 765 (1966). Nor does using their refusal against them in court. See South Dakota v. Neville, 459 U.S. 553, 563 (1983). And punishing that refusal with automatic license revocation does not violate drivers’ due process rights if they have been arrested upon probable cause, Mackey v. Montrym, 443 U.S. 1 (1979); on the contrary, this kind of summary penalty is “unquestionably legitimate.” Neville, supra, at 560. These cases generally concerned the Fifth and Fourteenth Amendments, but motorists charged with drunk driving have also invoked the Fourth Amendment’s ban on “unreasonable searches” since BAC tests are “searches.” See Birchfield, 579 U. S., at ___ (slip op., at 14). Though our precedent normally requires a warrant for a lawful search, there are well-defined exceptions to this rule. In Birchfield, we applied precedent on the “search-incident-to-arrest” exception to BAC testing of conscious drunk-driving suspects. We held that their drunk-driving arrests, taken alone, justify warrantless breath tests but not blood tests, since breath tests are less intrusive, just as informative, and (in the case of conscious suspects) readily available. Id., at ___ (slip op., at 35). We have also reviewed BAC tests under the “exigent circumstances” exception—which, as noted, allows warrantless searches “to prevent the imminent destruction of evidence.” Missouri v. McNeely, 569 U.S. 141, 149 (2013). In McNeely, we were asked if this exception covers BAC testing of drunk-driving suspects in light of the fact that blood-alcohol evidence is always dissipating due to “natural metabolic processes.” Id., at 152. We answered that the fleeting quality of BAC evidence alone is not enough. Id., at 156. But in Schmerber it did justify a blood test of a drunk driver who had gotten into a car accident that gave police other pressing duties, for then the “further delay” caused by a warrant application really “would have threatened the destruction of evidence.” McNeely, supra, at 152 (emphasis added). Like Schmerber, this case sits much higher than McNeely on the exigency spectrum. McNeely was about the minimum degree of urgency common to all drunk-driving cases. In Schmerber, a car accident heightened that urgency. And here Mitchell’s medical condition did just the same. Mitchell’s stupor and eventual unconsciousness also deprived officials of a reasonable opportunity to administer a breath test. To be sure, Officer Jaeger managed to conduct “a preliminary breath test” using a portable machine when he first encountered Mitchell at the lake. App. to Pet. for Cert. 60a. But he had no reasonable opportunity to give Mitchell a breath test using “evidence-grade breath testing machinery.” Birchfield, 579 U. S., at ___ (Sotomayor, J., concurring in part and dissenting in part) (slip op., at 10). As a result, it was reasonable for Jaeger to seek a better breath test at the station; he acted with reasonable dispatch to procure one; and when Mitchell’s condition got in the way, it was reasonable for Jaeger to pursue a blood test. As Justice Sotomayor explained in her partial dissent in Birchfield: “There is a common misconception that breath tests are conducted roadside, immediately after a driver is arrested. While some preliminary testing is conducted roadside, reliability concerns with roadside tests confine their use in most circumstances to establishing probable cause for an arrest. . . . The standard evidentiary breath test is conducted after a motorist is arrested and transported to a police station, governmental building, or mobile testing facility where officers can access reliable, evidence-grade breath testing machinery.” Id., at ___ (slip op., at 10). Because the “standard evidentiary breath test is conducted after a motorist is arrested and transported to a police station” or another appropriate facility, ibid., the important question here is what officers may do when a driver’s unconsciousness (or stupor) eliminates any reasonable opportunity for that kind of breath test. III The Fourth Amendment guards the “right of the people to be secure in their persons . . . against unreasonable searches” and provides that “no Warrants shall issue, but upon probable cause.” A blood draw is a search of the person, so we must determine if its administration here without a warrant was reasonable. See Birchfield, 579 U. S. at ___ (slip op., at 14). Though we have held that a warrant is normally required, we have also “made it clear that there are exceptions to the warrant requirement.” Illinois v. McArthur, 531 U.S. 326, 330 (2001). And under the exception for exigent circumstances, a warrantless search is allowed when “ ‘there is compelling need for official action and no time to secure a warrant.’ ” McNeely, supra, at 149 (quoting Michigan v. Tyler, 436 U.S. 499, 509 (1978)). In McNeely, we considered how the exigent-circumstances exception applies to the broad category of cases in which a police officer has probable cause to believe that a motorist was driving under the influence of alcohol, and we do not revisit that question. Nor do we settle whether the exigent-circumstances exception covers the specific facts of this case.[2] Instead, we address how the exception bears on the category of cases encompassed by the question on which we granted certiorari—those involving unconscious drivers.[3] In those cases, the need for a blood test is compelling, and an officer’s duty to attend to more pressing needs may leave no time to seek a warrant. A The importance of the needs served by BAC testing is hard to overstate. The bottom line is that BAC tests are needed for enforcing laws that save lives. The specifics, in short, are these: Highway safety is critical; it is served by laws that criminalize driving with a certain BAC level; and enforcing these legal BAC limits requires efficient testing to obtain BAC evidence, which naturally dissipates. So BAC tests are crucial links in a chain on which vital interests hang. And when a breath test is unavail-able to advance those aims, a blood test becomes essential. Here we add a word about each of these points. First, highway safety is a vital public interest. For decades, we have strained our vocal chords to give adequate expression to the stakes. We have called highway safety a “compelling interest,” Mackey, 443 U. S., at 19; we have called it “paramount,” id., at 17. Twice we have referred to the effects of irresponsible driving as “slaughter” comparable to the ravages of war. Breithaupt v. Abram, 352 U.S. 432, 439 (1957); Perez v. Campbell, 402 U.S. 637, 657, 672 (1971) (Blackmun, J., concurring in result in part and dissenting in part). We have spoken of “carnage,” Neville, 459 U. S., at 558–559, and even “frightful carnage,” Tate v. Short, 401 U.S. 395, 401 (1971) (Blackmun, J., concurring). The frequency of preventable collisions, we have said, is “tragic,” Neville, supra, at 558, and “astounding,” Breithaupt, supra, at 439. And behind this fervent language lie chilling figures, all captured in the fact that from 1982 to 2016, alcohol-related accidents took roughly 10,000 to 20,000 lives in this Nation every single year. See National Highway Traffic Safety Admin. (NHTSA), Traffic Safety Facts 2016, p. 40 (May 2018). In the best years, that would add up to more than one fatality per hour. Second, when it comes to fighting these harms and promoting highway safety, federal and state lawmakers have long been convinced that specified BAC limits make a big difference. States resorted to these limits when earlier laws that included no “statistical definition of intoxication” proved ineffectual or hard to enforce. See Birchfield, 579 U. S., at ___–___ (slip op., at 2–3). The maximum permissible BAC, initially set at 0.15%, was first lowered to 0.10% and then to 0.08%. Id., at ___, ___–___ (slip op., at 3, 6–7). Congress encouraged this process by conditioning the award of federal highway funds on the establishment of a BAC limit of 0.08%, see 23 U. S. C. §163(a); 23 CFR §1225.1 (2012), and every State has adopted this limit.[4] Not only that, many States, including Wisconsin, have passed laws imposing increased penalties for recidivists or for drivers with a BAC level that exceeds a higher threshold. See Wis. Stat. §346.65(2)(am); Birchfield, 579 U. S., at ___ (slip op., at 7). There is good reason to think this strategy has worked. As we noted in Birchfield, these tougher measures corresponded with a dramatic drop in highway deaths and injuries: From the mid-1970’s to the mid-1980’s, “the number of annual fatalities averaged 25,000; by 2014 . . . , the number had fallen to below 10,000.” Id., at ___ (slip op., at 6). Third, enforcing BAC limits obviously requires a test that is accurate enough to stand up in court, id., at ___–___ (slip op., at 3–5); see also McNeely, 569 U. S., at 159–160 (plurality opinion). And we have recognized that “[e]xtraction of blood samples for testing is a highly effective means of” measuring “the influence of alcohol.” Schmerber, 384 U. S., at 771. Enforcement of BAC limits also requires prompt testing because it is “a biological certainty” that “[a]lcohol dissipates from the bloodstream at a rate of 0.01 percent to 0.025 percent per hour. . . . Evidence is literally disappearing by the minute.” McNeely, 569 U. S., at 169 (opinion of Roberts, C. J.). As noted, the ephemeral nature of BAC was “essential to our holding in Schmerber,” which itself allowed a warrantless blood test for BAC. Id., at 152 (opinion of the Court). And even when we later held that the exigent-circumstances exception would not permit a warrantless blood draw in every drunk-driving case, we acknowledged that delays in BAC testing can “raise questions about . . . accuracy.” Id., at 156. It is no wonder, then, that the implied-consent laws that incentivize prompt BAC testing have been with us for 65 years and now exist in all 50 States. Birchfield, supra, at ___ (slip op., at 6). These laws and the BAC tests they require are tightly linked to a regulatory scheme that serves the most pressing of interests. Finally, when a breath test is unavailable to promote those interests, “a blood draw becomes necessary.” McNeely, 569 U. S., at 170 (opinion of Roberts, C. J.). Thus, in the case of unconscious drivers, who cannot blow into a breathalyzer, blood tests are essential for achieving the compelling interests described above. Indeed, not only is the link to pressing interests here tighter; the interests themselves are greater: Drivers who are drunk enough to pass out at the wheel or soon afterward pose a much greater risk. It would be perverse if the more wanton behavior were rewarded—if the more harrowing threat were harder to punish. For these reasons, there clearly is a “compelling need” for a blood test of drunk-driving suspects whose condition deprives officials of a reasonable opportunity to conduct a breath test. Id., at 149 (opinion of the Court) (internal quotation marks omitted). The only question left, under our exigency doctrine, is whether this compelling need justifies a warrantless search because there is, furthermore, “ ‘no time to secure a warrant.’ ” Ibid. B We held that there was no time to secure a warrant before a blood test of a drunk-driving suspect in Schmerber because the officer there could “reasonably have believed that he was confronted with an emergency, in which the delay necessary to obtain a warrant, under the circumstances, threatened the destruction of evidence.” 384 U. S., at 770 (internal quotation marks omitted). So even if the constant dissipation of BAC evidence alone does not create an exigency, see McNeely, supra, at 150–151, Schmerber shows that it does so when combined with other pressing needs: “We are told that [1] the percentage of alcohol in the blood begins to diminish shortly after drinking stops, as the body functions to eliminate it from the system. Particularly in a case such as this, where [2] time had to be taken to bring the accused to a hospital and to investigate the scene of the accident, there was no time to seek out a magistrate and secure a warrant. Given these special facts, we conclude that the attempt to secure evidence of blood-alcohol content in this case [without a warrant] was . . . appropriate . . . .” 384 U. S., at 770–771. Thus, exigency exists when (1) BAC evidence is dissipating and (2) some other factor creates pressing health, safety, or law enforcement needs that would take priority over a warrant application. Both conditions are met when a drunk-driving suspect is unconscious, so Schmerber controls: With such suspects, too, a warrantless blood draw is lawful. 1 In Schmerber, the extra factor giving rise to urgent needs that would only add to the delay caused by a warrant application was a car accident; here it is the driver’s unconsciousness. Indeed, unconsciousness does not just create pressing needs; it is itself a medical emergency.[5] It means that the suspect will have to be rushed to the hospital or similar facility not just for the blood test itself but for urgent medical care.[6] Police can reasonably anticipate that such a driver might require monitoring, positioning, and support on the way to the hospital;[7] that his blood may be drawn anyway, for diagnostic purposes, immediately on arrival;[8] and that immediate medical treatment could delay (or otherwise distort the results of) a blood draw conducted later, upon receipt of a warrant, thus reducing its evidentiary value. See McNeely, supra, at 156 (plurality opinion). All of that sets this case apart from the uncomplicated drunk-driving scenarios addressed in McNeely. Just as the ramifications of a car accident pushed Schmerber over the line into exigency, so does the condition of an unconscious driver bring his blood draw under the exception. In such a case, as in Schmerber, an officer could “reasonably have believed that he was confronted with an emergency.” 384 U. S., at 770. Indeed, in many unconscious-driver cases, the exigency will be more acute, as elaborated in the briefing and argument in this case. A driver so drunk as to lose consciousness is quite likely to crash, especially if he passes out before managing to park. And then the accident might give officers a slew of urgent tasks beyond that of securing (and working around) medical care for the suspect. Police may have to ensure that others who are injured receive prompt medical attention; they may have to provide first aid themselves until medical personnel arrive at the scene. In some cases, they may have to deal with fatalities. They may have to preserve evidence at the scene and block or redirect traffic to prevent further accidents. These pressing matters, too, would require responsible officers to put off applying for a warrant, and that would only exacerbate the delay—and imprecision—of any subsequent BAC test. In sum, all these rival priorities would put officers, who must often engage in a form of triage, to a dilemma. It would force them to choose between prioritizing a warrant application, to the detriment of critical health and safety needs, and delaying the warrant application, and thus the BAC test, to the detriment of its evidentiary value and all the compelling interests served by BAC limits. This is just the kind of scenario for which the exigency rule was born—just the kind of grim dilemma it lives to dissolve. 2 Mitchell objects that a warrantless search is unnecessary in cases involving unconscious drivers because warrants these days can be obtained faster and more easily. But even in our age of rapid communication, “[w]arrants inevitably take some time for police officers or prosecutors to complete and for magistrate judges to review. Telephonic and electronic warrants may still require officers to follow time-consuming formalities designed to create an adequate record, such as preparing a duplicate warrant before calling the magistrate judge. . . . And improvements in communications technology do not guarantee that a magistrate judge will be available when an officer needs a warrant after making a late-night arrest.” McNeely, 569 U. S., at 155. In other words, with better technology, the time required has shrunk, but it has not disappeared. In the emergency scenarios created by unconscious drivers, forcing police to put off other tasks for even a relatively short period of time may have terrible collateral costs. That is just what it means for these situations to be emergencies. IV When police have probable cause to believe a person has committed a drunk-driving offense and the driver’s unconsciousness or stupor requires him to be taken to the hospital or similar facility before police have a reasonable opportunity to administer a standard evidentiary breath test, they may almost always order a warrantless blood test to measure the driver’s BAC without offending the Fourth Amendment. We do not rule out the possibility that in an unusual case a defendant would be able to show that his blood would not have been drawn if police had not been seeking BAC information, and that police could not have reasonably judged that a warrant application would interfere with other pressing needs or duties. Because Mitchell did not have a chance to attempt to make that showing, a remand for that purpose is necessary. * * * The judgment of the Supreme Court of Wisconsin is vacated, and the case is remanded for further proceedings. It is so ordered. Notes 1 Wisconsin also authorizes BAC testing of drivers involved in accidents that cause significant bodily harm, with or without probable cause of drunk driving. See Wis. Stat. §343.305(3)2 (2016). We do not address those provisions. And while Wisconsin’s and other implied-consent laws permit urine tests, those tests are less common, see Birchfield v. North Dakota, 579 U. S. ___, ___, n. 1 (2016) (slip op., at 6, n. 1), and we do not consider them here. 2 Justice Sotomayor’s dissent argues that Wisconsin waived the argument that we now adopt, but the dissent paints a misleading picture of both the proceedings below and the ground for our decision. First, as to the proceedings below, the dissent contends that the sole question certified to the Wisconsin Supreme Court was “ ‘whether the warrantless blood draw of an unconscious motorist pursuant to Wisconsin’s implied consent law, where no exigent circumstances exist or have been argued, violates the Fourth Amendment.’ ” Post, at 3 (quoting App. 61). That is indeed how the intermediate appellate court understood the issue in the case, but the State Supreme Court took a broader view, as was its right. It regarded the appeal as presenting two questions, one of which was “whether a warrantless blood draw from an unconscious person pursuant to Wis. Stat. §343.305(3)(b) violates the Fourth Amendment.” See 383 Wis. 2d 192, 202–203, 914 N.W.2d 151,155–156 (2018). This broad question easily encompasses the rationale that we adopt today. Second, after noting that the State did not attempt below to make a case-specific showing of exigent circumstances, the dissent claims that our decision is based on this very ground. But that is not at all the basis for our decision. We do not hold that the State established that the facts of this particular case involve exigent circumstances under McNeely. Rather, we adopt a rule for an entire category of cases—those in which a motorist believed to have driven under the influence of alcohol is unconscious and thus cannot be given a breath test. This rule is not based on what happened in petitioner’s particular case but on the circumstances generally present in cases that fall within the scope of the rule. Those are just the sorts of features of unconscious-driver cases that Wisconsin brought to our attention, see Brief for Respondent 54–55; Tr. of Oral Arg. at 32–34, 48–51, which petitioner addressed, see Reply Brief at 14–15; Tr. of Oral Arg. at 15–20, 23–24, 29–31, 63–66. So it is entirely proper for us to decide the case on this ground. See Thigpen v. Roberts, 468 U.S. 27, 29–30 (1984). 3 While our exigent-circumstances precedent requires a “ ‘totality of the circumstances’ ” analysis, “the circumstances in drunk driving cases are often typical, and the Court should be able to offer guidance on how police should handle cases like the one before us.” McNeely, 569 U. S., at 166 (Roberts, C. J., concurring in part and dissenting in part). Indeed, our exigency case law is full of general rules providing such guidance. Thus, we allow police to proceed without a warrant when an occupant of a home requires “emergency assistance,” Brigham City v. Stuart, 547 U.S. 398, 403 (2006); when a building is on fire, see Michigan v. Tyler, 436 U.S. 499, 509 (1978); and when an armed robber has just entered a home, see United States v. Santana, 427 U.S. 38 (1976). “In each of these cases, the requirement that we base our decision on the ‘totality of the circumstances’ has not prevented us from spelling out a general rule for the police to follow.” McNeely, supra, at 168 (opinion of Roberts, C. J.). Neither does it prevent us here. 4 See NHTSA, Alcohol and Highway Safety: A Review of the State of Knowledge 167 (DOT HS 811 374, Mar. 2011). 5 See National Institutes of Health, U. S. National Library of Medicine, MedlinePlus, Unconsciousness (June 3, 2019), https://medlineplus. gov/ency/article/000022.htm (all Internet materials as last visited June 25, 2019). 6 Limmer et al., Emergency Care 598 (13th ed. 2016). 7 See id., at 593–594. 8 See J. Kwasnoski, G. Partridge, & J. Stephen, Officer’s DUI Handbook 142 (6th ed. 2013) (“[M]ost hospitals routinely withdraw blood from the driver immediately upon admittance”); see also E. Mitchell & R. Medzon, Introduction to Emergency Medicine 269 (2005) (“Serum glucose and blood alcohol concentrations are two pieces of information that are of paramount importance when an apparently intoxicated patient arrives at the [emergency room]”); Mayo Clinic, Alcohol Poisoning: Diagnosis & Treatment (2019), https://www.mayoclinic.org/diseases-conditions/alcohol-poisoning/diagnosis-treatment/drc-20354392. In this respect, the case for allowing a blood draw is stronger here than in Schmerber v. California, 384 U.S. 757 (1966). In the latter, it gave us pause that blood draws involve piercing a person’s skin. See id., at 762, 770. But since unconscious suspects will often have their skin pierced and blood drawn for diagnostic purposes, allowing law enforcement to use blood taken from that initial piercing would not increase the bodily intrusion. In fact, dispensing with the warrant rule could lessen the intrusion. It could enable authorities to use blood obtained by hospital staff when the suspect is admitted rather than having to wait to hear back about a warrant and then order what might be a second blood draw.

587.US.2018_17-8995

Petitioner Mont was released from federal prison in 2012 and began a 5-year term of supervised release that was scheduled to end on March 6, 2017. On June 1, 2016, he was arrested on state drug trafficking charges and has been in state custody since that time. In October 2016, Mont pleaded guilty to state charges. He then admitted in a filing in Federal District Court that he violated his supervised-release conditions by virtue of the new state convictions, and he requested a hearing. The District Court scheduled a hearing for November, but later rescheduled it several times to allow the state court to first sentence Mont. On March 21, 2017, Mont was sentenced to six years’ imprisonment, and his roughly 10 months of pretrial custody were credited as time served. On March 30, the District Court issued a warrant for Mont and set a supervised-release hearing. Mont then challenged the District Court’s jurisdiction on the ground that his supervised release had been set to expire on March 6. The District Court ruled that it had jurisdiction under 18 U. S. C. §3583(i) based on a summons it had issued in November 2016. It then revoked Mont’s supervised release and ordered him to serve an additional 42 months’ imprisonment to run consecutive to his state sentence. The Sixth Circuit affirmed on alternative grounds, holding that Mont’s supervised-release period was tolled under §3624(e), which provides that a “term of supervised release does not run during any period in which the person is imprisoned in connection with a conviction for a . . . crime unless the imprisonment is for a period of less than 30 consecutive days.” Because the roughly 10 months of pretrial custody was “in connection with [Mont’s] conviction” and therefore tolled the period of supervised release, the court concluded that there was ample time left on Mont’s term of supervised release when the March warrant issued. Held: Pretrial detention later credited as time served for a new conviction is “imprison[ment] in connection with a conviction” and thus tolls the supervised-release term under §3624(e), even if the court must make the tolling calculation after learning whether the time will be credited. Pp. 6–13. (a) The text of §3624(e) compels this reading. First, dictionary definitions of the term “imprison,” both now and at the time Congress created supervised release, may very well encompass pretrial detention, and Mont has not pressed any serious argument to the contrary. Second, the phrase “in connection with a conviction” encompasses a period of pretrial detention for which a defendant receives credit against the sentence ultimately imposed. “In connection with” can bear a “broad interpretation,” Merrill Lynch, Pierce, Fenner & Smith Inc. v. Dabit, 547 U.S. 71, 85, but the outer bounds need not be determined here, as pretrial incarceration is directly tied to the conviction when it is credited toward the new sentence. This reading is buttressed by the fact that Congress, like most States, instructs courts calculating a term of imprisonment to credit pretrial detention as time served on a subsequent conviction. See §3585(b)(1). Third, the text undeniably requires courts to retrospectively calculate whether a period of pretrial detention should toll a period of supervised release by including the 30-day minimum. The statute does not require courts to make a tolling determination as soon as a defendant is arrested on new charges or to continually reassess the tolling calculation throughout the pretrial-detention period. Its 30-day minimum-incarceration threshold contemplates the opposite. Pp. 6–8. (b) The statutory context also supports this reading. First, §3624(e) provides that supervised release “runs concurrently” with “probation or supervised release or parole for another offense,” but excludes periods of “imprison[ment]” served “in connection with a conviction.” This juxtaposition reinforces the fact that prison time is “not interchangeable” with supervised release, United States v. Johnson, 529 U.S. 53, 59, and furthers the statutory design of “successful[ly] transition[ing]” a defendant from “prison to liberty,” Johnson v. United States, 529 U.S. 694, 708–709. Second, it would be an exceedingly odd construction of the statute to give a defendant the windfall of satisfying a new sentence of imprisonment and an old sentence of supervised release with the same period of pretrial detention. Supervised release is a form of punishment prescribed along with a term of imprisonment as part of the same sentence. And Congress denies defendants credit for time served if the detention time has already “been credited against another sentence.” §3585(b). Pp. 8–10. (c) Mont’s argument that the statute’s present tense forbids any backward looking tolling analysis confuses the rule with a court’s analysis of whether that rule was satisfied. The present-tense phrasing does not address whether a judge must be able to make a supervised-release determination at any given time. Moreover, any uncertainty about whether supervised release is tolled matters little from either the court’s or the defendant’s perspective. As for the court, the defendant need not be supervised when he is held in custody; as for the defendant, there is nothing unfair about not knowing during pretrial detention whether he is also under supervised release. Pp. 10–12. 723 Fed. Appx. 325, affirmed. Thomas, J., delivered the opinion of the Court, in which Roberts, C. J., and Ginsburg, Alito, and Kavanaugh, JJ., joined. Sotomayor, J., filed a dissenting opinion, in which Breyer, Kagan, and Gorsuch, JJ., joined.

This case requires the Court to decide whether a convicted criminal’s period of supervised release is tolled—in effect, paused—during his pretrial detention for a new criminal offense. Specifically, the question is whether that pretrial detention qualifies as “imprison[ment] in connection with a conviction for a Federal, State, or local crime.” 18 U. S. C. §3624(e). Given the text and statutory context of §3624(e),we conclude that if the court’s later imposed sentence credits the period of pretrial detention as time served for the new offense, then the pretrial detention also tolls the supervised-release period. I A In 2004, petitioner Jason Mont began distributing cocaine and crack cocaine in northern Ohio. After substantial drug sales to a confidential informant and a search of his home that uncovered handguns and $2,700 in cash, a federal grand jury indicted Mont for multiple drug and firearm offenses. He later pleaded guilty to conspiring to possess with intent to distribute cocaine, and to possessing a firearm and ammunition after having been convicted of a felony. See 18 U. S. C. §922(g)(1) (2000 ed.); 21 U. S. C. §§841(a)(1), 846 (2000 ed.). The District Court sentenced Mont to 120 months’ imprisonment, later reduced to 84 months, to be followed by 5 years of supervised release. Mont was released from federal prison on March 6, 2012, and his supervised release was “slated to end on March 6, 2017.” 723 Fed. Appx. 325, 326 (CA6 2018); see 18 U. S. C. §3624(e) (a “term of supervised release commences on the day the person is released from imprisonment”). Among other standard conditions, Mont’s supervised release required that he “not commit another federal, state, or local crime,” “not illegally possess a controlled substance,” and “refrain from any unlawful use of a controlled substance.” Judgment in No. 4:05–cr–00229 (ND Ohio), Doc. 37, p. 111. Mont did not succeed on supervised release. In March 2015, an Ohio grand jury charged him with two counts of marijuana trafficking in a sealed indictment. Mont was arrested and released on bond while awaiting trial for those charges. Things only got worse from there. In October 2015, Mont tested positive for cocaine and oxycodone during a routine drug test conducted as part of his supervised release. But Mont’s probation officer did not immediately report these violations to the District Court; instead, the officer referred him for additional substance-abuse counseling. Mont proceeded to test positive in five more random drug tests over the next few months. He also used an “ ‘unknown’ liquid to try to pass two subsequent drug tests.” 723 Fed. Appx., at 326. In Jan- uary 2016, Mont’s probation officer finally reported the supervised-release violations, including Mont’s use of drugs and attempts to adulterate his urine samples. The violation report also informed the District Court about the pending state charges and the anticipated trial date of March 2016 in state court. The District Court declined to issue an arrest warrant at that time, but it asked to “ ‘be notified of the resolution of the state charges.’ ” Ibid.; see 18 U. S. C. §3606 (explaining that the District Court “may issue a warrant for the arrest” of the releasee for “violation of a condition of release”). On June 1, 2016, approximately four years and three months into his 5-year term of supervised release, Mont was arrested again on new state charges of trafficking in cocaine, and his bond was revoked on the earlier marijuana-trafficking charges. He was incarcerated in the Ma- honing County Jail and has remained in state custody since that date. Mont’s probation officer filed a report with the District Court stating that he had violated the terms of his release based on these new state offenses. The officer later advised the court that because Mont’s incarceration rendered him unavailable for supervision, the Probation Office was “toll[ing]” his federal supervision. App. 21. The officer promised to keep the court apprised of the pending state charges and stated that, if Mont were convicted, the officer would ask the court to take action at that time. In October 2016, Mont entered into plea agreements with state prosecutors in exchange for a predetermined 6-year sentence. The state trial court accepted Mont’s guilty pleas on October 6, 2016, and set the cases for sentencing in December 2016. Three weeks later, Mont filed a written admission in the District Court “acknowledg[ing]” that he had violated his conditions of supervised release “by virtue of his conviction following guilty pleas to certain felony offenses” in state court. Record in No. 4:05–cr–00229 (ND Ohio), Doc. 92, p. 419. Even though he had yet to be sentenced for the state offenses, Mont sought a hearing on the supervised-release violations at the court’s “earliest convenience.” Ibid. The court initially scheduled a hearing for November 9, 2016, but then, over Mont’s objection, rescheduled the hearing several times to allow for “the conclusion of the State sentencing.” App. 8; 723 Fed. Appx., at 327. On March 21, 2017, Mont was sentenced in state court to six years’ imprisonment. The judge “credited the roughly ten months that Mont had already been incarcerated pending a disposition as time served.” Id., at 327. The District Court issued a warrant on March 30, 2017, and ultimately set a supervised-release hearing for June 28, 2017. B Two days before that hearing, Mont challenged the jurisdiction of the District Court based on the fact that his supervised release had initially been set to expire on March 6, 2017. The court concluded that it had authority to supervise Mont, revoked his supervised release, and ordered him to serve an additional 42 months’ imprisonment to run consecutive to his state sentence. The court held that it retained jurisdiction to revoke the release under 18 U. S. C. §3583(i), which preserves, for a “reasonably necessary” period of time, the court’s power to adjudicate violations and revoke a term of supervised release after the term has expired “if, before its expiration, a warrant or summons has been issued on the basis of an allegation of such a violation.” The court further held that it retained authority to revoke Mont’s term of supervised release because it gave “notice by way of a summons” on November 1, 2016, when it originally scheduled the hearing. App. 22. The court also concluded that the delay between the guilty pleas in October 2016 and the hearing date in June 2017 was “reasonably necessary.” Id., at 24. The Sixth Circuit affirmed on alternative grounds. The court could find no evidence in the record that a summons had issued within the meaning of §3583(i). 723 Fed. Appx., at 329, n. 5. But because Circuit precedent pro- vided an alternative basis for affirmance, the court did not further consider the Government’s argument that the District Court retained jurisdiction under §3583(i). Instead, the court held that Mont’s supervised-release period was tolled while he was held in pretrial detention in state custody under §3624(e), which provides: “(e) Supervision After Release.— . . . The term of supervised release commences on the day the person is released from imprisonment and runs concurrently with any Federal, State, or local term of probation or supervised release or parole for another offense to which the person is subject or becomes subject during the term of supervised release. A term of supervised release does not run during any period in which the person is imprisoned in connection with a conviction for a Federal, State, or local crime unless the imprisonment is for a period of less than 30 consecutive days.” (Emphasis added.) Relying on Circuit precedent, the Sixth Circuit explained that when a defendant is convicted of the offense for which he was held in pretrial detention for longer than 30 days and “ ‘his pretrial detention is credited as time served toward his sentence, then the pretrial detention is “in connection with” a conviction and tolls the period of supervised release under §3624.’ ” Id., at 328 (quoting United States v. Goins, 516 F.3d 416, 417 (2008)). Because Mont’s term of supervised release had been tolled between June 2016 and March 2017, there was ample time left on his supervised-release term when the warrant issued on March 30, 2017. The Courts of Appeals disagree on whether §3624(e) tolls supervised release for periods of pretrial detention lasting longer than 30 days when that incarceration is later credited as time served on a conviction. Compare United States v. Ide, 624 F.3d 666, 667 (CA4 2010) (supervised-release period tolls); United States v. Molina-Gazca, 571 F.3d 470, 474 (CA5 2009) (same); United States v. Johnson, 581 F.3d 1310, 1312–1313 (CA11 2009) (same); Goins, supra, at 417 (same), with United States v. Marsh, 829 F.3d 705, 709 (CADC 2016) (supervised-release period does not toll); United States v. Morales-Alejo, 193 F.3d 1102, 1106 (CA9 1999) (same). We granted certiorari to resolve this split of authority. 586 U. S. ___ (2018). II We hold that pretrial detention later credited as time served for a new conviction is “imprison[ment] in connection with a conviction” and thus tolls the supervised-release term under §3624(e). This is so even if the court must make the tolling calculation after learning whether the time will be credited. In our view, this reading is compelled by the text and statutory context of §3624(e). A Section 3624(e) provides for tolling when a person “is imprisoned in connection with a conviction.” This phrase, sensibly read, includes pretrial detention credited toward another sentence for a new conviction. First, the definition of “is imprisoned” may well include pretrial detention. Both now and at the time Congress created supervised release, see §212(a)(2), 98Stat. 1999–2000, the term “imprison” has meant “[t]o put in a prison,” “to incarcerate,” “[t]o confine a person, or restrain his liberty, in any way.” Black’s Law Dictionary 681 (5th ed. 1979); 5 Oxford English Dictionary 113 (1933); accord, Black’s Law Dictionary 875 (10th ed. 2014). These definitions encompass pretrial detention, and, despite the dissent’s reliance on a narrower definition, post, at 5–7 (opinion of Sotomayor, J.), even Mont has not pressed any serious argument to the contrary. As the Sixth Circuit previously recognized, if imprisonment referred only to “confinement that is the result of a penalty or sentence, then the phrase ‘in connection with a conviction’ [would] becom[e] entirely superfluous.” Goins, supra, at 421. Second, the phrase “in connection with a conviction” encompasses a period of pretrial detention for which a defendant receives credit against the sentence ultimately imposed. The Court has often recognized that “in connection with” can bear a “broad interpretation.” Merrill Lynch, Pierce, Fenner & Smith Inc. v. Dabit, 547 U.S. 71, 85 (2006) (interpreting “in connection with the purchase or sale” broadly in the context of §10(b) of the Securities Exchange Act of 1934, 15 U. S. C. §78j(b)); see, e.g., United States v. American Union Transport, Inc., 327 U.S. 437, 443 (1946) (describing the phrase “in connection with” in the Shipping Act, 1916, 39Stat. 728, as “broad and general”). The Court has also recognized that “ ‘ in connection with’ is essentially indeterminate because connections, like relations, stop nowhere.” Maracich v. Spears, 570 U.S. 48, 59 (2013) (quotation altered). Here, however, we need not consider the outer bounds of the term “in connection with,” as pretrial incarceration is directly tied to the conviction when it is credited toward the new sentence. The judgment of the state court stated as much, crediting the pretrial detention that Mont served while awaiting trial and sentencing for his crimes against his ultimate sentence for those same crimes. This reading of “imprison[ment] in connection with a conviction” is buttressed by the fact that Congress, like most States, instructs courts calculating a term of imprisonment to credit pretrial detention as time served on a subsequent conviction. See 18 U. S. C. §3585(b)(1); Tr. of Oral Arg. 54 (statement of the Assistant Solicitor General representing that the same rule applies in 45 States and the District of Columbia). Thus, it makes sense that the phrase “imprison[ment] in connection with a conviction” would include pretrial detention later credited as time served, especially since both provisions were passed as part of the Sentencing Reform Act of 1984. See §212(a)(2), 98Stat. 2008–2009. If Congress intended a narrower interpretation, it could have easily used narrower language, such as “after a conviction” or “following a conviction.” See e.g., Bail Reform Act of 1984, §209(d)(4), 98Stat. 1987 (adding Federal Rule of Criminal Procedure 46(h), allowing courts to direct forfeiture of property “after conviction of the offense charged” (emphasis added)). We cannot override Congress’ choice to employ the more capacious phrase “in connection with.” Third, the text undeniably requires courts to retrospectively calculate whether a period of pretrial detention should toll a period of supervised release. Whereas §3624(e) instructs courts precisely when the supervised-release clock begins—“on the day the person is released”—the statute does not require courts to make a tolling determination as soon as a defendant is arrested on new charges or to continually reassess the tolling calculation throughout the period of his pretrial detention. Congress contemplated the opposite by including a minimum-incarceration threshold: tolling occurs “unless the imprisonment is for a period of less than 30 consecutive days.” §3624(e). This calculation must be made after either release from custody or entry of judgment; there is no way for a court to know on day 5 of a defendant’s pretrial detention whether the period of custody will extend beyond 30 days. Thus, at least some uncertainty as to whether supervised release is tolled is built into §3624(e) by legislative design. This fact confirms that courts should make the tolling calculation upon the defendant’s release from custody or upon entry of judgment. B The statutory context also supports our reading. Supervised release is “a form of postconfinement monitoring” that permits a defendant a kind of conditional liberty by allowing him to serve part of his sentence outside of prison. Johnson v. United States, 529 U.S. 694, 697 (2000). Recognizing that Congress provided for supervised release to facilitate a “transition to community life,” we have declined to offset a term of supervised release by the amount of excess time a defendant spent in prison after two of his convictions were declared invalid. United States v. Johnson, 529 U.S. 53, 59–60 (2000). As we explained: “The objectives of supervised release would be unfulfilled if excess prison time were to offset and reduce terms of supervised release” because “[s]upervised release has no statutory function until confinement ends.” Id., at 59. This understanding of supervised release informs our reading of the tolling provision. Consider §3624(e) itself. The sentence preceding the one at issue here specifies that supervised release “runs concurrently” with “probation or supervised release or parole for another offense.” §3624(e) (emphasis added). But the next sentence (the one at issue here) excludes periods of “imprison[ment]” served “in connection with a conviction.” The juxtaposition of these two sentences reinforces the fact that prison time is “not interchange- able” with supervised release. Id., at 59. Permitting a period of probation or parole to count toward supervised release but excluding a period of incarceration furthers the statutory design of “successful[ly] transition[ing]” a defendant from “prison to liberty.” Johnson, supra, at 708–709. Allowing pretrial detention credited toward another sentence to toll the period of supervised release is consistent with that design. Cf. A. Scalia & B. Garner, Reading Law 167 (2012) (explaining that “the whole-text canon” requires consideration of “the entire text, in view of its structure” and “logical relation of its many parts”). Second, it would be an exceedingly odd construction of the statute to give a defendant the windfall of satisfying a new sentence of imprisonment and an old sentence of supervised release with the same period of pretrial detention. Supervised release is a form of punishment that Congress prescribes along with a term of imprisonment as part of the same sentence. See generally §3583. And Congress denies defendants credit for time served if the detention time has already “been credited against another sentence.” §3585(b). Yet Mont’s reading of §3624(e) would deprive the Government of its lawfully imposed sentence of supervised release while the defendant is serving a separate sentence of incarceration—one often imposed by a different sovereign. Under our view, in contrast, time in pretrial detention constitutes supervised release only if the charges against the defendant are dismissed or the defendant is acquitted. This ensures that the defendant is not faulted for conduct he might not have committed, while otherwise giving full effect to the lawful judgment previously imposed on the defendant.[1] C In response to these points, Mont follows the D. C. Circuit in arguing that the present tense of the statute (“ ‘is imprisoned’ ”) forbids any backward looking tolling analysis. See Marsh, 829 F. 3d, at 709. Mont contends that, when a defendant is held in pretrial detention, a court cannot say at that moment that he “is imprisoned in connection with a conviction.” He relies on the Dictionary Act, which provides that “[i]n determining the meaning of any Act of Congress, unless the context indicates otherwise[,] words used in the present tense include the future as well as the present.” 1 U. S. C. §1. Mont’s argument confuses the rule (“any period in which the person is imprisoned in connection with a conviction”) with a court’s analysis of whether that rule was satisfied. Of course, the determination whether supervised release has been tolled cannot be made at the exact moment when the defendant is held in pretrial detention. Rather, the court must await the outcome of those separate proceedings before it will know whether “imprison[ment]” is tied to a conviction. But the statute does not require the court to make a contemporaneous assessment. Quite the opposite: As discussed, the statute undeniably contemplates that there will be uncertainty about the status of supervised release when a defendant has been held for a short period of time and it is unclear whether the imprisonment will exceed 30 days. There is no reason the statute would preclude postponing calculation just because the custody period extends beyond 30 days. Once the court makes the calculation, it will determine whether the relevant period ultimately qualified as a period “in which the person is imprisoned in connection with a conviction” for 30 or more days. In short, the present-tense phrasing of the statute does not address whether a judge must be able to make a supervised-release determination at any given time. Moreover, any uncertainty about whether supervised release is tolled matters little from either the court’s or the defendant’s perspective. As for the court, the defendant need not be supervised when he is held in custody, so it does not strike us as “odd” to make a delayed determination concerning tolling. Marsh, supra, at 710. The court need not monitor the defendant’s progress in transitioning back into the community because the defendant is not in the community. And if the court is concerned about losing authority over the defendant because of an impending conclusion to supervised release, it can simply issue a summons or warrant under §3583(i) for alleged violations. As for the defendant, there is nothing unfair about not knowing during pretrial detention whether he is also subject to court supervision. The answer to that question cannot meaningfully influence his behavior. A defendant in custody will be unable to comply with many ordinary conditions of supervised release intended to reacclimate him to society—for example, making restitution payments, attending substance-abuse counseling, meeting curfews, or participating in job training. The rules he can “comply” with are generally mandated by virtue of being in prison—for example, no new offenses or use of drugs. See §§3563(a)–(b) (listing mandatory and discretionary conditions). In this case, Mont’s supervised-release conditions required that he “work regularly at a lawful occupation” and “support his . . . dependants and meet other family responsibilities.” Judgment in No. 4:05–cr–00229 (ND Ohio), Doc. 37, at 111. Mont could not fulfill these conditions while sitting in an Ohio jail, and his probation officer correctly deemed him “unavailable for supervision.”[2] App. 21. III Applying §3624(e) to Mont, the pretrial-detention period tolled his supervised release beginning in June 2016. Mont therefore had about nine months remaining on his term of supervised release when the District Court revoked his supervised release and sentenced him to an additional 42 months’ imprisonment. And because §3624(e) independently tolled the supervised-release period, it is immaterial whether the District Court could have issued a summons or warrant under §3583(i) to preserve its authority. * * * In light of the statutory text and context of §3624(e), pretrial detention qualifies as “imprison[ment] in connection with a conviction” if a later imposed sentence credits that detention as time served for the new offense. Such pretrial detention tolls the supervised-release period, even though the District Court may need to make the tolling determination after the conviction. Accordingly, we affirm the judgment of the Sixth Circuit. It is so ordered. Notes 1 Our reading leaves intact a district court’s ability to preserve its authority by issuing an arrest warrant or summons under §3583(i) based on the conduct at issue in the new charges, irrespective of whether the defendant is later convicted or acquitted of those offenses. But preserving jurisdiction through §3583(i) is not a prerequisite to a court maintaining authority under §3624(e), nor does it impact the tolling calculation under §3624(e). 2 Although a defendant in pretrial detention is unable to be supervised, it does not necessarily follow that the defendant will be punished by his inability to comply with the terms of his supervised release if the detention period is not later credited as time served for a conviction. In that circumstance, the district court may always modify the terms of his supervision. See 18 U. S. C. §3583(e)(2). And, as the Government explained at oral argument, modification of supervised release may not be necessary to the extent that “the defendant can’t be deemed to have been required to” comply with the terms of supervised release while in custody. Tr. of Oral Arg. 45.

586.US.2018_17-587

John Guido and Dennis Rankin filed suit, alleging that the Mount Lemmon Fire District, a political subdivision in Arizona, terminated their employment as firefighters in violation of the Age Discrimination in Employment Act of 1967 (ADEA). The Fire District responded that it was too small to qualify as an “employer” under the ADEA, which provides: “The term ‘employer’ means a person engaged in an industry affecting commerce who has twenty or more employees . . . . The term also means (1) any agent of such a person, and (2) a State or political subdivision of a State . . . .” 29 U. S. C. §630(b). Initially, both Title VII of the Civil Rights Act of 1964 and the ADEA applied solely to private sector employers. In 1974, Congress amended the ADEA to cover state and local governments. A previous, 1972, amendment to Title VII added States and their subdivisions to the definition of “person[s],” specifying that those entities are engaged in an industry affecting commerce. The Title VII amendment thus subjected States and their subdivisions to liability only if they employ a threshold number of workers, currently 15. By contrast, the 1974 ADEA amendment added state and local governments directly to the definition of “employer.” The same 1974 enactment also amended the Fair Labor Standards Act (FLSA), on which many aspects of the ADEA are based, to reach all government employers regardless of their size. 29 U. S. C. §203(d), (x). Held: The definitional provision’s two-sentence delineation, set out in §630(b), and the expression “also means” at the start of §630(b)’s second sentence, combine to establish separate categories: persons engaged in an industry affecting commerce with 20 or more employees; and States or political subdivisions with no attendant numerosity limitation. The words “also means” in §630(b) add new categories of employers to the ADEA’s reach. First and foremost, the ordinary meaning of “also means” is additive rather than clarifying. See 859 F. 3d 1168, 1171 (case below) (quoting Webster’s New Collegiate Dictionary 34). The words “also means” occur dozens of times throughout the U. S. Code, typically carrying an additive meaning. E.g., 12 U. S. C. §1715z–1(i)(4). Furthermore, the second sentence of the ADEA’s definitional provision, §630(b), pairs States and their political subdivisions with agents, a discrete category that carries no numerical limitation. Reading the ADEA’s definitional provision, §630(b), as written to apply to States and political subdivisions regardless of size may give the ADEA a broader reach than Title VII, but this disparity is a consequence of the different language Congress chose to employ. The better comparator for the ADEA is the FLSA, which also ranks States and political subdivisions as employers regardless of the number of employees they have. The Equal Employment Opportunity Commission has, for 30 years, interpreted the ADEA to cover political subdivisions regardless of size, and a majority of the States impose age discrimination proscriptions on political subdivisions with no numerical threshold. Pp. 4–6. 859 F. 3d 1168, affirmed. Ginsburg, J., delivered the opinion of the Court, in which all other Members joined, except Kavanaugh, J., who took no part in the consideration or decision of the case.

Faced with a budget shortfall, Mount Lemmon Fire District, a political subdivision in Arizona, laid off its two oldest full-time firefighters, John Guido (then 46) and Dennis Rankin (then 54). Guido and Rankin sued the Fire District, alleging that their termination violated the Age Discrimination in Employment Act of 1967 (ADEA), 81Stat. 602, as amended, 29 U. S. C. §621 et seq. The Fire District sought dismissal of the suit on the ground that the District was too small to qualify as an “employer” within the ADEA’s compass. The Act’s controlling definitional provision, 29 U. S. C. §630(b), reads: “The term ‘employer’ means a person engaged in an industry affecting commerce who has twenty or more employees . . . . The term also means (1) any agent of such a person, and (2) a State or political subdivision of a State . . . .” The question presented: Does the ADEA’s numerosity specification (20 or more employees), applicable to “a person engaged in an industry affecting commerce,” apply as well to state entities (including state political subdivisions)? We hold, in accord with the United States Court of Appeals for the Ninth Circuit, that §630(b)’s two-sentence delineation, and the expression “also means” at the start of the second sentence, combine to establish separate categories: persons engaged in an industry affecting commerce with 20 or more employees; and States or political subdivisions with no attendant numerosity limitation. “[T]wenty or more employees” is confining language, but the confinement is tied to §630(b)’s first sentence, and does not limit the ADEA’s governance of the employment practices of States and political subdivisions thereof. I Initially, Title VII of the Civil Rights Act of 1964, 78Stat. 253, as amended, 42 U. S. C. §2000e et seq., which prohibits employment discrimination on the basis of race, color, religion, sex, and national origin, applied solely to private sector employers. The same was true of the ADEA, enacted three years later to protect workers against “arbitrary age discrimination.” 29 U. S. C. §621(b). As originally enacted, both Title VII and the ADEA imposed liability on “employer[s],” defined in both statutes to include “a person engaged in an industry affecting commerce” whose employees met a numerical threshold, but specifically to exclude governmental entities. 78Stat. 253 (Title VII); 81Stat. 605 (ADEA). In 1972, Congress amended Title VII to reach state and local employers. Under the revised provision of Title VII, “[t]he term ‘person’ includes one or more individuals, governments, governmental agencies, [and] political subdivisions,” also certain other specified entities, and “[t]he term ‘employer’ means a person engaged in an industry affecting commerce who has fifteen or more employees . . . .” 42 U. S. C. §2000e(a)–(b). For this purpose, amended Title VII defines “industry affecting commerce” to “includ[e] any governmental industry, business, or activity.” §2000e(h). The 1972 amendment to Title VII thereby extended the statute’s coverage to state and local government entities by defining them as “person[s].” In turn, as “person[s],” these entities meet Title VII’s definition of “employer” and are subject to liability only if they have at least 15 employees.[1] Two years later, in 1974, Congress amended the ADEA to cover state and local governments. Unlike in Title VII, where Congress added such entities to the definition of “person,” in the ADEA, Congress added them directly to the definition of “employer.” Thus, since 1974, the ADEA’s key definitional provision has read: “The term ‘employer’ means a person engaged in an industry affecting commerce who has twenty or more employees . . . . The term also means (1) any agent of such a person, and (2) a State or political subdivision of a State . . . .” 29 U. S. C. §630(b). In the same 1974 enactment, Congress amended the Fair Labor Standards Act (FLSA), on which parts of the ADEA had been modeled, to reach all government employers regardless of their size. See 88Stat. 58, 29 U. S. C. §203(d), (x). The parties dispute the proper reading of the ADEA following the 1974 amendment. Does “also means” add new categories to the definition of “employer,” or does it merely clarify that States and their political subdivisions are a type of “person” included in §630(b)’s first sentence? If the former, state and local governments are covered by the ADEA regardless of whether they have as many as 20 employees. If the latter, they are covered only if they have at least 20 employees. Federal courts have divided on this question. Compare Kelly v. Wauconda Park Dist., 801 F. 2d 269 (CA7 1986) (state and local governments are covered by the ADEA only if they have at least 20 employees); Cink v. Grant County, 635 Fed. Appx. 470 (CA10 2015) (same); Palmer v. Arkansas Council on Economic Educ., 154 F. 3d 892 (CA8 1998) (same); EEOC v. Monclova, 920 F. 2d 360 (CA6 1990) (same), with this case, 859 F. 3d 1168 (CA9 2017) (state and local governments are covered by the ADEA regardless of their number of employees). We granted certiorari to resolve the conflict. 583 U. S. __ (2018). II For several reasons, we conclude that the words “also means” in §630(b) add new categories of employers to the ADEA’s reach. First and foremost, the ordinary meaning of “also means” is additive rather than clarifying. As the Ninth Circuit explained, “ ‘also’ is a term of enhancement; it means ‘in addition; besides’ and ‘likewise; too.’ ” 859 F. 3d, at 1171 (quoting Webster’s New Collegiate Dictionary 34 (1973)). Indeed, reading “also” additively to create a separate category of “employer” seemed to this Court altogether fitting in EEOC v. Wyoming, 460 U. S. 226 (1983). There, we held that applying the ADEA to state and local governments does not encroach on States’ sovereignty or Tenth Amendment immunity. Id., at 240–242. In the course of so holding, we described the 1974 ADEA amendment as “extend[ing] the substantive prohibitions of the Act to employers having at least 20 workers [as opposed to 25 in the original version], and to the Federal and State Governments.” Id., at 233 (emphasis added). In this regard, we note, it is undisputed that the ADEA covers Federal Government entities, which our opinion in Wyoming grouped with state entities, regardless of the number of workers they employ. 29 U. S. C. §633a. Instructive as well, the phrase “also means” occurs dozens of times throughout the U. S. Code, typically carrying an additive meaning. See Brief for Respondents 11–13, and n. 2 (collecting citations). For example, 12 U. S. C. §1715z–1(i)(4), provides: “[T]he term ‘elderly families’ means families which consist of two or more persons the head of which (or his spouse) is sixty-two years of age or over or is handicapped. Such term also means a single person who is sixty-two years of age or over or is handicapped.” “[A] single person” plainly adds to, rather than clarifies, the preceding statutory delineation, “two or more persons.” Just so with States and their political subdivisions in the ADEA’s definition of “employer.” Notably, in §1715z–1(i)(4), Congress repeated the “sixty-two years of age or over or is handicapped” qualifier to render it applicable to “a single person.” In the ADEA, by contrast, Congress did not repeat the “twenty or more employees” qualifier when referencing state and local government entities. This Court is not at liberty to insert the absent qualifier. Furthermore, the text of §630(b) pairs States and their political subdivisions with agents, a discrete category that, beyond doubt, carries no numerical limitation. See Tr. of Oral Arg. 55–56. The Fire District does not gainsay that the 20-employee restriction applies to §630(b)’s first sentence. Its construction, however, would lift that restriction for the agent portion of the second sentence, and then reimpose it for the portion of that sentence addressing States and their political subdivisions. We resist a reading so strange.[2] The Fire District presses the argument that the ADEA should be interpreted in line with Title VII, which, as noted supra, at 3, applies to state and local governments only if they meet a numerosity specification. True, reading the ADEA as written to apply to States and political subdivisions regardless of size gives the ADEA, in this regard, a broader reach than Title VII. But this disparity is a consequence of the different language Congress chose to employ. See Gross v. FBL Financial Services, Inc., 557 U. S. 167, 174 (2009) (differences between Title VII’s and the ADEA’s language should not be ignored). The better comparator is the FLSA, on which many aspects of the ADEA are based. See 29 U. S. C. §626(b) (ADEA incorporates the “powers, remedies, and procedures” of the FLSA). Like the FLSA, the ADEA ranks States and political subdivisions as “employer[s]” regardless of the number of employees they have. The Fire District warns that applying the ADEA to small public entities risks curtailment of vital public services such as fire protection. Experience suggests otherwise. For 30 years, the Equal Employment Opportunity Commission has consistently interpreted the ADEA as we do today. EEOC Compliance Manual: Threshold Issues §2–III(B)(1)(a)(i), and n. 99. See also Kelly, 801 F. 2d, at 270, n. 1. And a majority of States forbid age discrimination by political subdivisions of any size; some 15 of these States subject private sector employers to age discrimination proscriptions only if they employ at least a threshold number of workers. See Brief for Respondents 28–29, and n. 6 (collecting citations). No untoward service shrinkages have been documented. In short, the text of the ADEA’s definitional provision, also its kinship to the FLSA and differences from Title VII, leave scant room for doubt that state and local governments are “employer[s]” covered by the ADEA regardless of their size. * * * For the reasons stated, the judgment of the Court of Appeals for the Ninth Circuit is Affirmed. Justice Kavanaugh took no part in the consideration or decision of this case. Notes 1 The Americans with Disabilities Act of 1990 defines “employer” in materially the same way as Title VII and accords “person . . . the same meaning” as in Title VII. 42 U. S. C. §12111(5), (7). 2 We need not linger over possible applications of the agent clause, for no question of agent liability is before us in this case.

586.US.2018_17-340

Petitioner New Prime Inc. is an interstate trucking company, and respondent Dominic Oliveira is one of its drivers. Mr. Oliveira works under an operating agreement that calls him an independent contractor and contains a mandatory arbitration provision. When Mr. Oliveira filed a class action alleging that New Prime denies its drivers lawful wages, New Prime asked the court to invoke its statutory authority under the Federal Arbitration Act to compel arbitration. Mr. Oliveira countered that the court lacked authority because §1 of the Act excepts from coverage disputes involving “contracts of employment” of certain transportation workers. New Prime insisted that any question regarding §1’s applicability belonged to the arbitrator alone to resolve, or, assuming the court could address the question, that “contracts of employment” referred only to contracts that establish an employer-employee relationship and not to contracts with independent contractors. The District Court and First Circuit agreed with Mr. Oliveira. Held: 1. A court should determine whether a §1 exclusion applies before ordering arbitration. A court’s authority to compel arbitration under the Act does not extend to all private contracts, no matter how emphatically they may express a preference for arbitration. Instead, antecedent statutory provisions limit the scope of a court’s §§3 and 4 powers to stay litigation and compel arbitration “accord[ing to] the terms” of the parties’ agreement. Section 2 provides that the Act applies only when the agreement is set forth as “a written provision in any maritime transaction or a contract evidencing a transaction involving commerce.” And §1 helps define §2’s terms, warning, as relevant here, that “nothing” in the Act “shall apply” to “contracts of employment of seamen, railroad employees, or any other class of workers engaged in foreign or interstate commerce.” For a court to invoke its statutory authority under §§3 and 4, it must first know if the parties’ agreement is excluded from the Act’s coverage by the terms of §§1 and 2. This sequencing is significant. See, e.g., Bernhardt v. Polygraphic Co. of America, 350 U.S. 198, 201–202. New Prime notes that the parties’ contract contains a “delegation clause,” giving the arbitrator authority to decide threshold questions of arbitrability, and that the “severability principle” requires that both sides take all their disputes to arbitration. But a delegation clause is merely a specialized type of arbitration agreement and is enforceable under §§3 and 4 only if it appears in a contract consistent with §2 that does not trigger §1’s exception. And, the Act’s severability principle applies only if the parties’ arbitration agreement appears in a contract that falls within the field §§1 and 2 describe. Pp. 3–6. 2. Because the Act’s term “contract of employment” refers to any agreement to perform work, Mr. Oliveira’s agreement with New Prime falls within §1’s exception. Pp. 6–15. (a) “[I]t’s a ‘fundamental canon of statutory construction’ that words generally should be ‘interpreted as taking their ordinary . . . meaning . . . at the time Congress enacted the statute.’ ” Wisconsin Central Ltd. v. United States, 585 U. S. ___, ___ (quoting Perrin v. United States, 444 U.S. 37, 42). After all, if judges could freely invest old statutory terms with new meanings, this Court would risk amending legislation outside the “single, finely wrought and exhaustively considered, procedure” the Constitution commands. INS v. Chadha, 462 U.S. 919, 951. The Court would risk, too, upsetting reliance interests by subjecting people today to different rules than they enjoyed when the statute was passed. At the time of the Act’s adoption in 1925, the phrase “contract of employment” was not a term of art, and dictionaries tended to treat “employment” more or less as a synonym for “work.” Contemporaneous legal authorities provide no evidence that a “contract of employment” necessarily signaled a formal employer-employee relationship. Evidence that Congress used the term “contracts of employment” broadly can be found in its choice of the neighboring term “workers,” a term that easily embraces independent contractors. Pp. 6–10. (b) New Prime argues that by 1925, the words “employee” and “independent contractor” had already assumed distinct meanings. But while the words “employee” and “employment” may share a common root and intertwined history, they also developed at different times and in at least some different ways. The evidence remains that, as dominantly understood in 1925, a “contract of employment” did not necessarily imply the existence of an employer-employee relationship. New Prime’s argument that early 20th-century courts sometimes used the phrase “contracts of employment” to describe what are recognized today as agreements between employers and employees does nothing to negate the possibility that the term also embraced agreements by independent contractors to perform work. And its effort to explain away the statute’s suggestive use of the term “worker” by noting that the neighboring terms “seamen” and “railroad employees” included only employees in 1925 rests on a precarious premise. The evidence suggests that even “seamen” and “railroad employees” could be independent contractors at the time the Arbitration Act passed. Left to appeal to the Act’s policy, New Prime suggests that this Court order arbitration to abide Congress’ effort to counteract judicial hostility to arbitration and establish a favorable federal policy toward arbitration agreements. Courts, however, are not free to pave over bumpy statutory texts in the name of more expeditiously advancing a policy goal. Rather, the Court should respect “the limits up to which Congress was prepared” to go when adopting the Arbitration Act. United States v. Sisson, 399 U.S. 267, 298. This Court also declines to address New Prime’s suggestion that it order arbitration anyway under its inherent authority to stay litigation in favor of an alternative dispute resolution mechanism of the parties’ choosing. Pp. 10–15. 857 F.3d 7, affirmed. Gorsuch, J., delivered the opinion of the Court, in which all other Members joined, except Kavanaugh, J., who took no part in the consideration or decision of the case. Ginsburg, J., filed a concurring opinion.

The Federal Arbitration Act requires courts to enforce private arbitration agreements. But like most laws, this one bears its qualifications. Among other things, §1 says that “nothing herein” may be used to compel arbitration in disputes involving the “contracts of employment” of certain transportation workers. 9 U. S. C. §1. And that qualification has sparked these questions: When a contract delegates questions of arbitrability to an arbitrator, must a court leave disputes over the application of §1’s exception for the arbitrator to resolve? And does the term “contracts of employment” refer only to contracts between employers and employees, or does it also reach contracts with independent contractors? Because courts across the country have disagreed on the answers to these questions, we took this case to resolve them. I New Prime is an interstate trucking company and Dominic Oliveira works as one of its drivers. But, at least on paper, Mr. Oliveira isn’t an employee; the parties’ contracts label him an independent contractor. Those agreements also instruct that any disputes arising out of the parties’ relationship should be resolved by an arbitrator—even disputes over the scope of the arbitrator’s authority. Eventually, of course, a dispute did arise. In a class action lawsuit in federal court, Mr. Oliveira argued that New Prime denies its drivers lawful wages. The company may call its drivers independent contractors. But, Mr. Oliveira alleged, in reality New Prime treats them as employees and fails to pay the statutorily due minimum wage. In response to Mr. Oliveira’s complaint, New Prime asked the court to invoke its statutory authority under the Act and compel arbitration according to the terms found in the parties’ agreements. That request led to more than a little litigation of its own. Even when the parties’ contracts mandate arbitration, Mr. Oliveira observed, the Act doesn’t always authorize a court to enter an order compelling it. In particular, §1 carves out from the Act’s coverage “contracts of employment of . . . workers engaged in foreign or interstate commerce.” And at least for purposes of this collateral dispute, Mr. Oliveira submitted, it doesn’t matter whether you view him as an employee or independent contractor. Either way, his agreement to drive trucks for New Prime qualifies as a “contract[ ] of employment of . . . [a] worker[ ] engaged in . . . interstate commerce.” Accordingly, Mr. Oliveira argued, the Act supplied the district court with no authority to compel arbitration in this case. Naturally, New Prime disagreed. Given the extraordinary breadth of the parties’ arbitration agreement, the company insisted that any question about §1’s application belonged for the arbitrator alone to resolve. Alternatively and assuming a court could address the question, New Prime contended that the term “contracts of employment” refers only to contracts that establish an employer-employee relationship. And because Mr. Oliveira is, in fact as well as form, an independent contractor, the company argued, §1’s exception doesn’t apply; the rest of the statute does; and the district court was (once again) required to order arbitration. Ultimately, the district court and the First Circuit sided with Mr. Oliveira. 857 F.3d 7 (2017). The court of appeals held, first, that in disputes like this a court should resolve whether the parties’ contract falls within the Act’s ambit or §1’s exclusion before invoking the statute’s authority to order arbitration. Second, the court of appeals held that §1’s exclusion of certain “contracts of employment” removes from the Act’s coverage not only employer-employee contracts but also contracts involving independent contractors. So under any account of the parties’ agreement in this case, the court held, it lacked authority under the Act to order arbitration. II In approaching the first question for ourselves, one thing becomes clear immediately. While a court’s authority under the Arbitration Act to compel arbitration may be considerable, it isn’t unconditional. If two parties agree to arbitrate future disputes between them and one side later seeks to evade the deal, §§3 and 4 of the Act often require a court to stay litigation and compel arbitration “accord[ing to] the terms” of the parties’ agreement. But this authority doesn’t extend to all private contracts, no matter how emphatically they may express a preference for arbitration. Instead, antecedent statutory provisions limit the scope of the court’s powers under §§3 and 4. Section 2 provides that the Act applies only when the parties’ agreement to arbitrate is set forth as a “written provision in any maritime transaction or a contract evidencing a transaction involving commerce.” And §1 helps define §2’s terms. Most relevant for our purposes, §1 warns that “nothing” in the Act “shall apply” to “contracts of employment of seamen, railroad employees, or any other class of workers engaged in foreign or interstate commerce.” Why this very particular qualification? By the time it adopted the Arbitration Act in 1925, Congress had already prescribed alternative employment dispute resolution regimes for many transportation workers. And it seems Congress “did not wish to unsettle” those arrangements in favor of whatever arbitration procedures the parties’ private contracts might happen to contemplate. Circuit City Stores, Inc. v. Adams, 532 U.S. 105, 121 (2001). Given the statute’s terms and sequencing, we agree with the First Circuit that a court should decide for itself whether §1’s “contracts of employment” exclusion applies before ordering arbitration. After all, to invoke its statutory powers under §§3 and 4 to stay litigation and compel arbitration according to a contract’s terms, a court must first know whether the contract itself falls within or beyond the boundaries of §§1 and 2. The parties’ private agreement may be crystal clear and require arbitration of every question under the sun, but that does not necessarily mean the Act authorizes a court to stay litigation and send the parties to an arbitral forum. Nothing in our holding on this score should come as a surprise. We’ve long stressed the significance of the statute’s sequencing. In Bernhardt v. Polygraphic Co. of America, 350 U.S. 198, 201–202 (1956), we recognized that “Sections 1, 2, and 3 [and 4] are integral parts of a whole. . . . [Sections] 1 and 2 define the field in which Congress was legislating,” and §§3 and 4 apply only to contracts covered by those provisions. In Circuit City, we acknowledged that “Section 1 exempts from the [Act] . . . contracts of employment of transportation workers.” 532 U. S., at 119. And in Southland Corp. v. Keating, 465 U.S. 1, 10–11, and n. 5 (1984), we noted that “the enforceability of arbitration provisions” under §§3 and 4 depends on whether those provisions are “part of a written maritime contract or a contract ‘evidencing a transaction involving commerce’ ” under §2—which, in turn, depends on the application of §1’s exception for certain “contracts of employment.” To be sure, New Prime resists this straightforward understanding. The company argues that an arbitrator should resolve any dispute over §1’s application because of the “delegation clause” in the parties’ contract and what is sometimes called the “severability principle.” A delegation clause gives an arbitrator authority to decide even the initial question whether the parties’ dispute is subject to arbitration. Rent-A-Center, West, Inc. v. Jackson, 561 U.S. 63, 68–69 (2010). And under the severability principle, we treat a challenge to the validity of an arbitration agreement (or a delegation clause) separately from a challenge to the validity of the entire contract in which it appears. Id., at 70–71. Unless a party specifically challenges the validity of the agreement to arbitrate, both sides may be required to take all their disputes—including disputes about the validity of their broader contract—to arbitration. Ibid. Applying these principles to this case, New Prime notes that Mr. Oliveira has not specifically challenged the parties’ delegation clause and submits that any controversy should therefore proceed only and immediately before an arbitrator. But all this overlooks the necessarily antecedent statutory inquiry we’ve just discussed. A delegation clause is merely a specialized type of arbitration agreement, and the Act “operates on this additional arbitration agreement just as it does on any other.” Id., at 70. So a court may use §§3 and 4 to enforce a delegation clause only if the clause appears in a “written provision in . . . a contract evidencing a transaction involving commerce” consistent with §2. And only if the contract in which the clause appears doesn’t trigger §1’s “contracts of employment” exception. In exactly the same way, the Act’s severability principle applies only if the parties’ arbitration agreement appears in a contract that falls within the field §§1 and 2 describe. We acknowledged as much some time ago, explaining that, before invoking the severability principle, a court should “determine[ ] that the contract in question is within the coverage of the Arbitration Act.” Prima Paint Corp. v. Flood & Conklin Mfg. Co., 388 U.S. 395, 402 (1967). III That takes us to the second question: Did the First Circuit correctly resolve the merits of the §1 challenge in this case? Recall that §1 excludes from the Act’s compass “contracts of employment of . . . workers engaged in . . . interstate commerce.” Happily, everyone before us agrees that Mr. Oliveira qualifies as a “worker[ ] engaged in . . . interstate commerce.” For purposes of this appeal, too, Mr. Oliveira is willing to assume (but not grant) that his contracts with New Prime establish only an independent contractor relationship. With that, the disputed question comes into clear view: What does the term “contracts of employment” mean? If it refers only to contracts that reflect an employer-employee relationship, then §1’s exception is irrelevant and a court is free to order arbitration, just as New Prime urges. But if the term also encompasses contracts that require an independent contractor to perform work, then the exception takes hold and a court lacks authority under the Act to order arbitration, exactly as Mr. Oliveira argues. A In taking up this question, we bear an important caution in mind. “[I]t’s a ‘fundamental canon of statutory construction’ that words generally should be ‘interpreted as taking their ordinary . . . meaning . . . at the time Congress enacted the statute.’ ” Wisconsin Central Ltd. v. United States, 585 U. S. ___, ___ (2018) (slip op., at 9) (quoting Perrin v. United States, 444 U.S. 37, 42 (1979)). See also Sandifer v. United States Steel Corp., 571 U.S. 220, 227 (2014). After all, if judges could freely invest old statutory terms with new meanings, we would risk amending legislation outside the “single, finely wrought and exhaustively considered, procedure” the Constitution commands. INS v. Chadha, 462 U.S. 919, 951 (1983). We would risk, too, upsetting reliance interests in the settled meaning of a statute. Cf. 2B N. Singer & J. Singer, Sutherland on Statutes and Statutory Construction §56A:3 (rev. 7th ed. 2012). Of course, statutes may sometimes refer to an external source of law and fairly warn readers that they must abide that external source of law, later amendments and modifications included. Id., §51:8 (discussing the reference canon). But nothing like that exists here. Nor has anyone suggested any other appropriate reason that might allow us to depart from the original meaning of the statute at hand. That, we think, holds the key to the case. To many lawyerly ears today, the term “contracts of employment” might call to mind only agreements between employers and employees (or what the common law sometimes called masters and servants). Suggestively, at least one recently published law dictionary defines the word “employment” to mean “the relationship between master and servant.” Black’s Law Dictionary 641 (10th ed. 2014). But this modern intuition isn’t easily squared with evidence of the term’s meaning at the time of the Act’s adoption in 1925. At that time, a “contract of employment” usually meant nothing more than an agreement to perform work. As a result, most people then would have understood §1 to exclude not only agreements between employers and employees but also agreements that require independent contractors to perform work. What’s the evidence to support this conclusion? It turns out that in 1925 the term “contract of employment” wasn’t defined in any of the (many) popular or legal dictionaries the parties cite to us. And surely that’s a first hint the phrase wasn’t then a term of art bearing some specialized meaning. It turns out, too, that the dictionaries of the era consistently afforded the word “employment” a broad construction, broader than may be often found in dictionaries today. Back then, dictionaries tended to treat “employment” more or less as a synonym for “work.” Nor did they distinguish between different kinds of work or workers: All work was treated as employment, whether or not the common law criteria for a master-servant relationship happened to be satisfied.[1] What the dictionaries suggest, legal authorities confirm. This Court’s early 20th-century cases used the phrase “contract of employment” to describe work agreements involving independent contractors.[2] Many state court cases did the same.[3] So did a variety of federal statutes.[4] And state statutes too.[5] We see here no evidence that a “contract of employment” necessarily signaled a formal employer-employee or master-servant relationship. More confirmation yet comes from a neighboring term in the statutory text. Recall that the Act excludes from its coverage “contracts of employment of . . . any . . . class of workers engaged in foreign or interstate commerce.” 9 U. S. C. §1 (emphasis added). Notice Congress didn’t use the word “employees” or “servants,” the natural choices if the term “contracts of employment” addressed them alone. Instead, Congress spoke of “workers,” a term that everyone agrees easily embraces independent contractors. That word choice may not mean everything, but it does supply further evidence still that Congress used the term “contracts of employment” in a broad sense to capture any contract for the performance of work by workers. B What does New Prime have to say about the case building against it? Mainly, it seeks to shift the debate from the term “contracts of employment” to the word “employee.” Today, the company emphasizes, the law often distinguishes between employees and independent contractors. Employees are generally understood as those who work “in the service of another person (the employer) under an express or implied contract of hire, under which the employer has the right to control the details of work perform- ance.” Black’s Law Dictionary, at 639. Meanwhile, independent contractors are sometimes described as those “entrusted to undertake a specific project but who [are] left free to do the assigned work and to choose the method for accomplishing it.” Id., at 888. New Prime argues that, by 1925, the words “employee” and “independent contractor” had already assumed these distinct meanings.[6] And given that, the company contends, the phrase “contracts of employment” should be understood to refer only to relationships between employers and employees. Unsurprisingly, Mr. Oliveira disagrees. He replies that, while the term “employment” dates back many centuries, the word “employee” only made its first appearance in English in the 1800s. See Oxford English Dictionary (3d ed., Mar. 2014), www.oed.com/view/Entry/61374 (all Internet materials as last visited Jan. 9, 2019). At that time, the word from which it derived, “employ,” simply meant to “apply (a thing) to some definite purpose.” 3 J. Murray, A New English Dictionary on Historical Principles 129 (1891). And even in 1910, Black’s Law Dictionary reported that the term “employee” had only “become somewhat naturalized in our language.” Black’s Law Dictionary 421 (2d ed. 1910). Still, the parties do share some common ground. They agree that the word “employee” eventually came into wide circulation and came to denote those who work for a wage at the direction of another. They agree, too, that all this came to pass in part because the word “employee” didn’t suffer from the same “historical baggage” of the older common law term “servant,” and because it proved useful when drafting legislation to regulate burgeoning industries and their labor forces in the early 20th century.[7] The parties even agree that the development of the term “employee” may have come to influence and narrow our understanding of the word “employment” in comparatively recent years and may be why today it might signify to some a “relationship between master and servant.”[8] But if the parties’ extended etymological debate persuades us of anything, it is that care is called for. The words “employee” and “employment” may share a common root and an intertwined history. But they also developed at different times and in at least some different ways. The only question in this case concerns the meaning of the term “contracts of employment” in 1925. And, whatever the word “employee” may have meant at that time, and however it may have later influenced the meaning of “employment,” the evidence before us remains that, as dominantly understood in 1925, a contract of employment did not necessarily imply the existence of an employer-employee or master-servant relationship. When New Prime finally turns its attention to the term in dispute, it directs us to Coppage v. Kansas, 236 U.S. 1, 13 (1915). There and in other cases like it, New Prime notes, courts sometimes used the phrase “contracts of employment” to describe what today we’d recognize as agreements between employers and employees. But this proves little. No one doubts that employer-employee agreements to perform work qualified as “contracts of employment” in 1925—and documenting that fact does nothing to negate the possibility that “contracts of employment” also embraced agreements by independent contractors to perform work. Coming a bit closer to the mark, New Prime eventually cites a handful of early 20th-century legal materials that seem to use the term “contracts of employment” to refer exclusively to employer-employee agreements.[9] But from the record amassed before us, these authorities appear to represent at most the vanguard, not the main body, of contemporaneous usage. New Prime’s effort to explain away the statute’s suggestive use of the term “worker” proves no more compelling. The company reminds us that the statute excludes “contracts of employment” for “seamen” and “railroad employees” as well as other transportation workers. And because “seamen” and “railroad employees” included only employees in 1925, the company reasons, we should understand “any other class of workers engaged in . . . interstate commerce” to bear a similar construction. But this argument rests on a precarious premise. At the time of the Act’s passage, shipboard surgeons who tended injured sailors were considered “seamen” though they likely served in an independent contractor capacity.[10] Even the term “railroad employees” may have swept more broadly at the time of the Act’s passage than might seem obvious today. In 1922, for example, the Railroad Labor Board interpreted the word “employee” in the Transportation Act of 1920 to refer to anyone “engaged in the customary work directly contributory to the operation of the railroads.”[11] And the Erdman Act, a statute enacted to address disruptive railroad strikes at the end of the 19th century, seems to evince an equally broad understanding of “railroad employees.”[12] Unable to squeeze more from the statute’s text, New Prime is left to appeal to its policy. This Court has said that Congress adopted the Arbitration Act in an effort to counteract judicial hostility to arbitration and establish “a liberal federal policy favoring arbitration agreements.” Moses H. Cone Memorial Hospital v. Mercury Constr. Corp., 460 U.S. 1, 24 (1983). To abide that policy, New Prime suggests, we must order arbitration according to the terms of the parties’ agreement. But often and by design it is “hard-fought compromise[ ],” not cold logic, that supplies the solvent needed for a bill to survive the legislative process. Board of Governors, FRS v. Dimension Financial Corp., 474 U.S. 361, 374 (1986). If courts felt free to pave over bumpy statutory texts in the name of more expeditiously advancing a policy goal, we would risk failing to “tak[e] . . . account of” legislative compromises essential to a law’s passage and, in that way, thwart rather than honor “the effectuation of congressional intent.” Ibid. By respecting the qualifications of §1 today, we “respect the limits up to which Congress was prepared” to go when adopting the Arbitration Act. United States v. Sisson, 399 U.S. 267, 298 (1970). Finally, and stretching in a different direction entirely, New Prime invites us to look beyond the Act. Even if the statute doesn’t supply judges with the power to compel arbitration in this case, the company says we should order it anyway because courts always enjoy the inherent authority to stay litigation in favor of an alternative dispute resolution mechanism of the parties’ choosing. That, though, is an argument we decline to tangle with. The courts below did not address it and we granted certiorari only to resolve existing confusion about the application of the Arbitration Act, not to explore other potential avenues for reaching a destination it does not. * When Congress enacted the Arbitration Act in 1925, the term “contracts of employment” referred to agreements to perform work. No less than those who came before him, Mr. Oliveira is entitled to the benefit of that same understanding today. Accordingly, his agreement with New Prime falls within §1’s exception, the court of appeals was correct that it lacked authority under the Act to order arbitration, and the judgment is Affirmed. Justice Kavanaugh took no part in the consideration or decision of this case. Notes 1 See, e.g., 3 J. Murray, A New English Dictionary on Historical Principles 130 (1891) (defining “employment” as, among other things, “[t]he action or process of employing; the state of being employed. The service (of a person). That on which (one) is employed; business; occupation; a special errand or commission. A person’s regular occupation or business; a trade or profession”); 3 The Century Dictionary and Cyclopedia 1904 (1914) (defining “employment” as “[w]ork or business of any kind”); W. Harris, Webster’s New International Dictionary 718 (1st ed. 1909) (listing “work” as a synonym for “employment”); Webster’s Collegiate Dictionary 329 (3d ed. 1916) (same); Black’s Law Dictionary 422 (2d ed. 1910) (“an engagement or rendering services” for oneself or another); 3 Oxford English Dictionary 130 (1933) (“[t]hat on which (one) is employed; business; occupation; a special errand or commission”). 2 See, e.g., Watkins v. Sedberry, 261 U.S. 571, 575 (1923) (agreement between trustee and attorney to recover bankrupt’s property); Owen v. Dudley & Michener, 217 U.S. 488, 494 (1910) (agreement between Indian tribe and attorneys to pursue claims). 3 See, e.g., Lindsay v. McCaslin (Two Cases), 123 Me. 197, 200, 122 A. 412, 413 (1923) (“When the contract of employment has been reduced to writing, the question whether the person employed was an independent contractor or merely a servant is determined by the court as a matter of law”); Tankersley v. Webster, 116 Okla. 208, 210, 243 P. 745, 747 (1925) (“[T]he contract of employment between Tankersley and Casey was admitted in evidence without objections, and we think conclusively shows that Casey was an independent contractor”); Waldron v. Garland Pocahontas Coal Co., 89 W. Va. 426, 427, 109 S.E. 729 (1921) (syllabus) (“Whether a person performing work for another is an independent contractor depends upon a consideration of the contract of employment, the nature of the business, the circumstances under which the contract was made and the work was done”); see also App. to Brief for Respondent 1a–12a (citing additional examples). 4 See, e.g., Act of Mar. 19, 1924, ch. 70, §5, 43Stat. 28 (limiting payment of fees to attorneys “employed” by the Cherokee Tribe to litigate claims against the United States to those “stipulated in the contract of employment”); Act of June 7, 1924, ch. 300, §§2, 5, 43Stat. 537–538 (providing same for Choctaw and Chickasaw Tribes); Act of Aug. 24, 1921, ch. 89, 42Stat. 192 (providing that no funds may be used to compensate “any attorney, regular or special, for the United States Shipping Board or the United States Shipping Board Emergency Fleet Corporation unless the contract of employment has been approved by the Attorney General of the United States”). See also App. to Brief for Respondent 13a (citing additional examples). 5 See, e.g., Act of Mar. 10, 1909, ch. 70, §1, 1909 Kan. Sess. Laws p. 121 (referring to “contracts of employment of auditors, accountants, engineers, attorneys, counselors and architects for any special purpose”); Act of Mar. 4, 1909, ch. 4, §4, 1909 Okla. Sess. Laws p. 118 (“Should the amount of the attorney’s fee be agreed upon in the contract of employment, then such attorney’s lien and cause of action against such adverse party shall be for the amount so agreed upon”); Act of Mar. 4, 1924, ch. 88, §1, 1924 Va. Acts ch. 91 (allowing extension of “contracts of employment” between the state and contractors with respect to the labor of prisoners); App. to Brief for Respondent 14a–15a (citing additional examples). 6 See, e.g., Atlantic Transp. Co. v. Coneys, 82 F. 177, 178 (CA2 1897); Nyback v. Champagne Lumber Co., 109 F. 732, 741 (CA7 1901). 7 See Carlson, Why the Law Still Can’t Tell an Employee When It Sees One and How It Ought To Stop Trying, 22 Berkeley J. Emp. & Lab. L. 295, 309 (2001) (discussing the “historical baggage” of the term “servant”); Broden, General Rules Determining the Employment Relationship Under Social Security Laws: After Twenty Years an Unsolved Problem, 33 Temp. L. Q. 307, 327 (1960) (describing use of the term “employer-employee,” in contradistinction to “master-servant,” in the Social Security laws). Legislators searched to find a term that fully encompassed the broad protections they sought to provide and considered an “assortment of vague and uncertain terms,” including “ ‘servant,’ . . . ‘employee,’ . . . ‘workman,’ ‘laborer,’ ‘wage earner,’ ‘operative,’ or ‘hireling.’ ” Carlson, 22 Berkeley J. Emp. & Lab. L., at 308. Eventually “ ‘employee’ prevailed, if only by default, and the choice was confirmed by the next wave of protective legislation—workers’ compensation laws in the early years of the Twentieth Century.” Id., at 309. 8 Black’s Law Dictionary 641 (10th ed. 2014); see also P. Durkin, Release Notes: The Changes in Empathy, Employ, and Empire (Mar. 13, 2014) (“Over time” the meaning of several employ-related words have “reflect[ed] changes in the world of work” and their meaning “shows an increasingly marked narrowing”), online at https://public.oed.com/blog/march-2014-update-release-notes/. 9 See, e.g., 1 T. Conyngton, Business Law: A Working Manual of Every-day Law 302–303 (2d ed. 1920); Newland v. Bear, 218 App. Div. 308, 309, 218 N.Y.S. 81, 81–82 (1926); Anderson v. State Indus. Accident Comm’n, 107 Ore. 304, 311–312, 215 P. 582, 583, 585 (1923); N. Dosker, Manual of Compensation Law: State and Federal 8 (1917). 10 See, e.g., The Sea Lark, 14 F.2d 201 (WD Wash. 1926); The Buena Ventura, 243 F. 797, 799 (SDNY 1916); Holt v. Cummings, 102 Pa. 212, 215 (1883); Allan v. State S. S. Co., 132 N.Y. 91, 99, 30 N.E. 482, 485 (1892) (“The work which the physician does after the vessel starts on the voyage is his and not the ship owner’s”). 11 Transportation Act of 1920, §§304, 307, 41Stat. 456; Railway Employees’ Dept., A. F. of L. v. Indiana Harbor Belt R. Co., Decision No. 982, 3 R. L. B. 332, 337 (1922). 12 The Act provided for arbitration between railroads and workers, and defined “employees” as “all persons actually engaged in any capacity in train operation or train service of any description.” Act of June 1, 1898, ch. 370, 30Stat. 424. The Act also specified that the railroads would “be responsible for the acts and defaults of such employees in the same manner and to the same extent as if . . . said employees [were] directly employed by it.” Id., at 425. See Dempsey, Transportation: A Legal History, 30 Transp. L. J. 235, 273 (2003).

587.US.2018_17-1174

Respondent Russell Bartlett was arrested by police officers Luis Nieves and Bryce Weight for disorderly conduct and resisting arrest during “Arctic Man,” a raucous winter sports festival held in a remote part of Alaska. According to Sergeant Nieves, he was speaking with a group of attendees when a seemingly intoxicated Bartlett started shouting at them not to talk to the police. When Nieves approached him, Bartlett began yelling at the officer to leave. Rather than escalate the situation, Nieves left. Bartlett disputes that account, claiming that he was not drunk at that time and did not yell at Nieves. Minutes later, Trooper Weight says, Bartlett approached him in an aggressive manner while he was questioning a minor, stood between Weight and the teenager, and yelled with slurred speech that Weight should not speak with the minor. When Bartlett stepped toward Weight, the officer pushed him back. Nieves saw the confrontation and initiated an arrest. When Bartlett was slow to comply, the officers forced him to the ground. Bartlett denies being aggressive and claims that he was slow to comply because of a back injury. After he was handcuffed, Bartlett claims that Nieves said “bet you wish you would have talked to me now.” Bartlett sued under 42 U. S. C. §1983, claiming that the officers violated his First Amendment rights by arresting him in retaliation for his speech—i.e., his initial refusal to speak with Nieves and his intervention in Weight’s discussion with the minor. The District Court granted summary judgment for the officers, holding that the existence of probable cause to arrest Bartlett precluded his claim. The Ninth Circuit reversed. It held that probable cause does not defeat a retaliatory arrest claim and concluded that Bartlett’s affidavit about what Nieves allegedly said after the arrest could enable Bartlett to prove that the officers’ desire to chill his speech was a but-for cause of the arrest. Held: Because there was probable cause to arrest Bartlett, his retaliatory arrest claim fails as a matter of law. Pp. 4–16. (a) To prevail on a claim such as Bartlett’s, the plaintiff must show not only that the official acted with a retaliatory motive and that the plaintiff was injured, but also that the motive was a “but-for” cause of the injury. Hartman v. Moore, 547 U.S. 250, 259–260. Establishing that causal connection may be straightforward in some cases, see, e.g., Mt. Healthy City Bd. of Ed. v. Doyle, 429 U.S. 274, but other times it is not so simple. In retaliatory prosecution cases, for example, the causal inquiry is particularly complex because the official alleged to have the retaliatory motive does not carry out the retaliatory action himself. Instead, the decision to bring charges is made by a prosecutor—who is generally immune from suit and whose decisions receive a presumption of regularity. To account for that “problem of causation,” plaintiffs in retaliatory prosecution cases must prove as a threshold matter that the decision to press charges was objectively unreasonable because it was not supported by probable cause. Hartman, 547 U. S., at 263. Pp. 5–7. (b) Because First Amendment retaliatory arrest claims involve causal complexities akin to those identified in Hartman—see, e.g., Reichle v. Howards, 566 U.S. 658; Lozman v. Riviera Beach, 585 U. S. ___—the same no-probable-cause requirement generally should apply. The causal inquiry is complex because protected speech is often a “wholly legitimate consideration” for officers when deciding whether to make an arrest. Reichle, 566 U. S., at 668. In addition, “evidence of the presence or absence of probable cause for the arrest will be available in virtually every retaliatory arrest case.” Ibid. Its absence will generally provide weighty evidence that the officers’ animus caused the arrest, whereas its presence will suggest the opposite. While retaliatory arrest cases do not implicate the presumption of prosecutorial regularity or necessarily involve multiple government actors, the ultimate problem remains the same: For both claims, it is particularly difficult to determine whether the adverse government action was caused by the officers’ malice or by the plaintiff’s potentially criminal conduct. Bartlett’s proposed approach disregards the causal complexity involved in these cases and dismisses the need for any threshold objective showing, moving directly to consideration of the officers’ subjective intent. In the Fourth Amendment context, however, this Court has “almost uniformly rejected invitations to probe [officers’] subjective intent,” Ashcroft v. al-Kidd, 563 U.S. 731, 737. A purely subjective approach would undermine that precedent, would “dampen the ardor of all but the most resolute, or the most irresponsible, in the unflinching discharge of their duties,” Gregoire v. Biddle, 177 F.2d 579, 581, would compromise evenhanded application of the law by making the constitutionality of an arrest “vary from place to place and from time to time” depending on the personal motives of individual officers, Devenpeck v. Alford, 543 U.S. 146, 154, and would encourage officers to minimize communication during arrests to avoid having their words scrutinized for hints of improper motive. Pp. 8–11. (c) When defining the contours of a §1983 claim, this Court looks to “common-law principles that were well settled at the time of its enactment.” Kalina v. Fletcher, 522 U.S. 118, 123. In 1871, when §1983 was enacted, there was no common law tort for retaliatory arrest based on protected speech. Turning to the “closest analog[s],” Heck v. Humphrey, 512 U.S. 477, 484, both false imprisonment and malicious prosecution suggest the same result: The presence of probable cause should generally defeat a First Amendment retaliatory arrest claim. Pp. 12–13. (d) Because States today permit warrantless misdemeanor arrests for minor criminal offenses in a wide range of situations—whereas such arrests were privileged only in limited circumstances when §1983 was adopted—a narrow qualification is warranted for circumstances where officers have probable cause to make arrests, but typically exercise their discretion not to do so. An unyielding requirement to show the absence of probable cause in such cases could pose “a risk that some police officers may exploit the arrest power as a means of suppressing speech.” Lozman, 585pp U. S., at ___. Thus, the no-probable-cause requirement should not apply when a plaintiff presents objective evidence that he was arrested when otherwise similarly situated individuals not engaged in the same sort of protected speech had not been. Cf. United States v. Armstrong, 517 U.S. 456, 465. Because this inquiry is objective, the statements and motivations of the particular arresting officer are irrelevant at this stage. After making the required showing, the plaintiff’s claim may proceed in the same manner as claims where the plaintiff has met the threshold showing of the absence of probable cause. Pp. 13–15. 712 Fed. Appx. 613, reversed and remanded. Roberts, C. J., delivered the opinion of the Court, in which Breyer, Alito, Kagan, and Kavanaugh, JJ., joined, and in which Thomas, J., joined except as to Part II–D. Thomas, J., filed an opinion concurring in part and concurring in the judgment. Gorsuch, J., filed an opinion concurring in part and dissenting in part. Ginsburg, J., filed an opinion concurring in the judgment in part and dissenting in part. Sotomayor, J., filed a dissenting opinion.

Respondent Russell Bartlett sued petitioners—two police officers—alleging that they retaliated against him for his protected First Amendment speech by arresting him for disorderly conduct and resisting arrest. The officers had probable cause to arrest Bartlett, and we now decide whether that fact defeats Bartlett’s First Amendment claim as a matter of law. I A Bartlett was arrested during “Arctic Man,” a weeklong winter sports festival held in the remote Hoodoo Mountains near Paxson, Alaska. Paxson is a small community that normally consists of a few dozen residents. But once a year, upwards of 10,000 people descend on the area for Arctic Man, an event known for both extreme sports and extreme alcohol consumption. The mainstays are high-speed ski and snowmobile races, bonfires, and parties. During that week, the Arctic Man campground briefly becomes one of the largest and most raucous cities in Alaska. The event poses special challenges for law enforcement. Snowmobiles, alcohol, and freezing temperatures do not always mix well, and officers spend much of the week responding to snowmobile crashes, breaking up fights, and policing underage drinking. Given the remote location of the event, Alaska flies in additional officers from around the State to provide support. Still, the number of police remains limited. Even during the busiest periods of the event, only six to eight officers are on patrol at a time. On the last night of Arctic Man 2014, Sergeant Luis Nieves and Trooper Bryce Weight arrested Bartlett. The parties dispute certain details about the arrest but agree on the general course of events, some of which were captured on video by a local news reporter. At around 1:30 a.m., Sergeant Nieves and Bartlett first crossed paths. Nieves was asking some partygoers to move their beer keg inside their RV because minors had been making off with alcohol. According to Nieves, Bartlett began belligerently yelling to the RV owners that they should not speak with the police. Nieves approached Bartlett to explain the situation, but Bartlett was highly intoxicated and yelled at him to leave. Rather than escalate the situation, Nieves left. Bartlett disputes that account. According to Bartlett, he was not drunk at that time and never yelled at Nieves. He claims it was Nieves who became aggressive when Bartlett refused to speak with him. Several minutes later, Bartlett saw Trooper Weight asking a minor whether he and his underage friends had been drinking. According to Weight, Bartlett approached in an aggressive manner, stood between Weight and the teenager, and yelled with slurred speech that Weight should not speak with the minor. Weight claims that Bartlett then stepped very close to him in a combative way, so Weight pushed him back. Sergeant Nieves saw the confrontation and rushed over, arriving right after Weight pushed Bartlett. Nieves immediately initiated an arrest, and when Bartlett was slow to comply with his orders, the officers forced him to the ground and threatened to tase him. Again, Bartlett tells a different story. He denies being aggressive, and claims that he stood close to Weight only in an effort to speak over the loud background music. And he was slow to comply with Nieves’s orders, not because he was resisting arrest, but because he did not want to aggravate a back injury. After Bartlett was handcuffed, he claims that Nieves said: “[B]et you wish you would have talked to me now.” 712 Fed. Appx. 613, 616 (CA9 2017). The officers took Bartlett to a holding tent, where he was charged with disorderly conduct and resisting arrest. He had sustained no injuries during the episode and was released a few hours later. B The State ultimately dismissed the criminal charges against Bartlett, and Bartlett then sued the officers under 42 U. S. C. §1983, which provides a cause of action for state deprivations of federal rights. As relevant here, he claimed that the officers violated his First Amendment rights by arresting him in retaliation for his speech. The protected speech, according to Bartlett, was his refusal to speak with Nieves earlier in the evening and his intervention in Weight’s discussion with the underage partygoer. The officers responded that they arrested Bartlett because he interfered with an investigation and initiated a physical confrontation with Weight. The District Court granted summary judgment for the officers. The court determined that the officers had probable cause to arrest Bartlett and held that the existence of probable cause precluded Bartlett’s First Amendment retaliatory arrest claim. The Ninth Circuit disagreed. 712 Fed. Appx. 613. Relying on its prior decision in Ford v. Yakima, 706 F.3d 1188 (2013), the court held that a plaintiff can prevail on a First Amendment retaliatory arrest claim even in the face of probable cause for the arrest. According to the Ninth Circuit, Bartlett needed to show only (1) that the officers’ conduct would “chill a person of ordinary firmness from future First Amendment activity,” and (2) that he had advanced evidence that would “enable him ultimately to prove that the officers’ desire to chill his speech was a but-for cause” of the arrest. 712 Fed. Appx., at 616 (internal quotation marks omitted). The court concluded that Bartlett had satisfied both requirements: A retaliatory arrest is sufficiently chilling, and Bartlett had presented enough evidence that his speech was a but-for cause of the arrest. The only causal evidence relied on by the court was Bartlett’s affidavit alleging that Sergeant Nieves said “bet you wish you would have talked to me now.” If that allegation were true, the court reasoned, a jury might conclude that the officers arrested Bartlett in retaliation for his statements earlier that night. The officers petitioned for review in this Court, and we granted certiorari. 585 U. S. ___ (2018). II We are asked to resolve whether probable cause to make an arrest defeats a claim that the arrest was in retaliation for speech protected by the First Amendment. We have considered this issue twice in recent years. On the first occasion, we ultimately left the question unanswered because we decided the case on the alternative ground of qualified immunity. See Reichle v. Howards, 566 U.S. 658 (2012). We took up the question again last Term in Lozman v. Riviera Beach, 585 U. S. ___ (2018). Lozman involved unusual circumstances in which the plaintiff was arrested pursuant to an alleged “official municipal policy” of retaliation. Id., at ___ (slip op., at 11). Because those facts were “far afield from the typical retaliatory arrest claim,” we reserved judgment on the broader question presented and limited our holding to arrests that result from official policies of retaliation. Id., at ___ (slip op., at 10). In such cases, we held, probable cause does not categorically bar a plaintiff from suing the municipality. Id., at ___–___ (slip op., at 11–12). We now take up the question once again, this time in a more representative case. A “[A]s a general matter the First Amendment prohibits government officials from subjecting an individual to retaliatory actions” for engaging in protected speech. Hartman v. Moore, 547 U.S. 250, 256 (2006). If an official takes adverse action against someone based on that forbidden motive, and “non-retaliatory grounds are in fact insufficient to provoke the adverse consequences,” the injured person may generally seek relief by bringing a First Amendment claim. Ibid. (citing Crawford-El v. Britton, 523 U.S. 574, 593 (1998); Mt. Healthy City Bd. of Ed. v. Doyle, 429 U.S. 274, 283–284 (1977)). To prevail on such a claim, a plaintiff must establish a “causal connection” between the government defendant’s “retaliatory animus” and the plaintiff’s “subsequent in- jury.” Hartman, 547 U. S., at 259. It is not enough to show that an official acted with a retaliatory motive and that the plaintiff was injured—the motive must cause the injury. Specifically, it must be a “but-for” cause, meaning that the adverse action against the plaintiff would not have been taken absent the retaliatory motive. Id., at 260 (recognizing that although it “may be dishonorable to act with an unconstitutional motive,” an official’s “action colored by some degree of bad motive does not amount to a constitutional tort if that action would have been taken anyway”). For example, in Mt. Healthy, a teacher claimed that a school district refused to rehire him in retaliation for his protected speech. We held that even if the teacher’s “protected conduct played a part, substantial or otherwise, in [the] decision not to rehire,” he was not entitled to reinstatement “if the same decision would have been reached” absent his protected speech. 429 U. S., at 285. Regardless of the motives of the school district, we concluded that the First Amendment “principle at stake is sufficiently vindicated if such an employee is placed in no worse a position than if he had not engaged in the [protected speech].” Id., at 285–286. For a number of retaliation claims, establishing the causal connection between a defendant’s animus and a plaintiff’s injury is straightforward. Indeed, some of our cases in the public employment context “have simply taken the evidence of the motive and the discharge as sufficient for a circumstantial demonstration that the one caused the other,” shifting the burden to the defendant to show he would have taken the challenged action even without the impermissible motive. Hartman, 547 U. S., at 260 (citing Mt. Healthy, 429 U. S., at 287; Arlington Heights v. Metropolitan Housing Development Corp., 429 U.S. 252, 270, n. 21 (1977)). But the consideration of causation is not so straightforward in other types of retaliation cases. In Hartman, for example, we addressed retaliatory prosecution cases, where “proving the link between the defendant’s retaliatory animus and the plaintiff’s injury . . . ‘is usually more complex than it is in other retaliation cases.’ ” Lozman, 585 U. S., at ___ (slip op., at 8) (quoting Hartman, 547 U. S., at 261). Unlike most retaliation cases, in retaliatory prosecution cases the official with the malicious motive does not carry out the retaliatory action himself—the decision to bring charges is instead made by a prosecutor, who is generally immune from suit and whose decisions receive a presumption of regularity. Lozman, 585 U. S., at ___–___ (slip op., at 8–9). Thus, even when an officer’s animus is clear, it does not necessarily show that the officer “induced the action of a prosecutor who would not have pressed charges otherwise.” Hartman, 547 U. S., at 263. To account for this “problem of causation” in retaliatory prosecution claims, Hartman adopted the requirement that plaintiffs plead and prove the absence of probable cause for the underlying criminal charge. Ibid.; see id., at 265–266. As Hartman explained, that showing provides a “distinct body of highly valuable circumstantial evidence” that is “apt to prove or disprove” whether retaliatory animus actually caused the injury: “Demonstrating that there was no probable cause for the underlying criminal charge will tend to reinforce the retaliation evidence and show that retaliation was the but-for basis for instigating the prosecution, while establishing the existence of probable cause will suggest that prosecution would have occurred even without a retaliatory motive.” Id., at 261. Requiring plaintiffs to plead and prove the absence of probable cause made sense, we reasoned, because the existence of probable cause will be at issue in “practically all” retaliatory prosecution cases, has “high probative force,” and thus “can be made mandatory with little or no added cost.” Id., at 265. Moreover, imposing that burden on plaintiffs was necessary to suspend the presumption of regularity underlying the prosecutor’s charging decision—a presumption we “do not lightly discard.” Id., at 263; see also id., at 265. Thus, Hartman requires plaintiffs in retaliatory prosecution cases to show more than the subjective animus of an officer and a subsequent injury; plaintiffs must also prove as a threshold matter that the decision to press charges was objectively unreasonable because it was not supported by probable cause. B Officers Nieves and Weight argue that the same no-probable-cause requirement should apply to First Amendment retaliatory arrest claims. Their primary contention is that retaliatory arrest claims involve causal complexities akin to those we identified in Hartman, and thus warrant the same requirement that plaintiffs plead and prove the absence of probable cause. Brief for Petitioners 20–30. As a general matter, we agree. As we recognized in Reichle and reaffirmed in Lozman, retaliatory arrest claims face some of the same challenges we identified in Hartman: Like retaliatory prosecution cases, “retaliatory arrest cases also present a tenuous causal connection between the defendant’s alleged animus and the plaintiff’s injury.” Reichle, 566 U. S., at 668. The causal inquiry is complex because protected speech is often a “wholly legitimate consideration” for officers when deciding whether to make an arrest. Ibid.; Lozman, 585 U. S., at ___ (slip op., at 9). Officers frequently must make “split-second judgments” when deciding whether to arrest, and the content and manner of a suspect’s speech may convey vital information—for example, if he is “ready to cooperate” or rather “present[s] a continuing threat.” Id., at ___ (slip op., at 9) (citing District of Columbia v. Wesby, 583 U. S. ___, ___ (2018) (slip op., at 10) (“suspect’s untruthful and evasive answers to police questioning could support probable cause”)). Indeed, that kind of assessment happened in this case. The officers testified that they perceived Bartlett to be a threat based on a combination of the content and tone of his speech, his combative posture, and his apparent intoxication. In addition, “[l]ike retaliatory prosecution cases, evidence of the presence or absence of probable cause for the arrest will be available in virtually every retaliatory arrest case.” Reichle, 566 U. S., at 668. And because probable cause speaks to the objective reasonableness of an arrest, see Ashcroft v. al-Kidd, 563 U.S. 731, 736 (2011), its absence will—as in retaliatory prosecution cases—generally provide weighty evidence that the officer’s animus caused the arrest, whereas the presence of probable cause will suggest the opposite. To be sure, Reichle and Lozman also recognized that the two claims give rise to complex causal inquiries for somewhat different reasons. Unlike retaliatory prosecution cases, retaliatory arrest cases do not implicate the presumption of prosecutorial regularity or necessarily involve multiple government actors (although this case did). Reichle, 566 U. S., at 668–669; Lozman, 585 U. S., at ___ (slip op., at 10). But regardless of the source of the causal complexity, the ultimate problem remains the same. For both claims, it is particularly difficult to determine whether the adverse government action was caused by the of- ficer’s malice or the plaintiff’s potentially criminal conduct. See id., at ___ (slip op., at 9) (referring to “the complexity of proving (or disproving) causation” in retaliatory arrest cases). Because of the “close relationship” between the two claims, Reichle, 566 U. S., at 667, their related causal challenge should lead to the same solution: The plaintiff pressing a retaliatory arrest claim must plead and prove the absence of probable cause for the arrest. Bartlett, in defending the decision below, argues that the “causation in retaliatory-arrest cases is not inherently complex” because the “factfinder simply must determine whether the officer intended to punish the plaintiff for the plaintiff’s protected speech.” Brief for Respondent 36–37; see also post, at 5 (Sotomayor, J., dissenting). That approach fails to account for the fact that protected speech is often a legitimate consideration when deciding whether to make an arrest, and disregards the resulting causal complexity previously recognized by this Court. See Reichle, 566 U. S., at 668; Lozman, 585 U. S., at ___ (slip op., at 9). Bartlett’s approach dismisses the need for any threshold showing, moving directly to consideration of the subjective intent of the officers. In the Fourth Amendment context, however, “we have almost uniformly rejected invitations to probe subjective intent.” al-Kidd, 563 U. S., at 737; see also Kentucky v. King, 563 U.S. 452, 464 (2011) (“Legal tests based on reasonableness are generally objective, and this Court has long taken the view that evenhanded law enforcement is best achieved by the application of objective standards of conduct, rather than standards that depend upon the subjective state of mind of the officer.” (internal quotation marks omitted)). Police officers conduct approximately 29,000 arrests every day—a dangerous task that requires making quick decisions in “circumstances that are tense, uncertain, and rapidly evolving.” Graham v. Connor, 490 U.S. 386, 397 (1989). To ensure that officers may go about their work without undue apprehension of being sued, we generally review their conduct under objective standards of reasonableness. See Atwater v. Lago Vista, 532 U.S. 318, 351, and n. 22 (2001); Harlow v. Fitzgerald, 457 U.S. 800, 814–819 (1982). Thus, when reviewing an arrest, we ask “whether the circumstances, viewed objectively, justify [the challenged] action,” and if so, conclude “that action was reasonable whatever the subjective intent motivating the relevant officials.” al-Kidd, 563 U. S., at 736 (internal quotation marks omitted). A particular officer’s state of mind is simply “irrelevant,” and it provides “no basis for invalidating an arrest.” Devenpeck v. Alford, 543 U.S. 146, 153, 155 (2004). Bartlett’s purely subjective approach would undermine that precedent by allowing even doubtful retaliatory arrest suits to proceed based solely on allegations about an arresting officer’s mental state. See Lozman, 585 U. S., at ___ (slip op., at 9). Because a state of mind is “easy to allege and hard to disprove,” Crawford-El, 523 U. S., at 585, a subjective inquiry would threaten to set off “broad-ranging discovery” in which “there often is no clear end to the relevant evidence,” Harlow, 457 U. S., at 817. As a result, policing certain events like an unruly protest would pose overwhelming litigation risks. Any inartful turn of phrase or perceived slight during a legitimate arrest could land an officer in years of litigation. Bartlett’s standard would thus “dampen the ardor of all but the most resolute, or the most irresponsible, in the unflinching discharge of their duties.” Gregoire v. Biddle, 177 F.2d 579, 581 (CA2 1949) (Learned Hand, C. J.). It would also compromise evenhanded application of the law by making the constitutionality of an arrest “vary from place to place and from time to time” depending on the personal motives of individual officers. Devenpeck, 543 U. S., at 154. Yet another “predictable consequence” of such a rule is that officers would simply minimize their communication during arrests to avoid having their words scrutinized for hints of improper motive—a result that would leave everyone worse off. Id., at 155. Adopting Hartman’s no-probable-cause rule in this closely related context addresses those familiar concerns. Absent such a showing, a retaliatory arrest claim fails. But if the plaintiff establishes the absence of probable cause, “then the Mt. Healthy test governs: The plaintiff must show that the retaliation was a substantial or motivating factor behind the [arrest], and, if that showing is made, the defendant can prevail only by showing that the [arrest] would have been initiated without respect to retaliation.” Lozman, 585 U. S., at ___ (slip op., at 8) (citing Hartman, 547 U. S., at 265–266).[1] C Our conclusion is confirmed by the common law approach to similar tort claims. When defining the contours of a claim under §1983, we look to “common-law principles that were well settled at the time of its enactment.” Ka- lina v. Fletcher, 522 U.S. 118, 123 (1997); Manuel v. Joliet, 580 U. S. ___, ___ (2017) (slip op., at 12) (common law principles “guide” the definition of claims under §1983). As the parties acknowledge, when §1983 was enacted in 1871, there was no common law tort for retaliatory arrest based on protected speech. See Brief for Petitioners 43; Brief for Respondent 20. We therefore turn to the common law torts that provide the “closest analogy” to retaliatory arrest claims. Heck v. Humphrey, 512 U.S. 477, 484 (1994). The parties dispute whether the better analog is false imprisonment or malicious prosecution. At common law, false imprisonment arose from a “detention without legal process,” whereas malicious prosecution was marked “by wrongful institution of legal process.” Wallace v. Kato, 549 U.S. 384, 389–390 (2007).[2] Here, both claims suggest the same result: The presence of probable cause should generally defeat a First Amendment retaliatory arrest claim. See generally Lozman, 585 U. S., at ___–___ (slip op., at 4–6) (Thomas, J., dissenting). Malicious prosecution required the plaintiff to show that the criminal charge against him “was unfounded, and that it was made without reasonable or probable cause, and that the defendant in making or instigating it was actuated by malice.” Wheeler v. Nesbitt, 24 How. 544, 549–550 (1861); see also Restatement of Torts §653 (1938). It has long been “settled law” that malicious prosecution requires proving “the want of probable cause,” and Bartlett does not argue otherwise. Brown v. Selfridge, 224 U.S. 189, 191 (1912); see also Wheeler, 24 How., at 550 (noting that “[w]ant of reasonable and probable cause” is an “element in the action for a malicious criminal prosecution”). For claims of false imprisonment, the presence of probable cause was generally a complete defense for peace officers. See T. Cooley, Law of Torts 175 (1880); 1 F. Hilliard, The Law of Torts or Private Wrongs 207–208, and n. (a) (1859). In such cases, arresting officers were protected from liability if the arrest was “privileged.” At common law, peace officers were privileged to make warrantless arrests based on probable cause of the commission of a felony or certain misdemeanors. See Restatement of Torts §§118, 119, 121 (1934); see also Cooley, Law of Torts, at 175–176 (stating that peace officers who make arrests based on probable cause “will be excused, even though it appear afterwards that in fact no felony had been committed”); see generally Atwater, 532 U. S., at 340–345 (reviewing the history of warrantless arrests for misdemeanors). Although the exact scope of the privilege varied somewhat depending on the jurisdiction, the consistent rule was that officers were not liable for arrests they were privileged to make based on probable cause. D Although probable cause should generally defeat a retaliatory arrest claim, a narrow qualification is warranted for circumstances where officers have probable cause to make arrests, but typically exercise their discretion not to do so. In such cases, an unyielding requirement to show the absence of probable cause could pose “a risk that some police officers may exploit the arrest power as a means of suppressing speech.” Lozman, 585 U. S., at ___ (slip op., at 10). When §1983 was adopted, officers were generally privileged to make warrantless arrests for misdemeanors only in limited circumstances. See Restatement of Torts §121, Comments e, h, at 262–263. Today, however, “statutes in all 50 States and the District of Columbia permit warrantless misdemeanor arrests” in a much wider range of situations—often whenever officers have probable cause for “even a very minor criminal offense.” Atwater, 532 U. S., at 344–345, 354; see id., at 355–360 (listing state statutes). For example, at many intersections, jaywalking is endemic but rarely results in arrest. If an individual who has been vocally complaining about police conduct is arrested for jaywalking at such an intersection, it would seem insufficiently protective of First Amendment rights to dismiss the individual’s retaliatory arrest claim on the ground that there was undoubted probable cause for the arrest. In such a case, because probable cause does little to prove or disprove the causal connection between animus and injury, applying Hartman’s rule would come at the expense of Hartman’s logic. For those reasons, we conclude that the no-probable-cause requirement should not apply when a plaintiff presents objective evidence that he was arrested when otherwise similarly situated individuals not engaged in the same sort of protected speech had not been. Cf. United States v. Armstrong, 517 U.S. 456, 465 (1996). That showing addresses Hartman’s causal concern by helping to establish that “non-retaliatory grounds [we]re in fact insufficient to provoke the adverse consequences.” 547 U. S., at 256. And like a probable cause analysis, it provides an objective inquiry that avoids the significant problems that would arise from reviewing police conduct under a purely subjective standard. Because this inquiry is objective, the statements and motivations of the particular arresting officer are “irrelevant” at this stage. Devenpeck, 543 U. S., at 153. After making the required showing, the plaintiff’s claim may proceed in the same manner as claims where the plaintiff has met the threshold showing of the absence of probable cause. See Lozman, 585 U. S., at ___ (slip op., at 8). * * * In light of the foregoing, Bartlett’s retaliation claim cannot survive summary judgment. As an initial matter, the record contains insufficient evidence of retaliation on the part of Trooper Weight. The only evidence of retaliatory animus identified by the Ninth Circuit was Bartlett’s affidavit stating that Sergeant Nieves said “bet you wish you would have talked to me now.” 712 Fed. Appx., at 616. But that allegation about Nieves says nothing about what motivated Weight, who had no knowledge of Bartlett’s prior run-in with Nieves. Cf. Lozman, 585 U. S., at ___ (slip op., at 10) (plaintiff “likely could not have maintained a retaliation claim against the arresting officer” when there was “no showing that the officer had any knowledge of [the plaintiff’s] prior speech”). In any event, Bartlett’s claim against both officers cannot succeed because they had probable cause to arrest him. As the Court of Appeals explained: “When Sergeant Nieves initiated Bartlett’s arrest, he knew that Bartlett had been drinking, and he observed Bartlett speaking in a loud voice and standing close to Trooper Weight. He also saw Trooper Weight push Bartlett back. . . . [T]he test is whether the information the officer had at the time of making the arrest gave rise to probable cause. We agree with the district court that it did; a reasonable officer in Sergeant Nieves’s position could have concluded that Bartlett stood close to Trooper Weight and spoke loudly in order to challenge him, provoking Trooper Weight to push him back.” 712 Fed. Appx., at 615 (citations and internal quotation marks omitted). Because there was probable cause to arrest Bartlett, his retaliatory arrest claim fails as a matter of law. Accordingly, the judgment of the United States Court of Appeals for the Ninth Circuit is reversed, and the case is remanded for further proceedings consistent with this opinion. It is so ordered. Notes 1 Justice Sotomayor would have us extend Mt. Healthy and rely on that “tried and true” approach as the exclusive standard in the retaliatory arrest context. See post, at 1–5, 14 (dissenting opinion). But not even respondent Bartlett argues for such a rule. And since our decisions in Hartman and Reichle, no court of appeals has applied that approach in retaliatory arrest cases of this sort. Justice Sotomayor criticizes the Court for spending “[m]uch of its opinion . . . analogizing to Hartman,” post, at 4, but of course Hartman is our precedent most directly on point. To the extent retaliatory arrest cases raise concerns distinct from that precedent, we have departed from Hartman to afford greater First Amendment protection. See infra, at 13–15. 2 For our purposes, we need not distinguish between the torts of false imprisonment and false arrest, which are “virtually synonymous.” 35 C. J. S., False Imprisonment §2, p. 522 (2009); see also Wallace, 549 U. S., at 388–389.

588.US.2018_18-457

Joseph Lee Rice III formed a trust for the benefit of his children in his home State of New York and appointed a fellow New York resident as the trustee. The trust agreement granted the trustee “absolute discretion” to distribute the trust’s assets to the beneficiaries. In 1997, Rice’s daughter, Kimberley Rice Kaestner, moved to North Carolina. The trustee later divided Rice’s initial trust into three separate subtrusts, and North Carolina sought to tax the Kimberley Rice Kaestner 1992 Family Trust (Trust)—formed for the benefit of Kaestner and her three children—under a law authorizing the State to tax any trust income that “is for the benefit of” a state resident, N. C. Gen. Stat. Ann. §105–160.2. The State assessed a tax of more than $1.3 million for tax years 2005 through 2008. During that period, Kaestner had no right to, and did not receive, any distributions. Nor did the Trust have a physical presence, make any direct investments, or hold any real property in the State. The trustee paid the tax under protest and then sued the taxing authority in state court, arguing that the tax as applied to the Trust violates the Fourteenth Amendment’s Due Process Clause. The state courts agreed, holding that the Kaestners’ in-state residence was too tenuous a link between the State and the Trust to support the tax. Held: The presence of in-state beneficiaries alone does not empower a State to tax trust income that has not been distributed to the beneficiaries where the beneficiaries have no right to demand that income and are uncertain to receive it. Pp. 5–16. (a) The Due Process Clause limits States to imposing only taxes that “bea[r] fiscal relation to protection, opportunities and benefits given by the state.” Wisconsin v. J. C. Penney Co., 311 U.S. 435, 444. Compliance with the Clause’s demands “requires some definite link, some minimum connection, between a state and the person, property or transaction it seeks to tax,” and that “the ‘income attributed to the State for tax purposes . . . be rationally related to “values connected with the taxing State,” ’ ” Quill Corp. v. North Dakota, 504 U.S. 298, 306. That “minimum connection” inquiry is “flexible” and focuses on the reasonableness of the government’s action. Id., at 307. Pp. 5–6. (b) In the trust beneficiary context, the Court’s due process analysis of state trust taxes focuses on the extent of the in-state beneficiary’s right to control, possess, enjoy, or receive trust assets. Cases such as Safe Deposit & Trust Co. of Baltimore v. Virginia, 280 U.S. 83; Brooke v. Norfolk, 277 U.S. 27; and Maguire v. Trefry, 253 U.S. 12, reflect a common principle: When a State seeks to base its tax on the in-state residence of a trust beneficiary, the Due Process Clause demands a pragmatic inquiry into what exactly the beneficiary controls or possesses and how that interest relates to the object of the State’s tax. Safe Deposit, 280 U. S., at 91. Similar analysis also appears in the context of taxes premised on the in-state residency of settlors and trustees. See, e.g., Curry v. McCanless, 307 U.S. 357. Pp. 6–10. (c) Applying these principles here, the residence of the Trust beneficiaries in North Carolina alone does not supply the minimum connection necessary to sustain the State’s tax. First, the beneficiaries did not receive any income from the Trust during the years in question. Second, they had no right to demand Trust income or otherwise control, possess, or enjoy the Trust assets in the tax years at issue. Third, they also could not count on necessarily receiving any specific amount of income from the Trust in the future. Pp. 10–13. (d) The State’s counterarguments are unconvincing. First the State argues that “a trust and its constituents” are always “inextricably intertwined,” and thus, because trustee residence supports state taxation, so too must beneficiary residence. The State emphasizes that beneficiaries are essential to a trust and have an equitable interest in its assets. Although a beneficiary is central to the trust relationship, the wide variation in beneficiaries’ interests counsels against adopting such a categorical rule. Second, the State argues that ruling in favor of the Trust will undermine numerous state taxation regimes. But only a small handful of States rely on beneficiary residency as a sole basis for trust taxation, and an even smaller number rely on the residency of beneficiaries regardless of whether the beneficiary is certain to receive trust assets. Finally, the State urges that adopting the Trust’s position will lead to opportunistic gaming of state tax systems. There is no certainty, however, that such behavior will regularly come to pass, and in any event, mere speculation about negative consequences cannot conjure the “minimum connection” missing between the State and the object of its tax. Pp. 13–16. 371 N. C. 133, 814 S.E.2d 43, affirmed. Sotomayor, J., delivered the opinion for a unanimous Court. Alito, J., filed a concurring opinion, in which Roberts, C. J., and Gorsuch, J., joined.

This case is about the limits of a State’s power to tax a trust. North Carolina imposes a tax on any trust income that “is for the benefit of” a North Carolina resident. N. C. Gen. Stat. Ann. §105–160.2 (2017). The North Carolina courts interpret this law to mean that a trust owes income tax to North Carolina whenever the trust’s beneficiaries live in the State, even if—as is the case here—those beneficiaries received no income from the trust in the relevant tax year, had no right to demand income from the trust in that year, and could not count on ever receiving income from the trust. The North Carolina courts held the tax to be unconstitutional when assessed in such a case because the State lacks the minimum connection with the object of its tax that the Constitution requires. We agree and affirm. As applied in these circumstances, the State’s tax violates the Due Process Clause of the Fourteenth Amendment. I A In its simplest form, a trust is created when one person (a “settlor” or “grantor”) transfers property to a third party (a “trustee”) to administer for the benefit of another (a “beneficiary”). A. Hess, G. Bogert, & G. Bogert, Law of Trusts and Trustees §1, pp. 8–10 (3d ed. 2007). As traditionally understood, the arrangement that results is not a “distinct legal entity, but a ‘fiduciary relationship’ between multiple people.” Americold Realty Trust v. ConAgra Foods, Inc., 577 U. S. ___, ___ (2016) (slip op., at 5). The trust comprises the separate interests of the beneficiary, who has an “equitable interest” in the trust property, and the trustee, who has a “legal interest” in that property. Greenough v. Tax Assessors of Newport, 331 U.S. 486, 494 (1947). In some contexts, however, trusts can be treated as if the trust itself has “a separate existence” from its constituent parts. Id., at 493.[1] The trust that challenges North Carolina’s tax had its first incarnation nearly 30 years ago, when New Yorker Joseph Lee Rice III formed a trust for the benefit of his children. Rice decided that the trust would be governed by the law of his home State, New York, and he appointed a fellow New York resident as the trustee.[2] The trust agreement provided that the trustee would have “absolute discretion” to distribute the trust’s assets to the beneficiaries “in such amounts and proportions” as the trustee might “from time to time” decide. Art. I, §1.2(a), App. 46–47. When Rice created the trust, no trust beneficiary lived in North Carolina. That changed in 1997, when Rice’s daughter, Kimberley Rice Kaestner, moved to the State. She and her minor children were residents of North Carolina from 2005 through 2008, the time period relevant for this case. A few years after Kaestner moved to North Carolina, the trustee divided Rice’s initial trust into three subtrusts. One of these subtrusts—the Kimberley Rice Kaestner 1992 Family Trust (Kaestner Trust or Trust)—was formed for the benefit of Kaestner and her three children. The same agreement that controlled the original trust also governed the Kaestner Trust. Critically, this meant that the trustee had exclusive control over the allocation and timing of trust distributions. North Carolina explained in the state-court proceedings that the State’s only connection to the Trust in the relevant tax years was the in-state residence of the Trust’s beneficiaries. App. to Pet. for Cert. 54a. From 2005 through 2008, the trustee chose not to distribute any of the income that the Trust accumulated to Kaestner or her children, and the trustee’s contacts with Kaestner were “infrequent.”[3] 371 N. C. 133, 143, 814 S.E.2d 43, 50 (2018). The Trust was subject to New York law, Art. X, App. 69, the grantor was a New York resident, App. 44, and no trustee lived in North Carolina, 371 N. C., at 134, 814 S. E. 2d, at 45. The trustee kept the Trust documents and records in New York, and the Trust asset custodians were located in Massachusetts. Ibid. The Trust also maintained no physical presence in North Carolina, made no direct investments in the State, and held no real property there. App. to Pet. for Cert. 52a–53a. The Trust agreement provided that the Kaestner Trust would terminate when Kaestner turned 40, after the time period relevant here. After consulting with Kaestner and in accordance with her wishes, however, the trustee rolled over the assets into a new trust instead of distributing them to her. This transfer took place after the relevant tax years. See N. Y. Est., Powers & Trusts Law Ann. §10–6.6(b) (West 2002) (authorizing this action). B North Carolina taxes any trust income that “is for the benefit of” a North Carolina resident. N. C. Gen. Stat. Ann. §105–160.2. The North Carolina Supreme Court interprets the statute to authorize North Carolina to tax a trust on the sole basis that the trust beneficiaries reside in the State. 371 N. C., at 143–144, 814 S. E. 2d, at 51. Applying this statute, the North Carolina Department of Revenue assessed a tax on the full proceeds that the Kaestner Trust accumulated for tax years 2005 through 2008 and required the trustee to pay it. See N. C. Gen. Stat. Ann. §105–160.2. The resulting tax bill amounted to more than $1.3 million. The trustee paid the tax under protest and then sued in state court, arguing that the tax as applied to the Kaestner Trust violates the Due Process Clause of the Fourteenth Amendment. The trial court decided that the Kaestners’ residence in North Carolina was too tenuous a link between the State and the Trust to support the tax and held that the State’s taxation of the Trust violated the Due Process Clause. App. to Pet. for Cert. 62a.[4] The North Carolina Court of Appeals affirmed, as did the North Carolina Supreme Court. A majority of the State Supreme Court reasoned that the Kaestner Trust and its beneficiaries “have legally separate, taxable existences” and thus that the contacts between the Kaestner family and their home State cannot establish a connection between the Trust “itself” and the State. 371 N. C., at 140–142, 814 S. E. 2d, at 49. We granted certiorari to decide whether the Due Process Clause prohibits States from taxing trusts based only on the in-state residency of trust beneficiaries. 586 U. S. ___ (2019). II The Due Process Clause provides that “[n]o State shall . . . deprive any person of life, liberty, or property, without due process of law.” Amdt. 14, §1. The Clause “centrally concerns the fundamental fairness of governmental activ- ity.” Quill Corp. v. North Dakota, 504 U.S. 298, 312 (1992), overruled on other grounds, South Dakota v. Wayfair, Inc., 585 U. S. ___, ___ (2018) (slip op., at 10). In the context of state taxation, the Due Process Clause limits States to imposing only taxes that “bea[r] fiscal relation to protection, opportunities and benefits given by the state.” Wisconsin v. J. C. Penney Co., 311 U.S. 435, 444 (1940). The power to tax is, of course, “essential to the very existence of government,” McCulloch v. Maryland, 4 Wheat. 316, 428 (1819), but the legitimacy of that power requires drawing a line between taxation and mere unjustified “confiscation.” Miller Brothers Co. v. Maryland, 347 U.S. 340, 342 (1954). That boundary turns on the “[t]he simple but controlling question . . . whether the state has given anything for which it can ask return.” Wisconsin, 311 U. S., at 444. The Court applies a two-step analysis to decide if a state tax abides by the Due Process Clause. First, and most relevant here, there must be “ ‘some definite link, some minimum connection, between a state and the person, property or transaction it seeks to tax.’ ” Quill, 504 U. S., at 306. Second, “the ‘income attributed to the State for tax purposes must be rationally related to “values connected with the taxing State.” ’ ” Ibid.[5] To determine whether a State has the requisite “minimum connection” with the object of its tax, this Court borrows from the familiar test of International Shoe Co. v. Washington, 326 U.S. 310 (1945). Quill, 504 U. S., at 307. A State has the power to impose a tax only when the taxed entity has “certain minimum contacts” with the State such that the tax “does not offend ‘traditional notions of fair play and substantial justice.’ ” International Shoe Co., 326 U. S., at 316; see Quill, 504 U. S., at 308. The “minimum contacts” inquiry is “flexible” and focuses on the reason- ableness of the government’s action. Quill, 504 U. S., at 307. Ultimately, only those who derive “benefits and protection” from associating with a State should have obligations to the State in question. International Shoe, 326 U. S., at 319. III One can imagine many contacts with a trust or its constituents that a State might treat, alone or in combination, as providing a “minimum connection” that justifies a tax on trust assets. The Court has already held that a tax on trust income distributed to an in-state resident passes muster under the Due Process Clause. Maguire v. Trefry, 253 U.S. 12, 16–17 (1920). So does a tax based on a trustee’s in-state residence. Greenough, 331 U. S., at 498. The Court’s cases also suggest that a tax based on the site of trust administration is constitutional. See Hanson v. Denckla, 357 U.S. 235, 251 (1958); Curry v. McCanless, 307 U.S. 357, 370 (1939). A different permutation is before the Court today. The Kaestner Trust made no distributions to any North Carolina resident in the years in question. 371 N. C., at 134–135, 814 S. E. 2d, at 45. The trustee resided out of State, and Trust administration was split between New York (where the Trust’s records were kept) and Massachusetts (where the custodians of its assets were located). Id., at 134, 814 S. E. 2d, at 45. The trustee made no direct investments in North Carolina in the relevant tax years, App. to Pet. for Cert. 52a, and the settlor did not reside in North Carolina. 371 N. C., at 134, 814 S. E. 2d, at 45. Of all the potential kinds of connections between a trust and a State, the State seeks to rest its tax on just one: the in-state residence of the beneficiaries. Brief for Petitioner 34–36; see App. to Pet. for Cert. 54a. We hold that the presence of in-state beneficiaries alone does not empower a State to tax trust income that has not been distributed to the beneficiaries where the beneficiaries have no right to demand that income and are uncertain ever to receive it. In limiting our holding to the specific facts presented, we do not imply approval or disapproval of trust taxes that are premised on the residence of beneficiaries whose relationship to trust assets differs from that of the beneficiaries here. A In the past, the Court has analyzed state trust taxes for consistency with the Due Process Clause by looking to the relationship between the relevant trust constituent (settlor, trustee, or beneficiary) and the trust assets that the State seeks to tax. In the context of beneficiary contacts specifically, the Court has focused on the extent of the in-state beneficiary’s right to control, possess, enjoy, or receive trust assets. The Court’s emphasis on these factors emerged in two early cases, Safe Deposit & Trust Co. of Baltimore v. Virginia, 280 U.S. 83 (1929), and Brooke v. Norfolk, 277 U.S. 27 (1928), both of which invalidated state taxes premised on the in-state residency of beneficiaries. In each case the challenged tax fell on the entirety of a trust’s property, rather than on only the share of trust assets to which the beneficiaries were entitled. Safe Deposit, 280 U. S., at 90, 92; Brooke, 277 U. S., at 28. In Safe Deposit, the Court rejected Virginia’s attempt to tax a trustee on the “whole corpus of the trust estate,” 280 U. S., at 90; see id., at 93, explaining that “nobody within Virginia ha[d] present right to [the trust property’s] control or possession, or to receive income therefrom,” id., at 91. In Brooke, the Court rejected a tax on the entirety of a trust fund assessed against a resident beneficiary because the trust property “[wa]s not within the State, d[id] not belong to the [beneficiary] and [wa]s not within her possession or control.” 277 U. S., at 29.[6] On the other hand, the same elements of possession, control, and enjoyment of trust property led the Court to uphold state taxes based on the in-state residency of beneficiaries who did have close ties to the taxed trust assets. The Court has decided that States may tax trust income that is actually distributed to an in-state beneficiary. In those circumstances, the beneficiary “own[s] and enjoy[s]” an interest in the trust property, and the State can exact a tax in exchange for offering the beneficiary protection. Maguire, 253 U. S., at 17; see also Guaranty Trust Co. v. Virginia, 305 U.S. 19, 21–23 (1938). All of the foregoing cases reflect a common governing principle: When a State seeks to base its tax on the in-state residence of a trust beneficiary, the Due Process Clause demands a pragmatic inquiry into what exactly the beneficiary controls or possesses and how that interest relates to the object of the State’s tax. See Safe Deposit, 280 U. S., at 91. Although the Court’s resident-beneficiary cases are most relevant here, similar analysis also appears in the context of taxes premised on the in-state residency of settlors and trustees. In Curry, for instance, the Court upheld a Tennessee trust tax because the settlor was a Tennessee resident who retained “power to dispose of” the property, which amounted to “a potential source of wealth which was property in her hands.” 307 U. S., at 370. That practical control over the trust assets obliged the settlor “to contribute to the support of the government whose protection she enjoyed.” Id., at 371; see also Graves v. Elliott, 307 U.S. 383, 387 (1939) (a settlor’s “right to revoke [a] trust and to demand the transmission to her of the intangibles . . . was a potential source of wealth” subject to tax by her State of residence).[7] A focus on ownership and rights to trust assets also featured in the Court’s ruling that a trustee’s in-state residence can provide the basis for a State to tax trust assets. In Greenough, the Court explained that the relationship between trust assets and a trustee is akin to the “close relationship between” other types of intangible property and the owners of such property. 331 U. S., at 493. The trustee is “the owner of [a] legal interest in” the trust property, and in that capacity he can incur obligations, become personally liable for contracts for the trust, or have specific performance ordered against him. Id., at 494. At the same time, the trustee can turn to his home State for “benefit and protection through its law,” id., at 496, for instance, by resorting to the State’s courts to resolve issues related to trust administration or to enforce trust claims, id., at 495. A State therefore may tax a resident trustee on his interest in a share of trust assets. Id., at 498. In sum, when assessing a state tax premised on the in-state residency of a constituent of a trust—whether beneficiary, settlor, or trustee—the Due Process Clause demands attention to the particular relationship between the resident and the trust assets that the State seeks to tax. Because each individual fulfills different functions in the creation and continuation of the trust, the specific features of that relationship sufficient to sustain a tax may vary depending on whether the resident is a settlor, beneficiary, or trustee. When a tax is premised on the in-state residence of a beneficiary, the Constitution requires that the resident have some degree of possession, control, or enjoyment of the trust property or a right to receive that property before the State can tax the asset. Cf. Safe Deposit, 280 U. S., at 91–92.[8] Otherwise, the State’s relationship to the object of its tax is too attenuated to create the “minimum connection” that the Constitution requires. See Quill, 504 U. S., at 306. B Applying these principles here, we conclude that the residence of the Kaestner Trust beneficiaries in North Carolina alone does not supply the minimum connection necessary to sustain the State’s tax. First, the beneficiaries did not receive any income from the trust during the years in question. If they had, such income would have been taxable. See Maguire, 253 U. S., at 17; Guaranty Trust Co., 305 U. S., at 23. Second, the beneficiaries had no right to demand trust income or otherwise control, possess, or enjoy the trust assets in the tax years at issue. The decision of when, whether, and to whom the trustee would distribute the trust’s assets was left to the trustee’s “absolute discretion.” Art. I, §1.2(a), App. 46–47. In fact, the Trust agreement explicitly authorized the trustee to distribute funds to one beneficiary to “the exclusion of other[s],” with the effect of cutting one or more beneficiaries out of the Trust. Art. I, §1.4, id., at 50. The agreement also authorized the trustee, not the beneficiaries, to make investment decisions regarding Trust property. Art. V, §5.2, id., at 55–60. The Trust agreement prohibited the beneficiaries from assigning to another person any right they might have to the Trust property, Art. XII, id., at 70–71, thus making the beneficiaries’ interest less like “a potential source of wealth [that] was property in [their] hands.” Curry, 307 U. S., at 370–371.[9] To be sure, the Kaestner Trust agreement also instructed the trustee to view the trust “as a family asset and to be liberal in the exercise of the discretion conferred,” suggesting that the trustee was to make distributions generously with the goal of “meet[ing] the needs of the Beneficiaries” in various respects. Art. I, §1.4(c), App. 51. And the trustee of a discretionary trust has a fiduciary duty not to “act in bad faith or for some purpose or motive other than to accomplish the purposes of the discretionary power.” 2 Restatement (Third) of Trusts §50, Comment c, p. 262 (2003). But by reserving sole discretion to the trustee, the Trust agreement still deprived Kaestner and her children of any entitlement to demand distributions or to direct the use of the Trust assets in their favor in the years in question. Third, not only were Kaestner and her children unable to demand distributions in the tax years at issue, but they also could not count on necessarily receiving any specific amount of income from the Trust in the future. Although the Trust agreement provided for the Trust to terminate in 2009 (on Kaestner’s 40th birthday) and to distribute assets to Kaestner, Art. I, §1.2(c)(1), App. 47, New York law allowed the trustee to roll over the trust assets into a new trust rather than terminating it. N. Y. Est., Powers & Trusts §10–6.6(b). Here, the trustee did just that. 371 N. C., at 135, 814 S. E. 2d, at 45.[10] Like the beneficiaries in Safe Deposit, then, Kaestner and her children had no right to “control or posses[s]” the trust assets “or to receive income therefrom.” 280 U. S., at 91. The beneficiaries received no income from the Trust, had no right to demand income from the Trust, and had no assurance that they would eventually receive a specific share of Trust income. Given these features of the Trust, the beneficiaries’ residence cannot, consistent with due process, serve as the sole basis for North Carolina’s tax on trust income.[11] IV The State’s counterarguments do not save its tax. First, the State interprets Greenough as standing for the broad proposition that “a trust and its constituents” are always “inextricably intertwined.” Brief for Petitioner 26. Because trustee residence supports state taxation, the State contends, so too must beneficiary residence. The State emphasizes that beneficiaries are essential to a trust and have an “equitable interest” in its assets. Greenough, 331 U. S., at 494. In Stone v. White, 301 U.S. 532 (1937), the State notes, the Court refused to “shut its eyes to the fact” that a suit to recover taxes from a trust was in reality a suit regarding “the beneficiary’s money.” Id., at 535. The State also argues that its tax is at least as fair as the tax in Greenough because the Trust benefits from North Carolina law by way of the beneficiaries, who enjoy secure banks to facilitate asset transfers and also partake of services (such as subsidized public education) that obviate the need to make distributions (for example, to fund beneficiaries’ educations). Brief for Petitioner 30–33. The State’s argument fails to grapple with the wide variation in beneficiaries’ interests. There is no doubt that a beneficiary is central to the trust relationship, and beneficiaries are commonly understood to hold “beneficial interests (or ‘equitable title’) in the trust property,” 2 Restatement (Third) of Trusts §42, Comment a, at 186. In some cases the relationship between beneficiaries and trust assets is so close as to be beyond separation. In Stone, for instance, the beneficiary had already received the trust income on which the government sought to recover tax. See 301 U. S., at 533. But, depending on the trust agreement, a beneficiary may have only a “future interest,” an interest that is “subject to conditions,” or an interest that is controlled by a trustee’s discretionary decisions. 2 Restatement (Third) of Trusts §49, Comment b, at 243. By contrast, in Greenough, the requisite connection with the State arose from a legal interest that necessarily carried with it predictable responsibilities and liabilities. See 331 U. S., at 494. The different forms of beneficiary interests counsels against adopting the categorical rule that the State urges. Second, the State argues that ruling in favor of the Trust will undermine numerous state taxation regimes. Tr. of Oral Arg. 8, 68; Brief for Petitioner 6, and n. 1. Today’s ruling will have no such sweeping effect. North Carolina is one of a small handful of States that rely on beneficiary residency as a sole basis for trust taxation, and one of an even smaller number that will rely on the residency of beneficiaries regardless of whether the beneficiary is certain to receive trust assets.[12] Today’s decision does not address state laws that consider the in-state residency of a beneficiary as one of a combination of factors, that turn on the residency of a settlor, or that rely only on the residency of noncontingent beneficiaries, see, e.g., Cal. Rev. & Tax. Code Ann. §17742(a).[13] We express no opinion on the validity of such taxes. Finally, North Carolina urges that adopting the Trust’s position will lead to opportunistic gaming of state tax systems, noting that trust income nationally exceeded $120 billion in 2014. See Brief for Petitioner 39, and n. 13. The State is concerned that a beneficiary in Kaestner’s position will delay taking distributions until she moves to a State with a lower level of taxation, thereby paying less tax on the funds she ultimately receives. See id., at 40. Though this possibility is understandably troubling to the State, it is by no means certain that it will regularly come to pass. First, the power to make distributions to Kaestner or her children resides with the trustee. When and whether to make distributions is not for Kaestner to decide, and in fact the trustee may distribute funds to Kaestner while she resides in North Carolina (or deny her distributions entirely). Second, we address only the circumstances in which a beneficiary receives no trust income, has no right to demand that income, and is uncertain necessarily to receive a specific share of that income. Settlors who create trusts in the future will have to weigh the potential tax benefits of such an arrangement against the costs to the trust beneficiaries of lesser control over trust assets. In any event, mere speculation about negative consequences cannot conjure the “minimum connection” missing between North Carolina and the object of its tax. * * * For the foregoing reasons, we affirm the judgment of the Supreme Court of North Carolina. It is so ordered. Notes 1 Most notably, trusts are treated as distinct entities for federal taxation purposes. Greenough, 331 U. S., at 493; see Anderson v. Wilson, 289 U.S. 20, 26–27 (1933). 2 This trustee later was succeeded by a new trustee who was a Connecticut resident during the relevant time period. 3 The state court identified only two meetings between Kaestner and the trustee in those years, both of which took place in New York. 371 N. C. 133, 143, 814 S.E.2d 43, 50 (2018). The trustee also gave Kaestner accountings of trust assets and legal advice concerning the Trust. Id., at 135, 814 S. E. 2d, at 45. 4 The trial court also held that North Carolina’s tax violates the dormant Commerce Clause. The state appellate courts did not affirm on this basis, and we likewise do not address this challenge. 5 Because North Carolina’s tax on the Kaestner Trust does not meet Quill’s first requirement, we do not address the second. 6 The State contends that Safe Deposit is no longer good law under the more flexible approach in International Shoe Co. v. Washington, 326 U.S. 310 (1945), and also because it was premised on the view, later disregarded in Curry v. McCanless, 307 U.S. 357, 363 (1939), that the Due Process Clause forbids “double taxation.” Brief for Petitioner 27–28, and n. 12. We disagree. The aspects of the case noted here are consistent with the pragmatic approach reflected in International Shoe, and Curry distinguished Safe Deposit not because the earlier case incorrectly relied on concerns of double taxation but because the beneficiaries there had “[n]o comparable right or power” to that of the settlor in Curry. 307 U. S., at 371, n. 6. 7 Though the Court did not have occasion in Curry or Graves to explore whether a lesser degree of control by a settlor also could sustain a tax by the settlor’s domicile (and we do not today address that possibility), these cases nevertheless reinforce the logic employed by Safe Deposit, Brooke v. Norfolk, 277 U.S. 27 (1928), Maguire v. Trefry, 253 U.S. 12 (1920), and Guaranty Trust Co. v. Virginia, 305 U.S. 19 (1938), in the beneficiary context. 8 As explained below, we hold that the Kaestner Trust beneficiaries do not have the requisite relationship with the Trust property to justify the State’s tax. We do not decide what degree of possession, control, or enjoyment would be sufficient to support taxation. 9 We do not address whether a beneficiary’s ability to assign a potential interest in income from a trust would afford that beneficiary sufficient control or possession over, or enjoyment of, the property to justify taxation based solely on his or her in-state residence. 10 In light of these features, one might characterize the interests of the beneficiaries as “contingent” on the exercise of the trustee’s discretion. See Fondren v. Commissioner, 324 U.S. 18, 21 (1945) (describing “the exercise of the trustee’s discretion” as an example of a contin-gency); see also United States v. O’Malley, 383 U.S. 627, 631 (1966) (de-scribing a grantor’s power to add income to the trust principal instead of distributing it and “thereby den[y] to the beneficiaries the privilege of immediate enjoyment and conditio[n] their eventual enjoyment upon surviving the termination of the trust”); Commissioner v. Estate of Holmes, 326 U.S. 480, 487 (1946) (the termination of a contingency changes “the mere prospect or possibility, even the probability, that one may have [enjoyment of property] at some uncertain future time or perhaps not at all” into a “present substantial benefit”). We have no occasion to address, and thus reserve for another day, whether a different result would follow if the beneficiaries were certain to receive funds in the future. See, e.g., Cal. Rev. & Tax. Code Ann. §17742(a) (West 2019); Commonwealth v. Stewart, 338 Pa. 9, 16–19, 12 A.2d 444, 448–449 (1940) (upholding a tax on the equitable interest of a beneficiary who had “a right to the income from [a] trust for life”), aff’d, 312 U.S. 649 (1941). 11 Because the reasoning above resolves this case in the Trust’s favor, it is unnecessary to reach the Trust’s broader argument that the trustee’s contacts alone determine the State’s power over the Trust. Brief for Respondent 23–30. The Trust relies for this proposition on Hanson v. Denckla, 357 U.S. 235 (1958), which held that a Florida court lacked jurisdiction to adjudicate the validity of a trust agreement even though the trust settlor and most of the trust beneficiaries were domiciled in Florida. Id., at 254. The problem was that Florida law made the trustee “an indispensable party over whom the court [had to] acquire jurisdiction” before resolving a trust’s validity, and the trustee was a nonresident. Ibid. In deciding that the Florida courts lacked jurisdiction over the proceeding, the Court rejected the relevance of the trust beneficiaries’ residence and focused instead on the “acts of the trustee” himself, which the Court found insufficient to support jurisdiction. Ibid. The State counters that Hanson is inapposite because the State’s tax applies to the trust rather than to the trustee and because Hanson arose in the context of adjudicative jurisdiction rather than tax jurisdiction. Brief for Petitioner 21, n. 9; Reply Brief 16–17. There is no need to resolve the parties’ dueling interpretations of Hanson. Even if beneficiary contacts—such as residence—could be sufficient in some circumstances to support North Carolina’s power to impose this tax, the residence alone of the Kaestner Trust beneficiaries cannot do so for the reasons given above. 12 The State directs the Court’s attention to 10 other state trust taxation statutes that also look to trust beneficiaries’ in-state residency, see Brief for Petitioner 6, and n. 1, but 5 are unlike North Carolina’s because they consider beneficiary residence only in combination with other factors, see Ala. Code §40–18–1(33) (2011); Conn. Gen. Stat. §12–701(a)(4) (2019 Cum. Supp.); Mo. Rev. Stat. §§143.331(2), (3) (2016); Ohio Rev. Code Ann. §5747.01(I)(3) (Lexis Supp. 2019); R. I. Gen. Laws §44–30–5(c) (2010). Of the remaining five statutes, it is not clear that the flexible tests employed in Montana and North Dakota permit reliance on beneficiary residence alone. See Mont. Admin. Rule 42.30.101(16) (2016); N. D. Admin. Code §81–03–02.1–04(2) (2018). Similarly, Georgia’s imposition of a tax on the sole basis of beneficiary residency is disputed. See Ga. Code Ann. §48–7–22(a)(1)(C) (2017); Brief for Respondent 52, n. 20. Tennessee will be phasing out its income tax entirely by 2021. H. B. 534, 110th Gen. Assem., Reg. Sess. (2017) (enacted); see Tenn. Code Ann. §67–2–110(a) (2013). That leaves California, which (unlike North Carolina) applies its tax on the basis of beneficiary residency only where the beneficiary is not contingent. Cal. Rev. & Tax. Code Ann. §17742(a); see also n. 10, supra. 13 The Trust also raises no challenge to the practice known as throwback taxation, by which a State taxes accumulated income at the time it is actually distributed. See, e.g., Cal. Rev. & Tax. Code Ann. §17745(b).

586.US.2018_17-1094

Respondent Troy Lambert filed a class action in federal court alleging that petitioner Nutraceutical Corporation’s marketing of a dietary supplement ran afoul of California consumer-protection law. On February 20, 2015, the District Court ordered the class decertified. Pursuant to Federal Rule of Civil Procedure 23(f), Lambert had 14 days from that point to ask the Court of Appeals for permission to appeal the order. Instead, he filed a motion for reconsideration on March 12, which the District Court denied on June 24. Fourteen days later, Lambert petitioned the Court of Appeals for permission to appeal the decertification order. Nutraceutical objected that Lambert’s petition was untimely because it was filed far more than 14 days from the February 20 decertification order. The Ninth Circuit held, however, that Rule 23(f)’s deadline should be tolled under the circumstances because Lambert had “acted diligently.” On the merits, the court reversed the decertification order. Held: Rule 23(f) is not subject to equitable tolling. Pp. 3–10. (a) Rule 23(f) is properly classified as a nonjurisdictional claim-processing rule, but that does not render it malleable in every respect. Whether a rule precludes equitable tolling turns not on its jurisdictional character but rather on whether its text leaves room for such flexibility. See Carlisle v. United States, 517 U.S. 416, 421. Here, the governing rules speak directly to the issue of Rule 23(f)’s flexibility and make clear that its deadline is not subject to equitable tolling. While Federal Rule of Appellate Procedure 2 authorizes a court of appeals for good cause to “suspend any provision . . . in a particular case,” it does so with a caveat: “except as otherwise provided in Rule 26(b).” Rule 26(b), which generally authorizes extensions of time, in turn includes the carveout that a court of appeals “may not extend the time to file . . . a petition for permission to appeal”—the precise type of filing at issue here. The Rules thus express a clear intent to compel rigorous enforcement of Rule 23(f)’s deadline, even where good cause for equitable tolling might otherwise exist. Precedent confirms this understanding. See Carlisle, 517 U.S. 416, and United States v. Robinson, 361 U.S. 220. Pp. 3–6. (b) Lambert’s counterarguments do not withstand scrutiny. Lambert argues that Rule 26(b)’s prohibition on extending the time to file a petition for permission to appeal should be understood to foreclose only formal extensions granted ex ante and to leave courts free to excuse late filings on equitable grounds after the fact. But this Court has already rejected an indistinguishable argument concerning Federal Rule of Criminal Procedure 45(b) in Robinson, and Lambert offers no sound basis for reading Rule 26(b) differently. Further, the 1998 Advisory Committee Notes to Rule 23(f) speak to a court of appeals’ discretion to decide whether a particular certification decision warrants review in an interlocutory posture, not to its determination whether a petition is timely. Finally, Lambert notes that every Court of Appeals to have considered the question would accept a Rule 23(f) petition filed within 14 days of the resolution of a motion for reconsideration that was itself filed within 14 days of the original order. Although his own reconsideration motion was not filed until after the initial 14 days had run, he cites the lower courts’ handling of such cases as evidence that Rule 23(f) is amenable to tolling. However, a timely motion for reconsideration affects the antecedent issue of when the 14-day limit begins to run, not the availability of tolling. See United States v. Ibarra, 502 U.S. 1, 4, n. 2. Pp. 6–9. (c) On remand, the Court of Appeals can address other preserved arguments about whether Lambert’s Rule 23(f) petition was timely even without resort to tolling. Pp. 9–10. 870 F.3d 1170, reversed and remanded. Sotomayor, J., delivered the opinion for a unanimous Court.

To take an immediate appeal from a federal district court’s order granting or denying class certification, a party must first seek permission from the relevant court of appeals “within 14 days after the order is entered.” Fed. Rule Civ. Proc. 23(f). This case poses the question whether a court of appeals may forgive on equitable tolling grounds a failure to adhere to that deadline when the opposing party objects that the appeal was untimely. The applicable rules of procedure make clear that the answer is no. I In March 2013, respondent Troy Lambert sued petitioner Nutraceutical Corporation in federal court, alleging that its marketing of a dietary supplement ran afoul of California consumer-protection law. The District Court for the Central District of California initially permitted Lambert to litigate on behalf of a class of similarly situated consumers. On February 20, 2015, however, the District Court revisited that decision and ordered the class decertified. From that point, Lambert had 14 days to ask the Court of Appeals for the Ninth Circuit for permission to appeal the order. See Fed. Rule Civ. Proc. 23(f). Instead of filing a petition for permission to appeal, Lambert informed the District Court at a status conference on March 2 (10 days after the decertification order) that he would “want to file a motion for reconsideration” in the near future. App. to Pet. for Cert. 74. The court told Lambert to file any such motion “no later than” March 12. Id., at 76. Neither Lambert nor the District Court mentioned the possibility of an appeal. Lambert filed his motion for reconsideration, in compliance with the District Court’s schedule, on March 12 (20 days after the decertification order). The District Court denied the motion on June 24, 2015. Fourteen days later, on July 8, Lambert petitioned the Court of Appeals for permission to appeal the decertification order. Nutraceutical’s response argued that Lambert’s petition was untimely because more than four months had elapsed since the District Court’s February 20 order decertifying the class, far more than the 14 days that Federal Rule of Civil Procedure 23(f) allows. App. 41. Notwithstanding the petition’s apparent untimeliness, the Court of Appeals “deem[ed] Lambert’s petition timely” because, in its view, the Rule 23(f) deadline should be “tolled” under the circumstances. 870 F.3d 1170, 1176 (CA9 2017). The Court of Appeals reasoned that Rule 23(f )’s time limit is “non-jurisdictional, and that equitable remedies softening the deadline are therefore generally available.” Ibid. Tolling was warranted, the court concluded, because Lambert “informed the [District Court] orally of his intention to seek reconsideration” within Rule 23(f)’s 14-day window, complied with the District Court’s March 12 deadline, and “otherwise acted diligently.” Id., at 1179. On the merits, the Court of Appeals held that the District Court abused its discretion in decertifying the class. Id., at 1182–1184. It reversed the decertification order. Id., at 1184. In accepting Lambert’s petition, the Court of Appeals “recognize[d] that other circuits would likely not toll the Rule 23(f) deadline in Lambert’s case.”[1] Id., at 1179. We granted certiorari. 585 U. S. ___ (2018). II When Lambert filed his petition, Federal Rule of Civil Procedure 23(f) authorized courts of appeals to “permit an appeal from an order granting or denying class-action certification . . . if a petition for permission to appeal is filed . . . within 14 days after the order is entered.”[2] The Court of Appeals held that Rule 23(f)’s time limitation is nonjurisdictional and thus, necessarily, subject to equitable tolling. While we agree that Rule 23(f) is nonjuris- dictional, we conclude that it is not subject to equitable tolling. Because Rule 23(f)’s time limitation is found in a procedural rule, not a statute, it is properly classified as a nonjurisdictional claim-processing rule. See Hamer v. Neighborhood Housing Servs. of Chicago, 583 U. S. ___, ___ (2017) (slip op., at 8).[3] It therefore can be waived or forfeited by an opposing party. See Kontrick v. Ryan, 540 U.S. 443, 456 (2004). The mere fact that a time limit lacks jurisdictional force, however, does not render it malleable in every respect. Though subject to waiver and forfeiture, some claim-processing rules are “mandatory”—that is, they are “ ‘unalterable’ ” if properly raised by an opposing party. Manrique v. United States, 581 U. S. ___, ___ (2017) (slip op., at 4) (quoting Eberhart v. United States, 546 U.S. 12, 15 (2005) (per curiam)); see also Kontrick, 540 U. S., at 456. Rules in this mandatory camp are not susceptible of the equitable approach that the Court of Appeals applied here. Cf. Manrique, 581 U. S., at ___ (slip op., at 8) (“By definition, mandatory claim-processing rules . . . are not subject to harmless-error analysis”). Whether a rule precludes equitable tolling turns not on its jurisdictional character but rather on whether the text of the rule leaves room for such flexibility. See Carlisle v. United States, 517 U.S. 416, 421 (1996). Where the pertinent rule or rules invoked show a clear intent to preclude tolling, courts are without authority to make exceptions merely because a litigant appears to have been diligent, reasonably mistaken, or otherwise deserving. Ibid.; see Kontrick, 540 U. S., at 458; United States v. Robinson, 361 U.S. 220, 229 (1960). Courts may not disregard a properly raised procedural rule’s plain import any more than they may a statute’s. See Bank of Nova Scotia v. United States, 487 U.S. 250, 255 (1988). Here, the governing rules speak directly to the issue of Rule 23(f )’s flexibility and make clear that its deadline is not subject to equitable tolling. To begin with, Rule 23(f) itself conditions the possibility of an appeal on the filing of a petition “within 14 days” of “an order granting or denying class-action certification.” Federal Rule of Appellate Procedure 5(a)(2) likewise says that a petition for permission to appeal “must be filed within the time specified.” To be sure, the simple fact that a deadline is phrased in an unqualified manner does not necessarily establish that tolling is unavailable. See Fed. Rule App. Proc. 2 (allowing suspension of other Rules for “good cause”); Fed. Rule App. Proc. 26(b) (similar); Fed. Rule Crim. Proc. 45(b) (similar); Fed. Rule Civ. Proc. 6(b) (similar). Here, however, the Federal Rules of Appellate Procedure single out Civil Rule 23(f) for inflexible treatment. While Appellate Rule 2 authorizes a court of appeals for good cause to “suspend any provision of these rules in a particular case,” it does so with a conspicuous caveat: “except as otherwise provided in Rule 26(b).” Appellate Rule 26(b), which generally authorizes extensions of time, in turn includes this express carveout: A court of appeals “may not extend the time to file . . . a petition for permission to appeal.” Fed. Rule App. Proc. 26(b)(1). In other words, Appellate Rule 26(b) says that the deadline for the precise type of filing at issue here may not be extended. The Rules thus express a clear intent to compel rigorous enforcement of Rule 23(f )’s deadline, even where good cause for equitable tolling might otherwise exist. Precedent confirms this understanding. Carlisle, 517 U.S. 416, and Robinson, 361 U.S. 220, both centered on Federal Rule of Criminal Procedure 45(b), an extension-of-time provision that parallels Appellate Rule 26(b). Carlisle addressed Rule 45(b)’s interaction with the time limit in Criminal Rule 29 for filing a postverdict motion for judgment of acquittal. See 517 U. S., at 419–423. Rule 45(b), as it was then written, made clear that “ ‘the court may not extend the time for taking any action’ ” under Rule 29, “ ‘except to the extent and under the conditions’ ” stated therein. Id., at 421. Because the Court found the text’s purpose to foreclose acceptance of untimely motions “plain and unambiguous,” the Court held that the District Court lacked that authority. Ibid. Likewise, in Robinson, the Court held that an earlier iteration of Rule 45(b) that said “ ‘the court may not enlarge . . . the period for taking an appeal’ ” prohibited a court from accepting a notice of appeal that was untimely filed. 361 U. S., at 224 (quoting Fed. Rule Crim. Proc. 45(b)). Because Rule 23(f) is not amenable to equitable tolling, the Court of Appeals erred in accepting Lambert’s petition on those grounds. III Lambert resists the foregoing conclusion on a variety of grounds. None withstands scrutiny. Most pertinently, Lambert argues that the above-mentioned Rules are less emphatic than they first appear. Rule 26(b)’s general grant of authority to relax time limits, he notes, refers both to “extend[ing]” the time to file a petition for permission to appeal and “permit[ting]” a petition to be filed after the deadline. See Fed. Rule App. Proc. 26(b) (“For good cause, the court may extend the time prescribed by these rules . . . to perform any act, or may permit an act to be done after that time expires” (emphasis added)). Rule 26(b)(1) then prohibits courts only from “extend[ing] the time to file,” while making no further mention of “permit[ting] an act to be done after that time expires.” In Lambert’s view, Rule 26(b)(1)’s prohibition on “extend[ing] the time to file” a petition for permission to appeal therefore should be understood to foreclose only formal extensions granted ex ante, and to leave courts free to excuse late filings on equitable grounds after the fact. Whatever we would make of this contention were we writing on a blank slate, this Court has already rejected an indistinguishable argument in Robinson. There, Rule 45(b) generally authorized both “ ‘enlarg[ing]’ ” a filing period and “ ‘permit[ting] the act to be done after [its] expiration,’ ” then specifically forbade “ ‘enlarg[ing] . . . the period for taking an appeal.’ ” 361 U. S., at 223. The lower court had accepted a late filing on the ground that to do so “would not be to ‘enlarge’ the period for taking an appeal, but rather would be only to ‘permit the act to be done’ after the expiration of the specified period.” Ibid.; see also id., at 230 (Black and Douglas, JJ., dissenting). This Court reversed, explaining that acceptance of the late filing did, in fact, “enlarge” the relevant filing period. Id., at 224. Lambert offers no sound basis for reading Rule 26(b) differently, and none is apparent. Cf. Torres v. Oakland Scavenger Co., 487 U.S. 312, 315 (1988) (“Permitting courts to exercise jurisdiction over unnamed parties after the time for filing a notice of appeal has passed is equivalent to permitting courts to extend the time for filing a notice of appeal”).[4] Likewise unavailing is Lambert’s reliance on the 1998 Advisory Committee Notes to Rule 23(f), which say that a petition “may be granted or denied on the basis of any consideration that the court of appeals finds persuasive.” Advisory Committee’s Notes on 1998 Amendments to Fed. Rule Civ. Proc. 23, 28 U. S. C. App., p. 815; see also Microsoft Corp. v. Baker, 582 U. S. ___, ___–___ (2017) (slip op., at 6–8). That comment, however, speaks to a court of appeals’ discretion to decide whether a particular certification decision warrants review in an interlocutory posture, not its determination whether a petition is timely. If anything, the comment serves as a reminder that interlocutory appeal is an exception to the general rule that appellate review must await final judgment—which is fully consistent with a conclusion that Rule 23(f)’s time limit is purposefully unforgiving. See Mohawk Industries, Inc. v. Carpenter, 558 U.S. 100, 106 (2009) (“The justification for immediate appeal must . . . be sufficiently strong to overcome the usual benefits of deferring appeal until litiga- tion concludes”); cf. Baker, 582 U. S., at ___ (slip op., at 6) (describing Rule 23(f) as “the product of careful calibration”).[5] Finally, Lambert notes that every Court of Appeals to have considered the question would accept a Rule 23(f) petition filed within 14 days of the resolution of a motion for reconsideration that was itself filed within 14 days of the original order. See 870 F. 3d, at 1177–1178, n. 3 (collecting cases). Although Lambert’s own reconsideration motion was not filed until after the initial 14 days had run,[6] he cites the lower courts’ handling of such cases as evidence that Rule 23(f) is indeed amenable to tolling. He further suggests that there is no basis for relaxing the 14-day limit in one situation but not the other. Lambert’s argument relies on a mistaken premise. A timely motion for reconsideration filed within a window to appeal does not toll anything; it “renders an otherwise final decision of a district court not final” for purposes of appeal. United States v. Ibarra, 502 U.S. 1, 6 (1991) (per curiam). In other words, it affects the antecedent issue of when the 14-day limit begins to run, not the availability of tolling. See id., at 4, n. 2 (noting that this practice is not “a matter of tolling”).[7] IV Lambert devotes much of his merits brief to arguing the distinct question whether his Rule 23(f) petition was timely even without resort to tolling. First, he argues that, even if his motion for reconsideration was not filed within 14 days of the decertification order, it was filed within the time allowed (either by the Federal Rules or by the District Court at the March 2 hearing). The timeliness of that motion, Lambert contends, “cause[d] the time to appeal to run from the disposition of the reconsideration motion, not from the original order.” Brief for Respondent 8; see id., at 9–18. Alternatively, he argues that the District Court’s order denying reconsideration was itself “an order granting or denying class-action certification” under Rule 23(f ). Id., at 8–9, 19–20. The Court of Appeals did not rule on these alternative grounds, which are beyond the scope of the question presented. Mindful of our role, we will not offer the first word. See United States v. Stitt, 586 U. S. ___, ___ (2018) (slip op., at 9); Pacific Bell Telephone Co. v. linkLine Communications, Inc., 555 U.S. 438, 457 (2009). If the Court of Appeals concludes that these arguments have been preserved, it can address them in the first instance on remand. * * * The relevant Rules of Civil and Appellate Procedure clearly foreclose the flexible tolling approach on which the Court of Appeals relied to deem Lambert’s petition timely. The judgment of the Court of Appeals is therefore reversed, and the case is remanded for further proceedings consistent with this opinion. It is so ordered. Notes 1 See Nucor Corp. v. Brown, 760 F.3d 341, 343 (CA4 2014); Fleischman v. Albany Med. Ctr., 639 F.3d 28, 31 (CA2 2011); Gutierrez v. Johnson & Johnson, 523 F.3d 187, 193, and n. 5 (CA3 2008); McNamara v. Felderhof, 410 F.3d 277, 281 (CA5 2005); Gary v. Sheahan, 188 F.3d 891, 892 (CA7 1999). 2 Rule 23(f ) has since been amended and now reads, in relevant part: “A court of appeals may permit an appeal from an order granting or denying class-action certification . . . . A party must file a petition for permission to appeal . . . within 14 days after the order is entered . . . .” The difference is immaterial for purposes of this case. 3 To be sure, this Court has previously suggested that time limits for taking an appeal are “mandatory and jurisdictional.” Budinich v. Becton Dickinson & Co., 486 U.S. 196, 203 (1988). As our more recent precedents have made clear, however, this Court once used that phrase in a “ ‘less than meticulous’ ” manner. Hamer, 583 U. S., at ___ (slip op., at 9); Kontrick v. Ryan, 540 U.S. 443, 454 (2004). Those earlier statements did not necessarily signify that the rules at issue were formally “jurisdictional” as we use that term today. 4 Lambert’s other textual arguments center on rules addressed to appeals as of right. See Brief for Respondent 9–12, 17–18, 36–38 (discussing, e.g., Fed. Rules App. Proc. 3 and 4). As noted above, Rules 5 and 26 specifically address petitions for permission to appeal from nonfinal orders such as the one at issue here. Lambert’s attempts to reason by implication from other, inapposite Rules therefore bear little weight. See Manrique v. United States, 581 U. S. ___, ___–___ (2017) (slip op., at 6–7). 5 Lambert also argues that interpreting Rule 23(f ) flexibly would be consistent with the Rules’ generally equitable approach. Brief for Respondent 21—27. But that simply fails to engage with the dispositive point here: Any such background preference for flexibility has been overcome by the clear text of the relevant rules. See, e.g., Young v. United States, 535 U.S. 43, 49 (2002). 6 Lambert argues that his counsel’s statements 10 days after the District Court’s decertification order constituted an oral motion for re-consideration, but the transcript belies any such claim. See App. to Pet. for Cert. 71 (requesting only “leave to file”); id., at 74 (informing the District Court that Lambert “will want to file” a reconsideration motion). 7 We therefore have no occasion to address the effect of a motion for reconsideration filed within the 14-day window. Moreover, because nothing the District Court did misled Lambert about the appeal filing deadline, see supra, at 1–2, we similarly have no occasion to address the question whether his motion would be timely if that had occurred. See Carlisle v. United States, 517 U.S. 416, 428 (1996); id., at 435–436 (Ginsburg, J., concurring) (discussing Thompson v. INS, 375 U.S. 384, 386–387 (1964) (per curiam), and Harris Truck Lines, Inc. v. Cherry Meat Packers, Inc., 371 U.S. 215, 216–217 (1962) (per curiam)). We also have no occasion to address whether an insurmountable impediment to filing timely might compel a different result. Cf. Fed. Rule App. Proc. 26(a)(3) (addressing computation of time when “the clerk’s office is inaccessible”).

586.US.2018_17-1307

Law firm McCarthy & Holthus LLP was hired to carry out a nonjudicial foreclosure on a Colorado home owned by petitioner Dennis Obduskey. McCarthy sent Obduskey correspondence related to the foreclosure. Obduskey responded with a letter invoking a federal Fair Debt Collection Practices Act (FDCPA or Act) provision, 15 U. S. C. §1692g(b), which provides that if a consumer disputes the amount of a debt, a “debt collector” must “cease collection” until it “obtains verification of the debt” and mails a copy to the debtor. Instead, McCarthy initiated a nonjudicial foreclosure action. Obduskey sued, alleging that McCarthy failed to comply with the FDCPA’s verification procedure. The District Court dismissed on the ground that McCarthy was not a “debt collector” within the meaning of the FDCPA, and the Tenth Circuit affirmed. Held: A business engaged in no more than nonjudicial foreclosure proceedings is not a “debt collector” under the FDCPA, except for the limited purpose of §1692f(6). Pp. 6–14. (a) The FDCPA regulates “‘debt collector[s].’” §1692a(6). Relevant here, the definition of debt collector has two parts. The Act first sets out the primary definition of the term “debt collector”: a “ ‘debt collector,’ ” it says, is “any person . . . in any business the principal purpose of which is the collection of any debts, or who regularly collects or attempts to collect, directly or indirectly, debts.” Ibid. The Act then sets forth the limited-purpose definition, which states that “[f]or the purpose of section 1692f(6) . . . [the] term [debt collector] also includes any person . . . in any business the principal purpose of which is the enforcement of security interests.” It is undisputed that McCarthy is, by virtue of its role enforcing security interests, at least subject to the specific prohibitions contained in §1692f(6). But only if McCarthy falls within the primary definition’s scope do the Act’s other provisions, including those at issue here, apply. Pp. 6–7. (b) Three considerations lead to the conclusion that McCarthy is not subject to the Act’s main coverage. First, and most decisive, is the text of the Act itself. The limited purpose definition says that “[f]or the purpose of section 1692f(6)” a debt collector “also includes” a business, like McCarthy, “the principal purpose of which is the enforcement of security interests.” §1692a(6) (emphasis added). This phrase, particularly the word “also,” strongly suggests that security-interest enforcers do not fall within the scope of the primary definition. If they did, the limited purpose definition would be superfluous. By contrast, under a reading that gives effect to every word of the limited-purpose definition, the FDCPA’s debt-collector-related prohibitions (with the exception of §1692f(6)) do not apply to those who, like McCarthy, are engaged in no more than security-interest enforcement. Second, Congress may well have chosen to treat security-interest enforcement differently from ordinary debt collection in order to avoid conflicts with state nonjudicial foreclosure schemes. Third, this Court’s reading is supported by legislative history, which suggests that the Act’s present language was the product of a compromise between competing versions of the bill, one which would have totally excluded security-interest enforcement from the Act, and another which would have treated it like ordinary debt collection. Pp. 7–10. (c) Obduskey’s counterarguments are unconvincing. First, he suggests that the limited-purpose definition is not superfluous because it was meant to cover “repo men”—a category of security-interest enforcers who he says would not otherwise fall within the primary definition of “debt collector.” The limited-purpose definition, however, speaks broadly of “the enforcement of security interests,” §1692a(6), not “the enforcement of security interests in personal property.” Second, Obduskey claims that the Act’s venue provision, §1692i(a), which covers legal actions brought by “debt collectors” to enforce interests in real property, only makes sense if those who enforce security interests in real property are debt collectors subject to all prohibitions and requirements that come with that designation. The venue provision, however, does nothing to alter the definition of a debt collector. Third, Obduskey argues that McCarthy engaged in more than security-interest enforcement by sending notices that any ordinary homeowner would understand as an attempt to collect a debt. Here, however, the notices sent by McCarthy were antecedent steps required under state law to enforce a security interest, and the Act’s (partial) exclusion of “the enforcement of security interests” must also exclude the legal means required to do so. Finally, Obduskey fears that this Court’s decision will permit creditors and their agents to engage in a host of abusive practices forbidden by the Act. But the Court must enforce the statute that Congress enacted, and Congress is free expand the FDCPA’s reach if it wishes. Pp. 10–14. 879 F.3d 1216, affirmed. Breyer, J., delivered the opinion for a unanimous Court. Sotomayor, J., filed a concurring opinion.

The Fair Debt Collection Practices Act regulates “ ‘debt collector[s].’ ” 15 U. S. C. §1692a(6); see 91Stat. 874, 15 U. S. C. §1692 et seq. A “ ‘debt collector,’ ” the Act says, is “any person . . . in any business the principal purpose of which is the collection of any debts, or who regularly collects or attempts to collect, directly or indirectly, debts.” §1692a(6). This definition, however, goes on to say that “[f]or the purpose of section 1692f(6)” (a separate provision of the Act), “[the] term [debt collector] also includes any person . . . in any business the principal purpose of which is the enforcement of security interests.” Ibid. The question before us concerns this last sentence. Does it mean that one principally involved in “the enforcement of security interests” is not a debt collector (except “[f ]or the purpose of section 1692f(6)”)? If so, numerous other provisions of the Act do not apply. Or does it simply reinforce the fact that those principally involved in the enforcement of security interests are subject to §1692f(6) in addition to the Act’s other provisions? In our view, the last sentence does (with its §1692f(6) exception) place those whose “principal purpose . . . is the enforcement of security interests” outside the scope of the primary “debt collector” definition, §1692a(6), where the business is engaged in no more than the kind of security-interest enforcement at issue here—nonjudicial foreclosure proceedings. I A When a person buys a home, he or she usually borrows money from a lending institution, such as a bank. The resulting debt is backed up by a “mortgage”—a security interest in the property designed to protect the creditor’s investment. Restatement (Third) of Property: Mortgages §1.1 (1996) (Restatement). (In some States, this security interest is known as a “deed of trust,” though for present purposes the difference is immaterial. See generally ibid.) The loan likely requires the homeowner to make monthly payments. And if the homeowner defaults, the mortgage entitles the creditor to pursue foreclosure, which is “the process in which property securing a mortgage is sold to pay off the loan balance due.” 2 B. Dunaway, Law of Distressed Real Estate §15:1 (2018) (Dunaway). Every State provides some form of judicial foreclosure: a legal action initiated by a creditor in which a court supervises sale of the property and distribution of the proceeds. Id., §16:1. These procedures offer various protections for homeowners, such as the right to notice and to protest the amount a creditor says is owed. Id., §§16:17, 16:20; Restatement §8.2. And in the event that the foreclosure sale does not yield the full amount due, a creditor pursuing a judicial foreclosure may sometimes obtain a deficiency judgment, that is, a judgment against the homeowner for the unpaid balance of a debt. National Consumer Law Center (NCLC), Foreclosures and Mortgage Servicing §§12.3.1–2 (5th ed. 2014). About half the States also provide for what is known as nonjudicial foreclosure, where notice to the parties and sale of the property occur outside court supervision. 2 Dunaway §17:1. Under Colorado’s form of nonjudicial foreclosure, at issue here, a creditor (or more likely its agent) must first mail the homeowner certain preliminary information, including the telephone number for the Colorado foreclosure hotline. Colo. Rev. Stat. §38–38–102.5(2) (2018). Thirty days later, the creditor may file a “notice of election and demand” with a state official called a “public trustee.” §38–38–101. The public trustee records this notice and mails a copy, alongside other materials, to the homeowner. §§38–38–102, 38–38–103. These materials give the homeowner information about the balance of the loan, the homeowner’s right to cure the default, and the time and place of the foreclosure sale. §§38–38–101(4), 38–38–103. Assuming the debtor does not cure the default or declare bankruptcy, the creditor may then seek an order from a state court authorizing the sale. Colo. Rule Civ. Proc. 120 (2018); see Colo. Rev. Stat. §38–38–105. (Given this measure of court involvement, Colorado’s “nonjudicial” foreclosure process is something of a hybrid, though no party claims these features transform Colorado’s nonjudicial scheme into a judicial one.) In court, the homeowner may contest the creditor’s right to sell the property, and a hearing will be held to determine whether the sale should go forward. Colo. Rules Civ. Proc. 120(c), (d). If the court gives its approval, the public trustee may then sell the property at a public auction, though a homeowner may avoid a sale altogether by curing the default up until noon on the day before. Colo. Rev. Stat. §§38–38–110, 38–38–104(VI)(b). If the sale goes forward and the house sells for more than the amount owed, any profits go first to lienholders and then to the homeowner. §38–38–111. If the house sells for less than what is owed, the creditor cannot hold the homeowner liable for the balance due unless it files a separate action in court and obtains a deficiency judgment. See §38–38–106(6); Bank of America v. Kosovich, 878 P.2d 65, 66 (Colo. App. 1994). Other States likewise prevent creditors from obtaining deficiency judgments in nonjudicial foreclosure proceedings. Restatement §8.2. And in some States, pursuing nonjudicial foreclosure bars or curtails a creditor’s ability to obtain a deficiency judgment altogether. NCLC, Foreclosures and Mortgage Servicing §12.3.2. B In 2007, petitioner Dennis Obduskey bought a home in Colorado with a $329,940 loan secured by the property. About two years later, Obduskey defaulted. In 2014, Wells Fargo Bank, N. A., hired a law firm, McCarthy & Holthus LLP, the respondent here, to act as its agent in carrying out a nonjudicial foreclosure. According to the complaint, McCarthy first mailed Obduskey a letter that said it had been “instructed to commence foreclosure” against the property, disclosed the amount outstanding on the loan, and identified the creditor, Wells Fargo. App. 37–38; see id., at 23. The letter purported to provide notice “[p]ursuant to, and in compliance with,” both the Fair Debt Collection Practices Act (FDCPA) and Colorado law. Id., at 37. (The parties seem not to dispute that this and other correspondence from McCarthy was required under state law. Because that is a question of Colorado law not briefed by the parties before us nor passed on by the courts below, we proceed along the same assumption.) Obduskey responded with a letter invoking §1692g(b) of the FDCPA, which provides that if a con- sumer disputes the amount of a debt, a “debt collector” must “cease collection” until it “obtains verification of the debt” and mails a copy to the debtor. Yet, Obduskey alleges, McCarthy neither ceased collecting on the debt nor provided verification. App. 22–23. Instead, the firm initiated a nonjudicial foreclosure action by filing a notice of election and demand with the county public trustee. Ibid.; see id., at 39–41. The notice stated the amount due and advised that the public trustee would “sell [the] property for the purpose of paying the indebtedness.” Id., at 40. Obduskey then filed a lawsuit in federal court alleging that the firm had violated the FDCPA by, among other things, failing to comply with the verification procedure. Id., at 29. The District Court dismissed the suit on the ground that the law firm was not a “debt collector” within the meaning of the Act, so the relevant Act requirements did not apply. Obduskey v. Wells Fargo, 2016 WL 4091174, *3 (D Colo., July 19, 2016). On appeal, the Court of Appeals for the Tenth Circuit affirmed the dismissal, concluding that the “mere act of enforcing a security interest through a non-judicial foreclosure proceeding does not fall under” the Act. Obduskey v. Wells Fargo, 879 F.3d 1216, 1223 (2018). Obduskey then petitioned for certiorari. In light of different views among the Circuits about application of the FDCPA to nonjudicial foreclosure proceedings, we granted the petition. Compare ibid. and Vien-Phuong Thi Ho v. ReconTrust Co., NA, 858 F.3d 568, 573 (CA9 2016) (holding that an entity whose only role is the enforcement of security interests is not a debt collector under the Act), with Kaymark v. Bank of America, N. A., 783 F.3d 168, 179 (CA3 2015) (holding that such an entity is a debt collector for the purpose of all the Act’s requirements), Glazer v. Chase Home Fin. LLC, 704 F.3d 453, 461 (CA6 2013) (same), and Wilson v. Draper & Goldberg, P. L. L. C., 443 F.3d 373, 376 (CA4 2006) (same). II A The FDCPA’s definitional section, 15 U. S. C. §1692a, defines a “debt” as: “any obligation or alleged obligation of a consumer to pay money arising out of a transaction in which the money, property, insurance, or services which are the subject of the transaction are primarily for personal, family, or household purposes.” §1692a(5) (emphasis added). The Act then sets out the definition of the term “debt collector.” §1692a(6). The first sentence of the relevant paragraph, which we shall call the primary definition, says that the term “debt collector”: “means any person . . . in any business the principal purpose of which is the collection of any debts, or who regularly collects or attempts to collect, directly or indirectly, debts owed or asserted to be owed or due another.” Ibid. The third sentence, however, provides what we shall call the limited-purpose definition: “For the purpose of section 1692f(6) [the] term [debt collector] also includes any person . . . in any business the principal purpose of which is the enforcement of security interests.” Ibid. The subsection to which the limited-purpose definition refers, §1692f(6), prohibits a “debt collector” from: “Taking or threatening to take any nonjudicial action to effect dispossession or disablement of property if— “(A) there is no present right to possession of the property . . . ; “(B) there is no present intention to take possession of the property; or “(C) the property is exempt by law from such dispossession or disablement.” The rest of the Act imposes myriad other requirements on debt collectors. For example, debt collectors may not use or threaten violence, or make repetitive annoying phone calls. §1692d. Nor can debt collectors make false, deceptive, or misleading representations in connection with a debt, like misstating a debt’s “character, amount, or legal status.” §1692e. And, as we have mentioned, if a consumer disputes the amount of a debt, a debt collector must “cease collection” until it “obtains verification of the debt” and mails a copy to the debtor. §1692g(b). No one here disputes that McCarthy is, by virtue of its role enforcing security interests, at least subject to the specific prohibitions contained in §1692f(6). The question is whether other provisions of the Act apply. And they do if, but only if, McCarthy falls within the scope of the Act’s primary definition of “debt collector.” B Three considerations lead us to conclude that McCarthy is not subject to the main coverage of the Act. First, and most decisive, is the text of the Act itself. As a preliminary matter, we concede that if the FDCPA contained only the primary definition, a business engaged in nonjudicial foreclosure proceedings would qualify as a debt collector for all purposes. We have explained that a home loan is an obligation to pay money, and the purpose of a mortgage is to secure that obligation. See supra, at 2. Foreclosure, in turn, is “the process in which property securing a mortgage is sold to pay off the loan balance due.” 2 Dunaway §15:1. In other words, foreclosure is a means of collecting a debt. And a business pursuing nonjudicial foreclosures would, under the capacious language of the Act’s primary definition, be one that “regularly collects or attempts to collect, directly or indirectly, debts.” §1692a(6). It is true that, as McCarthy points out, nonjudicial foreclosure does not seek “a payment of money from the debtor” but rather from sale of the property itself. Brief for Respondent 17 (emphasis added). But nothing in the primary definition requires that payment on a debt come “from a debtor.” The statute speaks simply of the “collection of any debts . . . owed or due.” §1692a(6). Moreover, the provision sweeps in both “direc[t]” and “indirec[t]” debt collection. Ibid. So, even if nonjudicial foreclosure were not a direct attempt to collect a debt, because it aims to collect on a consumer’s obligation by way of enforcing a security interest, it would be an indirect attempt to collect a debt. The Act does not, however, contain only the primary definition. And the limited-purpose definition poses a serious, indeed an insurmountable, obstacle to subjecting McCarthy to the main coverage of the Act. It says that “[f]or the purpose of section 1692f(6)” a debt collector “also includes” a business, like McCarthy, “the principal purpose of which is the enforcement of security interests.” §1692a(6) (emphasis added). This phrase, particularly the word “also,” strongly suggests that one who does no more than enforce security interests does not fall within the scope of the general definition. Otherwise why add this sentence at all? It is logically, but not practically, possible that Congress simply wanted to emphasize that the definition of “debt collector” includes those engaged in the enforcement of security interests. But why then would Congress have used the word “also”? And if security-interest enforcers are covered by the primary definition, why would Congress have needed to say anything special about §1692f(6)? After all, §1692f(6), just like all the provisions applicable to debt collectors, would have already applied to those who enforce security interests. The reference to §1692f(6) would on this view be superfluous, and we “generally presum[e] that statutes do not contain surplusage.” Arlington Central School Dist. Bd. of Ed. v. Murphy, 548 U.S. 291, 299, n. 1 (2006). By contrast, giving effect to every word of the limited-purpose definition narrows the primary definition, so that the debt-collector-related prohibitions of the FDCPA (with the exception of §1692f(6)) do not apply to those who, like McCarthy, are engaged in no more than security-interest enforcement. Second, we think Congress may well have chosen to treat security-interest enforcement differently from ordinary debt collection in order to avoid conflicts with state nonjudicial foreclosure schemes. As Colorado’s law makes clear, supra, at 3–4, state nonjudicial foreclosure laws provide various protections designed to prevent sharp collection practices and to protect homeowners, see 2 Dunaway §17:1. And some features of these laws are in tension with aspects of the Act. For example, the FDCPA broadly limits debt collectors from communicating with third parties “in connection with the collection of any debt.” §1692c(b). If this rule were applied to nonjudicial foreclosure proceedings, then advertising a foreclosure sale—an essential element of such schemes—might run afoul of the FDCPA. Given that a core purpose of publicizing a sale is to attract bidders, ensure that the sale price is fair, and thereby protect the borrower from further liability, the result would hardly benefit debtors. See 2 Dunaway §17:4. To be sure, it may be possible to resolve these conflicts without great harm to either the Act or state foreclosure schemes. See Heintz v. Jenkins, 514 U.S. 291, 296–297 (1995) (observing that the FDCPA’s protections may contain certain “implici[t] exception[s]”). But it is also possible, in light of the language it employed, that Congress wanted to avoid the risk of such conflicts altogether. Third, for those of us who use legislative history to help interpret statutes, the history of the FDCPA supports our reading. When drafting the bill, Congress considered a version that would have subjected security-interest enforcers to the full coverage of the Act. That version defined a debt collector as “any person who engages in any business the principal purpose of which is the collection of any debt or enforcement of security interests.” S. 918, 95th Cong., 1st Sess., §803(f ) (1977) (emphasis added). A different version of the bill, however, would have totally excluded from the Act’s coverage “any person who enforces or attempts to enforce a security interest in real or personal property.” S. 1130, 95th Cong., 1st Sess., §802(8)(E) (1977). Given these conflicting proposals, the Act’s present language has all the earmarks of a compromise: The prohibitions contained in §1692f(6) will cover security-interest enforcers, while the other “debt collector” provisions of the Act will not. These considerations convince us that, but for §1692f(6), those who engage in only nonjudicial foreclosure proceedings are not debt collectors within the meaning of the Act. III Obduskey makes several arguments to the contrary. But, on balance, we do not find them determinative. First, Obduskey acknowledges that unless the limited-purpose definition is superfluous, it must make some kind of security-interest enforcer a “debt collector” who would not otherwise fall within the primary definition. Reply Brief 11–13. But, according to Obduskey, “repo men”—those who seize automobiles and other personal property in response to nonpayment—fit the bill. See Black’s Law Dictionary 1493 (10th ed. 2014) (explaining that “repo” is short for “repossession,” which means “retaking property; esp., a seller’s retaking of goods sold on credit when the buyer has failed to pay for them”). This is so, he says, because repossession often entails only “limited communication” with the debtor, as when the repo man sneaks up and “tows a car in the middle of the night.” Brief for Petitioner 25–26, and n. 13. And because, according to Obduskey, the language of §1692f(6), which forbids “[t]aking or threatening to take any nonjudicial action to effect dispossession or disablement of property,” applies more naturally to the seizure of personal property than to nonjudicial foreclosure. (Emphasis added.) But we do not see why that is so. The limited-purpose provision speaks broadly of “the enforcement of security interests,” §1692a(6), not “the enforcement of security interests in personal property”; if Congress meant to cover only the repo man, it could have said so. Moreover, Obduskey’s theory fails to save the limited-purpose definition from superfluity. As we have just discussed, supra, at 7–8, if the Act contained only the primary definition, enforcement of a security interest would at least be an indirect collection of a debt. The same may well be true of repo activity, a form of security-interest enforcement, as the point of repossessing property that secures a debt is to collect some or all of the value of the defaulted debt. And while Obduskey argues that the language of §1692f(6) fits more comfortably with repossession of personal property than nonjudicial foreclosure, we think it at least plausible that “threatening” to foreclose on a consumer’s home without having legal entitlement to do so is the kind of “nonjudicial action” without “present right to possession” prohibited by that section. §1692f(6)(A). (We need not, however, here decide precisely what conduct runs afoul of §1692f(6).) We are also unmoved by Obduskey’s argument that repossession would not fall under the primary definition because it generally involves only limited communication with the debtor. For one thing, while some of the FDCPA’s substantive protections apply where there has been a “communicat[ion]” with a consumer, see, e.g., §1692c, the primary definition of debt collector turns on the “collection of . . . debts,” without express reference to communication, §1692a(6). For another, while Obduskey imagines a silent repo man striking in the dead of night, state law often requires communication with a debtor during the repossession process, such as notifying a consumer of a sale. NCLC, Repossessions §10.4 (9th ed. 2017). Second, Obduskey points to the Act’s venue provision, 15 U. S. C. §1692i(a), which states that “[a]ny debt collector who brings any legal action on a debt against any consumer shall . . . in the case of an action to enforce an interest in real property securing the consumer’s obligation, bring such action only in a judicial district” where the “property is located.” (Emphasis added.) This provision, he says, makes clear that a person who judicially enforces a real-property-related security interest is a debt collector; hence, a person who nonjudicially enforces such an interest must also be a debt collector. Indeed, he adds, this subsection “only makes sense” if those who enforce secu- rity interests in real property are debt collectors subject to all prohibitions and requirements that come with that designation. Brief for Petitioner 21. This argument, however, makes too much of too little. To begin with, the venue section has no direct application in this case, for here we consider nonjudicial foreclosure. And whether those who judicially enforce mortgages fall within the scope of the primary definition is a question we can leave for another day. See 879 F. 3d, at 1221–1222 (noting that the availability of a deficiency judgment is a potentially relevant distinction between judicial and nonjudicial foreclosures). More to the point, the venue provision does nothing to alter the definition of a debt collector. Rather, it applies whenever a “debt collector” brings a “legal action . . . to enforce an interest in real property.” §1692i(a)(1). In other words, the provision anticipates that a debt collector can bring a judicial action respecting real property, but it nowhere says that an entity is a debt collector because it brings such an action. Obduskey suggests that under our interpretation this provision will capture a null set. We think not. A business that qualifies as a debt collector based on other activities (say, because it “regularly collects or attempts to collect” unsecured credit card debts, §1692a(6)) would have to comply with the venue provi- sion if it also filed “an action to enforce an interest in real property,” §1692i(a)(1). Here, however, the only basis alleged for concluding that McCarthy is a debt collector under the Act is its role in nonjudicial foreclosure proceedings. Third, Obduskey argues that even if “simply enforcing a security interest” falls outside the primary definition, McCarthy engaged in more than security-interest enforcement by sending notices that any ordinary homeowner would understand as an attempt to collect a debt backed up by the threat of foreclosure. Brief for Petitioner 15–16; see Reply Brief 13. We do not doubt the gravity of a letter informing a homeowner that she may lose her home unless she pays her outstanding debts. But here we assume that the notices sent by McCarthy were antecedent steps required under state law to enforce a security interest. See supra, at 4. Indeed, every nonjudicial foreclosure scheme of which we are aware involves notices to the homeowner. See 2 Dunaway §17:4 (describing state procedures concerning notice of sale). And because he who wills the ends must will the necessary means, we think the Act’s (partial) exclusion of “the enforcement of security interests” must also exclude the legal means required to do so. This is not to suggest that pursuing nonjudicial foreclosure is a license to engage in abusive debt collection practices like repetitive nighttime phone calls; enforcing a security interest does not grant an actor blanket immunity from the Act. But given that we here confront only steps required by state law, we need not consider what other conduct (related to, but not required for, enforcement of a security interest) might transform a security-interest enforcer into a debt collector subject to the main coverage of the Act. Finally, Obduskey fears that our decision will open a loophole, permitting creditors and their agents to engage in a host of abusive practices forbidden by the Act. States, however, can and do guard against such practices, for example, by requiring notices, review by state officials such as the public trustee, and limited court supervision. See supra, at 3–4, 9. Congress may think these state protections adequate, or it may choose to expand the reach of the FDCPA. Regardless, for the reasons we have given, we believe that the statute exempts entities engaged in no more than the “enforcement of security interests” from the lion’s share of its prohibitions. And we must enforce the statute that Congress enacted. For these reasons, the judgment of the Court of Appeals is Affirmed.

587.US.2018_18-389

Respondent Brian Newton worked for petitioner Parker Drilling Management Services on drilling platforms off the California coast. Newton was paid for his time on duty but not for his time on standby, during which he could not leave the platform. Newton filed a class action in state court, alleging, as relevant here, that California’s minimum-wage and overtime laws required Parker to compensate him for his standby time. Parker removed the action to Federal District Court. The parties agreed that Parker’s platforms were subject to the Outer Continental Shelf Lands Act (OCSLA), which provides that all law on the Outer Continental Shelf (OCS) is federal law, administered by federal officials; denies States any interest in or jurisdiction over the OCS; and deems the adjacent State’s laws to be federal law only “[t]o the extent that they are applicable and not inconsistent with” other federal law, 43 U. S. C. §1333(a)(2)(A). The District Court concluded that the state laws relevant here should not be applied as federal law on the OCS because the Fair Labor Standards Act of 1938 (FLSA), a comprehensive federal wage-and-hour scheme, left no significant gap in federal law for state law to fill. It thus granted Parker judgment on the pleadings. The Ninth Circuit vacated and remanded. It held that state law is “applicable” under the OCSLA if it pertains to the subject matter at issue, a standard satisfied by California wage-and-hour laws. It also held that those state laws were not “inconsistent” with federal law because they were not incompatible with the federal scheme. Held: 1. Where federal law addresses the relevant issue, state law is not adopted as surrogate federal law on the OCS. Pp. 3–14. (a) After this Court held that the Federal Government has exclusive jurisdiction over the entire continental shelf, see, e.g., United States v. Louisiana, 339 U.S. 699, 705, Congress enacted the Submerged Lands Act, which ceded certain offshore lands to the coastal States, and passed the OCSLA, which affirmed the Federal Government’s exclusive control over the OCS. Pp. 3–4. (b) Newton argues that state law is “applicable” on the OCS whenever it pertains to the subject matter at issue and that it is “inconsistent” only if it would be pre-empted under ordinary pre-emption principles. Parker counters that state law is not “applicable” absent a gap in federal law that needs to be filled and that state law can be “inconsistent” with federal law even if it is possible to satisfy both sets of laws. Parker’s approach is more persuasive. This Court reads the statute’s words “ ‘in their context and with a view to their place in the overall statutory scheme.’ ” Roberts v. Sea-Land Services, Inc., 566 U.S. 93, 101. The Court’s pre-OCSLA decisions made clear that federal law controlled the OCS in every respect, and the OCSLA reaffirmed that role. Taken together, the OCSLA’s provisions convincingly show that state laws can be “applicable and not inconsistent” with federal law under §1333(a)(2)(A) only if federal law does not address the relevant issue. The OCSLA makes apparent “that federal law is ‘exclusive’ . . . and that state law is adopted only as surrogate federal law.” Rodrigue v. Aetna Casualty & Surety Co., 395 U.S. 352, 357. It borrows only certain state laws, which are then declared to be federal law and administered by federal officials. It would thus make little sense to treat the OCS as a mere extension of the adjacent State, where state law applies unless it conflicts with federal law. That type of pre-emption analysis applies only where overlapping, dual state and federal jurisdiction makes it necessary to decide which law takes precedence. But federal law is the only law on the OCS and there is no overlapping state and federal jurisdiction, so the reference to “not inconsistent” state laws presents only the question whether federal law has already addressed the relevant issue. If so, state law on the issue is inapplicable. Pp. 5–8. (c) This interpretation is supported by several other considerations. Pp. 8–14. (1) Newton’s interpretation—that the choice-of-law question on the OCS is the same as it would be in an adjacent State—would deprive much of the OCSLA of any import, violating the “ ‘cardinal principle’ of interpretation that courts ‘must give effect, if possible, to every clause and word of a statute.’ ” Loughrin v. United States, 573 U.S. 351, 358. Pp. 8–9. (2) This Court’s interpretation is consistent with the federal-enclave model and the historical development of the statute. The OCSLA treats the OCS as “an upland federal enclave.” Rodrigue, supra, at 366. Generally, when an area in a State becomes a federal enclave, “only the [state] law in effect at the time of the transfer of jurisdiction continues in force” as surrogate federal law, James Stewart & Co. v. Sadrakula, 309 U.S. 94, 100, provided that the state law does not conflict with “federal policy,” Paul v. United States, 371 U.S. 245, 269. Going forward, state law presumptively does not apply to the enclave. See Sadrakula, supra, at 100. As originally enacted, the OCSLA both treated the OCS as a federal enclave and adopted only the “applicable and not inconsistent” laws of the adjacent State in effect as of the Act’s effective date. This suggests that, like the general enclave rule, the OCSLA sought to make all OCS law federal yet also “provide a sufficiently detailed legal framework to govern life” on the OCS. Shell Oil Co. v. Iowa Dept. of Revenue, 488 U.S. 19, 27. Providing a sufficient legal structure to accomplish that purpose eliminated the need to adopt new state laws. The OCSLA’s text and context thus suggest that state law is not adopted to govern the OCS where federal law is on point. The later amendment of the OCSLA to adopt state law on an ongoing basis confirms the connection between the OCSLA and the federal enclave model. Pp. 9–11. (3) This Court’s interpretation accords with precedent construing the OCSLA. In Rodrigue, supra, at 352–353; Chevron Oil Co. v. Huson, 404 U.S. 97; and Gulf Offshore Co. v. Mobil Oil Corp., 453 U.S. 473, the Court viewed the OCSLA as adopting state law to fill in federal-law gaps. Pp. 11–14. 2. Under the proper standard, some of Newton’s present claims can be resolved, though others have not been analyzed by the Ninth Circuit. Some claims are premised on the adoption of California law requiring payment for all standby time. Because federal law already addresses this issue, California law does not provide the rule of decision on the OCS. To the extent Newton’s OCS-based claims rely on that law, they necessarily fail. Likewise, to the extent his OCS-based claims rely on the adoption of California’s minimum wage, the FLSA already provides for a minimum wage, so the state minimum wage is not adopted as federal law and does not apply on the OCS. Pp. 14–15. 881 F.3d 1078 and 888 F.3d 1085, vacated and remanded. Thomas, J., delivered the opinion for a unanimous Court.

The Outer Continental Shelf Lands Act (OCSLA), 67Stat. 462, 43 U. S. C. §1331 et seq., extends federal law to the subsoil and seabed of the Outer Continental Shelf and all attachments thereon (OCS). Under the OCSLA, all law on the OCS is federal law, administered by federal officials. The OCSLA denies States any interest in or jurisdiction over the OCS, and it deems the adjacent State’s laws to be federal law “[t]o the extent that they are applicable and not inconsistent with” other federal law. §1333(a)(2)(A). The question before us is how to determine which state laws meet this requirement and therefore should be adopted as federal law. Applying familiar tools of statutory interpretation, we hold that where federal law addresses the relevant issue, state law is not adopted as surrogate federal law on the OCS. I Respondent Brian Newton worked for petitioner Parker Drilling Management Services on drilling platforms off the coast of California. Newton’s 14-day shifts involved 12 hours per day on duty and 12 hours per day on standby, during which he could not leave the platform. He was paid well above the California and federal minimum wages for his time on duty, but he was not paid for his standby time. Newton filed a class action in California state court alleging violations of several California wage-and-hour laws and related state-law claims. Among other things, Newton claimed that California’s minimum-wage and overtime laws required Parker to compensate him for the time he spent on standby. Parker removed the action to Federal District Court. The parties agreed that Parker’s platforms were subject to the OCSLA. Their disagreement centered on whether the relevant California laws were “applicable and not inconsistent” with existing federal law and thus deemed to be the applicable federal law under the OCSLA. §1333(a)(2)(A). The District Court applied Fifth Circuit precedent providing that under the OCSLA, “state law only applies to the extent it is necessary ‘to fill a significant void or gap’ in federal law.” App. to Pet. for Cert. 51 (quoting Continental Oil Co. v. London Steam-Ship Owners’ Mut. Ins. Assn., 417 F.2d 1030, 1036 (1969)). It determined that the Fair Labor Standards Act of 1938 (FLSA), 52Stat. 1060, 29 U. S. C. §201 et seq., constitutes a comprehensive federal wage-and-hour scheme and thus left no significant gap for state law to fill. Because all of Newton’s claims relied on state law, the court granted Parker judgment on the pleadings. The Ninth Circuit vacated and remanded. It first held that state law is “ ‘applicable’ ” under the OCSLA whenever it “pertain[s] to the subject matter at hand.” 881 F.3d 1078, 1090, amended and reh’g en banc denied, 888 F.3d 1085 (2018). The court found that California wage-and-hour laws satisfied this standard and turned to “the determinative question in Newton’s case”: “whether California wage and hour laws are ‘inconsistent with’ existing federal law.” 881 F. 3d, at 1093. According to the Ninth Circuit, state laws are “inconsistent” with federal law under the OCSLA only “if they are mutually incompatible, incongruous, [or] inharmonious.” Ibid. (internal quotation marks omitted). Applying that standard, the court determined that no inconsistency exists between the FLSA and California wage-and-hour law because the FLSA saving clause “explicitly permits more protective state wage and hour laws.” Id., at 1097 (citing 29 U. S. C. §218(a)). Given the disagreement between the Fifth and Ninth Circuits, we granted certiorari. 586 U. S. ___ (2019). II Before the OCSLA, coastal States and the Federal Government disputed who had the right to lease submerged lands on the continental shelf. Some coastal States even asserted jurisdiction all the way to the outer edge of the shelf. See Shell Oil Co. v. Iowa Dept. of Revenue, 488 U.S. 19, 26 (1988). The disputes eventually reached this Court, which held in a series of decisions that the Federal Government has exclusive jurisdiction over the entire continental shelf. See United States v. California, 332 U.S. 19, 38–39 (1947); United States v. Louisiana, 339 U.S. 699, 705 (1950); United States v. Texas, 339 U.S. 707, 717–718 (1950). After these decisions, Congress divided jurisdiction over the shelf. In 1953, Congress enacted the Submerged Lands Act, 67Stat. 29, 43 U. S. C. §1301 et seq., which ceded to the coastal States offshore lands within a specified distance of their coasts. A few months later, Congress passed the OCSLA, which affirmed that the Federal Government exercised exclusive control over the OCS, defined as “all submerged lands” beyond the lands reserved to the States up to the edge of the United States’ jurisdiction and control. §1331(a). Specifically, the OCSLA declares that “the subsoil and seabed of the [OCS] appertain to the United States and are subject to its jurisdiction, control, and power of disposition.” §1332(1). The OCSLA then sets forth “detailed provisions for the exercise of exclusive jurisdiction in the area and for the leasing and development of the resources of the seabed.” United States v. Maine, 420 U.S. 515, 527 (1975); see §§1334–1354. Of primary relevance here, the OCSLA defines the body of law that governs the OCS. First, in §1333(a)(1), the OCSLA extends “[t]he Constitution and laws and civil and political jurisdiction of the United States” to the OCS. Section 1333(a)(1) provides that federal law applies “to the same extent as if the [OCS] were an area of exclusive Federal jurisdiction located within a State.” Then, §1333(a)(2)(A) provides: “To the extent that they are applicable and not inconsistent with this subchapter or with other Federal laws and regulations of the Secretary now in effect or hereafter adopted, the civil and criminal laws of each adjacent State, now in effect or hereafter adopted, amended, or repealed are declared to be the law of the United States for that portion of the subsoil and seabed of the outer Continental Shelf, and artificial islands and fixed structures erected thereon, which would be within the area of the State if its boundaries were extended seaward to the outer margin of the outer Continental Shelf . . . .” Section 1333(a)(2)(A) also states that “[a]ll of such applicable laws shall be administered and enforced by the appropriate officers and courts of the United States.” Finally, §1333(a)(3) emphasizes that “[t]he provisions of this section for adoption of State law as the law of the United States shall never be interpreted as a basis for claiming any interest in or jurisdiction on behalf of any State for any purpose over” the OCS. III A The question in this case is how to interpret the OCSLA’s command that state laws be adopted as federal law on the OCS “[t]o the extent that they are applicable and not inconsistent” with other federal law. §1333(a)(2)(A). Echoing the Ninth Circuit, Newton argues that state law is “applicable” on the OCS whenever it pertains to the subject matter at issue. Newton further argues that state law is only “inconsistent” with federal law if it is incompatible with the federal scheme. In essence, Newton’s argument is that state law is “inconsistent” only if it would be pre-empted under our ordinary pre-emption principles. Parker, on the other hand, argues that state law is not “applicable” on the OCS in the absence of a gap in federal law that needs to be filled. Moreover, Parker argues that state law can be “inconsistent” with federal law even if it is possible for a party to satisfy both sets of laws. Specifically, Parker contends that, although the FLSA normally accommodates more protective state wage-and-hour laws, such laws are inconsistent with the FLSA when adopting state law as surrogate federal law because federal law would then contain two different standards. B Although this is a close question of statutory interpretation, on the whole we find Parker’s approach more persuasive because “ ‘the words of a statute must be read in their context and with a view to their place in the overall statutory scheme.’ ” Roberts v. Sea-Land Services, Inc., 566 U.S. 93, 101 (2012). That rule is particularly relevant here, as the terms “applicable” and “not inconsistent” are susceptible of interpretations that would deprive one term or the other of meaning. If Newton is right that “applicable” merely means relevant to the subject matter, then the word adds nothing to the statute, for an irrelevant law would never be “applicable” in that sense. Cf. Ransom v. FIA Card Services, N. A., 562 U.S. 61, 70 (2011) (declining to interpret the word “applicable” in such a way that Congress “could have omitted the term . . . altogether”). And if Parker is right that “applicable” means “necessary to fill a gap in federal law,” it is hard to imagine circumstances in which “not inconsistent” would add anything to the statute, for a state law would rarely be inconsistent with a federal law that leaves a gap that needs to be filled. Moreover, when the OCSLA was enacted, the term “inconsistent” could mean either “incompatible,” as Newton contends, or merely “inharmonious,” as Parker argues. Webster’s New International Dictionary 1259 (2d ed. 1953); see also Funk & Wagnalls New Standard Dictionary 1245 (1957) (“logically discrepant” or “disagreeing” and “discordant”); The New Century Dictionary 811 (1953) (“self-contradictory” or “at variance”); 5 Oxford English Dictionary 173 (1933) (“incongruous” or “not agreeing in substance, spirit, or form”). In short, the two terms standing alone do not resolve the question before us. Particularly given their indeterminacy in isolation, the terms should be read together and interpreted in light of the entire statute. See Star Athletica, L. L. C. v. Varsity Brands, Inc., 580 U. S. ___, ___ (2017) (slip op., at 6) (“ ‘[I]nterpretation of a phrase of uncertain reach is not confined to a single sentence when the text of the whole statute gives instruction as to its meaning’ ”). Our pre-OCSLA decisions made clear that the Federal Government controlled the OCS in every respect, and the OCSLA reaffirmed the central role of federal law on the OCS. See supra, at 3–4. As discussed, the OCSLA gives the Federal Government complete “jurisdiction, control, and power of disposition” over the OCS, while giving the States no “interest in or jurisdiction” over it. §§1332(1), 1333(a)(3). The statute applies federal law to the OCS “to the same extent as if the [OCS] were an area of exclusive Federal jurisdiction located within a State.” §1333(a)(1). Accordingly, the only law on the OCS is federal law, and state laws are adopted as federal law only “[t]o the extent that they are applicable and not inconsistent with” federal law. §1333(a)(2)(A). Taken together, these provisions convince us that state laws can be “applicable and not inconsistent” with federal law under §1333(a)(2)(A) only if federal law does not address the relevant issue. As we have said before, the OCSLA makes apparent “that federal law is ‘exclusive’ in its regulation of [the OCS], and that state law is adopted only as surrogate federal law.” Rodrigue v. Aetna Casualty & Surety Co., 395 U.S. 352, 357 (1969). The OCSLA extends all federal law to the OCS, and instead of also extending state law writ large, it borrows only certain state laws. These laws, in turn, are declared to be federal law and are administered by federal officials. Given the primacy of federal law on the OCS and the limited role of state law, it would make little sense to treat the OCS as a mere extension of the adjacent State, where state law applies unless it conflicts with federal law. See PLIVA, Inc. v. Mensing, 564 U.S. 604, 617–618 (2011). That type of pre-emption analysis is applicable only where the overlapping, dual jurisdiction of the Federal and State Governments makes it necessary to decide which law takes precedence. But the OCS is not, and never was, part of a State, so state law has never applied of its own force. Because federal law is the only law on the OCS, and there has never been overlapping state and federal jurisdiction there, the statute’s reference to “not inconsistent” state laws does not present the ordinary question in pre-emption cases—i.e., whether a conflict exists between federal and state law. Instead, the question is whether federal law has already addressed the relevant issue; if so, state law addressing the same issue would necessarily be inconsistent with existing federal law and cannot be adopted as surrogate federal law. Put another way, to the extent federal law applies to a particular issue, state law is inapplicable. C Apart from §1333(a)(2)’s place in the overall statutory scheme, several other considerations support our interpretation, which accords with the standard long applied by the Fifth Circuit, see Continental Oil, 417 F. 2d, at 1036–1037. First, if Newton were correct that the choice-of-law question on the OCS is the same as it would be in an adjacent State, much of the OCSLA would be unnecessary. Second, our interpretation is consistent with the federal-enclave model—a model that the OCSLA expressly invokes—and the historical development of the statute. And third, the Court’s precedents have treated the OCSLA in accord with our interpretation. 1 Under Newton’s interpretation, state law would apply unless pre-empted by federal law, meaning that the OCS would be treated essentially the same as the adjacent State. See Tr. of Oral Arg. 49. But that interpretation would render much of the OCSLA unnecessary. For example, the statute would not have needed to adopt state law as federal law or say that federal law applies on the OCS as if it “were an area of exclusive Federal jurisdiction located within a State.” §§1333(a)(1)–(2). It could have simply defined which State’s law applied on the OCS and given federal officials and courts the authority to enforce the law. And the statute would not have needed to limit state laws on the OCS to those “applicable and not inconsistent” with federal law (as Newton understands those words), for irrelevant laws never apply and federal law is always “supreme,” U. S. Const., Art. VI, cl. 2. Newton’s interpretation deprives much of the statute of any import, violating the “ ‘cardinal principle’ of interpretation that courts ‘must give effect, if possible, to every clause and word of a statute.’ ” Loughrin v. United States, 573 U.S. 351, 358 (2014). 2 Further support for our interpretation comes from the statute’s treatment of the OCS as “an area of exclusive Federal jurisdiction located within a State”—i.e., as “an upland federal enclave.” §1333(a)(1); Rodrigue, supra, at 366. It is a commonplace of statutory interpretation that “Congress legislates against the backdrop of existing law.” McQuiggin v. Perkins, 569 U.S. 383, 398, n. 3 (2013). Generally, when an area in a State becomes a federal enclave, “only the [state] law in effect at the time of the transfer of jurisdiction continues in force” as surrogate federal law. James Stewart & Co. v. Sadrakula, 309 U.S. 94, 100 (1940). Existing state law typically does not continue in force, however, to the extent it conflicts with “federal policy.” Paul v. United States, 371 U.S. 245, 269 (1963); see Chicago, R. I. & P. R. Co. v. McGlinn, 114 U.S. 542, 547 (1885). And going forward, state law presumptively does not apply to the enclave. See Sadrakula, supra, at 100; see also Paul, supra, at 268; Pacific Coast Dairy, Inc. v. Department of Agriculture of Cal., 318 U.S. 285, 294 (1943). This approach ensures “that no area however small will be without a developed legal system for private rights,” while simultaneously retaining the primacy of federal law and requiring future statutory changes to be made by Congress. Sadrakula, supra, at 100; United States v. Tax Comm’n of Miss., 412 U.S. 363, 370, n. 12 (1973).[1] The original version of the OCSLA both treated the OCS as a federal enclave and adopted only the “applicable and not inconsistent” laws of the adjacent State that were in effect as of the effective date of the Act. 43 U. S. C. §1333(a)(2) (1970 ed.); see §1333(a)(1) (1970 ed.) (deeming the OCS “an area of exclusive Federal jurisdiction located within a State”). This textual connection between the OCSLA and the federal enclave model suggests that, like the generally applicable enclave rule, the OCSLA sought to make all OCS law federal yet also “provide a sufficiently detailed legal framework to govern life” on the OCS. Shell Oil, 488 U. S., at 27. Once that framework was established, federal law (including previously adopted state law) provided a sufficient legal structure to accomplish that purpose, eliminating the need to adopt new state laws. The federal-state balance in a typical federal enclave is quite different than in a State, and that difference is all the more striking on the OCS, which was never under state control. The text and context of the OCSLA therefore suggest that state law is not adopted to govern the OCS where federal law is on point. Although Congress later amended the OCSLA to adopt state law on an ongoing basis, this amendment only confirms the connection between the OCSLA and the federal enclave model. Beginning in 1825, when “federal statutory law punished only a few crimes committed on federal enclaves,” Congress enacted several Assimilative Crimes Acts (ACAs) that “borrow[ed] state law to fill gaps in the federal criminal law” on enclaves. Lewis v. United States, 523 U.S. 155, 160 (1998); see 18 U. S. C. §13(a) (criminalizing “any act or omission which, although not made punishable by any enactment of Congress, would be punish- able if committed or omitted within the jurisdiction of the” relevant State or territory). Mirroring the general enclave rule discussed above, the first ACA was limited to state laws in existence when the Act was passed. United States v. Sharpnack, 355 U.S. 286, 291 (1958). Because of this limitation, the initial ACA “gradually lost much of its effectiveness in maintaining current conformity with state criminal laws,” and Congress eventually provided for the adoption of the state laws in effect at the time of the crime. Id., at 291–292. After this Court upheld this ongoing adoption of state criminal law against a nondelegation challenge, see id., at 294, Congress amended the OCSLA to borrow state laws “ ‘in effect or hereafter adopted, amended, or repealed.’ ” §19(f), 88Stat. 2146. At the same time, Congress left unchanged the features of the OCSLA that we have emphasized above—i.e., that the only law on the OCS is federal, and that state law is adopted only when it is “applicable and not inconsistent” with existing federal law. Thus, we do not understand the statutory amendment to alter our conclusion. If anything, this history reinforces that the OCS should be treated as an exclusive federal enclave, not an extension of a State, and that the OCSLA, like the ACAs, does not adopt state law “where there is no gap to fill.” Lewis, supra, at 163. 3 Finally, our interpretation accords with the Court’s precedents construing the OCSLA. We first interpreted the OCSLA’s choice-of-law provision in Rodrigue v. Aetna Casualty & Surety Co., where we considered whether suits brought by the families of men killed on OCS drilling rigs could proceed under only the federal Death on the High Seas Act or also under state law. 395 U. S., at 352–353. We emphasized that under the OCSLA, the body of law applicable to the OCS “was to be federal law of the United States, applying state law only as federal law and then only when not inconsistent with applicable federal law.” Id., at 355–356. We explained that “federal law, because of its limited function in a federal system, might be inadequate to cope with the full range of potential legal problems,” and that the OCSLA “supplemented gaps in the federal law with state law through the ‘adoption of State law as the law of the United States.’ ” Id., at 357 (quoting §1333(a)(3)). We reiterated that the statutory language makes it “evident” “that federal law is ‘exclusive’ ” on the OCS and that “state law could be used to fill federal voids.” Id., at 357–358. After concluding that the Death on the High Seas Act did not apply to accidents on the OCS and thus left a gap related to wrongful deaths, we held that state law provided the rule of decision. We explained that “the inapplicability of the [federal Act] removes any obstacle to the application of state law by incorporation as federal law through” the OCSLA. Id., at 366. Two years later, in Chevron Oil Co. v. Huson, 404 U.S. 97 (1971), the Court again viewed the OCSLA as adopting state law to fill in federal-law gaps. In Huson, the question was whether federal admiralty law or a state statute governed a tort action arising from an injury that occurred on the OCS. Id., at 98–99. Describing Rodrigue’s analysis, we explained that where “there exists a substantial ‘gap’ in federal law,” “state law remedies are not ‘inconsistent’ with applicable federal law.” 404 U. S., at 101. We highlighted that “state law was needed” as surrogate federal law because federal law alone did not provide “ ‘a complete body of law,’ ” which is why “Congress specified that a comprehensive body of state law should be adopted by the federal courts in the absence of existing federal law.” Id., at 103–104. In other words, the OCSLA “made clear provision for filling in the ‘gaps’ in federal law.” Id., at 104. And because Congress had decided not to apply federal admiralty law on the OCS, leaving a gap on the relevant issue, we held that it was appropriate to “absor[b]” the state law as federal law. Id., at 104, 109. In Gulf Offshore Co. v. Mobil Oil Corp., 453 U.S. 473 (1981), we once again emphasized that “[a]ll law applicable to the [OCS] is federal law” and that the “OCSLA borrows the ‘applicable and not inconsistent’ laws of the adjacent States” “to fill the substantial ‘gaps’ in the coverage of federal law.” Id., at 480. We noted that under the OCSLA, the Federal Government “retain[ed] exclusive . . . control of the administration of the [OCS],” and that state law is incorporated “to fill gaps in federal law.” Id., at 479–480, n. 7. These precedents confirm our understanding of the OCSLA. Although none decided the precise question before us, much of our prior discussion of the OCSLA would make little sense if the statute essentially treated the OCS as an extension of the adjacent State. In Rodrigue, for example, there was no question that the state law at issue pertained to the subject matter or that the relevant federal law expressly preserved state laws regulating the same subject. See 395 U. S., at 355; 46 U. S. C. §767 (1964 ed.). Under Newton’s interpretation, that should have ended the case. Yet the Court instead analyzed at length whether the federal law extended to the OCS. See 395 U. S., at 359–366. It would be odd for our decisions to focus so closely on the gap-filling role of state law under the OCSLA if, as Newton argues, the existence of a federal-law gap is irrelevant. Our consistent understanding of the OCSLA remains: All law on the OCS is federal, and state law serves a supporting role, to be adopted only where there is a gap in federal law’s coverage. In sum, the standard we adopt today is supported by the statute’s text, structure, and history, as well as our precedents. Under that standard, if a federal law addresses the issue at hand, then state law is not adopted as federal law on the OCS.[2] IV Applying this standard, some of Newton’s present claims are readily resolvable. For instance, some of his claims are premised on the adoption of California law requiring payment for all time that Newton spent on standby. See Mendiola v. CPS Security Solutions, Inc., 60 Cal. 4th 833, 842, 340 P.3d 355, 361 (2015); Cal. Lab. Code Ann. §510(a) (West 2011). But federal law already addresses this issue. See 29 CFR §785.23 (2018) (“An employee who resides on his employer’s premises on a permanent basis or for extended periods of time is not considered as working all the time he is on the premises”); see also 29 U. S. C. §207(a). Therefore, this California law does not provide the rule of decision on the OCS, and to the extent Newton’s OCS-based claims rely on that law, they necessarily fail. Likewise, to the extent Newton’s OCS-based claims rely on the adoption of the California minimum wage (currently $12), Cal. Lab. Code Ann. §1182.12(b) (West Supp. 2019), the FLSA already provides for a minimum wage, 29 U. S. C. §206(a)(1), so the California minimum wage does not apply. Newton points out that the FLSA sets a minimum wage of “not less than . . . $7.25 an hour,” ibid. (emphasis added), and does not “excuse noncompliance with any Federal or State law . . . establishing a [higher] minimum wage,” §218. But whatever the import of these provisions in an ordinary pre-emption case, they do not help Newton here, for the question under the OCSLA is whether federal law addresses the minimum wage on the OCS. It does. Therefore, the California minimum wage is not adopted as federal law and does not apply on the OCS. Newton’s other claims were not analyzed by the Court of Appeals, and the parties have provided little briefing on those claims. Moreover, the Court of Appeals held that Newton should be given leave to amend his complaint. Because we cannot finally resolve whether Parker was entitled to judgment on the pleadings, we vacate the judgment of the Court of Appeals, and the case is remanded for further proceedings consistent with this opinion. It is so ordered. Notes 1 These general rules “may be qualified in accordance with agreements reached by the respective governments.” Sadrakula, 309 U. S., at 99; see also Paul, 371 U. S., at 268 (“[A] State may not legislate with respect to a federal enclave unless it reserved the right to do so when it gave its consent to the purchase by the United States”). 2 Of course, it is conceivable that state law might be “inconsistent” with federal law for purposes of §1333(a)(2) even absent an on-point federal law. For example, federal law might contain a deliberate gap, making state law inconsistent with the federal scheme. Or, state law might be inconsistent with a federal law addressing a different issue. We do not foreclose these or other possible inconsistencies.

588.US.2018_17-1705

Petitioners (collectively PDR) produce the Physicians’ Desk Reference, which compiles information about the uses and side effects of various prescription drugs. PDR sent health care providers faxes stating that they could reserve a free copy of a new e-book version of the Reference on PDR’s website. Respondent Carlton & Harris Chiropractic, a fax recipient, brought a putative class action in Federal District Court, claiming that PDR’s fax was an “unsolicited advertisement” prohibited by the Telephone Consumer Protection Act of 1991 (Telephone Act). 47 U. S. C. §227(b)(1)(C). The District Court dismissed the case, concluding that PDR’s fax was not an “unsolicited advertisement” under the Telephone Act. The Fourth Circuit vacated the District Court’s judgment. Based on the Administrative Orders Review Act (Hobbs Act), which provides that courts of appeals have “exclusive jurisdiction to enjoin, set aside, suspend (in whole or in part), or to determine the validity of” certain “final orders of the Federal Communication Commission,” 28 U. S. C. §2342(1), the Court of Appeals held that the District Court was required to adopt the interpretation of “unsolicited advertisement” set forth in a 2006 FCC Order. Because the Court of Appeals found that the 2006 Order interpreted the term “unsolicited advertisement” to “include any offer of a free good or service,” the Court of Appeals concluded that the facts as alleged demonstrated that PDR’s fax was an unsolicited advertisement. 883 F.3d 459, 467. Held: The extent to which the 2006 FCC Order binds the lower courts may depend on the resolution of two preliminary sets of questions that were not aired before the Court of Appeals. First, is the Order the equivalent of a “legislative rule,” which is “ ‘issued by an agency pursuant to statutory authority’ ” and has the “ ‘force and effect of law’ ”? Chrysler Corp. v. Brown, 441 U.S. 281, 302–303. Or is it the equivalent of an “interpretive rule,” which simply “ ‘advis[es] the public of the agency’s construction of the statutes and rules which it administers’ ” and lacks “ ‘the force and effect of law’ ”? Perez v. Mortgage Bankers Assn., 575 U.S. 92, ___. If the Order is the equivalent of an “interpretive rule,” it may not be binding on a district court, and a district court therefore may not be required to adhere to it. Second, did PDR have a “prior” and “adequate” opportunity to seek judicial review of the Order? 5 U. S. C. §703. If the Hobbs Act’s exclusive-review provision, which requires certain challenges to FCC orders to be brought in a court of appeals “within 60 days after” the entry of the order in question, 28 U. S. C. §2344, did not afford PDR a “prior” and “adequate” opportunity for judicial review, it may be that the Administrative Procedure Act permits PDR to challenge the Order’s validity in this enforcement proceeding. The judgment of the Court of Appeals is vacated, and the case is remanded for that court to consider these preliminary issues, as well as any other related issues that may arise in the course of resolving this case. Pp. 4–6. 883 F.3d 459, vacated and remanded. Breyer, J., delivered the opinion of the Court, in which Roberts, C. J., and Ginsburg, Sotomayor, and Kagan, JJ., joined. Thomas, J., filed an opinion concurring in the judgment, in which Gorsuch, J., joined. Kavanaugh, J., filed an opinion concurring in the judgment, in which Thomas, Alito, and Gorsuch, JJ., joined.

This case concerns two federal statutes, the Telephone Consumer Protection Act of 1991 (Telephone Act) and the Administrative Orders Review Act (Hobbs Act). The first statute generally makes it unlawful for any person to send an “unsolicited advertisement” by fax. 47 U. S. C. §227(b)(1)(C). The second statute provides that the fed- eral courts of appeals have “exclusive jurisdiction to enjoin, set aside, suspend (in whole or in part), or to determine the validity of” certain “final orders of the Federal Communication Commission.” 28 U. S. C. §2342(1). In 2006, the FCC issued an Order stating that the term “unsolicited advertisement” in the Telephone Act includes certain faxes that “promote goods or services even at no cost,” including “free magazine subscriptions” and “catalogs.” 21 FCC Rcd. 3787, 3814. The question here is whether the Hobbs Act’s vesting of “exclusive jurisdiction” in the courts of appeals to “enjoin, set aside, suspend,” or “determine the validity” of FCC “final orders” means that a district court must adopt, and consequently follow, the FCC’s Order interpreting the term “unsolicited advertisement” as including certain faxes that promote “free” goods. We have found it difficult to answer this question, for the answer may depend upon the resolution of two preliminary issues. We therefore vacate the judgment of the Court of Appeals and remand this case so that the Court of Appeals can consider these preliminary issues. I Petitioners (PDR Network, PDR Distribution, and PDR Equity, collectively referred to here as PDR) produce the Physicians’ Desk Reference, a publication that compiles information about the uses and side effects of various prescription drugs. PDR makes money by charging pharmaceutical companies that wish to include their drugs in the Reference, and it distributes the Reference to health care providers for free. In 2013, PDR announced that it would publish a new e-book version of the Reference. It advertised the e-book to health care providers by sending faxes stating that providers could reserve a free copy on PDR’s website. One of the fax recipients was respondent Carlton & Harris Chiropractic, a health care practice in West Vir- ginia. It brought this putative class action against PDR in Federal District Court, claiming that PDR’s fax violated the Telephone Act. Carlton & Harris sought statutory damages on behalf of itself and other members of the class. According to Carlton & Harris, PDR’s fax was an “unsolicited advertisement” prohibited by the Telephone Act. 47 U. S. C. §227(b)(1)(C). The Act defines “unsolicited advertisement” as “any material advertising the commercial availability or quality of any property, goods, or services which is transmitted to any person without that person’s prior express invitation or permission.” §227(a)(5). This provision says nothing about goods offered for free, but it does give the FCC authority to “prescribe regulations to implement” the statute. §227(b)(2). And, as we have said, the FCC’s 2006 Order provides that fax messages that “promote goods or services even at no cost, such as free magazine subscriptions, catalogs, or free consultations or seminars, are unsolicited advertisements under the [Telephone Act’s] definition. . . . ‘[F]ree’ publications are often part of an overall marketing campaign to sell property, goods, or services.” 21 FCC Rcd., at 3814. The Order also indicates, however, that faxes “that contain only information, such as industry news articles, legislative updates, or employee benefit information, would not be prohibited.” Ibid. The Order then sets forth “factors” the FCC “will consider” when determining whether “an informational communication” that contains advertising material is an “unsolicited advertisement.” Id., at 3814, n. 187. The District Court found in PDR’s favor and dismissed the case. It concluded that PDR’s fax was not an “unsolicited advertisement” under the Telephone Act. 2016 WL 5799301 (SD W. Va., Sept. 30, 2016). The court did recognize that the FCC’s Order might be read to indicate the contrary. Id., at *3. And it also recognized that the Hobbs Act gives appellate courts, not district courts, “exclusive jurisdiction” to “determine the validity of” certain FCC “final orders.” 28 U. S. C. §2342(1). Nonetheless, the District Court concluded that neither party had challenged the Order’s validity. 2016 WL 5799301, *3. And it held that even if the Order is presumed valid, a district court is not bound to follow the FCC interpretation announced in the Order. Id., at *4. In any event, the District Court also noted that a “careful reading” of the Order showed that PDR’s fax was not an “unsolicited advertisement” even under the FCC’s interpretation of that term. Ibid. Carlton & Harris appealed to the Fourth Circuit, which vacated the District Court’s judgment. 883 F.3d 459 (2018). The Court of Appeals held that “the jurisdictional command” of the Hobbs Act—that is, the word “exclusive”—“requires a district court to apply FCC interpretations” of the Telephone Act. Id., at 466. Thus, the District Court should have adopted the interpretation of “unsolicited advertisement” set forth in the 2006 Order. Ibid. And because the Order interpreted the term “advertisement” to “include any offer of a free good or service,” id., at 467, the facts as alleged demonstrated that PDR’s fax was an unsolicited advertisement. PDR filed a petition for certiorari. We granted certiorari to consider “[w]hether the Hobbs Act required the district court in this case to accept the FCC’s legal interpretation of the Telephone Consumer Protection Act.” 586 U. S. ___ (2018). II The Hobbs Act says that an appropriate court of appeals has “exclusive jurisdiction to enjoin, set aside, suspend (in whole or in part), or to determine the validity of . . . final orders of the Federal Communication Commission made reviewable by section 402(a) of title 47.” 28 U. S. C. §2342(1); see 47 U. S. C. §402(a) (making reviewable certain “orde[rs] of the Commission under” the Communications Act, of which the Telephone Act is part). It further provides that “[a]ny party aggrieved” may bring such a challenge in the court of appeals “within 60 days after” the entry of the FCC order in question. 28 U. S. C. §2344. Here, we are asked to decide whether the Hobbs Act’s commitment of “exclusive jurisdiction” to the courts of appeals requires a district court in a private enforcement suit like this one to follow the FCC’s 2006 Order interpreting the Telephone Act. The parties in this case did not dispute below that the Order is a “final order” that falls within the scope of the Hobbs Act. 883 F. 3d, at 464, n. 1. And we assume without deciding that the Order is such a “final order.” Even so, the extent to which the Order binds the lower courts may depend on the resolution of two preliminary sets of questions that were not aired before the Court of Appeals. First, what is the legal nature of the 2006 FCC Order? In particular, is it the equivalent of a “legislative rule,” which is “ ‘issued by an agency pursuant to statutory authority’ ” and has the “ ‘force and effect of law’ ”? Chrysler Corp. v. Brown, 441 U.S. 281, 302–303 (1979) (quoting Batterton v. Francis, 432 U.S. 416, 425, n. 9 (1977)). Or is it instead the equivalent of an “interpretive rule,” which simply “ ‘advis[es] the public of the agency’s construction of the statutes and rules which it administers’ ” and lacks “ ‘the force and effect of law’ ”? Perez v. Mortgage Bankers Assn., 575 U.S. 92, ___ (2015) (slip op., at 3) (quoting Shalala v. Guernsey Memorial Hospital, 514 U.S. 87, 99 (1995)). If the relevant portion of the 2006 Order is the equivalent of an “interpretive rule,” it may not be binding on a district court, and a district court therefore may not be required to adhere to it. That may be so regardless of whether a court of appeals could have “determin[ed]” during the 60-day review period that the Order is “vali[d]” and consequently could have decided not to “enjoin, set aside, [or] suspend” it. 28 U. S. C. §2342. And that may be so no matter what degree of weight the district court ultimately gives the FCC’s interpretation of the statute under Chevron U. S. A. Inc. v. Natural Resources Defense Council, Inc., 467 U.S. 837 (1984). We say “may” because we do not definitively resolve these issues here. Second, and in any event, did PDR have a “prior” and “adequate” opportunity to seek judicial review of the Order? 5 U. S. C. §703. The Administrative Procedure Act provides that “agency action is subject to judicial review in civil or criminal proceedings for judicial enforcement” except “to the extent that [a] prior, adequate, and exclusive opportunity for judicial review is provided by law.” Ibid. (emphasis added). We believe it important to determine whether the Hobbs Act’s exclusive-review provision, which requires certain challenges to FCC final orders to be brought in a court of appeals “within 60 days after” the entry of the order in question, 28 U. S. C. §2344, afforded PDR a “prior” and “adequate” opportunity for judicial review of the Order. If the answer is “no,” it may be that the Administrative Procedure Act permits PDR to challenge the validity of the Order in this enforcement proceeding even if the Order is deemed a “legislative” rule rather than an “interpretive” rule. We again say “may” because we do not definitively decide this issue here. III As we have said many times before, we are a court of “review,” not of “first view.” Cutter v. Wilkinson, 544 U.S. 709, 718, n. 7 (2005). Because the Court of Appeals has not yet addressed the preliminary issues we have described, we vacate the judgment of the Court of Appeals and remand this case so that the Court of Appeals may consider these preliminary issues, as well as any other related issues that may arise in the course of resolving this case. It is so ordered.