Peugh v. United States, 569 U.S. ___ (2013)

NOTICE: This opinion is subject to formal revision before publication in the preliminary print of the United States Reports. Readers are requested to notify the Reporter of Decisions, Supreme Court of the United States, Washington, D. C. 20543, of any typographical or other formal errors, in order that corrections may be made before the preliminary print goes to press.

SUPREME COURT OF THE UNITED STATES

_________________

No. 12–62

_________________

MARVIN PEUGH, PETITIONER v. UNITED STATES

on writ of certiorari to the united states court of appeals for the seventh circuit

[June 10, 2013]

Justice Sotomayor delivered the opinion of the Court, except as to Part III–C.[1]*

The Constitution forbids the passage of ex post facto laws, a category that includes “[e]very law that changes the punishment, and inflicts a greater punishment, than the law annexed to the crime, when committed.” Calder v. Bull, 3 Dall. 386, 390 (1798) (emphasis deleted). The U. S. Sentencing Guidelines set forth an advisory sentencing range for each defendant convicted in federal court. We consider here whether there is an ex post facto violation when a defendant is sentenced under Guidelines promulgated after he committed his criminal acts and the new version provides a higher applicable Guidelines sentencing range than the version in place at the time of the offense. We hold that there is.

I

Petitioner Marvin Peugh and his cousin, Steven Holle- well, ran two farming-related businesses in Illinois. Grainery, Inc., bought, stored, and sold grain; Agri-Tech, Inc., provided farming services to landowners and tenants. When the Grainery began experiencing cash-flow problems, Peugh and Hollewell engaged in two fraudulent schemes. First, they obtained a series of bank loans by representing falsely the existence of contracts for future grain deliveries from Agri-Tech to the Grainery. When they failed to pay back the principal on these loans, the bank suffered losses of over $2 million. Second, they ar- tificially inflated the balances of accounts under their con- trol by “check kiting,” or writing bad checks between their accounts. This scheme allowed them to overdraw an account by $471,000. They engaged in their illicit conduct in 1999 and 2000.

When their acts were uncovered, Peugh and Hollewell were charged with nine counts of bank fraud, in violation of 18 U. S. C. §1344. While Hollewell pleaded guilty to one count of check kiting, Peugh pleaded not guilty and went to trial, where he testified that he had not intended to defraud the banks. The jury found him guilty of five counts of bank fraud and acquitted him of the remaining counts.

At sentencing, Peugh argued that the Ex Post Facto Clause required that he be sentenced under the 1998 version of the Federal Sentencing Guidelines in effect at the time of his offenses, rather than under the 2009 version in effect at the time of sentencing. The two versions yielded significantly different results for Peugh’s applicable Guidelines sentencing range. Under the 1998 Guidelines, Peugh’s base offense level was 6. United States Sentencing Commission, Guidelines Manual §2F1.1 (Nov. 1998) (USSG). Thirteen levels were added for a loss amount of over $2.5 million, ibid., and 2 levels for obstruction of justice because of Peugh’s perjury at trial, see USSG §3C1.1 (Nov. 1998). The total offense level under the 1998 Guidelines was therefore 19. As a first-time offender, Peugh was in Criminal History Category I, and so his sentencing range under the 1998 Guidelines was 30 to 37 months. USSG, ch. 5, pt. A (Nov. 1998).

The 2009 Guidelines in effect when Peugh was sentenced in May 2010 assigned more severe consequences to his acts. First, the base offense level was raised from 6 to 7 for crimes, like Peugh’s, that have a statutory maximum term of imprisonment of 20 years or more. See USSG §2B1.1 (Nov. 2009); 18 U. S. C. §1344. Second, the enhancement for a loss exceeding $2.5 million was 18, a 5-level increase from the 1998 Guidelines. USSG 2B1.1 (Nov. 2009). After adding the 2-level enhancement for obstruction of justice, Peugh’s total offense level under the 2009 Guidelines was 27. With a Criminal History Cate- gory of I, Peugh’s sentencing range rose under the 2009 Guidelines to 70 to 87 months. USSG, ch. 5, pt. A (Nov. 2009). The low end of the 2009 Guidelines range was 33 months higher than the high end of the 1998 Guidelines range.

At the sentencing hearing, the District Court rejected Peugh’s argument that applying the 2009 Guidelines vio- lated the Ex Post Facto Clause, noting that it was foreclosed by Seventh Circuit precedent. App. 30 (discussing United States v. Demaree, 459 F. 3d 791 (2006)). The District Court declined to give Peugh a downward variance, concluding that “a sentence within the [G]uideline[s] range is the most appropriate sentence in this case,” App. 100. It sentenced Peugh to 70 months’ imprisonment, ibid., the bottom of the 2009 Guidelines range.

The Seventh Circuit, in keeping with its decision in Demaree, rejected Peugh’s ex post facto claim and affirmed his conviction and sentence. 675 F. 3d 736 (2012). We granted certiorari to resolve a conflict among the Courts of Appeals over whether the Ex Post Facto Clause may be violated when a defendant is sentenced under the version of the Sentencing Guidelines in effect at the time of sentencing rather than the version in effect at the time the crime was committed, and the newer Guidelines yield a higher applicable sentencing range.[2] 568 U. S. ___ (2012). We now reverse.

II

Prior to 1984, the broad discretion of sentencing courts and parole officers had led to significant sentencing disparities among similarly situated offenders. To address this problem, Congress created the United States Sentencing Commission. Mistretta v. United States, 488 U. S. 361, 362, 366–367 (1989). The Sentencing Reform Act of 1984, 98Stat. 1987, eliminated parole in the federal system and directed the Sentencing Commission to promulgate uniform guidelines that would be binding on federal courts at sentencing. Mistretta, 488 U. S., at 367. The Commission produced the now familiar Sentencing Guidelines: a system under which a set of inputs specific to a given case (the particular characteristics of the offense and offender) yielded a predetermined output (a range of months within which the defendant could be sentenced).

In United States v. Booker, 543 U. S. 220, 244 (2005), however, this Court held that mandatory Guidelines ran afoul of the Sixth Amendment by allowing judges to find facts that increased the penalty for a crime beyond “the maximum authorized by the facts established by a plea of guilty or a jury verdict.” See also Apprendi v. New Jersey, 530 U. S. 466, 490 (2000). The appropriate remedy for this violation, the Court determined, was to strike those portions of the Sentencing Reform Act that rendered the Guidelines mandatory. Booker, 543 U. S., at 245–258. Under the resulting scheme, a district court is still required to consult the Guidelines. See id., at 259–260, 264; 18 U. S. C. §3553(a)(4)(A). But the Guidelines are no longer binding, and the district court must consider all of the factors set forth in §3553(a) to guide its discretion at sentencing, see Booker, 543 U. S., at 259–260, 264. The Booker remedy, “while not the system Congress enacted,” was designed to “continue to move sentencing in Congress’ preferred direction, helping to avoid excessive sentencing disparities while maintaining flexibility sufficient to individualize sentences where necessary.” Id., at 264–265.

Our subsequent decisions have clarified the role that the Guidelines play in sentencing procedures, both at the district court level and when sentences are reviewed on appeal. First, “a district court should begin all sentencing proceedings by correctly calculating the applicable Guidelines range. As a matter of administration and to secure nationwide consistency, the Guidelines should be the starting point and the initial benchmark.” Gall v. United States, 552 U. S. 38, 49 (2007) (citation omitted). The district court must then consider the arguments of the par- ties and the factors set forth in §3553(a). Id., at 49–50. The district court “may not presume that the Guidelines range is reasonable,” id., at 50; and it “may in appropriate cases impose a non-Guidelines sentence based on dis- agreement with the [Sentencing] Commission’s views,” Pep- per v. United States, 562 U. S. ___, ___ (2011) (slip op., at 23) (citing Kimbrough v. United States, 552 U. S. 85, 109–110 (2007)). The district court must explain the basis for its chosen sentence on the record. Gall, 552 U. S., at 50. “[A] major departure [from the Guidelines] should be supported by a more significant justification than a minor one.” Ibid.

On appeal, the district court’s sentence is reviewed for reasonableness under an abuse-of-discretion standard. See id., at 51; Booker, 543 U. S., at 261–264. Failure to calculate the correct Guidelines range constitutes procedural error, as does treating the Guidelines as mandatory. Gall, 552 U. S., at 51. The court of appeals may, but is not required to, presume that a within-Guidelines sentence is reasonable. Rita v. United States, 551 U. S. 338, 347 (2007). The reviewing court may not apply a heightened standard of review or a presumption of unreasonableness to sentences outside the Guidelines range, although it “will, of course, take into account the totality of the circumstances, including the extent of any variance from the Guidelines range.” Gall, 552 U. S., at 49–51. We have in- dicated that “a district court’s decision to vary from the advisory Guidelines may attract greatest respect when” it is based on the particular facts of a case. Kimbrough, 552 U. S., at 109.[3] Overall, this system “requires a court to give respectful consideration to the Guidelines,” but it “permits the court to tailor the sentence in light of other statutory concerns as well.” Id., at 101 (internal quotation marks omitted).

Under 18 U. S. C. §3553(a)(4)(A)(ii), district courts are instructed to apply the Sentencing Guidelines issued by the United States Sentencing Commission that are “in effect on the date the defendant is sentenced.” The Sentencing Guidelines reiterate that statutory directive, with the proviso that “[i]f the Court determines that use of the Guidelines Manual in effect on the date that the defendant is sentenced would violate the [E]x [P]ost [F]acto [C]lause of the United States Constitution, the court shall use the Guidelines Manual in effect on the date that the offense of conviction was committed.” USSG §§1B1.11(a), (b)(1) (Nov. 2012). Whether the Ex Post Facto Clause was violated by the use of the more onerous Guidelines in effect on the date of Peugh’s sentencing is the question pre- sented here.

III

A

The Constitution prohibits both federal and state governments from enacting any “ex post facto Law.” Art. I, §9, cl. 3; Art. I, §10. The phrase “ ‘ex post facto law’ was a term of art with an established meaning at the time of the framing.” Collins v. Youngblood, 497 U. S. 37, 41 (1990). In Calder v. Bull, Justice Chase reviewed the definition that the term had acquired in English common law:

“1st. Every law that makes an action done before the passing of the law, and which was innocent when done, criminal; and punishes such action. 2d. Every law that aggravates a crime, or makes it greater than it was, when committed. 3d. Every law that changes the punishment, and inflicts a greater punishment, than the law annexed to the crime, when committed. 4th. Every law that alters the legal rules of evidence, and receives less, or different, testimony, than the law required at the time of the commission of the offence, in order to convict the offender.” 3 Dall., at 390 (emphasis deleted).

See also Carmell v. Texas, 529 U. S. 513, 521–525 (2000) (discussing Calder v. Bull and the common-law understanding of the term). Building on Justice Chase’s formulation of what constitutes an “ex post facto Law,” our cases “have not attempted to precisely delimit the scope of this Latin phrase, but have instead given it substance by an accretion of case law.” Dobbert v. Florida, 432 U. S. 282, 292 (1977).

At issue here is Calder’s third category of ex post facto laws, those that “chang[e] the punishment, and inflic[t] a greater punishment, than the law annexed to the crime, when committed.” 3 Dall., at 390. Peugh’s claim is that the Clause was violated because the 2009 Guidelines call for a greater punishment than attached to bank fraud in 2000, when his crimes were completed. The Government counters that because the more punitive Guidelines applied at Peugh’s sentencing were only advisory, there was no ex post facto problem.

Each of the parties can point to prior decisions of this Court that lend support to its view. On the one hand, we have never accepted the proposition that a law must increase the maximum sentence for which a defendant is eligible in order to violate the Ex Post Facto Clause. See, e.g., Lindsey v. Washington, 301 U. S. 397 (1937). More- over, the fact that the sentencing authority exercises some measure of discretion will also not defeat an ex post facto claim. See Garner v. Jones, 529 U. S. 244, 253 (2000). On the other hand, we have made it clear that mere speculation or conjecture that a change in law will retrospectively increase the punishment for a crime will not suffice to establish a violation of the Ex Post Facto Clause. See California Dept. of Corrections v. Morales, 514 U. S. 499, 509 (1995). The touchstone of this Court’s inquiry is whether a given change in law presents a “ ‘sufficient risk of increasing the measure of punishment attached to the covered crimes.’ ” Garner, 529 U. S., at 250 (quoting Morales, 514 U. S., at 509). The question when a change in law creates such a risk is “a matter of degree”; the test cannot be reduced to a “single formula.” Id., at 509 (internal quotation marks omitted).[4]

B

The most relevant of our prior decisions for assessing whether the requisite degree of risk is present here is Miller v. Florida, 482 U. S. 423 (1987), in which this Court considered an ex post facto challenge to a sentencing guidelines scheme implemented by the State of Florida. Under Florida’s system, a calculation under the guidelines yielded a presumptive sentencing range. Id., at 426. This range was assumed to be appropriate, and the sentencing judge had discretion to fix a sentence within that range “ ‘without the requirement of a written explanation.’ ” Ibid. (quoting Fla. Rule Crim. Proc. 3.701(d)(8) (1983)). If the court wished to depart from the guidelines range, however, it was required to give “clear and convincing reasons in writing for doing so.” 482 U. S., at 426. A within-guidelines sentence was unreviewable; a non-guidelines sentence was subject to appellate review. Ibid.

The petitioner in Miller had been sentenced under new guidelines that yielded a higher sentencing range than the guidelines that had been in place at the time of his crime, and he had received a sentence at the top of the new range. Ibid. This Court found an ex post facto violation. We emphasized that in order to impose the petitioner’s sentence under the pre-existing guidelines, the sentenc- ing judge would have been required to provide clear and convincing reasons in writing for the departure, and the sentence would then have been reviewable on appeal. Id., at 432. In contrast, because the sentence imposed was within the new guidelines range, it required no explanation and was unreviewable. Id., at 432–433. The fact that Florida’s guidelines “create[d] a high hurdle that must be cleared before discretion can be exercised” was sufficient to render the changed guidelines an ex post facto law. Id., at 435.

Miller thus establishes that applying amended sentencing guidelines that increase a defendant’s recommended sentence can violate the Ex Post Facto Clause, notwithstanding the fact that sentencing courts possess discretion to deviate from the recommended sentencing range. The sentencing scheme in Miller was designed to channel sen- tences for similarly situated offenders into a specified range. Its reason-giving requirements and standards of appellate review meant that while variation was possible, it was burdensome; and so in the ordinary case, a defendant would receive a within-guidelines sentence. Under the Florida system, therefore, an increase in the guidelines range applicable to an offender created a significant risk that he would receive a higher sentence.[5] The same principles apply here.

The post-Booker federal sentencing scheme aims to achieve uniformity by ensuring that sentencing decisions are anchored by the Guidelines and that they remain a meaningful benchmark through the process of appellate review. See Kimbrough, 552 U. S., at 107. As we have described, “district courts must begin their analysis with the Guidelines and remain cognizant of them throughout the sentencing process.” Gall, 552 U. S., at 50, n. 6 (emphasis added). Failing to calculate the correct Guidelines range constitutes procedural error. Id., at 51. A district court contemplating a non-Guidelines sentence “must con- sider the extent of the deviation and ensure that the justification is sufficiently compelling to support the degree of the variance.” Id., at 50. See also Pepper, 562 U. S., at ___ (Breyer, J., concurring in part and concurring in judgment) (slip op., at 1) (“[T]he law permits the court to disregard the Guidelines only where it is ‘reasonable’ for a court to do so” (citing Booker, 543 U. S., at 261–262)).

These requirements mean that “[i]n the usual sentencing, . . . the judge will use the Guidelines range as the starting point in the analysis and impose a sentence within the range.” Freeman v. United States, 564 U. S. ___, ___ (2011) (plurality opinion) (slip op., at 5). 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. (emphasis added). See also id., at ___ (Sotomayor, J., concurring in judgment) (slip op., at 2) (stating that outside the context of a Federal Rule of Criminal Procedure 11(c)(1)(C) plea agreement, “in the normal course the district judge’s calculation of the Guidelines range applicable to the charged offenses will serve as the basis for the term of imprisonment imposed”). That a district court may ultimately sentence a given defendant outside the Guidelines range does not deprive the Guidelines of force as the framework for sentencing. Indeed, the rule that an incorrect Guidelines calculation is procedural error en- sures that they remain the starting point for every sentencing calculation in the federal system.

Similarly, appellate review for reasonableness using the Guidelines as a benchmark helps promote uniformity by “tend[ing] to iron out sentencing differences.” Booker, 543 U. S., at 263. Courts of appeals may presume a within-Guidelines sentence is reasonable, see Rita, 551 U. S., at 347, and they may further “consider the extent of the deviation” from the Guidelines as part of their reason- ableness review, Gall, 552 U. S., at 51. As in Miller, then, the post-Booker sentencing regime puts in place proce- dural “hurdle[s]” that, in practice, make the imposition of a non-Guidelines sentence less likely. See 482 U. S., at 435.

This is a more difficult case than Miller, because there are relevant differences between Florida’s sentencing scheme and the current federal sentencing regime. The Florida Legislature had made a within-guidelines sentence unreviewable; whereas in the federal system, the courts of appeals may—but are not required to—presume that a within-Guidelines sentence is reasonable. And under Florida’s scheme, a sentencing court departing from the guideline range was required to provide “clear and convincing” reasons for the departure; whereas this Court has not, post-Booker, applied such an exacting across-the-board standard of review to variances. Rather, we have held that a district court varying from the Federal Guidelines should provide an explanation adequate to the extent of the departure. See Gall, 552 U. S., at 51.

But contrary to the arguments advanced by the Government and Justice Thomas’ dissent (hereinafter dissent), see Brief for United States 23–24; post, at 5–6, these differences are not dispositive. Although the federal system’s procedural rules establish gentler checks on the sentencing court’s discretion than Florida’s did, they nevertheless impose a series of requirements on sentencing courts that cabin the exercise of that discretion. Common sense indicates that in general, this system will steer district courts to more within-Guidelines sentences.

Peugh points to considerable empirical evidence indicating that the Sentencing Guidelines have the intended effect of influencing the sentences imposed by judges. Even 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. See United States Sentencing Commission, 2011 Sourcebook of Federal Sentencing Statistics, p. 63 (Figure G) (16th ed.) (USSC). In less than one-fifth of cases since 2007 have district courts imposed above- or below-Guidelines sentences absent a Government motion. See ibid. See also Baron-Evans & Stith, Booker Rules, 160 U. Pa. L. Rev. 1631, 1677 (2012). Moreover, the Sentencing Commission’s data indicate that when a Guidelines range moves up or down, offenders’ sentences move with it. See USSC, Final Quarterly Data Report, FY 2012, p. 32 (Figure C); USSC, Report on the Continuing Impact of United States v. Booker on Federal Sentencing, Pt. A, pp. 60–68 (2012).[6]

The federal system adopts procedural measures intended to make the Guidelines the lodestone of sentencing. A retrospective increase in the Guidelines range applicable to a defendant creates a sufficient risk of a higher sentence to constitute an ex post facto violation.

C

Our holding today is consistent with basic principles of fairness that animate the Ex Post Facto Clause. The Framers considered ex post facto laws to be “contrary to the first principles of the social compact and to every principle of sound legislation.” The Federalist No. 44, p. 282 (C. Rossiter ed. 1961) (J. Madison). The Clause ensures that individuals have fair warning of applicable laws and guards against vindictive legislative action. See Weaver v. Graham, 450 U. S. 24, 28–29 (1981); see also post, at 11–13. Even where these concerns are not directly implicated, however, the Clause also safeguards “a fundamental fairness interest . . . in having the government abide by the rules of law it establishes to govern the cir-cumstances under which it can deprive a person of his or her liberty or life.” Carmell, 529 U. S., at 533.

The Sentencing Guidelines represent the Federal Government’s authoritative view of the appropriate sentences for specific crimes. When Peugh committed his crime, the recommended sentence was 30 to 37 months. When he was sentenced, it was 70 to 87 months. “[T]he purpose and effect of the change in [the Guidelines calculation] was to increase the rates and length of incarceration for [fraud].” Miller, 482 U. S., at 431 (citing Florida Bar: Amendment to Rules of Criminal Procedure (3.701, 3.988—Sentencing Guidelines), 451 So. 2d 824, 824, n. (1984) (per curiam) (internal quotation marks and alterations omitted)). Such a retrospective increase in the measure of punishment raises clear ex post facto concerns. We have previously recognized, for instance, that a defendant charged with an increased punishment for his crime is likely to feel enhanced pressure to plead guilty. See Carmell, 529 U. S., at 534, n. 24; Weaver, 450 U. S., at 32. This pressure does not disappear simply because the Guidelines range is advisory; the defendant will be aware that the range is intended to, and usually does, exert controlling influence on the sentence that the court will impose.

We are therefore not persuaded by the argument advanced by the Government and also suggested by the dissent that the animating principles of the Ex Post Facto Clause are not implicated by this case. While the Government argues that the Sentencing Commission is insulated from legislative interference, see Brief for United States 42–44, our precedents make clear that the coverage of the Ex Post Facto Clause is not limited to legislative acts, see Garner, 529 U. S., at 247, 257 (recognizing that a change in a parole board’s rules could, given an adequate showing, run afoul of the Ex Post Facto Clause). It is true that we held, in Irizarry v. United States, 553 U. S. 708, 713–714 (2008), that a defendant does not have an “expectation subject to due process protection” that he will be sentenced within the Guidelines range. But, contrary to the dissent’s view, see post, at 11–13, the Ex Post Facto Clause does not merely protect reliance interests. It also reflects principles of “fundamental justice.” Carmell, 529 U. S., at 531.[7]

IV

The Government’s principal argument that there is no constitutional violation in this case is that the Sentencing Guidelines lack sufficient legal effect to attain the sta- tus of a “law” within the meaning of the Ex Post Facto Clause. Whereas the pre-Booker Guidelines “ha[d] the force and effect of laws,” Booker, 543 U. S., at 234, the post-Booker Guidelines, the Government contends, have lost that status due to their advisory nature. The dissent echoes this argument. Post, at 1–3, 6–8.

The distinction that the Government draws is necessar- ily a fine one, because our precedents firmly establish that changes in law need not bind a sentencing authority in order to violate the Ex Post Facto Clause. So, for example, a law can run afoul of the Clause even if it does not alter the statutory maximum punishment attached to a crime. In Lindsey v. Washington, 301 U. S. 397, this Court considered an ex post facto challenge to a Washington law altering the statutory penalty for grand larceny from a range of 0 to 15 years’ imprisonment to a mandatory term of 15 years’ imprisonment. Although the upper boundary of the sentencing court’s power to punish remained unchanged, it was enough that the petitioners were “deprived of all opportunity to receive a sentence which would give them freedom from custody and control prior to the expiration of the 15-year term.” Id., at 402 (emphasis added).

In addition, our cases make clear that “[t]he presence of discretion does not displace the protections of the Ex Post Facto Clause.” Garner, 529 U. S., at 253. In a series of cases, for example, this Court has considered the validity under the Ex Post Facto Clause of state laws altering the terms on which discretionary parole or early release was available to prisoners. See Garner, 529 U. S. 244; Morales, 514 U. S. 499; Weaver, 450 U. S. 24. Although these cases reached differing conclusions with respect to whether there was an ex post facto violation, in none of them did we indicate that the mere fact that the prisoner was not guaranteed parole but rather received it at the will of the parole board was fatal to his claim. See Garner, 529 U. S., at 253; Morales, 514 U. S., at 508–510, and n. 6; Weaver, 450 U. S., at 30–31.

The Government does not challenge these holdings but rather argues, in essence, that the Guidelines are too much like guideposts and not enough like fences to give rise to an ex post facto violation. It contrasts the Sentenc- ing Guidelines with the Florida system at issue in Miller, which, the Government indicates, really did place “a substantial legislative constraint on the judge’s exercise of sentencing discretion.” Brief for United States 21. But as we have explained at length, the difference between the federal system and the scheme the Court considered in Miller is one in degree, not in kind. The Florida system did not achieve its “binding legal effect,” Brief for United States 22, by mandating a within-guidelines sentence in every case. Rather, it achieved its “binding legal effect” through a set of procedural rules and standards for appellate review that, in combination, encouraged district courts to sentence within the guidelines. See Miller, 482 U. S., at 432–433. We have detailed all of the ways in which the federal sentencing regime after Booker does the same.[8]

The Government elaborates its argument that the Sentencing Guidelines do not have adequate legal force to constitute an ex post facto violation by reviewing the various features of the post-Booker sentencing regime that, in its view, tend to render the Guidelines purely advisory. As we have noted, district courts may not presume that a within-Guidelines sentence is reasonable; they may “in appropriate cases impose a non-Guidelines sentence based on a disagreement with the Commission’s views,” Pepper, 562 U. S., at ___ (slip op., at 23); and all sentences are reviewed under a deferential abuse-of-discretion standard. See supra, at 5–6.

While the Government accurately describes several attributes of federal sentencing after Booker, the conclusion it draws by isolating these features of the system is ultimately not supportable. On the Government’s account, the Guidelines are just one among many persuasive sources a sentencing court can consult, no different from a “policy paper.” Brief for United States 28. The Government’s argument fails to acknowledge, however, that district courts are not required to consult any policy paper in order to avoid reversible procedural error; nor must they “consider the extent of [their] deviation” from a given policy paper and “ensure that the justification is suffi- ciently compelling to support the degree of the variance,” Gall, 552 U. S., at 50. Courts of appeals, in turn, are not permitted to presume that a sentence that comports with a particular policy paper is reasonable; nor do courts of appeals, in considering whether the district court’s sentence was reasonable, weigh the extent of any departure from a given policy paper in determining whether the district court abused its discretion, see id., at 51. It is simply not the case that the Sentencing Guidelines are merely a volume that the district court reads with academic interest in the course of sentencing.

Of course, as the Government and the dissent point out, notwithstanding a rule that retrospective application of a higher Guidelines range violates the Ex Post Facto Clause, sentencing courts will be free to give careful consideration to the current version of the Guidelines as representing the most recent views of the agency charged by Congress with developing sentencing policy. See post, at 8 (citing Demaree, 459 F. 3d, at 795). But this does not render our holding “purely semantic.” Id., at 795. District courts must begin their sentencing analysis with the Guidelines in effect at the time of the offense and use them to calculate the sentencing range correctly; and those Guidelines will anchor both the district court’s discretion and the appellate review process in all of the ways we have described. The newer Guidelines, meanwhile, will have the status of one of many reasons a district court might give for deviating from the older Guidelines, a status that is simply not equivalent for ex post facto purposes.

Finally, the Government contends that a rule that the Ex Post Facto Clause is violated by the application of an increased Guidelines range would be in tension with this Court’s post-Booker cases and, indeed, would “largely undo . . . the Booker remedy” for the Sixth Amendment violation found there. Brief for United States 35. If the Guidelines are binding enough to trigger an ex post facto violation, the argument goes, then they must be binding enough to trigger a Sixth Amendment violation as well. The Government’s argument assumes that the Sixth Amendment and the Ex Post Facto Clause share a common boundary; that only where judge-found facts are the basis of a higher sentence in a manner that raises Sixth Amendment concerns can a set of sentencing rules be sufficiently determinate to run afoul of the Ex Post Facto Clause. But the Sixth Amendment and Ex Post Facto Clause inquiries are analytically distinct. Our Sixth Amendment cases have focused on when a given finding of fact is required to make a defendant legally eligible for a more severe penalty. Our ex post facto cases, in contrast, have focused on whether a change in law creates a “significant risk” of a higher sentence; here, whether a sentence in conformity with the new Guidelines is substantially likely. The Booker remedy was designed, and has been subsequently calibrated, to exploit precisely this distinction: it is intended to promote sentencing uniformity while avoiding a Sixth Amendment violation. In light of the statistics invoked by petitioner, see supra, at 12–13, it appears so far to be achieving this balance. Nothing that we say today “undo[es]” the holdings of Booker, Rita, Gall, Kimbrough, or our other recent sentencing cases.

*  *  *

The arguments put forward by the Government and the dissent cannot unseat the conclusion that Peugh’s case falls within Calder’s third category of ex post facto violations. “[T]he Ex Post Facto Clause forbids the [government] to enhance the measure of punishment by altering the substantive ‘formula’ used to calculate the applicable sentencing range.” Morales, 514 U. S., at 505. That is precisely what the amended Guidelines did here. Doing so created a “significant risk” of a higher sentence for Peugh, Garner, 529 U. S., at 251, and offended “one of the principal interests that the Ex Post Facto Clause was designed to serve, fundamental justice,” Carmell, 529 U. S., at 531.[9] For these reasons, we reverse the judgment of the Seventh Circuit and remand the case for further proceedings consistent with this opinion.

It is so ordered.

SUPREME COURT OF THE UNITED STATES

_________________

No. 12–62

_________________

MARVIN PEUGH, PETITIONER v. UNITED STATES

on writ of certiorari to the united states court of appeals for the seventh circuit

[June 10, 2013]

Justice Alito, with whom Justice Scalia joins, dissenting.

I agree with Justice Thomas that retroactive application of amended advisory Guidelines does not violate the Ex Post Facto Clause under our “sufficient risk” test. See California Dept. of Corrections v. Morales, 514 U. S. 499, 509 (1995). I do not have occasion in this case to reconsider that test’s merits or its relation to the original understanding of the Clause.

SUPREME COURT OF THE UNITED STATES

_________________

No. 12–62

_________________

MARVIN PEUGH, PETITIONER v. UNITED STATES

on writ of certiorari to the united states court of appeals for the seventh circuit

[June 10, 2013]

Justice Thomas, with whom the Chief Justice, Jus-tice Scalia, and Justice Alito join as to Parts I and II–C, dissenting.

The Constitution prohibits Congress from passing ex post facto laws. Art. I, §9, cl. 3. The retroactive application of the 2009 Guidelines did not alter the punishment affixed to petitioner’s crime and does not violate this pro-scription. I would affirm the Seventh Circuit’s decision denying petitioner’s ex post facto claim. Therefore, I respectfully dissent.

I

It is well established that an ex post facto law includes “[e]very law that changes the punishment, and inflicts a greater punishment, than the law annexed to the crime, when committed.” Calder v. Bull, 3 Dall. 386, 390 (1798) (opinion of Chase, J.). Under our precedents, the relevant inquiry for determining whether a law “inflicts a greater punishment,” is whether the “retroactive application of the change in [the] law created ‘a sufficient risk of increasing the measure of punishment attached to the covered crimes.’ ” Garner v. Jones, 529 U. S. 244, 250 (2000) (quoting California Dept. of Corrections v. Morales, 514 U. S. 499, 509 (1995)). The retroactive application of subsequently amended Guidelines does not create a “sufficient risk” of increasing a defendant’s punishment for two reasons. First, the Guidelines do not constrain the discretion of district courts and, thus, have no legal effect on a defendant’s sentence. Second, to the extent that the amended Guidelines create a risk that a defendant might receive a harsher punishment, that risk results from the Guidelines’ persuasive force, not any legal effect. The Guidelines help district judges to impose sentences that comply with §3553(a). The risk of an increased sentence is, in essence, the risk of a more accurate sentence—i.e., a sentence more in line with the statutory scheme’s penological goals. Guideline changes that help district courts achieve such pre-existing statutory sentencing goals do not create a risk of an increased sentence cognizable under the Ex Post Facto Clause. We have never held that government action violates the Ex Post Facto Clause when it merely influences the exercise of the sentencing judge’s discretion.

A

The Federal Sentencing Guidelines do not constrain the discretion of district courts. As we have said repeatedly, the Guidelines are “advisory.” United States v. Booker, 543 U. S. 220, 245 (2005) (remedial opinion for the Court by Breyer, J.). For this reason, district courts may not “presume” that a within-Guidelines sentence is appropriate. Gall v. United States, 552 U. S. 38, 50 (2007); see also Nelson v. United States, 555 U. S. 350, 352 (2009) (per curiam) (the Guidelines range is “not to be presumed reasonable”); Rita v. United States, 551 U. S. 338, 351 (2007) (“[T]he sentencing court does not enjoy the benefit of a legal presumption that the Guidelines sentence should apply”). Rather, district courts must “make an individualized assessment” of the appropriate sentence “based on the facts presented.” Gall, supra, at 50. Moreover, a district court may freely depart from the range recommended by the Guidelines based not only on “an individualized determination that [the Guidelines] yield an excessive sentence in a particular case,” but also based on “policy dis-agreement” with the Guidelines themselves. Spears v. United States, 555 U. S. 261, 264 (2009) (per curiam); see Pepper v. United States, 562 U. S. ___, ___ (2011) (slip op., at 23) (“[O]ur post-Booker decisions make clear that a district court may in appropriate cases impose a non-Guidelines sentence based on a disagreement with the Commission’s views”).

It is true that a district judge who “decides that an outside-Guidelines sentence is warranted” must “ensure that the justification is sufficiently compelling to support the degree of the variance” and that “a major departure should be supported by a more significant justification than a minor one.” Gall, 552 U. S., at 50. This does not demonstrate that the Guidelines constrain the judge’s dis-cretion, but rather comports with the notion that an explanation is essential for “meaningful appellate review.” Ibid. And, when a district court departs from the recommended range, the court of appeals may not presume that such a sentence is unreasonable. Id., at 47; id., at 41 (“[C]ourts of appeals must review all sentences—whether inside, just outside, or significantly outside the Guidelines range—under a deferential abuse-of-discretion standard”). While “[t]he applicable guideline [may] nudg[e] [the sentencing judge] toward the sentencing range,” “his freedom to impose a reasonable sentence outside the range is unfettered.” United States v. Demaree, 459 F. 3d 791, 795 (CA7 2006).

None of petitioner’s arguments to the contrary is persuasive. Petitioner first contends that the Guidelines constrain district courts’ discretion because improperly calculating the applicable guidelines is reversible error. Brief for Petitioner 20–21, and n. 7; 18 U. S. C. §3742(f); Cf. Gall, 552 U. S., at 51. This argument is a non sequitur. The Guidelines can only serve their advisory purpose if district courts consider the “range established” by the Guidelines, §3553(a)(4). For this reason, district courts must “begin all sentencing proceedings by correctly calculating the applicable Guidelines range.” Id., at 49. But the fact that courts must give due consideration to the recommendation expressed in the correct Guidelines does not mean that the Guidelines constrain the district court’s discretion to impose an appropriate sentence; it simply means that district courts must consider the correct vari-ables before exercising their discretion.

Petitioner next argues that the Guidelines limit district court discretion because sentences falling outside the Guidelines are more likely to be reversed for substantive unreasonableness. Brief for Petitioner 25. I doubt, however, that reversal is a likely outcome when a district judge can justify his sentence based on agreement with either of two Guidelines—the old or the new. If a dis- trict court calculated the sentencing range under the new Guidelines but sentenced the defendant to a below-Guidelines sentence that fell within the range provided by the old Guidelines, it would be difficult to label such a sentence “substantively unreasonable.” To do so would cast doubt on every within-Guidelines sentence issued un-der the old Guidelines. Similarly, it is hard to imagine that a court of appeals would reverse a sentence for substantive unreasonableness if it was above the range of the Guidelines in effect at the time of the offense but fell within the range of the most up-to-date Guidelines. This case provides an apt example. After considering all of the §3553(a)(2) factors, the District Court concluded that a sentence within the amended Guidelines range was “the most appropriate sentence in this case.” App. 100. The same sentence would undoubtedly be upheld on appeal if the District Court, on remand, once again determined that a sentence within the amended Guidelines was appropriate in light of all the facts. The essential point is that once new Guidelines have been promulgated, reasonableness review does not meaningfully constrain the discretion of district courts to sentence offenders within either of the two ranges.

The majority argues that our opinion in Miller v. Flor-ida, 482 U. S. 423 (1987), supports its conclusion that retroactive application of advisory Guidelines violates the Ex Post Facto Clause. See ante, at 9–10. But Miller leads to the opposite conclusion. There, Florida superimposed narrowed presumptive sentencing ranges on the statutory sentencing ranges for particular crimes. 482 U. S., at 425–426. If a judge imposed a sentence within that narrower presumptive range, he did not need to give a written explanation of his reasons for selecting that sentence, and the sentence was not subject to appellate review. Ibid. If the judge imposed a sentence outside the presumptive range, however, he was required to provide “ ‘clear and convincing reasons,’ ” id., at 426 (quoting Fla. Rule Crim. Proc. 3.701(d)(11) (1983)), based “on facts proved beyond a reasonable doubt,” that justified the departure, 482 U. S., at 432. In concluding that retroactive application of this scheme violated the Ex Post Facto Clause, we reasoned that the Florida guidelines did not “simply provide flexible ‘guideposts’ for use in the exercise of discretion: instead, they create[d] a high hurdle that must be cleared before discretion c[ould] be exercised.” Id., at 435.

The Court cites Miller for the proposition “that applying amended sentencing guidelines that increase a defendant’s recommended sentence can violate the Ex Post Facto Clause, notwithstanding the fact that sentencing courts possess discretion to deviate from the recommended sentencing range.” Ante, at 10. But that claim is not supported by Miller. The guidelines in Miller violated the Ex Post Facto Clause precisely because they constrained the sentencing judge’s discretion.

The Federal Guidelines, by contrast, do no such thing. Indeed, our post-Booker opinions have made abundantly clear that the Guidelines do not create a “high hurdle”— or any hurdle at all—“that must be cleared before discretion can be exercised.” Miller, 482 U. S., at 435. Rather, the Guidelines are “flexible ‘guideposts’ ” which inform the district courts’ discretion. Ibid. Accordingly, their retroactive application cannot constitute a violation of the Ex Post Facto Clause.

B

Notwithstanding the discretion district courts have to impose appropriate sentences anywhere within the statutory range, Guidelines do “influenc[e] the sentences imposed by judges.” Ante, at 12. But, the Guidelines do this by helping district courts impose sentences that are consistent with §3553(a). It is difficult to see how an advi-sory Guideline, designed to lead courts to impose sentences more in line with fixed statutory objectives, could ever constitute an ex post facto violation. But that is exactly what the Court concludes.

District courts are charged with imposing sentences that are “ ‘sufficient, but not greater than necessary’ to comply with the sentencing purposes set forth in” §3553(a). Pepper, 562 U. S., at ___ (slip op., at 13) (quoting §3553(a)). The district court’s task is to impose sentences that reflect the punitive goals of justice, deterrence, protection of the public, and rehabilitation. 18 U. S. C. §3553(a)(2). While easily stated, this goal is difficult to achieve. Enter the Sentencing Guidelines.

The Sentencing Reform Act of 1984 instructs the Sentencing Commission to promulgate Guidelines that reflect the “same basic §3553(a) objectives” that district courts must consider. Rita, 551 U. S., at 348; see also 28 U. S. C. §991(b)(1)(A). In crafting the Guidelines, the Commission began with “an empirical examination of 10,000 presentence reports setting forth what judges had done in the past.” Rita, supra, at 349 (citing United States Sentencing Commission, Guidelines Manual §1A1.1, comment., n. 3 (Nov. 2006) (USSG)). The Commission then “modif[ied] and adjust[ed] past practice in the interests of greater rationality, avoiding inconsistency, complying with congressional instructions, and the like.” Rita, supra, at 349. While an individual judge has limited experience upon which to draw, the Commission “has the capacity . . . to base its determinations on empirical data and national experience, guided by a professional staff with appropriate expertise.” Kimbrough v. United States, 552 U. S. 85, 109 (2007) (internal quotation marks omitted). And the Commission updates the Guidelines regularly as new information becomes available. It consults with “prosecutors, defenders, law enforcement groups, civil liberties associations, experts in penology, and others,” to ensure that the Guidelines continue to further §3553(a)’s goals. Rita, supra, at 350; see also Booker, 543 U. S., at 263 (noting that the Commission would “modify its Guidelines in light of what it learns, thereby encouraging what it finds to be better sentencing practices”).

In light of this extensive study, amendments to the Guidelines should produce sentencing ranges that better comport with the §3553(a) factors. If the Commission has fulfilled its mission of recommending sentences that are generally consistent with §3553(a)(2), then sentences should fall within the Guidelines range most of the time. This, in part, explains why within-Guidelines sentences are presumed, on appeal, to reflect a “discretionary decision” by the district court that “accords with the Commission’s view.” Rita, supra, at 351.

Again, this case furnishes a ready example. Prior to pe-titioner’s sentencing, Congress directed the Commission “to consider” whether fraud guidelines were “ ‘sufficient to deter and punish’ ” particular offenses, in light of increases to statutory maximum penalties for certain fraud crimes other than bank fraud. USSG App. C, Amdt. 653 (Reason for Amendment) (effective Nov. 1, 2003) (quoting White-Collar Crime Penalty Enhancement Act of 2002, §905(b)(2), 116Stat. 805). This produced amended Guidelines, which were based on the Commission’s further assessment of “economic crime issues over a number of years.” USSG App. C, Amdt. 617 (Reason for Amendment) (effective Nov. 1, 2001). With an amended Guidelines sentencing range, the District Court concluded that a within-Guidelines sentence was “the most appropriate sentence.” App. 100. Neither the statutory sentencing range nor §3553(a) changed between the time of petitioner’s offense and sentencing. Thus, it is quite incorrect to say that reliance on information reflected in the amended Guidelines violated the Ex Post Facto Clause.

This is underscored by the fact that even the Court’s holding—which requires district courts to calculate the Guidelines range in effect at the time of the offense—will not eliminate the “risk” of a higher sentence. The district judge remains free to consider the range produced by the amended Guidelines. See Demaree, 459 F. 3d, at 795 (“A judge is certainly entitled to take advice from the Sentencing Commission”). Thus, the mere fact that new Guidelines have been promulgated creates some risk of an increased sentence, even if district courts are required to calculate the Guidelines in effect at the time of the offense. Petitioner has presented no evidence indicating what portion of the risk of an increased sentence flows from the retroactive application of the amended Guidelines and what portion flows from their very existence. In the absence of such evidence, even if I agreed that advisory Guidelines could be ex post facto laws, which I do not, I would not find the “risk” of an increased sentence created by the retroactive application of the Guidelines to be “suf-ficient” for ex post facto purposes.

II

Today’s opinion also demonstrates the unworkability of our ex post facto jurisprudence. Under our current precedent, whenever a change in the law creates a “risk” of an increased sentence, we must determine whether the risk is “sufficient,” see Morales, 514 U. S., at 509, or sufficiently “ ‘significant,’ ” see ante, at 19, to violate the Ex Post Facto Clause. Our analysis under that test has devolved into little more than an exercise in judicial intuition. I would return to the original meaning of the Clause as stated in Justice Chase’s classic Calder formulation, under which laws of this sort are ex post facto only when they retro-actively increase the punishment “annexed to the crime.” 3 Dall., at 390.

A

This Court addressed the Ex Post Facto Clause a mere decade after the Constitution was ratified. In Calder, Justice Chase described four types of ex post facto laws. 3 Dall., at 390. As relevant, Justice Chase’s third category indicated that “[e]very law that changes the punishment, and inflicts a greater punishment, than the law annexed to the crime, when committed” violates the Ex Post Facto Clause. Ibid. Justice Chase’s emphasis on increases in the punishment “annexed to the crime” was grounded in the English common law and accurately reflected the original understanding of the Ex Post Facto Clause. See Part II–B, infra. Unfortunately, the Court rapidly deviated from this formulation. In Kring v. Missouri, 107 U. S. 221 (1883), the Court declared that “any law passed after the commission of an offence which . . . ‘in relation to that offence, or its consequences, alters the situation of a party to his disadvantage,’ is an ex post facto law.” Id., at 235 (quoting Justice Washington’s jury charge in United States v. Hall, 26 F. Cas. 84, 86 (No. 15,285) (CC Pa. 1809) (emphasis added). It took nearly a century for the Court to decide that Kring’s “departure from Calder’s explanation of the original understanding of the Ex Post Facto Clause was . . . unjustified.” Collins v. Youngblood, 497 U. S. 37, 49 (1990) (overruling Kring).

Following Collins’ disavowal of Kring, the Court held that a law is ex post facto if it “produces a sufficient risk of increasing the measure of punishment attached to the covered crimes.” Morales, supra, at 509. While Morales avoided the over-breadth of Kring’s “disadvantage the de-fendant” test, it failed to reconnect our ex post facto ju-risprudence to the original understanding of the term.[1]* The “sufficient risk” test also depends upon empirical analysis that cannot yield determinative answers and which courts are ill equipped to handle. See, e.g., Gar- ner, 529 U. S., at 255 (“When the rule does not by its own terms show a significant risk, the respondent must demonstrate, by evidence drawn from the rule’s practical implementation by the agency charged with exercising discretion, that its retroactive application will result in a longer period of incarceration than under the earlier rule”). More fundamentally, the “sufficient risk” test, like the “disadvantage the defendant” test, wrongly focuses on the particular sentence that the defendant might receive, rather than on the punishment “annexed to the crime.”

The practical difficulties with the test are apparent even from our application in Morales, where we considered an amendment to California’s parole procedures that allowed, under certain circumstances, the Board of Prison Terms to decrease the frequency of parole suitability hearings. Under the sufficient risk test, we were compelled to speculate about the possible effects of the new law on various individuals’ prison terms. Ultimately, we held that the amendment did not violate the Ex Post Facto Clause because the “narrow class of prisoners covered by the amendment [could not] reasonably expect that their prospects for early release on parole would be enhanced by the opportunity of annual hearings.” Morales, supra, at 512. But nothing in the text or history of the Ex Post Facto Clause suggests that it should hinge on the expectations that prisoners and defendants have about how many days they will spend in prison.

B

“Although the Latin phrase ‘ex post facto’ literally encompasses any law passed ‘after the fact,’ ” Collins, 497 U. S., at 41, the Court has long recognized that the phrase “was a term of art with an established meaning” at the time of the founding. Ibid. Blackstone offers the first key to understanding this “established meaning.” He explic-itly opposed laws that rendered innocent conduct crimi- nal after the fact. See 1 W. Blackstone, Commentaries *44 (hereinafter Blackstone). Such laws deprive citizens of notice and fair warning and are, therefore, an affront to man’s “reason and freewill.” Id., at *39; see id., at *46. Blackstone, thus, considered them illegitimate. Id., at *44; see also The Federalist No. 44, p. 301 (J. Cooke ed. 1961) (J. Madison) (“[E]x post facto laws . . . are contrary to the first principles of the social compact, and to every principle of sound legislation”). For this reason, ex post facto laws have rightly been described as “formidable instruments of tyranny,” id., No. 84, at 577 (A. Hamilton), and their prohibition a “bulwark in favour of the personal security of the subject,” Calder, supra, at 390 (opinion of Chase, J.).

Although Blackstone confined his discussion of ex post facto laws to those laws retroactively declaring innocent acts to be criminal, other authorities confirm that laws retroactively increasing the punishment were also understood to be ex post facto at the time of the founding. See, e.g., 2 R. Wooddeson, A Systematical View of the Laws of England, as treated in a Course of Vinerian Lectures 638 (1792) (discussing “acts of parliament, which principally affect the punishment, making therein some innovation, or creating some forfeiture or disability, not incurred in the ordinary course of law”); 3 J. Story, Commentaries on the Constitution of the United States §679, p. 486 (Abr. 1833) (The “prohibition” against ex post facto laws “reaches every law . . . whereby the act, if a crime, is aggravated in enormity, or punishment”). Justice Chase’s formulation reflects this understanding. Calder, 3 Dall., at 390 (“Every law that changes the punishment, and inflicts a greater punishment, than the law annexed to the crime, when committed” is ex post facto). Under this view, courts must compare the punishment affixed to the crime at the time of the offense with the punishment affixed at the time of sentencing. If the latter is harsher than the former, the court must apply the punishment in effect at the time of the offense.

At common law, it was quite easy to identify when a law retroactively increased the punishment, because the criminal law generally “prescribed a particular sentence for each offense.” Langbein, The English Criminal Trial Jury on the Eve of the French Revolution, in The Trial Jury in England, France, Germany 1700–1900, p. 36 (A. Schioppa ed. 1987). In a world of determinate sentencing, a retro-active increase in the punishment affixed to a crime rend-ers an act “punishable in a manner in which it was not punishable when it was committed,” Fletcher v. Peck, 6 Cranch 87, 138 (1810), which is sufficient for an ex post facto violation. The key point is that “the ex post facto [C]lause looks to the standard of punishment prescribed by a statute, rather than to the sentence actually imposed.” Lindsey v. Washington, 301 U. S. 397, 401 (1937).

Focusing on the punishment affixed by law, rather than on the specific sentence imposed, furthers the goals of notice and fair warning recognized by Blackstone as the rationales for the prohibition against ex post facto laws. See Ross’ Case, 19 Mass. 165, 170 (1824) (“A party ought to know, at the time of committing the offence, the whole extent of the punishment; for it may sometimes be a matter of calculation, whether he will commit the offence, considering the severity of the punishment”). Because increasing the punishment affixed to the crime deprives people of the opportunity to plan their conduct in light of the law, “[t]he enhancement of a crime, or penalty, seems to come within the same mischief as the creation of a crime or penalty; and therefore they may be classed together.” Calder, supra, at 397 (opinion of Paterson, J.).

Retroactive laws that merely create a risk that a defendant will receive a higher sentence, however, do not implicate traditional ex post facto concerns. An individual contemplating the commission of a given offense knows he may be sentenced anywhere within the legally prescribed range. He may hope to receive a lenient sentence, and he may even have good reasons for expecting leniency. But he does not have any guarantees. See Garner, 529 U. S., at 258 (Scalia, J., concurring in part in judgment) (“Discretion to be compassionate or harsh is inherent in the sentencing scheme, and being denied compassion is one of the risks that the offender knowingly assumes”). The law provides the defendant with only one assurance: He will be sentenced within the range affixed to his offense by statute. Legal changes that alter the likelihood of a particular sentence within the legally prescribed range do not deprive people of notice and fair warning, or implicate the concerns about tyranny that animated the adoption of the Ex Post Facto Clause.

C

The statutory range in effect at the time of petitioner’s offense remained in effect at his sentencing. The Guidelines sentencing range is not the punishment affixed to the offense. See in Part I–A, supra. Accordingly, sentencing petitioner under the amended Guidelines did not violate the Ex Post Facto Clause. Because the Court concludes otherwise, I respectfully dissent.