JUSTICE BREYER, concurring in part and concurring in the judgment.
I join Part III of the Court's opinion as to the second question presented. As to the first question presented, I agree with the Court's conclusion. And I agree with its opinion to the extent that it is consistent with this concurrence.
Like the majority, I believe Booker requires us to hold 18 U.S.C. § 3742(g)(2) unconstitutional. See ante, at 15-21; United States v. Booker, 543 U. S. 220 (2005); see also Apprendi v. New Jersey, 530 U. S. 466 (2000). And, like the majority, I believe that the law does not require a sentencing court to follow a Guideline policy statement that forbids taking account of postsentencing rehabilitation. United States Sentencing Commission, Guidelines Manual § 5K2.19 (Nov. 2010) (USSG). I would emphasize, however, that this conclusion does not leave a sentencing court free to disregard the Guidelines at will. To the contrary, the law permits the court to disregard the Guidelines only where it is "reasonable" for a court to do so. Booker, supra, at 261-262; Gall v. United States, 552 U. S. 38, 51-52 (2007); Kimbrough v. United States, 552 U. S. 85, 109 (2007). And an appellate court must be guided by the basic sentencing objectives of the statutes that create the Guidelines in determining whether, in disregarding the Guidelines, the sentencing court has acted unreasonably.
I
The Guideline in question consists of a policy statement that sets forth an exception to normal Guideline rules. Normally, the Guidelines authorize a sentencing judge to consider a departure from an ordinary Guidelines sentence in any case "where conduct significantly differs from the norm" to which "a particular guideline linguistically applies." USSG ch. 1, pt. A1, § 4(b) (discussing the Guidelines' general approach to departures). The policy statement at issue is one of a handful of Guideline rules that nonetheless forbid departure. It says that a defendant's "[p]ost-sentencing rehabilitative efforts, even if exceptional, . . . are not an appropriate basis for a downward departure when resentencing." USSG § 5K2.19. The policy statement thereby adds "Post-Sentencing Rehabilitative Efforts" to such factors as race, sex, national origin, creed, religion, and socioeconomic status, which the Guidelines absolutely prohibit the sentencing judge from taking into account. USSG ch. 1, pt. A1, § 4(b).
II
Can a sentencing court, despite this policy statement, take account of postsentencing rehabilitation in the particular circumstances that this case presents? I cannot find the answer to this question in the language of the sentencing statutes, in sentencing traditions, in the pre-Guidelines case of Williams v. New York, 337 U. S. 241 (1949), or in this Court's use of the word "advisory." As the majority points out, a sentencing statute forbids any "`limitation'" on the "`information concerning the background, character, and conduct'" that "`a court . . . may . . . consider.'" Ante, at 10 (quoting 18 U.S.C. § 3661 (emphasis deleted)). But this provision must refer to all relevant information. See USSG § 1B1.4 and comment. (generally incorporating § 3661, but noting that there are certain factors that should not be considered for any purpose). If the Guideline policy statement's absolute prohibition on consideration of postsentencing rehabilitation were legally binding, then information on that score (like information about race, religion, sex, or national origin) would fall outside the scope of this provision, for it would not be relevant. Thus, reference to the statute begs the question.
Nor can I find much help in the majority's reference to a sentencing "`tradition'" that considers "`every convicted person as an individual.'" Ante, at 9 (quoting Koon v. United States, 518 U. S. 81, 113 (1996)). That is because individualized sentencing is not the only relevant tradition. A just legal system seeks not only to treat different cases differently but also to treat like cases alike. Fairness requires sentencing uniformity as well as efforts to recognize relevant sentencing differences. Indeed, when Congress enacted the sentencing statutes before us, it focused upon the unfair way in which federal sentencing failed to treat similar offenders similarly. And Congress wrote statutes designed primarily (though not exclusively) to bring about greater uniformity in sentencing. See, e.g., Booker, supra, at 253-254. The statutes do so in large part through the creation of a system of Guidelines written by a Sentencing Commission, which Congress intended the courts to follow. See Mistretta v. United States, 488 U. S. 361 (1989) (Sentencing Commission constitutional); Rita v. United States, 551 U. S. 338, 348-349 (2007); 18 U.S.C. § 3553(a) (identifying relevant factors in sentencing, including uniformity).
The Williams case is similarly unhelpful. That is because Congress in the Sentencing Reform Act of 1984 — the law before us — disavowed the individualized approach to sentencing that that case followed. Williams emphasized the importance of a sentencing court's legal power to tailor punishment ability to fit the circumstances of each individual offender. 337 U. S., at 247 (emphasizing "modern concepts individualizing punishment"). But Congress, concerned that individualized sentencing had gone too far, wrote a new sentencing law designed to help correct "disparities" among similar defendants sentenced by different judges. See S. Rep. No. 98-225, p. 45 (1983) ("Sentencing disparities" are "unfair both to offenders and to the public"); id., at 38 (disparities "can be traced directly to the unfettered discretion the law confers on those judges and parole authorities responsible for imposing and implementing the sentence").
Booker's description of the Guidelines as "advisory" offers somewhat greater assistance — but only if that word is read in light of the Sixth Amendment analysis that precedes it. This Court has held that the Sixth Amendment forbids Congress (through the Commission) to create Guidelines that both (1) require judges (without juries) to find sentencing facts and also (2) tie those facts to the mandatory imposition of particular sentences. 543 U. S., at 226, 244; see also Apprendi, 530 U. S., at 490 ( Sixth Amendment requires jury findings in respect to factual matters that require judge to increase sentence); Blakely v. Washington, 542 U. S. 296, 303-304 (2004) (same in respect to a State's mandatory guidelines). In light of this Sixth Amendment prohibition, the Court, believing that Congress would not have intended to introduce new juries into each sentencing proceeding, excised the few particular provisions of the sentencing statutes that specified that application of the Guidelines was mandatory. Booker, 543 U. S., at 259. The Court believed that the relevant statutes remained workable without those few provisions, that their excision could further Congress' basic sentencing intentions, and that excision was more likely to do so than invalidation of the entire statutory scheme. With an occasional exception (such as the statutory provision we strike down today), there is no reason to think that the sentencing statutes as limited in Booker run afoul of the Sixth Amendment. Ibid. Booker made clear that the remaining statutory provisions, while leading us to call the Guidelines "advisory" (rather than "mandatory"), do not give a sentencing judge carte blanche to apply, or not to apply, the Guidelines as that judge chooses. Rather, the "district courts, while not bound to apply the Guidelines, must consult those Guidelines and take them into account when sentencing." Id., at 264. Moreover, Booker held that appellate court review of sentencing is valid. Booker explained that the "statutory language, the structure of the [Sentencing Reform Act], and the sound administration of justice," taken together, require appellate courts to apply "reasonableness standard[s]" of review. Id., at 260-261, 262 (internal quotation marks omitted). Reasonableness standards, we added, are "not foreign to sentencing law." Id., at 262. And the "Act has long required their use in important sentencing circumstances — both on review of departures . . . and on review of sentences imposed where there was no applicable Guideline." Ibid. See also id., at 261 (appellate courts will apply "a practical standard of review already familiar to appellate courts: review for `unreasonable[ness]'"); id., at 264 ("[C]ourts of appeals" will "review sentencing decisions for unreasonableness").
We have also indicated that, in applying reasonableness standards, the appellate courts should take account of sentencing policy as embodied in the statutes and Guidelines, as well as of the comparative expertise of trial and appellate courts. Thus, in Kimbrough, we observed that in light of the "discrete institutional strengths" of the Sentencing Commission and sentencing judges, "a district court's decision to vary from the advisory Guidelines may attract greatest respect when the sentencing judge finds a particular case `outside the "heartland" to which the Commission intends individual Guidelines to apply.'" 552 U. S., at 109 (quoting Rita, supra, at 351). We noted, however, that "while the Guidelines are no longer binding, closer review may be in order when the sentencing judge varies from the Guidelines based solely on the judge's view that the Guidelines range `fails properly to reflect § 3553(a) considerations' even in a mine-run case." 552 U. S., at 109.
III
Unlike the majority, I would decide the question Kimbrough left open. And I would follow its suggested framework for evaluating "reasonableness." As Kimbrough suggests, doing so takes proper account of the comparative institutional abilities of trial courts, appellate courts, and the Sentencing Commission. The trial court typically better understands the individual circumstances of particular cases before it, while the Commission has comparatively greater ability to gather information, to consider a broader national picture, to compare sentences attaching to different offenses, and ultimately to write more coherent overall standards that reflect nationally uniform, not simply local, sentencing policies.
Applying Kimbrough's suggested framework, I would reason very much as does the majority. The first question is whether a sentencing judge might sometimes take account of a (resentenced) offender's postsentencing rehabilitation — despite a Guideline policy statement that says never. I would find that it is reasonable for the judge to disregard the Guidelines' absolute prohibition, despite the Commission's comparatively greater policy-formation abilities. That is because the Guideline policy statement itself runs counter to ordinary Guideline sentencing policy, which rarely forbids departures and then for very strong policy reasons. Supra, at 2. See USSG ch. 1, pt. A1, § 4(b).
The Commission offers no convincing justification for creating this exception with respect to postsentencing rehabilitation. The Commission's commentary says that for a judge at resentencing to lower a sentence for this reason (reflecting good behavior while the case is on appeal) would conflict with the use of other mechanisms, such as "good-time" credits, for that purpose. But how is that so? A defendant, after sentencing but while his case is on appeal, may or may not be entitled to "good time." That may depend upon whether he remains on bail or upon particular "good-time" rules. Regardless, the resentencing judge can take account of any such matter. See also ante, at 26-28.
The Commission's commentary also suggests it would be inequitable to allow an offender who is being resentenced to receive any kind of credit for his good behavior, say while his case was on appeal. But why is that so? After all, the Guidelines permit a judge to take account of an offender's good behavior after arrest but before initial sentencing. That time period could last longer than the time taken up on appeal. Why should pretrial behavior count but appeal time behavior not count? Like the majority, I find this justification for the policy statement unconvincing. See ante, at 25-26.
The second question is whether, given the sentencing court's power to disregard the policy statement forbidding departures based on postsentencing rehabilitation, the facts and circumstances here could warrant a departure (or variance) for that reason. And the answer, in my view, is yes. This case presents unusual rehabilitative circumstances. As the majority observes: "By the time of his second resentencing in 2009, Pepper had been drug-free for nearly five years, had attended college and achieved high grades, was a top employee at his job slated for a promotion, had reestablished a relationship with his father, and was married and supporting his wife's daughter." Ante, at 14. These are case-specific facts and circumstances, and they are of the kind that should lead appellate courts to show their "greatest respect" for a sentencing decision, including a departure or variance, that rests upon them.
IV
In sum, the sentencing statutes, as we have interpreted them, require courts of appeals to review sentences for reasonableness, including sentences that depart or vary from a specific Guideline. The appellate courts should review those decisions more closely when they rest upon disagreement with Guidelines policy. Kimbrough, 552 U. S., at 109. They should review those decisions with greater deference when they rest upon case-specific circumstances that place the case outside a specific Guideline's "heartland." See ibid.; Rita, 551 U. S., at 351; Koon, 518 U. S., at 98-99.
By interpreting the sentencing statutes in this way, we can remain faithful to Congress' basic intent in writing them — despite the need to invalidate statutory provisions that conflict with the Sixth Amendment. The statutes create a Sentencing Commission with authority to develop sentencing policy embodied in the Guidelines. The Guidelines are to further the statutes' basic objective, namely greater sentencing uniformity, while also taking account of special individual circumstances, primarily by permitting the sentencing court to depart in nontypical cases. By collecting trial courts' reasons for departure (or variance), by examining appellate court reactions, by developing statistical and other empirical information, by considering the views of expert penologists and others, the Commission can revise the Guidelines accordingly. See USSG ch. 1, pt. A1, § 3. Trial courts, appellate courts, and the Commission all have a role to play in what is meant to be an iterative, cooperative institutional effort to bring about a more uniform and a more equitable sentencing system. See id., at 1-15. I would interpret the statutes before us accordingly.
JUSTICE ALITO, concurring in part, concurring in the judgment in part, and dissenting in part.
I join Part III of the opinion of the Court. I agree with the Court that the decision below cannot be affirmed on the basis of 18 U.S.C. § 3742(g), as amicus suggests. This provision was designed to function as part of the mandatory Guidelines scheme that the Court struck down in United States v. Booker, 543 U. S. 220, 258-265 (2005). Although amicus' argument is ingenious, even the sort of surgery sanctioned in Booker cannot transform this provision into one that can survive in the post-Booker world.
I also concur in the judgment to the extent that it holds that the decision below regarding evidence of postsentenc-ing rehabilitation must be reversed. That decision, which entirely precluded consideration of such evidence, was consistent with the policy statement in § 5K2.19 of the United States Sentencing Guidelines, but "[t]he Booker remedial decision . . . does not permit a court of appeals to treat the Guidelines' policy decisions as binding." Kimbrough v. United States, 552 U. S. 85, 116 (2007) (ALITO J dissenting)
Under. Booker, however, district judges are still required in almost all cases to give significant weight to the policy decisions embodied in the Federal Sentencing Guidelines. See 552 U. S., at 116; Gall v. United States, 552 U. S. 38, 61-67 (2007) (ALITO, J., dissenting). Congress delegated to the Sentencing Commission the authority to make policy decisions regarding federal sentencing, see 18 U.S.C. §§ 3553(a)(4), (5), and requiring judges to give significant weight to the Commission's policy decisions does not run afoul of the Sixth Amendment right that the mandatory Guidelines system was found to violate, i.e., the right to have a jury make certain factual findings that are relevant to sentencing.
While I continue to believe that sentencing judges should be required to give significant weight to all Guidelines provisions and policy statements, see Kimbrough, 552 U. S., at 116 (opinion of ALITO, J.), the Court in Kimbrough held that sentencing judges may not be required to give weight to some unusual policy decisions, see id., at 109-110 (majority opinion). And JUSTICE BREYER now makes a reasonable case that the particular policy statement involved in this case is distinguishable from almost all of the other rules that the Commission has adopted. See ante, p. ___ (opinion concurring in part and concurring in judgment). His position seems to me more consistent with Kimbrough than the Court's. It would at least prevent us from sliding all the way down the slippery slope that leads back to the regime of entirely discretionary federal sentencing that preceded the enactment of the Sentencing Reform Act of 1984, 98 Stat. 1987.
Anyone familiar with the history of criminal sentencing in this country cannot fail to see the irony in the Court's praise for the sentencing scheme exemplified by Williams v. New York, 337 U. S. 241 (1949), and 18 U.S.C. § 3661. By the time of the enactment of the Sentencing Reform Act in 1984, this scheme had fallen into widespread disrepute. See, e.g., Mistretta v. United States, 488 U. S. 361, 366 (1989) (noting "[f]undamental and widespread dissatisfaction with the uncertainties and the disparities" of this scheme); United States v. DiFrancesco, 449 U. S. 117, 142 (1980) ("It has been observed . . . that sentencing is one of the areas of the criminal justice system most in need of reform"); S. Rep. No. 98-223, p. 62 (1983) ("The shameful disparity in criminal sentences is a major flaw in the existing criminal justice system"). Under this system, each federal district judge was free to implement his or her individual sentencing philosophy, and therefore the sentence imposed in a particular case often depended heavily on the spin of the wheel that determined the judge to whom the case was assigned. See Bullington v. Missouri, 451 U. S. 430, 444, n. 16 (1981) ("There has been no attempt to separate policymaking from individual sentencing determinations" (internal quotation marks omitted)); M. Frankel, Criminal Sentences: Law Without Order 5 (1973) ("[T]he almost wholly unchecked and sweeping powers we give to judges in the fashioning of sentences are terrifying and intolerable for a society that professes devotion to the rule of law").
Insofar as 18 U.S.C. § 3661 permitted a sentencing judge to consider evidence of postsentencing rehabilitation, that provision was effectively modified by the subsequent enactment of the Sentencing Reform Act, which instructed the Sentencing Commission to adopt guidelines and policy statements that avoid "unwarranted sentencing disparities," 28 U.S.C. § 991(b)(1)(B); see also § 994(f), and which provided that sentencing courts "shall consider . . . any pertinent policy statement," 18 U.S.C. § 3553(a)(5).
Some language in today's opinion reads like a paean to that old regime, and I fear that it may be interpreted as sanctioning a move back toward the system that prevailed prior to 1984. If that occurs, I suspect that the day will come when the irrationality of that system is once again seen, and perhaps then the entire Booker line of cases will be reexamined.
JUSTICE THOMAS, dissenting.
I would affirm the Court of Appeals and uphold Pepper's sentence. As written, the Federal Sentencing Guidelines do not permit district courts to impose a sentence below the Guidelines range based on the defendant's postsentencing rehabilitation. See United States Sentencing Commission, Guidelines Manual § 5K2.19 (Nov. 2010) (USSG) Therefore I respectfully dissent
I agree with the Court that the law of the case doctrine did not control Pepper's resentencing. See ante, at 29-31. Although the charge to which Pepper pleaded guilty carried a mandatory minimum of 120 months' imprisonment, the mandatory minimum did not apply because he was eligible for safety-valve relief pursuant to 18 U.S.C. § 3553(f) (2000 ed.) and § 5C1.2 of the United States Sentencing Guidelines Manual (Nov. 2003) (USSG).
In United States v. Booker, 543 U. S. 220, 258-265 (2005), the Court rendered the entire Guidelines scheme advisory, a remedy that was "far broader than necessary to correct constitutional error." Kimbrough v. United States, 552 U. S. 85, 114 (2007) (THOMAS, J., dissenting). Because there is "no principled way to apply the Booker remedy," I have explained that it is "best to apply the statute as written, including 18 U.S.C. § 3553(b), which makes the Guidelines mandatory," unless doing so would actually violate the Sixth Amendment. Id., at 116; see Booker, supra, at 313-326 (THOMAS, J., dissenting in part); Gall v. United States, 552 U. S. 38, 61 (2007) (THOMAS, J., dissenting); Irizarry v. United States, 553 U. S. 708, 717 (2008) (THOMAS, J., concurring).
I would apply the Guidelines as written in this case because doing so would not violate the Sixth Amendment. The constitutional problem arises only when a judge makes "a finding that raises the sentence beyond the sentence that could have lawfully been imposed by reference to facts found by the jury or admitted by the defendant." Booker, supra, at 313 (opinion of THOMAS, J.). Pepper admitted in his plea agreement to involvement with between 1,500 and 5,000 grams of methamphetamine mixture, which carries a sentence of 10 years to life under 21 U.S.C. § 841(b)(1)(A)(viii). United States v. Pepper, 412 F. 3d 995, 996 (CA8 2005). Because Pepper has admitted facts that would support a much longer sentence than the 65 months he received, there is no Sixth Amendment problem in this case.
Pepper also stated that he understood both the 10-year statutory minimum and that the Government was making no promises about any exceptions. USSG § 5K1.1 provides that a court may depart from the Guidelines "[u]pon motion of the government stating that the defendant has provided substantial assistance in the investigation or prosecution of another person who has committed an offense." Pepper provided information to Government investigators and a grand jury concerning two other individuals involved with illegal drugs and guns.
Under a mandatory Guidelines regime, Pepper's sentence was proper. The District Court correctly calculated the Guidelines range, incorporated a USSG § 5K1.1 departure and the Government's motion under Federal Rule of Criminal Procedure 35(b), and settled on a 65-month sentence. Guideline § 5K2.19 expressly prohibits downward departures based on "[p]ost-sentencing rehabilitative efforts, even if exceptional." Nor is there any provision in the Guidelines for the "variance" Pepper seeks, as such variances are creations of the Booker remedy. I would therefore affirm the Court of Appeals' decision to uphold Pepper's sentence.
Although this outcome would not represent my own policy choice, I am bound by the choices made by Congress and the Federal Sentencing Commission. Like the majority, I believe that postsentencing rehabilitation can be highly relevant to meaningful resentencing. See ante, at 13-15. In light of Pepper's success in escaping drug addiction and becoming a productive member of society, I do not see what purpose further incarceration would serve. But Congress made the Guidelines mandatory, see 18 U.S.C. § 3553(b)(1), and authorized USSG § 5K2.19. I am constrained to apply those provisions unless the Constitution prohibits me from doing so, and it does not here.