Setser v. United States,
566 U.S. 231 (2012)

Annotate this Case
  • Syllabus  | 
  • Opinion (Antonin Scalia)  | 
  • Dissent (Stephen G. Breyer)



No. 10–7387



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

[March 28, 2012]

     Justice Breyer, with whom Justice Kennedy and Justice Ginsburg join, dissenting.

     The Sentencing Reform Act of 1984 seeks to reform federal sentencing practices by creating a federal Sentencing Commission instructed to develop and to promulgate federal Sentencing Guidelines. The provision of the Act here at issue concerns “multiple sentences.” See 18 U. S. C. §3584. It brings into focus a difficult Guidelines-related problem: How should a federal judge sentence an offender where the offender has been convicted of having violated several different statutes? The convictions may have taken place all at the same time. Or, some convictions might have taken place at an earlier time, the offender may already have been sentenced to prison, and indeed the offender may still be serving that sentence. The federal judge must decide the extent to which a sentence attached to one conviction should be served concurrently or consecutively with sentences attached to other convictions.

     An understanding of the nature of this general problem and the Sentencing Commission’s statutorily foreseen solutions will help the reader understand why, in my view, the better legal answer to the question before us is that a federal sentencing judge does not have the power to order that a “federal sentence be consecutive to an anticipated state sentence that has not yet been imposed.” Ante, at 1.


     The Sentencing Reform Act (SRA) has two overall objectives. See Barber v. Thomas, 560 U. S. __, __ (2010) (slip op., at 7); see also United States Sentencing Commission, Guidelines Manual §1A3, p. 1.2 (Nov. 1987) (USSG) (addressing statutory objectives). First, it seeks greater honesty in sentencing. Instead of a parole commission and a judge trying to second-guess each other about the time an offender will actually serve in prison, the SRA tries to create a sentencing system that will require the offender actually to serve most of the sentence the judge imposes. See Mistretta v. United States, 488 U. S. 361, 367 (1989) (“[The SRA] makes all sentences basically determinate”). Second, the Act seeks greater fairness in sentencing through the creation of Guidelines that will increase the likelihood that two offenders who engage in roughly similar criminal behavior will receive roughly similar sen-tences. See Barber, supra, at ___ (slip op., at 7) (noting that Congress sought to achieve, in part, “increased sentencing uniformity”).

     To implement these reforms, the SRA instructs the Commission to write Guidelines that inevitably move in the direction of increased “real offense” sentencing. See USSG §1A2, p. 1.1. (describing how statute, e.g., by insisting upon categories of offense behavior and offender characteristics, leads to this result). In principle, real offense sentencing would impose the same sentence upon different offenders who engage in the same real conduct irrespective of the statutes under which they are charged. Real offense sentencing, for example, would mean that two individuals, both of whom rob a bank and injure a teller, would receive the same sentence even if the Government charges one of them under a bank robbery statute and the other under an assault statute. See, e.g., USSG App. A (listing federal statutory offenses, while keying them to specific individual Guidelines that determine sentence based upon likely actual behavior). In the event, the Guidelines move the sentencing system in this direction while simultaneously recognizing that other factors require considerable modification of the real offense principle. See USSG §1A4(a) (“real offense vs. charge offense sentencing”).

     Nonetheless the “real offense” goal influenced the Act’s, and the Commission’s, objectives in respect to the sentencing of an offender with multiple convictions. Insofar as several convictions arise out of the same course of behavior, the sentencing judge should treat the crimes underlying the convictions as if they were all part of a single crime and sentence accordingly. But, insofar as the crimes underlying the convictions arise out of different courses of behavior, the sentencing judge should treat the crimes underlying the convictions as if they were not part of a single crime and should see that the ultimate sentence reflects that fact.

     To achieve these objectives is easier said than done. For one thing, it requires a definition of what counts as the same course of behavior. The Guidelines set forth that definition in §1B1.3, p. 1.17 (“Relevant Conduct”). For another thing, statutes and Guidelines that set forth related instructions must take into account the fact that sentencing-related circumstances can prove highly complex. To take a fairly simple example, suppose that a defendant is convicted of both robbery and impersonating a federal official, that he has engaged in a single course of behavior, but that neither the robbery nor the impersonation Guidelines take account of the other. Instructions about concurrent/consecutive sentences must give the judge an idea about what to do in such a case. They must also take account of the fact that a maximum penalty contained in a statute will trump a greater penalty contained in a Guideline. And they must tell the judge (faced with multiple convictions) what to do where that is so.

     Reflecting these, and other, complexities, the Guidelines contain complex instructions about how to sentence where the offender is convicted of “Multiple Counts,” see USSG §3D, or has previously been convicted of a crime for which he is “subject to an Un-discharged Term of Imprisonment,” see §5G1.3. The Guidelines also tell the sentencing judge how, through the use of partially concurrent and partially consecutive sentences, to build a total sentence that meets the Guidelines’ requirements. See §§5G1.2(d), 5G1.3.

     With this background it becomes easier to understand the statutory provisions before us. They reflect the fact that Congress expected sentencing judges, when faced with a defendant convicted of multiple crimes, to construct a sentence that would, at least to a degree, reflect the defendant’s real underlying behavior. Where two convictions reflect in whole or in part the same behavior, the overall sentences should reflect that fact, say by running concurrently.

     Accordingly, the statute says that “[m]ultiple terms of imprisonment imposed at the same time run concurrently unless the court orders or the statute mandates that the terms are to run consecutively.” 18 U. S. C. §3584(a). And that statement reflects the fact that often (but not always) multiple convictions after a single trial will reflect a sin-gle course of behavior (different aspects of which violate different criminal statutes). The statute also says that “[m]ultiple terms of imprisonment imposed at different times run consecutively unless the court orders that the terms are to run concurrently.” Ibid. This statement reflects the fact that several convictions imposed after different trials are more likely to reflect unrelated behaviors. In the first instance that the statute addresses, concurrent sentences are more likely to be appropriate; in the second, consecutive sentences are more likely to be appropriate. But that is not always so. Thus the statu- tory provisions assure sentencing judges that they retain the power to reach a different conclusion.

     At this point, I would ask the question that this case poses. Why does the statute say nothing about a sentencing judge imposing a sentence that might run consecu-tively with a sentence that a (typically different) judge has not yet imposed? The answer is this: Because the sentencing judge normally does not yet know enough about the behavior that underlies (or will underlie) a sentence that has not yet been imposed. Normally the sentencing judge does not know, for example, (a) what that sentence will be, (b) whether the behavior underlying that later sentence constitutes part of the same course of behavior that underlies the present sentence or, instead, is totally separate from the behavior underlying the present sentence, or (c) is partly the same and partly different. Even if the judge has an idea about what will happen, he does not know precisely what will happen; and precision in this matter is important.

     In a word, the sentencing judge normally does not yet know enough about what will happen in the sentencing-proceeding-yet-to-come to be able to construct a sentence that meets the Guidelines’ instructions and which, in doing so, helps to assure that different individuals who engage in the same criminal behavior will typically receive roughly comparable sentences.

     Of course, the Court is correct when it says that eventually the sentences will run (either wholly in or in part) concurrently or consecutively. And someone must decide how they will run. Ante, at 2–3. But the Court is not correct when it says that this someone should be the first federal sentencing judge. Rather, the Executive and Judicial Branches have devised a system that can draw upon the intentions of that first federal judge, while applying them in light of actual knowledge about what later happened. The Bureau of Prisons (BOP) in effect makes the consecutive/concurrent decision after considering, among other things, “any statement by the court that imposed the sentence,” including statements “concerning the purposes for which the sentence to imprisonment was determined to be warranted.” 18 U. S. C. §3621(b)(4)(A). And its program statement provides that it will review the “intent of the federal sentencing court” when deciding whether in effect to make an earlier federal, and later state, sentence concurrent or consecutive. Dept. of Justice, BOP, Program Statement 5160.05: Designation of State Institution for Service of Federal Sentence 4 (Jan. 16, 2003). The Bureau exercises this authority by designating (or refusing to designate) the state prison where an offender is or will be incarcerated pursuant to his state sentence as the place where he will serve his federal sentence. 18 U. S. C. §3621(b).

     This exercise of authority by the Executive Branch is not constitutionally surprising. After all, “federal sentencing” has “never . . . been thought to be assigned by the Constitution to the exclusive jurisdiction of any one of the three Branches of Government.” Mistretta, 488 U. S., at 364. And, until fairly recently the federal BOP decided (via parole) the far more global question of just how long (within broad limits) each imprisoned offender would serve. See id., at 367. Thus, the present Bureau involvement represents a further practical accommodation to a fact about the world, namely that the initial sentencing judge typically lacks important sentencing-related information about a second sentence that has not yet been imposed.


     Given the purposes and the mechanics of the SRA’s sentencing system, just described, the better reading of the “multiple sentences” provision is a reading that denies a sentencing judge the authority to “order that the federal sentence be consecutive to an anticipated state sentence that has not yet been imposed.” Ante, at 1. For one thing, nothing in the statute explicitly grants the judge that authority. The text refers to other circumstances, those that involve earlier or contemporaneous (multiple-count) convictions, while it does not refer to later imposed sentences.

     For another, exercise of any such authority would more likely hinder than advance the basic objectives of the SRA. As I have explained, supra, at 2–5, a sentencing judge typically needs detailed information when constructing a multiple-count or multiple-conviction Guideline sentence. The fact that the future sentence has not yet been imposed means that information will often be lacking, and that in turn means that the exercise of such authority would risk confusion and error. A sentencing judge who believes, for example, that the future conviction will be based upon different relevant conduct (and consequently orders a consecutive sentence) could discover that the second conviction rests upon the same relevant conduct (warranting a concurrent sentence). Mistakes of this kind increase the risk of sentencing disparity and, insofar as the first judge guesses wrong, they can mean a less honest sentencing process as well.

     Further, I can find no significant tradition (pre-Guideline or post-Guideline) of federal judges imposing a sentence that runs consecutively with a sentence not yet imposed. The Court refers to four Courts of Appeals cases for the proposition that “traditionally” a judge possessed this authority. Ante, at 4. The opinions in three of the cases are each about a page long and do not discuss the matter here at issue. (They assume, without significant discussion, the existence of the relevant sentencing authority.) See Anderson v. United States, 405 F. 2d 492, 493 (CA10 1969) (per curiam) (addressing the question whether a federal sentence runs from the date of its imposition or from the date of entry into federal custody); United States v. Kanton, 362 F. 2d 178, 179–180 (CA7 1966) (per curiam) (same); United States ex rel. Lester v. Parker, 404 F. 2d 40, 41 (CA3 1968) (per curiam) (addressing the question whether a sentence was insufficiently certain for purposes of due process). The fourth case, Salley v. United States, 786 F. 2d 546, 548 (CA2 1986), discusses the issue directly and takes the Court’s position. But, like the other three cases, it was decided before the Guidelines took effect (i.e., when the reasons for denying the authority were less strong). And, one judge on the panel disagreed in a separate opinion, and in my view has the better of the argument. See id., at 548–550 (Newman, J., concurring in result); see also United States v. Eastman, 758 F. 2d 1315, 1317 (CA9 1985) (holding that a judge lacks the here-relevant sentencing power). In any event, these instances are too few to constitute a “tradition.”

     In fact the Senate Committee Report accompanying the SRA provides strong evidence that there was no such tradition. S. Rep. No. 98–225 (1983). That Report thoroughly surveyed prior law. It says that the SRA is a “comprehensive statement of the Federal law of sentencing,” that it “describes in detail the kinds of sentences that may be imposed,” and that §3584 “provides the rules for determining the length of a term of imprisonment for a person convicted of more than one offense.” Id., at 50, 125–126. It further states that “[e]xisting law permits the imposition of either concurrent or consecutive sentences,” which practice it then describes as limited to two scenarios: “[t]erms of imprisonment imposed at the same time,” and those “imposed on a person already serving a prison term.” Id., at 126. It says the same when describing how §3584 is supposed to work. In neither place does it refer to a practice of, or any authority for, imposing a prison term that runs consecutively with a future term not yet imposed.

     In addition, a grant of such authority risks at least occasional incoherence. For example, the statute, after setting forth the court’s authority to impose a sentence of imprisonment that runs either concurrently or consecutively with other terms imposed in the same or in earlier proceedings, creates an exception that says: “except that the terms may not run consecutively for an attempt and for another offense that was the sole objective of the attempt.” 18 U. S. C. §3584(a). Now suppose the Court were right, and a sentencing judge had the authority to run a present term consecutively with a not-yet-imposed future term. Would it not be important to apply this same “attempt” exception in such instances as well? Indeed, the exception is phrased in categorical terms, and the legis-lative history in no way indicates that the exception ap-plies only occasionally. See S. Rep. No. 98–225, at 126 (“[C]onsecutive terms of imprisonment may not, contrary to current law, be imposed for [attempt] and for an offense that was the sole objective of the attempt” (emphasis added)). Yet it is difficult, if not impossible, to read the statute’s language as broadening the exception beyond the statutorily listed scenarios.

     Or, consider, for example, an offender tried for arguably related crimes in two different federal courts at two different times. The Court’s reading would not only allow the second judge to order concurrent service with the first sentence if warranted, as the statute explicitly permits, but it would also allow the first judge to make an analogous but anticipatory order based upon the sentence he expected the second judge would impose. But where complex forms of criminal behavior are at issue, these different judges may reach different conclusions. The result may well be conflict and confusion.

     Finally, as I said above, supra, at 5–6, a more practical solution to potential problems presented by a future sentencing proceeding lies closer at hand. The BOP has the statutory authority to effect concurrent service of federal and state sentences and is well situated to take into account both the intent of the first sentencing judge and the specific facts developed in the second sentencing. The relevant statute provides that “[t]he Bureau may designate any available penal or correctional facility . . . , whether maintained by the Federal Government or otherwise . . . .” 18 U. S. C. §3621(b). And in reliance on this authority, the Bureau has concluded that it has the power to “designat[e] . . . a state institution for concurrent service of a federal sentence.” Program Statement 5160.05, at 1. The Program Statement further provides that exercise of this power will be guided by, in part, “the intent of the federal sentencing court” in addition to “any other pertinent information regarding the inmate.” Id., at 4.

     The Court’s only criticism of this system is that it is less “natural” to read the statute “as giving the Bureau of Prisons what amounts to sentencing authority.” Ante, at 8. But what is unnatural about giving the Bureau that authority? The sentencing process has long involved cooperation among the three branches of Government. Mistretta, 488 U. S., at 364. And until the Guidelines the BOP itself decided, within broad limits, precisely how much prison time every typical offender would serve. Even today, it still decides that question within certain limits. 18 U. S. C. §3624 (2006 ed. and Supp. IV) (delegating to the BOP authority to calculate “good time credit,” which in effect reduces a prisoner’s term of incarceration); see also Barber, 560 U. S., at __ (slip op., at 1). Although Congress limited the Bureau’s authority in this respect, there is nothing unnatural about leaving the Bureau with a small portion of that authority—particularly where doing so helps significantly to alleviate a small, but important, technical problem in the application of the SRA’s sentencing system.

*  *  *

     Because the Court does not ask why the “multiple sentencing” provision leaves out the authority at issue—concerning the not-yet-imposed sentence—it reaches what I believe is the wrong result. Consequently, with respect, I dissent.

Disclaimer: Official Supreme Court case law is only found in the print version of the United States Reports. Justia case law is provided for general informational purposes only, and may not reflect current legal developments, verdicts or settlements. We make no warranties or guarantees about the accuracy, completeness, or adequacy of the information contained on this site or information linked to from this site. Please check official sources.