Concurring opinion filed by Circuit Judge HENDERSON.

Concurring opinion filed by Circuit Judge KAVANAUGH.

PER CURIAM.

Having been resentenced on remand from this court, Walter Henry and Charles Harrison again appeal their convictions and sentences stemming from their participation in a conspiracy to import and distribute heroin in the Washington, D.C. and Baltimore, Maryland metropolitan areas. They appeal their convictions on the ground that the trial court committed three evidentiary errors: (1) it admitted expert testimony based in part on testimonial hearsay in violation of Crawford v. Washington, 541 U.S. 36, 124 S. Ct. 1354, 158 L. Ed. 2d 177 (2004) (Claim One), (2) it admitted expert testimony based on unreliable methodology (Claim Two) and (3) it admitted the guilty plea of a non-testifying co-conspirator (Claim Three). Henry and Harrison also argue that they received ineffective assistance of counsel when their appellate counsel failed to raise Claims Two and Three on direct appeal.1 Finally, Henry and Harrison challenge their sentences on the ground that the district court applied the United States Sentencing Guidelines (Guidelines) in a mandatory fashion in violation of United States v. Booker, 543 U.S. 220, 125 S. Ct. 738, 160 L. Ed. 2d 621 (2005). For the reasons set forth below, we conclude that Henry and Harrison waived their evidentiary claims by failing to raise them at trial or on direct appeal. Furthermore, Henry's and Harrison's ineffective assistance of counsel claim must be raised on collateral review, see 28 U.S.C. § 2255, if at all. Consistent with our holding in United States v. Ayers, 428 F.3d 312 (D.C.Cir.2005), however, we will vacate the sentences and remand for resentencing because we cannot say that the district court's Booker error was harmless beyond a reasonable doubt.

We set forth in detail the facts surrounding the heroin conspiracy in United States v. Stover, 329 F.3d 859 (D.C.Cir. 2003), cert. denied, 541 U.S. 1018, 124 S. Ct. 2088, 158 L. Ed. 2d 635 (2004). Accordingly, we mention only those matters required for an understanding of the decision. On May 4, 1999, the Government charged Henry, Harrison and other individuals with conspiracy to possess with intent to distribute one kilogram or more of heroin. Nuri Lama, the conspiracy's ringleader, pleaded guilty and thereafter testified as a witness against his co-conspirators at their October 20, 1999 trial. Although the jury convicted Henry of possession with intent to distribute, it failed to reach a verdict on the drug conspiracy count against Henry and Harrison.2

On September 11, 2000, Henry and Harrison were retried on the drug conspiracy count. Because Lama had died between the two trials, the prosecution introduced evidence at the second trial that Lama had pleaded guilty to the conspiracy charge. The prosecution also introduced the expert testimony of Metropolitan Police Department Detective Tyrone Thomas who testified about the meanings of various code words used by the co-conspirators during telephone conversations intercepted by the FBI.

After a five-week trial, the jury convicted both Henry and Harrison of conspiracy to possess with intent to distribute one kilogram or more of heroin. In determining Henry's and Harrison's sentences under the then-mandatory Guidelines, the district court utilized a formula derived from Detective Thomas's expert testimony to calculate the amount of heroin for which Henry and Harrison were responsible. Based on its calculations, the court found each responsible for 39.4 kilograms of heroin, resulting in a base offense level of 38. The court then added four levels for the leadership roles of both Henry and Harrison in the conspiracy and two levels for possession of a firearm for a total offense level of 44. Combined with Henry's and Harrison's Criminal History Category of I, the Guidelines mandated a sentence of life imprisonment for both and the court sentenced them accordingly.

The co-conspirators, including Henry and Harrison, appealed their convictions and sentences. In Stover, we affirmed Henry's and Harrison's convictions but concluded that the district court had erroneously calculated the amount of heroin for which each should be held responsible. Accordingly, we vacated their sentences and remanded to the district court to recalculate the drug quantity. Stover, 329 F.3d at 876.

At their resentencing hearings,3 both Henry and Harrison argued that the Sixth Amendment prohibits judicial calculation of drug amounts at sentencing.4 The district court rejected their argument and, based on its revised calculation of the drug amounts, held both Henry and Harrison responsible for 27.3 kilograms of heroin. Again treating the Guidelines as mandatory, the court set both base offense levels at 36. It then added four levels for their managerial roles in the offense and two levels for possession of a firearm for a total offense level of 42. Combined with their Criminal History Category of I, Henry's and Harrison's Guidelines range was 360 months to life imprisonment and the district court again sentenced them to life imprisonment. Henry and Harrison filed timely notices of appeal.

We address separately Henry's and Harrison's evidentiary challenges, their ineffective assistance of counsel claim and their Booker challenge.

Although they failed to raise their evidentiary challenges at trial or on direct appeal, Henry and Harrison argue that we should nevertheless review them for plain error on this appeal after the resentencing remand. See Fed.R.Crim.P. 52(b). We disagree.

It is well-settled that "where an argument could have been raised on an initial appeal, it is inappropriate to consider that argument on a second appeal following remand." Nw. Ind. Tel. Co. v. FCC, 872 F.2d 465, 470 (D.C.Cir.1989) (citing Laffey v. Nw. Airlines, Inc., 740 F.2d 1071, 1089-90 (D.C.Cir.1984)); see also United States v. Ben Zvi, 242 F.3d 89, 95-96 (2d Cir.2001) (applying waiver to second appeal following resentencing remand); cf. United States v. Adesida, 129 F.3d 846, 849-50 (6th Cir.1997) (applying waiver at resentencing remand stage). The "widely-accepted" bar promotes procedural efficiency and prevents the "`bizarre result'" that "`a party who has chosen not to argue a point on a first appeal should stand better as regards the law of the case than one who had argued and lost.'" Nw. Ind. Tel., 872 F.2d at 470 (quoting Laffey, 740 F.2d at 1089-90). Although the "waiver principle is [not] an absolute preclusion to appellate review," Crocker v. Piedmont Aviation, Inc., 49 F.3d 735, 739 (D.C.Cir. 1995), we have stated that "discretion to waive a waiver is normally exercised only in `exceptional circumstances, where injustice might otherwise result,'" id. at 740 (quoting Eli Lilly & Co. v. Home Ins. Co., 794 F.2d 710, 717 (D.C.Cir.1986)).

Henry and Harrison have not demonstrated "exceptional circumstances" that excuse their failure to raise the evidentiary challenges either at trial or on direct appeal. Regarding Claim One—the allegedly erroneous admission of Thomas's expert testimony based in part on hearsay—Henry and Harrison argue that the Crawford decision, which the Supreme Court issued after their direct appeal,5 created a new legal rule that rendered the testimony inadmissible. While we have suggested that an intervening change in the law can constitute an "exceptional circumstance[]" that justifies waiving waiver, see Crocker, 49 F.3d at 740, the Crawford decision did not effect such a change with respect to the admissibility of Thomas's expert testimony. In Crawford, the Supreme Court altered the framework set forth earlier in Ohio v. Roberts, 448 U.S. 56, 100 S. Ct. 2531, 65 L. Ed. 2d 597 (1980), and held that the Confrontation Clause of the Sixth Amendment bars "testimonial" hearsay statements unless the declarant is unavailable to testify and the defendant had a prior opportunity to cross-examine him. Crawford, 541 U.S. at 68-69, 124 S. Ct. 1354. Crawford, however, did not involve expert witness testimony and thus did not alter an expert witness's ability to rely on (without repeating to the jury) otherwise inadmissible evidence in formulating his opinion under Federal Rule of Evidence 703.6 In other words, while the Supreme Court in Crawford altered Confrontation Clause precedent, it said nothing about the Clause's relation to Federal Rule of Evidence 703. Because Crawford does not represent an intervening change in the law regarding the admissibility of Thomas's expert testimony, no exceptional circumstance exists and Henry's and Harrison's Claim One is thus waived.

With respect to Claims Two and Three, Henry and Harrison appear to argue that we should address the merits in this appeal because their original appellate counsel acted ineffectively in failing to raise them on direct appeal. See Appellants' Br. at 35-36, 40. In particular, Henry and Harrison argue that the trial court erroneously permitted Detective Thomas to testify about the meanings of terms used by the co-conspirators in intercepted telephone conversations and erroneously admitted evidence that Lama pleaded guilty although Lama himself was unavailable to testify. In order to reach the merits of Claims Two and Three because counsel allegedly acted ineffectively, however, we would first need to determine whether counsel acted ineffectively. And because an ineffective assistance of appellate counsel claim must ordinarily be made on collateral review, see Part II.B infra, we decline to consider Henry's and Harrison's original appellate counsel's performance an exceptional circumstance that justifies waiving waiver. Accordingly, we do not reach the merits of Claims Two and Three.

Henry and Harrison argue that their original appellate counsel acted ineffectively in failing to raise Claims Two and Three on direct appeal. When a defendant raises an ineffective assistance of trial counsel claim for the first time on direct appeal, our "`general practice is to remand the claim for an evidentiary hearing.'" United States v. Moore, 394 F.3d 925, 931 (D.C.Cir.2005) (quoting United States v. Rashad, 331 F.3d 908, 909 (D.C.Cir.2003)). Other circuits require the defendant to pursue an ineffective assistance of trial counsel claim in collateral proceedings under 28 U.S.C. § 2255. See, e.g., United States v. Quintero-Barraza, 78 F.3d 1344, 1347 (9th Cir.1995); United States v. Matzkin, 14 F.3d 1014, 1017 (4th Cir.1994). Nevertheless, an ineffective assistance of trial counsel claim can be resolved on direct appeal if the trial record conclusively shows that counsel did or did not perform effectively. See, e.g., Moore, 394 F.3d at 931; Quintero-Barraza, 78 F.3d at 1347; Matzkin, 14 F.3d at 1017.

The question here is whether we should similarly treat a claim of ineffective assistance of appellate counsel raised in an appeal following a resentencing remand. On the one hand, such a claim is virtually unreviewable on direct appeal as appellate counsel will hardly assert his own ineffectiveness. Cf. United States v. Weaver, 281 F.3d 228, 234 (D.C.Cir.2002) (explaining that defendant can, with new counsel, raise ineffective assistance of trial counsel claim for first time on direct appeal because "trial counsel cannot be expected to argue his own ineffectiveness in a motion for a new trial"). Therefore, if the trial record sufficed, we could decide a claim of ineffective assistance of appellate counsel on appeal following a resentencing remand. Cf. Nw. Ind. Tel., 872 F.2d at 470 (explaining that waiver applies "where an argument could have been raised on an initial appeal" (emphasis added)); see also Ben Zvi, 242 F.3d at 96 (deciding ineffective assistance of appellate counsel claim on appeal after resentencing remand because "the underlying challenge" was "sufficiently presented" and "judicial efficiency would be served"). Unlike a claim of ineffective trial counsel that can be made on direct appeal, however, a claim of ineffective appellate counsel can be made only if a "second" appeal occurs—for example, as a result of a resentencing remand. In effect, a fortuity of the judicial process—whether we decide to remand for resentencing—would thus determine whether the defendant has an alternative to collateral review in pursuing an ineffective assistance of appellate counsel claim. Cf. Griffith v. Kentucky, 479 U.S. 314, 327-28, 107 S. Ct. 708, 93 L. Ed. 2d 649 (1987) (applying new Supreme Court rule retroactively in criminal case because it is "solely the fortuities of the judicial process" that determine case Court chooses to hear first on plenary review). We believe a uniform procedure should apply to all defendants with an ineffective assistance of appellate counsel claim and therefore we will not consider such a claim on appeal following remand for resentencing. Instead, a defendant with such a claim must pursue it on collateral review pursuant to 28 U.S.C. § 2255.

Finally, it is undisputed that the district court sentenced Henry and Harrison by applying the Guidelines in a mandatory fashion to increase his sentence beyond that which could have been imposed based solely on the facts found by the jury which is constitutional error under United States v. Booker, 543 U.S. 220, 125 S. Ct. 738, 160 L. Ed. 2d 621 (2005). See United States v. Simpson, 430 F.3d 1177, 1182-83 (D.C.Cir. 2005). Nevertheless, the Government argues that the error was harmless beyond a reasonable doubt and thus a resentencing remand is unnecessary. We disagree.

At their respective resentencing hearings, both Henry and Harrison raised a Sixth Amendment objection to their sentences. Accordingly, we review the sentences for harmless error under Federal Rule of Criminal Procedure 52(a) ("Any error, defect, irregularity, or variance that does not affect substantial rights must be disregarded."). That is, the Government must establish "beyond a reasonable doubt that the error complained of did not contribute to the sentence obtained." United States v. Coumaris, 399 F.3d 343, 351 (D.C.Cir.2005) (quotation omitted).

The Government maintains that the district court's sentencing error was harmless beyond a reasonable doubt because it imposed the maximum sentence in the Guidelines range notwithstanding its discretion to impose a lower sentence. That is, the district court decided to sentence Henry and Harrison to life imprisonment under a then-mandatory Guidelines range of 360 months to life imprisonment. Relying on Tenth Circuit precedent, Appellee's Br. at 47-48, the Government contends that if a judge imposed the maximum sentence within the then-mandatory Guidelines range, there is no reason to believe he would change the sentence on remand. See United States v. Riccardi, 405 F.3d 852, 876 (10th Cir.2005) (sentence imposed at top of Guidelines range pre-Booker constitutes harmless error because "[h]aving exercised his limited discretion under the pre-Booker system to give [the defendant] the highest permissible sentence, there is no reason to think the judge would exercise his now-greater discretion to reduce the sentence"); United States v. Waldroop, 431 F.3d 736, 743 (10th Cir.2005) (citing Riccardi for same proposition).

We do not believe that the pre-Booker imposition of a sentence at the top of a Guidelines range by itself constitutes harmless error. In United States v. Coles, we held that a Booker error constitutes plain error if "there would have been a materially different result, more favorable to the defendant, had the sentence been imposed in accordance with the post-Booker sentencing regime." 403 F.3d 764, 767 (D.C.Cir.2005). In so holding, we recognized that "[t]here undoubtedly will be some cases in which a reviewing court will be confident that a defendant has suffered no prejudice," id. at 769, and therefore remand would be unnecessary. For example, if a district judge imposed a sentence at the statutory maximum "`and [said] that if he could he would have imposed an even longer sentence, there would be no basis for thinking that if he had known that the sentencing guidelines [were] merely advisory he would have given the defendant a lighter sentence.'" Id. (quoting United States v. Paladino, 401 F.3d 471, 483 (7th Cir.2005)).

We also suggested in Coles, however, that the imposition of a sentence at the top of a Guidelines range without "the judge's characterization of the sentence," United States v. Tchibassa, 452 F.3d 918, 930 (D.C.Cir.2006), is "hardly conclusive," Coles, 403 F.3d at 769. We noted:

"A conscientious judge—one who took the guidelines seriously whatever his private views—would pick a sentence relative to the guideline range. If he thought the defendant a more serious offender than an offender at the bottom of the range, he would give him a higher sentence even if he thought the entire range too high."

Id. at 770 (quoting Paladino, 401 F.3d at 482). In other words, a trial judge treating the Guidelines as mandatory might have imposed the maximum sentence in a particular range not necessarily because he believed the defendant deserved that sentence but because he considered the defendant to be the most serious type of offender in the range. And if this analysis applies in the plain error context, where the burden of proving prejudice is on the defendant, it applies a fortiori in the context of constitutional harmless error, where the burden is on the Government to establish no prejudice beyond a reasonable doubt. Thus, we agree with our sister circuits that have held that a sentence at the top of a Guidelines range is not, in itself, enough to establish that a Booker error was harmless beyond a reasonable doubt. See United States v. Woods, 440 F.3d 255, 259 (5th Cir.2006); United States v. Cain, 433 F.3d 1345, 1348 (11th Cir. 2005).

Here, unlike in Coles, the defendants raised Sixth Amendment objections in the district court. Because we cannot conclude that the district court would have sentenced Henry and Harrison to life imprisonment irrespective of the mandatory nature of the Guidelines, the Government has not established that the error was harmless. We therefore vacate the sentences and remand for resentencing. See United States v. Baugham, 449 F.3d 167, 182-83 (D.C.Cir.2006); United States v. Brown, 449 F.3d 154, 159-60 (D.C.Cir. 2006); Ayers, 428 F.3d at 315-16.

For the foregoing reasons, we affirm Henry's and Harrison's convictions but we vacate the sentences and remand the case to the district court for resentencing under Booker and 18 U.S.C. § 3553(a).

So ordered.

KAREN LECRAFT HENDERSON, Circuit Judge, concurring.

I fully agree that we should affirm Henry's and Harrison's convictions. I am less certain, however, that we should remand for a second resentencing.

A Booker error is prejudicial if "there would have been a materially different result, more favorable to the defendant, had the sentence been imposed in accordance with the post-Booker sentencing regime."* United States v. Coles, 403 F.3d 764, 767 (D.C.Cir.2005). As Coles acknowledges, "[t]here undoubtedly will be some cases in which a reviewing court will be confident that a defendant has suffered no prejudice." Id. at 769. In such a case, "`there would be no basis for thinking that if [the judge] had known that the sentencing guidelines [were] merely advisory he would have given the defendant a lighter sentence,'" id. (quoting United States v. Paladino, 401 F.3d 471, 483 (7th Cir. 2005)), and thus no reason exists to remand for resentencing.

While I agree that a sentence imposed at the top of a Guidelines range does not without more constitute harmless error, the record in this case reveals more. Although the trial judge did not explicitly state that he would have imposed the same sentences were the Guidelines not mandatory, he has twice sentenced Henry and Harrison to the maximum sentence of life imprisonment. He originally imposed the mandatory sentence of life imprisonment. Upon remand for resentencing, the judge again sentenced Henry and Harrison to life imprisonment after calculating a lower Guidelines range of 360 months to life imprisonment. Therefore, the judge not only rejected a lower sentence within a particular Guidelines range, he rejected a lower sentence in a lower range. Indeed, a "conscientious judge" who considered mandatory life imprisonment to be excessive would necessarily impose a lighter sentence if a lower mandatory range applied. See Coles, 403 F.3d at 770. Because here the district judge chose not to impose a shorter sentence on the first remand—even with a lower Guidelines range—I believe there is no reason to think that he would reduce the sentence on a second remand with no mandatory Guidelines range.

Furthermore, at least with respect to Harrison, the district judge explained why the life sentence was appropriate. In United States v. Tchibassa, under plain error review, we found no prejudice where the district judge imposed the maximum sentence within the Guidelines range and stated that the sentence was "appropriate to serve as a warning to those who will kidnap Americans abroad and entirely appropriate for the type of actions that occurred here in depriving [the former hostage] not only of his freedom for two months, but basically of his life." 452 F.3d 918, 930 (D.C.Cir.2006) (quotation omitted) (emphasis original). As we explained in Tchibassa:

The judge's strong and unambiguous approval of the sentence imposed, based . . . on its deterrent effect and its proportionality to the crime committed, makes us confident that were the judge given the opportunity to resentence Tchibassa . . . he would not impose a sentence materially more favorable than the one he made plain he considered "appropriate."

Id. Similarly, the court stated at Harrison's resentencing hearing:

The Court finds that a sentence of life is appropriate in this case in light of the defendant's boasting of his lifestyle and his lifestyle, and the need for deterrence provides sufficient reason for the maximum penalty. Dealing in this amount of drugs should result in a sentence of life imprisonment, in this Court's view, and that will provide . . . some deterrence in the community if others were to understand that even though they've never been arrested before, if they deal in this amount of drugs they're going away for the rest of their lives.

7/21/04 Tr. 28-29 (emphasis added). Because the judge offered a "strong and unambiguous approval of the sentence imposed" upon Harrison, I think there is little, if any, reason to believe that he would impose a sentence more favorable to Harrison were he given the opportunity to resentence him. See Tchibassa, 452 F.3d at 930.

KAVANAUGH, Circuit Judge, concurring.

I join the Court's opinion and add this concurrence to note a few broader points about the path of post-Booker jurisprudence in the federal courts.

To review: In Booker, a five-Justice majority of the Supreme Court held that the United States Sentencing Guidelines were unconstitutional under the Fifth and Sixth Amendments to the extent that facts used to increase a criminal sentence (beyond what the defendant otherwise could have received) were not proved to a jury beyond a reasonable doubt. United States v. Booker, 543 U.S. 220, 226-27, 125 S. Ct. 738, 160 L. Ed. 2d 621 (2005) (Stevens, J., joined by Scalia, Souter, Thomas, and Ginsburg, JJ.). The logical upshot of this part of Booker (what is known as the Booker constitutional opinion) is that the Constitution is satisfied by a sentence in which sentencing facts are proved to a jury beyond a reasonable doubt.

In some tension with the Booker constitutional opinion, however, a different five-Justice majority of the Booker Court also held (in what is known as the Booker remedial opinion) that the constitutional problem with the Guidelines is more readily solved not by requiring sentencing facts to be proved to a jury beyond a reasonable doubt, but instead by making the Guidelines one factor in the district court's sentencing decision, along with other factors specified in 18 U.S.C. § 3553(a). Id. at 245-46, 260-61, 125 S. Ct. 738 (Breyer, J., joined by Rehnquist, C.J., and O'Connor, Kennedy, and Ginsburg, JJ.); cf. id. at 302, 125 S. Ct. 738 (Stevens, J., dissenting in part, joined by Scalia and Souter, JJ.) ("[B]y repealing the right to a determinate sentence that Congress established in the SRA, the Court has effectively eliminated the very constitutional right Apprendi sought to vindicate."). The Booker remedial opinion emphasized, however, that the sentencing court still "must consult" the Guidelines and "take them into account when sentencing." Id. at 264, 125 S. Ct. 738. The Booker remedial opinion also directed appellate courts to review district court sentences for "reasonableness"—a term not defined, but which the Court stated would help "to avoid excessive sentencing disparities while maintaining flexibility sufficient to individualize sentences where necessary." Id. at 264, 125 S. Ct. 738.

In light of the Booker remedial opinion and § 3553(a)'s requirement that district courts "shall consider" the Guidelines, as well as § 3553(a)'s express goal of avoiding unwarranted sentencing disparities, this Court and other federal courts after Booker have held that the Guidelines remain central to sentencing. In part because the "reasonableness" of a sentence is not self-defining and because the Booker remedial opinion said that appellate review would help maintain uniformity, appellate courts have relied on the Guidelines as the predominant substantive standard against which to measure a sentence's reasonableness. Indeed, many courts of appeals, including this one, have accorded a "presumption of reasonableness" to within-Guidelines sentences. See United States v. Dorcely, 454 F.3d 366, 376 (D.C.Cir. 2006); see generally United States v. Buchanan, 449 F.3d 731, 735-41 (6th Cir. 2006) (Sutton, J., concurring). And appeals courts have found many below-Guidelines sentences to be "unreasonable." The post-Booker appellate jurisprudence in turn has exerted further hydraulic pressure on district courts to rely heavily on the Guidelines in sentencing criminal defendants. It thus may be something of a misnomer to call the Guidelines "advisory" with respect to current sentencing practices given that appeals courts often assess the propriety of a district court sentence in part by reference to the Guidelines.

As we review what has happened since Booker, there is no denying that the post-Booker system in substance closely resembles the pre-Booker Guidelines system in constitutionally relevant respects. See Michael W. McConnell, The Booker Mess, 83 DENV. U.L. REV. 665, 678 (2006) ("All the things that troubled Sixth Amendment purists about the pre-Booker Guidelines system are unchanged."); see also Douglas A. Berman & Stephanos Bibas, Making Sentencing Sensible, 4 OHIO STATE J. CRIM. L. 37, 53 (2006); Douglas A. Berman, Tweaking Booker: Advisory Guidelines in the Federal System, 43 HOUS. L. REV. 341, 347-55 (2006). Four of the five Justices who joined the Booker remedial opinion, including its author Justice Breyer, did not find any constitutional problem with the Guidelines to begin with. So it is understandable that the current system as applied is not a major departure from the pre-Booker Guidelines system. Cf. Booker, 543 U.S. at 312-13, 125 S. Ct. 738 (Scalia, J., dissenting in part) (stating that Booker remedial opinion may convey message that "little has changed" from mandatory Guidelines system and posing question: "Will appellate review for `unreasonableness' preserve de facto mandatory Guidelines by discouraging district courts from sentencing outside Guidelines ranges?").

To be sure, district and appeals courts now take some additional and important procedural steps (as exemplified again by today's per curiam opinion). But the bottom line, at least as a descriptive matter, is that the Guidelines determine the final sentence in most cases. And notwithstanding the Booker constitutional opinion, many key facts used to calculate the sentence are still being determined by a judge under a preponderance of the evidence standard, not by a jury beyond a reasonable doubt. The oddity of all this is perhaps best highlighted by the fact that courts are still using acquitted conduct to increase sentences beyond what the defendant otherwise could have received—notwithstanding that five Justices in the Booker constitutional opinion stated that the Constitution requires that facts used to increase a sentence beyond what the defendant otherwise could have received be proved to a jury beyond a reasonable doubt.

In short, we appear to be back almost where we were pre-Booker. And if that is so—and if the lower courts' effort to harmonize the competing goals of the Booker opinions has become the jurisprudential equivalent of a dog chasing its tail—it makes sense to examine how current sentencing practices square not just with Booker but with underlying constitutional principles.

The disagreement in Booker (and in earlier cases such as Apprendi v. New Jersey, 530 U.S. 466, 120 S. Ct. 2348, 147 L. Ed. 2d 435 (2000), and Blakely v. Washington, 542 U.S. 296, 124 S. Ct. 2531, 159 L. Ed. 2d 403 (2004)) represents the collision of two starkly different conceptions of how the Fifth and Sixth Amendments apply to criminal sentencing.

The first conception of the Fifth and Sixth Amendments, which might be called the "deference-to-legislatures" model, generally defers to legislatures in defining crimes and enacting sentencing schemes. Under this interpretation, the Fifth and Sixth Amendments generally require that a jury find the elements of the crime (as defined by the legislature) beyond a reasonable doubt. As to sentencing, this approach gives legislatures wide discretion in crafting a mandatory or structured sentencing system; or adopting an unstructured system in which each sentencing judge possesses broad authority to assess a sentence based on the individual background, facts, and circumstances of the offense and offender; or choosing some approach in between. See generally Williams v. New York, 337 U.S. 241, 69 S. Ct. 1079, 93 L. Ed. 1337 (1949); McMillan v. Pennsylvania, 477 U.S. 79, 106 S. Ct. 2411, 91 L. Ed. 2d 67 (1986) (Rehnquist, J.) (opinion of the Court); Booker, 543 U.S. at 326-34, 125 S. Ct. 738 (Breyer, J., dissenting in part); Blakely, 542 U.S. at 314-26, 124 S. Ct. 2531 (O'Connor, J., dissenting); id. at 326-28, 124 S. Ct. 2531 (Kennedy, J., dissenting). Proponents of this approach argue that it has prevailed throughout most of our history, as courts have generally respected and adhered to legislative choices with respect to sentencing schemes. See Booker, 543 U.S. at 327-28, 125 S. Ct. 738 (Breyer, J., dissenting in part).

The second conception of the Fifth and Sixth Amendments, which might be termed the "real-elements-of-the-offense" model, rests on the constitutionally central role of the jury in the criminal process. This approach begins with the idea that no logical distinction exists between the elements of a crime and so-called sentencing facts that are used to increase a sentence. Because the Constitution requires that the Government prove the elements of a crime to a jury beyond a reasonable doubt, the Constitution also requires that the Government prove substantively similar sentencing facts (such as carrying a weapon during commission of a drug crime) to a jury beyond a reasonable doubt. To do otherwise, this view contends, would be to elevate form over substance and allow legislatures to evade the constitutional requirement that the prosecutor prove the elements of the crime to a jury beyond a reasonable doubt simply by re-labeling elements of the crime as sentencing factors. Under this jurisprudential approach, therefore, courts do not defer to a legislative choice to label a fact as a sentencing factor rather than an element of the crime. See Booker, 543 U.S. at 226-44, 125 S. Ct. 738 (Stevens, J., joined by Scalia, Souter, Thomas, and Ginsburg, JJ.); Harris v. United States, 536 U.S. 545, 572-83, 122 S. Ct. 2406, 153 L. Ed. 2d 524 (2002) (Thomas, J., dissenting); Apprendi, 530 U.S. at 498-99, 120 S. Ct. 2348 (Scalia, J., concurring).

There is an important qualification to this second approach, however, which may explain some of the conceptual and practical difficulty in this area. Despite requiring the jury to find beyond a reasonable doubt the facts used to increase a sentence, the adherents to the real-elements-of-the-offense approach allow purely discretionary sentencing schemes whereby judges "exercise broad discretion in imposing a sentence within a statutory range." Booker, 543 U.S. at 233, 125 S. Ct. 738; see also Apprendi, 530 U.S. at 481, 120 S. Ct. 2348. This concession creates an apparent anomaly: After all, discretionary sentencing systems appear to pose an even greater concern that key facts used to increase a sentence are found by judges—on the record or often silently—by a preponderance of the evidence rather than by juries beyond a reasonable doubt. See Apprendi, 530 U.S. at 548-49, 120 S. Ct. 2348 (O'Connor, J., dissenting) ("[O]ur approval of discretionary-sentencing schemes, in which a defendant is not entitled to have a jury make factual findings relevant to sentencing despite the effect those findings have on the severity of the defendant's sentence, demonstrates that the defendant should have no right to demand that a jury make the equivalent factual determinations under a determinate-sentencing scheme."); Kevin R. Reitz, The New Sentencing Conundrum, 105 COLUM. L. REV. 1082, 1119 (2005). Because the Court has long upheld discretionary sentencing schemes, Chief Justice Rehnquist stated for the Court in 1986 (before the Apprendi-Blakely-Booker cases): "We have some difficulty fathoming why the due process calculus would change simply because the legislature has seen fit to provide sentencing courts with additional guidance." McMillan, 477 U.S. at 92, 106 S. Ct. 2411.

Notwithstanding weighty arguments of the kind made by Chief Justice Rehnquist, the adherents to the real-elements-of-the-offense conception have maintained their approach—and continued to accept discretionary sentencing schemes as a constitutionally acceptable alternative. See Booker, 543 U.S. at 233, 125 S. Ct. 738. As a result, the real-elements-of-the-offense approach to the Constitution seems to mean the following: Legislatures may enact: (i) a discretionary sentencing scheme where the sentencing judge has complete discretion to impose a sentence within the legal range that applies to the crime found by the jury, and the judge may determine the sentence based on the judge's own subsidiary factual determinations, other considerations, or no stated rationale at all; or (ii) a mandatory sentencing scheme where the sentencing judge has no discretion to make factual determinations to increase a sentence. But legislatures, under this real-elements-of-the-offense approach, may not enact an intermediate sentencing scheme where the sentencing judge has structured discretion—in other words, where the sentencing judge must make factual determinations (such as "Did the defendant carry a gun during the drug transaction?") in order to increase a sentence.

How do post-Booker sentencing practices square with the various constitutional approaches described above?

If the deference-to-legislatures conception is correct, then current federal sentencing practices, which largely mirror pre-Booker practices, are obviously constitutionally permissible. Indeed, if this conception is correct, then the Booker constitutional opinion is incorrect and the Sentencing Guidelines should apply as promulgated and made mandatory by Congress.

If the real-elements-of-the-offense approach is correct, however, then current federal sentencing practices may be in tension with the Constitution. That is because the current system—in practice— works a lot like the pre-Booker system: District judges are obliged to apply the Guidelines, and certain facts used to increase a sentence (beyond what the defendant would have received based on the offense of conviction) are found by the judge, not by the jury beyond a reasonable doubt.