Argued and Submitted En Banc March 24, 2005-San Francisco, California
Before: Mary M. Schroeder, Chief Judge, Diarmuid F. O'Scannlain, Michael Daly Hawkins, Sidney R. Thomas, Kim McLane Wardlaw, William A. Fletcher, Raymond C. Fisher, Ronald M. Gould, Johnnie B. Rawlinson, Richard R. Clifton, and Carlos T. Bea, Circuit Judges.
Opinion by Judge Rawlinson; Partial Concurrence and Partial Dissent by Judge Wardlaw; Partial Concurrence and Partial Dissent by Judge Gould; Partial Concurrence and Partial Dissent by Judge O'Scannlain; Partial Concurrence and Partial Dissent by Judge Bea
This case requires us to apply the United States Supreme Court's ruling in United States v. Booker, 125 S.Ct. 738 (2005). In Booker, the Supreme Court struck down the sentencing scheme created by the Sentencing Reform Act of 1984 to the extent that the Act mandated the imposition of sentences predicated on facts not found by the jury or admitted by the defendant. To remedy the constitutional infirmity, the Court severed the mandatory portions of the Act, rendering its sentencing provisions, including the Sentencing Guidelines, effectively advisory. Left unresolved by Booker is the question of what relief, if any, is to be afforded to a defendant who did not raise a Sixth Amendment challenge prior to sentencing. We reheard this case en banc to address this issue for cases pending on direct review.
We are aware that our opinion is of considerable interest to the judges and practitioners in this Circuit who will face a myriad of issues post-Booker. We will not endeavor to foresee or address all potential ramifications of the Booker decision. However, we think it appropriate to amplify the context within which we decide this case in the hope of facilitating the resolution of pending cases.
We are, of course, not the only court of appeals to confront this issue. Our colleagues across the country have also wrestled with the aftermath of Booker. The difficulty of the matter is demonstrated by the fact that the various circuits have taken divergent approaches. We appreciate and have benefitted from their discussions in arriving at our own conclusion.
As described in more detail below, we hold that when we are faced with an unpreserved Booker error that may have affected a defendant's substantial rights, and the record is insufficiently clear to conduct a complete plain error analysis, a limited remand to the district court is appropriate for the purpose of ascertaining whether the sentence imposed would have been materially different had the district court known that the sentencing guidelines were advisory. If the district court responds affirmatively, the error was prejudicial and failure to notice the error would seriously affect the integrity, fairness and public reputation of the proceedings. The original sentence will be vacated by the district court, and the district court will resentence the defendant. If the district court responds in the negative, the original sentence will stand, subject to appellate review for reasonableness. See Booker, 125 S.Ct. at 769. In essence, we elect to follow the approach adopted by the Second Circuit in United States v. Crosby, 397 F.3d 103 (2d Cir. 2005).
I. Factual Background
Defendant Alfred Ameline pled guilty to knowingly conspiring to distribute methamphetamine in violation of 21 U.S.C. §§ 841(a)(1) and 846. The plea agreement approved by the court did not specify the quantity of methamphetamine involved, although at his change of plea hearing Ameline admitted that "some methamphetamine" was involved in the charged conduct. Ameline disputed the government's assertion that the amount of methamphetamine attributable to him was one and one-half kilograms.
The Presentence Report (PSR) prepared by the Probation Office attributed 1,079.3 grams of methamphetamine to Ame-line. That amount resulted in a base offense level of 32, after applying the drug equivalency table from the United States Sentencing Guidelines Manual (Guidelines) § 2D1.1(c). The probation officer's conclusion as to drug quantities was based solely on the investigative reports the officer had reviewed, and the PSR contained a summary of the salient portions of the reports. A two-level enhancement was recommended pursuant to § 2D1.1(b)(1) for possession of a firearm in connection with the charged offense, resulting in an adjusted offense level of 34. With the recommended three-level adjustment for acceptance of responsibility, the recommended total offense level was 31. With a criminal history category of I, the sentencing range recommended in the PSR was 108 months to 135 months.
Ameline filed objections to the PSR, challenging the amount of drugs attributed to him. He also denied the truth of the firearm allegations. However, he did not challenge the recommended drug quantity enhancement as violative of the Sixth Amendment. The probation officer dismissed Ameline's objections and reaffirmed his determination of the quantity of methamphetamine in the original PSR and his recommendation as to the weapons enhancement. Ameline objected to the final PSR finding in his sentencing memorandum to the court.
At the beginning of the sentencing hearing, before any witnesses were called, the district judge informed the parties how he intended to proceed:
It is the position of this court in this matter, as it is in all such cases, that the facts as recited in the presentence report are prima facie evidence of the facts set out there; that if the defendant challenges the facts set forth in the presentence report, it is the burden of the defendant to show that the facts contained in the report are either untruthful, inaccurate, or otherwise unreliable.
The district judge then asked defense counsel to call his first witness. However, before counsel called any witnesses, the court again reiterated its intention:
[I]t is my position that the statements in the presentence report, that is, statements of fact, are reliable on their face and prima facie evidence of the facts there stated. And I will be taking those into account to the extent relevant to the obligations that I have in fashioning sentence and fixing responsibility for drug quantities, if they are not overcome by other evidence presented at this hearing. Be guided accordingly.
Consistent with his objections, Ameline testified and presented witnesses to refute the drug amounts attributed to him in the PSR. The government contended that an even larger amount of drugs should be attributed to Ameline, based on transactions not included in the PSR recommendation. No specific testimony was directed toward the firearm enhancement.
At the conclusion of the sentencing hearing, the district court found that 1,603.60 grams of methamphetamine were attributable to Ameline. That finding resulted in a base offense level of 34, two levels higher than that recommended in the PSR. The PSR described two additional transactions, but the probation officer did not include those transactions in calculating the recommended drug amount. The district court, however, included the amounts involved in the described transactions, thus establishing a higher base offense level. The district court stated:
I should let all parties know that all findings are based upon a preponderance of the evidence standard and are established at least to that standard in the view of the court.
The district court found the § 2D1.1(b)(1) weapons enhancement "undisputed," and applied a two-level enhancement for an offense level of 36, but deducted three points for timely acceptance of responsibility, for a total offense level of 33. The district court sentenced Ameline to 150 months, in the middle of the 135 to 168-month Guidelines range.
Ameline appealed. In his opening brief, Ameline challenged the district court's allocation of the burden of proof and the reliability of the hearsay evidence used to prove drug quantity. Ameline did not initially contest the preponderance of the evidence standard employed by the district court or the propriety of judicial factfinding under a mandatory sentencing regime.
After the case was submitted for decision by a three-judge panel of our court, but before a decision was filed, the Supreme Court announced its decision in Blakely v. Washington, 124 S.Ct. 2531 (2004). In light of Blakely, our panel held that the determination of material sentencing facts by the district judge under a preponderance of the evidence standard, rather than by a jury as part of its verdict, violated Ameline's Sixth Amendment rights and amounted to reversible plain error. United States v. Ameline [Ameline I], 376 F.3d 967, 980 (9th Cir. 2004). The panel vacated Ameline's sentence and remanded with instructions that, if necessary, a jury must determine the amount of drugs attributable to Ameline and whether he possessed a weapon in connection with the offense. Id. at 983.
Within days of the filing of the panel decision in Ameline I, the Supreme Court granted certiorari and scheduled oral argument in Booker and a related case, United States v. Fanfan, 125 S.Ct. 12 (2004). Booker and Fanfan raised issues regarding the application of Blakely to federal sentencing.
That led our court to defer further action on this case until after the Supreme Court announced its decision in those cases.
After the Supreme Court's decision in Booker was announced, the panel issued an amended opinion. United States v. Ameline [Ameline II], 400 F.3d 646 (9th Cir. 2005). As before, the panel concluded that the district court had committed reversible plain error because Ameline's sentence "exceeded 'the maximum authorized by the facts established by a plea of guilty or a jury verdict.' " Id. at 653 (quoting Booker, 125 S.Ct. at 756). The panel vacated the original sentence and remanded for resentencing. Id. at 657-58.
We vacated the panel opinion and granted rehearing en banc. United States v. Ameline, 401 F.3d 1007 (9th Cir. 2005). Although Ameline did not challenge the constitutionality of the Guidelines in the district court or in his opening brief on appeal, we conclude, as did the three-judge panel, that it is nonetheless appropriate to permit him to raise those issues. See United States v. Valdez, 195 F.3d 544, 547 n.3 (9th Cir. 1999).
II. The Booker Decision
Before Booker, sentencing judges were bound by the Guidelines. See 18 U.S.C. § 3553(b)(1). After conviction, the Guidelines required the sentencing judge to make factual findings about the defendant and the offense and then, based on the conviction and facts found independently by the court, determine the appropriate sentencing range. Once the correct sentencing range was determined, departure from that range was authorized only for reasons stated in the Guidelines or where the court finds that there exists an aggravating or mitigating circumstance of a kind, or to a degree, not adequately taken into consideration by the Sentencing Commission in formulating the guidelines that should result in a sentence different from that described.
18 U.S.C. § 3553(b)(1).
 In Booker, the Supreme Court held that the Guidelines as constituted violated the Sixth Amendment. Booker, 125 S.Ct. at 756. That outcome followed from the conclusion that the Sixth Amendment precludes a judge from enhancing a sentence based on extra-verdict findings (other than the fact of prior conviction) in a mandatory sentencing regime. Id. at 748-49. The majority opinion, authored by Justice Stevens, observed that "[i]f the Guidelines as currently written could be read as merely advisory provisions that recommended, rather than required, the selection of particular sentences in response to differing sets of facts, their use would not implicate the Sixth Amendment." Id. at 750. Therefore, if a particularly prescient sentencing judge, pre-Booker, had made and used the same extra-verdict findings under the same mandatory guidelines regime, but made clear that he was treating the Guidelines as advisory rather than binding, no Sixth Amendment violation would have occurred under Booker. See id. ("[W]hen a trial judge exercises his discretion to select a specific sentence within a defined range, the defendant has no right to a jury determination of the facts that the judge deems relevant."); see also United States v. Coles, 403 F.3d 764, 768 (D.C. Cir. 2005); United States v. Paladino, 401 F.3d 471, 483 (7th Cir. 2005).
 A separate majority of the Court remedied the Sixth Amendment infirmity in the federal sentencing scheme by making the Guidelines effectively advisory. The remedial portion of Booker, authored by Justice Breyer, agreed that "without this provision-namely the provision that makes 'the relevant sentencing rules mandatory and imposes binding requirements on all sentencing judges'-the statute falls outside the scope" of the Sixth Amendment's jury trial requirement. Booker, 125 S.Ct. at 764 (citations and alteration omitted). Rather than engraft a jury trial requirement onto the mandatory sentencing guideline system, the remedial opinion severed from the Reform Act "the provision that requires sentencing courts to impose a sentence within the applicable Guidelines range (in the absence of circumstances that justify a departure) and the provision that sets forth standards of review on appeal, including de novo review of departures from the applicable Guidelines range." Id. (citations omitted).
 It is crucial for our current purpose to appreciate the distinction drawn by the Supreme Court and by us. Standing alone, judicial consideration of facts and circumstances beyond those found by a jury or admitted by the defendant does not violate the Sixth Amendment right to jury trial. A constitutional infirmity arises only when extra-verdict findings are made in a mandatory guidelines system. This conclusion has been reached by a majority of the appeals courts that have decided Sixth Amendment sentencing issues post-Booker. See United States v. Antonakopoulos, 399 F.3d 68, 75 (1st Cir. 2005); United States v. Williams, 399 F.3d 450, 458 (2d Cir. 2005); United States v. Mares, 402 F.3d 511, 518 (5th Cir. 2005); Paladino, 401 F.3d at 482-83 (7th Cir.); United States v. Pirani, No. 03-2871, 2005 WL 1039976, at *5 (8th Cir. Apr. 29, 2005) (en banc); United States v. Lawrence, No. 02-1259, 2005 WL 906582, at *12 (10th Cir. Apr. 20, 2005); United States v. Rodriguez, 398 F.3d 1291, 1300 (11th Cir. 2005); United States v. Smith, 401 F.3d 497, 499 (D.C. Cir. 2005) (per curiam).
III. Application of the Plain Error Standard of Review
 As explained, because the Sixth Amendment error was not raised in the district court, to warrant relief the error must constitute plain error. Plain error is "(1) error, (2) that is plain, and (3) that affects substantial rights." United States v. Cotton, 535 U.S. 625, 631 (2002) (citation, alteration and internal quotation marks omitted). If these three conditions of the plain error test are met, an appellate court may exercise its discretion to notice a forfeited error that (4) "seriously affects the fairness, integrity, or public reputation of judicial proceedings." Id. (citation and alteration omitted).1
 An error is plain if it is "contrary to the law at the time of appeal . . ." Johnson v. United States, 520 U.S. 461, 468 (1997). Ameline's claim of sentencing error meets this requirement, because Booker expressly invalidated the federal sentencing guidelines. See Booker, 125 S.Ct. at 755-56. The sentencing judge's enhancement of Ameline's sentence in reliance upon judge-made findings under the then-mandatory guidelines, was, therefore, constitutional error. See id.
A more vexing inquiry lies in the third prong of the plain error test: whether the error affected Ameline's substantial rights, that is, whether the outcome of Ameline's sentencing was affected by the erroneous enhancement of Ameline's sentence on the basis of judge-made findings in the mandatory guidelines regime. Ameline bears the burden of persuading us that his substantial rights were affected. He must establish "that the probability of a different result is sufficient to undermine confidence in the outcome of the proceeding." United States v. Dominguez Benitez, 124 S.Ct. 2333, 2340 (2004) (citations and internal quotation marks omitted). Because the error turns on the use of judge-found facts in a mandatory guidelines system and those guidelines are now advisory, Ameline must demonstrate a reasonable probability that he would have received a different sentence had the district judge known that the sentencing guidelines were advisory.
 The difficulty in assessing whether the sentencing error affected Ameline's substantial rights arises because the record does not provide an inkling of how the district court would have proceeded had it known that the Guidelines were advisory rather than mandatory. That is not surprising, since at the time of sentencing, the district court and the parties were operating under the reasonable belief that the Guidelines were mandatory. We surmise that the record in very few cases will provide a reliable answer to the question of whether the judge would have imposed a different sentence had the Guidelines been viewed as advisory. Although on occasion a district court judge has expressed frustration with the binding nature of the Guidelines, it was very rare for a judge, within the record of an individual case, to express that view. Pre-Booker, there simply would have been no need or practical reason for the judge to make such a record, since the judge could not have expected then that it would make a legal difference.
 We conclude that the best way to deal with this unusual situation is to follow the approach adopted by our colleagues on the Second, Seventh, and D.C. Circuits2 and ask the person who knows the answer, the sentencing judge.3 Rather than affirm a sentence that was unconstitutional and may have been prejudicial, we elect to remand to the district court to answer the question whether the sentence would have been different had the court known that the Guidelines were advisory. This is "[t]he only practical way (and it happens also to be the shortest, the easiest, the quickest, and the surest way) to determine whether" there was prejudice. Paladino, 401 F.3d at 483. If the district court responds affirmatively, the error undermines our confidence in the outcome. See id. at 484. On the other hand, if the district court responds in the negative and explains why on the record, the original sentence will stand, provided it is reasonable. See id.; cf. Booker, 125 S.Ct. at 769 (holding that both the Sixth Amendment ruling and the remedial interpretation of the Reform Act, including the reasonableness standard, apply to all cases pending on direct review). The remand contemplated under Crosby is a limited one designed to permit the sentencing judge to inform the reviewing court's analysis of whether the sentencing judge "would have imposed a materially different sentence" had he been aware of the now-advisory nature of the sentencing guidelines. Crosby, 397 F.3d at 117.
In Crosby, the Second Circuit relied on 18 U.S.C. § 3742(f) to support its approach that a third choice is available for election by an appellate court assessing error. Id. That subsection provides in pertinent part:
If the court of appeals determines that-
(1) the sentence was imposed in violation of law . . . the court shall remand the case for further sentencing proceedings with such instructions as the court considers appropriate . . .
18 U.S.C. § 3742(f)(1).
 The Second Circuit reasoned that the power to remand for resentencing necessarily encompasses the lesser power to order a limited remand. Crosby, 397 F.3d at 117. This is consistent with our case law. As we recently explained in United States v. Gunning, 401 F.3d 1145 (9th Cir. 2005), appellate courts are not precluded from limiting the scope of issues on remand, or the district court's consideration of evidence and arguments. Id. at 1148.
The Second Circuit viewed the limited remand as a solution to the dilemma facing appellate courts post-Booker: in the overwhelming majority of cases, we simply do not know whether the sentencing judge would have imposed the same sentence had he known that the sentencing guidelines were advisory, rather than mandatory. As an alternative to presuming prejudice, presuming non-prejudice, venturing to assess prejudice on a case-by-case basis or engaging in wholesale remands in cases where the record is not sufficiently developed to inform the appellate court's plain error analysis, the Second Circuit elected to remand to the district court. See United States v. Gonzalez, No. 04-1956, 2005 WL 1023059, at *3 (2d Cir. May 3, 2005) (citing Crosby, 397 F.3d at 117).
 We recognize that the Seventh and D.C. Circuits, while endorsing limited remands in Booker cases, have adopted a slightly different procedure from that in Crosby. Under the Second Circuit approach, if a district court judge determines that resentencing is warranted after remand from the court of appeals, he or she can simply vacate the sentence and resentence. Crosby, 397 F.3d at 120. The Seventh and D.C. Circuits, on the other hand, retain jurisdiction through the limited remand process, thus requiring the district court to indicate that it would have sentenced differently, the court of appeals to vacate the sentence, and finally, the court of appeals to remand to the district court for resentencing. Paladino, 401 F.3d at 484; Coles, 2005 WL 783069 at *7. The two procedures are very similar, but the Crosby procedure is less cumbersome. See Paladino, 401 F.3d at 484 ("Our procedure is not identical to that set forth in Crosby, though it is very close."); Coles, 2005 WL 783069 at *7 (D.C. Cir. Apr. 8, 2005) ("We note that the 'limited remand' procedures adopted by the Second and Seventh Circuits offer slightly different approaches."). Booker itself does not require the appellate court to undertake the additional step of vacating and remanding if the district court indicates that it wishes to resentence. Therefore, to facilitate the expeditious handling of Booker cases, we follow the Second Circuit in not retaining jurisdiction throughout the limited remand.
We acknowledge that this limited remand approach has not been adopted by all courts to face the post-Booker problem. Some circuits have held that when the reviewing court cannot determine whether the Sixth Amendment error was prejudicial, the defendant has not carried his burden and relief must be denied. See Antonakopoulos, 399 F.3d at 80 (1st Cir.); Mares, 402 F.3d at 521-22 (5th Cir.); Pirani, 2005 WL 1039976, at *6-7 (8th Cir.); Rodriguez, 398 F.3d at 1301 (11th Cir.). We do not quarrel with this position as a matter of doctrine; that is the way plain error review normally works.
As we indicated above, however, assigning such a burden to a defendant in this context requires him to demonstrate a sufficient probability that the district court would have imposed a different sentence under an advisory system, even though there would have been no reason for the sentencing court to so indicate. The sentencing court presumably knows the answer to the relevant question and would likely have made a record had it known at the time of sentencing that it would make a difference. Resolving unpreserved sentencing error through a limited remand is comparably easy and yields a result that is certain. See Williams, 399 F.3d at 457-59. If we decline to find out what the district court knows unless the defendant can make a showing of something over which he had no control, the defendant will surely feel abused, with some justification, and everyone will be left to wonder about whether the sentencing court might have acted differently. It seems to us that would itself undermine the fairness, integrity and public reputation of the judicial proceedings, something which we should try to avoid.4
Other circuits have, like our own decision in Ameline II, taken another approach. Notably, some have held that "where a defendant's sentence was enhanced based on facts neither admitted to nor found by a jury, . . . the defendant can demonstrate plain error and may be entitled to resentencing." Davis, 2005 WL 976941, at *1 (3d Cir.); Hughes, 401 F.3d at 548-49 (4th Cir.); Oliver, 397 F.3d at 379-80 (6th Cir.). Following that reasoning, prejudice is determined by comparing the sentence that the defendant could have received based solely on the jury's verdict (or on facts otherwise admitted by defendant) with the sentence that he actually received. If the former sentence would have been more favorable to the defendant, the defendant was prejudiced. See, e.g., Hughes, 401 F.3d at 548-49. As discussed above, however, we view judicial fact-finding as erroneous only when coupled with a mandatory guidelines system. Moreover, upon remand the district court will consult the guidelines as required by Booker and will be free to impose exactly the same sentence. "[I]f the judge would have imposed the same sentence even if he had thought the guidelines merely advisory . . . and the sentence would be lawful under the post-Booker regime, there is no prejudice to the defendant." Paladino, 401 F.3d at 483; accord Williams, 399 F.3d at 459; Antonakopoulos, 399 F.3d at 80-81. On balance, therefore, we conclude that the limited remand approach is preferable. As described by Judge Posner, it is the middle way between placing on the defendant the impossible burden of proving that the sentencing judge would have imposed a different sentence had the judge not thought the guidelines were mandatory and requiring that all defendants whose cases were pending when Booker was decided are entitled to be resentenced, even when it is clear that the judge would impose the same sentence and the court of appeals would affirm.
Paladino, 401 F.3d at 484-85.
We come at last to the fourth prong of plain error review. Here we examine whether a plain and demonstrably prejudicial error "seriously affects the fairness, integrity, or public reputation of [the] judicial proceedings." Johnson v. United States, 520 U.S. 461, 467 (1997) (citation and alteration omitted). This inquiry hinges on the question presented to the district court. If the district court would have imposed a different sentence in an advisory regime, we "exercise [our] discretion to notice [the] forfeited error," id., because "it is a miscarriage of justice to give a person an illegal sentence, just as it is to convict an innocent person," Paladino, 401 F.3d at 483. The original sentence will be vacated, and the district court, with the defendant present, will resentence in accordance with the Reform Act as modified by Booker. See Crosby, 397 F.3d at 120.
Our colleagues in dissent criticize our adoption of the approach articulated in Crosby, characterizing the limited remand procedure as contradicting Booker, abdicating our obligation to conduct appellate review, subsuming an inaccurate prejudice inquiry, disregarding district court judges who have left the bench, embracing illusory efficiencies, and encouraging cursory review. See e.g., Wardlaw Concurring/ Dissenting Opinion, p. 6375; Gould Concurring/Dissenting Opinion, pp. 6403, 6412, 6417. Despite our colleagues' thoughtful presentation of their views, we remain convinced that the limited remand procedure set forth in Crosby best resolves this unique issue that has arisen in the wake of the Supreme Court's holding in Booker.
Unfortunately, we cannot look to Booker for guidance in assessing plain error, because Booker did not address plain error in the context of resolving the issues in that case. See Antonakopoulos, 399 F.3d at 79 n.10 (noting that no plain error issue was before the Court in Booker). Although it is true that the Supreme Court did not approve a limited remand procedure, neither did it prohibit one. Similarly, we cannot conclude from the fact that Booker's case was remanded for resentencing that remand for resentencing is appropriate in all cases. The Supreme Court expressly instructed otherwise. See Booker, 125 S.Ct. at 769 ("Nor do we believe that every appeal will lead to a new sentencing hearing."). The fact remains that different approaches have been taken by courts post-Booker. We remain of the view that the Crosby approach is consistent with Booker.
 Nor do we agree that the limited remand procedure abdicates our obligation to conduct appellate review. To the contrary, the procedure provides for more accurate appellate review. Without the benefit of the district court's views, we would be left with review of a potentially misleading record. District court judges often make remarks at sentencing for purposes other than fact-finding. A district court judge may choose to say some encouraging words for the benefit of the defendant's family; a district court judge may decide to lecture the defendant with a warning. District court judges have also been known to make stray comments about the Guidelines during sentencing, without necessarily intending for them to be interpreted as meaning that a different sentence would have been imposed under a discretionary sentencing scheme. It would be a mistake for us to attribute fresh meaning to comments made in an entirely different context. It would also be a mistake to infer from a district court's silence that the district court would not have made a different decision under a different sentencing scheme. In sum, in this unusual context, our ability to assess plain error based on the cold record is significantly impaired. Although no system is perfect, our appellate plain error review will be better informed and more accurate by obtaining the district court's findings. Because a vacation of the sentence would be required if the district court would have imposed a different sentence under a discretionary sentencing system, it is also appropriate for us to direct the district court to proceed with resentencing if the Booker error prejudiced the defendant. Identifying legal error and providing direction to the district court on how to cure it is a quintessentially appellate function. Adopting a limited remand procedure to effectuate resolution of legal error does not abdicate our appellate responsibility in the least. Indeed, it is entirely consistent with our usual procedures in such circumstances. See, e.g., Gunning, 401 F.3d at 1148 (describing the prior holding in which the panel remanded to ask the district court "to make findings on the record if it had already considered the minor role adjustment, and to resentence if necessary.")
Further, despite our dissenting colleague's interpretation of the majority opinion as encompassing "every pre-Booker defendant asserting plain error," the limited remand is invoked only when it cannot be determined from the record whether the judge would have imposed a materially different sentence had he known that the Guidelines are advisory rather than mandatory. See Williams, 399 F.3d at 458-59 (observing that the limited remand procedure is appropriate where the record leaves uncertainty regarding what sentence would have been imposed absent error).5
Only after determining that the record did not sufficiently inform the reviewing court's analysis, and only after concluding that the first two prongs of the plain-error test were met did the court in Crosby remand the matter to the district court. See Williams, 399 F.3d at 460. Rather than being a dereliction, the remand was designed to avoid "leaving in place an error-infected sentence" due to the lack of an adequate record.6
See id. at 461.
Our dissenting colleague also takes us to task for "asking the wrong question of the district court," and for undertaking "to predict the likely outcome of a remand." However, the limited remand does not seek a response to the question of what the district court would do on remand. Instead, the limited remand seeks a response to the question of whether the district court would have imposed a materially different sentence at the time of sentencing had it known that the Guidelines were advisory rather than mandatory. See Paladino, 401 F.3d at 485. This inquiry mirrors the holdings of Dominguez Benitez, 124 S.Ct. at 2340, and Kotteakos v. United States, 328 U.S. 750, 764 (1946) (describing the appropriate inquiry as the effect on the jury's decision, i.e., the outcome).7
As noted, we cannot and should not attempt to anticipate and address every issue that may arise in the course of resolving pre-Booker sentencing appeals. Many of the "illusory efficiency" scenarios described by our dissenting colleagues are of that nature. Although it is true that a limited remand could result in a subsequent appeal, so could a resentencing. Nevertheless, the relative prospect of future appeals should not deter us from adopting a process that we view as facilitating reliable appellate review. And although efficiency is an important consideration in the administration of justice, see Government Emp. Ins. Co. v. Dizol, 133 F.3d 1220, 1226-27 (9th Cir. 1998) (recognizing the importance of judicial economy), it is not the most important consideration in the context of this case. More important is the opportunity under the Crosby approach to engage in a more accurately informed plain error review. See Coles, 403 F.3d at 770. Indeed, if efficiency were paramount, the procedure of choice would be one that declined to recognize most forfeited Booker errors. See e.g., Rodriguez, 398 F.3d at 1304 (explaining that "where the record does not provide 'any indication' that there would have been a different sentence," the defendant loses.).
Finally, we do not share the concern regarding cursory review on remand that is articulated in the "dissent from the Seventh Circuit's decision to adopt the same type of limited remand rule." We are confident that our conscientious and able colleagues on the district courts throughout this Circuit will continue to give each case the individualized attention it deserves.
IV. The Process to be Followed
 In hopes of making clear the process that we conclude should be followed, we outline it here. We begin with our own review. Booker explicitly stated that its holding applies to all cases pending on direct appeal. Booker, 125 S.Ct. at 769. Even where the briefs filed by the parties do not raise a Booker objection, we conclude that the issue may be raised and should be considered.
When faced with an unpreserved Booker/Fanfan error, the reviewing panel must first determine if an eligible party wants to pursue the subject. Although either the defendant or the government may raise the non-constitutional error that a sentence was erroneously imposed under guidelines believed to be mandatory, Booker/Fanfan 125 S.Ct. at 769, the Sixth Amendment objection - that the defendant's sentence was enhanced by judge-found facts under a mandatory Guidelines system - is available to the defendant alone.8 However, because we do not assume that every defendant will want to pursue resentencing, the limited remand procedure must include an opportunity for defendants with pending appeals to opt out of resentencing by promptly notifying the district court judge. See Crosby, 397 F.3d at 118.
If an eligible party seeks resentencing under Booker/ Fanfan, we will then engage in the plain error analysis described in this opinion. If that analysis leads the panel to the same dead end that we reach here, where it is not possible to reliably determine from the record whether the sentence imposed would have been materially different had the district court known that the Guidelines were advisory, we will remand to the sentencing court to answer that question.
In answering the question we pose, the district court need not determine or express what the sentence would have been in an advisory system. It is enough that the sentence would have been materially different. We agree with the Second Circuit that the "views of counsel, at least in writing," should be obtained. See Crosby, 397 F.3d at 120.
If the district court judge determines that the sentence imposed would not have differed materially had he been aware that the Guidelines were advisory, the district court judge should place on the record a decision not to resentence, with an appropriate explanation. A party wishing to appeal the order may file a notice of appeal as provided in Fed. R. App. P. 4(b).9
If the district court determines that the sentence imposed would have differed materially if the district court judge were applying the Guidelines as advisory rather than mandatory, the error was prejudicial, and the failure to notice the error would seriously affect the integrity, fairness and public reputation of the proceedings.10 In such a case, the original sentence will be vacated and the district court will resentence with the defendant present. In resentencing the defendant, the district court is permitted to take a fresh look at the relevant facts and the Guidelines consistent with Booker, the Sentencing Reform Act of 1984, Rule 32 of the Federal Rules of Criminal Procedure, and this opinion. See Gunning, 401 F.3d at 1148; see also United States v. Matthews, 374 F.3d 872, 880 (9th Cir. 2004). In either case, the defendant and the government have the right to appeal to this court the district court's decision, including a challenge to the sentence based on the reasonableness standard established in Booker. Booker, 125 S.Ct. at 765-66.
V. Erroneous Allocation of the Burden of Proof
 The government has conceded that the district court erred in placing the burden of proof on the defendant to disprove the factual basis for the base level offense and sentence enhancements sought by the government. The government "bear[s] the burden of proof for any fact that the sentencing court would find necessary to determine the base offense level." United States v. Howard, 894 F.2d 1085, 1090 (9th Cir. 1990). As we explained in Howard, "[s]ince the government is initially invoking the court's power to incarcerate a person, it should bear the burden of proving the facts necessary to establish the base offense level." Id. Here, by placing the burden on Ameline to disprove the factual statements made in the PSR, the district court improperly shifted the burden of proof to Ameline and relieved the government of its burden of proof to establish the base offense level.
Of course, the district court may rely on undisputed statements in the PSR at sentencing. United States v. Charles-worth, 217 F.3d 1155, 1160 (9th Cir. 2000). However, when a defendant raises objections to the PSR, the district court is obligated to resolve the factual dispute, see Fed. R. Crim. P. 32(i)(3)(B), and the government bears the burden of proof to establish the factual predicate for the court's base offense level determination. Howard, 894 F.2d at 1090. The court may not simply rely on the factual statements in the PSR.
 The government also bears the burden of proof when it seeks sentence enhancements. Id. at 1089. As we explained in Howard, the party seeking an adjustment bears the burden of proof. Id. Thus, when the defendant requests a downward adjustment, the defendant bears the burden of proof; when the government seeks an upward adjustment, it bears the burden of proof. Id. Here, the district court also erred by placing the burden of proof on the defendant to disprove the upward adjustment recommended in the PSR and sought by the government.
The fact that the Sentencing Guidelines have become discretionary following Booker does not alter this analysis. The district court's factual findings will determine the base offense level, which remains the starting point for determining the applicable guideline range for an offense under 21 U.S.C. § 841(a)(1). See U.S.S.G. § 2D1.1(c). When a defendant contests the factual basis of a PSR, the district court remains obligated to resolve the dispute before exercising its sentencing discretion under Booker. In resolving the factual dispute, the district court must continue to apply the appropriate burdens of proof, consistent with Howard.
We adopt the limited remand procedure articulated by the Second Circuit in Crosby to assess the existence of plain error in pre-Booker sentencing appeals.
 In this case, remand is also required to address the district court's error in misallocating the burden of proof at sentencing, in light of the presentation of evidence by Ameline challenging the drug amounts recommended in the PSR. To correct the Howard error, the district court must hold a new sentencing hearing in accordance with Federal Rule of Criminal Procedure 32(i). Accordingly, we vacate the sentence and remand for further proceedings consistent with this opinion.
SENTENCE VACATED and REMANDED.
WARDLAW, Circuit Judge, with whom GOULD, Circuit Judge, joins, and with whom O'SCANNLAIN and BEA, Circuit Judges, join with respect to Part I, concurring in part and dissenting in part:
The district court erred by treating the factual statements in the Presentence Report as presumptively accurate and placing the burden on Ameline to disprove them. See United States v. Howard, 894 F.2d 1085, 1090 (9th Cir. 1990). I therefore concur in the majority's decision to vacate Ameline's sentence and to remand his case for a new sentencing hearing pursuant to Federal Rule of Criminal Procedure 32. I respectfully dissent, however, from the majority opinion because it improperly delegates to the district courts our discretionary appellate function to conduct plain error review of unpreserved claims of Booker error. The resulting wholesale remand of hundreds, possibly thousands, of sentencing appeals elevates our administrative concerns over the law as pronounced by the Supreme Court in United States v. Booker, 125 S.Ct. 738 (2005), and over the interests of individual justice.
I. Flaws in the Majority's "Limited Remand" Approach
The majority accurately characterizes the Booker error. A constitutional infirmity arises only where a defendant's sentence is enhanced under a mandatory Guidelines system as a result of factual findings made by the sentencing judge beyond the facts admitted to by the defendant or found by a jury. The majority is also correct that, after Booker, such an error is plain. But those are only the first two prongs of the plain error test. We must also determine whether the plain error affected the defendant's "substantial rights." United States v. Cotton, 535 U.S. 625, 631 (2002). If this third prong is met, we may then exercise our discretion to notice the forfeited error, but only if the "error seriously affect[s] the fairness, integrity, or public reputation of judicial proceedings." Id. (quotations omitted).
Neither party to this litigation urged us to adopt the Second Circuit's Crosby approach, which relegates to the district courts the determination of the third and fourth prongs of plain error review. See United States v. Crosby, 397 F.3d 103 (2d Cir. 2005). Indeed, both parties advocated against our adoption of Crosby. Yet, after examining the analyses of our sister circuits, the majority selects Crosby for the reason that it is a "practical," "short," "easy," "quick," and "sure" way to dispose of unpreserved claims of Booker error. I agree that the majority's approach is "short," "quick," and "easy," at least for our court. It relieves our docket of "literally thousands of cases" (in the government's words). It relieves us of our obligation to review each of those cases individually. It permanently disposes of at least some portion of those cases, such as where a defendant will have already served his unconstitutional sentence by the time the district courts, with their own heavy dockets, absorb the avalanche of simultaneous remands and are able to make the time to decide whether or not to resentence. We will have removed several appeals from our own crowded docket by requiring the district courts to further crowd theirs, as they do our job. Although the burden-shifting approach the majority arrives upon may be "short," "easy," and "quick," it is neither correct nor just.
The majority's delegation of the third and fourth prongs of plain error review to the district courts has several flaws: (1) it contradicts what the Supreme Court said and did in Booker; (2) it abdicates our obligation as a reviewing court to conduct plain error review; (3) the question it poses for the district courts is not an accurate inquiry under the prejudice prong of plain error review; (4) it fails to make provision for the number of district judges who have left the bench or who have become otherwise incapacitated since imposing the unconstitutional sentences under review, and who, therefore, are incapable of answering its question; and (5) any efficiencies that it purportedly promises will prove illusory.
The majority's approach is at odds with what the Supreme Court said and did in Booker. There, the Court instructed "reviewing courts to apply ordinary prudential doctrines, determining, for example, whether the issue was raised below and whether it fails the 'plain-error' test." Booker, 125 S.Ct. at 769 (emphasis added). The Court directed appellate courts to fulfill their review function, and did not even hint that delegating that function to the district courts might be appropriate. "The Supreme Court phrased its instructions to reviewing courts in terms of what has long been accepted practice in the federal courts and did not suggest that this accepted practice must change in the wake of Booker." United States v. Milan, 398 F.3d 445, 454 (6th Cir. 2005).11
The majority's approach also conflicts with what the Supreme Court did in Booker. The Court reviewed Booker's case and found that the district court "imposed a sentence higher than the maximum authorized solely by the jury's verdict," a Sixth Amendment violation. Booker, 125 S.Ct. at 769. The Court remanded the case for resentencing, ordering that "[o]n remand, the District Court should impose a sentence in accordance with today's opinions, and, if the sentence comes before the Court of Appeals for review, the Court of Appeals should apply the standards set forth in this opinion." Id. The Court saw no need to ask the district court to determine in the first instance whether resentencing was warranted. It thus made clear that appellate courts should review an unpreserved claim of Booker error for whether it is plain error, and remand for resentencing in appropriate cases. Nor did the Court incorporate the remedial portion of its decision into the question of whether there was error-i.e., it did not ask whether the sentence would have been different under advisory Guidelines, reviewing the error in Booker's case without regard to its contemporaneous announcement that the Guidelines were advisory.
The significance of the Supreme Court's decision to remand in Booker is not undermined by the government's failure to argue plain error principles in that case. According to the majority, because the government had waived the argument that Booker's appeal was governed by the plain error standard of review, Booker provides no guidance for assessing unpreserved claims of Booker error. The majority's refusal to turn to Booker for guidance, however, disregards our law that to vacate and remand a sentence requires some determination of prejudice whether or not the issue is preserved; plain error principles change only the identity of the party who bears the burden of persuasion. See United States v. Olano, 507 U.S. 725, 734 (1993) (substantial rights prong of plain error analysis mirrors harmless error analysis but shifts the burden of persuasion with respect to prejudice to the defendant). Inherent in the Court's decision to remand Booker's case for resentencing is necessarily a decision that the error in his case was prejudicial, i.e., was not harmless and affected Booker's substantial rights. See Fed. R. Crim. P. 52(a) ("Any error . . . that does not affect substantial rights must be disregarded."). If the Court had not found the Booker error to be prejudicial, it could not have found that error to properly form the basis for vacating and remanding for resentencing.
Nor did the Court find any impediment to reversing Fanfan's sentence and allowing "the Government (and Fanfan should he so choose) [to] seek resentencing under the system set forth in [Booker]." Booker, 125 S.Ct. at 769. In Fanfan's case, the district court imposed a sentence lower than the sentence authorized by the jury's verdict and thus lower than what the then-mandatory Guidelines dictated. The Court understood that resentencing might be an undesirable option for Fanfan, as he risked the imposition of a higher sentence. Nevertheless, it readily permitted a remand for a full resentencing, even in the absence of a Sixth Amendment violation. Id.
2. Abdication of Our Appellate Review Function
Although the majority labels its approach a "limited remand," the approach is in fact an abdication of our appellate review function. "We have used the term 'limited remand' to describe a remand to the district court for proceedings prior to this court's consideration of the merits of an appeal." United States v. Washington, 172 F.3d 1116, 1118 (9th Cir. 1999) (emphasis added). Under the majority's approach, the district court conducts the third and fourth prongs of plain error review, resentences if it deems necessary, and reenters judgment. If a notice of appeal is filed, that judgment may become the subject of a new appeal, but our consideration of the merits of the defendant's original appeal is eliminated. Contrary to the majority's asserted reason for the remand-"to permit the sentencing judge to inform the reviewing court's analysis" of the third prong of plain error review, ante at 6360-the sentencing judge is under no obligation to actually inform us as to whether he "would have imposed a materially different sentence had he been aware of the now-advisory nature of the sentencing guidelines." Ante at 6360. Nor do we retain jurisdiction of the case. As a result, we will never make an informed appellate decision as to whether the Booker error affected the defendant's original sentencing proceeding, and the fourth, discretionary, decision with which we are charged under plain error review simply vanishes. Therefore, despite its best effort to make it appear otherwise, the majority is simply relinquishing to the district courts the last two prongs of plain error review.
As part of its efforts to justify its abdication of our appellate review function, the majority discusses at length several ways in which district judges create "potentially misleading records." Whether or not district judges create "potentially misleading records," it is a fact of life that appellate courts must review cold records; that's their job and the argument is no justification for abandoning that important work. This argument also conflicts with the majority's description of "The Process to be Followed," which requires that a three-judge panel review the entire "potentially misleading" record before ordering a "limited remand." Only "[i]f that analysis leads to the same dead end that [the majority] reach[es] here, where it is not possible to reliably determine from the record whether the sentence imposed would have been materially different" under the now-advisory Guidelines, ante at 6369, may the panel issue a "limited remand." The majority never describes at what point a panel may reach a "dead end," or when it will be impossible for a panel "to reliably determine" "whether the sentence imposed would have been materially different" under the now-advisory Guidelines. Ante at 6369. Unbelievably, the majority, acting in haste to dispose of unpreserved claims of Booker error, has left wide open the nature and extent of review in which any subsequent panel should engage before ordering a "limited remand."12
The majority also ignores our precedents that demonstrate that the limited remand device is used not to assign our traditional review function to the district court, but to elicit the district court's decision on a question it earlier missed, or to permit the district court to engage in its own traditional fact-finding function. In United States v. Gunning, 339 F.3d 948 (9th Cir. 2003) (per curiam), cited by the majority, Gunning argued that he was entitled to a two-level minor role adjustment under U.S.S.G. § 3B1.2(b). The district court made no findings regarding the adjustment. "Because it is for the district court to rule on sentencing issues in the first instance," we remanded for the court to make additional findings and to resentence if necessary. Id. at 949. We did not remand to the district court to allow it to determine whether it had committed a legal error in sentencing in the first place. Therefore, the majority's reliance on Gunning in support of its novel approach to plain error review is inapposite.
Similarly, in United States v. Hovsepian, 359 F.3d 1144 (9th Cir. 2004), we held that because the district court did not consider appellees' convictions or their possible affiliations with terrorist groups, the court failed to undertake a complete analysis of their naturalization applications. Id. at 1168. We remanded the case to the district court for additional fact-finding-a task well within the traditional province of district courts, see Fed. R. Civ. P. 52(a)-not for a determination of whether the district court committed legal error. Id. at 1169. To the contrary: We not only retained jurisdiction over the appeal, we further ordered:
Within 120 days of the issuance of the mandate, or such further time as this court may allow, the district court shall forward its additional findings of fact and conclusions of law to this court, with copies to the parties, so that we may review the district court's assessment of all the relevant facts in reaching a final disposition. Within 30 days after the district court forwards its findings and conclusions, any party may request this court's further review of the naturalization issue.
Id. (first emphasis added).
The majority cites no Supreme Court authority to support its delegation of the last two prongs of plain error review to the discretion of the district courts. Nor can it, because review for error, plain or otherwise, is an exclusive function of appellate courts. Booker itself makes this clear, as do a host of other Supreme Court decisions. See United States v. Dominguez Benitez, 124 S.Ct. 2333, 2340 (2004) (a defendant who seeks reversal on the ground that the district court committed plain error must satisfy the judgment of the "reviewing court" that the probability of a different result is sufficient to undermine confidence in the outcome of the proceeding); United States v. Vonn, 535 U.S. 55, 59 (2002) (a "reviewing court" must consider the effect of any error on substantial rights); Kotteakos v. United States, 328 U.S. 750, 763-64 (1946) (in deciding the prejudice prong of plain error, "appellate courts" must consider what effect the error had or reasonably may have had upon the outcome of the proceedings). As the Eleventh Circuit has recognized, "[t]he determination of plain error is the duty of courts of appeal, not district courts." Rodriguez, 398 F.3d at 1305; see also United States v. Mares, 402 F.3d 511, 522 (5th Cir. 2005) ("[W]e find no support for [the Crosby approach] in the Supreme Court plain error cases. Those cases place the obligation on the appellate courtsrather than the district courts-to determine the third prong of the plain error test.").
The majority's misreading of 18 U.S.C. § 3742(f) leads it to adopt the Second Circuit's erroneous rationale that "the power to remand for resentencing [under § 3742(f)] necessarily encompasses the lesser power to order a limited remand." Section 3742 governs appellate review of criminal sentences. As Justice Scalia noted in Booker: "Even the most casual reading of this section discloses that its purpose-its only purpose-is to enable courts of appeals to enforce conformity with the Guidelines." Booker, 125 S.Ct. at 791 (Scalia, J., dissenting).13 Thus, § 3742 imposes an obligation on appellate courts to review district courts' sentencing decisions. The majority misuses § 3742 by drawing an inference from that section, which was designed to curtail district courts' discretion,14 to support relinquishing to district courts our appellate review function.
The plain language of § 3742 demonstrates that it cannot carry the weight the majority assigns to it. Section 3742(f) provides, in pertinent part:
If the court of appeals determines that-
(1) the sentence was imposed in violation of law or imposed as a result of an incorrect application of the sentencing guidelines, the court shall remand the case for further sentencing proceedings with such instructions as the court considers appropriate.
18 U.S.C. § 3742(f)(1). This section mandates that "the court of appeals" make the determination that the sentence "was imposed in violation of the law." Id. The majority instead invites the district courts-the very courts whose decisions are under review-to participate in the predicate finding of legal error. Sending the third and fourth prongs of the plain error determination to the district court abrogates the very obligation of appellate review that § 3742(f) mandates. Ordering a "limited remand" is not a lesser power inherent in § 3742(f); it is an entirely different remedy that conflicts with congressional directives as to the review and disposition of sentencing appeals.
The Eleventh Circuit has also found illogic in the Second Circuit's-and now our own-reasoning that, because § 3742(f) authorizes remands for resentencing to remedy a sentence imposed in error, appellate courts necessarily have the "lesser" power to remand for a determination of whether the original sentence was contrary to law:
That conclusion does not follow at all. In cases of non-preserved error appellate courts lack the authority to remand for resentencing where the requirements of the plain error rule have not been met. The remands being ordered are not, as Crosby suggests, for the purpose of determining "whether to resentence," but for the purpose of determining whether the third prong of the plain error test has been met so that the unpreserved error may be noticed and the appealed sentence vacated. Neither § 3742(f) nor any other part of the Sentencing Reform Act purports to give appellate courts the authority to delegate the decision of whether there has been plain error to the very court whose judgment is being reviewed.
Rodriguez, 398 F.3d at 1306 (emphasis added).
By characterizing its approach as a "sure" and "certain" way to determine prejudice, the majority demonstrates its misapprehension of the correct standard for determining whether Booker error affected a defendant's substantial rights. As the Supreme Court most recently stated in Dominguez Benitez, to meet the third prong of the plain error test, "[a] defendant must . . . satisfy the judgment of the reviewing court, informed by the entire record, that the probability of a different result is sufficient to undermine confidence in the outcome of the proceeding." Dominguez Benitez, 124 S.Ct. at 2340 (quotations and citations omitted). Applying this prejudice test to Booker error, the correct inquiry is whether a reasonable probability exists that, but for the Booker error, the outcome of the defendant's sentencing proceeding would have been different so as to undermine our confidence in the sentence. Kotteakos, 328 U.S. at 764 (a reviewing court must determine "what effect the error had or reasonably may be taken to have had upon the outcome of the proceedings" (emphasis added)).
The majority's failure to grasp that the inquiry is objective, rather than subjective, leads to its notion that it is appropriate "to ask the sentencing judge." The majority clings to its erroneous view of the prejudice inquiry to defend the "limited remand" procedure as merely a last resort, after the appellate panel cannot find the answer to whether the district judge "would have imposed a materially different sentence" had the Guidelines been advisory. But that is an irrelevant question under the prejudice inquiry and invites the district judge himself to speculate about what he would have done in a different world. It is doubtful that most judges will even recall the circumstances of every sentencing hearing over which they have presided since Apprendi v. New Jersey, 530 U.S. 466 (2000). As appellate judges, we can and should objectively determine, based on a review of the record, whether it is reasonably probable that the defendant's sentence would have been different but for the Booker error. We do not need, nor is it appropriate to obtain, the subjective views of the sentencing judge. Because the inquiry is not subjective, the views of the sentencing judge are irrelevant. See United States v. Hughes, 401 F.3d 540, 551 (4th Cir. 2005) ("[T]he proper focus is on what actually happened as a result of the error, not what might happen in a subsequent proceeding on remand.").
3. Sentencing Post-Booker
Even if the majority could cite any legal precedent supporting its approach, the abbreviated sentencing determination it requires of the district court is not a substitute for the sentencing proceeding required after Booker. In the post-Booker world, district courts must consider the factors provided in 18 U.S.C. § 3553(a) in fashioning the appropriate sentence for the individual defendant. As stated by Justice Breyer in his remedial opinion for the Court:
Without the "mandatory" provision, the [Sentencing Reform Act] nonetheless requires judges to take account of the Guidelines together with other sentencing goals. See 18 U.S.C.A. § 3553(a) (Supp. 2004). The Act nonetheless requires judges to consider the Guidelines "sentencing range established . . . for the applicable category of offense committed by the applicable category of defendant," § 3553(a)(4), the pertinent Sentencing Commission policy statements, the need to avoid unwarranted sentencing disparities, and the need to provide restitution to victims, §§ 3553(a)(1), (3), (5)-(7) (main ed. and Supp. 2004). And the Act nonetheless requires judges to impose sentences that reflect the seriousness of the offense, promote respect for the law, provide just punishment, afford adequate deter-rence, protect the public, and effectively provide the defendant with the needed educational or vocational training and medical care. § 3553(a)(2) (main ed. and Supp. 2004).
Booker, 125 S.Ct. at 764-65.
District courts now also have the discretion to weigh a multitude of mitigating and aggravating factors that existed at the time of mandatory Guidelines sentencing, but were deemed "not ordinarily relevant," such as age, education and vocational skills, mental and emotional conditions, employment record, and family ties and responsibilities. See U.S.S.G. § 5H1.1-6; see also Koon v. United States, 518 U.S. 81, 95 (1996) ("Discouraged factors . . . are those not ordinarily relevant to the determination of whether a sentence should be outside the applicable guideline range. Examples include the defendant's family ties and responsibilities, his or her educational and vocational skills, and his or her military, civic, charitable, or public service record." (quotations and citations omitted)). Indeed, district courts may even consider factors that were precluded from consideration altogether prior to Booker, such as drug and alcohol dependence, lack of guidance as a youth, and personal financial difficulties and economic pressures. See U.S.S.G. §§ 5H1.4, 5H1.12, and 5K2.12; see also Koon, 518 U.S. at 95 ("The Commission lists certain factors that never can be bases for departure, [such as] lack of guidance as a youth; drug or alcohol dependence; and economic hardship . . . ." (citations omitted)).
The majority leaves open the possibility that these factors -now essential to sentencing consistent with 18 U.S.C. § 3553(a) and its directive that the sentence be "sufficient but not greater than necessary to comply with the purposes of sentencing"-might come to light in a sentencing hearing, but it does not require that a district court hold a hearing before deciding against resentencing the defendant. Indeed, "[i]n all too many instances, the process scripted by the [majority] will serve as an invitation for [a] district court to give only a superficial look at [an] earlier unconstitutionally[ ] imposed sentence." United States v. Paladino, 401 F.3d 471, 486 (7th Cir. 2005) (Ripple, J., dissenting from denial of rehearing en banc). Because under the majority's approach the appellate court will not perform plain error review, such a "quick look" sentencing decision would be subject only to the lesser "reasonableness" standard of review.
Defendants should be accorded the benefit of the complete discretion district courts now enjoy. They should have the opportunity to be heard on § 3553 and its newly relevant factors. Briefs of counsel are not a substitute for testimony, evidence, and argument. The majority fails to explain how, without an evidentiary hearing and briefing tantamount to resentencing by normal vacatur and remand procedures, a district judge could ever accurately answer the question as to whether he would have imposed a materially different sentence had he known that the Guidelines were advisory.
For example, Ameline's trial counsel made the strategic decision to focus on drug quantity and gun possession at Ameline's sentencing hearing because only those factors could have significantly decreased Ameline's offense level under the mandatory Guidelines. By contesting drug quantity and gun possession, Ameline potentially could have reduced his sentence by over 10 years because the drug quantity and gun possession determinations fixed the range within which Ameline was sentenced. By contrast, evidence as to Ameline's "background, character and conduct," see U.S.S.G. § 1B1.4, could have only affected the point within the Guidelines range at which Ameline would be sentenced. After Booker, however, a defendant's "background, character and conduct" are the very factors that might convince a judge to exercise his discretion to disregard the advisory Guidelines range, but only if, upon plain error review, we notice the error, vacate the defendant's sentence, and remand for resentencing. "In short, what the panel substitutes for the usual judicial reaction to an unconstitutionally[ ] imposed sentence is a process that simply is inadequate to the task." Paladino, 401 F.3d at 486 (Ripple, J., dissenting from denial of rehearing en banc).
4. Practical Problems
Chief among the practical problems presented by the majority's approach is that it fails to account for the fact that several district judges have retired, become disabled, or passed away since imposing the unconstitutional sentences currently under review. The recent retirement of the Honorable Lourdes G. Baird from the United States District Court for the Central District of California to become a private mediator is but one of many examples. See The Third Branch, Judicial Milestones, Vol. 37, Number 4 (April 2005), available at http://www.uscourts.gov (last visited May 1, 2004). A glaring imperfection in the majority's "quick" and "easy" approach is: What happens when we cannot "ask the sentencing judge?" The answer is that the Chief Judge of the district court will reassign the remanded case to a currently serving judge who cannot possibly answer the question the majority poses because he was not the original sentencing judge. The new sentencing judge will face additional practical problems in resolving how to handle the case. This will necessarily result in inconsistent applications of sentencing law, which will likely generate future due process and equal protection claims.
5. Illusory Efficiencies
The obvious attractiveness of the Crosby approach inheres in its purported efficiencies. As a reviewing court, however, our obligations under the Constitution are not always measured against a metric of efficiency or administrative convenience. See, e.g., Blakely v. Washington, 124 S.Ct. 2531, 2543 (2004) ("[O]ur decision cannot turn on whether or to what degree trial by jury impairs the efficiency . . . of criminal justice."); Universal Camera Corp. v. N.L.R.B., 340 U.S. 474, 490 (1951) ("Reviewing courts must be influenced by a feeling that they are not to abdicate the conventional judicial function."). As the Supreme Court cautioned in Booker itself, " 'delays, and little inconveniences in the forms of justice, are the price that all free nations must pay for their liberty in more substantial matters.' " Booker, 125 S.Ct. at 756 (quoting 4 Commentaries on the Laws of England 343-44 (1769)). In light of the important constitutional rights at issue, the minimal burdens that accompany individualized review of unpreserved claims of Booker error are worth bearing.
Moreover, any efficiency that the majority's approach may seemingly promise is likely to prove illusory. If a district judge chooses on remand not to resentence a defendant, we may be required to review the standing sentence for reasonableness. Review for "reasonableness" is not limited to consideration of the length of the sentence. Instead, the Booker Court instructed appellate courts, in determining reasonableness, to consider not only the length of the sentence but also the factors evaluated and the procedures employed by the district court. See Booker, 125 S.Ct. at 765-66; see also United States v. Webb, 403 F.3d 373, 382-83 (6th Cir. 2005) (after Booker appellate courts should "consider not only the length of the sentence but also the factors evaluated and the procedures employed by the district court in reaching its sentencing determination"). Ultimately, we may conclude that a sentence is unreasonable when the sentencing judge fails to explicitly consider the factors listed in 18 U.S.C. § 3553(a), or neglects to consult the applicable Guidelines range, and instead selects what he deems an appropriate sentence without such required consideration. In cases where a district judge elects not to hold a hearing before deciding not to resentence,15 however, it will be impossible for us to conduct a reasonableness review, for the judge's "appropriate explanation" for his decision not to resentence cannot replace the insight into the factors evaluated and the procedures employed by the judge that a transcript of a sentencing hearing would provide. This is especially true in light of the majority's failure to discuss what would be an "appropriate explanation." See Paladino, 401 F.3d at 488 (Kanne, J., dissenting from denial of rehearing en banc) ("The record in the case in which there was no resentencing (or hearing on the issue) will be impossible for us to review for reasonableness, if reasonableness is to be determined with regard to all of . . . 'the numerous factors that guide sentencing.' " (quoting Booker, 125 S.Ct. at 765-66)).
In addition to hindering our ability to review reimposed sentences for "reasonableness," the majority opens our docket to countless arguments that would be precluded if we had performed an individualized analysis. If we were to determine under an individualized analysis that a particular defendant did not meet his burden under the plain error standard, the case would not be remanded to the district court and there would be no second appeal. If we were to determine otherwise, the defendant's sentence would be vacated and his case remanded for a full resentencing. However, because the defendant would be entitled to all the procedures that accompany a full resentencing, such as the opportunity to "present all available accurate information bearing on mitigation of punishment," see United States v. Mack, 200 F.3d 653, 658 (9th Cir. 2000), the only issue on appeal likely would be the "reasonableness" of the defendant's new sentence. By contrast, under the majority's approach, if a district court elects not to hold a hearing before deciding to reimpose the pre-Booker sentence, the defendant could raise a host of claims in addition to the reasonableness of his sentence, such as the district court's failure to consider the factors set forth in 18 U.S.C. § 3553(a), and its violation of his right to allocution.
Therefore, while at first blush it may be enticing to an overworked and overscrutinized bench to rid ourselves of hundreds, perhaps thousands, of appeals, we may create much more work for ourselves down the road than if we had simply done it right in the first place. The Eleventh Circuit accurately describes Crosby as "requiring resentencing in order to determine whether resentencing is required."16 Rodriguez, 398 F.3d at 1305. It correctly observes that the Crosby procedure "undermines the Supreme Court's teaching that one of the principal purposes of the plain error rule is to avoid needless reversals and remands." Id. By adopting Crosby, the majority creates "the real possibility of another appeal and another remand on top of that." Id.
One of the reasons for plain error review is to reduce the burden on the judicial system. See Johnson v. United States, 520 U.S. 461, 468 (1997). The majority's approach will necessarily increase the workload of courts, first, at the district court level, and then, inevitably, at the appellate level. The better approach is to review these appeals as the Supreme Court instructed and as our appellate duty requires.
II. Individualized Review of Unpreserved Claims of Booker Error
Where the majority and I part company is with respect to the third and fourth prongs of the plain error test. The majority improperly delegates to the district courts the tasks of determining whether Booker error affected defendants' substantial rights, and whether to notice the forfeited error. As the First, Fifth, and Eleventh Circuits have held, however, we must analyze these aspects of plain error at the appellate level by reviewing the entire record of the sentencing proceedings. See United States v. Antonakopoulos, 399 F.3d 68, 75 (1st Cir. 2005); Mares, 402 F.3d at 522; Rodriguez, 398 F.3d at 1304.
1. "Substantial Rights"
For an error to affect "substantial rights," "the error must have been prejudicial: It must have affected the outcome of the district court proceedings." Olano, 507 U.S. at 734. "It is the defendant rather than the Government who bears the burden of persuasion with respect to prejudice." Id.
The Supreme Court has articulated several different formulations of a defendant's burden under the substantial rights test. See Dominguez Benitez, 124 S.Ct. at 2342 (Scalia, J. dissenting) ("By my count, this Court has adopted no fewer than four assertedly different standards of probability relating to the assessment of whether the outcome of trial would have been different if error had not occurred, or if omitted evidence had been included."). Most recently, in the context of a Rule 11 violation, the Court explained that a defendant "must show a reasonable probability that, but for the error, he would not have entered the plea." Id. at 2339. More generally, the Court stated:
A defendant must thus satisfy the judgment of the reviewing court, informed by the entire record, that the probability of a different result is sufficient to undermine confidence in the outcome of the proceeding.
Id. at 2340 (quotations and citations omitted).
Applying this prejudice test to Booker error, we must determine whether there is a reasonable probability that, but for the Booker error, the outcome of the defendant's sentencing proceeding would have been different so as to undermine our confidence in the sentence. Because the Booker Court held that an enhancement based on judge-found facts would not have implicated a constitutional violation if the Guidelines were advisory rather than mandatory, Booker, 125 S.Ct. at 750, in determining the prejudicial effect of the Booker error, we must focus upon the mandatory nature of the system under which the district judge enhanced the defendant's sentence rather than upon the use of judge-found facts in sentencing. See Rodriguez, 398 F.3d at 1303 ("The prejudice inquiry must focus on what has to be changed to remedy the error.").
We must also consider the profound impact of the Supreme Court's decision in Booker on federal sentencing, because the nature of the error may affect the application of the substantial rights prong. See Olano, 507 U.S. at 735. To understand the transforming nature of Booker, it helps to consider the nature of sentencing proceedings before the implementation of the mandatory federal Guidelines in 1987, which drastically altered the practice of sentencing in the federal criminal justice system.
Before the Guidelines were implemented, Congress had delegated virtually "unfettered discretion to the sentencing judge to determine what the sentence should be within the customarily wide range" of potential sentences prescribed by statute. Mistretta v. United States, 488 U.S. 361 (1989). The role of probation officers was to facilitate, rather than to circumscribe, judicial discretion. To this end, the primary purpose of the presentence report was to " 'aid the court in determining the appropriate sentence,' " and included " '[a]n assessment of the problems of the defendant and a consideration for the safety of the community.' " Kate Stith and José A. Cabranes, Judging Under the Federal Sentencing Guidelines, 91 NW U. L. Rev. 1247, 1250 (1997) (quoting Probation Div. Admin. Office of the U.S. Courts, Pub. No. 105, The Presentence Investigation Report 1 (1987)). The presentence report suggested a particular term of imprisonment, based on national sentencing statistics and the officer's informed judgment, but the role of the probation officer in sentencing was purely advisory. See id. at 1251. The sentencing judge could make use of the information in the presentence report or disregard it; the judge had complete discretion to balance the "various goals of sentencing, the relevant aggravating and mitigating circumstances, and the way in which these factors would be combined in determining a specific sentence." United States Sentencing Commission, Report to Congress: Downward Departures From the Federal Sentencing Guidelines, I, B-1 (2003) (quotations and citation omitted). The sentencing judge was free to impose a sentence anywhere within the statutory range, and could usually replace imprisonment with probation, if he so chose. See Mistretta, 488 U.S. at 365. Finally, judicial discretion was "unfettered" primarily because appellate review of the district judge's discretion was virtually non-existent. See Dorszynski v. United States, 418 U.S. 424, 431 (1974) ("[O]nce it is determined that a sentence is within the limitations set forth in the statute under which it is imposed, appellate review is at an end."). Judicial sentences that fell within the prescribed range "were virtually unreviewable on appeal." Mistretta, 488 U.S. at 363.
Booker is a signal event in the development of the law. It restored the complete discretion of sentencing judges to sentence anywhere within the statutory range, so long as the sentence is tethered to the congressional goals of sentencing set forth in 18 U.S.C. § 3553(a). As the D.C. Circuit recently stated, a sentencing court will have "committed no error . . . so long as the sentence was within the prescribed statutory range and is otherwise reasonable." United States v. Coles, 403 F.3d 764, 768-69 (D.C. Cir. 2005). Similarly, the Fourth Circuit, concluding that the district court engaged in "careful deliberation," held in a recent case that "because the district court sentenced [the defendant] within the statutory guidelines . . . we are of opinion [that] the sentence . . . is reasonable." United States v. Bartram, ___ F.3d ___, 2005 WL 994828 *4 (4th Cir. April 29, 2005). Moreover, the Sixth Circuit, also reviewing a sentence for post-Booker reasonableness, has adopted an abuse of discretion standard, thereby explicitly reverting to the standard applicable in the pre-Guidelines era. See United States v. Dalton, 404 F.3d 1029, 1032 (8th Cir. 2005) (stating that Booker "extended unreasonableness-the pre-PROTECT Act standard for reviewing departure sentences-to the review of all sentences," and proceeding to review the district court's sentencing decision for abuse of discretion).
Given that we have reverted to a sentencing system very close to that which existed before the Guidelines, it would be fundamentally unfair to require that a defendant demonstrate with 100 percent certainty that the district judge would have sentenced him differently. Rather, the relaxed inquiry of Dominguez Benitez is appropriate: we should determine whether there is a reasonable probability that, but for the Booker error, the outcome of the defendant's sentencing proceeding would have been different so as to undermine our confidence in the sentence.
Ameline, himself, has met this burden. Setting aside the Howard error, the Booker error so infected Ameline's sentencing hearing that Ameline's burden of showing a reasonable probability that without the judicial fact-finding and mandatory enhancements there would have been a different sentence is easily met. We need not seek the district court's guidance as to whether the Booker error affected Ameline's substantial rights. In Ameline's case, the sentencing judge made factual findings by a preponderance of the evidence and mandatorily enhanced Ameline's sentence above the Guidelines range applicable to the facts as admitted by Ameline in his plea agreement. Moreover, Ameline was sentenced under an unconstitutional mandatory scheme that forbade the exercise of the full measure of discretion that, after Booker, should be accorded to sentencing judges. And, due to the mandatory nature of the Guidelines when Ameline was sentenced, the district judge was precluded from complying with 18 U.S.C. § 3553(a)'s directive to "impose a sentence sufficient, but not greater than necessary," to comply with the statutory purposes of sentencing set forth in § 3553(a)(2). In addition, Ameline vigorously disputed the amount of methamphetamine that the government and the probation officer sought to attribute to his offense conduct. Ameline did not argue for reductions based on his "background, character and conduct," U.S.S.G. § 1B1.4, because, at that time, such factors were irrelevant to a determination of his Guidelines range. Finally, the record reflects that this was Ameline's first serious drug conviction and that, at the time of allocution before the district judge, he expressed remorse for his wrongful conduct. These circumstances are sufficient to satisfy the prejudice inquiry and to establish a violation of Ameline's substantial rights.
The Eleventh Circuit asked the right question in Rodriguez in determining whether Booker error affected Rodriguez's substantial rights: whether there is a reasonable probability that there would have been a different sentencing disposition had the sentencing judge been aware that the Guidelines were advisory. Nonetheless, it came up with the wrong answer, holding that, because in most cases the answer will be "we don't know," except in extraordinary circumstances, a defendant will not be able to meet his burden of showing that his substantial rights have been affected. Rodriguez, 398 F.3d at 1301. The Dominguez Benitez prejudice test is not limited to what the district judge said on the record; rather, our decision must be "informed by the entire record." Dominguez Benitez, 124 S.Ct. at 2340. Review of the entire record, including the presentence report, should in most cases reveal objective data from which we can determine to a reasonable probability whether a sentence would have differed but for the Booker error.
The Eleventh Circuit's overly demanding approach derives from a misapplication of the Supreme Court's plain error analysis in Jones v. United States, 527 U.S. 373 (1999), in which the Court considered an unpreserved jury instruction error. As noted above, the Supreme Court redefined the prejudice inquiry in Dominguez Benitez. See Dominguez Benitez, 124 S.Ct. at 2339. Furthermore, in applying Jones, the Eleventh Circuit failed to consider the unique nature of the sentencing process, which involves a choice of a point within a broad range of discretion rather than an either-or choice, like guilt or innocence. See Paladino, 401 F.3d at 482 ("Guilt is either-or; the defendant is either guilty or innocent. . . . But sentencing is not either-or; it is the choice of a point within a range established by Congress, and normally the range is a broad one.").
Under the Eleventh Circuit's approach, a pre-Booker defendant's right to resentencing depends largely upon whether the district court stated on the record that it felt constrained by the Guidelines or that it would have sentenced the defendant to a lower sentence if the Guidelines were advisory rather than mandatory. See Rodriguez, 398 F.3d at 1301 (holding that Rodriguez failed to meet his burden under the third prong of the plain error test because "[t]he record provides no reason to believe any result is more likely than the other"); see also United States v. Shelton, 400 F.3d 1325, 1333-34 (11th Cir. 2005) (applying Rodriguez and finding plain error only because the record reflected the sentencing judge's frustration with the constraints imposed by the Guidelines). Although explicit statements of frustration or similar sentiments in the record would facilitate a finding of prejudice, such statements by the sentencing judge are not required to show prejudice.
To the contrary: Because district judges were required to follow the mandatory Guidelines, it would be unreasonable to require defendants to point to criticism of, or disagreement with, the Guidelines to establish a violation of their substantial rights. Before Blakely, we, along with other courts, had repeatedly instructed sentencing courts to impose sentences within the applicable mandatory Guidelines range, with limited exceptions. "This well-established case law substantially undermined any need or incentive for sentencing courts [pre-Blakely] to note their objections and reservations in sentencing defendants under the then-mandatory Guidelines." United States v. Barnett, 398 F.3d 516, 528 (6th Cir. 2005).
Finally, at bottom, the Rodriguez court failed to meaningfully recognize Booker's profound impact on federal sentencing-restoring the discretion to the sentencing court that the mandatory Guidelines had stripped away. Nor did the Eleventh Circuit discuss the nature of the constitutional violation, which the Supreme Court has said can only be remedied by making the Guidelines advisory. The Eleventh Circuit gives short shrift to the very nature of the sentencing process and imposes a fundamentally unfair burden on defendants by requiring a defendant to demonstrate certainty in a different outcome under the new sentencing scheme to gain the benefit of a constitutionally sound sentencing process.
2. "Fairness and Integrity"
Unlike the majority, I believe we are required as an appellate court to analyze the fourth prong of the plain error test. I would hold that the Booker error affected the fairness and integrity of Ameline's sentencing proceedings. Although Ameline admitted to only a detectable amount of methamphetamine, and vigorously challenged the reliability of the hearsay evidence presented in the Presentence Report to increase his base offense level, the district court, under the mandatory Guidelines, imposed a sentence that violated Ameline's Sixth Amendment rights. As I have emphasized, the sentencing judge was constrained by the mandatory features of the Guidelines and had no opportunity to exercise the full measure of discretion that was restored by Booker. Letting Ameline's sentence stand "simply because it may happen to fall within the range of reasonableness unquestionably impugns the fairness, integrity, or public reputation of judicial proceedings." United States v. Hughes, 396 F.3d 374, 381 n.8 (4th Cir. 2005), amended on denial of reh'g en banc by 401 F.3d 540 (2005). "Moreover, declining to notice the error on the basis that the sentence actually imposed is reasonable would be tantamount to performing the sentencing function ourselves." Id. Therefore, I would hold that the district court's imposition of a 150-month sentence under the mandatory Guidelines in violation of Ameline's Sixth Amendment rights as construed by Booker was plain error under Cotton, and that the error is a ground independent of the Howard error for vacating Ameline's sentence and remanding his case for a new sentencing hearing.
I, therefore, dissent from the majority's "limited remand" approach to analyzing whether unpreserved claims of Booker error constitute reversible plain error. Although I agree that Ameline is entitled to a new Rule 32 sentencing hearing on the basis of the Howard error, I believe that he is entitled to that remedy, apart from the Howard error, because the Booker error in his case was plain error, reversible under Cotton.
GOULD, Circuit Judge, with whom WARDLAW, Circuit Judge, joins, and O'SCANNLAIN and BEA, Circuit Judges, join in part, concurring in part and dissenting in part:
I concur in Parts I and V of the majority opinion, and in the judgment vacating Ameline's sentence and remanding for the district court to give Ameline a new sentencing hearing consistent with our precedent in United States v. Howard, 894 F.2d 1085, 1090 (9th Cir. 1990). I also concur in Judge Wardlaw's partial concurrence and partial dissent concerning the proper approach under United States v. Booker, 125 S.Ct. 738 (2005), but write separately to elaborate my views.
The relief of resentencing for Ameline is required by Howard, 894 F.2d at 1090, and by the government's concession of error.17 An imprisoned person who shows a right to relief on appeal must receive it in due measure, else due process is not satisfied. See Hicks v. Oklahoma, 447 U.S. 343, 347 (1980) (holding that state appellate court deprived petitioner of liberty without due process of law by "simply affirm[ing] the sentence imposed . . . under the invalid mandatory [sentencing] statute"); see also Lassiter v. Dep't of Soc. Servs., 452 U.S. 18, 24 (1981) ("[D]ue process . . . . expresses the requirement of 'fundamental fairness.' ").
Although I join Judge Wardlaw's dissent concerning our procedures following Booker, I write separately to express my own views in disagreement with the majority's "limited remand" approach derived from the Second Circuit's United States v. Crosby, 397 F.3d 103, 117-18 (2d Cir. 2005). I think it wrong to delegate to the district courts the task of determining whether a defendant's substantial rights have been affected, because this is contrary to the Supreme Court's precedents on the plain error standard of review.
Our review is for plain error because Ameline did not object during his sentencing hearing on the ground that the mandatory Guidelines or the procedures used to determine the material sentencing facts were unconstitutional. See United States v. Cotton, 535 U.S. 625, 628-29 (2002). A defendant seeking relief under the plain error standard must show that there was:
(1) error, (2) that is plain, and (3) that affect[s] substantial rights. If all three conditions are met, an appellate court may then exercise its discretion to notice a forfeited error, but only if (4) the error seriously affect[s] the fairness, integrity, or public reputation of judicial proceedings.
Johnson v. United States, 520 U.S. 461, 466-67 (1997) (alterations in original) (internal quotation marks and citation omitted); see also United States v. Recio, 371 F.3d 1093, 1100 (9th Cir. 2004).
I agree with the majority's analysis of the first, second and fourth prongs of the plain error test. The first prong is met because Ameline's Sixth Amendment rights as construed in Booker were violated by the district court's imposition of a sentence that was enhanced based upon judicial fact-finding and was selected under a mandatory Guidelines system. See United States v. Olano, 507 U.S. 725, 732-33 (1993) ("Deviation from a legal rule is 'error' unless the rule has been waived."); Maj. Op. at 6358.
The second prong is met because it is clear after Booker that, given the mandatory nature of the Guidelines at the time of Ameline's sentencing, increasing Ameline's punishment based on facts not admitted by him or determined by a jury beyond a reasonable doubt was contrary to his Sixth Amendment rights. See Johnson, 520 U.S. at 468 ("[W]here the law at the time of trial was settled and clearly contrary to the law at the time of appeal[,] it is enough that an error be 'plain' at the time of appellate consideration."); Maj. Op. at 6358.
The fourth prong of the test, which limits our discretion to notice forfeited Booker errors to cases where the errors seriously affected the fairness and integrity of the defendants' sentencing proceedings, Johnson, 520 U.S. at 467, is met because "it is a miscarriage of justice to give a person an illegal sentence, just as it is to convict an innocent person." Maj. Op. at 6364 (citing United States v. Paladino, 401 F.3d 471, 483 (7th Cir. 2005)).
However, the majority errs in its analysis of the third prong of the plain error inquiry, and the majority's departure from our traditional appellate role may have negative implications for many cases. The majority follows the Second Circuit's "limited remand" approach, delegating to the district courts the task of determining whether a defendant's substantial rights have been affected. Crosby, 397 F.3d at 117-18. I cannot accept the majority's decision to adopt this approach, because it is inconsistent with the Supreme Court's precedents on what is required to satisfy the "substantial rights" test.
For an error to affect "substantial rights," it "must have been prejudicial: It must have affected the outcome of the district court proceedings." Olano, 507 U.S. at 734. "It is the defendant rather than the Government who bears the burden of persuasion with respect to prejudice." Id. To carry this burden, the "defendant must . . . satisfy the judgment of the reviewing court, informed by the entire record, that the probability of a different result is sufficient to undermine confidence in the outcome of the proceeding." United States v. Dominguez Benitez, 124 S.Ct. 2333, 2340 (2004) (internal quotation marks omitted) (holding that defendant claiming that plain Rule 11 error affected his substantial rights "must show a reasonable probability that, but for the error, he would not have entered the plea.").
Applying Dominguez Benitez's substantial rights test in the Booker context, the Eleventh Circuit explained that "[t]he prejudice inquiry must focus on what has to be changed to remedy the error." United States v. Rodriguez, 398 F.3d 1291, 1303 (11th Cir. 2005). I agree. The Supreme Court made clear in Booker that the error in pre-Booker sentencing cases is not simply that sentences were enhanced based on judge-found facts, but that judge-found facts were used to arrive at Guidelines sentences that were binding on the district courts. 125 S.Ct. at 750 (Stevens, J., Op. of the Ct.) ("If the Guidelines as currently written could be read as merely advisory provisions . . . their use would not implicate the Sixth Amendment."); id. at 764 (Breyer, J., Op. of the Ct.) ("With these two sections [that make the Guidelines mandatory] excised, the remainder of the Act satisfies the Court's constitutional requirements."). Thus, our assessment of the prejudicial effect of a Booker error turns on whether the record reflects that there is a "reasonable probability" that the outcome of Ameline's sentencing proceeding would have been different if the district court had considered the Guidelines to be advisory when sentencing, rather than on the mere fact that Ameline's sentence was enhanced based on judicial findings of fact. See United States v. Mares, 402 F.3d 511, 521 (5th Cir. 2005) (holding that under substantial rights prong court must inquire whether outcome of sentencing proceeding would have been different if sentencing judge had considered the Guidelines advisory); Paladino, 401 F.3d at 483 (same); United States v. Antonakopoulos, 399 F.3d 68, 75 (1st Cir. 2005) (same); Rodriguez, 398 F.3d at 1301 (same); Crosby, 397 F.3d at 117-18 (same). But see United States v. Davis, No. 02-4521, 2005 WL 976941, at *1-2 (3d Cir. Apr. 28, 2005) (holding that defendant can meet substantial rights prong by showing either that sentence was enhanced based on judge-found facts or that the district court erroneously believed Guidelines were mandatory at time of sentencing); United States v. Hughes, 401 F.3d 540, 548-49 (4th Cir. 2005) (holding that substantial rights prong was satisfied because defendant's sentence was increased beyond the sentence supported by the facts found by the jury or admitted by defendant); United States v. Oliver, 397 F.3d 369, 379-80 (6th Cir. 2005) (same).
Ameline's substantial rights were affected because extra-verdict facts were used to enhance his sentence under Guidelines thought to be mandatory. At the time of his sentencing, which took place before Booker restored sentencing discretion to the district courts, Ameline challenged the amount of methamphetamine that the government sought to use in calculating his base offense level. This was Ameline's first drug conviction,18 and Ameline expressed remorse for his wrongdoing.
Then there is the procedural error that affected Ameline's sentencing hearing. The district court, contrary to our holding in Howard, 894 F.2d at 1090, placed the burden on Ameline to disprove the drug quantities attributed to him in the Presentence Report. Had the district court required the government to meet its burden of proof to establish the factual predicate for the amount of methamphetamine it attributed to Ameline for purposes of determining the base offense level under U.S.S.G. § 2D1.1, the district court may have attributed a lesser drug quantity to Ameline. This is especially so, in light of the multiple layers of hearsay Ameline alleged that the district court relied upon in making its drug quantity determination. Maj. Op. at 6353-54. A lesser drug quantity in turn might have led the district court to use an entirely different Guidelines range, and may have also prompted the district court to exercise discretion, if any were available, to give Ameline a lighter sentence.
Indeed, the district court stated at Ameline's sentencing hearing that it was "somewhat baffl[ed]" by Ameline's "election to engage in this very serious and extremely detrimental-to-the-community conduct" because of Ameline's "history and the fact that [he was] of an age that [he was], and ha[d] in some ways held [him]self out as a responsible individual." That the district court expressed surprise that Ameline had trafficked in the large quantities of methamphetamine attributed to him by the Presentence Report bolsters the conclusion that there is a reasonable probability that the outcome of Ameline's sentencing proceeding would have been different if the district court had used the proper factfinding procedure and had not been bound by the Guidelines at the time of sentencing.
The government argues that Ameline cannot show a reasonable probability that he would have received a different sentence if the Guidelines had been discretionary at the time of his sentencing because the district court's comments suggested that the court deliberately opted not to exercise even the limited discretion that it had under the mandatory Guidelines, and instead imposed a sentence in the middle of the applicable Guidelines range after considering the circumstances of Ameline's offense conduct. In fashioning its sentence, the district court noted that Ameline's conduct had allegedly involved "multiple transactions of distribution of [an] unlawful substance . . . over an extended period of time," and expressed its view that this conduct was "very serious and extremely detrimental to the community" and "without question . . . hurt . . . many people."
While we should not minimize the seriousness of Ameline's wrongful conduct, these statements by the district court could be used to describe the heartland of any drug offense. In addition, it is a non sequitur for the government to argue that we can discern, from Ameline's middle-of-the-Guidelines-range sentence, whether the district court would have imposed the same sentence if free of the mandatory nature of the Guidelines. It is equally as likely that the district court had a policy of consistently sentencing in the middle of the Guidelines range. See United States v. Barnett, 398 F.3d 516, 528 (6th Cir. 2005) ("That the district court chose to sentence Barnett in the middle of [the mandatory Guidelines] range does not necessarily suggest that the district court would now feel that 265 months of imprisonment is the proper sentence for Barnett. Nor does it suggest that the court would not have sentenced Barnett to a lower sentence if it had the discretion, which it does now, to apply the Guidelines in an advisory fashion.").
I acknowledge that we cannot be certain what sentence the district court would have imposed had it treated the Guidelines as advisory only. As some of our sister circuits have already pointed out, "We just don't know." Rodriguez, 398 F.3d at 1301; see also Paladino, 401 F.3d at 483; Crosby, 397 F.3d at 118. The sentencing record often does not reflect the subtleties that may influence the district court's sentencing discretion or how the human dynamics of a sentencing hearing may influence the court's ultimate sentencing decision. Absent an express statement by the district court that it would have either sentenced a defendant more favorably or reached the same result if the Guidelines were not mandatory, we can never be certain how any particular district court would have exercised discretionary sentencing authority.
However, I disagree with the reasoning of those who favor a blanket approach to handling Booker plain error cases (whether they advocate denying relief or remanding in all cases) because of this uncertainty, for certainty has never been required to satisfy the prejudice inquiry in the context of plain error review. Rather, all that is required is that there be a "reasonable probability" that the error complained of affected the outcome of the proceedings. Dominguez Benitez, 124 S.Ct. at 2340.
Specifically with respect to a blanket denial rule such as the one urged by the government, it would be most unfair and unreasonable for us to ground our prejudice analysis on the fortuity of whether a defendant was sentenced by a particular sentencing judge who openly criticized the Guidelines or indicated that they were unduly harsh as applied to a particular defendant rather than one who felt constrained quietly to follow the law. Our judicial system is predicated on judges following the law and, so far as is humanly possible, keeping their personal opinions to themselves. See Coalition for Econ. Equity v. Wilson, 122 F.3d 692, 717 (9th Cir. 1997) (Norris, J., dissenting from denial of reh'g en banc) (quoting Planned Parenthood v. Casey, 505 U.S. 833, 868 (1992) (O'Connor, Souter & Kennedy, JJ., joint op.) ("It is the responsibility of all federal judges . . . to 'struggle to accept' [the burden of rendering decisions of which they disapprove] . . . . 'because [we] respect the rule of law.' ").19 And before Booker, we consistently required sentencing courts to comply with the mandates of the Guidelines. See, e.g., United States v. Chastain, 84 F.3d 321, 323-26 (9th Cir. 1996); United States v. Vilchez, 967 F.2d 1351, 1355 (9th Cir. 1992). "This well-established case law substantially undermined any need or incentive for sentencing courts pre-Booker to note their objections and reservations in sentencing defendants under the then-mandatory Guidelines." Barnett, 398 F.3d at 529. Given these realities, we cannot now require defendants seeking to establish that their substantial rights have been affected by Booker error to present us with these types of statements, which sentencing courts had no reason to make on the record under our prior case law. Cf. Antonakopoulos, 399 F.3d at 81 ("The existence of prejudice should not turn on how vocal the district judge was."). I would hold that Ameline has met his burden of showing that there is a reasonable probability that he would have received a different sentence if the district court had considered the Guidelines to be advisory when sentencing, notwithstanding that there is no affirmative statement by the district court to this effect.
The majority, however, does not express any holding on whether the substantial rights requirement has been satisfied here in Ameline's case. Instead the majority follows the Second Circuit in instituting a blanket "limited remand" rule for cases involving unpreserved Booker error where there is no clear statement from the district court in the record about what it would do if the Guidelines were merely advisory. Under this approach, all such cases will be remanded so that the district court can consider whether it would have imposed the same sentence if the Guidelines had been advisory at the time of sentencing. Maj. Op. at 6359-60; Crosby, 397 F.3d at 117-18. If the district court concludes that the sentence would have been the same, there is no prejudice to the defendant, and the initial sentence stands. Maj. Op. at 6359-60; Crosby, 397 F.3d at 117. But if the district court determines that it would have given the defendant a materially different sentence under the Booker rules, then the defendant's substantial rights were affected by the error, and he or she is entitled to resentencing. Maj. Op. at 6359-60; Crosby, 397 F.3d at 117. This "ask the district court" approach is wrong for many reasons.
First, I do not see how we can justify adopting a bright-line, "one size fits all" rule if we truly are "to apply [the] ordinary prudential doctrine[ ]" of plain error review, as Justice Breyer's remedial opinion in Booker admonished us to do. 125 S.Ct. at 769; see also Paladino, 401 F.3d at 487 (Ripple, J., dissenting from denial of reh'g en banc) ("The panel decision today [adopting limited remand bright-line rule for Booker cases] offers a superficially pragmatic, but not a principled, basis for adopting its novel approach to plain error analysis."). In United States v. Young, the Supreme Court declared that "[a] per se approach to plain-error review is flawed." 470 U.S. 1, 16 n.14 (1985). The Court further stressed that "when addressing plain error, a reviewing court cannot properly evaluate a case except by viewing such a claim against the entire record," and that "it is particularly important for appellate courts . . . not to extract from episodes in isolation abstract questions of evidence and procedure" because "[t]o turn a criminal trial into a quest for error no more promotes the ends of justice than to acquiesce in low standards of criminal prosecution." Id. at 16.
Most recently, in Dominguez Benitez, the Supreme Court reemphasized the necessity of a case-by-case, fact-intensive approach to the prejudice inquiry of the plain error test, by stating that a defendant seeking reversal for unpreserved Rule 11 error "must thus satisfy the judgment of the reviewing court, informed by the entire record, that the" error affected his or her substantial rights. 124 S.Ct. at 2340 (emphasis added). The Court thus rejected the standard this circuit had used in deciding whether a defendant was prejudiced by Rule 11 error because that standard "d[id] not allow consideration of any record evidence tending to show that a misunderstanding [occasioned by the error] was inconsequential to a defendant's decision, or evidence indicating the relative significance of other facts that may have borne on his choice regardless of any Rule 11 error." Id. at 2341. The majority's blanket rule for pre-Booker cases involving unpreserved Booker errors cannot be reconciled with these Supreme Court cases requiring case-specific assessments of prejudice in the context of plain error review.
Second, the majority's decision to remand for the district courts to make substantial rights determinations conflicts with the guidance of Booker, which expressly instructs that it is the duty of "reviewing courts," not district courts, to determine whether a sentencing error was plain. Booker, 125 S.Ct. at 769 ("[W]e expect reviewing courts to apply ordinary prudential doctrines, determining, for example, whether the issue was raised below and whether it fails the 'plain-error' test."); see also Mares, 402 F.3d at 522 ("[W]e find no support for [the limited remand] approach in the Supreme Court plain error cases. Those cases place the obligation on the appellate courts-rather than the district courts-to determine the third prong of the plain error test."); Rodriguez, 398 F.3d at 1305 ("The determination of plain error is the duty of courts of appeal, not district courts."). Even before Booker, the Supreme Court consistently indicated that plain error review should occur at the appellate level, see, e.g., Dominguez Benitez, 124 S.Ct. at 2340 (explaining that under the plain error standard, "[a] defendant must thus satisfy the judgment of the reviewing court, informed by the entire record, that the probability of a different result is 'sufficient to undermine confidence in the outcome' of the proceeding") (emphasis added); United States v. Vonn, 535 U.S. 55, 59 (2002) (holding that under the plain error standard, "a reviewing court may consult the whole record when considering the effect of any error on substantial rights") (emphasis added), and there is nothing in Booker that "suggest[s] that this accepted practice must change in the wake of Booker," United States v. Milan, 398 F.3d 445, 454 (6th Cir. 2005). On the contrary, the Supreme Court in Booker emphasized that "ordinary" plain error rules should apply. 125 S.Ct. at 769.
Tracking Crosby's rationale, the majority points to 18 U.S.C. § 3742(f) as authority for its assertion that the limited remand option "is available for election by an appellate court assessing [plain] error [in sentencing]," reasoning that "the power to remand for resentencing [under 18 U.S.C. § 3742(f)] necessarily encompasses the lesser power to order a limited remand" to permit the sentencing judge to determine whether the third prong of the plain error inquiry is met. Maj. Op. at 6360 (citing Crosby, 397 F.3d at 117-18). But the Eleventh Circuit has already observed that this "conclusion does not follow at all," because "[i]n cases of non-preserved error appellate courts lack the authority to remand for resentencing where the requirements of the plain error rule have not been met." Rodriguez, 398 F.3d at 1306. Assuming that the greater power to remand for resentencing does include the lesser power to do a limited remand, the majority puts the cart before the horse by seeking to justify a limited remand for the district court to determine the substantial rights issue on the ground that this power is encompassed by its power to do a full remand, when the majority cannot exercise full remand power unless and until all four requirements of the plain error test are satisfied. Even the plain language of § 3742(f), the subsection on which the majority relies, conditions the power to "remand the case for further sentencing proceedings" on "the court of appeals determin[ing] that . . . the sentence was imposed in violation of law." 18 U.S.C. § 3742(f)(1) (emphasis added).
The majority's attempt to bring this case within the rule recited in United States v. Gunning, 401 F.3d 1145 (9th Cir. 2005), about our ability to limit the scope of issues on remand, is unpersuasive. Maj. Op. at 6360-61. In Gunning, we ordered a limited remand on the issue of the defendant-appellant's eligibility for a minor role adjustment under the Sentencing Guidelines because we could not discern from the record whether that issue had been considered by the district court. 401 F.3d at 1146, 1148. But remanding in a particular case on an issue properly raised in the first instance at the district court level so that we might develop the record we need to dispose of an appeal is one thing. Establishing a general rule that requires remands in countless cases where we already have a record so that the district courts can handle the disposition of sentencing appeals themselves is quite another. Gunning is inapposite because the majority is not asking the district courts to provide us with additional information to aid our review of an issue that initially should have been raised at the district court level. Rather it is asking the district courts to conduct the plain error inquiry in our stead, applying what has always been an appellate standard of review to sentences they themselves had imposed. Maj. Op. at 6360-61.
Third, the majority argues that we should ask the sentencing judges what they would have done under advisory Guidelines because we cannot otherwise know whether defendants' rights were substantially affected by Booker errors. Maj. Op. at 6361-62 (expressing preference for limited remand option because it "yields a result that is certain"); id. at 6362-63. But this argument rests on the faulty premise that courts reviewing for plain Booker error must be able to determine prejudice to defendants asserting such error with certainty. As noted above, all that the Supreme Court requires in the context of plain error review is a "reasonable probability" of prejudice, Dominguez Benitez, 124 S.Ct. at 2340.
To illuminate this point, we can look not only to what the Supreme Court has said about the prejudice inquiry under the plain error test, but also to what it has done in the analogous context of conducting harmless error review of district court errors in sentencing under the Guidelines. See Olano, 507 U.S. at 734 (explaining that the plain error standard "normally requires the same kind of inquiry" as the harmless error standard except that the burden of persuasion is on the defendant instead of the government). In Koon v. United States, 518 U.S. 81, 113 (1996), the Supreme Court invalidated some of the factors that the district court had relied on to depart from the applicable sentencing range under the Guidelines. The Supreme Court then evaluated the record before it to determine if the error was harmless, and remanded for resentencing because "it [wa]s not evident [from the record] that the [district] court would have imposed the same sentence if it had relied only on [permissible grounds for departure]." Id. at 113-14; see also Williams v. United States, 503 U.S. 193, 202-03 (1992) (same). Despite that there was no way for the Supreme Court to be certain whether the district court would or would not have sentenced the defendant differently, the Court notably did not remand for the district court's input on whether the defendant's substantial rights were affected by the error.
Finally, the majority's rule does not effectively advance the purposes of the plain error doctrine to promote the efficient administration of justice while ensuring that the gravest injustices do not go unchecked. As the Supreme Court explained in Young, the plain error standard provides a "careful balancing of our need to encourage all trial participants to seek a fair and accurate trial the first time around against our insistence that obvious injustice be promptly redressed." 470 U.S. at 15; see also United States v. Wilson, 237 F.3d 827, 836 (7th Cir. 2001) ("The plain-error standard, which we are required to apply when a district court has not been given the first opportunity to correct alleged mistakes, strikes a balance among the proper functioning of the adversary system, efficiency in managing litigation, and the demands of justice.") (internal citation omitted).
Although the majority's per se rule delegating the substantial rights inquiry to the district courts might make our task of reviewing pre-Booker plain error cases easier at the outset, it regrettably may prove to be a less efficient approach in the long run than performing traditional case-by-case plain error review at the appellate level. Under the limited remand approach, a district court will have to "obtain the views of counsel, at least in writing," so attorneys can reargue sentencing issues as they would have if Booker had been the law at the time of the original sentencing. Maj. Op. at 6370; Crosby, 397 F.3d at 120. The district court will next evaluate the record as supplemented by the attorneys' arguments, and then, treating the Guidelines as discretionary, will decide if it would impose a materially different sentence. Maj. Op. at 6364; Crosby, 397 F.3d at 120. As the Eleventh Circuit observed in Rodriguez, this rule "essentially requires resentencing [albeit with truncated procedures] in order to determine whether resentencing is required," and then "add[s] the real possibility of another appeal and another remand on top of that," because the parties retain the right to appeal the district court's decision on whether it would have imposed a different sentence under an advisory sentencing system. 398 F.3d at 1305; see also Mares, 402 F.3d at 522 (agreeing with the Eleventh Circuit that the limited remand approach "has the potential of producing many needless remands and appeals from those remands," and voicing the belief that "the Supreme Court sought to avoid these extra steps in the judicial process by requiring appellate courts to answer all prongs of the plain error test").
Not only is the majority's rule unlikely to increase our efficiency in processing Booker plain error cases, it provides inadequate protection for those defendants with valid plain error claims. If we were to apply plain error review at the appellate level as the Supreme Court's precedents direct us to do, we would narrow the pool of cases for remand to those where the record demonstrates that there was a reasonable probability the defendant would have received a different sentence if the district court could have exercised discretion in sentencing. Because we would vacate the original sentences in such cases and require the district court to resentence in accordance with Booker, those defendants would get the full benefit of the Supreme Court's ruling. Thus application of traditional plain error review in this context of forfeited Booker error would properly ensure that judicial resources are only expended for resentencing in cases involving miscarriages of justice that plainly warrant correction, while also ensuring that the defendants in those select cases will be sentenced in a manner that does not violate their constitutional rights as construed by the Supreme Court in Booker.
With its limited remand approach to plain error review, the majority decrees that every pre-Booker defendant asserting plain error will receive a "quick look" at the district court level. But the risk of this widespread remedy is that those defendants that actually have plainly erroneous sentences might not get the full attention that they deserve. Judge Ripple from the Seventh Circuit articulated this concern persuasively in his dissent from the Seventh Circuit's decision to adopt the same type of limited remand rule, stating that such an approach "[i]n all too many instances . . . will serve as an invitation for the district court to give only a superficial look at the earlier unconstitutionally-imposed sentence," and that "[t]he constitutional right at stake hardly is vindicated by a looks-all-right-to-me assessment by a busy district court." Paladino, 401 F.3d at 486 (Ripple, J., dissenting from denial of reh'g en banc).
The majority claims that it is reluctant to speculate as to what the district court would have done under the Booker rules, but the solution the majority adopts does not really address this concern because the majority in essence is asking the district court to speculate in its stead. The inescapable reality is that:
Until the district court undertakes a new sentencing process-cognizant of the freedom to impose any sentence it deems appropriate as long as the applicable guidelines range and the 18 U.S.C. § 3553(a) factors are considered-the district court cannot accurately assess whether and how its discretion ought to be exercised. The panel's holding requires the [district] court to pre-judge and to pre-evaluate evidence it has not heard.
Because "the universe of factors that guides sentencing is larger than it was pre-Booker," id. at 488 (Kanne, J., dissenting from denial of reh'g en banc), there are many issues that defendants before Booker might not have raised or emphasized at their original sentencing hearings that the district courts can now freely consider under the advisory Guidelines system. Requiring the district court to obtain the views of counsel in writing does not make the abbreviated process outlined by the majority an adequate substitute for the sentencing process mandated by Booker. Sentencing involves "subtle issues as to how much emphasis ought to be given to particular facts and circumstances" that can be resolved "competently only after hearing witnesses and seeing the evidence." Id. at 486 (Ripple, J., dissenting from denial of reh'g en banc). Additionally, a district judge reviewing a cold, incomplete record supplemented by written arguments does not have the benefit of evaluating the relevant facts in the context of the human dynamics that come into play during sentencing hearings. In short, it is hard to see how even the original sentencing judges would be able to give an accurate assessment of how they would have acted had the Guidelines been advisory when they were sentencing pre-Booker defendants. We gain little by delegating to the district courts our duty as appellate judges to conduct plain error review, because the limited remand approach is neither more efficient nor more accurate than the conventional approach in determining whether a defendant's substantial rights have been affected by sentencing error.
I respectfully dissent.
O'SCANNLAIN, Circuit Judge, with whom Judge Bea joins, concurring in part and dissenting in part:
With hundreds of Booker-implicated cases awaiting disposition, we have chosen one of the worst possible vehicles in our inventory-one that presents a clear sentencing error unrelated to the constitutional issue for which we took it en banc to decide. I am fascinated that while we unanimously agree that Ameline's sentence must be vacated and remanded to the district court for resentencing because of the Howard error, the district court judge in this case is free to resentence on a clean slate and to apply Booker as he sees fit.
However, since the Court now promulgates the Second Circuit's "limited remand" without vacatur approach as our standard for all other pending Booker-related appeals, I would, instead, follow the jurisprudence of the Fifth and Eleventh Circuits as described in, respectively, United States v. Mares, 402 F.3d 511 (5th Cir. 2005), and United States v. Rodriguez, 398 F.3d 1291 (11th Cir. 2005), and simply apply plain error review with our usual vigor-to use the majority's language -"the way plain error review normally works." Maj. op. at 6362.
Forced to make the best of our predicament, I concur in Parts I, II and V of Judge Rawlinson's majority opinion and its result; I join Part I of Judge Wardlaw's lead dissent; and I join Judge Gould's dissent, with the exception of his analysis of the third prong of plain error review.
BEA, Circuit Judge, concurring in part and dissenting in part:
I agree that Ameline's sentence must be vacated and remanded due to the district court's error under United States v. Howard, 894 F.2d 1085, 1090 (9th Cir. 1990). I therefore concur in Parts I, II and V of the majority opinion. I also concur in the concurring and dissenting opinions of Judges O'Scannlain, Wardlaw (Part I), and Gould to the extent they conclude it is the responsibility of this court, not the district court, to determine whether "the probability of a different result is sufficient to undermine confidence in the outcome of the proceeding." United States v. Dominguez Benitez, 124 S.Ct. 2333, 2340 (2004).
I disagree with all of my colleagues, however, as to how this court should conduct its review in cases in which there is forfeited Booker error, but the record as to sentencing is not sufficiently developed to discern whether the error prejudiced the defendant. I therefore dissent from Parts III, IV and VI of the majority opinion, and the dissenting opinions that would restrict appellate court review of the effect of such Booker error to the record before the sentencing court.
Two considerations drive my analysis of how forfeited Booker error must be addressed. First, we must not abdicate our responsibility as the reviewing court to determine whether plain error has occurred. Second, the defendant should be given a proper opportunity to present evidence on the third element of the plain-error doctrine-to establish "that the probability of a different result is sufficient to undermine confidence in the outcome of the proceeding." Dominguez Benitez, 124 S.Ct. at 2340 (citations and internal quotation marks omitted); see also United States v. Olano, 507 U.S. 725, 731-32 (1993).
I. This Court-Not the District Court-Must Determine Prejudice
In Booker, the Supreme Court specifically stated that it expects the appellate courts to apply the plain-error doctrine to cases raising a Booker issue on appeal where no objection was made below:
[W]e must apply today's holdings-both the Sixth Amendment holding and our remedial interpretation of the Sentencing Act-to all cases on direct review. . . . That fact does not mean that we believe that every sentence gives rise to a Sixth Amendment violation. Nor do we believe that every appeal will lead to a new sentencing hearing. That is because we expect reviewing courts to apply ordinary prudential doctrines, determining, for example, whether the issue was raised below and whether it fails the "plain-error" test.
United States v. Booker, 125 S.Ct. 738, 769 (2005) (citations omitted, emphasis added). In contemplating that the plain-error doctrine would apply, the Supreme Court could have modified the doctrine if it had thought the district courts, rather than the reviewing courts, should determine the third prong. But the Supreme Court did not modify the plain-error doctrine in this manner, and this court may not do so. Thus, although I agree that a remand is a practical solution, and certainly the most convenient solution for the appellate courts, it is not what Booker and Dominguez Benitez tell us to do. We do not have the authority to overrule Supreme Court precedent and abdicate our non-delegable duty.20 Agostini v. Felton, 521 U.S. 203, 237 (1997).
II. The Proper Procedure
This is not the typical case involving the plain-error doctrine. I agree with Judges Wardlaw and Gould the record here is sufficient to allow this reviewing court to conclude both the third and fourth prongs of the Olano test are satisfied. Olano, 507 U.S. at 731-32.
However, the substantive law of what a court can consider at sentencing has changed since the pre-Booker sentencing hearings. The majority concludes that on remand the "views of counsel, at least in writing" should be obtained. Maj. Op. at 6370. But it doesn't do much good to allow counsel to argue reasons for a lighter sentence if the record does not contain facts from which such argument can be made. It is a commonplace that argument must be based on facts. This is especially so when the mandatory Sentencing Guidelines in effect before Booker prohibited the sentencing court from considering some of the facts that the court can consider now that the Sentencing Guidelines are merely discretionary. Compare, e.g., 18 U.S.C. § 3553(a) with U.S.S.G. § 5H1.
Only a prescient defendant would have made a factual record just in case the Guidelines were ruled to be discretionary. Few, if any, saw this result coming. And even if counsel in a particular case had foreseen this result, it is unlikely he would have made an offer of proof of facts which, at the time, the trial judge could not consider. Counsel would have had to risk the trial judge's ire for wasting the court's time and his clients' anxiety for upsetting the judge. In short, no counsel with trial experience would have offered the very evidence the Sentencing Guidelines explicitly said the trial judge could not consider.
What then are we to do now that such evidence is relevant if the defendant is re-sentenced? The defendant must be given a proper opportunity to present evidence to this court to prove he was prejudiced by the error, which the post-Booker decisions by the First, Fifth, and Eleventh Circuits do not give a defendant. See United States v. Mares, 402 F.3d 511 (5th Cir. 2005); United States v. Antonakopoulos, 399 F.3d 68 (1st Cir. 2005); United States v. Rodriguez, 398 F.3d 1291 (11th Cir. 2005). The defendant may want to present evidence on the many factors that are now relevant to sentencing that the sentencing judge could not consider under the mandatory Guidelines.21
Normally, documents or exhibits that were not filed with the district court, admitted into evidence, nor considered by the district court cannot be filed as part of the record. See, e.g. Kirshner v. Uniden Corp. of Am., 842 F.2d 1074, 1077 (9th Cir. 1988) (appellate courts consider only the "record before the trial judge when his decision was made") (emphasis in original; internal quotes omitted); Tonry v. Security Experts, Inc., 20 F.3d 967, 974 (9th Cir. 1994) (arbitrator's decision was improperly included in appellant's excerpts of record because it was never filed with the district court). Motions to supplement the record generally are not granted by this court if the evidence was not considered by the trial court below. See Southwest Center for Biological Diversity v. United States Forest Serv., 307 F.3d 964, 975 n.15 (9th Cir. 2002). Similarly, judicial notice is also of limited use in this context because the proffered fact "must be one not subject to reasonable dispute in that it is either (1) generally known within the territorial jurisdiction of the trial court or (2) capable of accurate and ready determination by resort to sources whose accuracy cannot reasonably be questioned." FED. R. EVID. 201.
How then can the defendant meet his burden? One way to present new evidence on appeal is to attach that evidence to a motion to remand. FED. R. APP. P. 27(a)(2)(B); see also United States v. Christy, 3 F.3d 765, 768 (4th Cir. 1993). Defendants should be given an opportunity to make a motion for either a full or a limited remand.22 Defendants would then have the opportunity to adduce evidence upon which they could argue the now-relevant sentencing factors at a re-sentencing and how this new evidence would likely affect their sentence. The evidence presented can be as simple as the defendant's affidavit or an affidavit by the attorney of what evidence he intends to adduce at the possible sentencing and upon which he was not able to rely before. The prosecution would then be free to attempt to rebut that evidence in its response to the motion. FED. R. APP. P. 27(a)(3).
Finally, in evaluating the defendant's evidence, the test should be an objective test of whether a reasonable judge would have given the defendant a lower sentence such that it "undermines our confidence" in the outcome. We should not use a subjective standard which focuses on the particular judge who first imposed sentence.
Clearly the current state of the law has engendered some confusion. The circuit courts cannot agree what is to be done, and even within our own circuit, there are numerous and divergent opinions. The result is that the very uniformity in sentencing which the Court strove to preserve even while dramatically changing the sentencing procedures is still a distant goal. Perhaps the Court will develop a workable solution in one or more of these post-Booker cases, quite possibly involving a change to the plain-error doctrine. But unless and until that happens, I believe we are bound to apply existing law to the Booker issues.
1. A different analysis will apply when a defendant preserves his Sixth Amendment claim by challenging the sentencing guidelines on constitutional grounds before the district court. See Pirani, 2005 WL 1039976, at *3-4; United States v. Fagans, No. 04-4845, 2005 WL 957187, at *2, (2d Cir. Apr. 27, 2005); Antonakopoulos, 399 F. 3d at 76.
2. Crosby, 397 F.3d at 120 (2d Cir.); Paladino, 401 F.3d at 483-84 (7th Cir.); Coles, 403 F.3d at 769-71 (D.C. Cir.). The Fifth Circuit has also expressed openness to this approach. See United States v. Pennell, No. 03-50926, 2005 WL 1030123, at *6 (5th Cir. May 4, 2005).
3. The district court judge who imposed the sentence in this case is still a sitting federal judge. We recognize that in some cases the original sentencing judge may no longer be available, due to death, disability or retirement, but the number of such cases should be very low. The sentencing judge's unavailability will not necessarily result in an inability to proceed. The record may reflect an admission that was undiscovered in the appellate process. There may be a negotiated sentence. The defendant may even elect not to challenge his sentence on remand. Because the situation of an unavailable judge is not before us, we leave that issue for resolution in a case that presents that issue.
4. We do not imply that all forfeited sentencing errors should be addressed by limited remand. Booker presents a unique situation. After Booker, we are left with hundreds of unconstitutional sentences pending on direct review, and we should not consign the vast majority of these defendants to serve illegal terms when we have an accurate way to resolve the plain error question. We need not speculate about the effect of the error; we can simply ask the sentencing judge.
5. In challenging the Second Circuit's reliance on 18 U.S.C. § 3742 as authority for ordering a limited remand, our colleague cites to dissenting opinions in Booker. Those dissents, of course, are not precedential. See Purcell v. BankAtlantic Fin. Corp., 85 F.3d 1508, 1513 (11th Cir. 1996) (noting that a dissenting Supreme Court opinion is not binding precedent and "does not tell us how a majority of the Court would decide."). The dissent also relies on the ipse dixit criticism in Rodriguez. However, we are persuaded that the Second Circuit has the better of this argument. See Crosby, 397 F.3d at 117 (noting that § 3742(f) provides that an appellate court may remand "with such instructions as the court considers appropriate."). Finally, the language in § 3742 conditioning the power to remand upon a determination that "the sentence was imposed in violation of law" does not undermine the Second Circuit's approach, because any sentence imposed under the mandatory guidelines system with extra-verdict findings was imposed in violation of law. See Booker, 125 S.Ct. at 748 (describing such a sentence as violating the Sixth Amendment.).
6. We acknowledge the existence of cases instructing reviewing courts to review the "entire record." See e.g., Dominguez Benitez, 124 S.Ct. at 2340. We have no quarrel with that directive in principle. However, that principle is of no assistance when review of the "entire record" nevertheless requires resort to rank speculation to complete the plain error analysis.
7. We note in passing that refraining from adopting the limited remand approach merely because the parties advocated against it is more of an abdication than remanding a matter to the district court to inform the plain error analysis.
8. In a case where the district court did not treat the sentencing guidelines as advisory but the defendant's sentence was not enhanced by extra-verdict findings -such as where there were no sentencing enhancements or where the defendant acknowledged the facts necessary to justify the enhancement-a different, non-constitutional error occurs. See United States v. Castillo, Nos. 02-3584 & 02-4344, 2005 WL 1023029 (7th Cir. May 3, 2005) (citing cases); Lawrence, 2005 WL 906582, at *12. Neither Ameline nor the government has raised the issue of non-constitutional error in this appeal.
9. A new appeal taken after the filing of the district court's order will be subject to the usual procedure pertaining to comeback cases, as provided in General Order 3.7, General Orders of the Ninth Circuit Court of Appeals (2005).
10. The parties and the district court may agree in a given case to proceed directly to a resentencing proceeding, without the need for submissions by counsel or separate consideration of the question of whether the previous sentence would have been different had the sentencing court known that the Guidelines were not mandatory.
11. Since Booker, the Supreme Court has vacated every one of our judgments that touches upon a Booker issue and remanded them for "further consideration in light of [Booker]." See, e.g., Alizondo v. United States, 125 S.Ct. 1000 (2005). It has not remanded them with instructions to remand to the district courts for participation in the appellate review function.
12. Under the majority's analysis, panels of our court remain free to review the record, apply the proper prejudice inquiry, find that prejudice exists, and remand for resentencing. The majority emphasizes that "the limited remand is invoked only when it cannot be determined from the record whether the judge would have imposed a materially different sentence had he known that the Guidelines are advisory rather than mandatory." Ante at 6366. The majority continues, stating that "[o]nly after determining that the record did not sufficiently inform the reviewing court's analysis," is the "limited remand" triggered. Ante at 6367. This appears to leave open the process utilized by the Eleventh Circuit in United States v. Rodriguez, 398 F.3d 1291, 1301 (11th Cir. 2005), where if the district judge openly indicated that he would not have imposed the sentence he did but for the mandatory Guidelines, then prejudice is found.
13. The continued viability of § 3742(f) itself is subject to question. As Justice Scalia wrote in Booker: If the Guidelines are no longer binding, one would think that the provision designed to ensure compliance with them would, in its totality, be inoperative. The Court holds otherwise. Like a black-robed Alexander cutting the Gordian knot, it simply severs the purpose of the review provisions from their text, holding that only subsection (e), which sets forth the determinations that the court of appeals must make, is inoperative, whereas all the rest of § 3742 subsists-including, mirabile dictu, subsection (f), entitled "Decision and disposition," which tracks the determinations required by the severed subsection (e) and specifies what disposition each of those determinations is to produce. This is rather like deleting the ingredients portion of a recipe and telling the cook to proceed with the preparation portion. Id.
14. To further constrain the exercise of discretion by district courts, appellate review of district courts' sentencing decisions was broadened by the Prosecutorial Remedies and Tools Against the Exploitation of Children Today Act of 2003 ("PROTECT Act"), Pub. L. No. 108-21, 117 Stat. 650 (2003). See Booker, 125 S.Ct. at 786-87 (Stevens, J., dissenting) ("Congress has rejected each and every attempt to . . . vest judges with more sentencing options. . . . Most recently, Congress' passage of the [PROTECT Act] reinforced the mandatory nature of the Guidelines by expanding de novo review of sentences to include all departures from the Guidelines . . . .").
15. Decisions in the aftermath of the Second Circuit's Crosby decision prove that district courts will elect not to hold hearings before deciding not to resentence. In one case, for example, a district judge for the Southern District of New York denied a defendant's request for a hearing, stating: Crosby provides specific guidance to the District Courts of this Circuit concerning cases remanded for possible re-sentencing. In deciding whether a defendant should be re-sentenced, the District Court should obtain the views of counsel, at least in writing, but need not require the presence of the Defendant. Moreover, the question of whether to re-sentence may be resolved with or without a hearing. Accordingly, defense counsel's current request for a conference is DENIED. The parties are directed to file any written submissions concerning the possible re-sentencing of defendant by May 15, 2005. United States v. Jasper, 2005 WL 774519 (S.D.N.Y. April 6, 2005) (citations and quotations omitted).
16. The Second Circuit has responded to this criticism by stating that "[t]he remand is for a determination of whether the original sentence would have been materially different, and only in that event does the remand lead to resentencing." United States v. Williams, 399 F.3d 450, 461 (2d Cir. 2005) (emphasis in original). This response misses the point, for the inquiry into whether the original sentence would have been "materially different" necessarily involves consideration of what sentence the district court would apply if it were to resentence under the advisory Guidelines.
17. The government during the en banc oral argument explicitly conceded that Howard required resentencing for Ameline: Q: How do you deal with . . . the fact that the district court put the burden on Ameline to disprove what was in the Presentence Report? I mean, isn't that a fundamental mistake that means whatever else is the grand architecture of post-Booker sentencing and what we do with it, that Mr. Ameline deserves another hearing? A: Yes. The district court in this case made an error [in shifting the burden of proof at sentencing to Ameline] that is totally separate and apart from the larger plain error questions on which this court granted rehearing en banc. . . . I agree that the case should go back for that reason because a starting point of discretionary sentencing with advisory Guidelines is a correctly calculated Guidelines range using the appropriate principles of judicial fact-finding. Q: Your argument is that we can't simply affirm on the basis of this record in this context? A: I would be reluctant to suggest that the Court should affirm because it's the government's position that in order for a court to have a properly-implemented advisory Guidelines sentence it needs to do the right procedural things in calculating the Guidelines sentence. And although it's conceivable that the court could find that the judge's error in shifting the burden to the defendant was ultimately in itself harmless . . . . I would acknowledge that at the outset of the hearing the judge said things about giving the PSR presumptive credence that probably went over the line.
18. Ameline only had one prior conviction in 1997 for "Issuing a Bad Check."
19. As Justice Potter Stewart sagely observed during a press conference at the Supreme Court at the time of his retirement: [T]he mark of a good Justice or any judge is one whose Opinions you can read and, after you have read them, you have no idea whether the judge was a man or a woman, a Republican or a Democrat, a Christian or a Jew, or-if a Christian, a Protestant or Catholic. You just know that he or she was a good judge. It is the first duty of a judge to remove from his judicial work his own social and philosophical and political or religious beliefs, and not to think of himself as being here as some great philosopher-king to just apply his own ideology. Barrett McGurn, America's Court: The Supreme Court and the People 105 (1997).
20. Additionally, for the reasons ably pointed out by Judges Wardlaw and Gould, a "quick look" remand may not be the easiest solution after all.
21. Such factors include: (1) specific offender characteristics under 18 U.S.C. § 3553(a) that were barred by U.S.S.G. § 5H1, such as: age; educational and vocational skills; employment record; family ties and responsibilities; military, civic, charitable or public service; employment-related contributions; prior good works; lack of guidance as a youth; and a disadvantaged upbringing; (2) the fact that "imprisonment is not an appropriate means of promoting correction and rehabilitation" under 18 U.S.C. § 3882(a); (3) comparative statistical evidence of sentencing for similar state crimes; and (4) any other information about the defendant that is relevant. See 18 U.S.C. § 3661 ("No limitation shall be placed on the information concerning the background, character, and conduct of a person convicted of an offense which a court of the United States may receive and consider for the purpose of imposing an appropriate sentence.").
22. Where the only issue in the case is a Booker issue, then the motion should be for a full remand because there is no need for this court to retain jurisdiction over the appeal. Where there are other issues in the case, the motion should be for a limited remand.