United States v. Padilla

LIPEZ, Circuit Judge,

with whom TORRUELLA, Circuit Judge, joins, dissenting.

I agree with the majority that the plain error in Padilla’s sentence is hot structural in nature. However, because I believe that it is meaningless to apply the- third step of the plain-error test in these circumstances, and -because I would resolve the fourth step in Padilla’s favor, I respectfully dissent from the majority’s decision, to leave the error uncorrected.

I. The third step (prejudice)

As the majority acknowledges, we are faced with comparing “two unknown variables: the limit the district court would have set on drug tests and the number of tests that the probation officer will demand.” Ante at 221. In the majority’s *226view, this circumstance makes it “nearly impossible” for Padilla to show prejudice. Id. Actually, the problem is deeper than that.

The plain-error test’s third step requires us to assess the probability that the error at issue affected the outcome of the proceeding in which it occurred. To meet the plain-error test, the prejudicial effect on the outcome of the proceeding must be “substantial and injurious.” United States v. Dominguez Benitez, 542 U.S. 74, -, 124 S.Ct. 2333, 2339, 159 L.Ed.2d 157 (2004) (internal quotation marks omitted). Typically, when a defendant comes to us alleging that the district court has committed plain error, we are able to ask meaningful questions: Would the judge or jury still have found the defendant guilty, or would the judge have imposed the same sentence, .or would the defendant still have entered a guilty plea, if the error had not occurred? We answer those questions “by viewing such a • claim against the entire record.... It is simply not possible for an appellate court to assess the seriousness of the claimed error by any other means.” United States v. Young, 470 U.S. 1, 16, 105 S.Ct. 1038, 84 L.Ed.2d 1 (1985). If we answer those questions “yes,” we conclude that there was no prejudice from the error.

Here, however, we cannot ask that kind of question. This case involves an erroneous delegation of authority to a probation officer. That delegation error was not a ruling preliminary to a guilty verdict,, or the imposition of a period of incarceration, or the acceptance of a guilty plea. We cannot assess the likelihood that a known outcome would have occurred even without the error. The error relates only to who imposes a condition of supervised release. For the purpose of evaluating the significance of the delegation error itself, the probability question that is so critical to the prejudice analysis of plain-error review is meaningless.

If we focus instead on the consequence of the improper delegation of authority (how many drug tests the defendant will be required to submit to during supervised release), the probability question is still meaningless. The improperly delegated authority has not yet been exercised. Thus we have two unknowns: the number of drug tests the judge never decided to require, and the number of drug tests the probation officer has yet to require. As an appellate court in these circumstances, we are deprived of any meaningful ability to review the plain error for prejudice. In other words, Padilla’s failure to show prejudice results not from any infirmities specific to his case, but from an inherent limitation on our ability to analyze the issue of prejudice. We are not applying a “test” when the subject of the test has zero chance of passing — and no defendant in Padilla’s situation will ever be able to show prejudice. In effect, the majority applies a per se rule which will always leave this kind of error uncorrected.

The Supreme Court’s logic in Sullivan v. Louisiana, 508 U.S. 275, 113 S.Ct. 2078, 124 L.Ed.2d 182 (1993), is instructive here. In that case, the district court had given the jury an erroneous reasonable-doubt instruction. Although many constitutional errors are amenable to harmless-error review, see Chapman v. California, 386 U.S. 18, 24, 87 S.Ct. 824, 17 L.Ed.2d 705 (1967), the Court held that the error in Sullivan was not because, “the entire premise of Chapman review is simply absent,” Sullivan, 508 U.S. at 280, 113 S.Ct. 2078. The Court explained:

There being no jury verdict of guilty-beyond-a-reasonable-doubt, the question whether the same verdict of guilty-beyond-a-reasonable-doubt would have been rendered absent the constitutional *227error is utterly meaningless. There is no object, so to speak, upon which harmless-error scrutiny can operate.

Id.

As I have said, I reject Padilla’s characterization of his sentencing error as structural. However, we are faced with a Sullivan-like problem here. Because we are comparing an unknown with an unknown, we have “no object, so to speak, upon which [the prejudice] scrutiny can operate.” Id. There are some questions to which no meaningful answer can be given — in other words, where the question’s premise simply does not apply. That is the situation here. We are not applying a “test” to facts, as if the outcome might depend on those facts. Thus, the prejudice analysis of plain-error review in a case such as this is simply illusory.

In concluding that this indeterminacy means that Padilla loses on the third step, the majority cites a statement from Jones v. United States: “Where the effect of an alleged error is so uncertain, a defendant cannot meet his burden of showing that the error actually affected his substantial rights.” 527 U.S. 373, 394-95, 119 S.Ct. 2090, 144 L.Ed.2d 370 (1999). Jones had been convicted of a capital offense and sentenced to death. On appeal, he claimed that an erroneous instruction had led the jury to believe that if it failed to recommend a sentence unanimously (whether life imprisonment or a death sentence), the court would then impose a sentence of less than life. 527 U.S. at 387, 119 S.Ct. 2090. The court in Jones knew the outcome of the process affected by the alleged error: that Jones had in fact been sentenced to death. That type of error lends itself to the traditional prejudice analysis of plain-error review, in which a court must assess the probability of the same outcome even if the error had not occurred. The Court engaged in such a probability analysis when it said that

even assuming that the jurors were confused over the consequences of deadlock, petitioner cannot show the confusion necessarily worked to his detriment. It 'is just as likely that the jurors, loath to recommend, a lesser sentence, would have compromised on a sentence of life imprisonment as on a death sentence.

Id. at 394-95, 119 S.Ct. 2090. For reasons that I have already explained, that traditional prejudice analysis does not apply to the delegation error at issue in this case. Here, we do not have the equivalent information: namely, how many drug tests probation will require. We are, in short, dealing with a different kind of plain-error problem. Thus the prejudice analysis in Jones has scant relevance to the prejudice issue here.

The government insists, however, that we are not dealing with a different kind of plain-error problem. It cites two cases which, in its view, demonstrate that even if a decision was made by the wrong person or institution, the Supreme Court nonetheless applies plain-error review as usual. Those cases are United States v. Cotton, 535 U.S. 625, 122 S.Ct. 1781, 152 L.Ed.2d 860 (2002) (drug quantity omitted from indictment and so- decided by judge instead of jury; Court reviewed for plain error); and Johnson v. United States, 520 U.S. 461, 117 S.Ct. 1544, 137 L.Ed.2d 718 (1997) (in prosecution for perjury, the element of materiality was decided by the judge instead of jury; Court reviewed for plain error). In each case, the judge made a decision that the jury should have made. The Court approached the claim of error by reviewing the entire record and concluding that, even if the question had been asked correctly of the jury rather than the judge, the outcome would be the same because the evidence was “overwhelming *228and uncontroverted.” Cotton, 535 U.S. at 634, 122 S.Ct. 1781; see also Johnson, 520 U.S. at 470, 117 S.Ct. 1544 .(evidence supporting materiality was “overwhelming”). Those cases did not, however, involve an improper delegation of authority which had never been exercised. They were thus amenable to a traditional hind of plain-error review; they only confirm that we are dealing here with a different kind of error, which necessitates a different approach to plain-error review.

II. The fourth step (public reputation of judicial proceedings)

The only meaningful plain-error question when reviewing a claim like Padilla’s involves the fourth step: whether the error so “seriously affects the fairness,- integrity or public reputation of judicial proceedings,” United States v. Olano, 507 U.S. 725, 736, 113 S.Ct. 1770, 123 L.Ed.2d 508 (1993) (internal quotation marks and citation omitted), that it cannot .stand. The majority dismisses this error as one “not of such magnitude or consequence that it would undermine faith in the judicial system were it to stand uncorrected.” Ante at 221. I cannot agree.

I first note one aspect of the history of this case. In United States v. Meléndez-Santana, 353 F.3d 93 (1st Cir.2003), where a panel of this court first decided that the delegation error at issue here was plain error, the government did not seek a rehearing by the panel or en banc. Since that case, as the majority describes, the government has repeatedly conceded the error in many eases and agreed to a remand for correction without asking for en banc review of the plain-error determination. See, e.g., United States v. Villafane-Jimenez, 410 F.3d 74, 88 (1st Cir.2005); United States v. Ayala-Pizarro, 407 F.3d 25, 29 (1st Cir.2005); United States v. Vega, 398 F.3d 149, 154 (1st Cir.2005). Even when the original panel in this case explicitly invited the government to seek a rehearing en banc, the government declined to do so. See United States v. Padilla, 393 F.3d 256, 261 (1st Cir.2004) (Campbell & Selya, JJ., concurring), vacated, 403 F.3d 780 (1st Cir.2005). At that point, the court sua sponte had to call for a rehearing en banc and order briefing.

As a procedural matter, that call was entirely appropriate. If we think we have made a legal error, we have a responsibility to correct it even if the parties are indifferent. Still, the costs of complying with the Meléndez-Santana rule arguably fell on government prosecutors, who had to go back to court and get the sentencing judge to establish the maximum number of drug tests. When the government displays little interest in avoiding this burden, it should warn us that our interest in avoiding it for them may be incompatible with the fourth step of plain-error review.

As we have said before, “[t]he ‘fairness, integrity or reputation’ plain-error standard is a flexible one and depends significantly on the nature of the error, its context, and the facts of the case.” United States v. Gandia-Maysonet, 227 F.3d 1, 6 (1st Cir.2000); see also United States v. Hoyle, 237 F.3d 1, 5 (1st Cir.2001) (same). This kind of delegation error — which leaves the defendant in an impossible position for showing prejudice, yet which is so easy for the prosecution and trial court to fix — meets that standard.

A. Sentencing errors are different

The Second Circuit has held that courts have some flexibility in applying the otherwise strict standards of plain-error review when faced with sentencing errors. See United States v. Sofsky, 287 F.3d 122, 125 (2d Cir.2002). We, in turn, have cited Sofsky for that proposition. See United States v. Cortes-Claudio, 312 F.3d 17, 24 *229(1st Cir.2002). One of the principle reasons for Sofsky’s conclusion was the relative ease of the burden of correction:

[Noticing unobjected-to errors that occur at trial precipitates an entire new trial that could have been avoided by a timely objection, whereas correcting a sentencing error results in, at most, only a remand for resentencing, or, as in this case, for a modification of the allegedly erroneous condition of supervised - release.

287 F.3d at 125. Here, of course, Padilla makes the same modest request for a modification in the conditions of his supervised release, and the same principle applies.

We have demonstrated our fidelity to this principle recently in our treatment of unpreserved Booker error. For instance, in United States v. Antonakopoulos, we began with a classic statement of the third prong. To show prejudice, we said that a defendant must “point to circumstances creating a reasonable probability that the district court would impose a different sentence more favorable to [him] under the new ‘advisory Guidelines’ Booker regime.” 399 F.3d 68, 75 (1st Cir.2005). Soon thereafter, we also said in United States v. Heldeman that we were “[g]uided by traditional plain error doctrine.” 402 F.3d 220, 224 (1st Cir.2005). Yet we acknowledged that, due to the special circumstances of unpreserved Booker error, “we are inclined not to be overly demanding as to proof of probability where, either in the existing record or by plausible proffer, there is reasonable indication that the district judge might well have reached a different result under advisory guidelines.” Id. We added that,

[a]fter all, it will be easy enough for the district judge on remand to say no with a minimum expenditure of effort if the sentence imposed, under the pre-Booker guidelines regime is also the one that the judge would have imposed under the more relaxed post-Booker framework:

Id.; see also United States v. Wilkerson, 411 F.3d 1, 8 (1st Cir.2005) (one reason for remand is that “it would be easy enough for [the” district judge] to say no with a minimum expenditure of effort”). Although Heldeman did not explicitly consider the fourth step of plain-error review in its analysis, I understand our consideration of the burden of correcting the error as particularly relevant to the fourth step. To the extent that Heldeman did not consider the fourth step at all, it shows that we have been sensitive to context in the application of plain-error doctrine-indeed, quite recently, -and in an instance where institutional concerns about the burden of error correction were important to the plain-error analysis.

In short, in the sentencing context, we have been able to adapt traditional plain-error review to unusual circumstances that the doctrine, and the Supreme Court’s explication of it, have not anticipated. We continue to cite both Antonakopoulos and Heldeman as our guides in this area of law. See, e.g., United States v. Lewis, 406 F.3d 11, 21 (1st Cir.2005).

B. Institutional concerns predominate

We should show that same flexibility here. The fourth step requires us to focus on the larger, institutional consequences of the error (as expressed by the phrase the “public reputation of judicial proceedings”), not only on the consequences of the • error for the defendant before us. Here, those institutional concerns should include the importance of courts following explicit rules that Congress has prescribed. Of course courts will sometimes go astray,6 as in this case. However, when the burden of *230getting it right is so' minimal, the obligation to make the correction is that much greater. We are not confronted here with the prospect of redoing trials, or plea proceedings, or even sentencing hearings. Instead, at most, the sentencing judge, with input from the prosecution and defense counsel, has to correct the improper delegation of authority to a probation officer and set the maximum number of drug tests.

- Congress has said, for reasons that may not be fully apparent to us, that judges rather than probation officers should set the maximum number of drug tests. We acknowledged that uncertainty about Congress’s rationale in Meléndez-Santana itself: “Legislative history does not reveal why Congress chose to go in a different direction” from a policy permitting judges .to delegate such decisions. United States v. Meléndez-Santana, 353 F.3d 93, 106 (1st Cir.2003). The majority views Congress’s commitment of these kinds of decisions to the courts ,a.s merely statutory and, in the final analysis, not particularly important. I acknowledge that, before 1994, probation officers had the discretion to set the maximum number of drug tests. Still, Congress has made a decision to alter that practice, and we should respect that choice by correcting the error that undermines it.

With a minimal expenditure of judicial resources, we can show that respect and thereby avoid a misapplication of the plain-error doctrine which, rather than preserving the integrity of plain-error review, repudiates our recent willingness to apply that doctrine flexibly in the sentencing context. Such flexibility does not threaten to open the floodgates of easy error correction, and burdensome retrials and re-sentencing, as the majority may fear. We will retain our ability to distinguish between types of error and the contexts in which they occur. The integrity of plain-error review does not suffer from its sensible application.

III. Conclusion

We should not appear to apply the plain-error doctrine when, in truth, we know that defendants are destined for failure. That is not a test, properly speaking — -that is a per se rule. The majority essentially opts for that per se approach and would leave the delegation error at issue here, and all similar errors, uncorrected. Respectfully, that is a mistake. We should exercise our discretion to correct this error, thereby carrying out a direct Congressional command with a minimal burden on prosecutors and the courts. The fourth step of plain-error review is designed to safeguard the reputation of the courts. Leaving this plain error uncorrected dis-serves that purpose.

. As I have said, a quick search through our recent cases shows that we are dealing with a *230common delegation error. See, e.g., Villafane-Jimenez, 410 F.3d at 88; Ayala-Pizarro, 407 F.3d at 29; Vega, 398 F.3d at 154.