concurring:

The majority holds that we may not use the supervisory power of the federal judiciary to remedy extreme appellate delay without a showing of trial prejudice. I write separately because, without analysis, the majority unnecessarily diminishes the supervisory power of the federal judiciary by accepting United States v. Hasting, 461 U.S. 499 , 103 S.Ct. 1974 , 76 L.Ed.2d 96 (1983), as controlling authority even though this is a case of appellate delay rather than trial error. In so doing, the majority effectively forecloses the use of supervisory power by the federal judiciary to remedy appellate delay without first considering whether this appeal can and should be decided on narrow prudential grounds.

The original panel relied upon United States v. Antoine, 906 F.2d 1379 (9th Cir.), cert. denied, 498 U.S. 963 , 111 S.Ct. 398 , 112 L.Ed.2d 407 (1990), in holding that we had the authority to use the supervisory power of the federal judiciary to dismiss Tucker’s indictment because he was subjected to extreme appellate delay through no fault of his own. The panel applied Antoine's four-part test for determining whether appellate delay violated due process and determined Tucker easily passed the first three parts of the test. See U.S. v. Tucker, 964 F.2d 952, 955 (9th Cir.1992), rehearing granted, 991 F.2d 1518 (9th Cir.1993) (reasoning that the length of the delay was longer than in Antoine , a dilatory court reporter caused almost all of the delay, and Tucker diligently pursued his appeal). To this extent the original panel’s opinion and the majority’s opinion agree. The opinions diverge, however, in applying the fourth part of the Antoine test: whether the defendant was prejudiced by the delay. See Antoine, 906 F.2d at 1382 .

Antoine’s prejudice element requires consideration of three factors: “1) oppressive incarceration pending appeal; 2) anxiety and concern of the convicted party awaiting the outcome of the appeal; and 3) impairment of the convicted person’s grounds for appeal or of the viability of his defense in case of retrial.” Id. Although the original panel determined that the first two factors did not apply to Tucker, it decided that the appellate delay experienced by Tucker satisfied the third factor even though he was not entitled to a retrial. Tucker, 964 F.2d at 955 . The en banc majority disagrees on this point, reasoning that Tucker does not satisfy the third Antoine factor — prejudice on retrial— because his appeal is meritless. In other words, Tucker can experience no prejudice on retrial because his original conviction must be affirmed on the merits. The majority’s logic is compelling, and I concede that the original panel’s opinion is flawed in this respect. Nonetheless, I cannot join the majority’s opinion because the majority’s analysis is flawed in a different respect. I disagree with the majority that United States *678 v. Hasting is authority for the proposition that a showing of trial-related prejudice is a necessary condition to our exercise of supervisory power to redress the problem of appellate delay.

There is an analytical void in the majority’s opinion. It makes no attempt to explain why Hasting , a case that involved prosecuto-rial misconduct at the trial level, should be read as controlling authority in this case of appellate delay. The majority simply reads — in my view, overreads — Hasting as authority for the sweeping proposition that “a federal court may not exercise its supervisory authority to reverse a conviction or dismiss an indictment absent prejudice to the defendant.” Majority Opinion p. 674. In Hasting the Supreme Court addressed a much narrower question: “whether, on this record, in a purported exercise of supervisory powers, a reviewing court may ignore the harmless-error analysis of Chapman.” 1 461 U.S. at 505 , 103 S.Ct. at 1978 (emphasis added). The Court held that the Seventh Circuit erred in reversing a conviction for prosecutorial misconduct at trial on the basis of its supervisory power without engaging in harmless error analysis. Id. at 505 , 103 S.Ct. at 1978 . In reversing the Seventh Circuit, the Supreme Court explained the harmless error rules: “ ‘[a]ll of these rules, state or federal, serve a very useful purpose insofar as they block setting aside convictions for small errors or defects that have little, if any, likelihood of having changed the result of the trial.’ ” Id. at 508, 103 S.Ct. at 1980 (quoting Chapman, 386 U.S. at 22, 87 S.Ct. at 827) (emphasis added). Notwithstanding the Court’s focus on the effect of error on the outcome of the trial, the majority offers no explanation whatsoever why Hasting cannot or should not be distinguished from this ease of extreme delay in the post-trial appeal. Nor does it explain the logic of measuring prejudice from appellate delay solely in terms of prejudice to the outcome of a trial. The only trials that appellate delay could conceivably prejudice are posi-appeal trials for the obvious reason that appellate delay can have no bearing on the outcome of a completed pre-appeal trial. Thus, in measuring prejudice from appellate delay solely in terms of prejudice to the outcome of a trial, the majority renders us helpless to provide relief from excessive appellate delay to the class of appellants most worthy of relief: those with meritorious appeals that entitle them not to a new trial, but to a reversal of their convictions, a dismissal of their indictments, and their unconditional release from prison.

We need look no further than Antoine for authority that we may measure prejudice from excessive appellate delay other than by its effect on the outcome of a trial. To be sure, Antoine recognizes the obvious potential of appellate delay for impairing the viability of the defense’s case on retrial. 906 F.2d at 1382 . Memories of key defense witnesses, for example, may fade, and some may even become unavailable to testify by the time of retrial for any number of reasons, including death. But Antoine goes on to recognize that appellate delay also has the potential for prejudice in a way that is unrelated to the outcome of trials: time spent in prison awaiting the outcome of the appeal can itself be a form of prejudice regardless of any impact the delay may have on retrial. Id. In its uncritical acceptance of Hasting as controlling authority, the majority fails to give any recognition to the hard fact that liberty interests are directly implicated by appellate delay. 2 Thus, by extending Hasting’s requirement of trial-related prejudice indiscriminately to cases of appellate delay, the majority mindlessly renders the federal judiciary powerless to use its supervisory power to combat the evils of appellate delay except in a single category of appeals: those *679 which result in a reversal and remand for retrial.

But what does any of that have to do with Tucker? Nothing really. His appeal is a loser anyway, so the six years spent in prison waiting for us to decide his appeal is time he would have served no matter what. Fair enough. But suppose Tucker had a winning-argument that there was insufficient evidence to sustain his conviction. Under the majority’s rule of decision — that in the absence of trial-related prejudice, Hasting forecloses the use of supervisory power to remedy extreme appellate delay — Tucker would still be out of luck. In other words, the majority’s indiscriminate application of Hasting to eases of appellate delay precludes the federal judiciary from using its supervisory power to provide some measure of relief to persons who languish in prison for years waiting for us to tell them that they are free to go home with no worries about a retrial because they should not have been sent to prison in the first place. Sorry, we would say, the Supreme Court told us in Hasting there is nothing we can do about the problem of extreme appellate delay unless the defendant wins a retrial rather than his unconditional freedom. I believe Hasting does not so limit our authority; appellate delay and trial error are conceptually incongruent.

In expanding Hasting’s rule of decision to cases of appellate delay, the majority forecloses the use of supervisory power to minimize the risk that appellate delay will take an unnecessary toll on the liberty of the innocent. For example, the majority eliminates the possibility of adopting a prophylactic rule that may benefit Tucker today, but would also benefit the innocent in the future. The reversal of Tucker’s conviction on the ground it has taken six years to decide his appeal would build into the appellate process a powerful incentive to provide all incarcerated defendants with speedy appeals. To be sure, the adoption of such a rule would sometimes benefit the guilty as well as the innocent, but that is the nature of a prophylactic rule. The exclusionary rule, for example, often benefits the guilty as well as the innocent, but we have long since decided we are willing to pay that social cost in order to promote compliance with the Fourth Amendment. Similarly, a rule that convictions would be reversed when appellate delay is extreme would promote speedy appeals. Indeed, such a prophylactic rule may be the only effective way to guard against the injustice of subjecting innocent persons to prolonged prison time while they wait for their convictions to be reversed on appeal.

I do not propose, however, that we adopt such a rule at this time and use it as a basis for reversing Tucker’s conviction. The fact that his meritless appeal took six years to process and decide does not, standing alone, justify the invocation of our supervisory power to adopt a rule that would promote speedy appeals. We simply have no reason to believe that excessive appellate delay is a significant enough problem in our circuit to justify adoption of such a rule at this time.

I believe we can and should dispose of Tucker’s appeal without addressing the broad question whether Hasting is controlling authority in cases of excessive appellate delay. For prudential reasons, -we should simply affirm Tucker’s conviction on the merits without discussing the nature and extent oí our authority to promote speedy appeals. In holding that Hasting’s rule of trial-related prejudice governs cases of appellate delay, the majority unnecessarily eviscerates that authority. While there may be room for disagreement about the existence and scope of the federal judiciary’s supervisory power, 3 it is settled law that such power has been entrusted to us. 4 While that authority should always be used sparingly, we should not lightly cabin our discretion to use it if the need arises.

. In Chapman v. California, 386 U.S. 18 , 87 S.Ct. 824 , 17 L.Ed.2d 705 (1967), the Supreme Court held that the harmless error rule applied under some circumstances to some constitutional violations. Id. at 22 , 87 S.Ct. at 827 .

. More perplexing is the majority’s failure to recognize the basic inconsistency in its reading of Hasting and its reliance on Antoine . If, as the majority holds, Hasting's trial-related definition of prejudice controls this case, the first two types of prejudice recognized by Antoine — prejudice in the form of oppressive incarceration and prejudice in the form of anxiety and concern while awaiting appeal — are placed beyond our reach.

. See, e.g., United States v. Rubio-Villareal, 967 F.2d 294, 301 (9th Cir.1992) (Wallace, J„ concurring) (stating that he was "unable to discover any basis for 'supervisory power,' and therefore question[ed] its validity.”).

. See, e.g., McNabb v. United States, 318 U.S. 332, 340 , 63 S.Ct. 608, 613 , 87 L.Ed. 819 (1943) (“Judicial supervision of the administration of criminal justice in the federal courts implies the duty of establishing and maintaining civilized standards of procedure and evidence.”).