474 U.S. 302 (1986)
delivered the opinion of the Court.
In this case we must decide, first, whether the Speedy Trial Clause of the Sixth Amendment1 applies to time during which respondents were neither under indictment nor subjected to any official restraint, and, second, whether certain delays occasioned by interlocutory appeals were properly weighed in assessing respondents’ right to a speedy trial. A divided panel of the Court of Appeals for the Ninth Circuit weighed most of the 90 months from the time of respondents’ arrests and initial indictment in November 1975 until the District Court’s dismissal of the indictment in May 1983 towards respondents’ claims under the Speedy Trial Clause. We conclude that the time that no indictment was outstanding against respondents should not weigh towards respondents’ speedy trial claims. We also find that in this case the delay attributable to interlocutory appeals by the Government and respondents does not establish a violation of the Speedy Trial Clause. Accordingly, we reverse the holding of the Court of Appeals that respondents were denied their right to a speedy trial.
In view of the nature of respondents’ claim, we state the factual and procedural history of this case in some detail. On November 14, 1975, pursuant to a tip from the Federal Bureau of Investigation, Oregon state troopers stopped two vehicles in search of several federal fugitives.2 After an exchange of gunfire and a motor chase, state troopers captured all but one of the respondents, Dennis Banks.3 Both vehicles were locked and impounded while federal and state authorities obtained search warrants.
Searches of the vehicles over the next two days disclosed 350 pounds of dynamite,4 6 partially assembled time bombs, 2,600 rounds of ammunition, 150 blasting caps, 9 empty hand grenades, and miscellaneous firearms.5 Oregon law enforcement officers, apparently unaware of the evidentiary consequences, adhered to their usual policy and destroyed the dynamite. A federal agent present at the destruction photographed the explosions. United States v. Loud Hawk, 628 F. 2d 1139, 1142 (CA9 1979). State officials also preserved wrappers from the dynamite casings.
A federal grand jury indicted respondents on November 25, 1975, on charges of possessing firearms and explosives. Trial in the United States District Court for the District of Oregon was set for the week of February 9, 1976. On December 22, 1975, a grand jury returned a five-count superseding indictment. This indictment charged all respondents with three counts relating to possession and transportation in commerce of an unregistered destructive device (the dynamite counts) and two counts relating to unlawful possession of firearms (the firearms counts).
Two days later, respondents filed a motion to suppress all evidence concerning the dynamite, arguing that federal and state officials had intentionally and negligently destroyed the dynamite before the defense had the opportunity to examine it. After initially denying respondents’ motion,6 and after two continuances at respondents’ behest,7 the District Court granted respondents’ motion to suppress on March 31, 1976. App. to Pet. for Cert. 157a. Three weeks later, the Government appealed the suppression order,8 and moved that trial on all counts be continued pending the outcome of the appeal. The District Court denied the Government’s request for a continuance, and when the case was called for trial, the Government answered “not ready.” Pursuant to Federal Rule of Criminal Procedure 48(b), the District Judge dismissed the indictment with prejudice. Six months had passed since the original indictment.
The Government immediately appealed the dismissal, and the two appeals were consolidated. The Court of Appeals heard argument on October 15, 1976, and a divided panel affirmed in an unreported opinion on July 26, 1977. App. to Pet. for Cert. 88a-118a. On the Government’s motion, the court voted on October 17, 1977, to hear the case en banc. On March 6, 1978, the Court of Appeals en banc remanded for findings of fact on whether federal officials participated in the destruction of the dynamite and whether respondents were prejudiced by its destruction. The court retained jurisdiction over the appeal pending the District Court’s findings. The District Court issued its findings on August 23, 1978, and the case returned to the Court of Appeals.
On August 7, 1979, the Court of Appeals reversed the suppression order and directed that the dynamite counts be reinstated. United States v. Loud Hawk, 628 F. 2d, at 1150. The court also held that although the Government could have gone to trial on the firearms counts pending the appeal, the District Court erred in dismissing those counts with prejudice. Id., at 1151. The Court of Appeals denied respondents’ petition for rehearing on October 1, 1979. Respondents petitioned for certiorari; we denied the petition on March 3, 1980. 445 U. S. 917. The mandate of the Court of Appeals issued on March 12, 1980, 46 months after the Government filed its notice of appeal from the dismissal of the indictment. Respondents were unconditionally released during that time.
Following remand, the District Court ordered the Government to reindict on the firearms charges.9 Respondents filed a number of motions during June and July of 1980 in response to the superseding indictment,10 including a motion to dismiss for vindictive prosecution. On August 8, 1980, the District Court granted the vindictive prosecution motion as to KaMook Banks and denied it as to respondents Dennis Banks, Render, and Loud Hawk. Both sides appealed. Respondents remained free on their own recognizance during this appeal.
The appeals were consolidated, and the Court of Appeals ordered expedited consideration. The court heard argument on January 7, 1981, but did not issue its decision until July 29, 1982. The court sustained the Government’s position on all issues. United States v. Banks, 682 F. 2d 841. Respondents’ petitions for rehearing were denied on October 5, 1982. Respondents again petitioned for certiorari, and we denied the petition on January 10, 1983. 459 U. S. 1117. The Court of Appeals’ mandate issued on January 31, 1983, almost 29 months after the appeals were filed.
The District Court scheduled trial to begin on April 11, 1983. The Government sought and received a continuance until May 3, 1983, because of alleged difficulties in locating witnesses more than seven years after the arrests. Subsequently, the court on its own motion continued the trial date until May 23, 1983, and then again rescheduled the trial for June 13. The record in this Court does not reveal the reasons for these latter two continuances. Defendants objected to each continuance.
On May 20, 1983, the District Court again dismissed the indictment, this time on the ground that respondents’ Sixth Amendment right to a speedy trial had been violated. 564 F. Supp. 691. The Government appealed, and unsuccessfully urged the District Court to request that the Court of Appeals expedite the appeal. On its own motion the court treated the appeal as expedited, and heard argument on January 4, 1984. A divided panel affirmed on August 30, 1984. 741 F. 2d 1184.11 We granted certiorari, 471 U. S. 1014 (1985), and now reverse.
The Government argues that under United States v. MacDonald, 456 U. S. 1 (1982), the time during which defendants are neither under indictment nor subject to any restraint on their liberty should be excluded — weighed not at all — when considering a speedy trial claim.12 Respondents contend that even during the time the charges against them were dismissed, the Government was actively pursuing its case and they continued to be subjected to the possibility that bail might be imposed. This possibility, according to respondents, is sufficient to warrant counting the time towards a speedy trial claim.
The Court has found that when no indictment is outstanding, only the “actual restraints imposed by arrest and holding to answer a criminal charge . . . engage the particular protections of the speedy trial provision of the Sixth Amendment.” United States v. Marion, 404 U. S. 307, 320 (1971) (emphasis added); see MacDonald, supra, at 9. As we stated in MacDonald: “The speedy trial guarantee is designed to minimize the possibility of lengthy incarceration prior to trial, to reduce the lesser, but nevertheless substantial, impairment of liberty imposed on an accused while released on bail, and to shorten the disruption of life caused by arrest and the presence of unresolved criminal charges.” 456 U. S., at 8.
During much of the litigation, respondents were neither under indictment nor subject to bail.13 Further judicial proceedings would have been necessary to subject respondents to any actual restraints. Cf. Klopfer v. North Carolina, 386 U. S. 213 (1967). As we stated in MacDonald: “[W]ith no charges outstanding, personal liberty is certainly not impaired to the same degree as it is after arrest while charges are pending. After the charges against him have been dismissed, ‘a citizen suffers no restraints on his liberty and is [no longer] the subject of public accusation: his situation does not compare with that of a defendant who has been arrested and held to answer.’” 456 U. S., at 9.
Respondents argue that the speedy trial guarantee should apply to this period because the Government’s desire to prosecute them was a matter of public record. Public suspicion, however, is not sufficient to justify the delay in favor of a defendant’s speedy trial claim. We find that after the District Court dismissed the indictment against respondents and after respondents were freed without restraint, they were “in the same position as any other subject of a criminal investigation.” MacDonald, supra, at 8-9. See Marion, supra, at 309. The Speedy Trial Clause does not purport to protect a defendant from all effects flowing from a delay before trial. The Clause does not, for example, limit the length of a pre-indictment criminal investigation even though “the [suspect’s] knowledge of an ongoing criminal investigation will cause stress, discomfort, and perhaps a certain disruption in normal life.” 456 U. S., at 9.
Nor does the fact that respondents were ordered to appear at the evidentiary hearing held on remand in the District Court during the first appeal — an appearance they waived— constitute the sort of “actual restraint” required under our precedents as a basis for application of the Speedy Trial Clause. Finally, we are not persuaded that respondents’ need for counsel while their case was technically dismissed supports their speedy trial claim. Although the retention of counsel is frequently an inconvenience and an expense, the Speedy Trial Clause’s core concern is impairment of liberty; it does not shield a suspect or a defendant from every expense or inconvenience associated with criminal defense.
We therefore find that under the rule of MacDonald, when defendants are not incarcerated or subjected to other substantial restrictions on their liberty, a court should not weigh that time towards a claim under the Speedy Trial Clause.
The remaining issue is how to weigh the delay occasioned by an interlocutory appeal when the defendant is subject to indictment or restraint. As we have recognized, the Sixth Amendment’s guarantee of a speedy trial “is an important safeguard to prevent undue and oppressive incarceration prior to trial, to minimize anxiety and concern accompanying public accusation and to limit the possibilities that long delay will impair the ability of an accused to defend himself.” United States v. Ewell, 383 U. S. 116, 120 (1966). These safeguards may be as important to the accused when the delay is occasioned by an unduly long appellate process as when the delay is caused by a lapse between the initial arrest and the drawing of a proper indictment, Ewell, supra, at 118-119, or by continuances in the date of trial, Barker v. Wingo, 407 U. S. 514, 517-518 (1972).
At the same time, there are important public interests in the process of appellate review. The assurance that motions to suppress evidence or to dismiss an indictment are correctly decided through orderly appellate review safeguards both the rights of defendants and the “rights of public justice.” Beavers v. Haubert, 198 U. S. 77, 87 (1905). The legislative history of 18 U. S. C. § 3731 “makes it clear that Congress intended to remove all statutory barriers to Government appeals and to allow appeals whenever the Constitution would permit.” United States v. Wilson, 420 U. S. 332, 337 (1975).
It is, of course, true that the interests served by appellate review may sometimes stand in opposition to the right to a speedy trial. But, as the Court observed in United States v. Ewell, supra, at 121:
“It has long been the rule that when a defendant obtains a reversal of a prior, unsatisfied conviction, he may be retried in the normal course of events. . . . [This rule] has been thought wise because it protects the societal interest in trying people accused of crime, rather than granting them immunization because of legal error at a previous trial, and because it enhances the probability that appellate courts will be vigilant to strike down previous convictions that are tainted with reversible error. . . . These policies, so carefully preserved in this Court’s interpretation given the Double Jeopardy Clause, would be seriously undercut by [an] interpretation given the Speedy Trial Clause [that raised a Sixth Amendment obstacle to retrial following successful attack on conviction].”
In Barker, we adopted a four-part balancing test to determine whether a series of continuances infringed upon the defendant’s right to a speedy trial. 407 U. S., at 530. That test assessed the “[l]ength of delay, the reason for the delay, the defendant’s assertion of his right, and prejudice to the defendant.” Ibid. (footnote omitted). The Barker test furnishes the flexibility to take account of the competing concerns of orderly appellate review on the one hand, and a speedy trial on the other. We therefore adopt this functional test to determine the extent to which appellate time consumed in the review of pretrial motions should weigh towards a defendant’s speedy trial claim. Under this test, we conclude that in this case the delays do not justify the “unsatisfactorily severe remedy of dismissal.” Id., at 522.
Barker’s first, third, and fourth factors present no great difficulty in application. The first factor, the length of delay, defines a threshold in the inquiry: there must be a delay long enough to be “presumptively prejudicial.” Id., at 530. Here, a 90-month delay in the trial of these serious charges is presumptively prejudicial and serves to trigger application of Barker’s other factors. Ibid.
The third factor — the extent to which respondents have asserted their speedy trial rights — does not support their position. Although the Court of Appeals found that respondents have repeatedly moved for dismissal on speedy trial grounds, 741 F. 2d, at 1192, that finding alone does not establish that respondents have appropriately asserted their rights. We held in Barker that such assertions from defendants are “entitled to strong evidentiary weight” in determining whether their rights to a speedy trial have been denied. 407 U. S., at 531-532. These assertions, however, must be viewed in the light of respondents’ other conduct.
Here, respondents’ speedy trial claims are reminiscent of Penelope’s tapestry.14 At the same time respondents were making a record of claims in the District Court for speedy trial, they consumed six months by filing indisputably frivolous petitions for rehearing and for certiorari after this Court’s decision in United States v. Hollywood Motor Car Co., 458 U. S. 263 (1982) (federal courts without jurisdiction to hear defendant’s interlocutory appeal from denial of motion to dismiss indictment). They also filled the District Court’s docket with repetitive and unsuccessful motions. See, e. g., n. 10, supra.
The Court of Appeals gave “little weight” to the fourth factor, prejudice to respondents. At most, the court recognized the possibility of “impairment of a fair trial that may well result from the absence or loss of memory of witnesses in this case.” 741 F. 2d, at 1193. See Barker, 407 U. S., at 532. That possibility of prejudice is not sufficient to support respondents’ position that their speedy trial rights were violated. In this case, moreover, delay is a two-edged sword. It is the Government that bears the burden of proving its case beyond a reasonable doubt. The passage of time may make it difficult or impossible for the Government to carry this burden.
The flag all litigants seek to capture is the second factor, the reason for delay. In Barker, we held that “different weights should be assigned to different reasons.” Id., at 531. While a “deliberate attempt to delay the trial in order to hamper the defense,” would be weighed heavily against the Government, a delay from “overcrowded courts” — as was the situation here — would be weighed “less heavily.” Ibid. Given the important public interests in appellate review, supra, at 313, it hardly need be said that an interlocutory appeal by the Government ordinarily is a valid reason that justifies delay. In assessing the purpose and reasonableness of such an appeal, courts may consider several factors. These include the strength of the Government’s position on the appealed issue, the importance of the issue in the posture of the case, and — in some cases — the seriousness of the crime. United States v. Herman, 576 F. 2d 1139, 1146 (CA5 1978) (Wisdom, J.). For example, a delay resulting from an appeal would weigh heavily against the Government if the issue were clearly tangential or frivolous. Ibid. Moreover, the charged offense usually must be sufficiently serious to justify restraints that may be imposed on the defendant pending the outcome of the appeal. Ibid.
Under Barker, delays in bringing the case to trial caused by the Government’s interlocutory appeal may be weighed in determining whether a defendant has suffered a violation of his rights to a speedy trial. It is clear in this case, however, that respondents have failed to show a reason for according these delays any effective weight towards their speedy trial claims. There is no showing of bad faith or dilatory purpose on the Government’s part. The Government’s position in each of the appeals was strong, and the reversals by the Court of Appeals are prima facie evidence of the reasonableness of the Government’s action. Moreover, despite the seriousness of the charged offenses, the District Court chose not to subject respondents to any actual restraints pending the outcome of the appeals.
The only remaining question is the weight to be attributed to delays caused by respondents’ interlocutory appeals. In that limited class of cases where a pretrial appeal by the defendant is appropriate, see, e. g., Hollywood Motor Car Co., supra, at 265-266, delays from such an appeal ordinarily will not weigh in favor of a defendant’s speedy trial claims. A defendant with a meritorious appeal would bear the heavy burden of showing an unreasonable delay caused by the prosecution in that appeal, or a wholly unjustifiable delay by the appellate court. A defendant who resorts to an interlocutory appeal normally should not be able upon return to the district court to reap the reward of dismissal for failure to receive a speedy trial. As one Court of Appeals has noted in the context of a District Court’s consideration of pretrial motions:
“Having sought the aid of the judicial process and realizing the deliberateness that a court employs in reaching a decision, the defendants are not now able to criticize the very process which they so frequently called upon.” United States v. Auerbach, 420 F. 2d 921, 924 (CA5 1969), rehearing denied, 423 F. 2d 676, cert. denied, 399 U. S. 905 (1970).
In the present case, respondents’ appeal was allowable under the law of the Ninth Circuit before our decision in Hollywood Motor Car, supra. But we find that their position was so lacking in merit that the time consumed by this appeal should not weigh in support of respondents’ speedy trial claim. Nor do we weigh the additional delay of six months resulting from respondents’ frivolous action in seeking rehearing and certiorari toward respondents’ speedy trial claim. See ibid., decided prior to these latter actions.
We cannot hold, on the facts before us, that the delays asserted by respondents weigh sufficiently in support of their speedy trial claim to violate the Speedy Trial Clause. They do not justify the severe remedy of dismissing the indictment. Accordingly, the judgment of the Court of Appeals for the Ninth Circuit is reversed.
It is so ordered.
Justice Marshall, with whom Justice Brennan, Justice Blackmun, and Justice Stevens join, dissenting.
The Court holds today that the Speedy Trial Clause of the Sixth Amendment does not apply to a Government appeal from a district court’s dismissal of an indictment, unless the defendant is incarcerated or otherwise under restraint during that appeal. The majority supports this result by equating the present case to United States v. MacDonald, 456 U. S. 1 (1982). That analysis, however, both ignores the considerable differences between this case and MacDonald and gives short shrift to the interests protected by the Speedy Trial Clause. I further disagree with the majority’s application of Barker v. Wingo, 407 U. S. 514 (1972), to the remaining appellate delays in this case.
The majority concludes that when an appeal arises out of the district court’s dismissal of an indictment, the lack of an outstanding indictment absolves the Government of its responsibility to provide a speedy trial. However, we have never conditioned Sixth Amendment rights solely on the presence of an outstanding indictment. Those rights attach to anyone who is “accused,”15 and we have until now recognized that one may stand publicly accused without being under indictment. The majority offers two reasons for concluding that respondents did not enjoy the right to a speedy trial during the Government’s appeals. First, respondents were suffering only “[p]ublic suspicion,” ante, at 311, and not a formal accusation. Second, they were not subject to “actual restraints” on their liberty. Both of these rationales are seriously flawed.
A
In United States v. Marion, 404 U. S. 307 (1971), we held that the Speedy Trial Clause does not apply until the Government, either through arrest or indictment, asserts probable cause to believe that a suspect has committed a crime. Before that time the individual, while possibly aware of the Government’s suspicion, is not “the subject of public accusation,” id., at 321, and his only protection against delay comes from the Due Process Clause and the applicable statute of limitations. The Court applied the same rationale in MacDonald, supra. In that case, military charges of murder against MacDonald, an Army officer, were dropped after an investigation. MacDonald was then given an honorable discharge, only to be indicted by a civilian grand jury nearly four years later for the same murders. The Court held that this delay did not implicate the speedy trial right because “the Speedy Trial Clause has no application after the Government, acting in good faith, formally drops charges.” Id., at 7. The Court reasoned that after the termination of the first formal prosecution, MacDonald was “in the same position as any other subject of a criminal investigation,” id., at 8-9, and thus was no more an “accused” than was the defendant in Marion before his arrest.
The same cannot be said of respondents in the present case.16 Unlike one who has not been arrested, or one who has had the charges against him dropped, respondents did not enjoy the protection of the statute of limitations while the Government prosecuted its appeals. That protection was an important aspect of our holding in Marion that prearrest delay is not cognizable under the Speedy Trial Clause. See 404 U. S., at 322-323. More importantly, in contrast to MacDonald, the Government has not “dropped” anything in this case.17 There has been at all relevant times a case on a court docket captioned United States v. Loud Hawk — I can think of no more formal indication that respondents stand accused by the Government.
The majority argues that while “the Government’s desire to prosecute [respondents] was a matter of public record,” that desire constituted only “[p]ublic suspicion” that is insufficient to call Sixth Amendment rights into play, citing Marion and MacDonald. Ante, at 311. The reason that the Government’s desire to prosecute in both of those cases did not constitute an “accusation,” however, is that the Government had not yet formalized its commitment. Indeed, in MacDonald, the Government dismissed the murder charges because it “concluded that they were untrue,” 456 U. S., at 10, n. 12, thus acknowledging that the first formal accusation had been a mistake and extinguishing the prior probable-cause determination. In the present case, the Government has made no such confession of error and continues to align its full resources against respondents in judicial proceedings.
The most telling difference between this case and MacDonald, however, is the fact that respondents’ liberty could have been taken from them at any time during the Government’s appeal. One of the primary purposes of the speedy trial right, of course, is to prevent prolonged restraints on liberty, id., at 8; Barker v. Wingo, 407 U. S., at 532, and the absence of any possibility of such restraints was a vital part of our MacDonald holding. See 456 U. S., at 9. In contrast, Congress has declared explicitly, in 18 U. S. C. §3731, that a person in respondents’ position shall be subject to the same restraints as an arrested defendant awaiting trial.18 Thus the District Court had the undoubted authority to condition respondents’ release on the posting of bail, or indeed to keep them in jail throughout the appeal, see 18 U. S. C. § 3142(e) (1982 ed., Supp. III). Respondents’ release could have been accompanied by restrictions on travel, association, employment, abode, and firearms possession, or conditioned on their reporting regularly to law enforcement officers and/or keeping a curfew. See § 3142(c). Considering all the circumstances, therefore, I believe that respondents’ position is most closely analogous to that of a defendant who has been arrested but not yet indicted.
B
As if acknowledging that the delay in this case is more analogous to postarrest, preindictment delay than to pre-arrest delay, the majority concedes that had respondents been incarcerated or forced to post bond during the Government’s appeals, the automatic exclusion rule of MacDonald would not apply. Ante, at 311, n. 13. Yet, inexplicably, the majority then suggests that the Speedy Trial Clause applies to postarrest, preindictment delay only when the defendant has been subjected to “‘actual restraints,’” ante, at 310, quoting Marion, 404 U. S., at 320 (emphasis added by majority opinion). The majority completely misreads Marion while creating a rule that is flatly inconsistent with our prior holdings.
We held in Marion that prearrest delay is not cognizable under the Speedy Trial Clause, but we certainly did not disturb the settled rule that the Government’s formal institution of criminal charges, whether through arrest or indictment, always calls the speedy trial right into play. See id., at 316-319; see also United States v. Gouveia, 467 U. S. 180, 185-186 (1984). Although it specified detention and bail as possible deleterious effects of a formal criminal charge, Marion nowhere suggested that it is the restraints themselves, rather than the assertion of probable cause, that constitute an accusation. Nor did we hold that a criminal charge has less constitutional significance when a defendant is released on recognizance rather than on bail. See 404 U. S., at 321, n. 12. The majority identifies no logic or precedent supporting its novel conclusion that a defendant who is arrested and released on bail is “accused,” while a defendant who is arrested and released without bail, on the same evidence, is not “accused.”19
Indeed, we have rejected precisely the interpretation of Marion that the majority now adopts. In Dillingham v. United States, 423 U. S. 64 (1975) (per curiam), we held that Marion does not require “actual prejudice” to invoke the speedy trial right for postarrest, preindictment delay. Such “actual prejudice” included the “actual restraints” that the majority now requires. The Court of Appeals in that case noted that the defendant was released on bond, but without any other restrictions, pending trial. After citing Marion, it held that “any increased strain on this man’s life which followed his arrest. . . does not rise to the level of substantial actual prejudice.” United States v. Palmer, 502 F. 2d 1233, 1237 (CA5 1974), rev’d sub nom. Dillingham v. United States, supra. We summarily rejected the “actual prejudice” rationale, and the majority gives no reason whatsoever for resurrecting it today.20
There can be no question that one who had been arrested and released under 18 U. S. C. § 3141(a) (1982 ed., Supp. III) would be entitled, under Marion, to the protections of the Speedy Trial Clause. Because respondents were by statute subject to the same restraints as that hypothetical defendant, I am at a loss to understand why they should enjoy less protection.
II
The majority also declines to hold the Government accountable for delay attributable to appeals during which respondents were under indictment. In doing so the majority emphasizes the second Barker factor — the reason for the delay, see 407 U. S., at 530. Because it concludes that “[t]here is no showing of bad faith or dilatory purpose on the Government’s part,” the majority declines to accord any “effective weight” to this factor in the speedy trial balance. Ante, at 316. In reaching this conclusion, it virtually ignores the most obvious “reason for the delay” in this case — the fact that the Court of Appeals was unable to decide these appeals in a reasonably prompt manner.
In Barker, we explained the application of the “reason for the delay” factor as follows:
“[D]ifferent weights should be assigned to different reasons. A deliberate attempt to delay the trial in order to hamper the defense should be weighted heavily against the government. A more neutral reason such as negligence or overcrowded courts should be weighted less heavily but nevertheless should be considered since the ultimate responsibility for such circumstances must rest with the government rather than with the defendant.” 407 U. S., at 531 (footnote omitted).
The majority’s application of this factor to the appellate delays in this case makes Government misconduct or bad faith a virtual prerequisite to a finding of a speedy trial violation. Seizing upon the approach of some of the Courts of Appeals,21 the majority analyzes the reason behind the appellate delay solely in terms of the reasonableness of the Government’s behavior in taking and prosecuting the appeal. This approach is inconsistent with the policies behind the speedy trial right. We recognized in Barker that the right protects both the defendant’s interest in fairness and society’s interest in providing swift justice. Id., at 519. Courts as well as prosecutors must necessarily work to promote those interests if they are to have any vitality. Because it is the Government as a whole — including the courts — that bears the responsibility to provide a speedy trial, the prosecutor’s good faith cannot suffice to discharge that responsibility.22
The Court of Appeals frankly admitted that “most of the delay must be attributed to the processes of this court,” 741 F. 2d 1184, 1191 (CA9 1984), a conclusion that is difficult to escape. This case involves appeals from pretrial rulings. The Court of Appeals had every reason to know that these appeals should have been ruled upon as expeditiously as possible. See that court’s Rule 20. Yet it took over five years for the Court of Appeals to decide two appeals, one of them “expedited.” No complicated analysis is needed to identify the reason for the delay in this case.
I would hold, simply, that a nonfrivolous appeal by any party permits a reasonable delay in the proceedings. The number and complexity of the issues on appeal, or the number of parties, might permit a greater or lesser delay in a given case. The government, not the defendant, must suffer the ultimate consequences of delays attributable to “overcrowded courts,” ibid., even at the appellate level.23 In the present case, the amount of time that the appeals consumed is patently unreasonable. I would therefore weigh the second Barker factor against the Government in this case.
III
The majority has seriously misapplied our precedents in concluding that delay resulting when the government appeals the dismissal of an indictment is excludable for speedy trial purposes unless the defendant is subjected to actual restraints during that appeal. Its application of Barker v. Wingo to this case also undercuts the very purpose of the speedy trial right. I respectfully dissent.