DAVID R. THOMPSON, Circuit Judge:
These are the consolidated appeals of defendants Gary Edward Aexander, Jonathan Harrington, Anthony F.- Hicks and Willie James Harris. The defendants appeal their convictions for conspiracy to commit robbery, in violation of 18 U.S.C. § 371; armed bank robbery, in violation of 18 U.S.C. § 2113(a) and (d); and use of a firearm during commission of a crime of violence, in violation of 18 U.S.C. § 924(c).
The defendants contend their Sixth Amendment right to trial by an impartial jury was violated because the district court (1) failed to strike for cause two prospective jurors who had previously been victims of robbery, and, (2) replaced a juror with an alternate on the last day of trial when the juror’s child became ill and had to be taken [1482]*1482to a hospital. They also argue the district court committed reversible error by admitting the government’s FDIC “Certificate of Proof of Insured Status.” as evidence that the victim bank was federally insured at the time of the robbery. In addition, defendant Hicks seeks reversal of his conviction on the ground that the district court erroneously admitted, for impeachment purposes, evidence of his prior drug and robbery convictions. Defendant Alexander contends the district court should have granted his motion for judgment of acquittal because he was not properly identified at trial.
The defendants also appeal their sentences. They argue their sentences should be recalculated because the sentence enhancement applied for robbery of a financial institution, pursuant to United States Sentencing Guidelines [hereinafter USSG] § 2B3.1(b)(l), deprived them of their right to substantive due process. They also contend other provisions of the guidelines under which they were sentenced unconstitutionally permitted double counting of various sentence enhancements. Defendant Harrington also argues he should have been given a two-point reduction in his base offense level for acceptance of responsibility. Finally, defendant Hicks argues the district court improperly enhanced his sentence based on his prior drug conviction, because his guilty plea to that crime was neither knowing, intelligent nor voluntary.
We have jurisdiction pursuant to 28 U.S.C. § 1291. We affirm • all convictions and all sentences.
FACTS
On January 24, 1992, four armed men — all wearing blue coveralls, gloves, and ski masks — burst into the First Interstate Bank in Victorville, California. They ordered everyone present in the bank to lie down, forced two of the bank’s employees to open the vault, and emptied cash from the vault into a duffle bag. After taking the money, the robbers fled in a van. The total amount stolen was $331,951.
The police had been alerted to the robbery by a passerby. As a result, officers arrived at the scene in time to pursue the van as it left the bank. A high-speed chase ensued in which the occupants of the van fired several shots at the police.
Eventually, the van pulled into a K-Mart parking lot. The four robbers got out of the van and split up. Two of them got into a Camaro and the other two got into a Camry. The vehicles sped away in different directions, with the police in pursuit. The chase continued at speeds up to 110 miles per hour, with more shots being fired at the police from the fleeing vehicles.
Eventually, the Camaro stalled; defendants Harrington and Alexander were arrested inside the car. The Camry also came to a stop, but its occupants continued their flight on foot. After a brief chase, police arrested defendants Harris and Hicks a short distance from the abandoned Camry.
Remaining pertinent facts are discussed in relevant portions of the discussion that follows.
DISCUSSION
I. Sixth Amendment Violations
A. Denial of Challenges for Cause
During jury selection, prospective jurors Mark Austin and Phyllis Kenny disclosed they had been victims of robberies. Austin said he “believed” he nonetheless could be fair and impartial despite the subject matter of the case — armed bank robbery. Upon additional interrogation, by the court, Austin stated unequivocally that he could put aside his own experience and act fairly.1 Prospec-[1483]*1483five juror Kenny twice said she “believed” she could be fair and impartial despite her experience, but never gave a more definitive statement.2
The defense challenged both prospective jurors Austin and Kenny for cause but the district court, satisfied that they could act fairly, denied the challenges. The defense then used peremptory challenges to excuse Austin and' Kenny. Ultimately, the defense exhausted its full allotment of peremptory challenges. The defendants now contend that the district court’s denial of their challenges for cause violated their Sixth Amendment right to an impartial jury, because it forced them to exercise peremptory challenges they otherwise could have used against other prospective jurors.
We reject the defendants’ argument. First, we note that, in Ross v. Oklahoma, 487 U.S. 81, 108 S.Ct. 2273, 101 L.Ed.2d 80 (1988), the Supreme Court specifically held that loss of a peremptory challenge because of a trial court’s improper failure to grant a challenge for cause does not amount to a violation of a constitutional right without a showing of prejudice. Id. at 88, 108 S.Ct. at 2278. As the Court explained:
[P]eremptory challenges are not of constitutional dimension. They are a means to achieve the end of an impartial jury. So long as the jury that sits is impartial, the fact that the defendant had to use a peremptory challenge to achieve that result does not mean the Sixth Amendment was violated.
Id. (citations omitted).
In the present case, the defendants removed prpspective jurors Austin and Kenny with peremptory challenges. They have not asserted that the jury which finally tried [1484]*1484them was in any way biased or prejudiced. Consequently, they have not shown any prejudice from the court’s denial of their challenges for cause.
In arguing they should not be required to show prejudice, the defendants rely on our decision in United States v. Claiborne, 765 F.2d 784, 799-800 (9th Cir.1985), cert. denied, 475 U.S. 1120,106 S.Ct. 1636, 90 L.Ed.2d 182 (1986), and the Fifth Circuit’s decision in United States v. Nell, 526 F.2d 1223, 1229 (5th Cir.1976).' These cases are not persuasive. We have specifically stated that, because it is a pre-Ross case, Claiborne is no longer good law. United States v. Baker, 10 F.3d 1374, 1404 n. 12 (9th Cir.1993), cert. denied, — U.S. -, 115 S.Ct. 330, 130 L.Ed.2d 289 (1994). Although the Fifth Circuit has not explicitly overruled Nell, it has recognized that a showing of prejudice is required for reversal of a conviction based on a trial court’s erroneous denial of a challenge for cause. See United States v. Mendoza-Burciaga, 981 F.2d 192, 197-98 (5th Cir. 1992) (denying a reversal because the defendant had not alleged prejudice from having to use a peremptory challenge to exclude a potential juror — no comment on Nell), cert. denied, — U.S. -, 114 S.Ct. 356, 126 L.Ed.2d 320 (1993). But see United States v. Bryant, 991 F.2d 171, 174 n. 3 (5th Cir.1993) (expressing doubt as to whether the Nell or Mendozar-Burciaga line of cases is controlling); United States v. Munoz, 15 F.3d 395, 398 n. 1 (5th Cir.1994) (suggesting that Ross' may not be applicable to review of federal criminal convictions on direct appeal).
Apart from the defendants’ failure to show prejudice, the district court determined that prospective jurors Austin and Kenny could serve impartially. This determination was neither an abuse of discretion nor manifest error. See United States v. Egbuniwe, 969 F.2d 757, 762 (9th Cir.1992) (Because “determination[s] of impartiality, in which demeanor plays such an important part, [are] particularly within the province of the trial judge,” an appellate court should not disturb a district court’s decision to deny challenges for cause without a showing of abuse of discretion or manifest error.) (quoting Ristaino v. Ross, 424 U.S. 589, 595, 96 S.Ct. 1017, 1020, 47 L.Ed.2d 258 (1976)). See also United States v. Baker, 10 F.3d 1374, 1403 (9th Cir.1993), cert. denied, — U.S. -, 115 S.Ct. 330, 130 L.Ed.2d 289 (1994); United States v. Poschwatta, 829 F.2d 1477, 1484 (9th Cir.1987), cert. denied, 484 U.S. 1064, 108 S.Ct. 1024, 98 L.Ed.2d 989 (1988).
Prospective juror Austin initially said he “believed” he could be impartial, but he ultimately stated definitively that he could separate his experience from the facts of the case and act fairly. The district court was not required to excuse him for cause based solely on his initial response. See United States v. Daly, 716 F.2d 1499, 1507 (9th Cir.1983) (finding no abuse of discretion in the district court’s refusal to remove for cause a prospective juror who initially said he would “try” to be impartial and ultimately stated, “Okay, I will do it”), cert. dismissed, 465 U.S. 1075, 104 S.Ct. 1456, 79 L.Ed.2d 773 (1984); Celestine v. Blackburn, 750 F.2d 353, 358-60 (5th Cir.1984) (affirming the district court’s decision in a murder trial not to remove for cause a prospective juror who was friendly with the victim’s granddaughter, when the prospective juror admitted it would be difficult to set aside her feelings but promised to “try,” and who ultimately stated definitively that she would be impartial), cert. denied, 472 U.S. 1022, 105 S.Ct. 3490, 87 L.Ed.2d 624 (1985).
Prospective juror Kenny’s case presents a closer question, because she never followed her initial response that she “believed” she could be fair with a more definite statement. The defendants bear the burden, however, of showing that Kenny was actually •biased, and that the district court abused its discretion or committed manifest error when it failed to excuse her for cause. Irvin v. Dowd, 366 U.S. 717, 723, 81 S.Ct. 1639, 1643, 6 L.Ed.2d 751 (1961). The defendants have not met their burden. Giving deference to the district court’s determination, because it was able to observe Kenny’s demeanor and assess her credibility, we cannot say the court abused its discretion or committed manifest error in concluding that when Kenny said she “believed” she could act impartially, this was equivalent to saying she would do so.
[1485]*1485B. Replacement of Juror
On the last full day of trial, the district court informed counsel for both sides that juror Ragsdale, ■ a single parent, was absent due to his daughter’s illness. The child was sick to the point of vomiting and had to be taken to a hospital. The court indicated Ragsdale would not be available for the remainder of the day, and his availability beyond that point was uncertain, because it depended on his daughter’s condition. Over unanimous objections by counsel for all defendants, the district court, pursuant to Rule 24(c) of the Federal Rules of Criminal Procedure, excused Ragsdale and replaced him with an alternate.3
The defendants argue the district court abused its discretion by excusing Ragsdale without adequate inquiry into the nature of his daughter’s illness and the- estimated duration of his consequent absence. They cite United States v. Smith, 550 F.2d 277, 285 (5th Cir.), cert. denied, 434 U.S. 841, 98 S.Ct. 138, 54 L.Ed.2d 105 (1977), for the proposition that, before a juror may be replaced by an alternate under Rule 24(c), sufficient facts must be presented to convince the trial judge that the juror’s ability to perform his duties has been impaired.
We review for abuse of discretion a district court’s decision to replace a juror with an alternate. United States v. Gay, 967 F.2d 322, 324 (9th Cir.), cert. denied, — U.S. -, 113 S.Ct. 359, 121. L.Ed.2d 272 (1992).
We note, first, that Smith, the very ease upon which the defendants rely, also held that a trial court’s exercise of discretion in excusing a juror under Rule 24(c) is not to be disturbed by an appellate court without a showing of bias or prejudice to the defendant. Id. In United States v. Trigg, 988 F.2d 1008, 1011 (9th Cir.1993), we recognized this limitation on Smith, and on that basis rejected the precise argument the defendants make here. Because the defendants “make[] no assertion[] that [they] suffered bias or prejudice, and do[ ] not claim that the alternate juror ... was not impartial,” their convictions would not be subject to reversal even were we to conclude that dismissal of' juror Ragsdale was an abuse of the district court’s discretion. United States v. Agramante, 980 F.2d 847, 850. (2d Cir.1992).
Moreover, we conclude the district court acted within its discretion in excusing juror Ragsdale. ‘Whether a juror’s absence is sufficiently disruptive to warrant removal is ... a function of the managerial complexity of the case, the flexibility of the court’s and parties’ schedules, and the availability of witnesses and other evidence.” Gay, 967 F.2d at 325.
This case involved four codefendants, lasted almost two weeks, and involved the testimony of numerous witnesses. The district court noted that if Ragsdale were not excused and replaced by an alternate juror, the trial would not conclude that day, but instead would be delayed not only for the remainder of the day, but by an upcoming three-day weekend; and trial could not resume on the ensuing Tuesday because another juror had previously scheduled an exam for that day. The delay would have been somewhat more than four days. •
Courts have approved the’replacement of jurors in cases of comparable complexity and similar potential delay. See United States v. Wilson, 894 F.2d 1245 (11th Cir.) (approving dismissal of ill juror in nine-week trial, where the juror would have been absent an indeterminate number of days), cert. denied, 497 U.S. 1029, 110 S.Ct. 3284, 111 L.Ed.2d 792 (1990); United States v. Armijo, 834 F.2d 132 (8th Cir.1987) (affirming replacement of juror who was involved in a ear accident when without replacement the five-day trial would have been delayed at least a week), cert. denied, 485 U.S. 990, 108 S.Ct. 1297, 99 L.Ed.2d 507 (1988); United States v. Stratton, 779 F.2d 820 (2d Cir.1985) (approving dismissal of juror who required four-and-a-half-day delay to observe religious holiday), cert. denied, 476 U.S. 1162, 106 S.Ct. 2285, 90 L.Ed.2d 726 (1986).
In Gay, we approved replacement of a juror who failed to appear after a two-month [1486]*1486recess because she was confused about the date trial was scheduled to resume. Although the juror lived only three hours from the courthouse, we held the complexity of the case — which “involv[ed] numerous parties, voluminous evidence, and lengthy trial proceedings” — as well as various rescheduling conflicts, justified the district court’s decision to replace the juror. 967 F.2d at 32S-24.4
In contrast, we held in United States v. Tabacca, 924 F.2d 906 (9th Cir.1991), that the district court erred in replacing a juror who claimed he was unable to attend jury deliberations because his wife had taken the car keys and left him no means of reaching the courthouse. In so holding, we specifically distinguished Wilson, Armijo, and Strat-ton, on the ground that, unlike those eases, in Tabacca the absent juror was certain to be available the very next day. Id. at 915. Here, the potential delay was uncertain. Moreover, the trial in Tabacca lasted only two-and-a-half days, with the government calling only three witnesses. Id. Here, as in Gay, the trial was much lengthier and more complex.5
We conclude the district court did not err in replacing juror Ragsdale.
II. Admission of FDIC Certificate of Insurance
Under 18 U.S.C. § 2113(f), the government must prove beyond a reasonable doubt that the victim bank was insured by the Federal Deposit Insurance Corporation (FDIC) on the date of the robbery. At trial, the government offered a “Certificate of Proof of Insured Status” as evidence of the federally insured status of the victim bank. This “Certificate of Proof’ consisted of a declaration by Patti C. Fox, Assistant Executive Secretary' of the FDIC, giving the history of the bank’s insured status and attesting that a diligent search of the official records of the FDIC revealed no entry indicating termination of the bank’s FDIC insurance on or before January 24, 1992 — the date of the robbery. The bank’s FDIC certificate of insurance, dated June 1, 1981, was appended to Fox’s declaration.
Over defense objection, the district court admitted both Fox’s declaration and the certificate of insurance as' hearsay statements admissible under Federal Rule of Evidence 803(24), the residual exception to the- hearsay rule.6 The defendants challenge the district court’s determination that these documents satisfy the requirements of Rule 803(24).
The same challenge the defendants make to the government’s use of Fox’s declaration and the bank’s certificate of insurance, to prove the bank’s federally insured status on the date of the crime, was squarely rejected by a recent decision of this court, albeit [1487]*1487for reasons different from the one on which the district court here relied. In United States v. Chapel, 41 F.Bd 1388 (9th Cir.1994), we held that a declaration virtually identical to the one offered by the government in this case was admissible pursuant to Federal Rule of Evidence 803(10).7 Id. at 1341. Moreover, the FDIC certificate of insurance appended to the declaration in Chapel was held to be admissible as «ore-hearsay. Id. at 1340 (citing United States v. Bellucci, 995 F.2d 157, 160-61 (9th Cir.1993), cert. denied, — U.S. -, 114 S.Ct. 2719, 129 L.Ed.2d 844 (1994)). As “a written statement[ ] which itself ‘affects the legal rights of the parties or is a circumstance bearing on conduct affecting their rights,’ [an FDIC certificate of insurance] falls outside the definition of hearsay.” Bellucci, 995 F.2d at 161 (quoting Fed.R.Evid. 801(c), advisory committee’s note, reprinted in 56 F.R.D. 183, 293 (1972)).
We held in Chapel that a certificate of insurance alone, if it antedates the crime, “does not adequately establish that the financial institution [was] FDIC insured on the date the charged offense occurred.” Chapel, 41 F.3d at 1340 (quoting United States v. Darrell, 828 F.2d 644, 648 (10th Cir.1987)) (emphasis omitted). However, we also held that an outdated certificate of insurance, when coupled with a declaration that no FDIC record indicates termination of the bank’s insured status prior to the date of the robbery, is “sufficient to establish the Bank was federally insured on the date of the robbery because it provefs] the Bank’s certificate of insurance remained valid even though it antedated the crime.” Id. at 1341.
Accordingly, we conclude the district court properly admitted both the certificate of insurance and Fox’s declaration — the former because it is not hearsay and the latter because it falls within Federal Rule of Evidence 803(10)’s “absence of a record” exception to the hearsay rule. Although these reasons are different from the reason given by the district court for admission of the challenged documents, we can affirm on any basis supported by the record. United States v. Click, 807 F.2d 847, 850 n. 5 (9th Cir.1987) (admission of evidence may be upheld even though district court’s decision was based on a different, even inappropriate, ground). The certificate of insurance and Fox’s declaration were both admissible and, taken together, are sufficient to establish that the bank was federally insured on the date it was robbed by the defendants.
III. Admission of Prior Convictions for Impeachment Purposes
Before trial, defendant Hicks filed a motion in limine to exclude evidence of his prior felony convictions for residential robbery and possession of rock cocaine for sale. The district court denied the motion, ruling that, if Hicks elected to testify, the evidence would be admissible for impeachment purposes, under Federal Rule of Evidence 609(a)(1), because its probative value outweighed its prejudicial effect.
At trial, Hicks chose to take the stand and present an alibi defense. He testified that he was in the vicinity of the arrests on the day of the robbery because he was scheduled to meet a friend there. He also said he ran away when he heard sirens and saw police cars because he was afraid of being arrested on two outstanding warrants for traffic violations. Purportedly for the same reason, he also gave the arresting officer a false name.
At the conclusion of his direct examination, Hicks again moved to exclude the evidence of his prior convictions. The district court adhered to its original ruling and denied the motion. On cross examination, the prosecution elicited testimony from Hicks regarding the nature and dates of both prior convictions.
[1488]*1488Hicks contends the district court erred in allowing the jury to hear evidence of his prior convictions. With regard to his prior robbery conviction, he argues United States v. Brackeen, 969 F.2d 827, 830 (9th Cir.1992) (en banc) (per curiam), stands for the proposition that a prior robbery conviction cannot be used to attack a defendant’s credibility. We disagree. Federal Rule of Evidence 609(a)' provides, in pertinent part, that evidence of prior convictions is admissible for purposes of attacking the credibility of a witness if the crime “(1) was [a felony], and the court determines the probative value of admitting this evidence outweighs its prejudicial effect to the defendant, or (2) involved dishonesty or false statement_” Brack-een held only that, in this circuit, bank robbery is not per se a crime of dishonesty, and therefore prior robbery convictions are not admissible for impeachment purposes under Rule 609(a)(2). Brackeen, 969 F.2d at 829. We did not foreclose in Brackeen the admission of a prior robbery conviction under the balancing test of Rule 609(a)(1).
Here, the government explicitly stated it intended to introduce both of Hicks’s prior convictions under Rule 609(a)(1), and the district court specifically ruled on that basis by applying the appropriate balancing test. If the district court did not abuse its discretion when it concluded Hicks’s prior convictions were more probative of his credibility than prejudicial to his defense, the evidence of both of his prior convictions was properly admitted. United States v. Browne, 829 F.2d 760, 762 (9th Cir.1987), cert. denied, 485 U.S. 991, 108 S.Ct. 1298, 99 L.Ed.2d 508 (1988) (noting that a district court’s decision to admit evidence of prior convictions is reviewed for an abuse of discretion).
In United States v. Cook, 608 F.2d 1175, 1185 n. 8 (9th Cir.1979) (en bane), cert. denied, 444 U.S. 1034, 100 S.Ct. 706, 62 L.Ed.2d 670 (1980), overruled on other grounds, Luce v. United States, 469 U.S. 38, 105 S.Ct. 460, 83 L.Ed.2d 443 (1984), we outlined five factors that should be considered in balancing the probative value of a prior conviction against its prejudicial impact for purposes of Rule 609(a)(1): (lj the impeachment value of the prior crime; (2) the point in time of conviction and the defendant’s subsequent history; (3) the similarity between the past crime and the charged crime; (4) the importance of the defendant’s testimony; and (5) the centrality of the defendant’s credibility. The government bears the burden of showing, based on these factors, that the proffered evidence’s probative value substantially outweighs its prejudicial effect. Browne, 829 F.2d at 763.
Hicks does not dispute that the first factor favors admission of both his prior convictions. We have previously stated that “prior convictions for robbery are probative of veracity.” United States v. Givens, 767 F.2d 574, 580 (9th Cir.), cert. denied, 474 U.S. 953, 106 S.Ct. 321, 88 L.Ed.2d 304 (1985). The same is true of prior convictions for drug offenses.
Hicks stipulates that both his prior crimes were sufficiently recent to satisfy the second Cook factor. He was convicted of residential robbery, and was sentenced to a four-year prison term, in 1987. Shortly after his parole in 1988, he committed the drug offense, for which he received another four-year prison sentence. Less than a year later, he was arrested for the present crime. “By its terms, Rule 609 allows for admissibility of such ... prior conviction[s] even where the defendant has been released for up to ten years.” Browne, 829 F.2d at 763. See Fed. R.Evid. 609(b).
Hicks concedes that, as to his prior drug offense, the third factor is satisfied because the drug offense is sufficiently different from the present bank robbery. With regard to the prior residential robbery, the district court held that offense was similar to the charged bank robbery and, therefore, the third factor weighed in favor of excluding it. However, we have held that even “a prior ‘bank robbery conviction [is] not inadmissa-ble per se, merely because the offense involved was identical to that for which [the defendant] was on trial.’ ” Browne, 829 F.2d at 763 (quoting United States v. Oaxaca, 569 F.2d 518, 527 (9th Cir.), cert. denied, 439 U.S. 926, 99 S.Ct. 310, 58 L.Ed.2d 319 (1978)). What matters is the balance of all five factors.
[1489]*1489Hicks contends that, contrary to the distriet court’s determination, the related fourth and fifth factors weigh against admission of his prior convictions. He contends his trial testimony was not particularly important and his credibility was not central to the case, because other evidence corroborated his alibi defense.8 We disagree. When a defendant takes the stand and denies having committed the charged offense, he places his credibility directly at issue. Browne, 829 F.2d at 764; Givens, 767 F.2d at 580; Oaxaca, 569 F.2d at 527.
In United States v. Bagley, 772 F.2d 482, 488 (9th Cir.1985), cert. denied, 475 U.S. 1023, 106 S.Ct. 1215, 89 L.Ed.2d 326 (1986), we held that admission of the defendant’s prior robbery convictions was an abuse of the district court’s discretion. But in that case we emphasized “the record [was] devoid of any evidence that [the defendant] intended to misrepresent his character or to testify-falsely as to his prior criminal record.” Id. Here, Hicks testified he ran from the police because he was afraid of being arrested on outstanding warrants for traffic violations. This testimony could reasonably have misled the jury into believing that, with the exception of some minor traffic infractions, Hicks had no previous trouble with the police. As we said in Cook:
[I]t is not surprising that the [district] court was unwilling to let a man with a substantial criminal history misrepresent himself to the jury, with the government forced to sit silently by, looking at a criminal record which, if made known, would give the jury a more comprehensive view of the trustworthiness of the defendant as a witness.
Cook, 608 F.2d at 1187.
We conclude that the district court properly balanced all five Cook factors and did not abuse its discretion in permitting the government to impeach Hicks with evidence of his prior drug and residential robbery convictions.
IV. Denial of Motion for Judgment of Ae-quittal
Defendant Alexander argues the district court should have granted his motion for judgment of acquittal because the evidence adduced at trial was insufficient as a matter of law to prove he was the same “Gary Alexander” who robbed the bank. We reject this argument.
Defendant Alexander was never directly identified by any witness in court.9 However, during trial, Deputy Sheriff Carol Best— who had followed the perpetrators from the site of the robbery and was present when the Camaro stalled and its occupants were arrested — responded affirmatively to the prosecutor’s question whether the men arrested were “Defendants Harrington and Alexander.” FBI Special Agent Robert M. Cross testified that he saw Alexander at the arrest scene while Alexander was sitting in the back of a Victorville police car, and saw him again after his incarceration. Cross also testified that, at the police station, he personally interviewed Alexander, at which time Alexander signed a waiver of his Miranda rights and made a full confession.
Alexander argues Deputy Best’s identification was inadequate because, on cross-examination, she admitted that at an earlier pretrial hearing she had been unable to identify him in court with certainty. Moreover, he contends her testimony wás ambiguous be-, cause it was presented before the defendants at the counsel table had been identified, and thus the jury could not have known which of the men seated there was Alexander. Alexander also argues Agent Cross’ testimony does not link him to the bank robbery because Cross never explicitly stated that the man he observed on the day of the robbery and later interrogated was the same man present in court during trial.
In reviewing a district court’s denial of a motion for judgment' of acquittal based [1490]*1490on insufficient evidence of identification, we apply the same test that the district court must employ in deciding whether to grant the motion in the first instance. United States v. Lucas, 963 F.2d 243, 247 (9th Cir. 1992); United States v. Kaplan, 554 F.2d 958, 963 (9th Cir.1977), cert. denied sub nom. Dolwig v. United States, 434 U.S. 1026, 98 S.Ct. 755, 54 L.Ed.2d 774 (1978); United States v. Leal, 509 F.2d 122, 125 (9th Cir. 1975). A motion for judgment of acquittal should be granted only if, viewing the evidence in the light most favorable to the government, no rational trier of fact could find beyond a reasonable doubt that the defendant is the person who committed the charged crime. Lucas, 963 F.2d at 247. “The government is entitled to all reasonable inferences that can be, drawn from the evidence.” Id.
Identification of the defendant as the person who committed the charged crime is always an essential element which the government must establish beyond a reasonable doubt. United States v. Telfaire, 469 F.2d 552, 555, 559 (D.C.Cir.1972); United States v. Fenster, 449 F.Supp. 435, 439 (E.D.Mich. 1978). However, in-court identification by a witness is not necessarily required. United States v. Fern, 696 F.2d 1269, 1276 (11th Cir.1983). “Identification can be inferred from all the facts and circumstances that are in evidence.” United States v. Weed, 689 F.2d 752, 754 (7th Cir.1982). “[A] witness need not physically point out a defendant so long as the evidence is sufficient to permit the inference that the person on trial was the person who committed the crime.” United States v. Darrell, 629 F.2d 1089, 1091 (5th Cir.1980).
For example, in-court identification is not necessary when the defendant’s attorney himself identifies his client at trial. United States v. Masters, 730 F.Supp. 686, 690 (W.D.N.C.1990). See also United States v. Hoelscher, 764 F.2d 491, 496 (8th Cir.1985) (defense attorney’s references to defendant as “Hoelscher” sufficed, along with a photograph and videotape of the defendant, as evidence of identification). Moreover, “the failure of any ... witnesses to point out that the wrong man had been brought to trial [can be] eloquent and sufficient proof of identity.” United States v. Weed, 689 F.2d at 755.
In the present case, Alexander’s attorney introduced him as “Gary Alexander,” Detective Owen Olson, who took possession of Alexander’s clothing at the Victorville police station, gestured toward the defense table during his testimony when he referred to “Gary Alexander.” Both Deputy Best and Special Agent Cross testified regarding their contacts with the Gary Alexander who robbed the bank, and neither suggested that the Gary Alexander on trial was not that same man. In fact, Deputy Best affirmatively stated the man she arrested for the bank robbery was “the Defendant Gary Alexander.”
We reject the contention that Deputy Best’s identification was ambiguous. The jury could reasonably have inferred that the man to whom Best referred was the Gary Alexander who was on trial. And, any confusion regarding the precise identity of “Defendant Alexander” was cleared up when his codefendants identified themselves.10
Deputy Best’s inability to identify Alexander at the pretrial hearing cuts against her trial testimony. But the identifying witness need not be free from doubt. United States v. Wilford, 493 F.2d 730, 732 n. 4 (3d Cir.), cert. denied, 419 U.S. 851, 95 S.Ct. 91, 42 L.Ed.2d 81 (1974). It was the function of the jury to consider Deputy [1491]*1491Best’s prior inability to identify Alexander and to discount her testimony to the degree it deemed appropriate. See United States v. Goode, 814 F.2d 1353, 1355 (9th Cir.1987).
That the government did not reopen its case so that Agent Cross could clarify his testimony linking Alexander to the bank robbery after Alexander moved for a judgment of acquittal is perhaps strategically surprising,' but it is not fatal. See Fenster, 449 F.Supp. 435 at 439,442 (denying a motion for judgment of acquittal despite the government’s surprising failure to cure irregularities in its identification evidence by reopening its case, but providing the government an opportunity to reconsider).
Viewing the evidence in the light most favorable to the verdict, and crediting the government with all the inferences that can be drawn from the evidence, we conclude that a rational trier of fact could find that defendant Alexander is the same man who, along with the other defendants, robbed the bank.11
V. Constitutionality of Sentence Enhancement for Robbery of a Financial Institution
USSG § 2B3.1(b)(l) requires that, if the property taken in a robbery belonged to a financial institution or post office, the convicted perpetrator’s base offense level be increased two levels. Because the victim of the robbery in this case was a bank, the district court enhanced the defendants’ sentences in accordance with this provision. The defendants challenge the constitutionality of § 2B3.1(b)(l) on substantive due process grounds, arguing that it arbitrarily and without legitimate reason punishes robbery from financial institutions and post offices more harshly than other robberies.
We review dé novo the constitutionality of a statute. United States v. Harding, 971 F.2d 410, 412 (9th Cir.1992), cert. denied, — U.S. -, 113 S.Ct. 1025, 122 L.Ed.2d 170 (1993). Because § 2B3.1(b)(l) implicates neither a fundamental right nor a suspect classification, we need only identify a rational basis for its specialized treatment of robberies involving financial institutions and post offices. United States v. Klein, 860 F.2d 1489, 1500-01 (9th Cir.1988). If a statute is not arbitrary, but implements a rational means of achieving a legitimate governmental end, it satisfies due process. Id. See also Nebbia v. New York, 291 U.S. 502, 537, 54 S.Ct. 505, 516, 78 L.Ed. 940 (1934).
The defendants bear the burden of proving the absence of a rational relationship betwéen § 2B3.1(b)(l) and a legitimate governmental objective. Harding, 971 F.2d at 413. “To challenge the Sentencing Guidelines successfully, the defendant^] ... must be able to demonstrate that the relevance of the distinction between [robberies of financial institutions and other robberies] is not even debatable.” Id.
The background material to § 2B3.1(b)(l) indicates the Sentencing Commission sought to punish robberies of financial institutions and post offices more severely because these entities typically keep large amounts of readily available cash, and are therefore particularly attractive robbery targets. See United States v. Smith, 905 F.2d 1296, 1301 (9th Cir.1990) (recognizing this motivation for enactment of the provision). The defendants argue this means-ends relationship is specious because other commercial establishments, besides financial institutions, and post offices, may also have large quantities of cash on hand. This may well be true, but the defendants bear the burden of demonstrating that the guidelines provision is irrational, not just imperfect. They have failed to do so. See United States v. McNeely, 20 F.3d 886, 888 (8th Cir.1994) (rejecting same argument). [1492]*1492VI. “Duplicative” Sentence Enhancements
The defendants challenge the district court’s imposition of various sentence enhancements as unconstitutionally duplicative, punishing them twice for the same conduct. They contend enhancements for robbery of a financial institution, under USSG § 2B3.1(b)(l), and for robbery resulting in a monetary loss exceeding $250,000, under USSG § 2B3.1(b)(6)(D), were both applied because the defendants stole a large amount of money.12 They further argue upward adjustments for both discharge of a firearm during commission of a robbery, under USSG § 2B3.1(b)(2)(A),. and assault on a law enforcement officer during the course of an offense or immediate flight therefrom, under USSG § 3A1.2(b), were imposed as a result of the shots the defendants fired at pursuing police officers.13 Similarly, they argue the shots fired at the police during the chase resulted in enhancement not only for assaulting an official victim, under § 3A1.2(b), but also for reckless endangerment of others during flight, under § 3C1.2.14
We review de novo the legality of a sentence, United States v. Guzman-Bruno, 27 F.3d 420, 422 (9th Cir.), cert. denied, — U.S. -, 115 S.Ct. 451, 130 L.Ed.2d 360 (1994), and the district court’s interpretation of the guidelines. United States v. Buenrostro-Torres, 24 F.3d 1173, 1174 (9th Cir.1994). We review for clear error a district court’s factual findings during the sentencing phase. Id.
Impermissible double counting occurs only when “one part of the Guidelines is applied to increase a defendant’s' punishment on account of a kind of. harm that has already been fully accounted for by application of another part of the Guidelines.” United States v. Reese, 2 F.3d 870, 895 (9th Cir. 1993), cert. denied, — U.S. -, 114 S.Ct. 928, 127 L.Ed.2d 220 (1994).
Each of the sentence enhancements applied in this case punished separate harms caused by distinct conduct on the part of the defendants.
A. Robbery of a Financial Institution and Monetary Loss Exceeding $250,000
USSG § 2B3.1(b)(l) authorizes a 2-level sentence enhancement if the offense involved robbery or attempted robbery of a financial institution or post office. Under USSG § 2B3.1(b)(6)(D), another three-level base offense increase is imposed for robbery resulting in a monetary loss exceeding- $250,-000. Contrary to the defendants’ contention, these two enhancements do not amount to double counting of a single act — namely, the theft of a large amount of money.
While the monetary loss provision increases the punishment to reflect the amount of money taken, the financial institution enhancement concerns only the type of institution robbed. As we said earlier, robberies of financial institutions and post offices are treated more severely than ordinary robberies because these institutions are likely to have large amounts of readily available cash and, consequently, are traditionally the prime targets of robbery. This does not mean, however, that the financial institution enhancement is geared toward punishing the theft of large quantities of cash. To the contrary, the § 2B3.1(b)(l) adjustment applies regardless of the amount of money actually stolen. See United States v. Smith, 905 F.2d 1296, 1301 (9th Cir.1990) (explaining that “[i]t is the nature of the institution and not the amount stolen that is legally operative in th[is] Guideline provision” because the dollar amount actually taken is largely “fortuitous”). In the present case, the institution robbed was a bank, and the amount stolen was $331,951. The defendants’ enhanced sentences properly reflected both facts.
[1493]*1493B. Discharge of a Firearm and Assault on an Official Victim
If a firearm is discharged during the commission of a robbery, USSG § 2B3.1(b)(2)(A) provides for an offense increase of seven levels. Additionally, USSG § 3A1.2(b) authorizes a 3-level increase if, during the course of the offense or immediate flight, the defendant assaults an officer in a manner creating substantial risk of injury. The defendants argue the district court, in-applying both provisions, doubly counted the shots they fired at police officers during their attempted getaway. We disagree.
The § 2B3'.l(b)(2)(A) enhancement was imposed as a result of the firearm discharged during the course of the robbery, while the defendants were still inside the bank.15 Punishment for the shots fired by the defendants while fleeing from the scene of the crime was. imposed by virtue of the official victim enhancement, § 3A1.2(b). Moreover, the official victim enhancement reflected not only the shooting, but also the fact that the defendants in the Camaro almost ran over a mo-torcyelé officer during their attempted getaway. This is obviously different conduct, causing a separate harm, from the shot fired inside the bank.
C. Assault on an Official Victim and Reckless Endangerment
USSG § 3C1.2 authorizes a 2-level increase in the base offense level “[i]f the defendant recklessly created a substantial risk of death or serious bodily injury to another person in the course of fleeing from a law enforcement officer.” Comment 1 to this section states that the enhancement may not be applied “where the offense guideline in Chapter Two, or another adjustment in Chapter Three, results in an equivalent or greater increase in offense level solely on the basis of the same conduct.” The defendants • contend that, because they were already punished for shooting at police officers under § 3A1.2(b), enhancement of their sentences under § 3C1.2 constituted double counting.
The Sentencing Commission’s comment 1 to § 3C1.2 requires that “[i]f both § 3A1.2(b) and § 3C1.2 apply to a defendant, the court must apply only the former and increase the offense level by three levels.” United States v. Sloley, 19 F.3d 149, 154 (4th Cir.), cert. denied, — U.S. -, 114 S.Ct. 2757, 129 L.Ed.2d 873 (1994). But this limitation exists, by the express terms of the comment, only when the same conduct makes both sections applicable. See United States v. Cabral-Castillo, 35 F.3d 182 (5th Cir.1994) (reversing the district court’s decision to enhance the defendant’s sentence under § 3C1.2 because he drove his vehicle at a high rate of speed, and under § 2Dl.l(b)(l) because he used a dangerous weapon, where the dangerous weapon was the car). In the present ease, the reckless endangerment provision applied not only because shots were fired during the attempted getaway, but also because of the risk of injury to civilians caused by the high-speed chase and by the defendants’ serious violations of other traffic laws.16 There was no double counting.
VII. Credit for Acceptance of Responsibili- ■ ty
The district court denied defendant Harrington’s request for a 2-point reduction in his offense level for acceptance of responsibility. Harrington appeals this decision, arguing that he confessed to the crime immediately after his arrest, and that he neither denied his guilt at trial nor presented a defense of any sort.
USSG § 3E1.1 authorizes a 2-point reduction in the base offense level “[i]f the defendant clearly demonstrates acceptance of responsibility for his offense.” By its plain language, § 3E1.1 places on the defendant the burden of “clearly” showing he is deserving of the reduction. We review for clear error a district court’s determination that a defendant is not entitled to a sentence reduction for acceptance of responsibility. United States v. Cox, 7 F.3d 1458, 1462 (9th Cir. [1494]*14941993); United States v. Molina, 934 F.2d 1440, 1450 (9th Cir.1991).
Ordinarily, a 2-point reduction for acceptance of responsibility will not apply to “a defendant who puts the government to its burden of proof at trial.” . USSG § 3E1.1, comment 2. However, a defendant need not plead guilty in order to qualify for a sentence reduction based on acceptance of responsibility.17 “In rare situations a defendant may clearly demonstrate an acceptance of responsibility for his criminal conduct even though he exercises his constitutional right to a trial.” Id. (emphasis added).
Comment 2 to § 3E1.1 indicates that such a “rare” situation may occur “where a defendant goes to trial to assert and preserve issues that do not relate to factual guilt (e,g., to make a constitutional challenge to a statute or a challenge to the applicability of a statute to his conduct).” Harrington contends he proceeded to trial only to challenge the severity of the guidelines which were applicable to him, not to assert his innocence. The record does not support this contention.
Harrington “declined to discuss the offense with the probation officer [and] indicated [the] statements] [he] made to officers were taken under duress.” Presentence Report at ¶ 75. These are not the actions of a person who acknowledges his complicity in a crime and accepts responsibility. Moreover, as the district court noted, a challenge to the guidelines could easily have been raised and preserved in ways other than going to trial. For example, Harrington could have pleaded guilty and then challenged the sentence imposed.
Harrington has not clearly demonstrated he was worthy of a reduction for acceptance of responsibility, and the district court did not clearly err in denying it. This is “the ordinary case [where] a defendant who pleads not guilty and is convicted [has] a difficult time convincing the court that his subsequent acceptance of responsibility for his actions is anything but a self-serving and insincere expression.” McKinney, 15 F.3d at 853. Compare United States v. Davis, 36 F.3d 1424 (9th Cir.1994) (affirming a denial of a reduction for acceptance of responsibility after finding “no affirmative evidence of contrition in the record” even though the defendant did not contest that he committed the offense, and went to trial only to present an entrapment defense), with McKinney, 15 F.3d at 852-55 (finding clear error in failing to grant a reduction for acceptance of responsibility where a defendant confessed, attempted to plead guilty before trial, exhibited confusion about the status of his plea when his attempt was rebuffed, and did not mount a substantive defense at trial).
VIII. Sentence Enhancement Based on Pri- or Drug Conviction
The district- court enhanced defendant Hicks’s sentence because of his 1988 conviction for possession of rock cocaine for sale. In doing so, the court rejected Hicks’s argument that his guilty plea to that crime was neither knowing, intelligent nor voluntary. Hicks contends this ruling was erroneous, and his prior drug offense should not have figured into calculation of the sentence for his present crime.
We need not decide the.merits of Hicks’s claim. Under Custis v. United States, — U.S. -, 114 S.Ct. 1732, 128 L.Ed.2d 517 (1994), a defendant is not entitled to mount a collateral attack on a prior conviction during sentencing for a subsequent crime, unless he contends that his prior conviction was obtained in violation of his Sixth Amendment right to the assistance of counsel.18 Hicks has not alleged a violation of his right to counsel.
Hicks argues Custis decided only that the Armed Career Criminal Act does not permit collateral attacks on prior convictions during sentencing for subsequent crimes, and is [1495]*1495therefore not controlling. This is an incomplete reading of Custis. The Court in Custis also held there is no right under the Constitution to challenge a prior conviction, except on the ground that the conviction was obtained in violation of the right to counsel. Id. at-, 114 S.Ct. at 1737-38.
Custis left open the possibility that, although the Constitution does not require a court to permit a collateral challenge to a prior conviction at sentencing, except for a violation of the right to counsel, such a challenge may be authorized by statute. We have held, however, that the guidelines do not provide independent authority for collateral attacks on prior convictions. United States v. Fondren, 32 F.3d 429, 430 (9th Cir.1994); see also Burrows, 36 F.3d at 885; United States v. Myers, 41 F.3d 531, 535 (9th Cir.1994). Therefore, Hicks may not challenge the legality of his prior drug conviction during sentencing for his present crime, except for a violation of his right to counsel. He asserts no such violation.
Hicks argues, in the alternative, that Custis is inapplicable to him because it was decided after both the crime and sentencing in his case. We disagree. The government raises Custis to affirm rather than to reverse the district court’s decision to enhance Hicks’s sentence based on his prior drug conviction. “[WJhatever might be true in other situations, in this unusual case in which the Supreme Court effectively changed the law after completion of the district court proceedings, the government should be permitted to rely on the new ruling to uphold the district court’s judgment.” United States v. Thomas, 42 F.3d 823, 825 (3d Cir. 1994).
IX. Conclusion
For the foregoing reasons, the convictions and sentences of all defendants are AFFIRMED.