concurring in part and concurring in the judgment:
I agree fully with all but Part III.A of the majority’s opinion. If, as the majority holds in Part III.B, the government demonstrates that the account holders whose accounts were targeted by Pham’s scheme incurred actual collateral expenses while convincing the banks to restore their account balances and correct errors in their credit histories, these expenses constitute reasonably foreseeable pecuniary harm arising from Pham’s conduct. These individuals could therefore be counted as the “victims” of Pham’s conduct for the purpose of a sentencing enhancement under U.S. Sentencing Guideline § 2Bl.l(b)(2). I part ways with the majority, however, with regard to Part III.A, where the majority holds, in the alternative, that both the financial institution and the account holder could be counted as a “victim” *724based on the same actual theft of funds, which was temporarily reflected in the account holder’s account balance but ultimately sustained by the bank. Although I agree that the temporary deprivation of access to one’s account funds is some kind of harm, it is not a pecuniary harm, and thus it is not a kind of harm accounted for under the Guidelines.
As the majority explains, the Guidelines allow an individual to be counted as a “victim” for the purpose of a sentencing enhancement under § 2Bl.l(b)(2) only if he or she “sustained any part of the actual loss” from the fraudulent scheme. U.S.S.G. § 2B1.1 cmt. n. 1. This loss must be “pecuniary,” which means “harm that is monetary or that otherwise is measurable in money,” and “does not include emotional distress, harm to reputation, or other non-economic harm.” Id. n. 3(A)(iii). In Part III.A, the majority holds that the account holders could be shown to have suffered such “actual loss” because “their accounts were debited,” even where “the losses sustained by the account holders turned out to be temporary” because these funds were later restored by the bank. Maj. Op. at 717, 718. The majority concedes that some reimbursements may be so speedy that the victim could not be said to have actually suffered or sustained the loss, but holds that a reimbursement that is not “instantaneous” is an actual pecuniary loss under the Guidelines. Id. at 719. The majority thus creates a vague standard for defining “actual loss,” whereby reimbursed account holders are “victims” if “some appreciable amount of time elapsed before the accounts were replenished,” but not if they were “fully reimbursed as soon as they notified their banks of the fraudulent activity.” Id. at 719, 719-20.
As the majority acknowledges, the losses to the account holders in this case turned out to be ephemeral. Pham and his co-defendants defrauded banks by impersonating legitimate account holders and presenting counterfeit checks in the account holders’ names, inducing the banks to give money to Pham and his co-defendants while improperly lowering the account balance of the victimized account holder. By restoring the account holders’ balances, the bank agreed that the money had not been validly drawn from these accounts. The reimbursement reflected the bank’s acknowledgment that it had committed error by improperly accepting a counterfeit check and improperly debiting an individual’s account for funds that were, in fact, fraudulently stolen from the bank itself. See United States v. Erpenbeck, 532 F.3d 423, 442 (6th Cir.2008) (“When a customer of a bank or a creditcard company is defrauded, the customer is generally protected by an agreement that the bank or company will handle any fraud based upon unauthorized charges against the customer’s account.... [Thus] a loss under this type of contractual agreement is necessarily temporary, and the customers are fully reimbursed.”).
I agree that it was reasonably foreseeable that the bank would mistakenly debit the holders’ accounts and that — as a result — the holders would, for a time, lose access to their account funds while the bank sorted out who should be the proper party to bear the loss. Assuming the bank fully reimburses the account holders, however — including refunding any lost interest or other charges caused by the improper debiting of funds — I do not believe it can fairly be said that the account holders have suffered any actual pecuniary harm from the temporary deprivation of access to their funds. Rather, they have sustained an inconvenience and a headache, which I do not dispute can be of serious proportions. To the extent the account holders’ inconvenience can be monetized in the form of lost vacation days, phone calls and *725letters to the bank and credit reporting agencies, unreimbursed expenses for replacement checks and so on, then I agree that the account holders have sustained a reasonably foreseeable, actual pecuniary loss that is sufficient to count them as “victims” under the Guidelines. See United States v. Abiodun, 536 F.3d 162, 168-69 (2d Cir.2008) (holding that reimbursed account holders may be “victims” under § 2B1.1 based on “loss of time” that can be measured in monetary terms). This is what I understand to be the holding of Part III.B of the opinion, which I fully join. To label the temporary deprivation of access to funds alone an “actual loss,” however, goes beyond the Guidelines’ clear — and narrow-definition of what makes someone a “victim” under § 2Bl.l(b)(2).
Moreover, I am concerned that the majority’s definition of “actual loss” in Part III.A risks double-counting both the number of victims and the amount of loss caused by the defendant. “Impermissible double counting occurs where 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 the application of another part of the Guidelines.” United States v. Thornton, 511 F.3d 1221, 1227-28 (9th Cir. 2008) (internal quotation marks omitted). The majority’s definition of “actual loss” means that both the account holder and the bank could be counted as “victims” based on the same harm, the pecuniary loss from the theft of funds. The majority acknowledges that this case differs from cases such as Thornton and United States v. Holt, 510 F.3d 1007, 1011-12 (9th Cir. 2007), where we upheld sentencing enhancements against double-counting challenges because the defendant’s sentence was enhanced under different guidelines that captured different aspects of the defendant’s harm. Here the majority proposes that a single harm — the pecuniary loss from the funds — constitutes an “actual loss” to two different parties under the same Guideline, simply because some time elapsed before one party agreed to absorb the loss. The pecuniary loss from the stolen funds, however, is fully absorbed by the bank. By also counting the account holder as a victim, the majority would either be double-counting the victims of the pecuniary loss, or else accounting for some other, non-pecuniary harm to the account holder, which is not permitted by the Guidelines.
More troubling is what this might mean in future cases for the calculation of the loss amount for purposes of a sentencing enhancement under § 2Bl.l(b)(l). If both the bank and the account holder, as the majority suggests, sustained an “actual loss” as a result of the theft of funds, then it is not clear why the amount stolen should not also be doubled for purposes of calculating a sentencing enhancement for the loss amount. Thus, under the majority’s rationale, a defendant who defrauded a bank of $500,000 could seemingly be punished as if he had stolen $1,000,000 — netting a two-level enhancement — because he would have caused $500,000 in “actual loss” to the bank and another $500,000 in “actual loss” to the account holders. See U.S.S.G. § 2Bl.l(b)(l). The majority points out that such double-counting is not implicated in Pham’s case because the government stipulated to the loss amount. See Maj. Op. at 718-19. This will not, of course, necessarily be so in future cases. This leads to the troubling proposition that defendants who steal by identity theft could be exposed to significantly higher punishments because the “actual loss” amount would be doubled, at least in cases where the bank understandably takes some time to verify that fraud occurred before agreeing to absorb the loss.
*726No other circuit has held that two victims can both be counted under the Guidelines based on the same “actual loss” of funds. In United States v. Lee; 427 F.3d 881, 894 (11th Cir.2005), the Eleventh Circuit rejected a defendant’s argument that the losses he caused should not be counted under the Guidelines because the victims managed to recover their stolen property from, the defendant himself The Lee court properly rejected this argument, which would have rendered a crime “victimless” if the victim recovered his stolen property. See id. at 895. Here, however, as in United States v. Yogar, 404 F.3d 967 (6th Cir.2005), our task is not to decide whether there was any victim who sustained the loss, but to choose between two possible victims. In Yogar, the Sixth Circuit suggested in dicta that there “may be situations in which a person could be considered a ‘victim’ under the Guidelines even though he or she is ultimately reimbursed” by the bank for the lost funds. Id. at 971. The Sixth Circuit did not explain how or why such a person could be considered a victim, nor explain what actual losses would have to be shown to make such a person a victim. The Sixth Circuit may well have simply anticipated our reasoning in Part III.B, in which I concur, which holds that unreimbursed collateral expenses incurred by the account holders could be sufficient to render these individuals “victims” under the Guidelines. In any event, nothing in the Sixth Circuit’s reasoning compels the conclusion that delay prior to reimbursement is alone sufficient to render the account holder a “victim” under the Guidelines, absent proof that the delay itself caused independent pecuniary harm.
The majority’s holding thus puts this court into tension with other circuits that have interpreted Yogar to mean that reimbursed victims of identity theft can be counted as “victims” under § 2Bl.l(b)(2) only when the government shows actual pecuniary losses apart from the temporary losses that were reimbursed. In United States v. Abiodun, 536 F.3d at 169, the Second Circuit held that reimbursed victims could be counted as “victims” due to the lost time involved in resolving shortfalls in their accounts, but only to the extent that “this ‘loss of time’ [can] be measured in monetary terms” and the value of the lost time is reflected in the calculation of total loss resulting from the defendants’ offenses. Similarly, the Fifth Circuit rejected the suggestion that “account holders are victims because they suffered pecuniary harm at the moment” that stolen funds are charged against their accounts. See United States v. Conner, 537 F.3d 480, 490 (5th Cir.2008). The Fifth Circuit joined the Eighth Circuit in holding that “only the party that ‘ultimately ’ bore the pecuniary harm from the offense suffered an ‘actual loss’ ” under the Guidelines. Id. (quoting United States v. Icaza, 492 F.3d 967, 970 (8th Cir.2007)). Consistent with the Second Circuit and our reasoning in Part III.B, the Fifth Circuit agreed that “[i]t is possible that with a proper evidentiary foundation ... unreim-bursed business losses [incurred while seeking reimbursement] could be considered ‘actual losses’ for the purposes of counting ‘victims.’ ” Id. at 491 (emphasis added). No court, however, has endorsed the broad proposition that the majority advances here, which is that mere delay prior to reimbursement — which has not itself been shown to have caused a separate, actual pecuniary loss — is sufficient to make someone a “victim” under the Guidelines.
There is no question that identity theft schemes such as Pham’s take a heavy toll on individual account holders, who may spend days or weeks working with the bank to investigate the fraud and restore their account funds, and then spend considerable time dealing with the fallout *727from their devastated credit histories. If the government can prove that such individuals incurred real, unreimbursed expenses as a foreseeable result of Pham’s conduct — even if these pecuniary losses are minimal, but so long as they are monetized — then I agree that Pham’s sentence can be enhanced under the Guidelines for his impact on these victims. Further, the district court always may “consider the large number of individual account holders affected by [Pham’s] crime as part of its consideration of § 3553(a) factors if the court decided to issue a non-guidelines sentence.” See Conner, 537 F.3d at 492. District judges therefore have ample tools to punish defendants for the very real harms they inflict on those whose identities become the vehicles for financial fraud. There is no need for us to create difficult-to-apply rules that would make the number of victims turn on the length of time the bank spent investigating the fraud before reimbursing the account holder, and that risk punishing the defendant twice for causing a single pecuniary harm through his theft of funds. Therefore, I respectfully cannot join the majority’s reasoning in Part III.A.