Bank President John Markert approved five nominee loans by the bank to friends and family of bank customer George Wintz. The proceeds were used to cover a $1.9 million overdraft in a Wintz checking account at the bank caused by the collapse of Wintz’s fraudulent check-kiting scheme. After the fraud was discovered, Markert and Wintz were charged with bank fraud, 18 U.S.C. § 1344; five counts of willful misapplication of bank funds by a bank officer, 18 U.S.C. § 656; and aiding and abetting and conspiring to commit those offenses. After a lengthy trial and three days of deliberations, the jury convicted Markert of willful misapplication but acquitted him of bank fraud. The jury convicted Wintz of bank fraud but acquitted him of aiding and abetting willful misapplication, and it acquitted Gregory Pederson, the commercial loan officer involved in the transactions, of all charges. At sentencing, the district court found that Markert’s offenses caused a loss equal to the total amount of the five nominee loans (approximately $1.9 million), resulting in a 16-level guidelines enhancement, and sentenced him to 42 months in prison. Markert appeals his conviction, arguing there was insufficient evidence and jury instruction error, and his sentence, challenging the court’s calculation of loss. We affirm the conviction but agree loss was erroneously calculated and remand for resentencing.
I. Background
We summarize the evidence at trial viewed in the light most favorable to the jury verdict. See United States v. Thomas, 422 F.3d 665, 667 (8th Cir.2005). In 2007, Markert was named President and CEO of Pinehurst Bank (“Pinehurst” or “the Bank”), a small bank in Saint Paul. Prior to joining Pinehurst, Markert was President of Northstar Bank, a community bank with branches in suburban Roseville and White Bear Lake. As Pinehurst’s President, Markert had unilateral authority to approve loans up to $250,000 to any one customer. Loans totaling between $250,000 and $500,000 required approval by a majority of a four-person Officer Loan Committee (“OLC”) dominated by Markert and officers he recruited from Northstar. Loans totaling over $500,000 required approval by the Bank’s Board of Directors.
*924Markert brought to Pinehurst his longtime friend and bank customer, George Wintz, the owner of trucking and warehouse entities named McCallum Transfer, Triangle Warehouse, and Cue Properties. Wintz opened two business checking accounts at Pinehurst, one for McCallum Transfer and one for Cue Properties; he left Triangle Warehouse’s account with Northstar. Markert approved a series of loans to Wintz, quickly reaching the $250,000 limit of his unilateral lending authority. Wintz soon reached the $500,000 limit of the OLC and sought additional loans. Markert personally vouched for Wintz, informing the Board that he had been Wintz’s banker for twenty-one years. The Board approved additional loans. By February 2009, Wintz reached the Bank’s legal lending limit of nearly $1.2 million. JoAnn Crowley, who had been Pinehurst’s Chief Financial Officer since the Bank opened in 2004, repeatedly warned Markert that Wintz may be “kiting checks,”1 using a system that allowed him to make deposits without coming to the bank. These concerns went unheeded. Indeed, Markert fired Crowley.
In early March, Northstar employees discovered that Wintz was kiting checks between his accounts at the two banks. On Thursday, March 5, Northstar returned to Pinehurst fifteen checks totaling nearly $1.9 million due to insufficient funds. Wintz had drawn these checks on his Triangle Warehouse account at North-star and deposited them into his McCallum Transfer account at Pinehurst, which had credited the deposits. With the deposit checks dishonored by Northstar, Pinehurst faced a nearly $1.9 million loss on Wintz’s overdrawn McCallum Transfer account. On Saturday, March 7, Markert and Wintz met with two of Wintz’s friends, both of whom had previously lent Wintz money, to ask for additional loans. Without disclosing the check-kiting, Markert advised that Wintz’s account at Pinehurst was overdrawn, and he was over the Bank’s lending limit. Markert said the Bank could fail, ending Markert’s banking career, unless Wintz came up with money to cover the overdraft. The friends expressed sympathy but refused to lend Wintz more money.
What Markert did next formed the basis for his misapplication conviction. By Monday, March 9, Wintz had persuaded five friends and family members to sign documents obligating them to repay the following loans by Pinehurst Bank:
• William McDonald, Wintz’s friend and former banker at Northstar, borrowed $200,000. McDonald was led to believe that Wintz needed a short-term loan to make payroll. Wintz’s overdraft and check-kiting were not disclosed.
• Julianne Lenertz, Wintz’s former girlfriend, borrowed $350,000, ostensibly to pay taxes owing from her sale of a business to Wintz in 2007 that Wintz had agreed to pay. Lenertz testified she knew nothing about the terms of the loan.
• Lance Edlin, Lenertz’s brother, borrowed $350,000, believing he was cosigning a loan to help pay his sister’s tax debt. He knew nothing of Wintz’s overdraft or check-kiting.
• Nancy Cook, Wintz’s long-time secretary, borrowed $500,000 in the name of Triangle Logistics, a company that Wintz “gave” to Cook in 2004. Cook *925testified that she agreed to borrow money but did not know why Wintz needed it. She did not know the amount of the loan or the interest rate before arriving at the Bank on March 9.
• Debra Strom, Wintz’s daughter, borrowed $500,000 in the name of Win Properties, another Wintz-owned company. Wintz made her President of Win Properties so she could sign loan documents.
Although the nominal borrowers signed loan documents obligating them to repay, each understood that Wintz was the real borrower and would be responsible for the principal and interest payments. Markert with the assistance of his allies on the OLC prepared and closed the five nominee loans on Monday, March 9. Disguised as investments in Cue Properties, the loan proceeds were first funneled into Wintz’s Cue Properties account, then immediately re-directed and credited to his McCallum Transfer account. After discussions with Markert on March 9, the Bank’s new CFO did not post the checks returned by North-star that would have caused a $1.9 million overdraft until March 10. No overdraft was recorded because by then the loan proceeds had infused the account with sufficient funds.
That week, the Board of Directors met for its monthly meeting. The five nominee loans were included in the monthly information packet, but nothing linked them to Wintz. Markert attended the meeting but did not tell the Board about Wintz’s near-overdraft or his check-kiting activities. There was also evidence that Markert and his OLC allies backdated the loan documents and concealed the nominee loans from the member of the OLC who was not part of Markert’s inner circle. In January 2010, during a routine audit of bank files, an auditor uncovered the true purpose of the five nominee loans. Markert was immediately terminated. Bank regulators reviewed the Bank’s loans to Wintz and determined it must book an additional $2.2 million of reserves for the increased loss exposure, which rendered the Bank insolvent. Three months later, regulators closed Pinehurst Bank.
II. Discussion
Markert was convicted of being an officer of a federally insured bank who “willfully misapplie[d] any of the moneys, funds or credits of such bank.” 18 U.S.C. § 656. Conviction of a bank officer under this statute requires proof that he “wil[l]fully misapplied funds for the benefit of himself or another person, for the purpose of defrauding or injuring the bank.” United States v. Barket, 530 F.2d 181, 186 (8th Cir.1975), cert. denied, 429 U.S. 917, 97 S.Ct. 308, 50 L.Ed.2d 282 (1976). On appeal, Markert argues that he was wrongly convicted because five nominee loans he caused the Bank to approve were simply “a series of intrabank transfers” in which (i) the Bank never lost control of the loan proceeds, and (ii) the effect of the transactions was to benefit, not to harm the Bank financially. Based on this interpretation of the transactions, Markert argues that the evidence was insufficient, and the district court’s instructions erroneous, on the essential elements of willful misapplication and intent to defraud. We address these two challenges in turn, reviewing questions of statutory interpretation de novo. See United States v. Reed, 668 F.3d 978, 982 (8th Cir.2012).
A. Willful Misapplication. It has been a crime to embezzle or “willfully misapply” the funds or credits of a federally-chartered or regulated bank since 1864. See Act of June 3, 1864, § 55, 13 Stat. 116. Because willful misapplication, unlike embezzlement, had no settled technical meaning at common law, the Supreme Court prescribed an important limitation on this *926potentially elastic term in United States v. Britton, 107 U.S. 655, 666-67, 2 S.Ct. 512, 27 L.Ed. 520 (1883):
We think the -willful misapplication made an offense by this statute means a misapplication for the use, benefit, or gain of the party charged, or of some company or person other than the [banking] association. Therefore, to constitute the offense of willful misapplication, there must be a conversion to his own use or the use of some one else of the moneys and funds of the association by the party charged.
Applying this standard, the Court in Brit-ton dismissed certain counts of the indictment because they “charge[d] maladministration of the affairs of the bank, rather than criminal misapplication of its funds.” Id.
True to Britton, many decisions of this court and our sister circuits have observed that “the gist of the offense of willful misapplication is the conversion of funds of a federally insured bank by [a bank officer] either to his own use or to the use of a third person, with the intent to injure or defraud the bank,” without limiting the offense to any common law or statutory definition of “conversion.” Barket, 530 F.2d at 186-87; see Dow v. United States, 82 F. 904, 906 (8th Cir.1897) (“conversion in some form”); Johnson v. United States, 95 F.2d 813, 817 (4th Cir.1938) (“conversion [of deposited loan proceeds] in some form”); United States v. Beran, 546 F.2d 1316, 1320 (8th Cir.1976) (“Conversion of bank funds ... is encompassed within the definition of criminal misapplication.”), cert. denied, 430 U.S. 916, 97 S.Ct. 1330, 51 L.Ed.2d 595 (1977). As we summarized this principle in United States v. Mohr, “ ‘Misapplication’ under 18 U.S.C. § 656 covers unauthorized and improper conduct, typically conversion of bank funds for personal use by a bank officer or third party.” 728 F.2d 1132, 1134 (8th Cir.) (emphasis added), cert. denied, 469 U.S. 843, 105 S.Ct. 148, 83 L.Ed.2d 87 (1984). The district court quoted this passage in instructing the jury on the five counts of willful misapplication.
1. Sufficiency of the Evidence. Overstating the extent to which prior cases have limited willful misapplication to the customary legal definition of “conversion,” Markert argues the trial evidence was insufficient because the government failed to prove “that the defendant at least temporarily deprive[d] the bank of the possession, control, or use of its funds.” United States v. Duncan, 598 F.2d 839, 858 (4th Cir.), cert. denied, 444 U.S. 871, 100 S.Ct. 148, 62 L.Ed.2d 96 (1979). The Bank never lost control of its funds, he asserts, because the five nominee loans were structured so that the nominal borrowers never controlled the loan proceeds, and “[n]ot a penny of the proceeds” was withdrawn from the Bank or depleted its assets. “This was a series of bookkeeping entries and nothing more,” Markert concludes, a classic example of maladministration of the Bank’s affairs, rather than criminal misapplication of its funds. We disagree.
We assume, and the government more or less concedes, the legal premise that willful misapplication required proof that the Bank lost some degree of control over the funds or credits alleged to be misapplied, whether or not the nominee loan proceeds were, in a technical sense, converted. But that being so, Markert’s argument is, on this trial record, factually unsound. His focus on the nominee borrowers’ lack of control is simply irrelevant. As the Supreme Court explained in Brit-ton, willful misapplication occurs when a bank officer converts the bank’s funds (in some manner) to his use, “or the use of some one else,” with intent to defraud the bank. Here, as in other cases where nominee loans were used to camouflage fraudu*927lent transactions, the “some one else” was the real borrower, check kiter Wintz,- not the nominee borrowers. The loan proceeds were unquestionably funds of the Bank. So the relevant question is, did Wintz have temporary control over those proceeds when they were credited to his McCallum Transfer account? No, argues Markert, because no money ever left the Bank when the proceeds were used to avoid the impending overdraft in that account.
This contention defies common sense and the realities of commercial banking. When a bank makes a loan and deposits the proceeds in a customer’s checking account, the customer acquires use and control of the funds, subject to the terms and conditions of the account relationship. Here, Wintz gained control of the nominee loan proceeds deposited into his account, but his pre-existing contractual relations gave the Bank immediate right to these funds. The Bank had paid third parties to honor checks previously written on the McCallum Transfer account, leaving that account with insufficient funds to cover the checks returned by Northstar. Under these circumstances, the Bank was contractually entitled to use the nominee loan proceeds to repay itself. Wintz, of course, immediately benefitted by avoiding a large overdraft that would have severely damaged his business and financial interests.
When the Bank paid third parties funds that the loan proceeds were used to repay is irrelevant to the willful misapplication analysis. For example, in United States v. Mullins, where the defendant bank officer used cashier’s checks, rather than new loans, to conceal a check-kiting scheme, the court concluded that the government proved willful misapplication of the amounts of the checks, finding “no merit” in the contention “that the funds were lost long before the issuance ... of these cashier’s checks.” 355 F.2d 883, 886 (7th Cir;), cert. denied, 384 U.S. 942, 86 S.Ct. 1465, 16 L.Ed.2d 540 (1966).2 Likewise, Markert' willfully misapplied loan proceeds for the control and benefit of customer Wintz. Compare United States v. Ness, 665 F.2d 248, 249, 251 (8th Cir.1981). As we noted over a century ago, “it is not always necessary that money should be actually withdrawn from a bank, to constitute a criminal misapplication of its funds.... ” Rieger v. United States, 107 F. 916, 930 (8th Cir.), cert. denied, 181 U.S. 617, 21 S.Ct. 923, 45 L.Ed. 1030 (1901); accord United States v. Rickert, 459 F.2d 352, 354 (5th Cir.1972) (“It is not necessary ... that cash actually leave the bank.”).
That the Bank suffered harm encompassed by the willful misapplication provision in § 656 is evident in other ways as well. By lending money to Wintz through nominees and concealing that fact from the Board of Directors, Markert caused the Bank to lend Wintz and his companies $2.2 million more than its federally regulated lending limit allowed. The government presented testimony that lending limits protect a bank’s financial safety by reducing the risk caused by committing too much capital to a single borrower. “Violation of state law or bank policy may indicate [willful] misapplication.” Mohr, 728 F.2d at 1134; accord United States v. Clark, 765 F.2d 297, 303 (2d Cir.1985) (nominee loans above the bank’s lending limits pose sufficient “risk of pecuniary loss” to constitute misapplication); United States v. Dougherty, 763 F.2d 970, 972-73 (8th Cir.1985) (using unapproved bankers’ acceptances to conceal overdrafts was will*928ful misapplication). Markert caused loan proceeds to be used for a purpose that the Board would not (and indeed could not) have approved. See United States v. Crabtree, 979 F.2d 1261, 1267 (7th Cir.1992), cert. denied, 510 U.S. 878, 114 S.Ct. 216, 126 L.Ed.2d 173 (1993); United States v. Woods, 877 F.2d 477, 480 (6th Cir.1989) (“the bank was effectively deprived,'without its consent, of control over its funds for an additional twenty-two months. That deprivation of control is sufficient to constitute misapplication of bank funds.”); United States v. Shively, 715 F.2d 260, 265-66 (7th Cir.1983) (willful misapplication “must require that the bank’s money have been used for a purpose that the bank would not have agreed to had it known what the purpose was”), cert. denied, 465 U.S. 1007, 104 S.Ct. 1001, 79 L.Ed.2d 233 (1984).
United States v. Michael, 456 F.Supp. 335, 342 (D.N.J.1978), on which Markert relies, is distinguishable because the court granted a motion to dismiss counts in an indictment charging willful misapplication, whereas here we have a full trial record. To the extent categorical statements in that opinion are contrary to our construction of the willful misapplication offense, we find Michael unpersuasive and decline to follow it. We have long held that withdrawal of funds from the victim bank is not a necessary element, so long as the requisite intent to defraud is proved.
2. A Misapplication Instruction Issue. Without objection, the district court instructed the jury that the five willful misapplication counts in the indictment required the government to prove that Markert “willfully misapplied the moneys or funds belonging to or intrusted to the custody or care of Pinehurst Bank.... with intent to defraud Pinehurst Bank.” The court further instructed:
The phrase “willfully misapplies” means the unauthorized, or unjustifiable or wrongful use of a bank’s funds. Misapplication includes the wrongful taking or use of money of the bank by a bank officer or employee for his own benefit or for the use and benefit of some other person.
“Misapplication of bank funds” covers unauthorized and improper conduct,3 typically conversion of bank funds for personal use by a bank officer or a third party.
Markert argues the court erred when it instructed that “willfully misapplies” means “the unauthorized or unjustifiable or wrongful use of a bank’s funds.” He posits that, by failing to instruct that willful misapplication requires proof that the Bank at least temporarily lost control or possession of its funds, the instruction included imprudent loans that do not involve the essential element of conversion, “thereby improperly criminalizing bad banking.” Markert acknowledges that our review is for plain error because he did not object to this instruction. See United States v. Gill, 513 F.3d 836, 852 (8th Cir.), cert. denied, 555 U.S. 1080, 129 S.Ct. 750, 172 L.Ed.2d 745 (2008).
We conclude that the district court’s willful misapplication instruction was not error, much less plain error, primarily for the reasons we reject Markert’s challenge to the sufficiency of the evidence on this issue. As we have explained, the evidence *929plainly showed that the Bank lost control of the nominee loan proceeds when they were deposited into Wintz’s McCallum Transfer checking account at the Bank. The court instructed the jury that, to convict Markert, it must find that he misapplied bank funds and did so with intent to defraud the Bank. At Markert’s request, the court further instructed that “[g]ood faith is a complete defense to each charge of ... misapplication of bank funds,” and that “evidence which establishes only that a person made a mistake in judgment or an error in management ... does not establish fraudulent intent.” When considered as a whole, these instructions defined the essential elements of the offense and sufficiently guarded against a conviction based on negligence, poor judgment, or “technical” violations of banking practices or regulatory requirements. The jury necessarily found, based on more than sufficient evidence, that Markert willfully and surreptitiously misapplied the nominee loan proceeds for the use and benefit of the real borrower, George Wintz, with the intent to defraud the Bank.
B. Intent to Defraud. Though no longer expressly stated in the statute, all courts agree that intent to defraud or injure the bank remains an essential element of the willful misapplication offense under 18 U.S.C. § 656. See, e.g., Barket, 530 F.2d at 186. Markert again argues insufficiency of the evidence and jury instruction error as to this element of the offense.
1. Sufficiency of the Evidence. Markert argues the evidence was insufficient to prove that he acted with the requisite intent to defraud or injure the Bank because intent to defraud requires proof of probable loss to the Bank, and the trial evidence demonstrated that the nominee loan proceeds benefitted the Bank because they (i) avoided a costly overdraft condition in Wintz’s McCallum Transfer account, and (ii) left the Bank at least some recourse against the five nominee borrowers. We disagree. First, the contention is legally unsound. As we have explained, knowing violation of rules designed to protect a bank’s financial safety and stability, such as legal lending limits, when done with the intent to deceive the bank (for example, to protect one’s employment or to conceal bank fraud by a customer), falls squarely within the offense of willful misapplication of bank funds. Second, the contention is factually unsound. Markert committed the Bank’s loan reserves to nominee loans the Board of Directors and bank regulators would not have approved, to avoid an overdraft condition resulting from the real borrower’s fraudulent check-kiting, a transaction that clearly created a probable risk of further loss to the Bank.
Intent to defraud “exists if a person acts knowingly and if the natural result of his conduct would be to ... defraud the bank even though this may not have been his motive.” Beran, 546 F.2d at 1321 (quotations omitted). That Markert intended or hoped to help Pinehurst avert an overdraft loss does not establish that he lacked an intent to defraud, though it may evidence lack of an intent to injure. Markert did much more than negligently approve imprudent nominee loans. He helped Wintz deceive the nominee borrowers, backdated loan documents, concealed the nominee loans from the fourth member of the OLC, and falsely structured the transactions as investments in Cue Properties to conceal the true purpose of the loans from both the putative borrowers and the Bank’s Board of Directors. All this is more than sufficient evidence for the jury to find the requisite intent to defraud. “The fact that [Markert] did not personally profit from his criminal conduct is not a legal excuse for his action.” Dougherty, 763 F.2d at 972.
*9302. The Intent to Defraud Instruction. Without objection, the district court defined the phrase “intent to defraud” to mean “to act knowingly with intent to deceive or cheat, for the purpose of causing a financial loss to someone else or bringing about a financial gain to the defendant or another.” On appeal, Markert argues the court plainly erred when it failed to instruct that, to find an intent to defraud, the jury must find that Markert’s actions subjected the bank to increased risk of pecuniary loss. We disagree. The court’s instruction was consistent with Eighth Circuit Model Criminal Jury Instruction 6.18.656 and with the instruction we approved in Dougherty, 763 F.2d at 973. In addition, the court instructed that “[g]ood faith is a complete defense to each charge of ... misapplication ... because it is inconsistent with the intent to defraud, which is an element of those charges.” There was no error, plain or otherwise.
C. The Nominee Loan Instruction. Count 2 of the indictment charged Markert and his co-defendants with bank fraud. As 18 U.S.C. § 1344(1) requires, Count 2 alleged “a scheme and artifice to defraud Pinehurst Bank” by issuing and concealing five loans to “straw borrowers ... commonly known as a ‘nominee loan’ scheme.” The district court addressed this allegation in Jury Instruction No. 20:
Count Two ... charges that the scheme to defraud ... was accomplished through nominee loans. Nominee loans are loans in which the nominal borrower was actually obtaining the money for a third party’s benefit. The mere transfer of a loan’s proceeds to a third party is not illegal. However, such a loan may be unlawful when the borrower and the bank officer fail to state the real borrower and recipient of the funds, thereby obtaining the loans by means of false pretenses when doing so with the intent to defraud the bank.
Markert joined Wintz’s timely objection and pursues the issue on appeal, so we review this instruction for abuse of discretion, recognizing the district court has wide latitude, so long as the entire charge, when considered as a whole, fairly and adequately contains the applicable law. See, e.g., United States v. Bevans, 496 F.2d 494, 499 (8th Cir.1974). Applying this standard, Instruction No. 20 was plainly a correct portion of the instructions relating to the bank fraud charges in Count 2. It carefully tracked our description of “nominee loans” in United States v. Willis, 997 F.2d 407, 410 n. 2 (8th Cir.1993), cert. denied, 510 U.S. 1050, 114 S.Ct. 704, 126 L.Ed.2d 670 (1994), a bank fraud case, with the addition — encouraged by defense counsel at the instruction conference — of the significant proviso, “when doing so with the intent to defraud the bank.”
On appeal, Markert of course shifts our focus away from Count 2, a charge of which he was acquitted. Rather, he now argues that this instruction was an abuse of the district court’s discretion in instructing the jury on the willful misapplication charges (Counts 3-7) because it failed to instruct that a defendant does not act with intent to defraud unless he knows that the nominal borrower lacks the financial ability to repay the nominee loan. In instructing on Counts 3-7, the court explained:
Counts Three through Seven charge that the misapplication of bank funds was accomplished by the approval and disbursement of bank funds for five nominee loans to borrowers of George Wintz. I have already defined “nominee loans” in Jury Instruction No. 20.
There was no objection to this instruction. Nor was the issue of the nominee borrowers’ ability to repay ever raised at the instruction conference. Therefore, our re*931view of the nominee loan instruction as it relates to Counts 3-7 is for plain error.
The district court’s definition of “nominee loans,” taken verbatim from our opinion in Willis, was unobjectionably simple — “Nominee loans are loans in which the nominal borrower was actually obtaining the money for a third party’s benefit.” What is more complex and can be debatable is defining when the use of nominee loans gives rise to criminal liability. Bank fraud and willful misapplication of bank funds are distinct offenses, but each has the essential element of intent to defraud. The district court properly included that element in Instruction No. 20 and then included intent to defraud in its further instructions on Counts 3-7. Thus, the nominee loan instruction in no way diluted the court’s instruction that the jury must find intent to defraud to convict Markert of willful misapplication.
We disagree with Markert’s assertion that a nominee borrower’s ability to repay always establishes the lack of intent to defraud. For example, when a bank officer uses nominee loans to conceal that he will use the proceeds for his own interests, “the financial condition of the putative borrower is irrelevant.” United States v. Steffen, 641 F.2d 591, 597 (8th Cir.), cert. denied, 452 U.S. 943, 101 S.Ct. 3091, 69 L.Ed.2d 959 (1981). The evidence in this case presented an analogous situation-use of nominee loans to conceal that the true borrower had exceeded his lending limits.4 We agree with Markert that the nominee borrower’s ability and intent to repay will be relevant in some cases. Compare United States v. Brennan, 994 F.2d 918, 923 (1st Cir.1993), with United States v. Parsons, 646 F.2d 1275, 1279-80 (8th Cir.1981). But in our view, the evidence in this case strongly suggested that the nominal borrowers’ ability to repay their loans was not essential to the question whether the scheme resulted in the willful misapplication of the loan proceeds. There was strong evidence the nominal borrowers were deceived about the purpose of the loans, did not know the loan terms, and some rather clearly lacked the ability to repay. As the district court’s instructions, considered as a whole, fairly and adequately presented the essential elements of the willful misapplication offense, we conclude there was no error, plain or otherwise, in the manner in which the court incorporated its nominee loan bank fraud instruction into the willful misapplication offenses charged in the indictment.
III. The Sentencing Issue
Prior to sentencing, the Revised Presentence Investigation Report recommended, without explanation, a finding that Markert’s willful misapplication offenses caused a loss equal to the total amount of the five nominee loans, nearly $1.9 million, resulting in a 16-level increase under the advisory Sentencing Guidelines. See U.S.S.G. § 2Bl.l(b)(l)(I). The government’s sentencing memorandum supported this recommendation on two grounds: (1) numerous cases have held that loss in check-kiting cases is calculated in this fashion, and Wintz’s check-kiting was relevant conduct in Markert’s offense; and (2) “the amount of loss in misapplication and fraudulent loan cases is the amount of funds misapplied,” citing United States v. Hulshof, 23 F.3d 1470 (8th Cir.1994). Markert’s sentencing memorandum argued, as he argues on appeal, that both the actual loss to the Bank resulting from the nominee loan transactions, and the loss he intended to inflict, were zero. As the Bank’s loss caused by Wintz’s completed check-kiting had already been incurred, he explained, the net *932loss — “the difference between what the victim paid and what the victim recovered” — was zero. If anything, he argued, the nominee loans improved the Bank’s financial position.
At sentencing, over Markert’s objection, the district court adopted the loss figure recommended in the PSR, explaining:
The misapplication of the bank funds through the nominee loan scheme was used to avoid detection or responsibility of an underlying check kite which totaled $1.8 million. The amount of funds misapplied is the amount of the loss in a misapplication case and here those nominee loans totaled $1.8 million.
This 16-level increase produced an advisory guidelines range of 87 to 108 months. The court varied downward and sentenced Markert to 42 months in prison. He appeals the fraud loss determination. We review the district court’s fact findings for clear error and its interpretation of the advisory guidelines de novo. See United States v. Holthaus, 486 F.3d 451, 454 (8th Cir.), cert. denied, 552 U.S. 939, 128 S.Ct. 343, 169 L.Ed.2d 241 (2007). Although the court granted a substantial downward variance, the 16-level increase, if it was procedural guidelines error, cannot be deemed harmless.
In defending the district court’s ruling that the amount of funds misapplied equals the amount of actual loss in a willful misapplication case, the government cites only one § 656 decision, Hulshof. That case is distinguishable because the misapplied funds at issue were not the proceeds of new loans by the victim bank. More importantly, Hulshof was decided before the Sentencing Commission adopted extensive revisions to the guidelines and commentary governing the calculation of loss for theft and fraud offenses, revisions intended to resolve several issues that had split the circuits. See U.S.S.G.App. C., Vol. II, Amend. 617, at 130-86. Under the revised guideline, “actual loss” is now defined as “the reasonably foreseeable pecuniary harm that resulted from the offense.” U.S.S.G. § 2B1.1, comment. (n.3(A)(i)). “Actual loss” under the revised guideline is a “net loss” concept based upon “the difference between what the victim paid and what the victim recovered plus any other forms of reasonably foreseeable pecuniary harm that resulted from the offense.” United States v. Hartstein, 500 F.3d 790, 798 n. 3 (8th Cir.2007), cert. denied, 552 U.S. 1102, 128 S.Ct. 939, 169 L.Ed.2d 734 (2008); § 2B1.1 comment. (n.3(E)); accord United States v. Smith, 951 F.2d 1164, 1167 (10th Cir.1991). As the Sentencing Commission explained:
The loss definition also provides for the exclusion from loss of certain economic benefits transferred to victims, to be measured at the time of detection [of the offense]. This provision codifies the “net loss” approach that has developed in the case law.... This crediting approach is adopted because the seriousness of the offense and the culpability of a defendant is better determined by using a net approach. This approach recognizes that the offender who transfers something of value to the victim! ] generally is committing a less serious offense than an offender who does not.
App. C. — Vol. II, Amend. 617, at 183; accord § 2B1.1 comment, (background).
The government’s actual loss contention as adopted by the district court ignored this change in the revised loss guideline by failing to acknowledge that the monetary value of the nominee loans the Bank received in exchange for the misapplied proceeds, measured at the time the misapplication offense was detected, must be credited against actual loss.5 The *933government has the burden to prove actual loss by the preponderance of the evidence. The trial record reflected considerable effort by the Bank to obtain repayment of the nominee loans, either by the nominal borrowers or by Wintz. The net value of those loans, measured at the time their nominal nature was detected, may be difficult to measure. “The court need only make a reasonable estimate of the loss.” § 2B1.1, comment, n. 3(C). But here no attempt was made to determine what value to credit in the actual loss calculation. A remand is required.
In a footnote to its appeal brief, the government alternatively contended that Markert’s actual loss should be the amount of the “float” when Wintz’s check-kiting scheme unraveled, a finding we upheld in an earlier case of check-kiting fraud, United States v. Whitehead, 176 F.3d 1030, 1042 (8th Cir.1999). We reject this contention. We did not adopt this as a categorical rule in Whitehead. More importantly, Markert was not charged with a check-kiting offense, and he was acquitted of the charge that he conspired with Wintz to commit bank fraud.
Under the revised guideline, the amount of loss continues to be the greater of actual or intended loss. See U.S.S.G. § 2B1.1, comment. (n.3(A)). Here, in response to Markert’s contention that the intended loss was zero because he had no subjective intent to harm the Bank, the government on appeal disclaimed any reliance on an intended loss theory. Thus, the issue turns on its flawed calculation of actual loss.
We affirm Markert’s conviction and remand for resentencing.
. Check kiting is a form of fraud designed to utilize the time, or "float,” needed for one bank to collect on checks drawn on accounts at other banks. By transferring funds between accounts at different banks using checks not supported by sufficient funds, the kiter fraudulently inflates the account balances, enabling him to write checks to third parties that the unsuspecting banks pay until the scheme is exposed.
. The dissent asserts that this opinion "does not cite to a single case that mirrors the facts we have before us.” In our view, Mullins is factually indistinguishable. No two complex fraud cases will ever be "mirror” images of each other.
. We have questioned the Eighth Circuit Model Jury Instruction that defines "misapplication” to include "unauthorized” conduct because simply using bank funds without authorization may lack the necessary mens rea. United States v. Robertson, 709 F.3d 741, 745 n. 3 (8th Cir.2013), interpreting 18 U.S.C. § 1163, a statute modeled on § 656. But that concern is not present here because the instructions clearly required the jury to find that Markert acted with intent to defraud.
. Moreover, although Markert did not receive the loan proceeds, there was evidence he viewed the transactions as necessary to preserve his employment.
. By contrast, a defendant's repayment of funds after the discovery of a fraud offense is *933not relevant to sentencing. See U.S.S.G. § 2B1.1 comment. (n.3(E)); United States v. Stennis-Williams, 557 F.3d 927, 930 (8th Cir.2009).