JORDAN, Circuit Judge

Andrew S. Ittleman, Christopher J. Rajotte, Fuerst Ittleman David & Joseph, PL, Miami, FL, for Plaintiffs-Appellants

Paul Andrew Allulis, Bruce R. Ellisen, U.S. Department of Justice, Chief Appellate Section Tax Division, Washington, DC, Michael W. May, Robert L. Welsh, U.S. Department of Justice, Tax Division, Washington, DC, U.S. Attorney Service - Middle District of Florida, U.S. Attorney's Office, Tampa, FL, for Defendant-Appellee

Before JORDAN and TJOFLAT, Circuit Judges, and BEAVERSTOCK, District Judge.

JORDAN, Circuit Judge: This appeal arises out of the Chapter 7 bankruptcy of Matthew and Kathleen Feshbach. Following a bench trial, the bankruptcy court concluded that the Feshbachs’ 2001 tax liability was not dischargeable under 11 U.S.C. § 523(a)(1)(C) because the Feshbachs "willfully attempted ... to evade or defeat" that liability. The district court affirmed, and the Feshbachs appealed. With the benefit of oral argument, we too affirm.

I

Mr. Feshbach is an investment professional and former hedge-fund manager. Beginning in the 1980s, he employed a strategy called "selling short against the box" that allowed him to delay the recognition of his taxable income from investments. The strategy worked well for Mr. and Mrs. Feshbach for several years. Without having to pay taxes on their "boxed-in" capital gains, they were able to build a $14 million home in Monte Sereno, California, in the 1990s. Mr. Feshbach believed that he would never have to pay taxes on those capital gains and that he could pass them on to his heirs untaxed.

Events in the late 1990s, however, forced Mr. Feshbach to close out his investment positions and incur a tax liability of $1,950,827 in 1999. Although the Feshbachs made some payments and received other credits during the 1999 tax year, they did not submit a payment with their tax return.

A

In June of 2001, the Feshbachs submitted an "offer-in-compromise" to the IRS to settle the outstanding tax debt from 1999 for $1 million, about half of what they owed. They proposed an immediate payment of $200,000, a payment of $300,000 upon the sale of their home in Bellaire, Florida (where they lived by that point), and payment of the balance over the next five years. The Feshbachs made a $200,000 payment toward their 1999 tax liability, consistent with their offer.

The IRS evaluates an offer-in-compromise based on a number of factors, including its collection potential, which is a function of the taxpayer's assets and anticipated future income minus "allowable expenses." See generally 26 C.F.R. § 301.7122-1(b) – (c). Testimony at trial established that allowable expenses reflect the taxpayer's standard of living, using national averages for household spending. Although the term seems to imply otherwise, an allowable expense is not a directive that a taxpayer limit his spending on particular goods or services. Allowable expenses instead help the IRS gauge an offer-in-compromise based on the taxpayer's income and lifestyle compared to that of the average American. Considering the Feshbachs’ income and allowable expenses, the IRS believed the offer was a "nonstarter." The Feshbachs submitted documents showing that the collection potential was not only greater than their offer, but also greater than the entire debt. According to the IRS, the Feshbachs were living "way over their allowable expenses" and their income was "high."

The Feshbachs withdrew their offer before the IRS could reject it. They opted instead for a temporary agreement in which they would make monthly payments of $1,000 while the IRS suspended its collection efforts. This arrangement was intended to give the Feshbachs breathing room to sell their Florida home and adjust their standard of living to free up cash for debt payments.

The Feshbachs had, however, dug a deeper hole. To make the $200,000 payment with their first offer, they liquidated some securities. Those transactions in turn yielded capital gains that contributed to a $3,247,839 tax liability for the 2001 tax year. By the time they filed their tax return for that year, the Feshbachs owed the IRS a total of $5,198,156.

That debt did not appear to be insurmountable for the Feshbachs. In 2001, Mr. Feshbach founded a hedge fund for "ultra-high net worth" investors, and that venture turned out to be highly profitable. Over the next nine years, the Feshbachs would earn over $13 million. Unfortunately for the IRS, the Feshbachs would also spend about $8.5 million on personal expenses and charitable contributions in that same period, leaving a balance of more than $3.8 million on their tax debt.

Back in 2002, though, the Feshbachs were optimistic about their hedge fund. They had a network of millionaire and institutional investors, and they continued to live in luxury while expanding their investment business. In 2002, for example, they had $764,498.78 in expenses, including a domestic payroll of $58,832.90 and household and personal expenses of $383,587.85.

B

The Feshbachs approached the IRS in 2002 with a second offer-in-compromise. This offer was higher in dollar value than the first but much lower as a percentage of their total liability. The Feshbachs by that point owed more than $6 million on their 1999 and 2001 taxes combined, including interest and penalties, and proposed to settle with the IRS for $1.25 million.

After reviewing their financial disclosures, an IRS officer determined that the Feshbachs were earning much more income than they had reported and noted that there "appear[ed] [to be] much more on [the] horizon." The Feshbachs had represented on a Form 433-A financial disclosure document that they were earning about $15,000 per month, or $180,000 annually. Yet they reported an income of $611,413 on their 2002 tax return. The following year, 2003, they reported an income of $738,608 on their tax return. And the IRS officer was right—there was indeed much more income on the horizon. For the next three tax years after they submitted their second offer-in-compromise, the Feshbachs reported more than $10 million in income.

As for the extant temporary agreement with the IRS, the Feshbachs had not yet sold their principal Florida residence as they had represented. They would stay in that home until 2008. Nor had they reduced their standard of living. They were paying only $1,000 per month toward their tax debt, while continuing to incur expenses that were "excessive" and "egregious" in the opinion of the IRS.

While the Feshbachs were paying $1,000 monthly to the IRS in 2002 and 2003, their total personal expenditures were over $1.5 million. That included the expenses listed above in 2002, and then, in 2003, another $106,150.42 on clothing and care, $16,283.58 on entertainment, $40,080.94 on groceries, and $56,497.01 on personal travel. The IRS officer who took over the case in 2003 believed that the second offer-in-compromise was evidence of a delay tactic. The IRS ended up rejecting that second offer, concluding that "there was no basis to compromise because the [Feshbachs] had the ability to fully pay the tax liability."

C

With their second offer-in-compromise declined, the Feshbachs approached the IRS in 2005 about an installment plan. The IRS explained that it would approve such an arrangement, if at all, only on the 2001 debt, and only after the Feshbachs paid the 1999 tax debt in full. The Feshbachs made two $50,000 payments in April and May of 2005 and obtained a loan to pay the balance of the 1999 tax debt. The IRS then approved a permanent installment plan in which the Feshbachs would pay $120,000 per quarter until the 2001 debt was fully satisfied.

From October of 2005 through January of 2008 the Feshbachs stayed on pace with the installment plan, paying $1.2 million over ten quarters. But in 2008 the financial crisis hit and Mr. Feshbach's health declined. The investment fund faltered and the Feshbachs began missing payments.

In September of 2008, the Feshbachs made a third offer-in-compromise of $120,000 on a $3.6 million balance for the 2001 tax debt. They proposed monthly payments of $2,500 over 48 months. Along with this offer they submitted another Form 433-A in which they claimed a monthly income of $833, or $9,996 annually. At that time, they were incurring monthly household expenses of over $12,000. An IRS officer testified at trial that he "knew either income was understated or expenses overstated or a combination of the both." In their next tax return, for the tax year 2008, the Feshbachs claimed an income of $193,205 (19 times more than what they stated on the Form 433-A).

While their third offer was pending from September 2008 to August 2009, the Feshbachs spent between $1,400 and $1,500 per month on entertainment, more than $4,000 per month for groceries, $4,000 per month for domestic help, and $4,500 per month on rent, in addition to other expenses. In 2009 and 2010, the Feshbachs spent a total of about $90,000 on household employee wages and $143,000 in charitable contributions. Mr. Feshbach made about $40,000 of those charitable contributions through his business in his son's name.

After a lengthy review process, the IRS concluded in December of 2010 that the Feshbachs had the means to pay $15,000 per month. The Feshbachs made four $15,000 payments toward a new installment plan but ceased payments altogether in May of 2011.

II

The Feshbachs filed for Chapter 7 bankruptcy in 2011, soon after they defaulted on their installment payments to the IRS. They initiated an adversary proceeding in the bankruptcy court seeking a determination that their 2001 tax liability was dischargeable. The government opposed discharge, arguing that the Feshbachs willfully attempted to evade and defeat that liability. It relied primarily on the Feshbachs’ total income and spending over the previous nine years. The government later argued that the Feshbachs had submitted "deliberate lowball offers-in-compromise which only served to delay" the payment of their taxes and had "obfuscate[d] ... their true financial picture in materials provided to the IRS."

After a bench trial, the bankruptcy court concluded in a comprehensive order that the Feshbachs willfully attempted to evade or defeat their 2001 tax liability. See In re Feshbach , 576 B.R. 660, 684 (Bankr. M.D. Fla. 2017). It found that the Feshbachs earned $13,056,518 from 2002 to 2010 and had the capacity to pay their 2001 tax debt in full. Yet they chose to spend more than $8.5 million on personal expenses and charitable contributions rather than paying their tax liability, leaving a $3.8 million balance. See id. at 668–69.

The bankruptcy court did not credit Mr. Feshbach's "self-serving" testimony that the couple incurred these myriad expenses to cultivate an appearance of wealth, attract clients, and generate income to pay down their tax debt. See id. at 675. Even if some portion of the personal expenses could be attributed to promoting their hedge fund business, the Feshbachs could not justify all of the personal spending as having a business purpose. Much of their spending had "no bearing on the [their] earning potential." Id. at 676. The house in Aspen and fine dining, for example, could have no business purpose because the Feshbachs testified that they never entertained potential clients outside of their home. Likewise, they failed to explain how $721,806.60 on personal travel expenses helped the business when the hedge fund already paid for Mr. Feshbach's business travel. And they did not explain how expensive private schooling for their son and large charitable gifts to the church where Mrs. Feshbach worked helped generate income. See id. at 676–77. The bankruptcy court concluded that the Feshbachs’ excessive discretionary spending constituted an attempt to evade or defeat taxes. See id. at 677–78.

The bankruptcy court also found that the Feshbachs acted willfully. See id. at 682. It rejected their argument that the offers-in-compromise demonstrated a good-faith desire to resolve the 2001 tax debt. Taken in the context of their high income and excessive personal spending, the offers-in-compromise were unrealistic. The Feshbachs had instead used the offer process "in a calculated manner" to delay the IRS collection efforts. See id. at 682. Therefore, even though the Feshbachs had not defrauded the IRS, they acted willfully when they redirected their funds away from paying their debt and toward their personal luxuries.

At the conclusion of the trial, the bankruptcy court had requested briefing on whether the Feshbachs’ tax debt could be partially discharged, in the event it determined that the Feshbachs would not have been able to satisfy their entire debt even if they had restrained their spending. See id. at 682–83. The bankruptcy court ultimately concluded that partial discharge would not have been permitted under § 523(a)(1)(C). See id. at 684. It observed that, while the question has not been resolved by the Eleventh Circuit, most courts have concluded that partial discharge is not allowed. See id. at 683.

The Feshbachs appealed, and the district court affirmed. It ruled that the bankruptcy court did not clearly err in finding that the Feshbachs willfully attempted to evade or defeat their 2001 tax liability. See Feshbach v. Dep't of Treasury, Internal Revenue Serv. , 594 B.R. 495, 500–01 (M.D. Fla. 2018). Like the bankruptcy court, the district court concluded that § 523(a)(1)(C) does not allow for a partial discharge. See id. at 501–02.

III

A Chapter 7 debtor receives a general discharge from all debts that arose before he filed a bankruptcy petition. See 11 U.S.C. § 727(b). This "fresh start" is available to the "honest but unfortunate debtor." In re Fretz , 244 F.3d 1323, 1326 (11th Cir. 2001) (quoting Grogan v. Garner , 498 U.S. 279, 286–87, 111 S.Ct. 654, 112 L.Ed.2d 755 (1991) ). To ensure that only the deserving debtor receives the benefits of discharge, Congress has enumerated several exceptions. As relevant here, there is no discharge "for a tax with respect to which the debtor ... willfully attempted in any manner to evade or defeat such tax." 11 U.S.C. § 523(a)(1)(C).

We have set forth a two-prong test for whether a tax debt is dischargeable under § 523(a)(1)(C). The government must prove by a preponderance of the evidence (1) that the debtor "attempted in any manner to evade or defeat [a] tax," and (2) that the attempt was done "willfully." In re Jacobs , 490 F.3d 913, 921 (11th Cir. 2007) ; Fretz , 244 F.3d at 1327.

To satisfy the conduct prong under § 523(a)(1)(C), the government must demonstrate that the debtor "engaged in affirmative acts to avoid payment or collection of taxes either through commission or culpable omission." Jacobs , 490 F.3d at 921 (internal quotation marks and citation omitted). The universe of evasive acts or omissions is broad: "Congress did not define or limit the methods by which a willful attempt to defeat and evade might be accomplished and perhaps did not define lest its effort to do so result in some unexpected limitation." Fretz , 244 F.3d at 1327 (quoting Spies v. United States , 317 U.S. 492, 499, 63 S.Ct. 364, 87 L.Ed. 418 (1943) ). See also In re Gardner , 360 F.3d 551, 557 (6th Cir. 2004) (explaining that cases interpreting § 523(a)(1)(c) "cast a wide net"). Mere nonpayment of taxes does not suffice to render a debt non-dischargeable, but failure to pay is relevant evidence within the "totality of conduct." Fretz , 244 F.3d at 1328 (quoting In re Fegeley , 118 F.3d 979, 983 (3d Cir. 1997) ).

Apparently demonstrating our affinity for multipart tests, we have adopted a three-prong test to determine whether a debtor's attempted evasion under § 523(a)(1)(C) is willful. The government must demonstrate that "(1) the debtor had a duty under the law, (2) the debtor knew he had that duty, and (3) the debtor voluntarily and intentionally violated that duty." In re Griffith , 206 F.3d 1389, 1396 (11th Cir. 2000) (en banc). It is undisputed that the Feshbachs had a duty to pay their 2001 tax liability and knew of that duty, so only the third prong is at issue in this case.

As the second court of review of a bankruptcy court judgment, we examine that judgment independently and employ the same standards of review as the district court. See In re Issac Leaseco, Inc. , 389 F.3d 1205, 1209 (11th Cir. 2004). We review the bankruptcy court's findings of fact for clear error and conclusions of law de novo . See Jacobs , 490 F.3d at 921. A bankruptcy court's determination that a taxpayer willfully attempted to evade or defeat a tax under § 523(a)(1)(C) is a question of fact subject to clear error review. See id. A finding of fact is clearly erroneous if, after reviewing the evidence, we are left with the definite and firm conviction that a mistake has been made. See Morrissette-Brown v. Mobile Infirmary Med. Ctr. , 506 F.3d 1317, 1319 (11th Cir. 2007). Stated differently, "[w]here there are two permissible views of the evidence, the factfinder's choice between them cannot be clearly erroneous." Anderson v. Bessemer City , 470 U.S. 564, 574, 105 S.Ct. 1504, 84 L.Ed.2d 518 (1985). Where findings "are based on determinations regarding the credibility of witnesses" at a bench trial, the reviewing court must afford substantial deference to those findings. See id. at 575, 105 S.Ct. 1504.

A

The Feshbachs argue that both courts below erred in concluding that their personal spending alone could satisfy the conduct prong of § 523(a)(1)(C). They contend that we must review that conclusion de novo because it expands our interpretation of the statute. We are not convinced that the Feshbachs’ characterization of the statute and our case law is accurate. See § 523(a)(1)(C) (encompassing attempts "in any manner"); Jacobs , 490 F.3d at 926 (explaining that the "conduct requirement is ... satisfied by [the debtor's] large discretionary expenditures"). But even if we assume that they are correct, we need not address the legal question of whether excessive personal spending alone can ever suffice. As we explain, there is evidence of other conduct that supports the bankruptcy court's finding that the Feshbachs attempted to evade their 2001 tax liability.

The bankruptcy court found that, while spending lavishly on personal luxuries instead of paying their taxes, the Feshbachs used the offer-in-compromise process to delay collection. That subsidiary finding was crucial to its analysis of the mental state prong and its determination that the Feshbachs acted willfully. But the finding touches on the conduct prong as well, as the bankruptcy court inferred willfulness from the Feshbachs’ actions. Cf. Spies , 317 U.S. at 499–500, 63 S.Ct. 364 (explaining that a jury may infer willfulness for criminal tax evasion from the defendant's conduct); United States v. Sorrentino , 726 F.2d 876, 880 (1st Cir. 1984) (explaining that the jury may infer willfulness for tax evasion "from the fact of underreporting [income] coupled with evidence of conduct by the defendant tending to mislead or conceal").

We may affirm on any ground supported by the record, see Ry. Labor Executives’ Ass'n v. S. Ry. Co. , 860 F.2d 1038, 1040 n.2 (11th Cir. 1988), and we consider the Feshbachs’ spending in this case, as did the bankruptcy court. See, e.g., In re Hamm , 356 B.R. 263, 277 (Bankr. S.D. Fla. 2006) (explaining that a debtor's "lavish lifestyle is [an] indicia of attempts to evade or default taxes such that they are excepted from discharge under § 523(a)(1)(C)," and collecting cases). But we also consider the Feshbachs’ conduct related to the offer-in-compromise process. As will become evident, these two patterns of conduct were interrelated and formed a larger mosaic of attempted evasion. See Fretz , 244 F.3d at 1328 (explaining that a court may evaluate the "totality of conduct to determine whether or not the debtor willfully attempted to evade or defeat taxes") (quoting Fegeley , 118 F.3d at 983 ). Considering their conduct as a whole, we conclude that the bankruptcy court did not clearly err in finding that the Feshbachs attempted to evade their 2001 tax debt.

Approaching the IRS with an offer-in-compromise is not, by itself, an act to evade taxes or evidence of attempted evasion. That is true even if the IRS rejects the taxpayer's offer. The negotiation process is flexible, and taxpayers must have leeway to present good-faith offers, even if the IRS concludes that an offer is too low or otherwise not acceptable. If we are quick to interpret rejected offers-in-compromise as bad-faith attempts to evade, then we risk discouraging settlement.

Having said that, we are not concerned that the result here will discourage settlement offers in the future. This is not a case of the IRS rejecting good-faith or facially reasonable offers-in-compromise. There is ample evidence that the Feshbachs approached the IRS with inadequate and unrealistic offers given their income and spending, and that they used the offer-in-compromise process to delay the payment of their taxes. See In re Colish , 289 B.R. 523, 533–34 (Bankr. E.D.N.Y. 2002) (finding that an attorney familiar with the offer-in-compromise process was attempting to delay the collection of his taxes by filing "serial" offers that were "clearly too low in relation to the tax obligations owed").

The bankruptcy court found that the Feshbachs were more sophisticated than the average taxpayer. Mr. Feshbach, for example, testified that he did his "homework" and was familiar with the Tax Code and IRS regulations. He would have understood, then, that a pending offer would halt IRS collection efforts. See 26 C.F.R. § 301.7122-1(g). See also In re Mitchell , 633 F.3d 1319, 1324 (11th Cir. 2011) ("A pending offer in compromise halts any collection activity by the IRS, and [the debtor] understood this."); Colish , 289 B.R. at 533–34 (finding that the debtor "knew that, while the offers were pending, he could forestall collection of all tax liabilities under consideration, which permitted him to delay filing bankruptcy and seeking discharge of his taxes").

The Feshbachs took advantage of the offer process by continuing to maintain an extravagant lifestyle. Before entering into the permanent installment agreement, for instance, they paid only $1,000 a month under a temporary agreement while spending hundreds of thousands of dollars on personal luxuries. It is true that they were abiding by the temporary agreement insofar as they were making the required payments. But they represented that they would reduce their expenses and sell their Florida home as a condition of the temporary arrangement. They never lowered their expenses and it took them several years to sell the home. In the two years governed by the temporary agreement, the Feshbachs spent over $1.5 million on personal expenditures. For that reason, the IRS rejected the second offer-in-compromise: the Feshbachs, "[w]hile paying a token $1,000.00 monthly to [the] IRS[,] ... continued to maintain the same lavish lifestyle."

At trial, the Feshbachs tried to justify their personal spending as business expenses, and they claimed they were unable to sell the Florida home despite honest efforts. The bankruptcy court did not credit this evidence. It found that many of the personal expenses had no business purpose whatsoever. It also concluded, from Mr. Feshbach's testimony, that the couple likely never intended to sell the home in the first place. We afford substantial deference to these findings, as they are based on the credibility of witnesses at a bench trial. See Anderson , 470 U.S. at 575, 105 S.Ct. 1504.

There is also some suggestion that the Feshbachs obfuscated their financial picture to the IRS during the offer-in-compromise process. The bankruptcy court did not expressly address this issue, but the government raised it below. Although we cannot say with 100% certainty that the Feshbachs deliberately lied about their income or expenses, we cannot ignore the vast disparities between the income they reported to the IRS during the settlement process and the income they actually earned. The inference that the Feshbachs clouded their income and spending bolsters the bankruptcy court's finding that the Feshbachs used the offer-in-compromise process as a delay tactic.

Contemporaneous notes and trial testimony show that IRS officers believed the incomes the Feshbachs reported on the Forms 433-A were inconsistent with their actual incomes. An IRS officer reviewing the first offer had difficulty reconciling the Feshbachs’ claimed and actual income because of the unexplained $50,000 monthly deposits to their bank accounts. With respect to their second and third offers, the Feshbachs’ tax returns were also telling. On their 2002 and 2008 tax returns, the Feshbachs claimed income about 3.5 and 19 times greater than what they had represented months earlier. In 2002, the Feshbachs predicated their offer on an annual income of $180,000, but on their next tax return they reported an income of $611,413. In September of 2008, the Feshbachs offered to settle their debt for a few cents on the dollar based on an annual income of $9,996, yet they reported an annual income of $193,205 on their subsequent tax return.

In their briefing below, the Feshbachs argued that their income was accurate when reported but that the delays in the IRS review process and volatility in their business led to disparities with their tax returns. This argument is unconvincing under the circumstances. The notion that the Feshbachs faced such difficult times that they needed to settle their tax debt for cents on the dollar is difficult to reconcile with their consistently excessive spending. Another taxpayer in the Feshbachs’ shoes—whose annual income really had dropped from the millions to $180,000 and eventually to $9,996—would cut back on the personal chef, the dining-out, the expensive private schooling, or the other recurring non-business expenses. The Feshbachs never did that. In the two years after reporting an income of $180,000 on a Form 433-A, the Feshbachs spent $1.5 million on themselves. And when they submitted their 2008 offer of $120,000—just 3.3% of their total debt—the Feshbachs were spending over $12,000 per month on household goods and services.

Nor does it appear to be a coincidence that this happened multiple times. The Feshbachs submitted meager offers to the IRS at conveniently low points in their highly profitable business venture. Nothing in the record, other than post-hoc argument, can explain the fortuitous and significant jumps in income soon after their offers were submitted. The bankruptcy court found that Mr. Feshbach was a sophisticated taxpayer and investor who had done his homework combing through IRS statutes and regulations. That is reason to be suspicious of the timing and content of these offers.

The Feshbachs also argued below that the $180,000 income on which they predicated their second offer-in-compromise was based on an income average from the prior four years, 1998 through 2001, modified to account for their "incurrence of phantom income in 1999 and 2001." Whether that is an accurate four-year average, properly accounting for "phantom income," we cannot say for sure. But the Feshbachs’ argument is seriously weakened by their 1998 income of more than $700,000. To end up with a four-year average of $180,000 after earning $700,000 in the first year, the Feshbachs would have needed to earn less than $20,000 over the next three years (from 1999 to 2001). Yet they earned over $5,000,000 in 1999 and around $8.5 million in 2001. It is difficult to believe that all this money was "phantom" income. We see no evidence that Feshbachs lived as though they were earning almost no income over any three-year period. On the contrary, an IRS officer concluded in 2001 that the Feshbachs were living way over their allowable expenses.

In any event, even if the Feshbachs’ evidence and arguments could have permitted a different result, they did not compel it. Our review of the bankruptcy court's conclusion is for clear error. Having examined the record, we find substantial evidence of attempted evasion and conclude that the bankruptcy court did not clearly err.

B

We now turn to the mental state prong of § 523(a)(1)(C). The Feshbachs make two arguments. First, they contend that the bankruptcy court applied the incorrect legal standard of mens rea . They assert that the government was required to prove that their overspending was undertaken with the specific intent to evade taxes. Second, they argue that even under the standard applied by the bankruptcy court, the government failed to prove that they acted with the requisite mental state. We disagree on both points.

In Fegely , 118 F.3d at 984, the Third Circuit distinguished between the criminal provisions of the Tax Code, which require proof of fraudulent intent, and "civil willfulness" under § 523(a)(1)(c), which requires only a "voluntary, conscious, and intentional" attempt to violate a known duty to pay taxes. In Fretz , 244 F.3d at 1330, we adopted the civil willfulness standard, relying on Fegely . From that point, "the established rule in this Circuit [has been] that a debtor's tax debts are non-dischargeable if the debtor acted knowingly and deliberately in his efforts to evade his tax liabilities." Mitchell , 633 F.3d at 1328. A showing of criminal fraudulent intent is not required. See id. Accord Alan N. Resnick & Henry J. Sommer, Collier on Bankruptcy ¶ 523.07[4] (16th ed. 2020); 3 William L. Norton III, Norton Bankruptcy Law & Procedure § 57:7 (3d ed. April 2020). We also note that the Feshbachs’ exact argument was rejected in United States v. Coney , 689 F.3d 365, 374 (5th Cir. 2012), and that decision relied on Fretz , 244 F.3d at 1330, and similar cases from our sister circuits.

Having concluded that the bankruptcy court applied the correct legal standard, we review for clear error its finding that the Feshbachs acted willfully. If the totality of the circumstances tends to show that they acted voluntarily and intentionally in their attempts to evade or defeat the payment of taxes, the third prong of the mental state test is satisfied. See Griffith , 206 F.3d at 1396–97 ; Fegeley , 118 F.3d at 983–84.

In our view, the bankruptcy court did not clearly err in finding that the Feshbachs acted willfully. For one, we have held that excessive discretionary spending constitutes circumstantial evidence of willfulness. See Mitchell , 633 F.3d at 1329. See also In re Peterson , 317 B.R. 556, 564 (Bankr. N.D. Ga. 2004) (explaining that "a debtor's spending habits during the tax years in issue are relevant under the totality test" with respect to the mental state prong). In addition, the bankruptcy court also inferred willfulness from the Feshbachs’ exploitation of the offer-in-compromise process.

The Feshbachs cite numerous pieces of evidence that, in their view, demonstrate a good-faith and genuine effort to pay their taxes. We do not weigh that evidence in the first instance, as our review is for clear error, but their arguments are not convincing in any event.

For example, the Feshbachs contend that the bankruptcy court "disregarded" the trial testimony of IRS officers that their offers were not dilatory. But there were also the contemporaneous notes of the IRS officers who believed the offers were intended to delay collection. The bankruptcy court, after having listened to the testimony, credited the latter. It did not clearly err in doing so.

The Feshbachs also argue that the IRS manual and internal regulations set forth guidelines to evaluate evasive offers and point out that the officers never rejected their offers as dilatory. They also assert that their efforts to compromise yielded substantial penalties and interest, which they say suggests a lack of willful evasion. But even accepting those contentions at face value, they are probative at best; they do not show that the bankruptcy court committed clear error. The Feshbachs’ several subsidiary points and arguments do not carry the day.

IV

At the conclusion of the bench trial, the bankruptcy court asked the parties to brief the issue of whether a partial discharge of the 2001 tax liability was possible. In its order denying discharge, it explained that partial discharge would have been relevant "if, for example, the [c]ourt determined that the Feshbachs would not have been able to satisfy their entire tax debt even if they had restrained their spending." In re Feshbach , 576 B.R. at 682–83. It also concluded that partial discharge was not available. See id. at 684.

The bankruptcy court did not find that the Feshbachs were unable to pay the entire debt, personal spending aside. Nor do we find anything in the record suggesting as much, particularly given that the Feshbachs earned $13 million in income over nine years, that their tax debt was less than half of that amount, and that the IRS determined several times that they had the capacity to pay their debt in full. Partial discharge is therefore not at issue, and we have no reason to reach the legal question of whether § 523(a)(1)(c) would permit it.

V

Without reaching the question of partial discharge, we affirm the bankruptcy court's judgment that the Feshbachs’ 2001 tax debt was non-dischargeable under § 523(a)(1)(c).

AFFIRMED.