United States v. Hoffman

Full title: UNITED STATES OF AMERICA, Plaintiff - Appellee Cross-Appellant v. PETER M…

Court: UNITED STATES COURT OF APPEALS FOR THE FIFTH CIRCUIT

Date published: Aug 9, 2018

No. 16-30104 (5th Cir. Aug. 9, 2018)

Opinion

No. 16-30104 c/w 16-30226 c/w 16-30013 c/w 16-30527

08-09-2018

UNITED STATES OF AMERICA, Plaintiff - Appellee Cross-Appellant v. PETER M. HOFFMAN Defendant - Cross-Appellee MICHAEL P. ARATA; SUSAN HOFFMAN, Defendants - Appellants Cross-Appellees UNITED STATES OF AMERICA, Plaintiff - Appellee-Cross- Appellant v. PETER M. HOFFMAN, Defendant - Appellant-Cross- Appellee UNITED STATES OF AMERICA, Plaintiff - Appellant v. PETER M. HOFFMAN; MICHAEL P. ARATA, Defendants - Appellees UNITED STATES OF AMERICA, Plaintiff - Appellee Cross-Appellant v. PETER M. HOFFMAN, Defendant - Appellant Cross-Appellee MICHAEL P. ARATA; SUSAN HOFFMAN, Defendants - Cross Appellees

GREGG COSTA, Circuit Judge

Appeals from the United States District Court for the Eastern District of Louisiana Before KING, DENNIS, and COSTA, Circuit Judges. :

With its colorful history and rich cultural stew, Louisiana has long been a popular setting for works of fiction, including movies. In recent years the state has also tried to become a place where films are made. That effort enjoyed considerable success. The Curious Case of Benjamin Button, Django Unchained, Twelve Years a Slave, The Dallas Buyer's Club, and Dawn of the Planet of the Apes are some recent films of note shot in New Orleans. Believe it or not, in one recent year (2013) Louisiana surpassed even California as the most popular locale for filming major-studio productions. Mike Scott, Louisiana Outpaces Los Angeles, New York, and All Others in 2013 Film Production, Study Shows, TIMES-PICAYUNE (Mar. 10, 2014). This development led some to call New Orleans "Hollywood South." Id.

State tax credits for the film industry spurred much of this growth. Id. ("[M]ake no mistake: The state's tax-credit program . . . is largely responsible for the surge in local productions."). They also provided an incentive for fraud. A jury found that to be the case for Peter Hoffman, Michael Arata, and Susan Hoffman. It credited the government's allegations that they submitted fraudulent claims for tax credits, mostly by (1) submitting false invoices for construction work and film equipment or (2) using "circular transactions" that made transfers of money between bank accounts look like expenditures related to movie production. Their principal challenge to those convictions is an argument that the tax credits are not property within the meaning of the mail and wire fraud statutes but are instead akin to the video poker licenses the Supreme Court rejected as a basis for federal prosecution in Cleveland v. United States, 531 U.S. 12 (2000). If we conclude that the credits are property subject to the federal fraud statutes, defendants also contend that the evidence was insufficient to convict because they made a good-faith effort to comply with a state program riddled with gray areas.

While the defendants seek to undo their convictions, the government is unhappy with the sentences of probation that all three received. So it too appeals, arguing that the substantial downward variances exceeded the district court's discretion. The government also contends that the district court improperly vacated a number of the jury's guilty verdicts.

I.

The Hoffmans and Arata owned and jointly operated Seven Arts Pictures Louisiana, LLC (Seven Arts). Each of them was also involved in several other film-related ventures. Through their companies, defendants purchased a "dilapidated mansion" at 807 Esplanade in New Orleans, intending to renovate the structure and turn it into a postproduction facility where films are edited and prepared for final release. To offset the cost of this project, Seven Arts applied for film infrastructure tax credits with the state.

A.

Louisiana enacted the Motion Picture Incentive Tax Credit in 1992 to encourage local development of the movie and television industry. La. Rev. Stat. § 47:6007. In its initial form, the law authorized investors to claim a credit for 50% to 70% of losses sustained during in-state film production. In other words, it was a "safety net" for bad film investments. John Grand, Motion Picture Tax Incentives: There's No Business Like Show Business, STATE TAX NOTES at 791 (Mar. 13, 2006). The state legislature extended the program in 2002, permitting investors to claim tax credits for money spent on profitable projects. La. Rev. Stat. § 47:6007(C)(1) (2002). The next year saw further amendment, this time allowing investors to sell or transfer the tax credits. Id. § 47:6007(C)(4) (2003). This was an important innovation because many investors—those like Peter Hoffman who resided in California—did not themselves owe Louisiana taxes. Nontransferable credits had been of little value to these numerous out-of-state producers.

The program was again amended in 2005 (and extended in 2007), when the legislature authorized income tax credits for state-certified infrastructure and production projects.¹ See generally La. Rev. Stat. § 47:6007(C) (2005). Projects with total base investment exceeding $300,000 could qualify for tax credits worth up to 40% of in-state expenditures. Id. § 47:6007(C)(1)(b)(i), (iii); see also Dep't of Revenue, Policy Servs. Div., 2005 Regular Legislative Session: Legislative Summaries 5 (Jan. 13, 2006), http://www.rev.state.la.us/publications/lsls(2005).pdf.

1 The film infrastructure tax credits central to this case lapsed in 2009, though investors can still obtain credits for film production. See Loren C. Scott & Assocs., Inc., The Economic Impact of Louisiana's Entertainment Tax Credit Programs ii, 1-2 (Apr. 2013), https://louisianaentertainment.gov/assets/ENT/docs/2013_OEID_Program_Impact_Report%20_FINAL.pdf. But in June 2017 Louisiana lawmakers placed a long-term spending cap "on the tax breaks for Hollywood South" and imposed a 2025 sunset provision on the entire program. Associated Press, Louisiana's Film Tax Credit Program to Continue, with a Cap, TIMES-PICAYUNE (June 2, 2017).

Louisiana's Office of Entertainment Industry Development, a component of the Department of Economic Development, administered the program. Issuance of film tax credits was a two-step process. First, the applicant had to file an initial application for tax credits and obtain a precertification letter from the state agencies. See Red Stick Studio Dev., L.L.C. v. Louisiana, 56 So. 3d 181, 183-84 (La. 2011). After receiving that authorization, the applicant still had to submit a cost report tallying its expenditures, accompanied by an audit from an independent accountant. Id. at 183 n.4. After a review of those materials, the same state agencies determined whether the expenditures should be certified and tax credits issued.

For infrastructure projects, qualifying expenditures could include the purchase, construction, and use of tangible items directly related to Louisiana film production. The law defined "base investment" as the "actual investment made and expended," while "expended in the state" meant "property which is acquired from a source within the state and . . . services procured and performed in the state." La. Rev. Stat. § 47:6007(B)(1), (3) (2007). And the state could recapture tax credits if it found that "monies for which an investor received tax credits . . . [we]re not invested in and expended with respect to a state-certified production . . . and with respect to a state-certified infrastructure project." Id. § 47:6007(E)-(F) (2007).

B.

Such was the statutory and administrative landscape facing Peter, Arata, and Susan as they sought to develop 807 Esplanade.² A bank loaned them $3.7 million for the project, $1.7 million of which was earmarked to purchase the property while the remainder was placed in an account that could be drawn on to make payments for construction and renovation. From its inception, Seven Arts sought to lower the cost of the 807 Esplanade project via various tax credits. Beyond the film credits, for example, it sought "historic rehabilitation tax credits." In October 2007, Arata submitted the company's initial film credit application to the state, which included a cost estimate of $9 million, a business plan, and a contractor's agreement.

2 We refer to Peter and Susan Hoffman by their first names to avoid confusion.

The state issued a precertification letter in May 2008. The letter contained a caveat that it did not guarantee any tax credits would be issued. But it did note that the project as described "appear[ed] to meet the criteria of a State-Certified Infrastructure Project," subject to administrative rules that may be released at a future date. The letter also placed certain restrictions on the tax credit certification. Namely, Seven Arts had until the end of 2008 to earn credits on the project, unless it spent $4.5 million prior to that date (in which case future credits might be possible). It also mentioned that before any credits could be "certified and released" at least $2.25 million (25%) in base investment must have been spent on film-related infrastructure. That 25% had to be used for "the creation of infrastructure specifically designed for motion picture production," not on the purchase of land or preexisting facilities. But tax credits could be earned on so-called "multiple-use facilities" once the production facility was complete.

C.

As the precertification letter emphasized, it did not authorize the issuance of tax credits. That could only occur based on the "actual amount expended by the project," verification of that amount by an independent auditor, and final approval by state authorities. To satisfy those critical final steps, the defendants submitted three cost reports and audits. Misrepresentations in those reports, the ones mentioned earlier that involved fake invoices and circular transactions, are what led to this prosecution.

In October 2008, two months prior to the expenditure deadline, Peter and Arata hired an auditing firm to review project expenditures. Katherine Dodge, the auditor, requested additional information, like bank transactions showing the company's transfers to vendors. Arata emailed Regions Bank with a request to forward withdrawal and deposit slips to Dodge. But it was too late. The next day Dodge's firm withdrew based on her concerns.

Seven Arts soon replaced her with auditor Katie Davis of the Malcom Dienes firm. Peter and Arata provided Davis with the company's general ledger, which noted a $7.42 million capital contribution from the parent company—Seven Arts Pictures, Inc.—along with vendor invoices and receipt of payment confirmations signed by Damon Martin and Leo Duvernay. These documents made it appear as though the company had made payments out of the capital contribution to Martin, owner of Departure Studios, for film equipment and to Duvernay, the project's general contractor, for construction. But bank statements, which were not included just as they had not been sent to the first auditor, revealed that those transactions were in reality withdrawals and deposits of the same funds. They were, in other words, "circular transactions" that the government argued were intended to trick state authorities into believing that Martin and Duvernay had been paid when they had not.

In February 2009, Arata sent the first cost report, which claimed $6,531,202 in qualifying expenditures through October 2008, along with the auditor's statement verifying that amount, to the state. Lacking access to the bank records, the audit verified that $1,027,090 had been paid to Martin and $1,749,257 to Duvernay. The report also listed a $3.7 million payment to purchase and renovate 807 Esplanade, nearly the entire balance of the remaining expenditures claimed.

Louisiana authorities certified and "paid out" tax credits worth $1,132,480.80 in June 2009. That amount was substantially below 40% of the claimed expenditures because the $3.7 million building purchase was "deemed multiuse" and therefore ineligible for credits until the project was complete. After certification, Seven Arts "cashed in," to use the district court's words, by selling the credits to third-party taxpayers.

About two months after Louisiana issued the credits, concerns about Peter fabricating invoices led Arata to send a letter terminating his day-to-day participation in Seven Arts and other projects in which he acted as Peter's lawyer. Arata also reported his concerns about the invoices to the President and CFO of the Seven Arts parent company. He did not, however, report this to state authorities in accordance with ethics advice he received from a lawyer. Nor did he mention his concerns in his letter to Peter. Instead, he invoked the time-honored excuse of needing to devote more time to his family (his son), as well as to his other business interests. Because Arata retained an ownership stake in Seven Arts through his interest in Voodoo Studios, LLC, he stated in the letter to Peter that he would still "assist with the renovation and completion of 807 Esplanade as my time permits."

So Peter on his own submitted the company's second cost report to state authorities in January 2010. That report, audited by the Dienes firm, claimed almost $6 million in expenditures related to 807 Esplanade from November 2008 to September 2009, an amount in addition to that already certified in June 2009. The purported expenditures included $2,302,860 in construction costs paid to Duvernay, $807,202 for audio equipment, $705,587 for interest payments on a $10 million loan from Seven Arts Filmed Entertainment LA, LLC (SAFELA), $400,000 in project management fees to Leeway Properties, Inc. (a Susan Hoffman entity), $350,000 in legal and notary fees for Peter and Arata, $250,000 for construction finance supervision, and $150,000 for Leeway office space. For differing reasons, the government at trial challenged the legitimacy of these expenditures. For example, Seven Arts had supported the construction payments with a Duvernay-signed invoice that the company created only in anticipation of the Dienes audit. Duvernay testified that the fees were not actually paid to his company but that he signed the invoice anyway because Susan told him that the document "was just for [Peter's] own records." The request for legal fees shows that Arata was not completely out of the loop despite sending the letter. After receiving an invoice for the legal fees relating to 807 Esplanade, Arata sent one of his business partners an email saying, "[Peter] wants to submit this for tax credits. Ha!" He continued, "And since I was not his lawyer for the deal, it makes it even better. What he could submit and what is actual are the bills he got from Guy Smith, even the Jones Walker bills. But instead, he . . . puts me down as receiving $150K in fees! Love it."

After Peter submitted the second cost report, state officials asked forensic accountant Michael Daigle to analyze both rounds. As part of his investigation, Daigle contacted the Dienes firm about concerns he had developed. As a result of that interaction, the firm took the "very unique" step of recalling its audits associated with both cost reports. It recalled the first audit over Peter's objection. Withdrawing the first audit, he thought, would be "extremely damaging to the purchasers for value of the credits already certified." Those fears were not unfounded. After the Dienes firm withdrew its audits, the state revoked the previously issued credits, declined to issue new credits for the second cost report, and conveyed the problems unearthed during Daigle's investigation to the state inspector general.

The company's attempts to earn film tax credits on 807 Esplanade were thus battered by the waves of the Daigle investigation, the audit withdrawal, and the tax credit revocation. Nevertheless, Seven Arts persisted. By June 2012, 807 Esplanade was complete and the site functioned as a film production and postproduction facility. The company retained a new firm, Silva Gurtner & Abney LLC, to conduct an audit for a third cost report, this one covering October 2007 to June 2012. In other words, Seven Arts wanted to claim tax credits not only for the period after September 2009 but also for the time covered in the first two (rejected) cost reports. Of the $11,945,184 in claimed expenditures, the Silva firm deemed $11,785,934 "qualified." It even certified a number of expenditures that were similar or identical to those the state had rejected in the second cost report.

At the state's request, Daigle also conducted a forensic review of the company's third cost report. After reviewing the Silva audit, Daigle concluded that the company's qualifying expenditures totaled $2,743,319.18 by the end of 2008, which would mean maximum allowable project expenditures of $5,486,638.36 for tax credit purposes, per the limitations outlined in the state's precertification letter. Daigle cast doubt on the $3,842,355 in related party transactions contained in the Silva audit. Even excluding that amount, however, the company's total qualifying infrastructure expenditures—based on the acquisition and construction costs for 807 Esplanade—exceeded that maximum allowable amount, making it eligible for up to $2,194,655.34 in tax credits.³ Having apparently never faced a similar situation and relying on Daigle, the state decided to "reestablish" the tax credits issued after submission of the first cost report, thereby avoiding punishment of third-party purchasers of Seven Arts credits.

3 At trial, however, Daigle testified that his best estimate of qualifying expenditures for Seven Arts was roughly $4.2 million, which equates to allowable tax credits of about $1.6 million.

The state inspector general enlisted the help of the FBI and began investigating the company's tax credits. This led the Silva firm to withdraw, revise, and then reissue its July 2012 audit in order to disclose uncertainties about the legitimacy of certain expenditures.

D.

The joint state and federal investigation led to the filing of criminal charges. No model of restraint, the indictment contains 25 counts. It charges Peter with one count of conspiracy to commit mail and wire fraud, nineteen counts of wire fraud, and one count of mail fraud. It charges Arata with one count of conspiracy, nineteen counts of wire fraud, one count of mail fraud, and four counts of making false statements to the FBI. And it charges Susan with one count of conspiracy, fifteen counts of wire fraud, and one count of mail fraud.

During the two-week trial, the government sought to prove that the defendants fabricated invoices and shifted money in and out of accounts to make it appear as though Seven Arts had actually spent money on film infrastructure when it had not. The defendants countered that in the face of a difficult-to-interpret statutory regime they had made efforts to comply with state custom and practice as established by the acceptance of prior tax credit applications.

The jury did not buy that defense. It convicted Peter on all 21 counts. It convicted Arata of 13 counts—conspiracy, seven counts of wire fraud, one count of mail fraud, and four counts of making a false statement. Reflecting that Susan's name was "scarcely mentioned" during the trial, the jury found her guilty only of one count each of conspiracy, wire fraud, and mail fraud.

The defendants moved for judgments of acquittal. In a lengthy opinion, the district court granted Peter's motion with respect to five counts of wire fraud (Counts 2, 3, 4, 5, and 7) but denied the remainder; granted Arata's motion with respect to all but the conspiracy count (Count 1) and one count of wire fraud (Count 6); and denied Susan's motion. The district court then denied defendants' motions for new trial, both with respect to their remaining convictions and for all counts in the event that this court were to reverse the acquittals.

The district court imposed sentences far below those suggested by the Sentencing Guidelines. The Guidelines recommended sentences of roughly 14 to 17 years for Peter, 9 to 11 years for Arata, and 4 to 5 years for Susan. But the district court placed all of them on probation—five years for Peter,⁴ four for Arata, and three for Susan.

4 At the sentencing hearing, the district court sentenced Peter to six years' probation. The judgment shows a sentence of five years, which is the statutory maximum. 18 U.S.C. § 3561(c)(1).

The government also sought forfeiture of the issued tax credits and restitution on behalf of the state. The district court ordered forfeiture in the amount of $223,434.25. But in a ruling not challenged on appeal, it denied the government's motion for restitution because the state, in its view, ended up suffering no "actual, pecuniary loss." Even if it had initially suffered a loss in issuing tax credits due to fraud, the court concluded the state did not ultimately lose money because Seven Arts eventually made infrastructure expenditures on 807 Esplanade entitling the company to an amount of credits at least equal to those issued.

II.

The parties raise numerous issues in their cross appeals. We begin with the one that would wipe away all the conspiracy and fraud counts: defendants' contention that the Louisiana tax credits are not "property" covered by the federal fraud statutes. Their vehicle for raising this issue was a motion to dismiss the indictment, see FED. R. CRIM. P. 12(b)(3), the denial of which we review de novo, United States v. Cooper, 714 F.3d 873, 876-77 (5th Cir. 2013).

The mail and wire fraud statutes, which have the same elements other than the jurisdictional hook of the mailing or interstate wire, criminalize schemes "to defraud, or for obtaining money or property by means of false or fraudulent pretenses, representations, or promises." 18 U.S.C. §§ 1341, 1343. Property, as ordinarily understood, extends to every kind of valuable right and interest. See Pasquantino v. United States, 544 U.S. 349, 356 (2005) (citing Leocal v. Ashcroft, 543 U.S. 1, 9 (2004)). Under the common law of fraud, and the even more venerable law of common sense, "[t]he right to be paid money has long been thought to be a species of property." Id. at 356 (citing BLACKSTONE, COMMENTARIES ON THE LAWS OF ENGLAND 153-55 (1768)). Common law fraud encompassed both defrauding a victim of money and of her entitlement to that money because of the "economic equivalence between money in hand and money legally due." Id. That the victim happened to be the government, instead of a private party, did not negate that economic injury. Id.

The Supreme Court set forth these principles in considering whether a scheme to defraud Canada of excise tax revenue by smuggling liquor into the country violated the wire fraud statute. Id. at 353. By evading taxes that would have been due had the liquor imports been declared, the defendants inflicted a "straightforward" economic injury akin to "embezzl[ing] funds from the Canadian treasury." Id. at 356-57. Indeed, a country "could hardly have a more 'economic' interest than in the receipt of tax revenue." Id. at 357. Smuggling goods to deprive a government of tax revenue via a fraudulent scheme that used interstate wires was thus held to constitute wire fraud. Id. at 357. Although Pasquantino involved depriving a foreign government of tax revenue, prosecutors have also successfully used the mail and wire fraud statutes against schemes to defraud state and local governments of tax revenue. See Fountain v. United States, 357 F.3d 250, 260 (2d Cir. 2004) (deeming taxes owed to states and the federal government property within the meaning of the mail and wire fraud statutes); see also United States v. Louper-Morris, 672 F.3d 539, 557 (8th Cir. 2012); United States v. Frederick, 422 F. App'x 404, 405 (6th Cir. 2011) (both involving schemes to defraud states of tax revenue); Matthew D. Lee, Chicago Restaurant Tax Case Highlights Broad DOJ Authority, LAW360 (May 25, 2016), https://www.law360.com/articles/800503/chicago-restaurant-tax-case-highlights-broad-doj-authority (discussing case in which restaurant owner pleaded guilty to wire fraud for failing to pay state taxes on cash transactions); cf. Hemi Grp., LLC v. City of New York, 559 U.S. 1, 4 (2010) (evaluating a suit in which New York City brought RICO charges, based on predicate acts of mail and wire fraud, because defendant allegedly caused the loss of "tens of millions of dollars in unrecovered cigarette taxes").

From Pasquantino's holding that tax revenue is property under the fraud statutes, it follows that Louisiana's tax credits can also be the object of a scheme to defraud. As tax credits reduce the dollars otherwise owed to the state, lying to obtain them has the same effect as lying to evade taxes: the state collects less money. Indeed, the drain on Louisiana finances caused by the film tax credit regime—$282.6 million in just one year (2016)—led the state to curtail the program. Tyler Bridges, New Study of Louisiana Film Tax Credit Program Again Finds Expensive, "Significant Hit" to Budget, ADVOCATE (Apr. 10, 2017).⁵ Fraud in connection with obtaining those tax credits can affect the state's books as much as fraud used to evade paying Louisiana income taxes. Either situation implicates the state's interest in taxes owed that Pasquantino recognizes as property.

5 Tax breaks on three projects alone—The Green Lantern, The Twilight Saga: Breaking Dawn, and HBO's True Detective—cost the state nearly $85 million. Louisiana's Film Tax Credit Program to Continue, with a Cap, supra note 1. At least the Louisiana season of True Detective was the good one.

Tax credits are also the functional equivalent of government spending programs. See Drew Desilver, The Biggest U.S. Tax Breaks, PEW RES. CTR. (Apr. 6, 2016), http://www.pewresearch.org/fact-tank/2016/04/06/the-biggest-u-s-tax-breaks/ ("[S]uch special-purpose breaks are effectively the same as directing spending"). That is why economists treat tax deductions and credits as "tax expenditures." See Tax Policy Center, Briefing Book: A Citizen's Guide to the Fascinating (Though Often Complex) Elements of the Federal Tax System, BROOKINGS INSTITUTION, http://www.taxpolicycenter.org/briefing-book/what-are-tax-expenditures-and-how-are-they-structured. Viewing tax credits in this light further highlights their economic impact. Consider one of the largest tax expenditures in the federal tax code, the home mortgage interest deduction which totaled $77 billion in 2016. Desilver, supra. The impact on the government's coffers would be the same if, instead of offering that deduction, it sent taxpayers $77 billion in grants to help them pay their home loans. As defendants conceded at oral argument, fraud in connection with obtaining a state government grant is undoubtedly subject to wire fraud prosecution. Because there is no bottom-line difference between a government spending program and a tax credit, there is no economic rationale for treating the former as property but not the latter. When it comes to depriving the government of revenue—property under Pasquantino—there thus is no meaningful distinction between fraudulently claiming a tax credit, fraudulently obtaining a public grant, or fraudulently failing to report income.

The congruity of these three situations involving the public fisc is further evident from looking to an example from the private sector. Everyone would recognize that plane tickets are property of an airline. That means obtaining them via deceit is fraud. See United States v. Morris, 348 F. App'x 2, 3-4 (5th Cir. 2009) (discussing wire fraud conviction of an airline employee who fraudulently issued 1,011 tickets and sold them for her benefit). But so too, we have recognized, is swindling reward miles that can be redeemed for free flights. United States v. Loney, 959 F.2d 1332, 1336 (5th Cir. 1992); see also United States v. James, 616 F. App'x 753, 755 (5th Cir. 2015) (affirming wire fraud conviction for "discount fraud" that allowed defendant to purchase less expensive computers). The reason is that revenue lawfully owed the airline is taken in both situations. Loney, 959 F.2d at 1336-37; cf. Felder's Collision Parts, Inc. v. All Star Advertising Agency, Inc., 777 F.3d 756, 763 (5th Cir. 2015) (reducing a seller's revenue by the amount of a rebate in a predatory pricing case). A tax credit is the public sector equivalent of a coupon; it reduces the amount that is otherwise owed.

In an attempt to avoid these basic economic principles, the defendants invoke Cleveland v. United States, 531 U.S. 12 (2000). It does not give us much pause. Another federal fraud prosecution out of Louisiana, Cleveland involved misrepresentations on applications for state video poker licenses. The Court held that the license was not property in the regulator's hand. Id. at 20. It rejected the argument that a state's "intangible rights" to decide who is eligible to operate poker machines created a property interest; that interest "amount[ed] to no more and no less than Louisiana's sovereign power to regulate." Id. at 23. As for the government's attempt to fit the licenses into the traditional category of an economic property interest, it could not show any financial harm resulting from the effort to trick the state into issuing a license. Id. at 22 ("Tellingly . . . the Government nowhere alleges that Cleveland defrauded the State of any money to which the State was entitled by law."). Quite the opposite in fact: the company that misrepresented its eligibility for the license paid the state more than $1.2 million. Id. So unlike lies to obtain tax credits, Cleveland's lies to establish eligibility for the poker license generated revenue for Louisiana even though they resulted in the regulatory harm of allowing those deemed unworthy to operate the machines. Cleveland's rejection of that regulatory harm as property does not undermine the conclusion that the drain on a state's treasury resulting from schemes to unlawfully obtain tax credits deprives the state of a classic property interest. See, e.g., Louper-Morris, 672 F.3d at 557 (affirming mail and wire fraud convictions involving scheme to defraud Minnesota of education tax credits); United States v. Lefkowitz, 125 F.3d 608, 614, 617 (8th Cir. 1997) (affirming mail and wire fraud convictions for a scheme to falsely obtain tax credits for low-income housing); Frederick, 422 F. App'x at 405 (addressing mail fraud prosecution for scheme to obtain Michigan Homestead Property Tax Credits).

A case we decided after Cleveland does seem closer to this one at first blush because it involves tax credits. See United States v. Griffin, 324 F.3d 330, 354 (5th Cir. 2003). Griffin held that "unissued" federal tax credits were not property of a state agency under the mail and wire fraud statutes. Id. at 355. But the unique nature of the program it considered, in which the state merely allocated federal tax credits, means no state property was at risk. The state agency, the Texas Department of Housing and Community Affairs, did not have a property interest in the tax credits that offset federal income tax obligations. Id. at 338. Under that program, the federal government allotted a certain amount of tax credits to Texas; the state housing agency's job was to then assign those credits to low-income housing developments within the state. Id. at 338, 354. The fraud arose in connection with a preapplication to the state agency seeking an allocation of some of the credits. Id. at 352-54. The credits would not actually issue until years later, if and when the project was completed. Id. at 355. We emphasized this feature of the Griffin fraud—that it did not result in the issuance of any tax credits, only an allocation of them. Id. at 354-55. We also noted the more fundamental point that even if the credits had issued, their fraudulent issuance would not have caused economic harm to Texas because the credits "offset [] federal income tax obligations." Id. at 355.

Unraveling the cooperative federalism arrangement in Griffin shows that it follows directly from Cleveland. The state's role as an allocator of federal tax credits meant it was acting much like the licensor in Cleveland: deciding which applicants would best serve the state's regulatory interests, decisions that did not directly implicate the state's finances. If anything, as in Cleveland the fraud in Griffin netted money for the state because the company receiving the allocation had to pay an application fee and a $40,000 commitment fee.⁶ Id. at 340, 355. Griffin thus rightly recognized that the fraud to obtain an allocation of federal tax credits could not have deprived Texas of property.⁷

6 As in Cleveland, prosecutors argued that Texas had a property interest because the conduct it was approving would provide economic benefits, such as the application fees, to the state. Compare Griffin, 324 F.3d at 355, with Cleveland, 531 U.S. at 21-22. This missed the fundamental point that the fraudulent conduct must deprive the victim of property, not provide it with property.

7 Griffin addressed only whether the state housing agency had a property interest in the credits. 324 F.3d at 354-55. The better argument would have been that the federal government had a property interest in those credits. Indeed, Griffin did not disturb the conviction for conspiracy to steal federal funds under 18 U.S.C. § 666, which applies to theft of federal "property." Id. at 345-46. The indictment had also listed the United States as one of the victims of the mail fraud, see id. at 352, but for whatever reason the government only defended that conviction in our court on the ground that Texas had a property interest in the credits, id. at 353-55.

Griffin does not provide a defense against this prosecution because the film tax credits do reduce state coffers. And the scheme alleged here did not end with misrepresentations in connection with obtaining precertification for the credits. It continued with falsehoods in the three Seven Arts cost reports, which caused Louisiana authorities to certify and actually issue transferable credits. Because Louisiana was administering its own tax credits, the fraudulent issuance of those credits would deplete the state treasury.⁸ That means Louisiana has a property interest in the tax credits. Stealing them via fraud has the same economic effect on the state as "embezzle[ing] funds from the [] treasury." Pasquantino, 544 U.S. at 356.

8 That Seven Arts completed the infrastructure work at a later date and might have been entitled to the credits then—the basis for awarding no restitution—does not provide a defense to mail fraud. The scheme to defraud need not result in loss to the victim. United States v. McMillan, 600 F.3d 434, 450 (5th Cir. 2010). That is because what is unlawful is engaging in the scheme to defraud, even if it turns out to "be absolutely ineffective." Durland v. United States, 161 U.S. 306, 315 (1896).

We also reject the defendants' argument that application of the wire and mail fraud statutes to Louisiana's film tax credit program raises unprecedented federalism or due process concerns. As to the federalism issue, defendants concede that these federal statutes can combat fraud in connection with evading state taxes or obtaining state benefits. We do not see how state tax credits raise any greater concerns about federal intrusion in state policymaking than those far more prevalent traditional state tax and spending programs. Regulatory complexity is not limited to tax credits. And recourse to federalism is not a great fit with this case. The state did not indicate that it thought the defendants' creation of false invoices and use of circular transactions was allowed under state law. To the contrary, it sought the assistance of federal law enforcement to investigate potential crimes, which made sense as complex interstate schemes (the Hoffmans resided in California) are one of the more strongly rooted bases for federal criminal law.

This prosecution also does not raise notice concerns under the Due Process Clause. The honest services aspect of mail fraud has given rise to vagueness challenges. See, e.g., Skilling v. United States, 561 U.S. 358, 367 (2010) (construing the honest-services statute beyond its "core meaning . . . would encounter a vagueness shoal"). But the classic property conception of fraud has not. See Daniel W. Hurson, Comment, Mail Fraud, the Intangible Rights Doctrine, and the Infusion of State Law: A Bermuda Triangle of Sorts, 38 HOUS. L. REV. 297, 303-10 (2001) (contrasting prosecutions for schemes "whose purpose was to deprive another of money or property," a "basic purpose[]" of the mail fraud statute since its inception, with courts' long struggle to define schemes that deprive another of intangible rights); cf. Skilling, 561 U.S. at 412 ("As to fair notice, whatever the school of thought concerning the scope and meaning of [scheme or artifice to defraud], it has always been as plain as a pikestaff that bribes and kickbacks constitute honest-services fraud." (quoting Williams v. United States, 341 U.S. 97, 101 (1951)) (cleaned up)). That is because lying to cheat another party of money has been a crime since long before Congress passed the first mail fraud statute making it a federal offense in 1872. Courtney Chetty Genco, Note, What Happened to Durland?: Mail Fraud, RICO, and Justifiable Reliance, 68 NOTRE DAME L. REV. 333, 337, 345-47 (1992) (identifying the common law crime of "cheating" as a precursor to mail fraud). Although defendants focus on a lack of clarity in the administration of Louisiana's tax credit program, vagueness challenges look to whether the elements of the offense provide sufficient notice. See Connally v. Gen. Constr. Co., 269 U.S. 385, 391 (1926). The government did not have to prove violations of state law. United States v. Foshee, 606 F.2d 111, 113 (5th Cir. 1979). The elements the jury had to find included terms like misrepresentations and property that have deep roots in both criminal and civil law. As we once stated, fraud "needs no definition; it is as old as falsehood and as versable as human ingenuity." Weiss v. United States, 122 F.2d 675, 681 (5th Cir. 1941). Defendants point to no court that has held that the elements of property-based mail fraud are vague, and we see no basis for being the first to do so.

The district court correctly found the tax credits are property subject to prosecution under the mail and wire fraud statutes. This prosecution alleging the use of fabricated invoices and misleading bank transactions to obtain a financial benefit lies at the historic core of the federal fraud statutes and neither offends due process nor exceeds federal power.

III.

Having rejected the defendants' global challenge to the prosecution's theory, we consider their fact-based challenges to the specific counts of conviction. But our sufficiency review does not just entail the usual posture of a defendant seeking to set aside convictions. Because the district court granted judgment of acquittals on a number of counts—five for Peter and eleven for Arata—the government also appeals, seeking reinstatement of those convictions that it believes the evidence supported.⁹ Whether we are looking at the verdicts the district court sustained or those it threw out, our standard of review is the same. We conduct a de novo review of the evidence in determining whether it was sufficient to convict. See United States v. Danhach, 815 F.3d 228, 235 (5th Cir. 2016). In conducting that review, we weigh the evidence "in a light most deferential" to the jury verdict and give the party that convinced the jury the benefit of all reasonable inferences. United States v. Lucio, 428 F.3d 519, 522 (5th Cir. 2005); see United States v. Ingles, 445 F.3d 830, 834-35 (5th Cir. 2006). Consequently, we "must affirm the verdict unless no rational juror could have found guilt beyond a reasonable doubt." United States v. Sanjar, 876 F.3d 725, 744 (5th Cir. 2017).

9 We perform our duty and review all of the acquitted counts the government appeals. We note, however, that a successful appeal will have no practical effect for most of the counts. This is especially true when it comes to Peter. Because his Guidelines range already captured the full amount of intended loss in this scheme and any other conceivable enhancements, reinstating some convictions would not change Peter's range. So what is the point of trying to convict him of 21 counts? Doing so is inconsistent with DOJ policy. The U.S. Attorneys' Manual counsels that to "promote the fair administration of justice, as well as the perception of justice" prosecutors should charge "as few separate counts as are reasonably necessary"—it sets a default ceiling of 15—so long as that does not jeopardize a successful prosecution or prevent the court from fully capturing a defendant's sentencing exposure. U.S. DEP'T OF JUSTICE, UNITED STATES ATTORNEYS' MANUAL: CRIMINAL RESOURCE MANUAL § 215 (1997). A single count of wire fraud encompasses Peter's Guideline range in allowing a sentence up to 20 years in prison. 18 U.S.C. § 1343.

A.

In assessing the sufficiency of the evidence, we first discuss conspiracy, then mail and wire fraud, and finally false statements.

Count 1: The jury convicted all three defendants of conspiracy to commit mail and wire fraud. As this offense was charged under the general conspiracy statute (18 U.S.C. § 371) rather than the one specific to fraud offenses (18 U.S.C. § 1349),¹⁰ the government had to prove an agreement to commit the fraud offense, the defendants' knowledge of the unlawful objective and willful agreement to join the conspiracy, and an overt act by a member of that conspiracy to further the unlawful goal. United States v. Mauskar, 557 F.3d 219, 229 (5th Cir. 2009). The district court upheld the conspiracy convictions. We too are of the opinion that the direct and circumstantial evidence was sufficient to prove the existence of an agreement to defraud the state of film infrastructure tax credits, the commission of overt acts meant to further that scheme, and the willful involvement of each defendant.

10 The substantive difference is that the section 1349 conspiracy does not require an overt act. Sanjar, 876 F.3d at 737. It also carries a punishment of twenty years as opposed to the five years of the general conspiracy statute. Compare 18 U.S.C. § 1343, with 18 U.S.C. § 371.

Peter's contribution is the most straightforward as he was involved in the creation and dissemination of all three Seven Arts cost reports. Emblematic of that involvement is his role in substantiating Seven Arts' expenditures related to construction and equipment. As one example, Peter and Arata opened bank accounts for Duvernay (construction) and Martin (equipment) into which Seven Arts supposedly made "payments." But bank records indicate that those payments were almost immediately returned to Seven Arts—that is, they were not really payments at all. To memorialize the "payments," Peter created invoices showing about $2 million in construction and just over $1 million in equipment costs, and he convinced Duvernay and Martin to sign them even though they had not been paid anything close to the listed amounts. Duvernay signed the invoice because Peter "convinced [him] to sign it, saying it was just for his records," while Martin testified that the equipment described in his invoice was just a "dream list" created at the behest of Seven Arts. Another example of Peter's steering of the scheme comes from the confusion over legal fees during the preparation of the second cost report. The company's auditor, after conferring with Arata, told Peter that she was removing over $200,000 in unsupported legal fees. Peter objected and sent Arata an email urging him to send the auditor the SAFELA operating agreement as proof of the fees so she would not "get any more suspicious." The agreement supposedly showed that Arata was paid for his legal work by giving his company Voodoo an equity interest in SAFELA. But as the jury knew and we have already mentioned, Arata told his business partner that he "was not the lawyer for the deal" and laughed at the notion that Peter wanted to submit his legal fees for tax credits. This is just a sampling of the abundant evidence that allowed the jury to conclude that Peter was part of, indeed the leader of, the fraud conspiracy.

Arata's case is more complicated, in part because as we discuss below the district court concluded he withdrew from the conspiracy after the submission of the company's first cost report. But as support for the jury's conclusion that he joined the conspiracy, Arata's fingerprints are all over that first report. In anticipation of the submission, Arata and Peter opened the Duvernay and Martin bank accounts. To facilitate circular transactions using those accounts, Arata took out a $400,000 loan through another of his businesses and put that money into the Seven Arts account; that money then was bounced between the Seven Arts and Duvernay/Martin accounts to make it appear as though the company had made payments for construction and equipment costs when it had not. The Seven Arts bank statements make clear that these circular transactions constituted "both withdrawals and deposits," but that was not reflected in the company's general ledger, which would form "the building block of an audit." The ledger instead showed the deposits as capital contributions from the parent company of Seven Arts. Arata and Peter provided that ledger, along with Seven Arts-generated invoices and confirmation of the supposed capital contribution, to an auditor. This is enough—and there is more—to support the jury's view that Arata was part of the conspiracy.

As suggested by her conviction on just one count of wire fraud and one of mail fraud, Susan's involvement in the tax credit scheme is less apparent. Although witnesses did not focus on her, some evidence of her knowing participation comes from a December 2009 certification she signed in her capacity as president of Leeway Properties. It says that Seven Arts paid her company $700,000 for, among other things, management fees and office space. The jury was entitled to view the $400,000 of that labelled a project management fee as a fiction. The state's forensic auditor Michael Daigle asked Susan a series of questions about her scheduling and budgeting responsibilities on the project—responsibilities that could justify the management fee—but she could not provide answers. As far as he could tell, her duties were limited to "interior design decisions on painting and carpeting and things like that." Though Daigle acknowledged that $400,000 may under some circumstances be a reasonable management fee "that was clearly not the case here." Yet Susan signed an affidavit in 2009 in which she claimed to have spent in excess of 1,000 hours supervising construction at 807 Esplanade.¹¹ The same inflated billing can be seen for the $150,000 that Seven Arts supposedly paid Leeway for office space. Peter said that the space, which rented for $4,250 a month, included an office for Duvernay and a courtyard that was roughly three times the size of the office, where construction materials were kept. But Duvernay testified that he worked for Susan on her 900 Royal Street property, renovating its upper levels, and used a space at that location as an office. He described it as a "12-by-12 room. That was it." By the time of his testimony, Duvernay had rented a 300 or 350 square foot office—as opposed to the roughly 150 square foot office at 900 Royal—from Susan at 906 Royal Street for $600 per month. Daigle thought the office space rental "transaction lacked economic substance." And given her later rental arrangement with Duvernay, Susan should have recognized that as well. A rational juror could infer that $400,000 and $150,000 were not reasonable sums for management fees and office space, respectively, and that Susan knew as much when she certified these expenses.

11 Even that affidavit somewhat contradicted a 2011 email in which Susan listed, albeit from memory, a "very rough account" of time spent on the project that amounted to about 500 hours.

We therefore uphold defendants' conspiracy convictions. This decision has ramifications for the fraud counts we are about to discuss. As the district court recognized, this evidence showing willful participation in the conspiracy to commit wire and mail fraud also establishes the intent to defraud necessary for the substantive fraud offenses. That intent does not evaporate because the project might have later spent money that made it eligible for the tax credits at a subsequent point in time. If the defendants intended to submit false cost reports to obtain property they were not then entitled to—and they did—then they engaged in fraud. To illustrate this point, consider a teller who embezzles from the bank. If the teller plans to pay the money back a year later, that does not mean there was not intent in the first place to deprive the bank of its property. Contingencies are just that; future plans to make a victim whole do not mean a crime was not committed (later conduct that makes the victim whole can be a mitigating factor at sentencing as we later discuss).

An even more significant consequence of upholding the conspiracy convictions is Pinkerton liability. Because the court gave a co-conspirator liability instruction, any of the three conspirators is liable for any acts of mail and wire fraud committed during the conspiracy that were foreseeable and that furthered the agreement. Sanjar, 876 F.3d at 743 (citing Pinkerton v. United States, 328 U.S. 640 (1946)). Those two conditions of Pinkerton liability do not add much if anything to what the conspiracy and fraud offenses already require. With an established agreement to commit mail and wire fraud, it is going to be foreseeable that mail and wire fraud might occur. Id. at 743-44 (explaining that foreseeability is usually not disputed when "Pinkerton liability is extending only to the substantive offense that is the object of the conspiracy"). As for the requirement that the substantive offense further the conspiracy, mail and wire fraud have a built-in element requiring that the specific act charged furthered the scheme. 18 U.S.C. §§ 1341, 1343.

The district court relied on this co-conspirator liability to uphold some of the fraud convictions. But it relied on another principle of conspiracy law— that a conspirator can withdraw from the enterprise—in refusing to do so for Arata on the counts occurring in the later stages of the conspiracy. This ruling arose in an unusual, perhaps even novel, posture. Withdrawal from a conspiracy is an affirmative defense on which the defendant bears the burden of proof (by a preponderance). Smith v. United States, 568 U.S. 106, 112-13 & n.5 (2013); United States v. Heard, 709 F.3d 413, 427-28 (5th Cir. 2013). The defendant typically, if not always, puts this issue before the jury by requesting an instruction on withdrawal. This is done often enough to warrant a pattern charge in our circuit. FIFTH CIRCUIT PATTERN JURY INSTRUCTIONS (CRIMINAL) § 2.18. Arata did not request that instruction or otherwise argue withdrawal at trial. He did not even argue it in his post-trial motion for acquittal. There was a suggestion of withdrawal by Arata's counsel at the hearing on that motion, but even that was not an express claim of withdrawal. Despite the argument not being raised at trial or in post-trial motions, the district court ruled that Arata withdrew when he sent the August 6, 2009 letter to Peter ending his day-to-day involvement in the project. Notably, the district court treated itself as the factfinder on withdrawal, framing the issue as whether "[t]he preponderance at trial proved that Mr. Arata terminated his relationship with Mr. Hoffman after suspecting him of fabricating invoices in July or August 2009."

At a minimum, the district court should have evalauted the withdrawal defense with the deference that would have been required had the jury rejected it, which its verdict implicitly did. There is an argument that Arata's failure to seek a withdrawal instruction or otherwise raise the issue at trial forfeited his ability to use the theory to limit his conspiracy offense found by the jury. That is what typically happens when an affirmative defense is not timely asserted. See Biddinger v. Comm'r of Police of New York, 245 U.S. 128, 135 (1917) ("The statute of limitations is a defense and must be asserted on the trial by the defendant in criminal cases . . . ."); United States v. Bey, 725 F.3d 643, 646 (7th Cir. 2013) (finding that defendant waived entrapment for appellate review after he withdrew his proposed jury instruction on it); United States v. Haney, 318 F.3d 1161, 1163 (10th Cir. 2003) (finding that defendant waived any claim that his conviction should be overturned for lack of a duress instruction because, in part, he did not raise that defense during trial). As withdrawal from a conspiracy is an affirmative defense, it is typically governed by the procedural rules governing such defenses. Smith, 568 U.S. at 113 (relying in the withdrawal context on the common-law rule that the defendant bears the burden of proving an affirmative defense). But we need not decide whether Arata forfeited the withdrawal defense. Even if he did not, his failure to ask for a withdrawal instruction cannot put him in a better position to undo the jury's verdict than he would be in had he requested it. Given the absence of any jury determination that Arata left the conspiracy, we can overrule part of the verdict and find withdrawal only if Arata can show that is the only reasonable view of the evidence. Cf. United States v. Barton, 992 F.2d 66, 70 (5th Cir. 1993) (emphasizing, for a case in which insanity was raised as a defense, the deference a reviewing court gives "where the jury has found against a party having the burden of proof by clear and convincing evidence").

It is not. Withdrawal requires a deliberate attempt to disassociate from the unlawful enterprise. Heard, 709 F.3d at 428. Perhaps a factfinder could have found that intent and action in the letter Arata sent Peter. But that is not the only reasonable way to view it. Although Arata sent the letter and reduced his participation after that point, he did not completely abandon ship. He did not even cease all direct involvement in the fraudulent aspects of the business. Sending the letter did not stop Arata five months later from helping Peter at a critical stage. Arata complied with Peter's request and sent an email to the auditor attaching the SAFELA operating agreement showing Voodoo's 40% stake to verify the supposed fees he had been paid for legal work. From this a factfinder was free to conclude that Arata had not left the conspiracy by "disavow[ing] or defeat[ing]" its purpose but was instead continuing to help it along even if only from the sidelines. Smith, 568 U.S. at 113 (quoting Hyde v. United States, 225 U.S. 347, 369 (1912)); see also id. at 112-13 (noting that even "[p]assive participation in the continuing scheme is not enough to sever the meeting of the minds that constitutes conspiracy"). Because the evidence did not compel a finding that Arata withdrew, there was no basis to disrupt the jury's verdict that he was a full-fledged conspirator. The jury verdicts against him on the fraud counts can thus be sustained under Pinkerton if one conspirator committed the individual offense.

B.

To prove those fraud offenses, the government had to show (1) a scheme to defraud that employed false material representations, (2) the use of mail or interstate wires in furtherance of the scheme, and (3) the specific intent to defraud. See United States v. Kuhrt, 788 F.3d 403, 413-14 (5th Cir. 2015); United States v. Brooks, 681 F.3d 678, 700 (5th Cir. 2012). One aspect of these elements is the basis for a number of our rulings below, so it is worth explaining now. The mailing or wire need not contain a falsehood. That act, which serves as the unit of prosecution, just needs to further the fraudulent scheme. Judge Brown explained the point this way six decades ago:

The thing sent through the mails need not, as impliedly urged, be a cunning deceptive appeal which causes another to give up money or property. It can be, and frequently is, a wholly innocent thing or innocuous in itself, such as the deposit of a check, transmission of a check from a collecting to a drawee bank, or the like. The thing

which is condemned is (1) the forming of the scheme to defraud, however and in whatever form it may take, and (2) a use of the mails in its furtherance. If that is satisfied, more is not required.

Gregory v. United States, 253 F.2d 104, 109 (5th Cir. 1958); accord United States v. Martin, 228 F.3d 1, 16 (1st Cir. 2000); United States v. Green, 786 F.2d 247, 249 (7th Cir. 1986). So while the mailing or wire must promote the scheme in some manner, it need not contain a falsehood. See United States v. Tencer, 107 F.3d 1120, 1125 (5th Cir. 1997) ("Even a routine or innocent mailing may supply the mailing element as long as it contributes to the execution of the scheme."). An interstate email that says "Meet me at the bowling alley tonight" can serve as the necessary wire if the parties planned the fraud while bowling a few frames that evening.¹²

12 This principle means not much is needed to multiply wire fraud counts once the government has proven the scheme to defraud with its requisite intent. With today's rampant use of email and other technology that often crosses state lines, it will usually not be hard to identify scores of wires that further a scheme. Then again, adding all these counts packs little additional punishment punch—one count of wire fraud already allows a sentence up to 20 years—so there will rarely be a reason to go overboard. See supra note 9.

With this understanding of the limited role of the "in furtherance" requirement, we consider the specific counts.

Counts 2, 4, and 7 (wire fraud): The jury convicted Peter and Arata of wire fraud on Counts 2, 4, and 7. Each count concerns an email Arata sent from a Yahoo account either to auditors at the Dienes firm or to state officials in support of the company's first cost report. Both the senders and recipients resided in Louisiana. The district court overturned these convictions because it did not believe there was evidence to establish that the emails travelled outside the state. We disagree.

To prove that the email crossed state lines, the government called Yahoo paralegal Sherry Hoyt. When Hoyt was asked whether Yahoo had any email servers in Louisiana between 2008 and the present, she responded "No." When asked whether an email would have to leave the state if it was sent from someone in Louisiana using a Yahoo account to someone else in Louisiana, she responded "Yes." Defense counsel did an effective job on cross of showing the limits of Hoyt's technical knowledge. For example, when Peter's counsel asked whether Hoyt had any training in email message routing, she responded "No." And when Arata's counsel asked whether an email from a Yahoo account to a non-Yahoo account could be routed through a non-Yahoo server located in Louisiana, Hoyt responded "I don't know."

This impeachment could have led jurors to conclude the government did not prove the interstate nexus. But that did not happen, and we cannot displace the jury's contrary credibility determination. That is what the finding of an interstate email amounted to. The jury heard direct testimony that Yahoo emails had to leave the state. If they believed Hoyt, the wire element was established because the testimony of a single witness is sufficient proof of a fact. United States v. Bowen, 818 F.3d 179, 186 (5th Cir. 2016). That is true even when that testimony is from an accomplice testifying in exchange for a benefit, testimony the jury is told must be viewed with "caution" and "great care." Id.; see FIFTH CIRCUIT PATTERN JURY INSTRUCTIONS (CRIMINAL) § 1.14. Hoyt had no comparable incentive to lie, and the gaps in her knowledge did not make it "incredible" to believe her testimony, especially when no contrary evidence was presented at trial (an attorney's unadopted question is not evidence, so there was no evidence that the email could have been routed intrastate). Bowen, 818 F.3d at 186 (explaining that a jury's acceptance of a cooperator's testimony should be rejected only if it was "incredible"). The verdicts on these three counts will be reinstated.¹³

13 These emails also clearly furthered the scheme to defraud Louisiana. Count 2 is an email to the company's auditor containing a payment certification related to construction costs, Count 4 sent the company's first cost report to Louisiana, and Count 7 is the subsequent tender to the state of documentation supporting claimed equipment expenditures.

Count 3 (wire fraud): The jury convicted Peter and Arata of wire fraud on Count 3, which is a February 25, 2009 email sent by Arata to Peter attaching the Seven Arts general ledger. There was no interstate problem with this wire; it was sent from Arata in Louisiana to Peter in California. The district court instead saw a problem with the "in furtherance" requirement. Because the ledger had previously been sent to the Dienes firm and Peter already had it in his possession, it concluded the email "in no way sought to further the scheme." Review of this count thus involves the "in furtherance of" requirement that we have already explained is what connects the jurisdictional act of sending a wire to the fraud. The use of the wires "need not be an essential element" of the scheme; it can further the fraud as long as it is "incident to an essential part of the scheme, or a step in the plot." United States v. Dowl, 619 F.3d 494, 499 (5th Cir. 2010) (quoting Schmuck v. United States, 489 U.S. 705, 710-11 (1989)).

To achieve this scheme's goal of swindling the tax credits, the defendants had to submit cost reports, audit information, and supporting documents to state authorities. Those submissions were thus an essential part of the scheme even though that is not what the law requires for the wire. The email that is the subject of Count 3 was a step in verifying those critical submissions before they were sent. In it, Arata wrote Peter, "We should go through [the Seven Arts ledger] carefully and make sure they are capturing all of the expenses." Of course, the more expenses that were captured, the larger the tax credit. The cost report was not submitted until the next day, February 26. Because this email was sent to guarantee that Arata and Peter maximized the expenditures that would be submitted to the state, the jury's finding of guilt on Count 3 was proper.

Count 5 (wire fraud): Count 5 concerns an email Arata sent to auditor Davis and Peter, attaching the "executed SAFELA Operating Agreement evidencing Voodoo's 40% interest in this entity." Arata closed the email by saying he hoped the operating agreement "helps you and Peter wrap up the [Seven Arts] audit" for the second cost report. The district court acquitted Arata of this count on the ground that there was no actual evidence, only speculation, of his intent to defraud in sending the operating agreement. Regarding Peter, it held that the government produced no evidence that the legal fees were actually improper, which implies there was no "material falsehood."

The district court again put more weight on the "in furtherance" requirement than it has to carry. There need not be intent to defraud particular to each wire but only with respect to the overall scheme. See Tencer, 107 F.3d at 1125; Gregory, 253 F.2d at 109. The district court found that required intent for both Arata and Peter in refusing to acquit them on at least some of the fraud counts as well as the conspiracy count. But it imposed an unnecessary element in requiring that the particular attachment to the email evince fraud. To illustrate this point, an email with no attachment that only said "Please finish your review of the operating agreement so we can wrap up the audit" would be one that furthered the scheme, as it would be a step toward filing the cost report. As this is all that the law requires for the wire, this email that also had an attachment was a step in the successful execution of the scheme because it put the conspirators one step closer to completing the audit. The government did not have to prove that the legal fees standing alone were false or fraudulent.¹⁴ Because the email in Count 5 furthered the fraud scheme, we reverse the postverdict acquittal.

14 In any event, we note there was evidence to that effect. See supra page 9-10 and infra page 39.

Count 6 (wire fraud): We now arrive at a fraud count on which the district court upheld a guilty verdict. It is an email from Seven Arts employee Mark Halvorson to a state official that contained construction and film equipment invoices related to the company's first cost report.¹⁵ The district court concluded that the expenditures claimed in the email revealed an intent to defraud. We agree as there was evidence of circular transactions between accounts that had no legitimate business purpose yet made it look like "payments" for construction work had been made. But the bigger point is the one we are repeating: the wire need not be independently fraudulent to further the overall fraud. We affirm the district court's refusal to acquit Peter and Arata on this count.

15 Another part of the law on mailings and wires is that the government need not prove that defendants personally used or intended the use of those communications. It is enough that they knew the use would follow in the "ordinary course of business" or could reasonably be foreseen. United States v. Ingles, 445 F.3d 830, 835 (5th Cir. 2006). The attachments in Halvorson's email supported the company's position on the first cost report, which was the subject of a chain of earlier emails between Peter and Stelly (with Arata carbon copied). Peter and Arata knew or at the very least could reasonably foresee that this email would be sent.

Counts 8, 10, & 12 (wire fraud): The jury convicted only Peter on these counts. Each concerns emails either he or Seven Arts Vice President Marcia Matthew sent to auditor Davis, with attachments showing proof of expenditures related to the second cost report. The district court upheld the convictions. Peter again argues a lack of fraud specific to these expenditures but, as we have explained, that showing is not necessary. As long as the scheme to defraud employed misrepresentations, a truthful email that helped advance the scheme can be the basis for a wire fraud conviction. It turns out the jury could have viewed these claimed expenditures as fraudulent—for example, the invoice that is the subject of Count 12 was supposedly for construction expenses yet includes $350,000 in legal and notary fees and $250,000 for auditors—but that was not necessary. We affirm these three convictions.

Count 9 (wire fraud): The wire in Peter's Count 9 conviction is an email Davis sent to Peter about payment confirmation letters and film equipment purchases that were necessary to complete the Dienes firm audit for the second cost report. Peter argued below that his communication with Davis was part of the normal "give and take" of the audit, that he made no material falsehoods, and that the email was not in furtherance of a scheme to defraud. The district court disagreed, finding the jury entitled to determine that he employed false representations to the auditors so that fake equipment purchases would be included in the report. Again, this finding was not necessary though it is supportable as there was testimony from two witnesses—Richard Conway and Simon Ellson—that the claimed purchase of film equipment from a British company never happened. Regardless, the email was a step in completing the audit, which was itself a step in gaining approval for the second cost report. Because the email furthered the scheme to fraudulently obtain tax credits, the conviction will be upheld.

Count 11 (wire fraud): The jury convicted both Peter and Susan on Count 11. This count concerns a December 2009 certification, created by Seven Arts and signed by Susan. It lists payments for office space, consulting fees, and project management fees totaling $700,000. Matthew sent the payment confirmation to Davis. The district court upheld the conviction. As is the case with the communications we have already discussed, sending this certification to the company's auditor furthered the scheme to obtain tax credits. We will not disturb the district court's denial of the motions for acquittal on this count.

Count 13 (wire fraud): The jury convicted Peter and Arata on Count 13. It is a December 2009 email from Matthew to Davis that included invoices for roughly $350,000 in legal work allegedly done by Peter and Arata on 807 Esplanade as well as a loan agreement between Seven Arts and Susan's New Moon Pictures. The district court overturned Arata's conviction on this count but upheld Peter's.

The district court found that Peter was not entitled to acquittal because a rational juror could have concluded that the evidence supported a finding that he created a nonexistent $10 million loan from New Moon and fake draws on that loan, supported by circular transactions, in order to claim interest expenditures. But that finding of fraud specific to this email was not necessary for the reasons we have discussed. Emailing the invoices and loan agreement furthered the scheme to obtain tax credits. Peter's conviction will stand. The district court granted an acquittal as to Arata because (1) the evidence was insufficient to show he provided support for the legal fees and (2) he withdrew from the conspiracy before the loan agreement was sent to the auditor. Our earlier rejection of the withdrawal rationale takes care of both of these concerns. As Peter caused those documents to be sent in furtherance of the scheme, Pinkerton means Arata is also liable for that foreseeable act that furthered the conspiracy he was still part of.

Counts 14-20 (wire fraud): The jury convicted Peter of wire fraud on these seven counts. They are emails discussing material related to the company's second or third cost reports. Counts 14 through 18 concern documentation for expenditures eventually included in the second cost report submission. Counts 19 and 20 relate to the same for the third cost report. The district court, noting that Peter merely raised factual disputes that the jury was entitled to resolve in the government's favor, denied his motion for acquittal on these counts. As finalizing the expenditures submitted to the state was a core part of the fraud, these wires easily furthered Peter's efforts to defraud Louisiana of tax credits. We affirm the district court's denial of the motions to acquit on these counts.

Count 21 (mail fraud): The jury convicted all three defendants on Count 21. The mailing was a package Peter sent to forensic auditor Michael Daigle containing materials supporting the second cost report. The district court upheld the convictions of Peter and Susan, and so do we. As should be apparent by now, this type of communication furthered the fraud because it was an attempt to convince the auditor to approve the expenses. Because Arata was still a member of the conspiracy at this time, the verdict against him should also stand.

C.

In Counts 22 through 25, the jury convicted Arata of making false statements to the FBI during an interview in January 2014. The district court disagreed with those verdicts and acquitted Arata on all of them. The findings of guilt should be reinstated if there was sufficient evidence to show that Arata (1) knowingly and willfully (2) made a statement (3) that was false, (4) material, (5) and within the jurisdiction of the FBI. United States v. Hoover, 467 F.3d 496, 499 (5th Cir. 2006). The viability of these jury verdicts turns largely on the "knowing" and "falsity" elements.

Count 22: Essentially for the reasons the district court provided in granting the Rule 29 motion, we affirm its ruling on this count. The jury found Arata lied when he said he "terminated his relationship" with Peter in the summer of 2009. Arata had sent a letter ending the attorney-client relationship and his day-to-day involvement in 807 Esplanade, so to support the verdict the government has to advance a broad theory of "relationship" that includes Arata's limited involvement in the second cost report and other business ventures that continued after that point. But the termination letter, which Arata provided to the FBI, contemplates a number of ways in which his relationship with Peter would continue outside the attorney-client context. In light of these circumstances, there was not evidence to support a finding that Arata knowingly provided a false statement in saying he terminated his relationship with Peter.

Count 23: The jury convicted Arata on this count for saying he was "not aware" of the legal fee expenses the company claimed in the second cost report. The district court's acquittal relied heavily on the FBI agent's acknowledgment that there was no evidence Arata saw the second cost report before its submission. But the government correctly points out that one need not see a document to be aware of it. There was certainly evidence that Arata knew Peter intended to submit the legal fees in the second report. As we noted, Arata told his business partner that "[Peter] wants to submit [the legal fees] for tax credits. Ha! . . . [S]ince I was not his lawyer for the deal, it makes it even better. What he could submit and what is actual are the bills he got from [other attorneys]. But instead, he . . . puts me down as receiving $150K in fees! Love it." And there is evidence to show that Arata knew Peter went through with it as he helped conceal this fraud. The auditor Davis testified that Arata sent her the operating agreement showing his company Voodoo received a stake in SAFELA; that verified the fees as Voodoo's equity interest was how Arata was paid for the supposed legal work. How can a person verify something they are not aware of? This evidence is sufficient to support the jury's view that Arata lied when he said he was not aware of the legal fees claimed in the second report.

Counts 24: This count concerned Arata's statement that the film equipment listed in the company's first cost report "had been 'acquired' in that [it] would be contributed to 807 Esplanade by the vendor as a business partner." The equipment had neither been acquired nor contributed. Yet the district court determined no evidence existed to support a finding that Arata lied in voicing his belief that the Departure equipment deal "would be" completed.

There is evidence to support the district court's negative view of this count. The equipment-for-ownership deal with Martin's Departure Studios fell through only after the first cost report was submitted. Departure sent Seven Arts a list of film equipment valued at over $1 million in September 2008. Peter signed an affidavit in November of that year attesting to the fact that Seven Arts "acquired" film equipment that "w[ould] be delivered" upon completion of 807 Esplanade. Martin also testified to his understanding that equipment would be delivered to 807 Esplanade and that Departure would be paid for it. In fact, Martin was under the impression that he would be a 25% partner in the business, though he admitted that the arrangement "was not formalized." When asked at trial whether he believed the film equipment transaction was "a real deal," Martin responded "Yes."

But there is also evidence in the other direction, and that is enough to require deference to the jury's finding the inculpatory evidence more compelling. Most powerfully, evidence showed that Arata was a party to the creation of fake equipment purchase invoices and payment certifications that he then forwarded to the company's auditors. Why engage in this fraud if Arata believed that Department Studios would in fact contribute the equipment and that Seven Arts had already "acquired" it? If that were the case, legitimate documentation would exist. Further support for the jury's finding is found in Arata's indication in January 2009, before the first cost report submission, that Seven Arts was already storing in its California office sound mixing and editing equipment purchased from Departure Studios. That was not true. We reverse the acquittal on Count 24.

Count 25: The final alleged false statement is Arata's saying that he "thought he fully disclosed both sides of the transactions for construction and equipment expenditures to the auditors." The government argued, and the jury agreed, that Arata instead had purposely concealed those transactions. The district court vacated the conviction because of its view that there was insufficient evidence to support Arata's intentional concealment of the circular transactions.

There was sufficient support for the jury's contrary view that Arata fully disclosed only part of the transactions—the "first half" consisting of the outgoing payments but not the money coming back into the accounts. In an email to himself, for example, Arata attached invoices documenting both sides of multiple circular transactions between Seven Arts and Departure, which were routed through Regions Bank. He did the same for the Duvernay transactions. But documentation showing credits to the company's account, as opposed to debits from it, was stripped from the emails provided to auditor Katherine Dodge. The government admits that Arata did disclose the "second halves" of these transactions to the auditor, albeit in the illegible form of carbon copies of handwritten bank tickets. This stark contrast between the clean documents showing the outgoing money and the barely discernible ones showing that money coming back is not consistent with "fully disclos[ing]" the circular transactions. And Arata never stated in the body of the emails to the auditor that the money cycled through the accounts. This is enough to get the government past the low hurdle of sufficiency review, with "fully" doing a lot of the work to show the falsity of the statement. We also reinstate this conviction.

* * *

This is how things stand after the sufficiency review. Peter is convicted of all 21 counts, which includes one count of conspiracy (Count 1), nineteen counts of wire fraud (Counts 2-20), and one count of mail fraud (Count 21). Arata is convicted of one count of conspiracy (Count 1), six counts of wire fraud (Counts 2-7), one count of mail fraud (Count 21), and three counts of making a false statement (Counts 23-25). Susan is convicted of one count of conspiracy (Count 1), one count of wire fraud (Count 11), and one count of mail fraud (Count 21).

IV.

We next address defendants' motions for a new trial. The district court denied defendants the "exceptional remedy of a new trial," even on the contingency—now realized—that we were to reverse its acquittals.

Unlike the sufficiency review we just conducted, which evaluates a jury's findings and thus gives no deference to the trial judge, the decision on a new trial motion is entrusted to the discretion of the district court so we will reverse it only on an abuse of that leeway. United States v. Piazza, 647 F.3d 559, 564 (5th Cir. 2011). The trial court may grant a new trial "if the interest of justice so requires." FED. R. CRIM. P. 33(a). New trial requests generally take two forms. The first, like sufficiency review, focuses on the evidentiary support for the verdict, with the movant having to show that the verdict is so strongly against the weight of the evidence that it affects the defendant's substantial rights. United States v. Wright, 634 F.3d 770, 775 (5th Cir. 2011). A new trial request can also be based on procedural problems with the trial if they caused a miscarriage of justice. Id. The defendants pursued both avenues in the trial court but only appeal the ruling as to the alleged procedural defects.

A.

Defendants raise four grounds for a new trial. Arata argues that the government repeatedly used improper trial tactics, leading to an unjust verdict.¹⁶ When a new trial is sought for prosecutorial misconduct, any improper remark must impact the defendant's substantial rights. United States v. Rice, 607 F.3d 133, 138 (5th Cir. 2010).

16 In their reply, Peter and Susan advanced arguments of retroactive misjoinder, admission of privileged communications, and prosecutorial misconduct that were only "noticed" in the table of contents of their opening brief. Failure on appeal to adequately brief an issue waives it. Procter & Gamble Co. v. Amway Corp., 376 F.3d 496, 499 n.1 (5th Cir. 2004). Citing an issue in the table of contents but then not addressing it in the body of the brief obviously does not constitute adequate briefing. See FED. R. APP. P. 28(a)(8)(A) (noting that an appellant's argument must contain "contentions and the reasons for them, with citations to the authorities and parts of the record on which the appellant relies"). The reason we do not allow new arguments in a reply is that the other side does not have a chance to respond. That problem exists when all the opening brief does is provide one sentence on an issue in the table of contents. Trying to raise an argument only by listing it in the table of contents is also an end run around page limits.

In its opening statement, the government said the defendants "utterly abused the Louisiana film tax credit program, and in the process they took advantage of and exploited every human being that they could." The district court called this "[f]alse theater." But such theater was not so far afield from the theory and evidence the government presented throughout trial; evidence that it turns out was sufficient to sustain multiple convictions against Arata. As such, the district court acted within its discretion in deciding that any hyperbole did not require a new trial. The same is true for Arata's complaints about how the government framed questions to witnesses. For example, it asked Arata's business partner whether Arata said he "called Katie Davis and told her that the operating agreement was substantiation for the legal fees?" Davis had not testified to that fact, though she speculated that Arata submitted the operating agreement to support payment of the disputed legal fees. Though the district court deemed these questions "utterly inappropriate," it determined that Arata could not show that they "caused any prejudice" in the scope of a trial that lasted two weeks. We agree.

B.

The Hoffmans point to three trial court rulings in seeking a new trial. They first contend that the instruction telling the jury it is "not necessary that the government prove that the defendants violated, or intended to violate a Louisiana state legal duty, law rule or regulation" amended the indictment. Improper constructive amendment occurs when the jury is allowed "to convict the defendant upon a factual basis that effectively modifies an essential element of the offense charged." United States v. Cooper, 714 F.3d 873, 878 (5th Cir. 2013) (quoting United States v. Gonzalez, 436 F.3d 560, 577 (5th Cir. 2006)). That did not happen. Contrary to the Hoffmans' contention, the indictment did not charge them with violating state law. It charged them with making various misrepresentations—lies about the company's expenditures, the creation of purchase invoices, and the purpose of circular transactions. Using such lies in furtherance of a scheme to defraud violates federal law regardless whether they independently violate state law. See United States v. Dotson, 407 F.3d 387, 393 (5th Cir. 2005).

C.

The Hoffmans also argue they are entitled to a new trial because the district court rejected their proposed jury instructions, including one on the meaning of the tax credit statute. As long as the instructions it gives accurately state the law, a district court is given "substantial latitude" in the particulars of how it instructs the jury. United States v. Richardson, 676 F.3d 491, 506-07 (5th Cir. 2012). On top of that is the discretion it receives here because this issue is being raised in a challenge to the denial of a new trial motion. That deference is not pierced by the failure to instruct on details about the tax credit law given that the instruction accurately informed the jury of the elements of mail and wire fraud. United States v. Cessa, 856 F.3d 370, 376 (5th Cir. 2017) (finding no abuse of discretion when the court gave the correct instruction even if defendant's requested addition was also legally accurate).

D.

The Hoffmans' final basis for a new trial is the district court's decision to exclude expert testimony about the film tax credits. They contend the testimony would have highlighted the confusing nature of the regulations and thus shed light on their intent to defraud (that is, their lack thereof). Deference to trial court rulings in this area again poses a significant hurdle. Cf. United States v. Guerrero, 768 F.3d 351, 365 (5th Cir. 2014) (acknowledging that Daubert decisions are reviewed for abuse of discretion and should not be disturbed unless "manifestly erroneous" (quoting United States v. Norris, 217 F.3d 262, 268 (5th Cir. 2000))). It appears the district court could have allowed this testimony so long as it was focused on descriptions of the tax credit regime and not opinions about the defendants' mindsets. Compare United States v. Calvin, 39 F.3d 1299, 1309 (5th Cir. 1994) ("By disallowing that testimony the district court deprived [the defendant] of an opportunity to present critical evidence that he lacked fraudulent intent in assisting with the transactions."), and United States v. Davis, 471 F.3d 783, 789 (7th Cir. 2006) ("Experts are permitted to testify regarding how their government agency applies rules as long as the testimony does not incorrectly state the law or opine on certain ultimate legal issues in the case."), with FED. R. EVID. 704 (noting that an expert witness testifying in a criminal case "must not state an opinion about whether the defendant did or did not have a mental state . . . that constitutes an element of the crime charged or of a defense," even though opinion testimony embracing "an ultimate issue" is not generally objectionable).

But we need not determine whether the district court exceeded the bounds of its discretion in excluding the testimony because the byzantine nature of the tax credit program was otherwise conveyed to the jury. As the district court noted, "evidence at trial showed that [the] then-newly passed film infrastructure tax law was implemented haphazardly and in a manner rife with disorder." Plenty of witnesses involved in the creation and evaluation of the cost reports—including Seven Arts employees, auditors, state officials, and business partners (actual and contemplated)—made this point that the Hoffmans contend undermines a finding of fraudulent intent. What is more, Peter, a self-professed tax lawyer, testified at length about his understanding of the statute's language and purpose. So did Arata, also a lawyer, who noted that the state did not even pass rules interpreting the statute until 2010. Thus any error in not allowing the expert to testify did not cause substantial prejudice.

* * *

No ruling during the trial caused a miscarriage of justice. There is no basis for redoing it.

V.

Having upheld the jury's verdict in large part, we now consider sentencing. The government argues the probation sentences are unreasonable in light of the much greater sentences recommended by the Guidelines. Appellate review of the substantive reasonableness of a sentence is "highly deferential." United States v. Campos-Maldonado, 531 F.3d 337, 339 (5th Cir. 2008). It is not enough that "the appellate court might reasonably have concluded that a different sentence was appropriate." Gall v. United States, 552 U.S. 38, 51 (2007). An abuse of discretion must be shown to undo the decision of the trial judge who is in the best position to weigh the sentencing factors. Id. at 51-53. Even sentences like these that are outside the Guidelines range are reviewed with deference, though they are not entitled to the presumption of reasonableness that a within-Guidelines sentence may be afforded on appellate review. Id. at 51.

The dissenting opinion emphasizes this discretion. But while considerable deference is due the sentencing court given the bespoke nature of criminal punishment, the Supreme Court preserved a role for appellate review when it ruled that the Guidelines were only advisory. See United States v. Booker, 543 U.S. 220, 261-265 (2005). Rather than reverting to the pre-Guidelines situation when there was essentially no reasonableness review of a sentence, Koon v. United States, 518 U.S. 81, 96 (1996), the Court concluded that appellate review would assist in "avoid[ing] excessive sentencing disparities while maintaining flexibility sufficient to individualize sentences where necessary," Booker, 543 U.S. at 264-65. Consistent with that concern about disparities, appellate courts "may consider the extent of the deviation" from the Guidelines when performing their limited function as a check on extreme ones. Gall, 552 U.S. at 51.

A.

For Peter, the gap is colossal between the custodial sentence the Guidelines recommended, a range 168 to 210 months,¹⁷ and the 60 months of probation he received.¹⁸ His counsel acknowledged that he was not aware of our court's considering any challenge to a sentence in which the downward variance was so great. This chasm between the Guidelines' view of the appropriate sentence and the district court's, with its ramifications for the sentencing disparities that Congress instructs courts to avoid, see 18 U.S.C. § 3553(a)(6), is an important factor in considering whether the district court exceeded its discretion. Gall, 552 U.S. at 49-50 (noting it is "uncontroversial that a major departure should be supported by a more significant justification than a minor one"). But what ultimately matters is whether its assessment of the statutory sentencing factors was reasonable, so we consider both the reasons why the district court thought probation was warranted and the reasons why the Guidelines and government think prison time is necessary.

17 There is some suggestion that the district court did not determine a final Guidelines range. But the Statement of Reasons it signed after the sentencing hearing confirms that the district court adopted the Presentence Report's recommended range of 168 to 210 months.

18 Our reinstatement of five guilty verdicts on which the district court had acquitted Peter does not pose an obstacle to our review of the sentences for the counts on which the court did enter judgment. The reversals turned on issues like whether there was sufficient evidence that particular emails crossed state lines or furthered the scheme. None of these questions affect Peter's overall culpability. His Guidelines calculation captured the loss attributable to the entire scheme, so the reinstated counts will not affect that. It is for this reason that we voiced skepticism about the need to charge and convict Peter of all 21 counts. See supra note 9. As the reinstatement of the additional counts does not alter the Guidelines or change any other sentencing consideration, we will review the reasonableness of the sentences that were entered. Cf. United States v. Weingarten, 713 F.3d 704, 712 (2d Cir. 2013) ("[I]f the vacatur of a count of conviction has altered the 'factual mosaic related to' the remaining counts, on remand 'the court must reconsider the sentence imposed on the count or counts affected by the vacatur . . . as well as on the aggregate sentence.'" (quoting United States v. Rigas, 583 F.3d 108, 118-19 (2d Cir. 2009))).

Why was Peter's Guidelines range so high? To the base offense level for fraud offenses, the Guidelines added enhancements because the intended loss exceeded $3.5 million, the scheme was sophisticated, Peter led it, he abused his position of trust as a lawyer to facilitate the fraud, and he obstructed justice by lying at trial. These facts are relevant to numerous statutory factors courts must consider, including the "nature and circumstances of the offense," "history and characteristics of the defendant," and "need for the sentence imposed to reflect the seriousness of the offense." 18 U.S.C. § 3553(a)(1), (a)(2)(A). It is also noteworthy that this was not Peter's first brush with the law as is often true in white-collar cases;¹⁹ he has a 1997 federal conviction, albeit a misdemeanor, for delivering a false tax return.

19 Federal fraud defendants are less likely to have criminal history than any other category of offenders except those convicted of child pornography. UNITED STATES SENTENCING COMMISSION, THE CRIMINAL HISTORY OF FEDERAL OFFENDERS 4-6 (2018).

So why did the district court believe probation was appropriate? The main reason seems to have been what it described as a "serious dispute" that the project may have eventually been entitled to even more tax credits than were fraudulently obtained with the first cost report. When pronouncing sentence it also noted a related concern about inconsistency in the state's view about how much it lost, as well as Peter's "health issues," the fact that his prior federal conviction was a misdemeanor, and its view that a sentence of probation "is sufficient to deter other criminal conduct."

We disagree with that final assessment about the deterrent value of Peter's sentence. Giving probation to the leader of a sophisticated, multimillion dollar fraud scheme—particularly a defendant undeterred by a previous term of probation for a federal economic crime and who also lied at trial—perpetuates one of the problems Congress sought to eliminate in creating the Sentencing Commission: that sentencing white-collar criminals to "little or no imprisonment . . . creates the impression that certain offenses are punishable only by a small fine that can be written off as a cost of doing business." United States v. Martin, 455 F.3d 1227, 1240 (11th Cir. 2006) (quoting S. Rep. No. 98-225, at 76 (1983)); see also Mistretta v. United States, 488 U.S. 361, 375 n.9 (1989) (noting the Senate Report's view that sentencing had been too lenient for white-collar criminals); Brent E. Newton, The Story of Federal Probation, 53 AM. CRIM. L. REV. 311, 315 & n.29 (2016) (reciting the extensive legislative history showing that Congress intended for many white-collar defendants to receive prison time).²⁰ This ineffective deterrence is especially concerning given that scholars believe there is a greater connection in white collar cases between sentencing and future as financial crimes are "more rational, cool, and calculated than sudden crimes of passion or opportunity." Martin, 455 F.3d at 1240 (quoting Stephanos Bibas, White-Collar Plea Bargaining and Sentencing After Booker, 47 WM. & MARY L. REV. 721, 724 (2005)). Another problem with probation in multimillion dollar fraud cases is that it undermines public confidence in whether the justice system is "do[ing] equal right to the poor and to the rich" as our oath requires. 28 U.S.C. § 453. For these reasons, we have repeatedly expressed a "distaste for sentencing that reflects different standards of justice being applied to white and blue collar criminals," United States v. Saleh, 257 F. App'x 740, 745 (5th Cir. 2007) (citing United States v. Andrews, 390 F.3d 840, 848 (5th Cir. 2004)); see also United States v. Mueffelman, 470 F.3d 33, 40 (1st Cir. 2006) (noting the need to minimize "discrepancies between white- and blue-collar offenses").

20 The Guidelines recognize this history:

Under pre-guidelines sentencing practice, courts sentenced to probation an inappropriately high percentage of offenders guilty of certain economic crimes, such as theft, tax evasion, antitrust offenses, insider trading, fraud, and embezzlement, that in the Commission's view are "serious."

The Commission's solution to this problem has been to write guidelines that classify as serious many offenses for which probation previously was frequently given and provide for at least a short period of imprisonment in such cases. The Commission concluded that the definite prospect of prison, even though the term may be short, will serve as a significant deterrent, particularly when compared with pre-guidelines practice where probation, not prison, was the norm.

U.S.S.G. ch. 1, pt. A(4)(d).

Peter's second sentence of probation in the federal system does not deter large-scale fraud or reflect the serious nature of either this offense or economic crimes generally. 18 U.S.C. § 3553(a)(1), (a)(2)(A)-(B). It also results in significant and unwarranted sentencing disparities with others engaged in frauds of similar magnitude who receive sentences at least in the ballpark of what the Guidelines recommend. Id. § 3553(a)(6). Some of the reasons the district court gave for its sentence, especially the uncertainty about whether Louisiana ultimately suffered any loss, are sound reasons for a downward variance, even a substantial one. But this is not a case in which the court went 50%, or even 75%, below the Guidelines range.²¹ It went from roughly 15 years in prison to zero. In reviewing the reasonableness of a Booker sentence, the Supreme Court recognized that "custodial sentences are qualitatively more severe than probationary sentences of equivalent terms." Gall, 552 U.S. at 43, 48, 59-60. Here the substantial qualitative difference between custody and probation is combined with a drastic reduction in the length of the punishment—168 months to 60 months. Other courts of appeals have vacated variances of much lesser degree that benefitted white-collar defendants.²² What is more, none of those defendants had a prior white-collar conviction and most of them accepted responsibility by pleading guilty. See supra note 22.

21 The dissenting opinion treats the sentence as a 72% variance. Dissenting Op. at 7. It does this by equating 60 months of probation with 60 months in custody (so 60 is a 72% reduction from the 168 low end of the Guidelines). That notion is easily dispelled almost every day in this circuit when defendants plea for probation at sentencing hearings. And as discussed above, treating sentences of probation and custody as equivalent is also at odds with the views of Congress and the Supreme Court. Finally, even ignoring the qualitative differences, a 72% variance is much more substantial than many cases in which courts have found downward variances in white-collar cases to be unreasonable. See infra note 22.

22 See, e.g., United States v. Morgan, 635 F. App'x 423, 448-52 (10th Cir. 2015) (finding that the case "cries out for appellate intervention" and requiring resentencing because a noncustodial sentence, in the face of a 41 to 51 month guidelines range, would not deter public officials from soliciting bribes); United States v. Hayes, 762 F.3d 1300, 1307-10 (11th Cir. 2014) (finding sentence of three years of probation unreasonable when Guidelines range was 41 to 51 months even though defendant had cooperated); United States v. Musgrave, 761 F.3d 602, 609 (6th Cir. 2014) (noting, in vacating the district court's one-day prison sentence in the face of a 57 to 71 months guidelines range, that "Congress understood white-collar criminals to be deserving of some period of incarceration, as evidenced by its prohibition on probationary sentences in this context"); United States v. Kuhlman, 711 F.3d 1321, 1325, 1328-29 (11th Cir. 2013) (finding probation sentence unreasonable when Guidelines range was 57 to 71 months); United States v. Peppel, 707 F.3d 627, 635 (6th Cir. 2013) (finding seven days plus three years supervised release unreasonable when Guidelines range was 97 to 121 months); United States v. Engle, 592 F.3d 495, 497-98, 501-04 (4th Cir. 2010) (vacating a sentence of 48 months' probation for tax evasion when the Guidelines range was 24 to 30 months in prison); United States v. Livesay, 587 F.3d 1274, 1279 (11th Cir. 2009) (vacating a sentence of 60 months' probation in light of defendant's 78 to 97 months Guidelines range and holding that "any sentence of probation would be unreasonable given the magnitude and seriousness" of his conduct); Martin, 455 F.3d at 1230, 1241-42 (vacating, when defendant's Guidelines range was 108 to 135 months imprisonment, a seven-day sentence after the court of appeals had previously rejected a sentence of 60 months' probation).

The dissenting opinion ignores Peter's criminal history as well as other factors favoring a meaningful sentence such as Peter's lying in court, using his position as a lawyer to facilitate the fraud, and leading a sophisticated conspiracy. 18 U.S.C. § 3553(a)(1) (listing "history and characteristics of the defendant" and "nature and circumstances of the offense" as factors to consider in imposing a sentence). It instead focuses on the district court's later conclusion in its restitution order that Louisiana did not end up suffering a loss as a justification for the extreme variance. No doubt loss is a key—often the key—factor in sentencing a fraud defendant. But it is not the exclusive concern. Congress and the Sentencing Commission have commanded that courts conduct a holistic evaluation that includes the troubling features of Peter's conduct and history we have just mentioned. Id. § 3553(a); U.S.S.G. § 2B1.1.

Even just considering loss, Peter is not the Chamber of Commerce hero the dissenting opinion makes him out to be. In talking only about the state's actual loss, it neglects that Peter would have stolen millions from the state if it had not detected his scheme. Dissenting Op. at 3-5. The Guidelines say to use intended loss when that is greater than actual loss, U.S.S.G. § 2B1.1 app. note 3(A) & 3(A)(ii), the reason being that a fraudster's intent reflects his culpability, ROGER W. HAINES, JR. ET AL., FEDERAL SENTENCING GUIDELINES HANDBOOK 275 (2002) (explaining that "intended loss is a direct measurement of culpable mental state"); Frank O. Bowman, Coping With "Loss": A Re-examination of Sentencing Federal Economic Crimes Under the Guidelines, 51 VAND. L. REV. 461, 558-60 (1998) (explaining that a focus on intended loss makes sense for "moral and utilitarian considerations"). Indeed, that is why it has long been against the law to attempt a crime even if one does not succeed. Id. at 559 ("The Sentencing Commission provided an increase in offense level for 'intended loss' for the same reasons that substantive criminal liability is imposed for inchoate crimes like attempt and conspiracy."); see also Francis Bowes Sayre, Criminal Attempts, 41 HARV. L. REV. 821, 822-837 (1928) (tracing the criminalization of attempts back to the Star Chamber and treatise of Sir Edward Coke).

Judged by this telling measure of culpability, Peter tried to steal $2 million from Louisiana beyond what his project earned when all was said and done. The district court credited testimony from a state auditor indicating that expenditures on the project would have ended up qualifying Seven Arts for about $1.6 million in tax credits even though the claims were false when submitted. That figure exceeds the roughly $1.1 million issued and later revoked, which is what the dissenting opinion emphasizes.²³ But it neglects that Peter submitted false claims totaling more than $9.1 million, 40% of which would have resulted in over $3.6 million in credits. Only the state's vigilance in discovering the circular transactions and phony expenditures kept it from being cheated out of the additional millions.²⁴

23 Notably, for the first application which is the only one the state approved, the project did not end up earning the all the credits it received. It was entitled to only $860,000 according to the state auditor the district court credited. That is why in calculating forfeiture the court used $272,480.80 as the amount Seven Arts received above what it ended up earning on the first application. So looking just at the first application, the state did lose money.
The district court found that state did not ultimately lose money on the entire project because it would have qualified for $1.6 million in credits. As discussed above, however, that is far less than the $3.6 million in credits that Peter sought and would have fraudulently received had the state not detected his fraud.

24 Even if the project ended up receiving all the credits that Peter sought, submitting false claims to obtain the credits before they were earned ran a significant risk that Louisiana would not be made whole. As is the case for any business enterprise, it was far from a guarantee that the facility would end up being built. Any number of economic, personal, regulatory, or—this being New Orleans—weather-related hardships could have prevented the completion of the project. Part of why intended loss is relevant to a sentencing court is that it captures "the degree of risk the defendant's behavior posed." Bowman, supra, at 560.

That this fortuity of having been caught should not fully excuse Peter's complex scheme can be shown with an analogy to a "blue collar" theft. Consider a thief who steals $2 million dollars of jewelry from a store. If police catch him leaving the store and recover the stolen goods, is it likely that a no-harm-no-foul argument would result in a sentence of probation? Of course not. Looking only at actual loss in fraud cases where the fraudster is caught in the act is thus another implicit way in which "different standards of justice [may be] applied to white and blue collar criminals." Saleh, 257 F. App'x at 745. The victim being made whole can certainly be a mitigating factor at sentencing, but it does not justify the degree of leniency afforded Peter given his attempt to receive millions more in tax credits than the project ever earned. See Bowman, supra, at 559 (explaining that the law punishes attempts in part because luck plays a role in whether people engaging in equally blameworthy conduct succeeds).

Determining the outer boundaries of a sentencing judge's discretion is admittedly a judgment call. But looking at the entire sentencing landscape, we readily conclude that this sentence exceeded those bounds. Peter's scheme was a serious one that involved creating bogus financial transactions in an effort to mislead a state agency into issuing almost $4 million in tax credits. One only needs to have read this opinion to see the tangled web of financial maneuvers Peter wove. Add to that his criminal history, perjury, and use of his position as a lawyer to further the crime. The result is that giving Peter probation was a variance too far.

We vacate the sentences of probation and remand for resentencing on those counts, along with the ones we reinstated, consistent with the principles we have just discussed.

B.

If our review of Peter's sentence shows the limits of a district court's sentencing discretion, our review of Susan's demonstrates its extent. She too received a sentence of probation (three years). But her Guidelines range was much lower than Peter's; it recommended a prison term of 46 to 57 months. This reflects her far less substantial role in the offense. As the district court observed, witnesses "scarcely mentioned" her during the trial. Whereas Peter dove head first into the fraud, Susan just dipped her toes in it. That is enough to sustain her convictions for the reasons we have explained. But a person's role in the offense is a critical factor in sentencing. In addition to not being a leader of the fraud, Susan does not have any criminal history, did not commit perjury, and did not abuse a position of trust. To be sure, even if not nearly as great as Peter's, the downward variance she received was substantial. It is of similar scope to some we just cited that other courts have vacated. See Engle, 592 F.3d at 495. But the extent of a variance is just one consideration in reviewing the substantive reasonableness of a sentence. That review is highly factbound, so one can also find decisions affirming downward variances similar to the one Susan received. See, e.g., United States v. Rowan, 530 F.3d 379, 380-81 (5th Cir. 2008) (affirming a sentence of 60 months supervised release on child pornography charges despite a guidelines range of 46 to 57 months). And on the flip side, we have upheld a number of upward variances of similar and sometimes much greater magnitude. See, e.g., United States v. Hebert, 813 F.3d 551, 561-63 (5th Cir. 2015) (upholding a variance of 1214% from the high end of the Guidelines range); United States v. Urbina, 542 F. App'x 398, 398-99 (5th Cir. 2013) (affirming a 60-month sentence that was 329% higher than the Guidelines range maximum); United States v. Brantley, 537 F.3d 347, 349-50 (5th Cir. 2008) (upholding a 180-month sentence that was 253% higher than the maximum end of the Guidelines range); United States v. Mejia-Huerta, 480 F.3d 713, 717-18, 723 (5th Cir. 2007) (affirming a 120-month sentence that exceeded the high end of the Guidelines range by 344%). Booker discretion is not a one-way street. We defer to both upward and downward variances so long as the district court provides an explanation tailored to the statutory sentencing factors that is not outside the bounds of reasonableness. It did so in sentencing Susan to probation.²⁵

25 The government also alleges the district court committed procedural error in deciding Susan's sentence, but we reject that claim.

C.

Arata's Guidelines range was, at 108 to 135 months, higher than Susan's but lower than Peter's. That is consistent with his relative role in the scheme. We do not address the substantive reasonableness of his probation sentence, however, because the reinstatement of certain counts may influence sentencing. Cf. Weingarten, 713 F.3d at 712 (explaining how an altered "factual mosaic" may affect resentencing). For at least one thing, our reversal of some of the false statement counts means that Arata lied to the FBI in connection with the investigation. Obstruction of justice is a relevant sentencing consideration. We thus vacate his probation sentences without opining on their propriety to allow the district court to sentence him in the first instance under the new landscape resulting from our sufficiency review.

VI.

The final issue is forfeiture. Arata and the government challenge the district court's $223,434.25 award.²⁶ As opposed to restitution which is remedial, forfeiture is punitive. The aim of a forfeiture award is to take any ill-gotten gains from a defendant. See United States v. Taylor, 582 F.3d 558, 566 (5th Cir. 2009).

26 The Hoffmans brief forfeiture only in their reply, when they challenge the amount awarded and also a couple procedural aspects of the order. As with some of their new trial arguments, see supra note 16, they only include these issues in the table of contents of their opening brief. That is not sufficient. Id. In any event, we note that there is no problem with the timing of the forfeiture, and the Hoffmans cannot establish plain error with respect to the district court's failure to ask whether the parties wanted a jury to decide forfeiture. United States v. Valdez, 726 F.3d 684, 699 (5th Cir. 2013) (noting that the failure to inquire whether parties wanted the jury to decide forfeiture did not meet the third and fourth requirements for plain-error correction when evidence supported the court's award).

Arata contends no forfeiture should have been awarded based on his view that the project ultimately qualified for more tax credits than it received. The government argues the award should have reflected the full $1,132,480.80 in issued credits without any reduction for amounts the project ultimately earned or the street value of the credits. In calculating forfeiture, the district court started with that $1,132.480.80 in issued credits. It then subtracted the $860,000 in tax credits a state accountant testified Seven Arts was entitled to. This put the amount at $272,480.80. The court then valued the tax credits in light of the 82 cents on the dollar the company received when it sold them. Applying that ratio to the illegal credits received resulted in the award of $223,434.25 (272,480.80 x .82).

We find no clear error in this calculation. The district court was entitled to offset the forfeiture with the amount of credits Seven Arts ultimately earned according to the state accountant, a number Arata says was too low and the government too high. Using that figure and the adjustment for the market value of the credits was a reasonable means of ascertaining what the defendants gained from their fraud, which is the measure of forfeiture.²⁷ We affirm the forfeiture award.

27 Honeycutt v. United States held that defendants could not be held jointly and severally liable for proceeds derived from narcotics offenses that the defendants did not themselves acquire. 137 S. Ct. 1626, 1630 (2017). We have since applied that holding to forfeiture for health care fraud. See Sanjar, 876 F.3d at 748-50. The defendants do not invoke Honeycutt, however, perhaps because all three were Seven Arts co-owners and therefore "acquired" the ill-gotten tax credits.

* * *

To recap our many rulings: We AFFIRM the district court's denial of defendants' motions to dismiss the indictment. We AFFIRM the district court's denial of defendants' motions for judgment of acquittal, and AFFIRM in part and REVERSE in part the district court's grant of defendants' motions for judgment of acquittal. We AFFIRM the district court's denial of defendants' motions for new trial. We AFFIRM the district court's forfeiture award. Finally, we VACATE Peter Hoffman's sentence and REMAND for resentencing. We AFFIRM Susan Hoffman's sentence. And we VACATE Arata's sentence and REMAND for resentencing. JAMES L. DENNIS, Circuit Judge, concurring in part and dissenting in part:

Though I concur in most of the majority's opinion, I respectfully disagree with the conclusion that the district court abused its discretion by sentencing Peter Hoffman to five years of probation and a $40,000 fine. Because I would affirm Peter's sentence, I respectfully dissent as to part V(A).

1. The Importance of Sentencing Judges' Discretion

As the Supreme Court explained in Gall v. United States: "the sentencing judge is in a superior position to find facts and judge their import under [18 U.S.C.] § 3553(a) in the individual case. The judge sees and hears the evidence, makes credibility determinations, has full knowledge of the facts, and gains insights not conveyed by the record." 552 U.S. 38, 51-52 (2007) (citations and quotation marks omitted). "It has been uniform and constant in the federal judicial tradition for the sentencing judge to consider every convicted person as an individual and every case as a unique study in the human failings that sometimes mitigate, sometimes magnify, the crime and the punishment to ensue." Id. at 52 (citations and quotation marks omitted). Accordingly, "the Court of Appeals should . . . give[] due deference to [a] District Court's reasoned and reasonable decision that the § 3553(a) factors, on the whole, justified the sentence." Id. at 59-60. These § 3553(a) factors include "the nature and circumstances of the offense" and "the history and characteristics of the defendant," as well as the need "to reflect the seriousness of the offense" and "provide just punishment." 18 U.S.C. § 3553(a). The statute prescribes that district courts "shall impose a sentence sufficient, but not greater than necessary, to comply with [these] purposes." Id.

2. The District Court Did Not Abuse Its Discretion

As the majority acknowledges, the district court in this case committed no procedural error: it correctly calculated the applicable Guidelines range, allowed both parties to present argument on what they believed to be an appropriate sentence, considered the § 3553(a) factors, and explained its reasoning before issuing Peter's sentence. The remaining question for this Court is thus whether the resulting sentence was substantively reasonable—i.e., whether the district court abused its discretion in determining that the § 3553(a) factors supported a sentence of probation and a considerable variance from the Guidelines range. After only briefly addressing the uniquely unusual facts of this case (which I detail below), the majority decrees that the sentence of five years of probation and a fine of $40,000 effectively reduced the sentence to "zero" and was "a variance too far." Maj Op. at 51, 55. Respectfully, I must conclude the majority is mistaken, as its analysis fails to apply the requisite deference to the district court's decision. Notably, we "must review all sentences—whether inside, just outside, or significantly outside the Guidelines range—under a deferential abuse-of-discretion standard." Gall, 552 U.S. at 41.

This case features all the hallmarks the Supreme Court has indicated require appellate courts to grant considerable deference to district courts' determination of sentences. As the district court explicitly stated at Peter's sentencing, it was "intimately familiar" with the circumstances of this case: it oversaw a lengthy jury trial and subsequently issued a detailed, 124-page ruling on defendants' motions for acquittal. In this order, the district court noted that the defendants "spent more than $5 million turning the dilapidated mansion at 807 Esplanade Avenue into a state-of-the-art post-production film studio (a studio that is in operation today and has serviced post-production needs for movies and television series)." Not only did this studio ultimately earn the tax credits the defendants received, the credits received were ultimately less than what the defendants were entitled to.²⁸ At sentencing, the district court appeared to accept the defense's related assertion that the state was not a victim, and instead "got exactly what [it] asked for:" a completed, multimillion-dollar post-production studio. Recognizing these unusual circumstances, the district court reached a critical conclusion: "This is not an ordinary fraud case."

28 These determinations were further supported in Peter's PSR, which stated that a downward variance may be appropriate because, among other reasons, "The infrastructure project involved in the instant offense was actually completed and audits confirmed the tax credits released to the project were ultimately earned and were in fact less than the credits the project actually earned when it was later completed."

The majority, in contrast, gives short shrift to these unique extenuating circumstances. The majority acknowledges in its brief introduction that, thanks to newly developed post-production infrastructure funded through its tax credit incentive program, Louisiana has "enjoyed considerable success" in its efforts to "become a place where films are made." Maj. Op. at 2. However, when evaluating the seriousness of Peter's conduct, it then fails to take into account how Louisiana has benefited, and continues to benefit from, completed film infrastructure projects like this one. Completed post-production studios like 807 Esplanade were precisely what Louisiana elected to invest in when it codified its intention "to encourage development in Louisiana of a strong capital and infrastructure base for motion pictures[s] . . . in order to achieve an independent, self-supporting [film post-production] industry." LA. REV. STAT. § 47:6007. Ultimately, in Peter's case, the state not only got what it bargained for: it got it at a discount. The majority glosses over these critical, mitigating facts, instead reweighing Peter's sentencing factors to emphasize aggravating circumstances. See United States v. McElwee, 646 F.3d 328, 343 (5th Cir. 2011) (appellate court not entitled to reweigh sentencing factors (citing Gall, 552 U.S. at 51)); see also Gall, 552 U.S. at 51, 59 ("The fact that the appellate court might reasonably have concluded that a different sentence was appropriate is insufficient to justify reversal of the district court. . . . [I]t is not for the Court of Appeals to decide de novo whether the justification for a variance is sufficient or the sentence reasonable.").

That the state ultimately suffered no loss is all the more significant because Peter's Guidelines range was only as high as it was because of an 18-level enhancement for the $3.6 million "intended loss" calculated in the PSR, a loss the state did not actually incur. In arriving at this figure, however, the PSR did not acknowledge that it included in this loss amount expenditures the defendants had not timely made, but did ultimately make, in order to complete the promised post-production studio. The district court was entitled to, and did, consider that the economic reality differed greatly from the PSR's high loss calculation. Indeed, the district court relied on the PSR itself in doing so. The PSR cautioned that the loss figure it proposed "does not accurately reflect, and appears to over-estimate, the damage caused to the victim in the instant offense." Consistent with the PSR's suggestion that this lack of actual harm could reasonably warrant a downward variance, the district court determined that a Guidelines range based in part on this questionably relevant "intended loss" figure significantly overstated the seriousness of Peter's conduct. Cf. United States v. Huber, 462 F.3d 945, 950-51 (8th Cir. 2006) (holding, where district court departed downward in light of defendant's small "net profit," that "[t]he district court did not clearly err in its factual determination that the high value of the laundered funds led to a base offense level that substantially overstated the seriousness of the offense").

Therefore, though I share the majority's concerns about preferential treatment for white-collar criminals, I disagree with its implication that this is a classic example of letting a white-collar criminal off easy. I conclude instead, that in light of these extenuating circumstances that rendered this a no-loss, victimless crime, the district court was within its discretion to treat Peter differently than it ordinarily would the leader of a large-scale fraud scheme. See United States v. Williams, 517 F.3d 801, 810 (5th Cir. 2008) (noting that a variance deserves "greatest respect" when the facts of a case are out of the ordinary). This case, in which neither the state nor any other institutions or individuals suffered any loss, but in fact received the benefit of a completed, state-of-the-art post-production facility, is not at all like a case such as United States v. Martin, in which the Eleventh Circuit found a downward departure unreasonable because the defendant's crimes "resulted in over a billion dollars of loss harming thousands of victims;" were "major economic crimes that harmed not only individual victims but also many institutions and companies;" and were "peculiarly corrosive to the economic life of the community, as demonstrated by the deleterious effects the large-scale fraud in this case had on the healthcare industries and securities markets." 455 F.3d 1227, 1239 (11th Cir. 2006).

Finally, though the majority does acknowledge that this case presents "sound reasons for a downward variance, even a substantial one," it then proceeds to imply that any non-custodial sentence is a "variance too far" that 64 effectively reduces the sentence to "zero." Maj. Op. at 51, 55. However, Gall specifically rejected this argument, declaring that viewing a probation sentence as a "100% departure" inappropriately "gives no weight to the substantial restriction of freedom involved in a term of supervised release or probation." 552 U.S. at 48. As Gall further

Offenders on probation are nonetheless subject to several standard conditions that substantially restrict their liberty. . . . Probationers may not leave the judicial district, move, or change jobs without notifying, or in some cases receiving permission from, their probation officer or the court. They must report regularly to their probation officer, permit unannounced visits to their homes, refrain from associating with any person convicted of a felony, and refrain from excessive drinking.

Id. Accordingly, I reject the majority's implication that Peter's five-year probation sentence is insubstantial.

The unusual circumstances of this no-loss, victimless case, combined with Peter's age, health conditions, and non-felonious criminal history, justified the court's decision to impose the variance that it did.²⁹ Without reweighing the sentencing factors, which it is well-established we may not do, I cannot conclude that the district court abused its discretion.

29 The PSR also noted that Peter's personal characteristics could justify a downward variance: "The defendant is 66 years old and has never been convicted of any felonious criminal conduct and has a significant history of gainful employment . . . . The likelihood that he will commit further crimes is minimal."

3. The Majority Fails to Give Downward Variances the Deference

Our Circuit Consistently Gives Upward Variances

As a final but not insignificant note: this court consistently upholds sentences that vary upwardly from defendants' Guidelines ranges, citing district courts' considerable discretion in weighing the 18 U.S.C. § 3553 factors and determining appropriate sentences. See, e.g., United States v. Nguyen, 854 F.3d 276, 284 (5th Cir. 2017) (noting our court's "'highly deferential' review for substantive reasonableness"); United States v. Brantley, 537 F.3d 347, 349-50 (5th Cir. 2008); United States v. Williams, 517 F.3d 801, 808-13 (5th Cir. 2008); see also United States v. Guadian, 724 F. App'x 329 (5th Cir. 2018) (upholding a 180-month sentence imposed for a non-violent, no weapons involved marijuana trafficking offense with a Guidelines calculation of 63-78 months). The majority states that we should accord equal deference to downward variances, noting correctly that "Booker discretion is not a one-way street." Maj. Op at 56. Consistent with this principle, the majority affirms Susan Hoffman's downward variance, reasoning that "we have upheld a number of upward variances of similar and sometimes much greater magnitude." Maj. Op. at 56 (citing United States v. Hebert, 813 F.3d 551, 561-63 (5th Cir. 2015) (affirming 1214% upward variance); United States v. Urbina, 542 F. App'x 398, 398-99 (5th Cir. 2013) (329% upward variance); Brantley, 537 F.3d at 349-50 (253% upwards); United States v. Mejia-Huerta, 480 F.3d 713, 717-18 (5th Cir. 2007) (344% upwards)). However, applying the same formula to determine the percentage decrease here as the majority used to determine the percentage increase in the cases that it cites, it becomes clear that each upward variance also significantly outstrips the 72% decrease the district court applied when sentencing Peter to five years of probation with a $40,000 fine.³⁰ Though Peter's variance was admittedly more considerable than Susan's, the majority's own calculations demonstrate that it is still considerably less extreme than the upward variances we have consistently upheld.

30 To calculate the percentage increase or decrease between two numbers, as the majority does, the numerical increase or decrease is divided by the original number, then multiplied by 100. Thus: Percentage Increase = (New Number - Original Number) / Original Number x 100; Percentage Decrease = (Original Number - New Number) / Original Number x 100. Relative change and difference, WIKIPEDIA, https://en.wikipedia.org/wiki/Relative_change_and_difference (last updated Apr. 14, 2018).

The fact is, it is only the exceptionally rare case in which this court finds an upward variance substantively unreasonable. See United States v. Gerezano-Rosales, 692 F.3d 393, 401 (5th Cir. 2012) (finding a 108-month sentence substantively unreasonable because the district court increased the within-Guidelines sentence it had just imposed by three years based on its belief that defendant responded "disrespectfully" to the sentence). If our court is to continue to accord great deference to district courts' decisions to impose upward variances, we must certainly also do so when reviewing downward variances.

***

For these reasons, though I agree with most of the majority's diligent and well-thought opinion, I respectfully dissent as to Part V(A) vacating Peter Hoffman's sentence as unreasonable.