Skatteforvaltningen v. Markowitz

Opinion

25-916-cv Skatteforvaltningen v. Markowitz

In the

United States Court of Appeals for the Second Circuit

August Term, 2025

No. 25-916

SKATTEFORVALTNINGEN,

Plaintiff-Appellee,

v.

RICHARD MARKOWITZ, JOCELYN MARKOWITZ, JOHN VAN MERKENSTEIJN, ELIZABETH VAN MERKENSTEIJN, BERNINA PENSION PLAN, BASALT VENTURES LLC ROTH 401(K) PLAN, AVANIX MANAGEMENT LLC ROTH 401(K) PLAN, HADRON INDUSTRIES LLC ROTH 401(K) PLAN, CAVUS SYSTEMS LLC ROTH 401(K) PLAN, STARFISH CAPITAL MANAGEMENT LLC ROTH 401(K) PLAN, VOOJO PRODUCTIONS LLC ROTH 401(K) PLAN, AZALEA PENSION PLAN, OMINECA PENSION PLAN, BATAVIA CAPITAL PENSION PLAN, CALYPSO INVESTMENTS PENSION PLAN, ROUTT CAPITAL PENSION PLAN, RJM CAPITAL PENSION PLAN, MICHELLE INVESTMENTS PENSION PLAN, REMECE INVESTMENTS LLC PENSION PLAN, XIPHIAS LLC PENSION PLAN, TARVOS PENSION PLAN, Defendants-Appellants.

On Appeal from the United States District Court for the Southern District of New York.

ARGUED: MAY 13, 2026 DECIDED: AUGUST 31, 2026

Before: NARDINI, LEE, and ROBINSON, Circuit Judges.

Defendants-Appellants Richard and Jocelyn Markowitz, John and Elizabeth van Merkensteijn, and various pension funds they control appeal the entry of judgments against them following a jury verdict finding them guilty of defrauding Skatteforvaltningen (“Skat”), the tax authority of the Kingdom of Denmark, by submitting false claims for tax refunds that Skat fulfilled. Before trial in the United States District Court for the Southern District of New York (Lewis A. Kaplan, District Judge), defendants conceded that they were never entitled to the refunds they requested under the U.S.-Denmark tax treaty, but contended that they had been deceived by their London-based trading partner into believing that they owned shares of stocks in Danish companies and that Danish tax had been withheld from dividends issued by those Danish companies. In defendants’ telling, they were accordingly unaware that the tax refund claims submitted on their behalf to recover taxes purportedly paid to Skat were false. The jury evidently did not believe defendants’ side of the story.

On appeal, defendants argue primarily that Skat’s entire suit is barred by the common law revenue rule, which prohibits courts from hearing actions by foreign nations to enforce their foreign tax laws. They also argue that the district court abused its discretion when it

3

excluded several pieces of evidence that supposedly showcased their non-fraudulent states of mind, and that there is insufficient evidence to support the fraud judgments against Jocelyn Markowitz and Elizabeth van Merkensteijn, the wives of Richard and John, which Skat pursued via an agency theory of liability.

All of defendants’ challenges fail. Because defendants concededly never received any dividends on Danish equities and thus never owed or paid any foreign taxes, Skat’s suit does not seek to enforce foreign tax laws. Defendants may have exploited the Danish system of dividend tax withholding and the U.S.-Denmark tax treaty to defraud Skat. But that does not mean that Skat’s attempt to recover the funds it was defrauded into disbursing was a claim for the collection of foreign taxes within the meaning of the revenue rule.

Likewise, we discern no abuse of discretion in the district court’s evidentiary rulings. And there was ample evidence from which the jury could have and did conclude that Jocelyn and Elizabeth formed agency relationships with their husbands and that the scope of those relationships encompassed the fraud perpetrated on Skat.

Accordingly, we AFFIRM the judgment of the district court.

MARC A. WEINSTEIN (Neil J. Oxford, William R. Maguire, Gregory C. Farrell, on the brief), Hughes Hubbard & Reed LLP, New York, NY, for Plaintiff-Appellee.

4

ANDREW WEINER, Kostelanetz LLP, Washington, D.C. (Nicholas Bahnsen, Kostelanetz LLP, Washington, D.C., Sharon McCarthy, Kostelanetz LLP, New York, NY, on the brief), for Defendants-Appellants.

WILLIAM J. NARDINI, Circuit Judge:

Defendants-Appellants Richard and Jocelyn Markowitz, John and Elizabeth van Merkensteijn, and various pension funds they control appeal the entry of judgments against them following a jury verdict finding them guilty of defrauding Skatteforvaltningen (“Skat”), the tax authority of the Kingdom of Denmark, by submitting false claims for tax refunds that Skat fulfilled. Before trial in the United States District Court for the Southern District of New York (Lewis A. Kaplan, District Judge), defendants conceded that they were never entitled to the refunds they requested under the U.S.-Denmark tax treaty, but contended that they had been deceived by their London-based trading partner into believing that they owned shares of stocks in Danish companies and that Danish tax had been withheld from dividends issued by those Danish companies. In defendants’ telling, they were accordingly unaware that the tax refund claims submitted on their behalf to recover taxes purportedly paid to Skat were false. The jury evidently did not believe defendants’ side of the story.

5

On appeal, defendants argue primarily that Skat’s entire suit is barred by the common law revenue rule, which prohibits courts from hearing actions by foreign nations to enforce their foreign tax laws. They also argue that the district court abused its discretion when it excluded several pieces of evidence that supposedly showcased their non-fraudulent states of mind, and that there is insufficient evidence to support the fraud judgments against Jocelyn Markowitz and Elizabeth van Merkensteijn, the wives of Richard and John, which Skat pursued via an agency theory of liability.

All of defendants’ challenges fail. Because defendants concededly never received any dividends on Danish equities and thus never owed or paid any foreign taxes, Skat’s suit does not seek to enforce foreign tax laws. Defendants may have exploited the Danish system of dividend tax withholding and the U.S.-Denmark tax treaty to defraud Skat. But that does not mean that Skat’s attempt to recover the funds it was defrauded into disbursing was a claim for the collection of foreign taxes within the meaning of the revenue rule.

Likewise, we discern no abuse of discretion in the district court’s refusal to admit certain evidence under the hearsay exception in Federal Rule of Evidence 804(b)(1) or its exclusion of other evidence, mostly relating to defendants’ communications with their attorneys, under Rule 403. Finally, we hold that there was ample evidence from which the jury could have and did conclude that Jocelyn and Elizabeth formed agency relationships with their husbands and that the scope of those relationships encompassed the fraud perpetrated on Skat.

6

Accordingly, we AFFIRM the judgment of the district court.

I. Background

Unless otherwise indicated, the following background information was presented to the jury in the form of exhibits and testimony at trial.

A. Factual history

Defendants Richard Markowitz, a former investment banker, and John van Merkensteijn, a former corporate lawyer (except where otherwise indicated, “defendants”), along with two other individuals, formed a small investment firm called Argre Management in 2005. In 2008, Markowitz was introduced to the concept of dividend tax arbitrage, a trading strategy that exploits how a country taxes dividends issued by domestic companies to nonresident shareholders. In short, most every country has a withholding tax on passive income, including dividends. When a dividend is issued by a company in Country A to a shareholder that is a resident of Country B, the dividend is subject to Country A’s ordinary withholding tax, which is often 30%. However, treaties between countries can minimize or eliminate the withholding tax that the shareholder in Country B owes to Country A. For example, a tax treaty between Country A and Country B might stipulate that certain entities resident in Country B owe only 10% tax on dividends issued by companies in Country A. Because Country A automatically withholds 30% of dividends issued by its domestic companies, under the tax treaty, eligible Country B investors would be entitled to money representing

7

the 20% that is rightfully theirs under the tax treaty—that is, a refund of the 20% they overpaid. In some countries, including Denmark, investors receive a certificate confirming the dividend withholding tax deduction from the custodian of their shares. The investors then make a corresponding claim to the tax authority that imposed the withholding tax to claim their refund.

In 2010, Markowitz spoke with London-based investor Sanjay Shah and others at Shah’s firm, Solo Capital, about an investment opportunity related to dividend arbitrage. At the time, Solo Capital was organizing the “Broadgate fund” in Ireland to capitalize on a reduced tax rate on dividends in Germany under the Germany- Ireland tax treaty. Merrill Lynch would serve as the prime broker for the transaction and Pricewaterhouse Coopers (“PwC”) would be the auditor. Argre invested approximately $10 to $15 million in the Broadgate Fund, including approximately $1 to $2 million apiece from Markowitz and van Merkensteijn. The trade turned a profit. Merrill Lynch and PwC subsequently terminated their involvement in the Broadgate trading because they were concerned that the level of borrowing was excessive relative to the contemplated transaction.

In 2011, Solo Capital proposed another arbitrage transaction to Argre, this time under the U.S.-Germany tax treaty. The planned transaction would involve the purchase of shares via a non-taxable U.S. entity that was entitled to a zero-tax rate on dividends in Germany under the U.S.-Germany tax treaty (versus the merely reduced rate in the Broadgate transaction). Argre hired Michael Ben- Jacob, a lawyer at Kaye Scholer in New York, to advise on the

8

proposed transaction. 1 Ben-Jacob introduced Argre to a charitable entity, Ezra Academy, that would serve as the non-taxable account holder and developed the structure for the transaction: Argre and other investors would provide Ezra Academy with the capital to purchase the German shares in exchange for a percentage of the profits from the dividend tax arbitrage. Deutsche Bank served as custodian for the transaction (the “Ezra transaction”).

Two months into the trading, Deutsche Bank sent a termination notice to Solo Capital. As Markowitz testified, the bank “decided after an internal decision . . . that they did not want to be participating in the dividend arbitrage transactions involving German shares [due to] reputational risks to Deutsche Bank.” App’x at 688. Markowitz explained that around the time of the Ezra transaction, Germany was in the process of “altering their rules and regulations regarding dividend withholding tax . . . such that by the end of 2011” dividend arbitrage transactions would not be “consistent with the laws of the land in Germany.” Id.

The Ezra transaction did not turn out to be profitable because the German government ultimately refused to issue refunds, and Ezra Academy eventually withdrew its refund requests. See App’x at 658. Although Ezra Academy had hedged its positions and was able to sell the German shares, its investors lost money due to the transaction costs involved.

1 Skat subsequently sued Ben-Jacob for his role in the fraud at issue here, but he settled shortly before trial.

9

i. The Danish trading

Argre and Solo Capital turned their attention to Belgium and Denmark. 2 As relevant here, a treaty between the United States and Denmark provides for the full refund of tax withheld on dividends (at the time, 27%) to shareholders that are United States pension plans, which are exempt from taxation. In re Skat Tax Refund Scheme Litig., 356 F. Supp. 3d 300, 308 (S.D.N.Y. 2019) (“Skat I”). Argre and Solo Capital aimed to exploit this zero-tax rate, and Argre accordingly established several pension plans that would engage in this trading, preceded by the creation of sponsoring limited liability companies (“LLCs”).

Because neither Merrill Lynch nor Deutsche Bank—nor any comparable institution—was willing to serve as custodian for the latest dividend arbitrage transactions, Solo Capital became a custodian registered in the United Kingdom. The contemplated trading in Belgium and Denmark differed in a key respect from the prior transactions: Unlike the other transactions, which required investors to put up the capital needed to finance the stock purchases, “Solo Capital stated that it would not require any margin or upfront funding of futures contracts.” Appellants’ Br. at 10. That is, Argre would not need to invest any money in order to reap the rewards of the trading.

2 Argre’s trading activities in Belgium spawned a lawsuit by the Belgian tax authority asserting similar claims to those at issue here. See In re Kingdom of Belgium, Fed. Pub. Serv. Fin. Pension Plan Litig., 680 F. Supp. 3d 460, 465 (S.D.N.Y. 2023). That litigation is ongoing.

10

This was—supposedly—possible thanks to a trading strategy that proceeded as follows. First, Solo Capital, on behalf of the pension plans, would purport to buy large numbers of shares in Danish companies that had announced they would pay a dividend before the ex-dividend date, which is the date on which a company determines which shareholders will be entitled to receive a dividend. The pension funds would purportedly hedge those purchases against the potential that their value would decline by selling exchange-traded futures contracts on the stock. The next day, Solo Capital (on behalf of the pension plans) would lend out that same stock to a stock borrower on a short-term basis for cash collateral that was equal to the purchase price of the stock, thereby “financing” the stock purchases just before the settlement date, which is the date on which a buyer of stock (here, the pension plans) has to provide the cash to the seller and the seller has to tender the shares to the buyer. The purpose of this transaction structure was to give the pension plans the right to a dividend at exactly the right moment—namely, the moment at which the defendants would be entitled to a dividend that would be taxed by the Danish government (i.e., the ex-dividend date). Then, the pension plans would “unwind” the transactions by selling their shares and using the proceeds to refund the cash collateral that was borrowed from the stock borrowers.

We pause here to note that one of the major focuses at trial was the degree to which the mechanics of the trading—including the fact that Argre did not have to invest any money of its own in order to claim entitlements to refunds—would have made it clear to Argre

11

that the entire trading scheme was fraudulent. As discussed below in Section III.A, defendants disputed the degree to which Markowitz and van Merkensteijn were aware that the strategy did not result in the bona fide purchase of Danish stocks. For now, we merely point out that, as defendants conceded before trial, the true reason that Solo Capital did not require any upfront funding for the transaction— which would have required the (at least temporary) purchase of millions of Danish shares—was because its trading strategy did not result in the actual purchase of any Danish shares. As Skat’s expert explained, “[t]hey were closed loop, circular transactions in which a seller ‘sold’ shares it did not have to a [pension] plan that purported to cover that sale by supposedly borrowing those same shares from the plan . . . which never had the shares to begin with.” App’x at 525. As it turned out, the sellers that had supposedly sold the initial shares to the pension plans had themselves borrowed those shares from a stock loan intermediary. And from where did that stock loan intermediary source those shares? From the very same pension plans. Id. at 525–26, 529–30. The trading thus represented a closed “loop,” with no real stocks ever entering the picture: As Skat’s expert summarized, “not only were there no shares, but money didn’t move outside . . . [or] inside the Solo platform. It was all book fictitious entries.” Id. at 526.

The fictitious “trading” began in August 2012, when six of the initial Argre plans (supposedly) collectively purchased tens of millions of shares of the Danish company TDC for approximately $249 million. After these initial (non-existent) trades, the Argre

12

partners created several more pension plans to participate in the scheme, including plans for which their wives were the sole beneficiaries. After Markowitz and van Merkensteijn parted ways with the other Argre partners in 2014, they created 40 new pension plans in order to continue to participate in the arbitrage “trading,” without the fear of getting sued for using plans that former partners “had worked with and worked on behalf of.” App’x at 604. As with the original pension plans, the beneficiaries of these plans included Markowitz and van Merkensteijn’s friends and family. See, e.g., id. at 607 (Markowitz confirming that several LLCs and pension plans were formed for the benefit of his sister and brother-in-law). Across all of the trading from 2012 to 2015, the pension plans purported to buy $75 billion worth of Danish stock.

ii. The reclaim applications

The key piece to the profitability of these structured transactions was the pension plans’ supposed receipt of dividends and concomitant reclaim applications to the Danish government. Thus, in order to submit those applications, Solo Capital-related custodian entities generated dividend statements for the pension plans that purported to list the name of a Danish security the plan owned, the number of shares, the gross dividend, the tax, and the net dividend. The statements represented that the pension plan’s account was credited with an amount equal to the net dividend—the gross dividend less 27%. The statements were subsequently submitted (via a reclaim agent) to Skat, alongside a refund claim application, a cover letter, and an IRS form that certified the pension plans’ U.S. tax

13

residency. Altogether, the documents submitted to Skat represented that the pension plans were the beneficial owners of the relevant stock on the ex-dividend date, that they had received dividends, and that they had paid taxes on those dividends. Each of those representations was false; as defendants conceded at the outset of trial, the trading strategy executed by Solo Capital did not in fact result in the beneficial ownership of any Danish shares. Nevertheless, between 2012 and mid-2015, Skat made hundreds of millions in payments corresponding to the amounts claimed by the pension plans.

In July 2015, the U.K. government alerted Skat to a suspected fraud concerning dividend taxes and identified the reclaim agents and custodians involved, including those linked to Solo Capital. The scheme the U.K. authorities described was similar to a scheme described in a whistleblower complaint submitted to Skat in June 2015. On August 6, 2015, Skat halted all refund payments while it investigated the alleged fraud.

B. Procedural history

Between May and June 2018, Skat filed 140 complaints in eleven different federal judicial districts in the United States alleging that various defendants had defrauded it of millions of dollars by submitting reclaim applications that falsely represented they were entitled to tax refunds. On October 3, 2018, the complaints were consolidated and assigned to Judge Kaplan by the Judicial Panel on Multidistrict Litigation. Skat subsequently filed several additional complaints, including the ones against Richard and Jocelyn

14

Markowitz and John and Elizabeth van Merkensteijn, and their pension plans. Skat claimed that the MDL defendants obtained refunds of tax withholdings on dividends by fraud, aiding and abetting of fraud, and negligent misrepresentation, and sought return of the tax refunds based on mistake, unjust enrichment, and equitable recoupment.

The MDL defendants moved to dismiss Skat’s claims on the basis of the common law revenue rule, which prohibits courts from hearing actions by foreign nations to enforce their foreign tax laws. The district court denied the motion to dismiss, and the MDL defendants’ subsequent motion for summary judgment, reasoning that if Skat “can prove that the defendants never in fact owned the relevant Danish stocks . . . the revenue rule would not apply because the substance of the claims would be for garden variety commercial fraud,” not for a violation of Danish tax law. Skat I, 356 F. Supp. 3d at 308; see also In re Customs and Tax Admin. of the Kingdom of Denmark (SKAT) Tax Refund Litig., 2023 WL 8039623, at *9–*10 (S.D.N.Y. Nov. 20, 2023) (“Skat II”) (summary judgment decision). As the district court explained, Skat did not allege that defendants “participated in tax evasion or otherwise violated Danish tax law.” Skat II, 2023 WL 8039623, at *10. Instead, it alleged that defendants stole money by “pretending that the plans were entitled to tax refunds because they owned Danish shares and received Danish dividends[] on which Danish tax had been paid.” Id. Such a claim, according to the district court, did not fall within the ambit of the revenue rule.

15

Skat also brought fraud claims against Solo Capital and Shah in the United Kingdom in 2018. In April 2021, Justice Andrew Baker of the English High Court dismissed Skat’s claims under the English equivalent of the common law revenue rule—termed “Dicey Rule 3”—because the claims “in substance, sought indirectly to enforce . . . Danish revenue law.” 3 Skat II, 2023 WL 8039623, at *5 (internal quotation marks omitted). In February 2022, the English Court of Appeal reversed Justice Baker’s decision, holding that Dicey Rule 3 did not bar Skat’s claims. Id. On November 8, 2023, the English Supreme Court affirmed the decision of the Court of Appeal. Id.

Back in the United States, trial proceeded in January 2025 on Skat’s fraud, negligent misrepresentation, and restitution-based claims against the Markowitzes, the van Merkensteijns, and their pension plans. 4 At the beginning of trial, defendants conceded that, as a factual matter, Solo Capital did not acquire shares of Danish stock when it purported to execute the dividend arbitrage transactions that undergirded the reclaim applications filed with Skat on behalf of their pension plans. In other words, they abandoned the argument they made at summary judgment, that they were indeed the beneficial

3 The term “Dicey Rule 3” refers to a formulation of the common law revenue rule in a highly respected English treatise on the conflicts of law. See 1 Dicey, Morris & Collins on the Conflict of Laws 107 (15th ed. 2012) (“Rule 3 – English courts have no jurisdiction to entertain an action: (1) for the enforcement, either directly or indirectly, of a penal, revenue or other public law of a foreign State; or (2) founded upon an act of state.”). Although the formulation of the rule remains unchanged in the latest version of the Dicey treatise, it has been renumbered as Rule 20. See 1 Dicey, Morris & Collins on the Conflict of Laws 291-92 (16th ed. 2022).

4 These defendants were selected to be the first of multiple bellwether trials

for entities that defrauded Denmark via false refund claims.

16

owners of the shares, and that they therefore made no material misstatements to the Danish government. Thus, as to the fraud claims, the sole issue at trial was whether Markowitz and van Merkensteijn—individually and as agents of their wives—believed that the statements made to Skat in the form of the reclaim applications and associated documents were true when made, and whether any false statements were made negligently. (Defendants also denied making those false statements negligently.) Defendants therefore adduced evidence purporting to show—and argued to the jury—that at the time the trading was ongoing, they had been deceived by Shah into believing the stocks had actually been purchased. Markowitz and van Merkensteijn testified that they believed that shares were actually being bought, settled, and custodied at Solo Capital, including because they did not know the identity of the entities that initially sold them their shares (and thus did not know those entities were merely stock loan intermediaries that had supposedly sourced the shares from defendants’ own pension plans) and because they had no idea that the dividend statements created by Solo Capital were fabricated.

Skat, for its part, lacked direct evidence that Markowitz and van Merkensteijn were aware of the fraudulent nature of the scheme. Skat therefore introduced evidence purporting to show a mountain of red flags that—it argued—would have made clear to experienced investors such as Markowitz and van Merkensteijn that the trading was fraudulent.

17

For example, as Skat argued to the jury during closing arguments, (1) Solo Capital (and other entities Shah created) served as custodian for these transactions—rather than a large bank—and began to do so only after the custodians (and auditors) for the prior transactions pulled out; (2) although the pension plans that Argre set up had zero or negative balances, the first “trade” supposedly purchased over $530 million in stock (a strategy that repeated over the remaining trades); (3) the trades were implausibly large for the no-name broker that was used as a middleman; (4) the terms of the purported stock loans (the supposed existence of which Markowitz and van Merkensteijn relied on for their claim that they were unaware that Danish stock had not actually been purchased) were highly irregular; (5) liquidity “magically appeared for every trade,” such that “[e]very time [defendants] supposedly bought stock, it just so happened that some stock loan counterparty in the Cayman Islands was interested in borrowing that same amount of shares for the same amount of money,” App’x at 931; (6) as to the transactions that supposedly hedged the risk of the stock supposedly being bought, no party ever posted any margin (i.e., collateral); (7) in each instance where the pension plans supposedly lent their share purchases to (newly-created) stock borrowers, the cash collateral was exactly equal to the purchase price of the shares, notwithstanding changes in market price between the date the stocks were purchased and the date they were lent; (8) and the sheer amount of shares Solo purported to purchase was implausible, such as one set of March 2013 trades in which the Argre plans supposedly purchased approximately 10% of all the outstanding shares of the Danish company Novo Nordisk, a

18

feat that only very large financial institutions, such as BlackRock, State Street, and Bank of New York Mellon, might have been able to do.

Skat also put on evidence of what it described as efforts by Markowitz and van Merkensteijn to disguise the nature of the scheme, which would constitute evidence of knowledge and intent. For example, as Skat likewise argued to the jury, (1) there is no legitimate reason for a single person to have multiple pension plans open in his or her name—much less six, the maximum number Argre signed an individual up for in the course of this trading—because the limits on 401(k) contributions are per person and not per plan; (2) Markowitz and van Merkensteijn instructed their lawyers to name their newly formed LLCs (and concomitant pension plans) with “dissimilar” names that were not all “financial sounding,” like “capital” and “management,” and to instead use words incorporating “rocks” or “fish” and to append “manufacturing, productions, [or] technology” to the LLC names, App’x at 605; and (3) email evidence suggested that the real reason for the numerous pension plans was to keep the reclaims below one million euros per stock so as to avoid raising suspicion.

In turn, defendants put on evidence and argued to the jury that “the so-called red flags were not as suspicious as [Skat’s expert] made them out to be and further that they relied in good faith on countervailing information that the trading was legitimate.” Appellants’ Br. at 16. Nevertheless, the jury found each of the defendants liable on all of Skat’s claims.

19

The district court entered judgments against the individual defendants and their numerous pension plans totaling over $476 million based on Skat’s gross payments on their reclaim applications plus prejudgment interest and minus credits for sums Skat had recovered through other means. In its memorandum concerning the judgment, the district court explained that “[t]he amounts of the money judgments warranted as a matter of law on the fraud claims are exactly equal to those that would be warranted on the negligent misrepresentation claims and would exceed those that would be warranted on the restitution claims,” so it entered judgment “on the fraud claims alone.” Dist. Ct. Dkt. 1526, at 1. It also held that “in the hypothetical absence of liability on the fraud claims, it would enter judgments for the plaintiff and against each consolidated defendant in the precise amounts on the negligent misrepresentation claims.” Id. It additionally stated that “in the hypothetical absence of liability on both the fraud and negligent misrepresentation claims, it would enter judgments in the appropriate cases for the plaintiff and against each consolidated defendant on the restitution claims” in a manner laid out in a series of tables. Id. at 1–2.

Defendants now appeal, arguing primarily that the common law revenue rule bars this entire suit. They also argue that the district court committed evidentiary errors warranting a new trial, and that there is insufficient evidence to support the fraud verdicts against Jocelyn Markowitz and Elizabeth van Merkensteijn. We address each argument in turn.

20

II. The common law revenue rule

Defendants argue that the common law revenue rule bars this suit because the substance of Skat’s claims seeks to enforce foreign (i.e., Danish) tax law. We review this purely legal argument—raised and denied at the motion to dismiss and summary judgment stages of this litigation—de novo. See Keeling v. Hars, 809 F.3d 43, 47 (2d Cir. 2015).

A. Legal background

The common law revenue rule is a principle inherited—as the name suggests—from English courts that “barred courts from enforcing the tax laws of foreign sovereigns.” Pasquantino v. United States, 544 U.S. 349, 352 (2005). “Since the late 19th and early 20th century, courts have treated the common-law revenue rule as a corollary of the rule that, as Chief Justice Marshall put it, ‘[t]he Courts of no country execute the penal laws of another.’” Id. at 360–61 (quoting The Antelope, 10 Wheat. 66, 123, 6 L.Ed. 268 (1825)). As Justice Thomas explained in Pasquantino, the most recent Supreme Court case addressing the revenue rule:

The rule against the enforcement of foreign penal statutes . . . tracked the common-law principle that crimes could only be prosecuted in the country in which they were committed. The basis for inferring the revenue rule from the rule against foreign penal enforcement was an analogy between foreign revenue laws and penal laws.

21

Courts first drew that inference in a line of cases prohibiting the enforcement of tax liabilities of one sovereign in the courts of another sovereign, such as a suit to enforce a tax judgment. The revenue rule’s grounding in these cases shows that, at its core, it prohibited the collection of tax obligations of foreign nations. Unsurprisingly, then, the revenue rule is often stated as prohibiting the collection of foreign tax claims.

Id. at 361 (internal footnotes and citations omitted).

In Pasquantino, the Court considered whether a federal wire fraud prosecution targeting defendants for their scheme to smuggle liquor into Canada to evade Canada’s alcohol import taxes was barred by the common law revenue rule. The Court first determined that defendants’ conduct fell within the meaning of the wire fraud statute, 18 U.S.C. § 1343, because it involved the use of domestic wires to perpetrate a scheme to defraud Canada of its property—namely its right to uncollected excise taxes. Id. at 355–59.

Turning to defendants’ revenue rule argument, the Court held that the prosecution was not barred because no common law revenue rule cases decided by the date of enactment of the wire fraud statute “held or clearly implied that the revenue rule barred the United States from prosecuting a fraudulent scheme to evade foreign taxes,” id. at 360, and because the purposes of the revenue rule did not otherwise suggest that the prosecution was improper, id. at 368. In its review of relevant common law cases, the Court emphasized that the case before it involved “a criminal prosecution brought by the United

22

States in its sovereign capacity” to deter and punish domestic criminal conduct—namely, use of domestic wires to perpetrate fraud, rather than to recover a foreign tax liability, id. at 362, 364, rendering it fundamentally different from cases the object of which was to collect “money that would pay foreign tax claims,” id. at 364.

As relevant here, the Court also noted that the suit “enforce[d]”

Canadian tax law “in an attenuated sense” insofar as it, like earlier cases voiding contracts aimed at circumventing foreign tax law, “encouraged the payment of foreign taxes.” Id. at 366–67. But, as the Court explained, the “revenue rule never proscribed all enforcement of foreign revenue law,” at times permitting “indirect recognition” of foreign tax laws. Id. at 366, 368.

Having concluded that common law precedent did not pose a “clear bar” to the prosecution, the Court turned to “whether the purposes of the revenue rule, as articulated in the relevant authorities, suggest differently.” Id. at 368. The Court concluded that they did not. The “prosecution pose[d] little risk of causing the principal evil against which the revenue rule was traditionally thought to guard: judicial evaluation of the policy-laden enactments of other sovereigns.” Id. And although the prosecution required the “recogni[tion of] foreign law to determine whether the defendant violated U.S. law,” the Court deferred to the Executive branch’s conclusion, in electing to bring the prosecution, that such recognition did not threaten “international friction.” Id. at 369.

23

B. Discussion

Pasquantino and this Circuit’s cases teach that revenue rule issues are resolved on a case-by-case, fact-specific basis rather than by reflexive application of categorical rules. We ask, first, whether the substance of the suit seeks to enforce foreign tax law. This inquiry is informed by analogous common law cases (if any). If we conclude that prior cases pose no clear bar to the suit, we next consider whether permitting the suit contravenes the purposes of the revenue rule, with particular attention to whether the case would involve the judicial adjudication of issues of foreign tax law, which is the most significant policy rationale undergirding the rule.

As discussed below, we conclude that Skat’s suit neither seeks the domestic enforcement of foreign tax law nor threatens the evils the revenue rule is designed to protect against.

i. Whether Skat’s suit constitutes direct enforcement of Danish tax law

The revenue rule bars suits where “the substance of the claim is, either directly or indirectly, one for tax revenues,” Att’y Gen. of Canada v. R.J. Reynolds Tobacco Holdings, Inc., 268 F.3d 103, 130 (2d Cir. 2001) (“Canada”), such that “the whole object of the suit is to collect tax for a foreign revenue, and that this will be the sole result of a decision in favour of the plaintiff.” Id. at 131 (quoting United States v. Harden, [1963] S.C.R. 366, 372–73 (Can.)). “What matters is not the form of the action, but the substance of the claim.” Id. at 130. As we have explained, “[a] suit directly seeks to enforce foreign tax laws

24

when a judgment in favor of the plaintiffs would require the defendants to reimburse them for lost tax revenues. In contrast, indirect enforcement occurs when a foreign state seeks a remedy that would give extraterritorial effect to its tax laws”; for example, “a suit seeking damages based on law enforcement costs is an attempt to shift the cost of enforcing the tax laws onto the defendants.” European Cmty. v. RJR Nabisco, Inc., 355 F.3d 123, 131 (2d Cir. 2004) (“EC I”) (Sotomayor, J.) (discussing Canada), cert. granted, judgment vacated and remanded, 544 U.S. 1012 (2005). 5 Here, defendants conceded that they never beneficially owned Danish stocks, and thus they conceded that they were not issued dividends from which Denmark withheld taxes subject to refund. In this way, they abandoned any argument that Skat’s suit would entail an adjudication of whether they did in fact beneficially own Danish stocks, which in turn would determine whether they had any tax liability under Danish law—a determination that, they contended at the outset of this litigation, would indirectly enforce Danish tax law. See Skat II, 2023 WL 8039623, at *7–*10 (concluding that resolving disputed questions about beneficial ownership under Danish law would not constitute impermissible indirect enforcement). Accordingly, we consider only their claim that Skat’s suit directly enforces Danish tax law.

5 Pasquantino was decided while a petition for certiorari in EC I was pending. After Pasquantino was issued, the Supreme Court vacated EC I and remanded for reconsideration in light of Pasquantino. This Court decided to reinstate its prior judgment because “the factors that led the Pasquantino Court to hold the revenue rule inapplicable to [18 U.S.C.] § 1343 smuggling prosecutions are missing here.” European Cmty. v. RJR Nabisco, Inc., 424 F.3d 175, 181 (2d Cir. 2005) (“EC II”) (Sotomayor, J.).

25

That argument proceeds as follows: According to defendants, Skat “erroneous[ly]” issued tax refunds, and this suit represents Skat’s effort to recover the mistakenly issued refunds. Appellants’ Br. at 22. Because those refunds were “issued in the normal course of administering the Danish tax system,” defendants construe Skat’s efforts to recover the stolen funds as “tax claims” by a foreign sovereign that are barred by the revenue rule. Id. at 22–23; see Pasquantino, 544 U.S. at 361.

We disagree. In contrast to cases in which a foreign sovereign has attempted to use the U.S. courts to collect taxes a defendant unduly evaded (and therefore owed), see Canada, 268 F.3d at 131; European Cmty. v. RJR Nabisco, Inc., 424 F.3d 175, 181 (2d Cir. 2005) (“EC II”) (Sotomayor, J.), these defendants concededly never owed any Danish taxes at all. Defendants admit that they never owned any Danish shares and thus received no dividends from which any Danish taxes were withheld such that they might be entitled to claim a refund. They admit that the “refund claims” submitted on their behalf were works of fiction aimed at convincing Skat to hand over money. We agree with the district court that Solo Capital’s conduct (on defendants’ behalf) is more akin to “garden variety commercial fraud” than tax evasion, and thus that Skat’s efforts to recover the money stolen does not amount to a tax collection claim. Skat I, 356 F. Supp. 3d at 308. “The fact that the victim of the fraud happens to

26

be the government, rather than a private party, does not lessen the injury.” Pasquantino, 544 U.S. at 356. 6

Take an example. Suppose that Solo Capital had submitted an entirely fabricated invoice for services rendered to Skat, complete with supporting documents, and Skat paid the invoice. If Skat brought suit in the United States to recover the funds it paid due to the fraudulent invoice, the revenue rule would obviously not be implicated. Defendants have committed exactly the same genre of fraud, regardless of whether the instrumentality they used was the Danish tax system.

Our conclusion is bolstered by the fact that this suit does not threaten the “principal evil” the revenue rule is designed to protect against: “judicial evaluation of the policy-laden enactments of other sovereigns.” Pasquantino, 544 U.S. at 368. Because defendants conceded that they never owned Danish shares as a factual matter, and thus that no dividend was ever issued to them and no taxes were

6 The U.K. Supreme Court reached the same conclusion in Skat’s twin litigation in England. As the Supreme Court reasoned, the “substance” of Skat’s claim was not for “sums due as tax in Danish law” nor a claim that the defendants were liable to Skat because “they cheated [Skat] out of tax which was due to it.” Add. to Appellee Br. at 34. There, as here, there was no allegation of any “unpaid tax.” Id. Instead, the substance of the claim was “to recover payments made by [Skat] which were induced by fraud and to which the recipients were not entitled on any basis.” Id. at 35. Defendants assert that because the U.K. Supreme Court did not consider the constitutional separation of powers concerns animating (in part) the U.S. revenue rule, including applicable treaties, the U.K. decision is “distinguishable.” Reply Br. at 12. As discussed below, we do not perceive those concerns to militate strongly (if at all) in defendants’ favor.

27

ever withheld, the sole issue the jury was called upon to decide was whether defendants were aware that they never owned any shares at the time they submitted false reclaim applications. Thus, the jury (like the district court) was never called upon to interpret or apply Danish tax law, and indeed, the success of Skat’s claims hinges not upon a question of Danish tax law but instead upon the intent of the defendants.

Indeed, we have previously reasoned that a suit that needed to prove only that defendants intended to defraud a foreign government did not implicate the revenue rule precisely because the validity of a foreign tax law was not at issue. In United States v. Trapilo, we held that prosecuting various New York-based defendants for a scheme to defraud the Canadian government of tax revenue did not implicate the common law revenue rule. 130 F.3d 547, 551–53 (2d Cir. 1997). This was so, we explained, because proving defendants’ intent to defraud “does not hinge on whether or not [they] were successful in violating Canadian revenue law.” Id. at 552. Absent any “obligation to pass on the validity of Canadian revenue law,” the revenue rule was not implicated. Id. at 553. Just so here: The only issue the jury needed to decide was whether defendants intended to defraud the Danish government.

Our cases subsequent to Trapilo do not undermine our conclusion that this case does not involve direct enforcement of foreign tax law. We have indeed twice held that the revenue rule bars civil suits by foreign sovereigns under the Racketeer Influenced and Corrupt Organizations Act, 18 U.S.C. §§ 1961–1968 (“RICO”),

28

attempting to recover tax revenues lost—that is, taxes evaded—due to smuggling and money laundering (and related misdeeds). See Canada, 268 F.3d at 131; EC I, 355 F.3d at 127. But, crucially, in those cases foreign sovereigns sought to recover damages representing taxes that had been illegally evaded under those sovereigns’ tax laws—that is, to recover for violations of their tax laws. See Canada, 268 F.3d at 105–06 (seeking damages based on, inter alia, revenue lost “from the evasion of tobacco duties and taxes”); EC I, 355 F.3d at 127; EC II, 424 F.3d at 182 (holding that Pasquantino did not disturb the holding of EC I). In such a case, the damages claim quite obviously constituted an effort by a foreign sovereign to directly enforce its tax laws.

Even though this case similarly features a civil fraud claim brought by a foreign sovereign, here, by contrast, Skat does not seek the enforcement of Danish tax laws. The parties agreed that defendants did not owe Danish taxes, and they did not pay Danish taxes. Indeed, they agree that they were not bona fide Danish taxpayers at all. If, as in Canada, Skat’s damages claim arose from taxes unlawfully unpaid by defendants, its suit would indeed implicate the revenue rule because the “substance” of the claim would be a claim for tax collection. But that is not the case here. Although the Danish tax code provided the means by which defendants stole money from Skat, that choice on behalf of defendants does not render Skat’s suit a tax suit.

Defendants complain that our holding would carve out a “fraud exception” to the revenue rule. But, as explained, we do not

29

find the revenue rule inapplicable to Skat’s suit simply because its claims sound in fraud. Rather, we conclude that the revenue rule is inapplicable in light of the nature of the specific fraud scheme here. Notably, all parties agree that if defendants’ fraud were intermingled with violations of tax law, such as the inflation of refund claims or the fabrication of business expenses to obtain a deduction, Skat’s suit would indeed be barred by the revenue rule. But that is not the case here.

Further, the single case defendants cite in support of the argument that fraud cases necessarily fall within the ambit of the revenue rule fails to carry the weight they ascribe to it. Banco Do Brasil, S.A. v. A.C. Israel Commodity Co., 12 N.Y.2d 371 (1963), concerned an action brought by the state-owned Bank of Brazil “to recover damages for a conspiracy to defraud the Government of Brazil of American dollars by illegally circumventing the foreign exchange regulations of Brazil.” Id. at 374. The New York Court of Appeals held that, contrary to the plaintiff’s argument, the Bretton Woods Agreement (a multilateral treaty to which the United States was signatory) did not provide an avenue for recovery for the bank. See id. at 376. Having held there was no cause of action, the Court of Appeals also pointed out that “there is a remedial consideration which bars recovery in this case,” namely that the bank (an instrumentality of Brazil) was “seeking, by use of an action for conspiracy to defraud, to enforce what is clearly a revenue law,” and so regardless of the text of the Bretton Woods Agreement, recovery was not available because “one State does not enforce the revenue

30

laws of another.” Id. at 377. This (arguable dicta) stands for the uncontroversial proposition that suits to recover evaded taxes are barred by the revenue rule, not that fraud claims that happen to involve taxation authorities but that do not involve questions of tax liability fall within the scope of the revenue rule.

Defendants also argue that the tax treaties between the United States and Denmark, and the United States and Japan, treat recovery of fraudulent refunds as a “tax claim,” and so Skat’s claim must be read as a tax claim that is barred by the revenue rule. Appellants’ Br. at 23. As discussed below, tax treaties with other countries are relevant to the revenue rule analysis (in addition to their role as general interpretive aids) to the extent they embody the decisions of the branches of our government that are constitutionally tasked with dealing with foreign affairs and to the extent adhering to their terms enables international harmony and upholds the separation of powers. See Canada, 268 F.3d at 122; EC II, 424 F.3d at 182 n.9 (noting that in Pasquantino “[t]he Supreme Court did not criticize Canada for relying on [] treaties as evidence of congressional intent” when it observed, citing Canada, that the terms of U.S. tax treaties did not convince the Court that petitioners’ scheme fell outside the terms of the wire fraud statute). Thus, to the extent we were to adopt an interpretation of Skat’s suit that was at odds with the terms of a U.S. treaty, we might indeed run afoul of the purposes of the revenue rule.

Setting aside the foreign policy considerations (for now), we do not think the U.S.-Denmark and U.S.-Japan treaties shed much light on the proper framing of Skat’s claim. The most recent amendment

31

to the U.S.-Japan treaty provides that the state-parties will assist one another “in the collection of taxes,” but, specifically:

[O]nly in the collection of the following revenue claims: (a) a revenue claim in respect of a company . . . [and] (b) a revenue claim in respect of an individual. However, if the individual is a national of the Contracting State from which assistance is requested (hereinafter referred to as the ‘requested State’) at the time the application for assistance is received, assistance shall be lent only for revenue claims with respect to which the individual or a person acting on behalf of the individual: (i) has filed a fraudulent tax return or a fraudulent claim for refund; (ii) has willfully failed to file a tax return to evade taxes; or (iii) has transferred assets into the requested State to avoid collection of the revenue claim.

Protocol Amending the Convention Between the Government of the United States of America and the Government of Japan for the Avoidance of Double Taxation and the Prevention of Fiscal Evasion with Respect to Taxes on Income, Japan-U.S., art. XIII, Jan. 24, 2013, T.I.A.S. No. 19-830.

Defendants argue that the agreement to provide collection assistance for certain revenue claims, including an individual’s “fraudulent claim for refund,” implies that any effort by a state to recover money disbursed pursuant to such a fraudulent claim is a “revenue claim.” But this interpretation ignores the first paragraph of the relevant article, which makes clear that assistance is provided pursuant to the treaty only for the collection of taxes. Reading the two

32

paragraphs of the article together, it is clear that the treaty contemplates collection assistance where an inflated refund claim resulted in an underpayment of taxes, resulting in taxes actually owed. Thus, because the treaty defines a revenue claim as a claim “for the collection of taxes,” its definition lacks any analytical force here, where Denmark does not seek to collect any unpaid or improperly refunded taxes.

Defendants’ argument based on the U.S.-Denmark tax treaty misses the mark for similar reasons. That treaty covers certain taxes specified in article 2 of the treaty—namely, U.S. federal income and excise taxes, and Danish income taxes and taxes imposed under the Hydrocarbon Tax Act. Convention for the Avoidance of Double Taxation and the Prevention of Fiscal Evasion with Respect to Taxes on Income, U.S.-Den., art. 2, Aug. 19, 1999, T.I.A.S. No. 13056. Article 27, in turn, sets forth the procedures under which one state may request assistance from the other state “in the collection of taxes referred to in Article 2,” which the treaty defines as a “revenue claim,” but clarifies that “[n]o assistance will be provided” if the taxpayer at issue can prove that, at the relevant time, it was a citizen or entity of the state from which assistance was requested. Id. art. 27(8). Defendants argue that this language implies that Skat cannot sue them or their pension plans because they are U.S. entities. We disagree. The treaty covers only certain Danish taxes that a taxpayer might owe, none of which can fairly be said to encompass the funds Skat disbursed due to fraud. The treaty speaks not at all to Skat’s ability (or inability) to bring a fraud claim relating to taxes never owed.

33

At bottom, defendants have failed to point us to any case in which a defrauded foreign sovereign taxing authority was barred from recovery in the United States against defendants who never owed taxes, much less a well-established rule that would block such an action. See Pasquantino, 544 U.S. at 360; id. at 368 (“Even if the present prosecution is analogous to the indirect enforcement cases on which petitioners rely, those cases do not yield a rule sufficiently well established to narrow the wire fraud statute in the context of this criminal prosecution.”). Nor have they pointed us to any other authority making clear that claims Skat brought in this suit constitute the direct enforcement of Danish tax law.

We therefore turn to whether permitting this suit undermines the purposes of the revenue rule. See id. at 368; Canada, 268 F.3d at 109 (examining “the policies underlying the revenue rule, and the rule’s congruence with the international tax policies pursued by . . . our government” to determine whether they “support[] the conclusion that the revenue rule is applicable” to Canada’s suit).

ii. Whether the Suit undermines the purposes of the revenue rule—namely, respect for sovereignty, international relations, and the separation of powers

Pasquantino counsels that where precedent does not provide a clear bar to a given case, the Court should turn to “whether the purposes of the revenue rule . . . suggest differently.” 544 U.S. at 368; see also EC I, 355 F.3d at 131–32 (“Even if the substance of the claim invokes foreign tax laws, the revenue rule will not be triggered where

34

the sovereignty and extraterritoriality concerns that inform the rule’s application are not present.”). Those purposes, as enunciated in Pasquantino, include barring judicial evaluation of the policy enactments of other sovereigns to reduce the risk of international friction. Id. at 368–69. And, as discussed above, although the text of particular tax treaties was not at issue in Pasquantino, Canada instructs that tax treaties constitute policy judgments by the “the political branches of our government” to which we must defer because “foreign relations policy” is an area “where the political branches have primacy.” Canada, 268 F.3d at 119.

None of these considerations counsels in favor of application of the revenue rule here. First, because defendants abandoned their only argument that would have required the court to interpret and apply Danish law—namely, their argument that they beneficially owned Danish shares—the need for any judicial evaluation of Danish law was entirely obviated. 7 Thus, as discussed, in the particular circumstances of this case, the “principal evil” prevented by the revenue rule is not present. Pasquantino, 544 U.S. at 368.

Second, as to the role of the political branches and the concomitant importance of respecting the separation of powers, we have already explained why we do not think Congress and the Executive have spoken (much less clearly spoken) against permitting suits of this kind via the tax treaties with Denmark and Japan.

7 We do not intend to imply that if defendants had argued that they had beneficially owned the stocks the revenue rule would apply—a question we need not address. The potential tax liability here would have arisen not from mere ownership of the stock, but from issuance of a dividend.

35

Because defendants’ policy arguments flow entirely from their interpretation of those two treaties, which we have rejected, we are likewise unpersuaded by defendants’ insistence that permitting this suit weakens the political branches’ authority over tax treaty negotiations or that it will engender ill will by giving Denmark a right that it did not bargain for, unlike (according to defendants) Japan. To the contrary, foreign relations are arguably furthered by permitting a government defrauded by United States citizens to recover the funds it was induced into dispatching; shutting the courthouse doors to foreign sovereigns defrauded by American citizens would more likely create rather than prevent “international friction.” Id. at 369.

We are forced to speculate about the impact of this case on international relations and the separation of powers because the Executive branch never weighed in on this case. We acknowledge that the Executive’s lack of involvement here renders this case different from Pasquantino (and Trapilo), where the Court’s conclusion as to the policy implications of the government’s suit was grounded in part in its deference to the Executive’s opinion on the litigation before it. See Pasquantino, 544 U.S. at 369; EC II, 424 F.3d at 181 (noting that “the involvement of the United States government was a key factor in determining the outcome of Pasquantino” and contrasting it with the civil RICO lawsuit “brought by foreign governments” it was considering). Nevertheless, as discussed, we do not see the foreign policy implications of Skat’s suit as conflicting with the revenue rule’s purposes, nor have we been presented with any arguments to that effect aside from defendants’ limited treaty-based arguments.

36

We therefore hold that, on the particular facts of this very particular case, the revenue rule does not bar Skat’s lawsuit.

III. Evidentiary arguments

Defendants next argue that the district court erroneously excluded evidence that would have bolstered their argument to the jury that Markowitz and van Merkensteijn were unaware of Solo Capital and Shah’s fraud and instead acted with “non-fraudulent states of mind.” Appellants’ Br. at 35. Because defendants conceded at the beginning of trial that the refund claims indeed falsely represented their ownership of Danish stocks, receipt of dividends from which taxes were withheld, and concomitant entitlement to refunds, their sole defense was that they had been bamboozled into believing that the refund claims submitted on their behalf were legitimate. Thus, the excluded evidence—like the rest of the evidence presented at trial—would be probative only to the extent it shed light on the “central question in this case”: “whether Richard and John were in on” the scheme orchestrated by Shah “or whether they themselves were deceived.” Id. at 30.

Preserved challenges to the admissibility of evidence are reviewed for abuse of discretion, “and we will disturb an evidentiary ruling only where the decision to admit or exclude evidence was manifestly erroneous.” United States v. Kandic, 134 F.4th 92, 99 (2d Cir. 2025) (quoting United States v. McGinn, 787 F.3d 116, 127 (2d Cir. 2015)). Even where the district court abused its discretion, “[w]e will reverse only where the improper admission or exclusion of evidence

37

affects a substantial right of one of the parties.” Phoenix Assocs. III v. Stone, 60 F.3d 95, 104–05 (2d Cir. 1995) (internal quotation marks omitted); accord Rossbach v. Montefiore Med. Ctr., 81 F.4th 124, 137 (2d Cir. 2023). “Whether an evidentiary error implicates a substantial right depends on the likelihood that the error affected the outcome of the case.” Utica Mut. Ins. Co. v. Munich Reinsurance Am., Inc., 7 F.4th 50, 62 (2d Cir. 2021) (citation omitted). “Thus, an evidentiary error in a civil case is harmless unless the appellant demonstrates that it is likely that in some material respect the factfinder’s judgment was swayed by the error.” Id. at 62–63 (citation modified).

We discuss each of defendants’ challenges in turn.

A. The district court’s refusal to admit Shah’s testimony from the United Kingdom litigation under Rule 804(b)(1)

Defendants first argue that the district court erroneously refused to admit testimony from Skat’s United Kingdom litigation under Rule 804(b)(1)’s hearsay exception for former testimony offered against a party (here, Skat) who had an opportunity and similar motive to develop the testimony. Defendants sought to admit excerpts from Shah’s U.K. cross-examination because Shah was not deposed in this case 8 and was unavailable at trial.

8 The district court denied defendants’ motion (which Skat joined conditionally) seeking issuance of a letter of request for international judicial assistance to obtain testimony from Shah. As the district court explained, the April 19, 2024, motion was filed after its disposition of the parties’ summary judgment motions and mere weeks after its pretrial order. The district court determined that

38

A party had a “similar motive” within the meaning of Rule 804(b)(1) if it was “on the same side of the same issue” and had a “substantially similar degree of interest in prevailing on that issue” in both proceedings. United States v. DiNapoli, 8 F.3d 909, 912 (2d Cir. 1993). “[I]n assessing similarity of motive under Rule 804(b)(1),” the court “must consider whether the party resisting the offered testimony at a pending proceeding had at a prior proceeding an interest of substantially similar intensity to prove (or disprove) the same side of a substantially similar issue.” Id. at 914–15.

By way of background, Skat sued Shah, Solo Capital, and several other entities in England for fraud over the same conduct at issue here. Notably, several pension plans were named as defendants, but not any of the pension plans controlled by the individual defendants here. During Skat’s nearly two-week-long cross-examination of Shah, Skat asked Shah whether he had explained to Argre that Solo Capital had not actually acquired Danish shares. Shah replied: “I don’t recall explaining that to them, no. I explained to them the activities of the pension plan from the pension plan’s perspective.” App’x at 1507. Skat further asked why Argre would “have been willing to participate in a scheme where they couldn’t possibly have understood why it was that they didn’t need

issuing the request at that late date risked delaying the trial, particularly because defendants refused to assure the court that trial would proceed as scheduled regardless of their ability to depose Shah. It also concluded that the testimony defendants sought from Shah would be cumulative of other evidence and witnesses available in the United States. Defendants have not challenged this ruling on appeal.

39

to obtain any funding” unless Shah explained the mechanics of the trading to them, a process he called “the loop.” Id. Shah replied:

No, I didn’t explain to [Argre] the loop. The loop is quite a recent concept for me and it was necessary for them to understand that they were lending the shares out in order to raise cash to purchase the shares. I think that is all they needed to know. They seemed to be happy with that. I don’t have any recollection of any hesitancy on their side. As well as that, they were also quite resourceful in terms of getting their own advice and they were—if needed, they were able to fund these accounts. But we told them that on this occasion they didn’t need to do that.

Id. In response to a question from Skat, Shah asserted that he had given Argre less than “half the story,” and instead had given Argre “probably an eighth of the story.” Id.

As discussed, at trial defendants did not argue that the “loop”

resulted in actual legal entitlements to Danish stocks (and their dividends), but instead argued that “almost all of that loop was entirely invisible” and “purposely hidden” from Markowitz and van Merkensteijn. App’x at 420. Thus, in defendants’ view, these portions of Shah’s testimony amount to “direct evidence” that Shah “never informed Richard and John that Solo Capital did not acquire Danish Shares.” Appellants’ Br. at 32. As to Rule 804(b)(1) specifically, defendants argue that Skat had the same motive when it was questioning Shah in the U.K. litigation, notwithstanding the absence of these defendants in the U.K. lawsuit, because the other pension

40

plans that were named as defendants also denied knowledge of Shah’s fraud. Put another way, defendants argue that because Skat had an interest in proving other unrelated pension plan defendants’ knowledge of Shah’s fraud, it had—and acted on—a substantially similar motive with respect to the Argre plans.

We cannot say that the district court abused its discretion when it determined that the issues in the U.K. litigation were insufficiently similar within the meaning of Rule 804. There was no need for Skat to establish the “knowledge and intent” of Markowitz and van Merkensteijn in the U.K. litigation, as neither Markowitz and van Merkensteijn nor any entities they controlled were named as defendants there. App’x at 879. Instead, as the district court explained, “the issue of whether Shah disclosed [the true nature of the trading] to the defendants in this case was extremely incidental to the London case,” because the “London case did not involve the liability of any of these defendants.” Id. at 906 (emphasis added). The district court recognized that if Skat had elicited an admission on cross that Shah had disclosed the nature of the fraud to the defendants, that “would have been of some help on credibility” in the U.K. litigation, but overall “any impact on the outcome of the London case would be highly speculative.” Id. Despite defendants’ protestations, we agree with the district court that “wanting to prevail,” Appellants’ Br. at 34– 35, is not a sufficiently similar motive where, as here, a plaintiff seeks relief against two wholly unrelated sets of defendants and where the only issue at trial is the mental state of one set.

41

We therefore hold that the district court did not abuse its discretion when it excluded this testimony.

B. The district court’s exclusion of evidence under Rule 403

Defendants next argue that the district court abused its discretion when it excluded evidence under Rule 403 that showed they received legal advice and submitted certain forms to the Internal Revenue Service (“IRS”). Rule 403 provides that a district court “may exclude relevant evidence if its probative value is substantially outweighed by a danger of . . . unfair prejudice, confusing the issues, misleading the jury, undue delay, wasting time, or needlessly presenting cumulative evidence.” Fed. R. Evid. 403. “Evidence is said to have probative value if it tends to prove or actually proves a proposition.” United States v. Garnes, 102 F.4th 628, 638 (2d Cir. 2024) (quoting United States v. Jamil, 707 F.2d 638, 642 (2d Cir. 1983)).

“We grant considerable deference to a district court’s determination to exclude relevant evidence under this rule, recognizing that Rule 403 calls for ‘on-the-spot balancing.’” Id. at 635 (quoting Sprint/United Mgmt. Co. v. Mendelsohn, 552 U.S. 379, 384 (2008)). District courts are “well-positioned to carry out [such balancing] due to their proximity to the detailed circumstances of the case. Accordingly, we will reverse a district court’s ruling on admissibility under Rule 403 as an abuse of discretion only if its conclusion is ‘arbitrary or irrational.’” Id. (quoting United States v. Awadallah, 436 F.3d 125, 131 (2d Cir. 2006)).

42

“[I]n making a Rule 403 determination, a court may consider other available means of proof or ‘evidentiary alternatives.’” Id. at 638 (quoting Old Chief v. United States, 519 U.S. 172, 184 (1997)). “The probative worth of any particular bit of evidence is obviously affected by the scarcity or abundance of other evidence on the same point.” Old Chief, 519 U.S. at 185 (quoting, in a parenthetical, 22 C. Wright & K. Graham, Federal Practice and Procedure § 5250, at 546–47 (1978)).

i. The Danish law firm opinion sent to Solo Capital that was shared with Argre

In June 2012, a Solo Capital employee forwarded to Markowitz a legal opinion from a Danish law firm. The opinion evaluated the Danish tax implications of contemplated transactions involving the purchase of Danish equities by a United States pension fund before the ex-dividend date. The letter had two top-line conclusions: first, that the contemplated transactions “should not attract any Danish tax other than withholding tax on dividends,” App’x at 712, and second, that a U.S. pension fund should be entitled to claim a full refund of the Danish dividend withholding. Defendants sought to introduce this letter into evidence, which Skat opposed. In support of defendants’ argument that the letter should be admitted, Markowitz testified that he understood the trading described in that letter to be “virtually identical to [the arbitrage trading they] had been doing” in Germany and that the letter was “critically important” to his decision to proceed with arbitrage trading in Denmark. Id. at 705.

Following argument, the district court excluded the letter. As the district court explained, because defendants had conceded that as

43

a factual matter they were not entitled to any refunds, and accordingly that the assumptions underlying the letter—which assumed, inter alia, the actual purchase of Danish shares—did not apply to the transactions that defendants actually engaged in, the letter could be relevant at trial only if it supported defendants’ claim that they did not have culpable knowledge of Shah’s fraud. The district court took the view that the letter had little probative value for that proposition; it showed at most that Shah lied to his tax lawyers and that those (faulty) assumptions were repeated in the law firm’s letter. At the end of the day, the court explained, the jury would believe Markowitz’s testimony (and the other defendants’ testimony) that he was unaware of the fraud or they would not, and “they are not going to do it . . . to any greater degree or with any greater probability with this opinion” than without it. Id. at 713.

On appeal, defendants argue (wishfully) as to probative value that the “opinion confirmed Richard’s understanding that Solo Capital would acquire Danish shares and showed that the dividend arbitrage trading strategy worked, further allaying any potential concerns.” Appellants’ Br. at 37. But the district court did not exceed its discretion in concluding that this forwarded opinion, prepared at the behest of Solo Capital, does not tend to make it more likely that the jury would have believed Markowitz’s testimony that he thought the transactions were legitimate. The letter, which did not reflect legal advice that Markowitz himself had requested, details the legal ramifications of transactions that concededly never occurred. Its probative value was therefore minimal, and the district court

44

appropriately balanced it against the comparatively higher likelihood that admitting the letter would “waste a tremendous amount of time.” App’x at 713.

We also conclude that any purported error here was harmless.

Lore v. City of Syracuse, 670 F.3d 127, 155 (2d Cir. 2012) (“[A]n erroneous evidentiary ruling warrants a new trial only when a substantial right of a party is affected, as when a jury’s judgment would be swayed in a material fashion by the error.” (internal quotation marks omitted)). Notably, Markowitz was allowed to testify that receiving this letter was “critically important” to his decision to proceed with the Danish trading, communicating (permissibly or not) that sign-off from Danish lawyers was crucial to his decision to move forward. App’x at 705. This testimony supported defendants’ trial theme that they endeavored to comply with the law—and that they were deceived into engaging in illegal trading—which is the same purpose introducing the letter itself would have served. And Markowitz’s own testimony, of course, is more probative as to his state of mind than the actual opinion letter itself.

We therefore reject the argument that the district court abused its discretion when it declined to admit the letter, and further hold that any putative error was harmless in light of Markowitz’s testimony about the impact of the opinion on his state of mind.

45

ii. Certain emails with Argre’s law firm

Defendants also attempted to show their “non-fraudulent states of mind” by introducing evidence of their communications with their own law firm Kaye Scholer. Although defendants disclaimed a formal advice-of-counsel defense, they sought to introduce evidence of communications with Kaye Scholer to show “that they had a good faith belief that the representations made in the reclaim forms submitted to S[kat] were accurate.” In particular, they point to “[t]he lengths to which [they] went to hire attorneys to ensure that the pension plans were properly qualified, confirm that the plans satisfied the requirements of the tax treaty between the United States and Denmark, make certain the trade strategy adhered to relevant U.S. and foreign securities laws, and involve lawyers in the execution of the strategy.” Dist. Ct. Dkt. 1233 at 3, 6. Before trial, Skat filed a motion in limine seeking to preclude defendants from arguing that Kaye Scholer’s involvement in and advice concerning their dealings with Solo Capital evidenced their good faith in making representations to Skat about their ownership of shares under Danish law. Id. at 7, 20.

The district court granted Skat’s motion. It explained, first, that communications with Kaye Scholer were clearly irrelevant to defendants’ mental state because Kaye Scholer—an American law firm—advised defendants only about American legal issues and therefore could not shed light on whether defendants knew their representations about stock ownership under Danish law were false.

46

App’x at 366. 9 Thus, according to the district court, the raft of emails defendants sought to introduce had little probative value, and “risk[ed] suggesting to the jury that, because lawyers were involved to some degree with one aspect of events, the defendant was entitled to conclude that he was acting within the law with respect to some other aspect of events.” Id. (internal quotation marks omitted). Thus, the court ruled that defendants could not argue that “Kaye Scholer’s involvement in and advice concerning their scheme evidences their good faith in making the beneficial ownership representations to S[kat],” but may “introduce such evidence insofar as they seek to illustrate the mechanics of the trading strategy and show the jury that they did not ‘act[] alone’ in filing their claims with S[kat].” Id. at 367.

At trial, the district court excluded some—but not all—of defendants’ emails with Kaye Scholer in line with its earlier preclusion ruling. Overall, as defendants themselves acknowledge, the emails show at most that “Richard and John were forthright with Kaye Scholer and, in turn, that Kaye Scholer understood the structure and purpose of the transactions on which Richard and John were asking for advice.” Appellants’ Br. at 41.

We agree with the district court that the introduction of the excluded emails would have shed little light on the only issue relevant at trial: the degree to which Richard and John knew of Shah’s fraud. For example, one of the excluded emails concerns “tax and

9The district court allowed defendants to seek to introduce evidence of one instance in which Kaye Scholer advised Markowitz about Danish securities reporting requirements.

47

compliance issues” surrounding payments to a Shah-controlled Cayman Islands entity. App’x at 1457. In the email, Kaye Scholer advised defendants that they saw no reason defendants should not make a payment to the entity, subject to a few points of U.S. law— namely, tax law and the Patriot Act. Id. This email, like the other emails advising defendants on U.S. law, sheds no light on defendants’ state of mind as to their ownership of shares of Danish stocks, as the district court correctly observed. See also App’x at 1235, 1285–87 (excluded emails detailing the request and receipt of legal advice relating to reporting requirements under U.S. tax law). Similarly, defendants’ emails with Kaye Scholer about the Ezra transaction shed little to no light on “the soundness of the later transactions involving Danish shares,” Appellants’ Br. at 42, both because the circumstances surrounding that trading were materially different—including the initial involvement of a well-known institutional custodian and the actual investment of millions of dollars in capital—and because defendants’ state of mind vis-à-vis those earlier transactions sheds minimal if any light on their beliefs about their ownership of shares under Danish law. See App’x at 1227–32 (excluded emails from Markowitz to Kaye Scholer forwarding a diagram prepared by Solo Capital explaining the mechanics of the contemplated Ezra transaction); id. at 1953–62 (excluded email summarizing advice on U.S. law issues that Kaye Scholer had rendered to defendants to date relating to ex-dividend trading, around the time of the Ezra transaction). We therefore agree with the district court that the excluded emails with Kaye Scholer had minimal probative value and threatened significant prejudice to the extent that they suggested that

48

defendants’ communication with counsel about dividend arbitrage trading implied that defendants are innocent of fraud. The district court thus acted well within its discretion in excluding them.

Moreover, contra defendants’ arguments on appeal that the district court “nipped nearly all” evidence of defendants’ engagement with Kaye Scholer “in the bud,” Appellants’ Br. at 40–41, the trial record is replete with references to Markowitz and van Merkensteijn’s reliance on Kaye Scholer, both in the form of admitted exhibits and statements to the jury. Any suggestion that evidence of their reliance on the firm was wholly absent at trial is belied by defendants’ own closing statements, which repeatedly referenced defendants’ communications with Kaye Scholer in an attempt to communicate that they had no knowledge that they were engaged in fraudulent conduct. See, e.g., App’x at 963 (“[Richard and John] are talking to the lawyers, ladies and gentlemen. Let’s just think about that. It’s an email between Rich and John and their lawyers, where one of the lawyers is proposing names [for new LLCs and pension plans]. In your experience, do people openly discuss efforts to conduct a coverup with their lawyers? It’s just silly.”); id. at 981 (“And what you’ve seen also is that many of the reclaim forms that are filed by these friends and family have Kaye Scholer’s names all over them, the same law firm on all of these different reclaim applications for all of these different pension funds. They are not spreading them around to try to make it look secret or hidden. They are perfectly comfortable with the same law firm on all of these reclaim applications.”). We therefore also hold that any claimed error in the refusal to admit the challenged

49

emails was harmless in light of the evidence and argumentation presented to the jury regarding Markowitz and van Merkensteijn’s communications with their lawyers.

iii. Certain tax forms submitted to the IRS

The final evidentiary ruling defendants challenge is the district court’s exclusion of specific tax forms they submitted to the IRS reporting on their pension plans’ trading activities. The district court admitted a summary stipulation that defendants disclosed to the U.S. government “foreign bank and brokerage accounts; the balances in those accounts; and the volume of foreign securities traded, including Danish securities,” App’x at 1951, but defendants suggest that seeing the actual filings (with the attendant attestations that the information “is true and correct to the best of my knowledge and belief” and warnings that violating the reporting requirements is a crime) would have communicated to the jury “their good faith belief that Solo Capital acquired Danish shares on their behalf consistent with the reclaim applications submitted to S[kat]” and “that defendants were not secretive about their dividend arbitrage trading.” Appellants’ Br. at 48–49. According to defendants, the exclusion of these tax forms “gutted support for defendants’ position that they did not intend to submit false reclaim applications or know of Mr. Shah’s fraud resulting in the false reclaim applications.” Id. at 50.

We, like the district court, do not see how defendants’

submissions to the IRS carry any probative value for the question whether they knew that Shah was abusing Danish law for their

50

benefit, much less how the jury’s receipt of jargon-laden tax forms would have communicated anything more about their intent than the summary stipulation itself conveyed. After all, the average juror presumably knows that lying to the IRS is a perilous choice. Similarly, it seems perfectly plausible that defendants would have wanted to make sure they did not run afoul of U.S. tax law even while using it to perpetrate tax fraud in Denmark: After all, the only reason the trading was contemplated at all was to take advantage of preferential U.S. tax treatment for pension plans. In short, the actual tax forms simply do not make any disputed fact more or less true.

Additionally, as with the evidence relating to Kaye Scholer, the evidence that was admitted—here, the stipulation, along with Markowitz’s testimony about the disclosures defendants were required to and did make to the U.S. government—permitted defense counsel to emphasize in closing that they made required disclosures to the U.S. government, which counsel stated was “further proof that they believed the trading was real.” App’x at 953. It is difficult to imagine what additional arguments counsel could have made armed with the actual tax forms, and defendants do not convincingly articulate one on appeal. We therefore hold that the district court did not abuse its discretion by admitting only the parties’ stipulation about defendants’ compliance with U.S. tax law.

IV. Sufficiency of the evidence as to Jocelyn Markowitz and Elizabeth van Merkensteijn

Finally, defendants argue that there is insufficient evidence to support the fraud verdicts against Jocelyn Markowitz and Elizabeth

51

van Merkensteijn. “We review challenges to the sufficiency of the evidence in support of a jury’s verdict in the light most favorable to the prevailing party and draw all reasonable inferences in that party’s favor.” Moore v. Rubin, 160 F.4th 271, 289 (2d Cir. 2025). 10 Thus, “[a] party challenging the sufficiency of the evidence bears a heavy burden. We overturn a verdict only if there is such a complete absence of evidence supporting the verdict that the jury’s findings could only have been the result of sheer surmise and conjecture or if a reasonable juror could not arrive at the verdict in light of the overwhelming amount of evidence in favor of the losing party.” Id. (internal quotation marks and citation omitted).

Skat pursued an agency theory of liability as to Jocelyn and Elizabeth, seeking to prove that they entered into agency relationships with their husbands relating to the Danish trading which would render them legally responsible for fraudulent acts committed by Richard and John.

An agent’s authority may be apparent or actual; and actual authority may be express or implied. See Bailey v. City of New York, 214 N.Y.S.3d 58, 61 (2d Dep’t 2024). Apparent authority arises from words or conduct by a principal that when reasonably interpreted causes a third party to believe that the principal has authorized someone to act as her agent. Minskoff v. Am. Exp. Travel Related Servs.

10 Defendants argue that our review is de novo, and in a certain sense they are right – we review the district court’s evaluation of evidentiary sufficiency de novo. But the district court, like an appellate court, reviews the jury’s verdict with great deference.

52

Co., 98 F.3d 703, 708 (2d Cir. 1996). It “is normally created through the words and conduct of the principal as they are interpreted by a third party, and cannot be established by the actions or representations of the agent.” Id. Here, there was no allegation that Jocelyn or Elizabeth (the purported principals) personally communicated with Skat (the third party), so the parties appear to agree that Skat’s theory hinges on the existence of Richard and John’s actual authority.

Under New York law (which the parties do not dispute applies here), an agency relationship imbued with actual authority “results from a manifestation of consent by one person to another that the other shall act on his behalf and subject to his control, and the consent by the other to act.” N.Y. Marine & Gen. Ins. Co. v. Tradeline (L.L.C.), 266 F.3d 112, 122 (2d Cir. 2001) (internal quotation marks omitted). In turn, “a principal is liable to third parties for the acts of an agent operating within the scope of his . . . authority.” Citibank, N.A. v. Nyland (CF8) Ltd., 878 F.2d 620, 624 (2d Cir. 1989); see also Adler v. Helman, 564 N.Y.S.2d 828, 830 (3d Dep’t 1991). 11 “The scope of an agent’s actual authority is determined by the intention of the principal or, at least, by the manifestation of that intention to the agent.” Vista Food Exch., Inc. v. Comercial De Alimentos Sanchez, 147 F.4th 73, 90 (2d

11 A principal may also subsequently ratify an act of her agent, and such ratification may be express or implied. See Angerosa v. White Co., 290 N.Y.S. 204, 215 (4th Dep’t 1936). “[R]atification may be implied where the principal retains the benefit of an unauthorized transaction with knowledge of the material facts.” Standard Funding Corp. v. Lewitt, 89 N.Y.2d 546, 552 (1997); see also 148 S. Emerson Partners, LLC v. 148 S. Emerson Assocs., LLC, 70 N.Y.S.3d 213, 215 (2d Dep’t 2018).

53

Cir. 2025) (quoting Wen Kroy Realty Co. v. Pub. Nat’l Bank & Tr. Co. of N.Y., 260 N.Y. 84, 89 (1932)). “An agent has actual authority to take action designated or implied in the principal’s manifestations to the agent and acts necessary or incidental to achieving the principal’s objectives, as the agent reasonably understands the principal’s manifestations and objectives when the agent determines how to act.” Id. (quoting Restatement (Third) of Agency § 2.02(1) (2006)). “The general rule is that an agent employed to do an act is deemed authorized to do it in the manner in which the business entrusted to him is usually done.” Id. (emphasis omitted) (quoting Wen Kroy, 260 N.Y. at 89–90).

Thus, to prove that Jocelyn and Elizabeth were liable for fraud, Skat had to adduce sufficient evidence for a jury to find by a preponderance that an agency relationship existed between Jocelyn and Richard and between Elizabeth and John, and that Richard and John were authorized by their wives to submit refund claims to Skat. We hold that there was ample evidence from which the jury could have and did infer that (1) Elizabeth and Jocelyn entered into agency relationships with their husbands in relation to their investment accounts and (2) that Richard and John’s submission of (false) refund claims to Skat was within the scope of that authority.

First, Elizabeth and Jocelyn created limited liability companies with no apparent business purpose and used them to create pension plans, despite the fact that both women contributed to employersponsored 401(k)s when they were previously employed and were therefore presumably aware that it would be odd for one person to

54

have multiple pension plans and to have pension plans unconnected to any particular employer. See, e.g., App’x at 829 (Jocelyn testifying that she created Calypso Investments LLC without a “specific business in mind,” though she testified that she later acted through it when helping her husband set up his office and plan business dinners); id. at 828 (Jocelyn testifying that she created the Calypso Investments Retirement Plan); id. (Jocelyn testifying that she contributed to a 401(k) plan while employed at UBS Securities); id. at 834–36 (Elizabeth testifying that she created the Azalea LLC without any intention to do business through that entity and that she created the Azalea pension plan without an understanding of how the plan would be funded despite having contributed to a 401(k) plan when she was previously employed at an advertising agency).

Further, Elizabeth and Jocelyn clearly authorized their husbands to use the pension plans they created to engage in trading activities. See App’x at 828–29 (Jocelyn testifying that she understood the plan would be “involved in an investment strategy with foreign securities”); id. at 836–37 (Elizabeth confirming that she authorized her husband and one of her husband’s employees to take action on behalf of the Azalea plan and to pursue “the transaction”). And Jocelyn, who previously worked in investment banking, testified that she eventually learned that the plan would engage in transactions in Danish shares and that there would be a “tax component” to the transactions. Id. at 828–29.

Nevertheless, according to defendants, because Elizabeth and Jocelyn testified that they merely signed documents creating the

55

numerous pension plans and LLCs central to this scheme—at their husbands’ urging without reviewing the documents—no agency relationship was formed. Instead, they refer to this course of conduct as “the normal division of labor that occurs between spouses,” Appellants’ Br. at 53, and assert that because marriage without more does not create an agency relationship, Skat failed to present sufficient evidence of an agency relationship. See, e.g., B-Sharp Musical Prods., Inc. v. Haber, 899 N.Y.S.2d 792, 794 (1st Dep’t 2010) (“[A]n agency relationship may not be implied or inferred solely by reason of the marital relationship of the couple.”) (citations omitted)).

Of course, reasonable people could disagree about whether creating LLCs (unconnected to any real businesses) and using those LLCs to create pension plans (unconnected to a particular employer) for the purpose of making investments is everyday spousal behavior. But Elizabeth’s and Jocelyn’s testimonies provide far more evidence to establish the existence of agency relationships than the mere fact of their marriages. Jocelyn testified that she gave her husband full authority to make financial decisions, including those related to the pension plans, on her behalf: She explained that he “handled all of the financials and finances” for the family, that she “deferred [reading financial documents] to Rich,” that she “[a]lways” relied on him with respect to the pension plan details, although he would bring “anything major” to her attention, and that she signed what he gave her without reading it. App’x at 829. Likewise, Elizabeth testified that “John handled all the finance,” id. at 834, that she relied on her husband to make investments on her behalf, id. at 840, that she was

56

generally unaware of the details of the pension plans and signed what her husband gave her, and that despite being named as a trustee of a pension plan, she authorized her husband to act on her behalf, id. at 836. The jury could have reasonably concluded that, based on the facts in this case—including the wives’ active participation in the creation of these entities, whose only purpose was to effectuate this trading activity—there was an agency relationship between the husbands and wives. And based on the wives’ consenting to their husbands acting on their behalf in financial matters generally and the pension plans specifically, a reasonable jury could have further concluded that the actions the husbands took were within the scope of that relationship.

Defendants also argue that any fraudulent activity by Richard and John cannot be imputed to Jocelyn and Elizabeth because such fraud would have been outside the scope of the agency relationship. They “do not dispute the well-settled rule that a principal, even if innocent, is liable for acts of fraud that are within the scope of an agent’s actual or apparent authority.” Chubb & Son Inc. v. Consoli, 726 N.Y.S.2d 398, 400 (1st Dep’t 2001). Instead, they contend that because—in their view—Jocelyn and Elizabeth “did not manifest or intend anything beyond investing for retirement,” there is insufficient evidence that they gave their husbands authority to defraud Skat. Appellants’ Br. at 54–55. But this is not the degree of specificity the law of agency requires. As we have explained, agents have actual authority to take actions that are “designated or implied” in the principal’s communications to their agent as well as acts “necessary

57

or incidental to achieving the principal’s objectives,” so long as the agent adopts a reasonable understanding of the principal’s objectives. Vista Food, 147 F.4th at 90 (emphases added); see also Angerosa v. White Co., 290 N.Y.S. 204, 208–09 (4th Dep’t 1936) (The “‘[s]cope of authority’ includes not only the actual authorization conferred upon the agent by his principal, but also that which has apparently or impliedly been delegated to him.”). Where an agent’s conduct falls within the scope of his authority, “[a]gency law presumes imputation even where the agent acts less than admirably, exhibits poor business judgment, or commits fraud.” Kirschner v. KPMG LLP, 15 N.Y.3d 446, 465 (2010) (citation omitted). Thus, for example, “[a] principal who gives his agent authority to solicit a sale and accepts the fruits of his efforts will be held responsible for the fraudulent as well as the fair means by which the contract was obtained, if such instrumentalities are in line with the accomplishment of the object of the agency.” Angerosa, 290 N.Y.S. at 209.

So too here. Elizabeth and Jocelyn authorized their husbands to engage in transactions that would increase their wealth. One way Richard and John endeavored to fulfill that mission was by engaging in dividend tax arbitrage which—if done legally and successfully— might have indeed generated returns in their wives’ investment accounts. See, e.g., App’x at 597 (Richard testifying that the Calypso Investments LLC Pension Plan “was formed to do dividend arbitrage trading and other investments as well”). That—as the jury found— the husbands chose to fulfill their mission via fraud does not mean that Elizabeth and Jocelyn cannot be held responsible for that fraud

58

where, as they freely admitted, they had given their husbands wideranging authority over their investment accounts.

Thus, defendants’ argument that the sum total of Jocelyn and Elizabeth’s conduct manifested only a desire to “invest[] for retirement” but not to “defraud Skat” fails to undermine the jury’s verdict. Appellants’ Br. at 55. The non-binding cases on which defendants rely for the proposition that Richard and John’s conduct fell outside of the scope of any agency relationship do not persuade us otherwise. 12

12 The wife in In re Budnick had been asked by her husband to ask her father for a loan, but she embezzled funds from a third party instead and “kept the . . . [e]mbezzlements a secret from” her husband. 469 B.R. 158, 175 (Bankr. D. Conn. 2012). The scope of the agency relationship there was tightly defined: the husband wished for his wife to ask a specific third party for a loan. Because the wife’s conduct was clearly outside that defined scope, a bankruptcy judge held that the husband could not be liable under an agency theory. Here, by contrast, the wives gave their husbands open-ended authority to conduct trading on their behalf, an instruction the husbands did not defy.

And United States v. Bowcut, 287 F.2d 654 (9th Cir. 1961), on which defendants also rely, did not concern an agency theory of liability but instead the government’s assertion that it had adduced evidence of a wife’s fraudulent intent because she had signed fraudulent tax returns. Id. at 656–57. The Ninth Circuit rejected that assertion because there was no other evidence of the wife’s knowledge or fraudulent intent and no available authority suggesting that a signature was sufficient to adduce intent. Id. at 657. The court also observed that a wife is not “normally [] an active and knowing participant in the representations contained in a joint tax return,” so it doubted that “any rational inference of knowledge or fraudulent intent can be drawn from this fact.” Id. Setting the court’s antiquated observation aside, the Ninth Circuit’s refusal to impute fraudulent intent to a taxpayer in a suit for equitable recoupment based solely on her signature says nothing about whether a jury may permissibly determine that

59

At bottom, the question whether Richard and John were acting within the scope of their agency was a question for the jury, see Giglio v. Rubin, 104 N.Y.S.2d 263, 266 (1st Dep’t 1951) (“[T]he scope of the agent’s authority is ordinarily a question of fact for the jury.”); see also Uy v. A. Hussein, 131 N.Y.S.3d 70, 73 (2d Dep’t 2020); Time Warner City Cable v. Adelphi Univ., 813 N.Y.S.2d 114, 116 (2d Dep’t 2006); Patterson v. Khan, 659 N.Y.S.2d 90, 91 (2d Dep’t 1997), and the jury permissibly answered the question in Skat’s favor.

In a final bid to undermine the jury’s verdict on this score, defendants appear to suggest that there was some error in the jury instructions relating to Elizabeth and Jocelyn because the instructions purportedly “conflate[d] direct and vicarious liability” and the verdict form did not specifically ask the jury to make a finding about the existence and scope of the agency relationships between Jocelyn, Elizabeth, and their husbands. Appellants’ Br. at 56–57. According to defendants, these errors “leav[e] one to guess how the jury reached its verdict.” Id. at 57. But defendants made no such objection to the instructions below, nor did they request an agency-specific section of the jury verdict form. We therefore review this argument for fundamental error, reversing only if the instructions contravened “an established rule of law,” the error was “so serious and flagrant that it goes to the very integrity of the trial,” and the resulting charge

the fraud of an agent was within the scope of his authority under New York law. The jury here was entitled to find that the sum of Elizabeth and Jocelyn’s conduct here—yes, effectuated in part via their signatures—rendered the actions taken by their husbands comfortably within the scope of their authority.

60

“deprived the jury of adequate legal guidance to reach a rational decision” or “completely misled the jury as to the governing law.” Innomed Labs, LLC v. ALZA Corp., 368 F.3d 148, 155 (2d Cir. 2004) (internal quotation marks and citations omitted); accord Kotler v. Jubert, 986 F.3d 147, 157–58 (2d Cir. 2021).

We perceive no such error here. The jury instructions explained that one way a defendant could be liable for making false statements to Skat was under an agency theory of liability. The instructions reiterated that Skat asserted that “Mr. Markowitz and Mr. van Merkensteijn, in submitting reclaim applications, acted as agents for their spouses,” explained different forms of an agent’s authority, and instructed that principals are liable for “false statements that an agent makes . . . within the scope of the agent’s authority.” App’x at 1007–08. The relevant section of the (lengthy) verdict form, in turn, asked whether Skat proved that Jocelyn and Elizabeth “made or caused to be made the false statements” in their investment plans’ refund applications to Skat. Id. at 1086-87, 1094–95. Thus, read in context, defendants’ assertion that the jury’s finding of liability was somehow confused or otherwise deficient is wholly speculative, and runs afoul of the presumption that juries follow the instructions given to them. See CSX Transp., Inc. v. Hensley, 556 U.S. 838, 841 (2009). It was made abundantly clear to the jury that Skat’s fraud claims against Jocelyn and Elizabeth lie in vicarious and not

61

direct liability, 13 and defendants’ belated attempts to muddy the waters are insufficient to disturb the jury’s verdict.

V. Conclusion

In sum, we hold as follows:

1. Skat’s lawsuit is not barred by the common law revenue rule because it does not seek to directly enforce foreign tax laws.

2. The district court did not err by refusing to admit testimony from Skat’s U.K. litigation under Federal Rule of Evidence 804(b)(1).

3. The district court did not err by excluding certain pieces of evidence under Rule 403.

4. The evidence is sufficient to support the fraud judgments against Jocelyn Markowitz and Elizabeth van Merkensteijn.

Accordingly, we AFFIRM the judgment of the district court.

13 See, e.g., App’x at 944–45 (“We are not here suggesting that Mrs.

Markowitz or Mrs. van Merkensteijn knew of a fraud or intended to defraud S[kat], but that’s not the end of the story. They authorized their husbands to act on their behalf and for their benefit. . . . Their husbands and Mr. LaRosa were acting as what’s called under the law their agents. And when your agent commits a fraud on your behalf and for your benefit and you authorize them to act, you are on the hook as though you committed it yourself. That is why they are here on fraud claims in addition to our claims to get the money back that their plans got.”).