Summary
holding that when a transaction is only designed to produce tax gains instead of real gains, "it is an act of tax evasion that, even if technically compliant, lies outside of the intent of the Tax Code and so lacks economic substance"
Summary of this case from Benenson v. Comm'r of Internal RevenueOpinion
No. 16-1282
12-16-2016
SANTANDER HOLDINGS USA, INC., and Subsidiaries, f/k/a Sovereign Bancorp., Inc., Plaintiff, Appellee, v. UNITED STATES of America, Defendant, Appellant.
Judith A. Hagley, with whom Caroline D. Ciraolo, Acting Assistant Attorney General, Tax Division, Diana L. Erbsen, Deputy Assistant Attorney General, Tax Division, Gilbert S. Rothenberg, Richard Farber, Washington, DC, and Carmen M. Ortiz, United States Attorney, were on brief, for appellant. Jonathan S. Massey, Washington, DC, with whom Leonard A. Gail, Chicago, IL, Paul J. Berks, Massey & Gail LLP, Rajiv Madan, and Skadden, Arps, Slate, Meagher & Flom LLPwere on brief, for appellee. Martin S. Kaufman, Scarsdale, NY, on brief for Atlantic Legal Foundation, amicus curiae. Scott P. Martin, Gibson, Dunn & Crutcher LLP, Washington, DC, Kate Comerford Todd, Steven P. Lehotsky, Washington, DC, Warren Postman, and U.S. Chamber Litigation Centeron brief for Chamber of Commerce of the United States of America, amicus curiae. Derek T. Ho, Bradley E. Oppenheimer, William H. Milliken, Kellogg, Huber, Hansen, Todd, Evans & Figel, P.L.L.C., Washington, DC, and K. Richard Fosteron brief for Financial Services Roundtable, amicus curiae.
Judith A. Hagley , with whom Caroline D. Ciraolo , Acting Assistant Attorney General, Tax Division, Diana L. Erbsen , Deputy Assistant Attorney General, Tax Division, Gilbert S. Rothenberg , Richard Farber , Washington, DC, and Carmen M. Ortiz , United States Attorney, were on brief, for appellant.
Jonathan S. Massey , Washington, DC, with whom Leonard A. Gail , Chicago, IL, Paul J. Berks , Massey & Gail LLP , Rajiv Madan , and Skadden, Arps, Slate, Meagher & Flom LLP were on brief, for appellee.
Martin S. Kaufman , Scarsdale, NY, on brief for Atlantic Legal Foundation, amicus curiae.
Scott P. Martin , Gibson, Dunn & Crutcher LLP , Washington, DC, Kate Comerford Todd , Steven P. Lehotsky , Washington, DC, Warren Postman , and U.S. Chamber Litigation Center on brief for Chamber of Commerce of the United States of America, amicus curiae.
Derek T. Ho , Bradley E. Oppenheimer , William H. Milliken , Kellogg, Huber, Hansen, Todd, Evans & Figel, P.L.L.C. , Washington, DC, and K. Richard Foster on brief for Financial Services Roundtable, amicus curiae.
Before Lynch and Selya, Circuit Judges, and Burroughs, District Judge.
Of the District of Massachusetts, sitting by designation.
LYNCH, Circuit Judge.
Under the Internal Revenue Code, taxpayers receive, subject to various technical requirements, credits against owed U.S. income tax for every dollar paid to a foreign country for taxable international business transactions of economic substance. See26 U.S.C. §§ 901 –909. Over the past decade, some banks have engaged in complicated transactions the very purpose of which is to generate a foreign tax credit in order to take advantage of the U.S. deductions, and have done so at the expense of the U.S. taxpayer.
This case concerns whether Sovereign Bancorp, Inc., later acquired by Santander Holdings USA, Inc. (together, "Sovereign"), a U.S. taxpayer, is entitled to a refund from the Internal Revenue Service ("IRS") after the IRS began disallowing its claim for foreign tax credits and imposing accuracy-related penalties in 2008. The credits at issue here were claimed for tax years 2003 to 2005 for taxes arranged to be paid to the United Kingdom as part of a Structured Trust Advantaged Repackaged Securities ("STARS") transaction that Sovereign had engaged in. This STARS transaction was initiated in 2003 and was scheduled to last five years, but it ended early, in July 2007, when STARS and similar transactions became the subject of heightened scrutiny from the IRS. SeeDetermining the Amount of Taxes Paid for Purposes of Section 901, 72 Fed. Reg. 15,081 (proposed Mar. 30, 2007). Sovereign and Barclays Bank ("Barclays"), which is chartered in the United Kingdom, were the two parties to the transaction at issue.
Sovereign brought suit to obtain a refund from the IRS in the District of Massachusetts in 2009. The amount of the refund sought is approximately $234 million in taxes, penalties, and interest. Sovereign asserts that it is entitled to foreign tax credits against its U.S. taxes for taxes it paid to the United Kingdom as part of the STARS transaction at issue. As the government concedes, the STARS transaction complied on its face with then-existing U.S. statutory and regulatory requirements. But the government opposes the refund, arguing that the STARS transaction here is an "abusive tax shelter" and so amounts to a transaction that fails the common law economic substance test.
Congress and the IRS have long been concerned with taxpayers inappropriately seeking foreign tax credits. IRS regulations proposed in 2007 and finalized in 2011 prohibited STARS transactions, but not retroactively. SeeDetermining the Amount of Taxes Paid for Purposes of Section 901, 72 Fed. Reg. 15,081, 15,084 (proposed Mar. 30, 2007) ; Determining the Amount of Taxes Paid for Purposes of the Foreign Tax Credit, 76 Fed. Reg. 42,036 (July 18, 2011) (codified at 26 C.F.R. pt. 1). The regulations reflect an understanding that STARS transactions and similar complex financial structures for which foreign tax credits are sought both pose a danger to the federal fisc and do not serve the purposes intended by Congress in enacting the foreign tax credit regime. Those purposes include avoiding double taxation and enabling the conduct of business affairs abroad by U.S. firms. SeeH.R. Rep. No. 83–1337, at 4103 (1954) ("The [foreign tax credit] provision was originally designed to produce uniformity of tax burden among United States taxpayers, irrespective of whether they were engaged in business in the United States or engaged in business abroad."). This case involves a STARS transaction that took place before such transactions were forbidden by regulation, and no one contends the 2011 regulation applies. This decision thus directly affects only that transaction.
During roughly the same period as the transaction at issue here, from 2001 to 2007, other U.S. banks also entered into STARS transactions with Barclays. They similarly sought tax credits, and the IRS similarly opposed them. In Bank of New York Mellon Corp. v. Commissioner(BNY), 801 F.3d 104, 107 (2d Cir. 2015), the Second Circuit affirmed a judgment disallowing the credits claimed by Bank of New York Mellon for its STARS transaction with Barclays. Using somewhat different reasoning, the Federal Circuit in Salem Financial, Inc. v. United States, 786 F.3d 932, 951, 954–55 (Fed. Cir. 2015), also upheld a determination disallowing credits claimed by Branch Banking & Trust Corporation for a STARS transaction with Barclays. Both circuit court opinions contain extensive factual descriptions of the STARS transactions, which also largely characterize the transaction at issue here. A third case, involving a Wells Fargo STARS transaction, was tried in a federal district court in the Eighth Circuit. SeeWells Fargo & Co. v. United States, 143 F.Supp.3d 827, 842 (D. Minn. 2015). After trial, a jury found that the transaction lacked economic substance.
We note that while we discuss the findings of the Second Circuit in BNY, our opinion does not rely in any sense on the earlier opinion of the tax court in that case. SeeBank of N.Y. Mellon Corp.v. Comm'r, 140 T.C. 15, as amended by 106 T.C.M. (CCH) 367 (T.C. 2013). Because we do not rely on that opinion, we need not address Sovereign's argument that the judge in that case suffered from a conflict of interest, a claim the government vigorously disputes.
Although Sovereign argues on appeal that the transaction in BNYis distinguishable, it conceded below that the transaction was "very similar" to the one at issue here. And while Sovereign contends that the bank in Salemadopted a different litigation strategy than the one pursued by Sovereign, it does little to demonstrate that the STARS transaction in Saleminvolved any materially different facts.
The Massachusetts district court in this case awarded summary judgment to Sovereign. It first entered partial summary judgment for Sovereign on the issue of whether a payment Sovereign received from Barclays should be considered income to Sovereign in calculating the STARS Trust transaction's profit. Santander Holdings USA, Inc. & Subsidiaries v. United States(Santander I), 977 F.Supp.2d 46, 48 (D. Mass 2013). It then entered judgment for Sovereign after finding as a matter of law that the Trust and Loan transactions had economic substance, and so Sovereign was entitled to interest-related deductions on expenses for the Loan transaction and a refund on the disallowed foreign tax credits claimed for the Trust transaction and the penalties imposed by the IRS. Santander Holdings USA, Inc. v. United States(Santander II), 144 F.Supp.3d 239, 248 (D. Mass. 2015). The court also denied the government's cross-motion for partial summary judgment in its favor on a number of issues, including whether Sovereign's U.K. taxes should be regarded as expenses in any calculation of Sovereign's profit from the STARS transaction. Id. at 242–44, 248. The government appeals from the grant of summary judgment to Sovereign and the denial of its cross-motion.
Through concessions made by both the government and Sovereign, the appeal has been considerably simplified. The government no longer contends that it is entitled to a jury trial on the tax refund claim; it seeks a jury trial only on the penalties claim. The government also does not contend any longer that the district court improperly excluded evidence, or that there are any material disputes of fact, or that summary judgment was entered prematurely. Rather, the government agrees that the controlling issue is one of law and argues that its cross-motion for summary judgment as to the Trust portion of the STARS transaction should have been allowed. Sovereign, for its part, agrees, for the purposes of summary judgment, that the proper focus is on the Trust transaction alone, and not on the Loan transaction.
The government also argued to the district court that the foreign tax credits claimed by Sovereign should be denied on the basis of two "substance over form" doctrines, the "step transaction" and "conduit" doctrines, but the district court rejected the argument. Santander II, 144 F.Supp.3d at 244. As the government focuses its appeal on the economic substance doctrine, we do not consider the district court's rejection of the government's substance-over-form argument.
The parties have agreed for purposes of this appeal that the Trust transaction should be analyzed separately from the Loan transaction. The bank in Salemsimilarly accepted the bifurcation of the tax consequences of the Trust transaction and the Loan transaction for purposes of that appeal. Salem, 786 F.3d at 940.
The government no longer contests the economic substance of the Loan transaction, as long as the Loan transaction is analyzed separately from the Trust transaction, and does not appeal the district court's decision that Sovereign may claim certain interest-expense deductions.
We hold that the district court committed reversible error and that the government is entitled to summary judgment in its favor as to the economic substance of the STARS Trust transaction. We largely agree with the reasoning of the Federal Circuit opinion in Salem in rejecting the claims that the Trust transaction had economic substance and substantially rely on its analysis.
I.
We give a brief description of the transaction and then of this Circuit's economic substance test.
A. The STARS Transaction
Sovereign entered into the STARS transaction with Barclays in 2003. U.S. banks were then aware of the tax risks of being denied the full amount of U.S. foreign tax credits. See, e.g., Salem, 786 F.3d at 937. Like other STARS transactions, the one Sovereign entered into had, as the district court put it, a "Rube Goldberg" complexity. Santander I, 977 F.Supp.2d at 48. We explain it briefly and rely on BNY and Salem for further details.
In 2003, Sovereign first created a Trust (the Trust half of the transaction) into which it ultimately contributed about $6.7 billion of its U.S.-located income-producing assets. The trustee of the Trust was, by its terms, a U.K. citizen, a fact which subjected the Trust to U.K. taxes. The U.K. taxes were at a rate of 22%. The Trust was also subject to U.S. federal income tax at a rate of 35%, but it could claim a tax credit for the taxes paid to the United Kingdom. The Trust was structured, therefore, to receive foreign tax credits for the amount paid in tax on the Trust to the United Kingdom. It is undisputed that Sovereign paid all U.K. taxes for which it claimed U.S. tax credits.
Barclays acquired an interest in the Trust for $750 million in November 2003 at the Trust's initial creation and acquired an additional $400 million interest almost a year later, when Sovereign added additional funds to the Trust. Significantly, Barclays was required to sell its interest in the Trust back to Sovereign for $1.15 billion at the end of the transaction.Sovereign treated this $1.15 billion contribution from Barclays as a Loan (the Loan half of the STARS transaction) for accounting and regulatory purposes, including in all of Sovereign's filings to the Securities and Exchange Commission and the Office of Thrift Supervision. The offsetting agreements that converted Barclays's purchase of an interest in the Trust into the Loan effectively resulted in Barclays lending Sovereign $750 million at a floating monthly rate of LIBOR plus 50 basis points and $400 million at LIBOR plus 25 basis points.
LIBOR stands for "Intercontinental Exchange London Interbank Offered Rate." Salem, 786 F.3d at 937 n.1. LIBOR "is a benchmark rate that some of the world's leading banks charge each other for short-term loans." Id.
When it first marketed this transaction to potential counterparties, Barclays did not include this Loan component. SeeSalem, 786 F.3d at 936. The government suggests that Barclays added the Loan to "disguise the true nature of the [transaction] and permit U.S. taxpayers to justify STARS as low-cost funding." Sovereign asserts that "there is no evidence ... that any non-loan transaction was ever offered to (or considered by) Sovereign," and that "[t]o the extent Barclays may have proposed a non-loan transaction to other banks, the evidence shows they were uninterested in it." Because we must analyze the Loan and Trust transactions separately, this dispute is immaterial.
The Trust engaged in a series of actions that generated a U.K. tax benefit for Barclays. The Trust distributed funds to a Barclays Blocked Account, which Barclays could not access, but which allowed Barclays to formally hold the funds in its name. The Barclays Blocked Account then immediately returned the funds to the Trust. Barclays owed U.K. taxes on the distributions made to the Barclays Blocked Account, but, importantly, Barclays was entitled to a tax credit for the U.K. tax paid on this income by the Trust, and Barclays also was permitted to deduct its re-contributions to the Trust as a tax loss. The combination of the tax credit and deduction "creat[ed] a net tax deduction for Barclays that it could use to offset tax on other income unrelated to [the STARS transaction]."
In exchange, Barclays paid Sovereign a monthly sum, referred to as the "Barclays" or "Bx" payment. The amount of the Bx payment was calculated to equal 50% of the U.K. tax Sovereign paid on the Trust's income. In a sense, the 50% was a return to Sovereign of half of its tax payment, whether or not it was technically a rebate. The Bx payment was "netted against Sovereign's interest obligation" on the Loan.
The benefits for both parties can be illustrated by a hypothetical also employed by the Second and Federal Circuits. See BNY, 801 F.3d at 111 ; Salem, 786 F.3d at 938. Assume $100 of income in the Trust for a given month. Through its ownership interest in the Trust and the Trust's structure, Barclays would be liable for a 30% U.K. corporate tax on the Trust's income, amounting to $30. BNY, 801 F.3d at 111. Barclays would then claim a credit for the 22% U.K. tax paid on the Trust by Sovereign amounting to $22, bringing Barclays's own tax liability down to $8. Id. The Trust would set aside $22 to settle the U.K. tax owed by Sovereign; the remaining $78 would be shuttled into and out of the Barclays Blocked Account. Id. Sovereign would claim a U.S. foreign tax credit for the $22 it paid in U.K. taxes. Id.
Upon redistributing the $78 to the Trust from the Barclays Blocked Account, Barclays would claim a trading loss deduction on the $78 which, at the corporate tax rate of 30%, would amount to $23.40. Id. Barclays would make a Bx payment to Sovereign calculated to be half of the 22% U.K. tax paid by Sovereign, which would amount to $11. Id. Barclays then deducted this payment at the 30% U.K. tax rate as well, resulting in a $3.30 deduction. Id. In the end, Barclays would save $7.70 in taxes for each $100 of Trust income ($23.40 – $8 – $11 + $3.30), and Sovereign would save $11 (the amount of the Bx payment calculated against Sovereign's U.K. tax exposure). Id. Both parties ultimately reduced their tax exposure—Barclays through the various deductions generated by the Trust transaction and Sovereign through the Bx payment.
B. The Economic Substance Doctrine
The federal income tax is, and always has been, based on statute. The economic substance doctrine, like other common law tax doctrines, can thus perhaps best be thought of as a tool of statutory interpretation, as then-Judge Breyer characterized it in his opinion for this court in Dewees v. Commissioner, 870 F.2d 21, 35–36 (1st Cir. 1989).
Sovereign argues that the foreign tax credit area is so heavily populated with IRS regulation that there is no need for any further regulation by the courts under the guise of the economic substance doctrine. On these facts, we reject the proposition. In practical terms, it takes time for the government to analyze a new problem, come up with a solution, and promulgate regulations. "The endless ingenuity of taxpayers in attempting to avoid taxes means that there will be a first time for everything," Wells Fargo, 143 F.Supp.3d at 838, and the economic substance test guards against abuse of loopholes that Congress and the IRS have not anticipated.
As one commentator says:
A related ... claim is that the legislature assumes that long-standing common law doctrines such as economic substance will be used to interpret the statutes it enacts. Under this claim, the doctrines have been implicitly adopted as part of the statute—at least where the statute does not indicate otherwise.
Joseph Bankman, The Economic Substance Doctrine, 74 S. Cal. L. Rev. 5, 11 (2000).
The common law economic substance doctrine traces back to the Supreme Court's decision in Gregory v. Helvering, 293 U.S. 465, 55 S.Ct. 266, 79 L.Ed. 596 (1935). The Court there looked beyond the fact that a corporate reorganization technically complied with the statutory requirement and found that it lacked economic substance. Id. at 468–70, 55 S.Ct. 266. It found as such because the reorganization was:
In 2010, Congress enacted a statutory economic substance test. See26 U.S.C. § 7701(o). The statutory test was not made retroactive. Our analysis, however, is not in conflict with that test, as Congress specified that the 2010 codification would be applied as courts have previously and consistently applied the economic substance doctrine. Id.§ 7701(o)(5)(C). If the codification reveals anything about congressional intent as to pre–2010 STARS transactions, it supports our conclusion.
an operation having no business or corporate purpose—a mere device which put on the form of a corporate reorganization as a disguise for concealing its real character, and the sole object and accomplishment of which was the consummation of a preconceived plan, not to reorganize a business or any part of a business, but to transfer a parcel of corporate shares to the petitioner.
Id. at 469, 55 S.Ct. 266. The Court reached this conclusion from the fact that "the transaction upon its face lies outside the plain intent of the statute." Id. at 470, 55 S.Ct. 266.
The Court clarified the doctrine further in Frank Lyon Co. v. United States, 435 U.S. 561, 98 S.Ct. 1291, 55 L.Ed.2d 550 (1978), where it reversed the Eighth Circuit's decision that a sale-and-leaseback transaction did not meet the economic substance test. Id. at 584, 98 S.Ct. 1291. The Court explained that "[i]n applying this doctrine of substance over form, the Court has looked to the objective economic realities of a transaction rather than to the particular form the parties employed." Id. at 573, 98 S.Ct. 1291 (emphasis added).
The First Circuit has addressed challenges to the economic substance of transactions in a number of cases, although the cases often have not invoked the "economic substance doctrine" by that name. See, e.g., Stone v. Comm'r, 360 F.2d 737 (1st Cir. 1966) ; Fabreeka Prods. Co. v. Comm'r, 294 F.2d 876 (1st Cir. 1961) ; Granite Tr. Co. v. United States, 238 F.2d 670 (1st Cir. 1956). This court has been particularly wary of inquiring into the subjective motivations of taxpayers: "[U]nless Congress makes it abundantly clear, we do not think tax consequences should be dependent upon the discovery of a purpose, or a state of mind, whether it be elaborate or simple." Fabreeka Prods. Co., 294 F.2d at 878.
Dewees is our most recent significant case on the economic substance doctrine. There, this court upheld a tax court decision that a "loss [the petitioners] incurred while engaged in ‘straddle’ trading on the London Metals Exchange was not an ‘ordinary loss' deductible from their income." Dewees, 870 F.2d at 22. The tax court held that the loss was not deductible because the straddle trades were sham transactions and not "entered into for profit" within the meaning of section 108 of the Internal Revenue Code. Id.
This court upheld the tax court's decision for four principal reasons. We emphasized that the case was one of some 1,100 consolidated by the tax court, from the general pattern of which the tax court could infer that the transactions were designed to avoid taxes; that the promotional material for the transactions focused exclusively on their tax effects; that although margin accounts were opened for the transactions, none of the investors in any of the transactions ever received a margin call; and that no investor ever made a net profit or "was ever asked to pay a loss, beyond the initial margin deposit" for the transactions. Id. at 31. We rejected the petitioner's argument that we must analyze the taxpayer's subjective motivation under the relevant statutory framework. Id. at 34. Among other reasons, the court noted that the tax court had concluded that the transactions were "shams in substance," and that "[c]ase law makes clear that a taxpayer cannot deduct a ‘sham transaction’ loss, irrespective of his subjective profit motive." Id. at 35.
To the extent that similar evidence is in the record for this case, it supports our conclusion. As in Dewees, we have examined the pattern that has emerged from comparable STARS transactions. We have used Sovereign's and Barclays's communications to each other about the transaction and, in particular, their emphasis on the connection of the Bx payment to Sovereign's U.K. taxes and the Trust transaction's "tax risk," to conclude that the Trust transaction had no objective purpose outside its tax effect. And we too have noted that the transaction at issue here was structured such that it exposed neither party to realistic non-tax risk.
Dewees instructs that the economic substance doctrine is centered on discerning whether the challenged transaction objectively "lies outside the plain intent of the [relevant statutory regime]." Id. at 29 (quoting Gregory, 293 U.S. at 470, 55 S.Ct. 266 ). It further instructs that a transaction fails the economic substance test if, "though [it] actually occurred and technically complied with the tax code, [it] w[as] mere[ly a] device[ ] to avoid tax liability." Id. at 30 ; see alsoSchussel v. Werfel, 758 F.3d 82, 97 (1st Cir. 2014) (noting that courts may "disregard the form of transactions that have no business purpose or economic substance beyond tax evasion"). In other words, when a transaction "is one designed to produce tax gains ... [not] real gains," Dewees, 870 F.2d at 31 —such as when the challenged transaction has no prospect for pre-tax profit—then it is an act of tax evasion that, even if technically compliant, lies outside of the intent of the Tax Code and so lacks economic substance.
II.
"We review orders granting or denying summary judgment de novo." Fithian v. Reed, 204 F.3d 306, 308 (1st Cir. 2000). "The general characterization of a transaction for tax purposes is a question of law subject to review." Frank Lyon, 435 U.S. at 581 n.16, 98 S.Ct. 1291.
In its first partial summary judgment decision, the district court rejected the government's argument that the Bx payment was in effect a tax rebate. Santander I, 977 F.Supp.2d at 50. The district court concluded instead that the Bx payment as a matter of law was income to Sovereign. Id. at 52. The Federal Circuit reached the same conclusion as the district court in our case and held that the Bx payment must be counted as income under the logic of Old Colony Trust Co. v. Commissioner, 279 U.S. 716, 49 S.Ct. 499, 73 L.Ed. 918 (1929) ; IES Industries, Inc. v. United States, 253 F.3d 350 (8th Cir. 2001) ; and Compaq Computer Corp. & Subsidiaries v. Commissioner, 277 F.3d 778 (5th Cir. 2001). Salem, 786 F.3d at 944–46. By contrast, the Second Circuit accepted the government's argument. BNY, 801 F.3d at 121–22.
We see no need to address the government's characterization of the Bx payment as a rebate, not income, because we hold that whether the Bx payment is best characterized as a rebate or as income, Sovereign's argument still fails. The STARS Trust transaction itself does not have a reasonable prospect of creating a profit without considering the foreign tax credits, and, as a result, it is not a transaction for which Congress intended to give the benefit of the foreign tax credit. This conclusion mirrors that of the Federal Circuit in Salem, and we reach it largely for the reasons stated there. Salem, 786 F.3d at 946–55. We agree with that court that we must "assess [the] transaction's economic reality, and in particular its profit potential, independent of the expected tax benefits." Id. at 948. Using similar reasoning, we find that the Trust transaction is "shaped solely by tax-avoidance features," id. at 942 (quoting Stobie Creek Invs. LLC v. United States, 608 F.3d 1366, 1375 (Fed. Cir. 2010) ), that "lack a bona fide business purpose," id. Most importantly, we agree with the Federal Circuit that the Trust transaction is profitless, id. at 949, and that it is "not the type of transaction Congress intended to promote with the foreign tax credit system," id. at 954.
The Trust transaction is profitless because the "profit" to Sovereign from the Bx payment comes at the expense of exposure to double the Bx payment's value in U.K. taxes. To return briefly to the $100 hypothetical: even if Sovereign receives an $11 Bx payment from Barclays (half of the $22 paid by Sovereign to the United Kingdom at its 22% tax rate), the Trust transaction lacks a reasonable potential (or any potential) of generating profit because the $11 Bx payment is earned at the expense of the $22 U.K. tax. In other words, every $1 the Trust transaction earns through the Bx payment costs $2 from the transaction costs of subjecting the Trust transaction to U.K. tax. When the primary transaction cost of the Bx payment, the U.K. taxes, are factored into the pre-tax profitability calculation, the Trust transaction is plainly profitless. Sovereign's "profit" comes from the foreign tax credits it claims for the U.K. taxes combined with a Bx payment calculated as half its U.K. tax liability.
Because exposure to U.K. taxation was the necessary and sufficient condition of the Bx payment, the U.K. taxes were an expense incurred by Sovereign for the "profit" generated by the Trust transaction. And when the U.K. taxes are recognized as expenses, there is no pre-tax profit, and the Trust transaction lacks a cardinal feature of an economically substantial transaction: a reasonable prospect of pre-tax profit.
Sovereign and the district court rely heavily on Compaqand IESfor the proposition that foreign taxes should not be treated as expenses. Santander II, 144 F.Supp.3d at 242–44. Those cases did not analyze STARS transactions and so are distinguishable factually. We agree with the Salemcourt's analysis of this issue as to the Trust transaction. 786 F.3d at 947–49.
Nor does our conclusion that Sovereign's U.K. taxes should be considered expenses contradict the Supreme Court's holding in Old Colony. Old Colony did not involve foreign taxes and says nothing about whether foreign tax liability may ever be considered an expense. SeeOld Colony, 279 U.S. at 716, 49 S.Ct. 499.
Accordingly, we conclude both that the STARS Trust transaction had no objective non-tax economic benefit and that Congress, in creating the foreign tax credit regime, did not intend that it would cover this type of generated transaction. Exposure to U.K. taxation for the purpose of generating U.S. foreign tax credits was the Trust transaction's whole function.
SeeJohn P. Steines, Jr., Subsidized Foreign Tax Credits and the Economic Substance Doctrine, 70 Tax Lawyer (forthcoming 2017) ("[I]t is virtually impossible for a dispassionate analyst to reasonably conclude that Congress intended to surrender more revenue than that captured by the foreign government in a holistic sense where the U.S. taxpayer and the counterparty split the remaining spoils solely by reason of carefully exploited inconsistent international tax rules in an otherwise unprofitable transaction that is an overly complicated version of an orthodox deal that would not have given rise to credits at all."). A copy of this article was filed with the court and disclosed to the parties.
Our conclusion that the Trust transaction lacks economic substance is entirely consistent with our statement in Dewees that "taxpayers may lawfully structure transactions that seek real gains in a way that also maximizes tax advantages." 870 F.2d at 32. Again, this situation does not involve private parties structuring an agreement to benefit both parties and only then seeking to maximize the tax benefits. The Bx payments do not come into fruition until and unless Sovereign pays the U.K. taxes (for which it will seek a 100% credit on its U.S. taxes).
Indeed, the record demonstrates that the Bx payment is inextricably linked to the deliberate incurring of Sovereign's U.K. tax liability. Barclays and Sovereign made clear to each other that the Bx payment would be calculated based on Sovereign's U.K. tax liability and the credits that Barclays would then be able to claim. An internal communication between the parties stated that the Bx payment would allow "Barclays [to] share[ ] U.K. tax credits with Sovereign."
The STARS scheme is profitable only because Sovereign plans to obtain U.S. tax credits; that is, the whole existence of the Trust transaction depends on getting a U.S. tax credit. There is otherwise no business reason to engage in the transaction. As the Salem court found:
The evidence thus supports the trial court's finding that the STARS Trust was a "prepackaged strategy" created to generate U.S. and U.K. tax benefits for [the counterparty] and Barclays. Barclays agreed to bear half of [the counterparty's]
U.K. tax expense under the transaction in exchange for an opportunity to claim substantial U.K. tax benefits for itself (through the trading loss deduction). [The counterparty], on the other hand, benefited by claiming a foreign tax credit equal to the entire amount of the Trust's U.K. taxes while "getting back one-half of the U.K. tax" from Barclays. Absent those tax advantages, the STARS transaction would never have occurred.
786 F.3d at 952 (citation omitted). Here, Sovereign subjected its property and income to U.K. taxation only because it anticipated it could avoid U.S. taxes through the resulting U.S. tax credit.
The Trust transaction did not advance the Tax Code's interest in providing foreign tax credits in order to encourage business abroad or in avoiding double taxation. Nor does disallowing foreign tax credits for the STARS Trust transaction interfere with the United Kingdom's authority. After all, it was the U.K. authorities who in 2005 first called STARS transactions to the attention of the IRS as a potential impermissible tax shelter.
Moreover, there is no tension between denying foreign tax credits for the STARS Trust transaction and the U.S.–U.K. tax treaty: As the government correctly notes, the treaty requires the grant of foreign tax credits "subject to the limitations of the laws of the United States." Convention with Great Britain and Northern Ireland regarding Double Taxation and Prevention of Fiscal Evasion, art. 24, July 24, 2001, S. Treaty Doc. No. 107–19. Among those limitations, of course, are "anti-abuse principles" such as the economic substance doctrine. SeeTreasury Dep't, Technical Explanation of the Convention Between the Government of the United States of America and the Government of the United Kingdom of Great Britain and Northern Ireland for the Avoidance of Double Taxation and the Prevention of Fiscal Evasion with Respect to Taxes on Income and on Capital Gains14. See alsoDel Commercial Props., Inc.v. Comm'r, 251 F.3d 210, 214 (D.C. Cir. 2001) ("[I]f the sole purpose of a transaction with a foreign corporation is to dodge U.S. taxes, [a] treaty cannot shield the taxpayer from the fatality of the [substance-over-form] step-transaction doctrine.").
Of course, as the government readily admits, some transactions that are not immediately profitable without tax benefits, such as investments in "nascent technologies," may have economic substance. SeeSalem, 786 F.3d at 950. But the Trust transaction is not comparable to such transactions because it does not "meaningfully alter[ ] the taxpayer's economic position (other than with regard to the tax consequences)." Id.
Moreover, unlike long-term investments that may not initially turn a profit, but which have economic substance, the Trust transaction lacks any real economic risk. The Salem court pointed out that Barclays ran little risk of having to pay the Bx payment in absence of the anticipated U.K. tax benefits because the counterparty indemnified Barclays should that happen. Id. at 943–44. The Bx payments were not truly independent of the expected U.K. tax effects. The counterparty's "ability to benefit economically from the Bx payments depended on Barclays'[s] receipt of its expected tax benefits, which in turn depended on the Trust's U.K. tax payments." Id. at 944.
Here, too, Sovereign and Barclays "developed contractual remedies and took other steps to minimize the risk of [a divergence between actual effects and the pre-engineered outcome of the Bx payment's relationship to the U.K. taxes]." Unlike transactions that have survived an economic substance challenge, such as the sale-and-leaseback structure in Frank Lyon, 435 U.S. at 577, 98 S.Ct. 1291, the STARS Trust transaction posed no non-tax risks to Sovereign. Instead, Sovereign's internal discussions focused on the "risk" of being unable to claim foreign tax credits for the U.K. taxes on the Trust transaction, and it informed the Federal Deposit Insurance Corporation that it would "bear the United States tax risk" of the transaction.
Further, we agree with the government that Sovereign's U.K. tax was artificially generated through a series of circular cash flows through the Trust and was the quid pro quo for the Bx payment. The assets in the Trust never effectively left Sovereign's control, nor did they perform any function when placed in the Trust that they could not without the Trust—other than, of course, creating the tax effect that made possible the Bx payment. Indeed, when calculating the profit potential of the STARS transaction, Sovereign deducted the income from the Trust assets, as that income would have been earned without the Trust's existence.
Resorting to the uncontroversial principle that the foreign tax credit regime was designed to avoid double taxation does not help Sovereign. If mere invocation of that principle were enough, every tax avoidance scheme would pass muster. After all:
the fact that the transactions produced a net gain to the taxpayer after taking both the foreign taxes and the foreign tax credit into account says nothing about the economic reality of the transactions, because all tax shelter transactions produce a gain for the taxpayer after the tax effects are taken into account—that is why taxpayers are willing to enter into them and to pay substantial fees to the promoters.
Salem, 786 F.3d at 948.
Equally fundamental to the purpose of granting foreign tax credits is the related principle that those credits are extended only to legitimate business transactions. SeeH.R. Rep. No. 83–1337, at 4103 (1954) ("The [foreign tax credit] provision was originally designed to produce uniformity of tax burden among United States taxpayers, irrespective of whether they were engaged in business in the United States or engaged in business abroad." (emphasis added)). The Trust transaction provided no business for Sovereign. It furnished Barclays with a tax benefit, which Barclays in turn shared with Sovereign, effectively giving Sovereign a tax benefit of its own when combined with the anticipated foreign tax credits. The Trust transaction was not a legitimate business and lacked economic substance.
III.
We reverse the judgment of the district court as to the economic substance of the Trust transaction and the foreign tax credits claimed for the Trust transaction and remand for judgment to be entered for the United States on the refund claim and for a trial limited to the penalties issue. Costs are awarded to the appellant.