KOZINSKI, Circuit Judge:
It's a timeless and tiresome question of American tax law: Is a transaction debt or equity? The extremes answer themselves. The classic equity investment entitles the investor to participate in management and share the (potentially limitless) profits — but only after those holding preferred interests have been paid. High risk, high reward. The classic debt instrument, by contrast, entitles an investor to preferred and limited payments for a fixed period. Low risk, predictable reward. But a vast hinterland of hybrid financial arrangements lurks in the middle.
This case puts us in that awkward position with an unusual twist. In the textbook example, a taxpaying corporation wants an investment to be treated as debt so it can deduct the interest payments. Here, Hewlett-Packard ("HP") wants its investment in a foreign entity to be treated as equity, so that HP will be entitled to the foreign tax credits that the entity — a so-called "FTC generator" — produces. The United States taxes the worldwide income of domestic corporations, but gives them a credit against their domestic taxes for foreign taxes they (or a subsidiary) pay. FTC generators are entities that churn out foreign credits for U.S. multinationals, which companies typically desire if they pay foreign taxes at a lower average rate than domestic taxes. See Stafford Smiley & Michael Lloyd, Foreign Tax Credit Generators, 39 J. Corp. Tax'n 3, 4-5 (2012). No small sum is on the line: The transaction here saved HP (and lost Treasury) millions of dollars.
But HP is entitled to the foreign tax credits only if it owned at least 10% of the voting stock and received dividends — in other words, if the investment was really equity, not debt. I.R.C. § 902(a). So, was it?
The tax borscht at issue was cooked up in the 1990s by AIG Financial Products. The arrangement took advantage of the fact that contingent interest — interest payments that depend on future developments, and may never be paid at all — was immediately taxable in the Netherlands but not in the United States. This allowed AIG to create a Dutch company — called Foppingadreef Investments, or "FOP" — that would (and could) do little else than purchase contingent interest notes. The entity's preferred shares would be owned by an American company, which would receive all the dividends from the notes, and thus be entitled to claim foreign tax credits for FOP's Dutch taxes. ABN, a Dutch bank, would own FOP's common shares and sell it the contingent notes.
Because the accrued contingent interest was taxable in the Netherlands but not in the United States, FOP would generate "excess" foreign credits that the American investor could use to offset American taxes on other foreign profits. And, because FOP could do little beside purchase contingent interest notes, the preferred stock guaranteed, in essence, a fixed stream of payments to the holder of the preferred shares. FOP did not, and indeed could not, have any general creditors.
In 1996, AIG sold FOP's preferred stock to HP for a little over $200 million. HP contemporaneously purchased a put from ABN, which gave HP the right to sell its shares to ABN in 2003 or 2007. HP claimed millions in foreign tax credits between 1997 and 2003. The company then exercised its option, sold its preferred shares back to ABN, and reported a sweet capital loss of more than $16 million.
HP again appeals.
We hazard a few observations on this split. First, the distinction between fact and law is notoriously fuzzy, and can turn as much on convention as logic. See, e.g., Nathan Isaacs, The Law and the Facts, 22 Colum. L. Rev. 1 (1922). Second, calling this a mixed question rather than a factual one doesn't add much focus: If it's a mixed question, we still ask whether the trial court "based its ruling on an erroneous view of the law or on a clearly erroneous assessment of the evidence." Cooter & Gell v. Hartmarx Corp., 496 U.S. 384, 405, 110 S.Ct. 2447, 110 L.Ed.2d 359 (1990). But this just means that "[w]hen an appellate court reviews a district court's factual findings, the abuse of discretion and clearly erroneous standards are indistinguishable." Id. at 401, 110 S.Ct. 2447. Thus, calling this a "mixed question" succeeds only in pushing the conceptual conundrum back one step: Are we reviewing a factual finding or not?
A better approach is to ask whether the purposes of a deference regime apply in debt/equity cases. They do. Corporate tax planning involves abstruse transactions that generalist appellate courts are ill-equipped to untangle; the need for "practical human experience" and the "multiplicity of relevant factual elements" cry out for deference to the Tax Court.
And, like other circuits, we use a multi-factor test to decide whether an instrument is best characterized as debt or equity. Our test, incredibly, involves the unguided weighing of no fewer than eleven non-exclusive factors — whether there is a fixed maturity date, whether the investor participates in management and so on.
The parties expend considerable effort arguing over whether "most" of the relevant factors point one way or the other. But our test isn't a bean-counting exercise. Instead, it's best understood as a non-exhaustive list of circumstances that are often helpful in guiding a court's factual determination. And, while such a free-floating inquiry is hardly a paragon of judicial predictability, it's the necessary evil of a tax code that mistakes a messy spectrum for a simple binary, and has repeatedly failed to offer the courts statutory or regulatory guidance. See Howard E. Abrams & Richard L. Doernberg, Federal Corporate Taxation 75 (5th ed. 2002).
Admittedly, our circuit's test is somewhat confusing in its treatment of taxpayer intent. Intent is listed as one of the eleven factors, but we've also said that the test is "primarily directed" at discovering the intent of the parties to the transaction. A.R. Lantz Co., 424 F.2d at 1333; see also Bauer, 748 F.2d at 1367-68. To make matters worse, one of our important precedents contains a garbled attempt to clarify the issue: "However, analysis of the factors previously enumerated, including to subjective resolution of the ultimate the objective expression of intent, leads issue: Whether the parties in fact intended the advance to create debt rather than equity." A. R. Lantz Co., 424 F.2d at 1333-34.
We think the best way to read our precedent is as follows: Our test is "primarily directed" at determining whether the parties subjectively intended to craft an instrument that is more debt-like or equity-like. A quest for subjective intent always requires objective evidence, hence the eleven factors. On this account, all factors on the list could be described as "evidence of intent." Direct, objective evidence of intent — say, an email from an executive stating he wishes to create an unalloyed
Our precedent's preoccupation with intent is nonetheless a little puzzling, since it suggests that a taxpayer could achieve debt treatment for an instrument that functions as equity (or vice versa), so long as he had the right state of mind in crafting the instrument. Were we writing on a barren slate, we might say that our test is simply directed at determining whether an instrument functions more like debt or equity. There's nothing magical about intent. Nonetheless, we believe our circuit's roundabout intent-based test merges with this simple function test in all but a few outlandish cases. Cf. United States v. Powell, 955 F.2d 1206, 1212 (9th Cir. 1991) ("[A] jury is not precluded from considering the reasonableness of the interpretation of the law in weighing the credibility of the claim that the [defendants] subjectively believed that the law did not require that they file income tax returns."); Cheek v. United States, 498 U.S. 192, 203-04, 111 S.Ct. 604, 112 L.Ed.2d 617 (1991) ("[T]he more unreasonable the asserted beliefs or misunderstandings are, the more likely the jury will consider them to be nothing more than simple disagreement with known legal duties imposed by the tax laws."). Crucially, both tests allow us to resist taxpayer gamesmanship: The clever taxpayer who designs a debt instrument but proclaims an unpersuasive intention to own equity would fail via either path.
With this background in place, we have no difficulty concluding that the Tax Court didn't err in finding that HP's investment in FOP is best characterized as debt. While the factors point in different directions, the Tax Court committed no clear error in considering or weighing them. It appropriately found that the formal labels attached to the documents didn't settle the inquiry. Instead, of particular importance to the Tax Court was the de facto presence of a fixed maturity date, and HP's de facto creditor's rights. The Tax Court concluded that the deal had a de facto maturity date because HP had an overwhelming economic incentive to divest itself of FOP after 2003: After that year, FOP would have negative earnings, thereby preventing HP from claiming foreign tax credits. HP knew this, and never expected to stay in the transaction after 2003. HP's income was also highly predictable: It was entitled to semiannual payments equal to 97% of the after-tax base interest on the notes, and had a contractual remedy against ABN and, if ABN failed to pay interest on the notes, FOP as well. While payment of the dividends was contingent on FOP's earnings, the transaction was arranged such that FOP's earnings were all but predetermined. In short, HP's investment earned it a limited return for a fixed period, and the Tax Court made no error in concluding that the investment was debt.
The tax code and regulations allow for the deduction of bona fide losses, I.R.C. § 165(a), but fees paid for a tax shelter cannot be deducted, see, e.g., Wells Fargo & Co. v. United States, 641 F.3d 1319, 1330 (Fed. Cir. 2011). If the IRS denies a deduction, the taxpayer bears the burden of proving by a preponderance of the evidence that the IRS determination was incorrect. Welch v. Helvering, 290 U.S. 111, 115, 54 S.Ct. 8, 78 S.Ct. 212 (1933). If the Tax Court concludes that the taxpayer has not met that burden, we must uphold that judgment if it is based on a permissible view of the evidence. See, e.g., MacDonald v. Kahikolu, Ltd., 581 F.3d 970, 976 (9th Cir. 2009).
The Tax Court's judgment — that HP's purported loss was really a fee paid for a tax shelter — was certainly based on a permissible view of the evidence. Indeed, internal HP communications referred explicitly to the "fee" that HP needed to pay AIG in order to participate in the FOP deal. A clawback agreement even obligated AIG to compensate HP if HP didn't get its desired tax results.
HP almost got its desired tax results.