Reuven S. Avi-Yonah (firstname.lastname@example.org) is the Irwin I. Cohn Professor of Law at the University of Michigan Law School.
In this installment of Reflections With Reuven Avi-Yonah, part 1 of a two-part series, Avi-Yonah examines the quickly evolving pattern of successful litigation for the IRS, focusing on key cases from the past year.
In her final 2023 column,1 Mindy Herzfeld summarizes the main tax cases of 2023: Moore,2 Liberty Global,3 Christensen,4 YA Global,5 Farhy,6 Bittner,7 Coca-Cola,8 and 3M.9 It is remarkable what a good year this has been for the IRS and Justice Department litigators. Of these eight cases, they won four, lost three, and one is pending. But of the three losses, two (Christensen and Farhy) may plausibly be reversed on appeal, one (Bittner) is on a small issue, and in the pending case (Moore), the government seems likely to prevail.
These results may be a doubtful record for a private sector firm, but they stand in sharp contrast to a long string of IRS losses in major cases that has only begun reversing recently. It raises the question: What has changed?
This is particularly true for transfer pricing, probably the area of tax law that consumes the most IRS litigation resources. In the 40 years between 1979 (when the agency won in Dupont)10 and 2019 (when it won in Altera),11 the IRS lost U.S. Steel,12 HCA,13 Ciba-Geigy,14 Searle,15 Lilly,16 Bausch & Lomb,17 Sundstrand,18 Merck,19 Compaq,20 UPS,21 DHL,22 Xilinx,23 Veritas,24 and Amazon.25 Its only major achievement in that period was settling Glaxo26 for $3.4 billion in 2006. In Medtronic,27 the IRS lost in the Tax Court in 2016, won a partial reversal on appeal, lost again on remand, and a second appeal is now pending — along with two other large transfer pricing cases (Facebook and Perrigo).28 Even the victory in Altera, which was hard fought, was on a limited issue (the inclusion of stock options in the pool of costs to be shared under a cost sharing agreement), although it was significant because it reversed the previous IRS defeat in Xilinx.
The first sign that things were changing came in 2020, when the IRS decisively won Coca-Cola.29 This case is on appeal. In 2023 it was followed by two more IRS transfer pricing victories, suggesting that the first victory was not a fluke.
In 3M,30 the Tax Court with a reviewed 9-8 decision upheld the IRS’s determination that 3M could not avoid tax on royalties from its Brazilian subsidiary by relying on a nonpublicly promulgated Brazilian prohibition on paying royalties above 1 percent. The parties had stipulated that the arm’s-length rate was 6 percent, and the Tax Court held that this was the correct rate under section 482.31
The blocked income issue was governed by a 1994 regulation that set the standards for when the taxpayer could rely on a foreign ban on payments to avoid current taxation.32 The taxpayer conceded that it did not meet the procedural requirements of the regulation.33 Instead, the taxpayer argued that the regulation was invalid and, under the case law that preceded the regulation, the IRS was precluded from relying on section 482 to tax it on blocked income.
The Tax Court rejected this administrative law challenge to the regulation, reversing its position in Altera. This result is important because it reverses the trend of government losses on the Administrative Procedure Act issue. But the most interesting part of the opinion deals with the substance of the blocked income problem. The problem was created by the Supreme Court in First Security Bank, which held that the IRS could not allocate insurance commissions from a related party to a bank because the bank was statutorily forbidden from underwriting insurance.34 This rule was thereafter applied by the Tax Court to a Spanish law prohibiting payment of royalties to a related party and to a private Saudi decree setting the price of oil charged by Aramco to its affiliates.35
The majority opinion struggled valiantly to distinguish these precedents. It argued that the case law relied on the statement in the 482 regulations that the taxpayer had complete power to allocate the income in question, which does not appear in the 1995 version of the regulation that governed the tax year at issue in the case (2006). It also argues that the enactment of the commensurate with income (CWI) standard in 1986 governs the outcome in 3M and distinguishes it from the pre-1986 case law.
Neither of these arguments is persuasive. The definition of control in section 482 is famously broad and governs both the pre- and post-1986 results. Under that broad definition, it does not matter whether the taxpayer has the “complete power” to allocate the income under an external nontax legal prohibition. The point of section 482 is to produce a clear reflection of the taxpayer’s income from transactions with related parties.
Nor is CWI relevant to the result. CWI was intended to ensure that income from the transfer of intangibles be taxed to the transferor even if there were no arm’s-length comparables that required this outcome. CWI does not refer to blocked income, and there is no basis for assuming that Congress had blocked income in mind when it enacted CWI, as is clear from the legislative history that 3M cites extensively.36
The basic problem is that the blocked income cases were wrongly decided. Fundamentally, it should not matter for tax purposes whether there was nontax law that prohibits payments because the point is to accurately measure the taxpayer’s income. Tax law frequently creates results that differ from nontax law — for example, in the assignment of income cases, there was no dispute that the assignments were valid as a matter of contract law. The same rule applies to deemed dividends under subpart F and global intangible low-taxed income, which are includable in income even if there is a foreign prohibition on payment of actual dividends.37
Obviously, the Tax Court could not reach this result because First Security is a Supreme Court opinion. But as Chief Judge Kathleen M. Kerrigan stated in her concurrence in 3M:
First, First Security Bank is distinguishable on its face and therefore does not control this case. In Commissioner v. First Security Bank . . . the Court was interpreting a domestic Federal banking law which implicitly prohibits national banks from acting as insurance agents in places with a population of 5,000 or more. Here the Tax Court is tasked with interpreting a foreign law. Additionally, the law considered in First Security Bank was one of general application whereas the blocked income regulation has a specific use: It is aimed at restrictions that bar payments only to foreign companies affiliated with the local business. I find these factual distinctions to be significant, confining First Security Bank’s holding to the circumstances presented in that case.
The basic problem in extending First Security to foreign law is that, as is well known from the foreign tax credit context, foreign law is frequently adjusted to meet the interests of U.S. taxpayers that are major investors in a foreign country. That is particularly true in a case like Aramco,38 when only one U.S. taxpayer is involved. The Saudi oil minister was probably fully aware that his decree requiring Aramco to sell oil to related refineries at below-market prices (but permitting those refineries to resell at market prices) conferred a huge tax benefit on the most important foreign investor in Saudi Arabia.
Thus, what the Tax Court should have done is reverse its rulings in Procter & Gamble and the oil cases, distinguishing First Security. It should have held that the whole concept of blocked income is inconsistent with the statutory purpose of section 482, which is to clearly reflect income for tax purposes.
3M is certain to be appealed, given the closeness of the outcome. In preparation for this appeal, Treasury should revoke the blocked income regulation, thereby mooting the APA argument. Instead, it should argue that the Eighth Circuit ignore the decisions of the circuit courts in Proctor & Gamble and Aramco and should hold that foreign legal prohibitions on the payment of income have nothing to do with the outcome under section 482. After all, the pre-2017 “trapped income” phenomenon has shown that large multinationals have no problem financing U.S. activities (including paying taxes) even when unable to access their foreign-source income.
II. Coca Cola II
3M was followed by another IRS victory involving the blocked income regulation. In Coca-Cola,39 the Tax Court considered the remaining issue in the case: whether the blocked income provision should apply to eliminate the amount the government said was owed by Coca-Cola because Brazilian law limited the amount of royalty and technology transfer fees a subsidiary could pay its foreign parent. Judge Albert G. Lauber held that while Brazilian law may have restricted the type of payment — that is, a royalty — it did not place limitations on the amount of the payment. In other words, the $900 million that was properly characterized as a U.S. taxable dividend could have been paid as a dividend from the Brazilian subsidiary, even if not as a royalty.
These two cases represent remarkable achievements for the IRS litigators. In the next column, I will analyze their victories in the non-transfer pricing cases decided last year, YA Global and Liberty Global. And I have previously argued that the decision in Christensen, which they lost, is clearly erroneous and should be reversed.40
What has changed? Because it is not a matter of increased resources (the new IRS funding from 2022 could not yet have influenced these litigation outcomes), I believe it is the increased quality of the government litigators. And this was on full display in the Supreme Court argument in Moore, in which Solicitor General Elizabeth Prelogar was amazing, fluently citing subsections and showing full command of the tax law. While most tax cases will never attract such attention at the highest level, we can hope that, with increased funding, the IRS will not be so frequently out-litigated by private firms with infinitely superior resources.
26 GlaxoSmithKline Holdings (Americas) Inc. v. Commissioner, No. 5750-04 (T.C. 2006).
28 Facebook Inc. v. Commissioner, No. 21959-16; Perrigo Co. v. United States, No. 1:17-cv-00737 (W.D. Mich. 2021). On these cases, see Ryan Finley, “What’s at Stake in the Medtronic II Appeal, Part 1,” Tax Notes Int’l, Oct. 2, 2023, p. 59; Finley, “What’s at Stake in the Medtronic II Appeal, Part 2,” Tax Notes Int’l, Oct. 9, 2023, p. 221; Finley, “Defending the Income Method as Arm’s Length in Facebook,” Tax Notes Int’l, May 8, 2023, p. 683; and Finley, “Separating Transfer Pricing From Economic Substance in Perrigo,” Tax Notes Int’l, Jan. 23, 2023, p. 435.
29 Coca-Cola Co. v. Commissioner, 155 T.C. No. 10 (2020).
36 For the history of CWI, see Avi-Yonah, “The Rise and Fall of Arm’s Length: A Study in the Evolution of U.S. International Taxation,” 15 Va. Tax Rev. 89 (1995), which is cited extensively in 3M, 160 T.C. No. 3.
38 Texaco Inc. & Subs. v. Commissioner, 98 F.3d 825 (5th Cir. 1996).