Reuven S. Avi-Yonah is the Irwin I. Cohn Professor of Law at the University of Michigan. He would like to thank Kimberly Clausing, Jeffery Kadet, and Fadi Shaheen for helpful comments.
In this article, Avi-Yonah looks at the Tax Court’s recent Medtronic II decision and considers how the new corporate alternative minimum tax may provide a solution to the problem of adjudicating transfer pricing in the absence of comparables.
Copyright 2022 Reuven S. Avi-Yonah.
All rights reserved.
The recent Tax Court decision in Medtronic II illustrates once again the problem of adjudicating transfer pricing in the absence of comparables. The existing solutions to the profit-shifting problem — namely transfer pricing, subpart F, and global intangible low-taxed income — are all problematic. However, the new corporate alternative minimum tax may offer a roadmap to a possible solution.
The Basic Problem: Profit Shifting
The fundamental problem with outbound U.S. international taxation is profit shifting out of the United States. That problem, in turn, stems from respecting the separateness for tax purposes of U.S. parent corporations and their controlled foreign corporations. Until 2017, CFCs were generally not subject to tax on their foreign-source income until it was repatriated to the parent, resulting in a significant tax incentive to shift profits out of the United States. The traditional way to combat this form of profit shifting was transfer pricing. However, as illustrated again by the recent Tax Court opinion in Medtronic II,1 the arm’s-length standard that underlies transfer pricing analysis is broken, and it cannot be fixed in the absence of comparable transactions, which rarely exist.2
To address this problem, the United States has adopted a series of anti-profit-shifting regimes, the most important of which are subpart F, enacted in 1961, and the GILTI regime, enacted in 2017. But both these regimes are flawed. Instead, the new corporate alternative minimum tax may offer a better way forward.
Transfer pricing enforcement in the United States dates to the original enactment of what is now section 482 in 1928, when mandatory consolidation of CFCs with the parent was abandoned. In the 1930s, the new arm’s-length standard — which was developed in the treaty context — was added to section 482, and in 1968 the arm’s-length standard was given practical meaning by regulations embodying the three traditional methods relying on exact comparables. By the 1980s, it was clear that such comparables were difficult to find, and in 1995 the current transfer pricing regulations that included the comparable profits method and profit split, which relied on much less exact comparables, were adopted.
The Tax Court’s August 18 decision in Medtronic II, on remand from the Eighth Circuit, illustrates once again that current transfer pricing rules are inadequate to address the profit-shifting problem.
In the Medtronic case, the taxpayer licensed its intellectual property to its Puerto Rico subsidiary, Medtronic Puerto Rico Operations Co. (MPROC), for a low royalty rate. The IRS objected, arguing that MPROC was merely engaged in manufacturing and that the bulk of the profit belonged to Medtronic, where the IP was generated. The taxpayer responded that this analysis ignored MPROC’s crucial role in ensuring quality control, which was essential to maintaining market share in the medical device industry. The Tax Court originally accepted the taxpayer’s argument, rejected the IRS’s proposed adjustment based on the CPM, and used instead the taxpayer’s preferred method, the comparable uncontrolled transaction method, based on a litigation settlement between the taxpayer and Siemens Pacesetter Inc.3 However, the Tax Court made several adjustments to the CUT, reaching a royalty rate identical to the previous settlement between the taxpayer and the IRS. Essentially, the Tax Court seemed to be telling the parties that they could have prevented the waste of judicial resources (the opinion is 144 pages long) by continuing for 2005-2006 the same memorandum of understanding they had reached for previous years.
However, the Eighth Circuit reversed the Tax Court, ruling that the Pacesetter agreement was not comparable, based on the comparability factors in the regulations. The result was that neither side’s method could be adopted, and the case was remanded to the Tax Court.4
On remand, Tax Court Judge Kathleen Kerrigan adopted the Eighth Circuit’s approach, rejecting both the taxpayer’s CUT and the government’s CPM for being based on inadequate comparables. Instead, she calculated a royalty rate (48.8 percent) that fell roughly in between the ones suggested by the taxpayer (22 percent) and the IRS (67 percent).
Medtronic falls into a familiar pattern of pre-1995 transfer pricing cases.5 Initially, in Medtronic I, Kerrigan relied on the CUT method because that gives a judge some basis to establish the correct transfer price. This result is similar to cases like U.S. Steel6 and Bausch and Lomb,7 in which flawed comparables were used to reach a result favorable to the taxpayer.8 But when the Eighth Circuit correctly rejected the CUT method as well as the CPM, because as usual no adequate comparables were available, Kerrigan was forced on remand to essentially invent a royalty rate out of nothing. As a previous judge held long ago in another case that rejected all the comparables:
Even though we have rejected respondent’s 100-percent allocation of taxable income from [a subsidiary] to petitioner, the evidence indicates overwhelmingly that an allocation is necessary and proper in this case. Unfortunately, there is little quantitative evidence in this record upon which we can determine what a reasonable allocation of profits would be. Neither party has been particularly helpful to the Court in this regard. However, we must do the best with what we have.9
Or as another judge stated in a different transfer pricing case, “the evidence referred to supports [the result] in the weakest possible way. [We are] making bricks without straw.”10
The basic problem in all these cases is the arm’s-length standard. As has been repeatedly shown, when there are no comparables (which is the case in most transfer pricing cases involving unique intangibles, which are the most significant ones), there is no basis for deciding a case based on the arm’s-length standard because there are no comparable transactions showing what unrelated parties would have agreed to at arm’s length.11
Subpart F was originally enacted to address this limitation of transfer pricing, as illustrated by the Dupont decision.12 The most important innovation of subpart F was the base company rule, which was designed to address structures such as the one used in DuPont, in which goods were sold to a CFC in a low-tax jurisdiction and immediately resold at a higher price to CFCs in high-tax jurisdictions.
Subpart F functioned adequately as a backstop to transfer pricing until the 1980s. Under conditions of globalization and increased reliance on intangibles, it was not able to combat profit shifting because it assumed that offshore manufacturing would take place in high-tax jurisdictions, which was no longer true given tax competition and the rise of China as a manufacturing base. This helped make multinationals comfortable with unrelated contract manufacturers and, by extension, contract manufacturing business models.13 In addition, the adoption in 1996 of check-the-box regulations undermined even the original base company rule by enabling multinationals to designate CFCs as branches. There was also the congressional mandate to allow, beginning in 2006, foreign-to-foreign profit shifting through a new “look-through” rule and increased reliance on cost sharing, which enabled multinationals to shift IP out of the United States. The result was the nearly unlimited ability of multinationals to shift profits from both the United States and high-tax jurisdictions to low-tax jurisdictions, so that by 2017, U.S. multinationals had accumulated $3 trillion in low-tax jurisdictions.
In 2017 the United States adopted a participation exemption that, for the first time, enabled U.S. multinationals to repatriate foreign profits without paying tax on the dividend. However, to prevent profit shifting, the United States also adopted GILTI, which imposed current taxation on CFCs and treated all CFCs as a single unit.
GILTI significantly limits the profit-shifting potential of the participation exemption because it is imposed on all foreign profits above an exemption based on tangible assets, and most U.S. multinationals do not have many tangible assets offshore. But GILTI suffers from two significant flaws. The first and most important one is the lower rate (10.5 percent), which is half the domestic rate of 21 percent and therefore encourages profit shifting. The second problem is the ability to combine high- and low-tax jurisdictions, so if the taxpayer operates in a jurisdiction with a rate above the GILTI rate, it has an incentive to shift profits to a lower-tax jurisdiction to reach an average foreign tax rate of 13.125 percent, which, with foreign tax credits, eliminates the GILTI tax.
Empirical studies have demonstrated that although GILTI has reduced profit shifting compared with the pre-2017 situation, because of these two flaws it has not eliminated it.14
The new corporate AMT offers a promising alternative to GILTI. It includes the income of all CFCs in the base, and it does not have a different rate for foreign and domestic income. Moreover, it does not contain any exclusions for tangible property of CFCs, so that unlike GILTI, there is no incentive to build factories and create jobs offshore. Thus, conceptually there is no incentive to shift profits out of the United States under the corporate AMT.15
However, the corporate AMT has its own limitations. First, it only applies to the largest U.S. corporations, those with average income over $1 billion (although most of the large U.S. multinationals are included). Second, it is an AMT, and many U.S. multinationals would not be subject to it because their effective regular tax rate is higher than the corporate AMT rate, especially because the corporate AMT (despite being based on book income) allows deviations such as expensing domestic investments that significantly reduce its rate. Most importantly, the corporate AMT rate is too low to be an adequate substitute for the corporate tax, as evidenced by the revenue it is expected to raise, which is less than one-tenth of the revenue from the regular corporate tax.
Despite its flaws, the corporate AMT offers the most promising path to a permanent solution to the profit-shifting problem because it finally disregards the separateness of CFCs from the parent and applies the same rate to the U.S. and foreign income of the entire group.
To make the corporate AMT an effective anti-profit-shifting regime, the threshold should be lowered (for example, to $100,000 average annual income, as originally suggested by Sen. Elizabeth Warren, D-Mass.), and the rate should be increased to 21 percent or higher.16 If those two steps are taken, the regular corporate tax could be repealed in its entirety, and what is now the corporate AMT could be adopted as a substitute rather than as an alternative to the regular corporate tax. At a rate of 21 percent, the corporate AMT will raise more revenue than the existing corporate tax, since its base is broader.
Repealing the regular corporate tax also means repealing section 482 (as it applies to U.S. corporations), subpart F, and GILTI. The risk is that some future Congress might choose to reinstate the differential tax rate on foreign profits, and that in the absence of these safeguards, profit shifting might become rampant again. But several competing considerations seem to address this concern. First, if the corporate AMT rate is set at 21 percent, it will be at the lower end of the OECD average, which weakens the competitiveness argument for lowering the tax rate on foreign income. Second, the likely adoption of pillar 2 of BEPS 2.0 by our major trading partners will establish a 15 percent floor on their taxes, and if the United States adopts a 21 percent tax that applies to the entire income of its multinationals, other G-7 countries may follow suit. Third, after this kind of structural change is adopted, it is less likely to be reversed by future Congresses, as illustrated by the equivalence of the dividends and capital gains rate since 2003 (despite frequent rate changes for both). If a future Congress is concerned with competitiveness, it should reduce the overall corporate rate rather than reduce only the rate on foreign income, because it is the overall rate that determines the cost of capital. Finally, it is hard to base any tax reform proposal on trying to predict what future Congresses might do.
2 But see Coca-Cola v. Commissioner, 155 T.C. 145 (2020), in which a comparable was successfully used by the IRS. For a discussion, see Reuven S. Avi-Yonah and Gianluca Mazzoni, “Coca-Cola: A Decisive IRS Transfer Pricing Victory, at Last,” Tax Notes Int’l, Dec. 14, 2020, p. 1419.
4 Medtronic Inc. v. Commissioner, 900 F.3d 610 (8th Cir. 2018). See Avi-Yonah, “Medtronic: Has the Tide Turned for Transfer Pricing?” Tax Notes Int’l, Apr. 6, 2020, p. 21.
5 See generally 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).
7 Bausch & Lomb Inc. v. Commissioner, 92 T.C. 525 (1989), aff’d, 933 F.2d 1084 (2d Cir. 1991).
8 For a discussion, see id.
11 Avi-Yonah, supra note 4.
12 See Avi-Yonah and Nir Fishbien, “Once More, With Feeling: TRA 17 and Original Intent of Subpart F,” Tax Notes, Nov. 13, 2017, p. 959; Fishbien, “From Switzerland With Love: Surrey’s Papers and the Original Intent(s) of Subpart-F,” 38 Va. Tax Rev. 1 (2018).
13 See Jeffery M. Kadet, “BEPS Primer: Past, Present, and Future,” Tax Notes Int’l, July 6, 2020, p. 51; Avi-Yonah, “Tax Competition and Multinational Competitiveness: The New Balance of Subpart F,” Tax Notes Int’l, Apr. 19, 1999, p. 1575.
14 Kimberly A. Clausing, “Profit Shifting Before and After the Tax Cuts and Jobs Act,” 73(4) Nat’l Tax J. 1233-1266 (2020).
15 If a U.S. multinational is in an excess credit position (that is, its foreign tax rate exceeds the U.S. tax rate), there is still an incentive to shift profit out of the United States even if the U.S. rate on domestic and foreign income is the same, as long as the FTC limit of section 904 is not applied on a country-by-country basis. For example, assume U.S. corporation X has $100 of domestic income taxed at 15 percent. Foreign country A’s tax rate is 20 percent. Foreign country B’s tax rate is 10 percent. The United States taxes foreign income under the corporate AMT at 15 percent with FTC. Now assume X also has $100 of income from foreign country A subject to tax at 20 percent. X will pay $15 to the United States and $20 to country A, so it will get the FTC and have excess credits of $5. Total tax is $35/$200, or 17.5 percent. If X earns another $100 from the United States, it will pay $30 on U.S.-source income and $20 on foreign income for a total tax of $50/$300, or 16.6 percent, with $5 excess credits. If instead, X earns another $100 from foreign country B, it will pay $15 to the United States, $20 to country A, and $10 to country B. Total tax is $45/$300, or 15 percent, with no excess credits because of blending. Therefore, there is an incentive to shift profits from the United States to foreign country B.
However, (a) this problem does not arise if the U.S. multinational is in a excess limit position (that is, its foreign tax rate is lower than the U.S. tax rate), and this result is more likely if the United States raises the corporate AMT rate to 21 percent, as suggested in the article; (b) the same result (getting rid of the excess credit) can be accomplished by shifting income from the high-tax to the low-tax foreign country, rather than from the United States to the low-tax foreign country; and (c) if pillar 2 is adopted, foreign countries will have an incentive to adopt a 15 percent tax rate, which is equivalent to the corporate AMT rate, in which case no profit-shifting issue arises.