Tax Notes logo

Viewing the GILTI Tax Rates Through a Tax Expenditure Lens

Posted on Dec. 12, 2022
Stephen E. Shay
Stephen E. Shay
Robert J. Peroni
Robert J. Peroni
J. Clifton Fleming Jr.
J. Clifton Fleming Jr.

J. Clifton Fleming Jr. is the Ernest L. Wilkinson Chair and professor of law at the Brigham Young University J. Reuben Clark Law School, Robert J. Peroni is the Fondren Foundation Centennial Chair for Faculty Excellence and professor of law at the University of Texas School of Law at Austin, and Stephen E. Shay is the Paulus Endowment Senior Tax Fellow at Boston College Law School. The authors are thankful for helpful comments from participants in works-in-progress workshops at the Vienna University of Economics and Business on October 4; the Southeastern Association of Law Schools annual meeting in Destin, Florida, on August 31; and the Law and Society Association annual meeting in Lisbon, Portugal, on July 14.

In this article, the authors explain that for U.S. corporations earning foreign-source active business income through controlled foreign corporations, pre-Tax Cuts and Jobs Act tax deferral planning is now largely obsolete and has been replaced with planning that centers on avoiding subpart F income and maximizing global intangible low-taxed income. They also explain that the low GILTI tax rates are a tax expenditure that fares poorly under cost-benefit analysis, even when taking into account the new U.S. corporate minimum tax and the possibility of modifications that would make the GILTI regime pillar 2 compliant.

Copyright 2022 J. Clifton Fleming Jr., Robert J. Peroni, and Stephen E. Shay.
All rights reserved.

I. Introduction

Since its enactment, the modern U.S. income tax has used a so-called worldwide system that attempts to reach both the U.S.-source and foreign-source income of U.S. residents (that is, U.S. citizens, U.S. resident aliens, and domestic corporations), with some exceptions.1 Of course, the foreign-source income also often bears a foreign income tax, and since 1918, the United States has mitigated this double taxation by allowing a credit for the foreign income tax.2 Thus, if the foreign tax was at least equal to the U.S. tax before the foreign tax credit, the United States would not collect any revenue, but if the U.S. tax exceeded the foreign tax, it would collect the excess, commonly referred to as a residual tax.3 Before the 2017 Tax Cuts and Jobs Act, U.S. multinational corporations focused their outbound tax planning on reducing or eliminating this residual tax by using two related tax expenditures: deferral,4 and nearly unlimited cross-crediting of foreign income taxes on income from one country against U.S. tax on income taxed by another country at an effective rate below the U.S. rate.5 This article explains how the TCJA rendered that planning largely obsolete and replaced it with a new approach dictated by the hybrid global intangible low-taxed income and limited partial exemption regime. More precisely, this article describes the TCJA’s effect on tax planning for U.S. residual tax on foreign-source active business income earned by U.S. parent corporations through the activities of their controlled foreign corporations and examines the application of tax expenditure analysis to the changed situation.

II. Before the TCJA

Until the TCJA’s effective date,6 the U.S. residual tax on CFC income could often be deferred until repatriation — that is, until the CFC made dividend distributions to its U.S. parent corporation or the U.S. parent sold shares of the CFC at a price that reflected the CFC’s accumulated income. This tax deferral reduced the effective U.S. tax rate on foreign subsidiary income. The degree of reduction depended on both the deferral period’s length and the applicable discount rate.7 The beneficial effect of deferral could be enhanced by cross-crediting — that is, by crediting foreign tax paid in excess of the U.S. tax on income earned in high-tax foreign countries against the U.S. residual tax due on income earned in low-tax foreign countries. To the extent of that cross-crediting, the residual tax was eliminated rather than merely applied at a reduced rate.8

The foregoing features of the pre-TCJA system were limited by section 951(a), contained in subpart F of the code,9 which imposes an immediate U.S. tax on the 10-percent-or-more U.S. shareholders’ pro rata shares of specified types of foreign income earned by a CFC. This is principally passive income and business income generated by using a low-taxed CFC as a base for selling goods produced, or services performed, in a foreign country other than the CFC’s country of incorporation.10 To be covered, the sales must be of products purchased from or sold to a related person, and the services must be performed for or on behalf of a related person.11 Unless subject to a branch rule exception,12 sales of goods manufactured by the CFC are excepted, even if sold to a related person or to customers outside the CFC’s country of incorporation and regardless of whether the CFC’s manufacturing operations occur outside its country of incorporation.13 Thus, a great deal of U.S. residual tax on CFC active business income could be deferred under these largely ineffective anti-deferral rules.

Indeed, it was reasonable to argue that before the TCJA, tax planning by U.S. multinational enterprises for investments in low-tax foreign countries focused on (1) making certain that the deferral of U.S. residual tax was possible and prolonged, (2) avoiding subpart F’s negative impact on deferral, and (3) enhancing the benefit of deferral through cross-crediting and aggressive transfer pricing. From this viewpoint, deferral was the main driver of outbound tax planning by U.S. MNEs before the TCJA.

Although the Congressional Budget Office and the staff of the Joint Committee on Taxation consistently regarded deferral and its preferential effective rate of U.S. residual tax as a tax expenditure,14 deferral apologists argued that this was wrong and that deferral should not be subjected to the cost-benefit scrutiny accompanying tax expenditure analysis. Instead, so the argument went, the deferral benefit was merely a “natural” consequence of the income tax’s fundamental structure,15 which, consistent with the credulous Moline Properties doctrine,16 generally treats shareholders and corporations as separate taxpayers, even when the corporation has only one shareholder and virtually no meaningful activity other than respecting corporate formalities.17

III. After the TCJA

In the aftermath of the TCJA, the importance of deferral in U.S. outbound multinational income tax planning has shrunk dramatically for several reasons. First, the TCJA generally required that U.S. parent corporations include in income all income accumulated by their CFCs after 1986 and not previously taxed under the subpart F provisions.18 These amounts can now be distributed to U.S. parents without incurring a further U.S. tax as previously taxed earnings and profits.19 Thus, there is no U.S. tax advantage to U.S. parent corporations from continuing to hold these accumulated earnings in CFCs, and to this extent, deferral has become an obsolete planning tool.

The next factor in deferral’s decline is the TCJA’s adoption of a feature that imposes an immediate, nondeferred tax that after a 50 percent deduction is at an effective 10.5 percent rate (13.125 percent after 2025) on GILTI of CFCs that is attributable to U.S. shareholders that are U.S. corporations.20 For this purpose, a U.S. shareholder is the owner (directly, indirectly through a foreign entity, or constructively) of at least 10 percent of the vote or value of the CFC’s stock.21

In simplified terms, GILTI is most of a CFC’s foreign-source income that is tested income — that is, income that is not subpart F income, U.S. trade or business income, or foreign-source income of the CFC that is subject to an effective rate of foreign income tax greater than 18.9 percent (high-foreign-taxed income).22 However, tested income is GILTI only to the extent that it is net tested income (that is, tested income in excess of tested losses of loss CFCs) and only to the extent that net tested income exceeds an exemption amount equal to 10 percent of the U.S. shareholder’s share of CFCs’ investment in tangible depreciable business property23 (referred to in the code as the net deemed tangible income return (NDTIR)). Thus, neither the subpart F income regime nor GILTI applies to a CFC’s NDTIR or high-foreign-taxed income when either is attributable to a domestic corporate 10 percent shareholder.

Most importantly for purposes of this article, the U.S. tax on GILTI is not deferred. It is imposed on a U.S. shareholder of the CFC for the CFC’s year in which the tested income underlying the GILTI is earned by the CFC. Moreover, a U.S. corporate parent can receive tax-free distributions from the CFC of both (1) the CFC’s income that bears the GILTI tax24 and (2) the CFC’s income that escaped the GILTI tax because of the exemption for NDTIR and high-foreign-taxed income.25 Thus, for both a CFC’s GILTI and its business income covered by the exemption, there is no U.S. tax for a corporate U.S. shareholder to defer, and it is pointless for those shareholders to accumulate these amounts in a CFC for tax reasons.

Moreover, the GILTI tax does not apply to income that is subject to current inclusion under section 951(a)(1) of the subpart F regime. U.S. parent corporations continue to bear an immediate U.S. tax on that income at the regular U.S. corporate tax rate,26 and that income can then be distributed tax-free to a CFC’s U.S. parent.27 Thus, for income covered by section 951(a)(1), no U.S. tax is susceptible to deferral, and accumulation of that income in a CFC is pointless as a tax strategy.

Finally, section 245A provides that CFC foreign-source income that escapes immediate taxation under the GILTI regime (including the NDTIR and high-foreign-taxed income) or under section 951(a)(1) can be distributed tax-free by a CFC to U.S. corporations that are U.S. shareholders as defined above.28 Consequently, for that income, there is no U.S. tax to defer, and it is pointless to accumulate the income in a CFC for tax reasons.

The foregoing discussion means that deferral effectively is no longer available to U.S. parent corporations that are 10 percent shareholders in CFCs. If a U.S. corporate parent does not meet the 10 percent stock threshold for U.S. shareholder status, neither the section 245A deduction nor sections 951 and 951A will apply to the U.S. parent, and the U.S. tax on the U.S. shareholder’s share of the CFC’s income is subject to deferral, just as it was under pre-TCJA law.29 The amount of CFC earnings attributable to individuals and less-than-10-percent U.S. corporate shareholders likely is a relatively small portion of the foreign-source income earned by U.S. corporations operating through CFCs, which is the focus of this article.

Thus, the role of deferral for U.S. parent corporations has been largely eliminated by the TCJA, albeit at considerable cost in forgone revenue. That, however, does not mean that the foreign-source active business income earned by U.S. parent corporations through CFCs is no longer the beneficiary of favorable U.S. tax treatment. To the contrary, the reduced U.S. tax rate that was achieved indirectly through deferral has been replaced with (1) the zero U.S. effective tax rate that applies under the GILTI regime to the sum of a CFC’s NDTIR plus any of the CFC’s other foreign-source non-subpart F income that is not tested income (including, most importantly, income that has been subject to foreign income tax at an effective rate greater than 18.9 percent — that is, 90 percent of the 21 percent tax rate in section 11),30 and (2) the 10.5 percent GILTI effective rate applicable to a CFC’s foreign-source non-subpart F income that exceeds the foregoing zero-taxed income. Thus, the tax planning focus of U.S. parent corporations has shifted from achieving deferral to maximizing the amount of low-taxed foreign income that qualifies for the 0 and 10.5 percent GILTI rates.31

IV. Tax Expenditure Characterization

The favorable GILTI rates are not achieved indirectly through reliance on the separateness of corporations and shareholders, as was the case with the reduced residual tax rate that was accomplished through deferral and cross-crediting under prior law (and current law to the extent deferral still applies). Instead, the advantageous GILTI rates are applied directly and transparently in the relevant provisions of the code: sections 951A(b), 951(c)(2)(A)(i)(III), 959(a), and 245A,32 which effectively apply a zero U.S. tax rate to NDTIR and high-foreign-taxed income; and section 250, which allows a deduction for 50 percent of the GILTI inclusions of any domestic corporation. Thus, the argument that deferral is not a tax expenditure because it is merely the natural consequence of a fundamental structural feature of the income tax — the separateness of corporations and shareholders — is unavailable for the low GILTI effective rates. Those rates appear to be indisputable tax expenditures that should be subjected to the cost-benefit analysis required by tax expenditure analysis and repealed if they fail under it.

To be specific, the TCJA’s application of an effective U.S. income tax rate of 0 percent after the section 245A dividends received deduction, instead of the full 21 percent rate, to a CFC’s NDTIR plus the CFC’s foreign-source income that has been subject to an effective foreign income tax rate greater than 18.9 percent but less than 21 percent,33 is a subsidy to the underlying foreign activities and clearly a tax expenditure. The same conclusion applies regarding the TCJA’s application of the effective 10.5 percent rate (after the section 250 deduction) to a CFC’s GILTI that exceeds the foregoing zero-rate-eligible amount.34 The section 245A 100 percent dividends received deduction does not achieve the appropriate tax policy objective of avoiding double U.S. corporate taxation as do other dividends received deductions. It is in substance a 0 percent tax rate. Similarly, the 50 percent section 250 deduction involves no economic outlay and in substance exempts 50 percent of net tested income to reach GILTI. While nominally taxed at a full rate, in substance what generally is thought of as GILTI (but really is net tested income) is taxed 50 percent at a full 21 percent rate and 50 percent exempted.

Proponents of the preferential GILTI rates might argue that they are merely tax cuts and tax cuts are not tax expenditures. That point clearly applies to the TCJA’s general reduction of individual and corporate tax rates in sections 1 and 11. General tax rate reductions are not viewed as having a subsidy effect and are not tax expenditures. The 0 and 10.5 percent GILTI rates, however, apply to a discrete class of income: CFC active business income that is attributable to the CFC’s corporate U.S. shareholders. A targeted reduction of this type properly is characterized as a subsidy in the form of complete or partial exemption of a category of income that is indisputably distinct from a broad rate cut, such as the TCJA’s substitution of a flat 21 percent corporate rate for the graduated corporate income tax rates that reached a top rate of 35 percent under pre-TCJA law.

The flip side of the preceding argument that GILTI rates are not a tax expenditure is that because the effective rate of U.S. tax under deferral could be less than 10.5 percent, and even zero when cross-crediting was used, the zero rate on NDTIR and high-foreign-taxed income is not a change from the deferral results under prior law and that the 10.5 percent GILTI tax rate on income exceeding NDTIR and high-foreign-taxed income could be a tax increase when compared with deferral outcomes. Therefore, it might be argued, the GILTI regime actually increases U.S. taxation of foreign income earned through CFCs and does not deliver a subsidy. Consequently, there is no tax expenditure involved. This argument fails, however, because it ignores that the baseline U.S. rate on corporate business income is now 21 percent and that the 0 and 10.5 percent GILTI effective rates are significant reductions from that baseline (or the 100 percent and 50 percent exemptions are exceptions to the general rule of income inclusion). In other words, the GILTI regime may increase the taxation of some foreign income generated by CFC activities, as compared with pre-TCJA law, but this new regime still provides a substantial tax subsidy for a significant portion of foreign-source income earned through a CFC as compared with the 21 percent rate imposed on U.S.-source income earned by U.S. corporations and on U.S. corporations’ pro rata shares of foreign-source income that is subpart F income.35

The case for treating the GILTI section 250 deduction and the section 245A dividends received deduction as tax expenditures seems overwhelming regardless of the position one might have taken in the debate over whether the indirect rate reduction achieved before the TCJA through deferral and cross-crediting was a tax expenditure.36 The 0 and 10.5 percent CFC dividend and GILTI effective rates cannot be sheltered from tax expenditure analysis by insisting that they are a natural or normal part of an income tax structure, that they are “mere” rate cuts, or that they are an overall tax increase. They are pure subsidies for a defined class of foreign income earned by a limited subset of taxpayers.37 If they are to remain in the code, they must be able to withstand the rigors of cost-benefit analysis.38

V. Applying Cost-Benefit Analysis

The GILTI regime imposes significant direct and indirect revenue and efficiency costs, as detailed below.

A. Distortive Effect

The GILTI regime can inefficiently distort the business location decision by causing an investment in a low-tax foreign country that has an inferior pretax return, when compared with an investment in the United States, to swing to a superior after-tax return.

Example: USCo must choose between building a new U.S. factory that will produce an annual pretax return of $1.1 million and a new factory in Newlandia that will produce an annual pretax return of $1 million. If the factory is built in Newlandia, the foreign tax rate will be zero under a business development regime (that is, a so-called tax holiday), and the maximum U.S. tax rate under the GILTI regime will be 10.5 percent even if the NDTIR is zero.39 If the factory is built in the United States, the U.S. tax rate will be 21 percent. The after-tax results are as shown in the table.

U.S. Factory

$1.1 million before tax

$869,000 after tax

Newlandia Factory

$1 million before tax

$895,000 after tax (assuming zero NDTIR and zero high-foreign-taxed foreign income)

Thus, the preferential GILTI tax rate has placed the economically inferior Newlandia investment in a superior after-tax position. This provides a tax incentive for USCo to pursue the less economically desirable Newlandia investment. To the extent of the NDTIR (as well as the reduced GILTI effective rate), GILTI also encourages Newlandia and other source countries to engage in tax competition to attract U.S. investment.

Moreover, the preceding example made the unrealistic assumption that the Newlandia investment would have a zero NDTIR. If, however, USCo spent $5 million on the new depreciable factory building (ignoring the amount spent on the underlying land), the NDTIR would equal 10 percent of that sum; the amount taxable under the GILTI regime would only be $1 million - $500,000 NDTIR = $500,000; the U.S. tax would fall to 10.5 percent * $500,000 = $52,500; the after-tax return of the Newlandia factory would increase from $895,000 to $947,500; and the U.S. tax incentive to locate the factory in Newlandia would substantially increase. Indeed, every additional $1,000 spent on the Newlandia factory building would decrease the GILTI tax base by 10 percent of that same amount (that is, $100) and save $10.50 of U.S. tax. Thus, not only can the GILTI regime distort USCo’s decision about whether to locate the new factory in the United States or Newlandia, it can also provide an incentive for USCo to invest more in a Newlandia factory than it otherwise would.

B. Profit Shifting

The GILTI regime confers a tax preference on foreign-source non-subpart F income earned though a foreign subsidiary that is a CFC. The result is an incentive to earn low-foreign-taxed foreign income, avoid subpart F, and engage in aggressive transfer pricing to shift income from the U.S. parent corporation’s 21 percent U.S. rate to the more favorable GILTI effective rate. Stated differently, when compared with the alternative of taxing CFC income at the normal 21 percent U.S. corporate rate, the GILTI regime places continued stress on U.S. transfer pricing enforcement, which is one of the most problematic, uncertain, and expensive elements of the U.S. international income tax system for both taxpayers and the IRS.

VI. Competitiveness

The most frequently used argument for incurring the GILTI regime’s preceding costs is that the GILTI tax rates (zero on the NDTIR plus high-foreign-taxed income and 10.5 percent on the remaining GILTI tax base) are necessary to make U.S. corporations competitive in low-tax foreign markets. The argument can be stated as follows: Local businesses in a low-tax foreign country pay only the local income tax on their in-country profits. The same is true of foreign corporations that operate in the low-tax country but are resident in a country that exempts foreign-source income from residence-country tax. Without the favorable GILTI tax rates, U.S. parent corporations would be unduly disadvantaged when competing in low-tax foreign countries because, in addition to the low foreign tax, they would pay a current home-country residual tax on their foreign profits that are covered by the GILTI regime, while their local and exemption-country competitors would pay only the low foreign tax. Therefore, the argument goes, the United States should provide a compensating tax subsidy for the foreign-source active business income of U.S. resident companies.

This argument by U.S. GILTI proponents that the GILTI tax rates are necessary for U.S. multinationals to compete in the global marketplace is unsupported by empirical evidence that there is an international competitiveness problem that is solved by the GILTI regime. Claims by GILTI advocates that a competitiveness problem exists are rendered questionable at best by the extensive overseas success of many U.S. businesses.40 While taxation affects business decisions, where is the objective evidence (as opposed to anecdotes and special pleading) of a systemic negative competitiveness outcome that would result from taxing CFC income at the regular corporate rate instead of being attributable to labor cost differentials, product quality differences, regulatory differences, and other nontax factors? Stated differently, if industries face an international competitiveness problem, what is the evidence that U.S. taxation (after taking into account FTCs) is the cause, and how is the GILTI subsidy the best solution? Answers to these questions have not been forthcoming from proponents of the GILTI subsidy.41

Of course, a GILTI advocate might shift ground by conceding that U.S. businesses are competing effectively abroad but then arguing that this success is because of the generous tax assistance provided by the pre-TCJA U.S. international tax regime, that withdrawal of this aid without a compensating substitute would have caused U.S. businesses to flounder in foreign markets, and that tax assistance should be continued in the form of the section 245A dividends received deduction and the section 250 GILTI deduction. This argument fails, however, because there is an absence of evidence that the ongoing competitive success of U.S. MNEs depends on special tax benefits given only to foreign-source income.

Moreover, the GILTI regime is a poorly targeted subsidy device. For example, it is fully available regardless of whether the beneficiary has little competition in the foreign country (such as a pharmaceutical company selling one-of-a-kind, patent-protected drugs) or faces fierce competition. Also, GILTI assistance is fully available even if the U.S. corporate beneficiary’s principal competitor in a particular foreign country is a resident of the United States. The competitive struggles among U.S. information technology companies and U.S. soft drink producers in foreign markets are examples. As these points illustrate, the GILTI regime is a poorly designed way to enhance competitiveness vis-à-vis significant foreign competitors.

Even if we were to stipulate that the United States faces a systemic international competitiveness problem, it is doubtful that providing tax assistance to U.S. MNEs ranks very high among potential remedies. For example, strengthening public education in the United States holds greater promise of effective results.42

Finally, even if U.S. MNEs have a competitiveness problem for which the GILTI regime is an efficacious response, that is not the end of the cost-benefit inquiry. GILTI assistance must also compete against other uses of government revenue. To be specific, should large U.S. corporations receive an expensive government subsidy,43 which redounds substantially to the benefit of their shareholders, when the United States is grappling with budget deficits that raise affordability obstacles to continuing the child tax credit, expanding healthcare benefits, mitigating the harm of climate change, providing educational assistance, and attending to other pressing needs? For example, federal revenue spent on U.S. K-12 education for fiscal 2019 totaled $60.3 billion,44 while the staff of the JCT scored the GILTI provisions as a $45.4 billion tax expenditure for fiscal 2020.45 Regrettably, the readily available data do not allow a direct year-to-year comparison, but when the GILTI subsidy for fiscal 2020 is scored as a tax expenditure that is more than 75 percent of total federal spending in the preceding year on K-12 education, a salient question of priorities is presented.

VII. The Way Forward

For the above reasons, we conclude that the GILTI regime is a tax expenditure that apparently fails cost-benefit analysis. The solution seems straightforward. By substituting the regular 21 percent section 11 tax rate in place of the 0 percent tax rate that applies to NDTIR and high-foreign-taxed income and the 10.5 percent tax rate that applies to GILTI, the undesirable tax expenditure feature of the GILTI regime can be eliminated. Subpart F and GILTI can then be combined into a single integrated anti-deferral regime, which would be a welcome simplification.46

VIII. The New Corporate Minimum Tax

The recently enacted Inflation Reduction Act (P.L. 117-169) creates a new 15 percent corporate alternative minimum tax that generally applies to corporate adjusted financial statement (book) income (other than book income of an S corporation, a regulated investment company, or a real estate investment trust) if the corporation has average annual adjusted financial statement income of more than $1 billion for the three-tax-year period ending with the tax year preceding the tax year in question.47 If the minimum tax exceeds the corporation’s section 11 tax plus its base erosion and antiabuse tax liability, the corporation’s tax liability for the year is the sum of the excess plus its section 11 tax and its BEAT liability.48

The JCT estimated that the new minimum tax would apply to only about 150 corporations per year before narrowing amendments were made during the legislative process.49 More recent estimates suggest that only about 125 corporations will be affected each year.50 Either way, the coverage of the new minimum tax is miniscule compared with the more than 1.4 million C corporations that are exposed to the regular 21 percent corporate income tax rate,51 and very small in relation to the approximately 2,900 corporations that report GILTI.52 Thus, even if the new corporate AMT increases the effective tax rate on GILTI to 15 percent for some corporations, it does not change the conclusion of this article that the tax treatment of GILTI is a pure subsidy for a discrete class of foreign income earned by a limited subset of taxpayers. The case for treating the GILTI section 250 deduction as a tax expenditure that must be subjected to cost-benefit analysis remains compelling.

IX. The Proposed Pillar 2 GLOBE

The OECD/G-20 inclusive framework’s proposed pillar 2 global anti-base-erosion model rules would substantially restrict tax competition by imposing a global minimum tax rate of 15 percent on income earned by large multinational groups (generally, with consolidated revenue of €750 billion or more). Under pillar 2, when income of an in-scope company (and its affiliates in a country) is taxed at a rate below 15 percent, pillar 2 effectively permits other countries to impose an additional tax to bring the effective rate to the agreed minimum under an income inclusion rule applied to an ultimate parent company or, if the ultimate parent company’s country does not adopt pillar 2, under an undertaxed profits rule that may be imposed by any participating country in which an affiliate of the ultimate parent company is a tax resident. Based on the most recent IRS data from 2019, it is reasonable to expect that somewhere in the range of 1,150 U.S. companies might be subject to the pillar 2 minimum tax.53

The United States has supported pillar 2, but as a result of opposition from the U.S. multinational business community transmitted through U.S. senators, the United States has failed to adopt rules that would be sufficiently convergent with pillar 2 rules to constitute GILTI as a qualifying IIR for pillar 2 purposes. Even if that were the case, under the preceding analysis, we would still consider the section 245A and GILTI section 250 deductions to constitute tax expenditures in the relevant U.S. context because of the difference between the 15 percent pillar 2 rate and the 21 percent U.S. corporate rate.

The only conclusion to be drawn from the U.S. multinational community’s political opposition to pillar 2 is that it apparently believed pillar 2 would die or be materially delayed if not adopted by the United States.54 Rather than support the far more level playing field that would result from international cooperation under pillar 2, the U.S. multinational community, thinking it could kill or delay the project, opted for a position that would leave it arguably disadvantaged by GILTI for whatever period would elapse until pillar 2 or a substitute is adopted by other countries.

The U.S. failure to adopt a pillar 2-compliant GILTI regime, however, materially risks a large portion of the potentially in-scope U.S. multinationals paying pillar 2 top-up taxes to other countries if a relatively small number of countries adopt pillar 2 rules in their domestic law. Canada, France, Germany, Italy, the Netherlands, Spain, Switzerland, and the United Kingdom have announced or otherwise put forward plans to adopt pillar 2 rules without waiting for other countries to act, including the EU as a whole or the United States.55 Based on IRS country-by-country data, U.S. companies in the scope of pillar 2 have many reporting entities in those countries: Canada (995), France (593), Germany (681), Italy (553), the Netherlands (626), Spain (515), Switzerland (451), and the United Kingdom (836). If each of these countries (or even a plurality) applies a UTPR, the top-up tax for a large portion of in-scope U.S. multinationals will potentially be paid to those countries instead of to the United States under a pillar 2-convergent U.S. IIR. The opposition of the U.S. multinational business community and its congressional allies to U.S. adoption of a pillar 2-compliant GILTI regime risks a self-inflicted revenue loss on the United States. The multinationals risk the level playing field that they apparently do not want, preferring lower GILTI tax rates instead. By taking this position, the multinationals — having already acquiesced in the adoption of GILTI with a section 250 deduction in the TCJA — would appear to further undermine their arguments that U.S. residence-country tax is a major inhibiting factor in their ability to compete outside the United States.

X. Conclusion

Before the TCJA, the outbound income tax planning of U.S. MNEs was arguably focused on maximizing the subsidy provided by deferral of U.S. residual tax (particularly when combined with cross-crediting). The TCJA rendered deferral largely irrelevant and replaced it with strategies to maximize the amount of CFC income that qualifies for preferential tax rates provided by the GILTI regime. These rates are indisputably a tax expenditure that should be subjected to the same cost-benefit analysis that applies to direct government expenditures. This conclusion is not weakened by either the recently adopted U.S. corporate minimum tax or the possibility of the United States modifying the GILTI regime to make it pillar 2 compliant.

The GILTI rates fare poorly under cost-benefit analysis. They are defended principally as a subsidy to offset a broad international competitiveness challenge faced by U.S. corporations. However, the existence of that general problem is empirically unsupported, and if it exists, the GILTI rates are a poorly targeted remedy. More important, it seems unwise to divert revenue to trying to improve the international profitability of large U.S. corporations at the expense of other pressing U.S. needs.


1 See, e.g., section 245A (dividends received deduction for the foreign-source portion of dividends paid by a foreign corporation to 10-percent-or-more U.S. corporate shareholders, enacted as part of the 2017 Tax Cuts and Jobs Act) and section 911 (exclusion for a limited amount of foreign-source earned income of qualified U.S. individuals).

2 See Robert J. Peroni, Karen B. Brown, and J. Clifton Fleming Jr., Taxation of International Transactions: Materials, Text, and Problems 369-370, 410 (2021). The credit is supposed to be limited to the amount of the pre-foreign tax credit U.S. income tax on the relevant category of foreign-source income, but defective source and expense allocation rules and rules allowing cross-crediting within a limitation category undermine this limitation. See sections 861(b) and 904(d); and former section 863(b) (before amendment by the TCJA). For an explanation of the dynamics of cross-crediting and the tax policy issues that it raises, see Fleming, Peroni, and Stephen E. Shay, “Worse Than Exemption,” 59 Emory L.J. 79, 113-118, 132-145 (2009).

3 Under section 904(a), the effective rate comparison is in relation to the U.S. net income tax base. See sections 904(a) and 861(b).

4 Deferral of U.S. residual tax on foreign-source business income is a tax expenditure because it creates a time-value-of-money tax saving by deviating from the baseline norm of current taxation of business income. See Fleming and Peroni, “Reinvigorating Tax Expenditure Analysis and Its International Dimension,” 27 Va. Tax. Rev. 437, 528-541 (2008).

5 Before the TCJA, the FTC had only two limitation baskets, and the result was virtually unlimited cross-crediting (except for passive income). See Fleming, Peroni, and Shay, supra note 2, at 134. Virtually unlimited cross-crediting is a tax expenditure because it generates tax savings that are more generous than could be obtained under a theoretically correct per-item FTC limitation system or under a next-best per-country limitation, which is a pragmatic approximation of a per-item system. See Fleming and Peroni, supra note 4, at 543-546. A multi-basket limitation system, such as the prior law version of section 904(d) enacted as part of the Tax Reform Act of 1986, reduces, but does not eliminate, the tax expenditure feature in the FTC limitation system because it still allows excessive cross-crediting of high foreign taxes on some foreign-source business income against the U.S. residual tax on low-taxed foreign business income.

6 The TCJA became generally effective for tax years starting after 2017. The dividend exemption provision, however, was effective for distributions after December 31, 2017, so multinational taxpayers with non-calendar fiscal years were presented with a potential tax avoidance opportunity. Their foreign subsidiaries could make distributions of appreciated assets in the so-called effective date doughnut hole period, between January 1, 2018, and the beginning of the relevant fiscal year, that would step up the basis of distributed assets without giving rise to U.S. taxable income after application of the section 245A dividends received deduction. The IRS issued final regulations in 2020 seeking to curb this avoidance planning. Under those regulations, a portion of a dividend attributable to an extraordinary disposition during the effective date doughnut hole period is ineligible for the section 245A deduction (the ineligible ED amount). Reg. section 1.245A-5(b).

7 For an explanation of how deferral reduced the effective tax rate, see Fleming, Peroni and Shay, supra note 2, at 96-104.

8 See id.

9 See sections 951-965.

11 Id.

12 See section 954(d)(2) and reg. section 1.954-3(b); see also Whirlpool Financial Corp. v. Commissioner, 154 T.C. 142 (2020), aff’d, 19 F.4th 944 (6th Cir. 2021), reh’g denied, Nos. 20-1899 and 20-1900 (6th Cir. 2022), cert. denied, No. 22-9 (U.S. Nov. 21, 2022) (applying branch rule to find foreign base company sales income under section 954(d)(2)).

13 See reg. section 1.954-3(a)(4)(i). The manufacturing exception was dramatically extended in reg. section 1.954-3(a)(4)(i) to take account of activities of contract manufacturers if the CFC made a “substantial contribution” through its employees (including by oversight and direction) to the manufacturing. This rendered an already porous regime close to harmless outside unusual fact patterns.

14 See, e.g., Office of Management and Budget, “Executive Office of the President, Budget of the United States Government, Fiscal Year 2008,” Analytical Perspectives 287 (2007); and JCT, “Estimates of Federal Tax Expenditures for Fiscal Years 2007-2011,” JCS-3-07, at 24 (Sept. 24, 2007).

16 Moline Properties Inc. v. Commissioner, 319 U.S. 436 (1943) (holding that a corporation will be respected as a separate taxpayer if it conducts any business activity or was formed for a business purpose).

17 See id. at 438-439.

18 See section 965. These earnings were taxed at extremely favorable effective rates (after section 965 deductions) of 15.5 percent for earnings held in cash or equivalents and 8 percent for earnings held in other assets. Moreover, the payment could be spread over eight years (back-ended) without interest, so the effective rate was even lower than the apparent 8 percent and 15.5 percent rates. Before the enactment of the TCJA, we expressed our view that transition generosity was inappropriate from a tax policy point of view. See Fleming, Peroni, and Shay, “Getting From Here to There: The Transition Tax Issue,” Tax Notes, Apr. 3, 2017, p. 69. The generosity of the section 965 deduction transformed what had been deferral to partial exemption and without any accounting for previously deducted amounts allocated to foreign income that now became partially exempt. See Shay, “Addressing an Opaque Foreign Income Subsidy With Expense Disallowance,” Tax Notes Federal, Aug. 2, 2021, p. 699 (arguing that under section 265, expenses allocable to foreign income exempted by reason of exemptive deductions should be disallowed); Patrick Driessen, “Getting Foreign Deferral’s Epitaph Right,” Tax Notes Federal, June 15, 2020, p. 1883 (revenue estimate of loss materially understated).

20 See sections 250 and 951A. For a U.S. shareholder that is not a U.S. corporation, the tax on GILTI is at the normal marginal tax rates in section 1 (which has a top marginal tax rate of 37 percent for tax years starting before 2026 and 39.6 percent for tax years starting after 2025). The section 250 deduction is available to a U.S. shareholder who is an individual only if that individual elects to be taxed as a C corporation under section 962. The deductions against GILTI and foreign-derived intangible income under section 250 when aggregated may not exceed taxable income, and if they do, the excess taxable income reduces each of them pro rata based on the amounts of FDII and GILTI. Section 250(a)(2). Throughout, this article uses the term “effective rate” in relation to post-TCJA GILTI and section 245A to reflect the section 250 and section 245A deductions’ reduction of the nominal statutory rate. In many cases, the effective rate of tax measured as the actual tax paid divided by the income of the CFC after allocable U.S. deductions will be different from this simplified calculation.

21 See sections 951(b), 958, and 318.

22 See section 951A(c)(2)(A)(i)(III); and reg. section 1.951A-2(c)(1)(iii) and (c)(7) (expanding the CFC high-tax exception by election to otherwise tested income). Exclusions from tested income also include any dividend received from a related person (section 951A(c)(2)(A)(i)(IV)) and foreign oil and gas extraction income (section 951A(c)(2)(A)(i)(V)), but these income categories are of lesser importance to this article’s analysis. Also, subpart F income is excluded from tested income (section 951A(c)(2)(A)(i)(II)) but is exposed to current taxation at regular rates by section 951(a)(1) (that is, the section 250 deduction does not apply to a subpart F income inclusion by a domestic corporation that is a U.S. shareholder).

24 See sections 951A(f)(1)(A) and 959(a).

26 See section 11 (as mentioned earlier, the section 250 deduction does not apply to a domestic corporation’s subpart F income inclusion). For a U.S. shareholder who is not a U.S. corporation, the U.S. tax on subpart F income is imposed at the regular tax rates in section 1.

28 The section 245A deduction is not available to an individual U.S. shareholder of a CFC even if that individual makes a section 962 election to be taxed as a C corporation. For these shareholders, the NDTIR and the high-foreign-taxed portion of the CFC earnings are included in income when distributed by the CFC. Also, the section 245A deduction does not apply to a foreign corporation that is a passive foreign investment company for the U.S. corporate shareholder and that is not a CFC. See section 245A(b)(2). Earnings that are eligible for the section 245A deduction are taken into account when distributed, as are earnings that are not eligible for the section 245A deduction because they are in excess of the foreign-source portion of the distribution or are within the exception for a non-CFC PFIC. These limited amounts continue to be subject to deferral even for a corporate U.S. shareholder.

29 And, deferral still applies to an individual or domestic corporation that owns stock in a CFC but does not meet the 10 percent stock threshold for U.S. shareholder status.

31 See Fleming, Peroni, and Shay, “Expanded Worldwide Versus Territorial Taxation After the TCJA,” Tax Notes, Dec. 3, 2018, p. 1173 nn. 18-20.

32 The section 245A deduction represents a dividend exemption form of territorial taxation of international income. For an article arguing that territoriality is a tax expenditure, see Driessen, “Would Territoriality Be a Tax Expenditure?Tax Notes, Feb. 2, 2015, p. 647.

33 If the foreign tax in relation to the U.S. tax base is 21 percent or greater, the U.S. FTC will typically eliminate the U.S. residual tax.

34 These tax expenditures are increased to the extent that expense deductions are allowed for expenses allocable to the zero-rate-eligible amount and not partially disallowed for expenses allocable to GILTI taxed at the low 10.5 percent rate. See generally Shay, Reuven S. Avi-Yonah, Driessen, Fleming, and Peroni, “Why R&D Should Be Allocated to Subpart F and GILTI,” Tax Notes Federal, June 23, 2020, p. 2081, 2084.

35 In its most recent tax expenditure report, the JCT scored the GILTI rates as a $295.1 billion tax expenditure for the 2020-2024 period. See JCT, “Estimates of Federal Tax Expenditures for Fiscal Years 2020-2024,” JCX-23-20, at 24 (Nov. 5, 2020).

36 See id.

37 There were 1,477,196 U.S. corporation tax returns filed for 2019, but only 2,891 of the returns that claimed an FTC reported GILTI for 2019. See IRS Statistics of Income, “Corporation Income Tax Returns Complete Report — 2019,” at 9 (July 2022); and IRS SOI, “U.S. Corporation Returns With a Foreign Tax Credit — 2019,” Table 3 (July 2022).

38 A code provision’s classification as a tax expenditure should have no normative implications. The only consequences of that classification are to quantify (as part of the budgeting process) the cost of the code provision in terms of forgone revenue and to subject the code provision to the same cost-benefit analysis that should be applied to direct expenditures. See Fleming and Peroni, supra note 4, at 444-445.

39 An NDTIR greater than zero will result in a U.S. effective tax rate of less than 10.5 percent. In this example, for reasons discussed later, we assume that the section 55(b) corporate alternative minimum tax does not apply and that neither Newlandia nor another country in which USCo has an affiliate implemented a pillar 2 qualified domestic top-up tax or undertaxed profits tax. See OECD, “Tax Challenges Arising From the Digitalisation of the Economy — Global Anti-Base Erosion Model Rules (Pillar Two)” (2021).

40 See Office of the U.S. Trade Representative, “Section 301 Investigation, Report on France’s Digital Services Tax” (Dec. 2, 2019) (reporting testimony by representatives of U.S. multinational businesses and market evidence of the success of U.S. businesses in the areas of digital advertising and digital platforms). The same companies that the Office of the U.S. Trade Representative argues are discriminated against by the French digital services tax because of their market successes (Google (Alphabet), Apple, Facebook (Meta), and Amazon) are among the largest beneficiaries of the GILTI tax expenditure. For a critical analysis of the Office of the U.S. Trade Representative’s discussion of international tax norms in the report on France’s DST, see Shay, “Trade Enforcement Tools and International Taxation: A Digital Services Tax Case Study,” in Elgar Companion to WTO (forthcoming).

41 See Fleming and Peroni, supra note 4, at 535; and Kimberly A. Clausing, testimony before the Senate Finance Committee, at 3 (Oct. 3, 2017) (“In terms of the ability to generate after-tax profits and market dominance, U.S. multinational companies are already quite competitive.”); see also IMF, OECD, United Nations, and the World Bank Group, “Options for Low Income Countries’ Effective and Efficient Use of Tax Incentives for Investment,” 11-12 (2015) (10 out of 14 surveys of investors in low-income countries show tax incentives are redundant — that is, unnecessary to attract the investment — in 70 percent or more of cases).

42 “Broader notions of competitiveness emphasize the fundamentals that determine the health and well-being of our broader economy. Are workers well-educated, and do they have the skills required to earn wages in the global economy?” Clausing, supra note 41. “Human capital theory is the now widely accepted idea that education, training, and other forms of learning are investments that pay off in the future. . . . [T]he technology for producing foundational skills such as a numeracy is well understood, and resources are the main constraint.” David J. Deming, “Four Facts About Human Capital,” 36 J. Econ. Persp. 75-76 (Summer 2022). “Empirical studies show that the growth in income inequality is largely due to differences in educational attainment.” Joann M. Weiner, “Conversations: Harvey S. Rosen,” Tax Notes, Nov. 26, 2007, p. 857, at 859.

43 See JCT, supra note 35.

44 U.S. Department of Education, Institute of Education Sciences, “Revenues and Expenditures for Public Elementary and Secondary Education: FY 19,” at 2 (June 2021).

45 JCT, supra note 35, at 24. We recognize that tax expenditure estimates are not equivalent to revenue estimates, and we use this example merely to illustrate scale.

46 Further, as we have discussed in prior articles, we would recommend revising section 904 to incorporate a per-country FTC limitation (in place of the basket limitation system) and revising section 960(d) to eliminate the 20 percent indirect FTC haircut for foreign taxes properly attributable to tested income under section 951A.

47 Sections 55(b)(2) and 59(k).

48 Section 55(a). The new corporate AMT applies to tax years beginning after 2022.

50 See Martin A. Sullivan, “Tax Credits and Depreciation Relief Slash Burden of New Corporate AMT,” Tax Notes Federal, Aug. 22, 2022, p. 1185, at 1186; and Sullivan, “Identifying Corporations Likely to Pay the New Corporate AMT,” Tax Notes Federal, Aug. 22, 2022, p. 895.

51 See IRS SOI citations, supra note 37.

52 See id.

53 For 2019, 1,140 U.S. companies with revenue of $850 million or above filed a Form 8975, “Country-by-Country Report.” IRS SOI, “SOI Tax Stats — Country-by-Country Report,” Table 1B Country-by-Country Report (Form 8975): Tax Jurisdiction Information (Schedule A: Part I) Limited to Reporting Entities With Positive Profit Before Income Tax by Major Geographic Region and Selected Tax Jurisdiction, Tax Year 2019.

54 Their congressional allies actively and directly encouraged Hungary to veto pillar 2 in the EU. See letter from two Republican subcommittee leaders on the House Ways and Means Committee to Hungarian Ambassador Szabolcs Takacs (June 20, 2022) (thanking Hungary for opposing pillar 2); and Jeff Stein, “GOP Officials Back Hungary’s Resistance to Global Tax Deal, Bucking Biden,” The Washington Post, July 1, 2022. See also Richard Rubin, “Sen. Joe Manchin Balks at Global Minimum Tax Championed by Biden,” The Wall Street Journal, July 6, 2022 (“‘We’re not going to go down that path overseas right now, because the rest of the countries won’t follow,’ Mr. Manchin told the radio host Hoppy Kercheval. ‘And we’ll put all of our international companies in jeopardy, which harm the American economy. Can’t do that. So we took that off the table.’”).

55 Laurence Norman, “France, Germany and Other EU Countries Pledge to Advance Minimum Tax,” The Wall Street Journal, Sept. 9, 2022 (including France, Germany, Italy, the Netherlands, and Spain); Sarah Paez, “Switzerland Seeks Input on Temporary Pillar 2 Regulation,” Tax Notes Int’l, Aug. 22, 2022, p. 953; HM Treasury, OECD Pillar 2 Consultation (Jan. 2, 2022).


Copy RID