Coordinating Pillar 2 With the U.S. GILTI Regime

I. Executive Summary

The OECD and G-20 inclusive framework on base erosion and profit shifting continues its work on a global anti-base erosion (GLOBE) proposal under pillar 2 to address the tax challenges arising from the digital economy. Pillar 2 is expected to consist of four components, which together are intended to ensure that large internationally operating businesses pay a minimum level of tax regardless of where they are headquartered or operate. Those components are:

• an entity-based income inclusion rule (IIR) applied to subsidiaries and branches treated as entities;

• an entity-based switchover rule applied to branches subject to a tax treaty;

• a payment-based undertaxed payments rule (UTPR) applied to deductible payments made to related entities not subject to a tax treaty; and

• a payment-based subject-to-tax rule applied to deductible payments made to related entities subject to a tax treaty.

While many readers may be familiar with the U.S. version of an IIR (the U.S. global intangible low-taxed income regime adopted in 2017), the anticipated pillar 2 framework is expected to contain several important elements that might put the GILTI regime at odds with jurisdictions implementing pillar 2. The question then arises of how the U.S. GILTI regime will interact with pillar 2 components on their implementation. This paper explores various options available to policymakers to coordinate other jurisdictions’ potential pillar 2 rules with the GILTI regime. Those options include grandfathering, treating the U.S. GILTI regime as a pillar-2-compliant IIR, and other coordination methods that can be implemented unilaterally or multilaterally.

Grandfathering, a relatively administrable and transparent approach, would coordinate pillar 2 rules with the U.S. GILTI regime, while acknowledging that the GILTI regime predates pillar 2 and thus necessarily will differ from the pillar 2 IIR. The appropriateness of grandfathering the U.S. GILTI regime ultimately depends on whether it achieves the policy objectives of pillar 2. As discussed below, GILTI is intended to subject to a reduced rate of U.S. tax foreign earnings that otherwise are subject to no or little tax, similar to the purpose of the GLOBE proposal to ensure that the profits of internationally operating businesses are subject to a minimum rate of tax. The similarity between some design choices made for the U.S. GILTI regime and for the pillar 2 IIR further indicate how closely the GILTI regime achieves the policy objectives of pillar 2. In fact, in many cases, the U.S. GILTI regime could be viewed as even more onerous than the pillar 2 IIR.

Absent grandfathering, the GILTI regime also could be coordinated if the pillar 2 criteria for determining a pillar-2-compliant IIR are sufficiently flexible to encompass GILTI. As a result of the inevitable variations among countries’ IIRs, the governing standard for pillar 2 compliance should be substantial similarity, not identicality. Substantial similarity generally may be determined in three ways: a quantitative analysis, a qualitative analysis, and a list of compliant regimes. A quantitative analysis requiring a hypothetical tax liability to be computed as if the IIR of the yielding jurisdiction applied would seem extremely burdensome. Other approaches that evaluate a particular rule’s compatibility on a qualitative basis that does not require evaluating the facts and circumstances of a particular taxpayer would be more administrable. Establishing (and potentially periodically reevaluating) a list of compliant jurisdictions also would require a qualitative analysis. In this context, whether GILTI is compliant with a benchmark pillar 2 IIR would be determined by analyzing its design features vis-à-vis the benchmark rule.

Ideally, the rules for determining whether an IIR regime is compliant with pillar 2 should take into account several factors, including efficacy, simplicity, administrability, transparency, and minimization of double taxation. Each design aspect of an IIR is expected to affect the IIR’s overall reach and extent. Some of the variations in IIRs arising from those design aspects will be insubstantial and thus should be given little weight when coordinating multiple pillar 2 regimes. Other variations, although substantial, can be addressed by treating the IIR as partially compliant and limiting application of other countries’ pillar 2 regimes to the noncompliant part.

For example, differences in base, rate, blending, measures addressing time-varying volatility, and relevant foreign taxes frequently should be considered insubstantial because, although they alter the operational impact of an IIR, they do not alter the policy objectives of the IIR or its capacity to address those objectives. By contrast, differences in scope and carveouts may be substantial but likely can be readily isolated so that the pillar 2 regimes of other jurisdictions would be applied only on a limited basis — that is, to the noncompliant part.

In comparison to the pillar 2 IIR, the U.S. GILTI regime’s scope, rate, measures addressing time-varying volatility, and relevant foreign taxes in many cases will be broader or more onerous. On the other hand, the regime’s base, blending, and carveouts could be viewed as less onerous than those under the pillar 2 IIR, but they are unlikely to be deemed substantial because they do not cause the regime’s policy objectives, or its achievement of those objectives, to differ from the objectives of pillar 2. Even if substantial differences were found to exist, most could be readily isolated and thus should require only a limitation on the application of other countries’ pillar 2 rules, rather than treating GILTI as noncompliant. Deeming GILTI partially compliant would also affect the extent to which other components of pillar 2, such as the undertaxed payments and subject to tax rules, find application.

Finally, the complexity of the U.S. GILTI regime and its layering on top of the intricate U.S. international tax system requires a holistic approach to evaluate whether and to what extent the regime should be considered pillar-2-compliant and the consequences of that determination. Importantly, measuring the U.S. tax imposed because of GILTI will require taking into account the particular aspects of the U.S. international tax system, including limitations on the U.S. foreign tax credit (in addition to the 20 percent reduction in creditable foreign taxes) that could result in additional U.S. tax costs even when the foreign effective rate equals or exceeds the GILTI rate threshold (currently, 13.125 percent).

If action to grandfather GILTI or treat it as a pillar-2-compliant IIR is not taken on a multilateral or unilateral level, coordination rules that apply to any other jurisdiction’s IIR would be expected to equally apply to the U.S. GILTI regime. The last section of this article discusses potential coordination methods and attempts to illustrate the implications of ordering rules and their absence. Also, because variations in the domestic implementation of the pillar 2 rules are expected, it also attempts to analyze those essential IIR design choices available to policymakers and systematically evaluate their implications.

II. Introduction

The development of pillar 2 is reportedly moving at a fast pace (or at least more quickly than that of pillar 1),¹ yet, in the words of the inclusive framework, “significant work still remains.”² The design features of pillar 2 that are to be settled include the crucial question of coordinating different jurisdictions’ pillar 2 rules.

Pillar 2 consists of multiple rules requiring coordination between them. The IIR is expected to take priority over the UTPR, potentially limiting the scope of the UTPR’s application to cases in which a compliant IIR does not apply. Hence, whether a particular IIR is in compliance with pillar 2 becomes a crucial matter.³

The U.S. version of an IIR — that is, the U.S. GILTI regime — could be odds with jurisdictions implementing pillar 2. The question then arises of how the U.S. GILTI regime will interact with the components of pillar 2 on implementation. Multiple stakeholders have requested coordination rules specific to the GILTI rules.⁴ The U.S. government has already expressed its preference for a GILTI-like pillar 2 solution and recently reiterated its support for the pillar 2 project.⁵

This article explores the various options available to policymakers to coordinate the U.S. GILTI regime with other jurisdictions’ potential pillar 2 rules, including grandfathering and treating the GILTI regime as a pillar-2-compliant IIR. It also explores a framework for variations in domestic implementations of pillar 2 and considerations for coordinating those implementations with the GILTI regime.

III. Overview

In early 2019 the OECD released a policy note⁶ and public consultation document on addressing the challenges arising from the digital economy,⁷ followed by a program of work in May⁸ and a pillar 2 consultation document⁹ in November.

In essence, the pillar 2 proposal consists of two main rules: an IIR and a base-eroding payments tax (later labeled a UTPR). The IIR would function as a foreign minimum tax, ensuring the income of a foreign branch or controlled entity is subject to a minimum rate of tax. The UTPR would deny a deduction or impose source-based taxation for payments to related parties that were subject to tax below a minimum rate.¹⁰ Those two rules are complemented by two treaty-based rules: a switchover rule that generally would permit a residence jurisdiction to switch from an exemption to a credit method when the profits attributable to a permanent establishment are subject to an effective rate below the minimum rate, and a subject-to-tax rule that would restrict some treaty benefits to only when an item of income is taxed at a minimum rate.

The pillar 2 IIR is expected to adopt either a jurisdictional or global blending approach.¹¹ A top-down approach also is expected to be adopted under which the IIR applies to the ultimate parent if the ultimate parent’s jurisdiction has an IIR. If the ultimate parent jurisdiction does not have an IIR, the pillar 2 IIR will apply at the level of entities that are immediately below the ultimate parent in the ownership structure.¹² Moreover, the pillar 2 IIR tax base is expected to be computed by using financial accounts with some adjustments.¹³

Because the pillar 2 rules would be implemented in domestic laws and tax treaties, the inclusive framework has said pillar 2 should incorporate a coordination or ordering rule to avoid the risk of double taxation. It has also said it is working on coordination aspects, along with designing the priority of pillar 2 rules and their interaction with other rules, including existing BEPS measures. Recent reports suggest that significant progress has been made and that the IIR will take priority over the UTPR.¹⁴

In the meantime, countries have begun introducing domestic rules inspired by pillar 2, signaling the urgency of the need for coordination.¹⁵ The GILTI regime was introduced with major U.S. tax reform in 2017, so U.S. corporations already have been navigating a type of IIR for more than two years.

Whether the U.S. GILTI regime, which influenced the starting point for the IIR,¹⁶ and other countries’ regimes are deemed compliant with pillar 2 will depend on the design choices made. It is anticipated that the GILTI regime may be considered an acceptable regime and included in a compliant regimes list (often referred to as a “whitelist”).¹⁷ Even so, the extent and mechanics of the interaction between GILTI and pillar 2 are unknown.

For jurisdictions without an IIR or UTPR, the pillar 2 components are expected to be introduced via domestic legislation. Given the broad policy options available, it seems important to categorically analyze design features that must exist for a regime to be considered compliant with a multilateral foreign minimum tax.

The OECD is expected to develop model legislation for jurisdictions to use as the basis for domestic legislation, and it might also explore a multilateral convention.¹⁸ If multilateral agreement is reached, because of differences in domestic implementation and existing tax regimes, no two sets of pillar 2 rules will be identical even under a prescriptive agreement. If a multilateral agreement is not reached, the inclusive framework is expected to issue detailed recommendations on how to implement pillar 2 rules unilaterally. Jurisdictions that then introduce those rules supposedly would have more design options, so greater differences could arise. The inclusive framework guidance could include a framework for evaluating pillar 2 compliance, including a list of compliant regimes, to facilitate unilateral coordination of pillar 2 regimes based on multilateral consensus, and grandfather the U.S. GILTI regime.

IV. Coordinating Pillar 2 and GILTI

GILTI predates the inclusive framework’s work on pillar 2, so many of its design choices may be decidedly different than those in pillar 2. As discussed below, some of those choices will generally make the GILTI regime more onerous than the pillar 2 IIR, both in terms of administrability and substantive tax liability, while others will generally make it less onerous. And others still will have mixed results depending on the taxpayer’s facts and circumstances. But none of those choices materially distinguishes the policy objectives of the U.S. GILTI regime from those of the pillar 2 IIR. Accordingly, pillar 2 rules should coordinate with the GILTI regime in a manner similar to how the pillar 2 rules coordinate with other pillar 2 IIRs.

A. Grandfathering GILTI

One method for coordinating pillar 2 rules with the U.S. regime is to grandfather GILTI. Grandfathering would treat any multinational group subject to the GILTI regime at its parent level as exempt from further application of pillar 2 rules. Importantly, grandfathering should not require a determination of whether and to what extent the GILTI regime complies with pillar 2 because it acknowledges that GILTI predates the inclusive framework’s work on pillar 2 and does not need to be amended to achieve the pillar’s policy objectives. In short, under grandfathering, U.S. multinationals would not be subject to other pillar 2 rules.¹⁹

The appropriateness of grandfathering the U.S. GILTI regime ultimately depends on whether the regime achieves the pillar 2 policy objectives. There may be some differences in opinion over those objectives, but the pillar 2 consultation document frames the GLOBE proposal as “ensuring that the profits of internationally operating businesses are subject to a minimum rate of tax.” The GILTI regime has a similar aim: to subject to a reduced rate of U.S. tax foreign earnings that are otherwise subject to little or no tax.²⁰

The similarity between the design choices made for the U.S. GILTI regime and the pillar 2 IIR might indicate how closely the GILTI regime achieves the policy objectives of pillar 2. In that respect, some aspects of GILTI are likely more onerous than the pillar 2 IIR will be. For example, the GILTI regime does not have a size-based threshold for application, can apply to subsidiaries not controlled by a particular U.S. shareholder, does not include measures to address time-based volatility or timing differences, and applies to the full extent of most subsidiary income (rather than as a top-up tax) with limitations on available foreign tax credits.²¹

Further, the GILTI regime’s base is determined under U.S. tax principles, rather than a simpler but less precise measure such as financial accounts, and, despite employing global blending, it necessitates a rigorous calculation of income on a per-entity basis. Also, the GILTI regime does not include any particular simplification measure similar to those reportedly considered for the design of pillar 2, such as safe harbors for high-tax jurisdictions or de minimis profit exclusion mechanism.²² Also, any U.S. multinational enterprise or foreign parented MNE involving an intermediary U.S. corporation is required to calculate its GILTI liability each year regardless of its historical effective tax rate. In other words, the GILTI rules do not include a mechanism that would quickly eliminate the compliance burden with the GILTI rules under certain circumstances.

Other aspects may be different but less consequential, such as the GILTI regime containing a substance-based carveout measured using tangible asset basis rather than some other formulaic, substance-based carveout.

At least one aspect of GILTI — its overall blending — may be less onerous than the pillar 2 IIR, if the inclusive framework ultimately adopts jurisdictional blending.

A detailed analysis of those choices should not be necessary for grandfathering,²³ but because most aspects of the U.S. GILTI regime are likely similar to or more onerous than the pillar 2 IIR, the regime effectively achieves the pillar’s policy objectives.

As mentioned, grandfathering the U.S. GILTI regime is based in part on the regime predating pillar 2. As a result, subsequent changes to the GILTI regime could call grandfathering into question. When subsequent changes do not materially alter the GILTI regime’s efficacy in achieving the policy objectives of pillar 2, however, there would not appear to be a compelling case for denying grandfathering. For example, there have been proposals to modify the GILTI regime by raising the effective rate threshold, limiting blending, or eliminating the substance-based carveout. But those proposals would only make the regime more onerous and would not limit its ability to ensure that the profits of internationally operating businesses are subject to a minimum rate of tax.

Grandfathering coordinates pillar 2 rules with the U.S. GILTI regime while acknowledging that GILTI predates pillar 2 and thus necessarily will differ from the pillar 2 IIR. It is also relatively administrable for U.S. taxpayers and tax administrators (who have been working with the GILTI regime for multiple years), non-U.S. tax administrators (who generally would not need to apply new pillar 2 rules to subsidiaries of U.S. multinational entities), and the inclusive framework (by obviating the need to develop a framework for pillar 2 compliance that accounts for design variations between GILTI and the IIR, which could be very burdensome).

The lessons learned from the relatively recent experience with the U.S. GILTI regime — a regime that not only predates pillar 2 discussions, but also is the starting point of the IIR in pillar 2 — should inform policymakers so that appropriate pillar 2 design choices are made. As discussed below, some policymakers want to deviate from some aspects of the U.S. GILTI regime, so it is conceivable that the pillar 2 IIR will differ from GILTI.

Further, grandfathering the U.S. GILTI regime arguably would be a transparent approach because it would make clear the particular treatment provided to the only modern IIR (at least known to the authors) in place before pillar 2 discussions began. If, for example, possible future law changes are a concern, grandfathering could be subject to periodic evaluation, similar to the mechanism for evaluating jurisdictions deemed compliant with pillar 2. Even so, as discussed below, grandfathering is not the only method for coordinating pillar 2 rules with the GILTI regime.

B. Treating GILTI as a Pillar-2-Compliant IIR

Another method for coordinating pillar 2 rules with the U.S. GILTI regime is to treat the GILTI regime as a pillar-2-compliant IIR. As discussed below, the standard for pillar 2 compliance should be substantial similarity to, not identicality with, a benchmark. Consequently, determining whether and to what extent GILTI, like any other IIR, is a pillar-2-compliant regime requires understanding whether it is substantially similar to a benchmark IIR. Substantial similarity depends on how a given regime addresses or disregards the various design aspects.

1. Standard of Compliance

As discussed above, it would be prudent to expect that some level of variation will always exist among IIRs introduced by different jurisdictions.²⁴ That is so even when a great level of detail is decided by multilateral action. Vast variations in existing taxation regimes and practices around the globe would inevitably come into play. Moreover, tax administration practices greatly differ from jurisdiction to jurisdiction, potentially resulting in different interpretations and administration of a common rule. Thus, in the absence of a universal rulemaker and administrator, it is unrealistic to search for identicality among IIRs. Instead, in determining whether one jurisdiction would yield to another’s rule, substantial similarity arguably should be the standard.

a. Substantial Similarity

The extent to which a rule is substantially similar to a benchmark rule would intrinsically depend on whether the tested rule reaches the benchmark rule’s underlying policy objectives. For that reason alone, it is critical that the objectives of pillar 2 are clearly described and agreed on. Indeed, the lack of a clear expression of the pillar’s underlying goals has been noted by commentators.²⁵

Simplicity and administrability would also be important factors. Unsurprisingly, the inclusive framework, taxpayers, and tax practitioners have emphasized those aspects, which are also critical from a tax administration perspective.²⁶ Evaluating a rule’s compatibility with pillar 2 would require the different capacities of tax administrations worldwide, particularly regarding international taxation.²⁷ Considering design simplicity would mean that some level of precision — material or immaterial — would need to be forsaken.

Finally, using substantial similarity, rather than identicality, as a standard would allow flexibility in implementing pillar 2 regimes and variability among those regimes. That flexibility would need to be balanced against the needs for achieving the pillar 2 policy objectives and creating sufficiently simple and administrable rules.

Substantial similarity could be determined in at least three ways: a hypothetical computation, a qualitative analysis, or a list of compliant regimes.

First, a hypothetical computation of a foreign minimum tax under the IIR of the yielding jurisdiction could be required. Tax liability would be computed as if the IIR of that jurisdiction applied. If the liability under the applying jurisdiction is close enough to the hypothetical tax on the income inclusion of the yielding jurisdiction (for example, not less than 90 percent of the liability), it would be deemed substantially similar. That method would require computations under separate sets of rules and would likely be extremely burdensome.

Second, a qualitative analysis of design considerations could be made by taking into account the relative importance of the rule’s various features. As discussed below, the design choices available to policymakers yield different outcomes. While some aspects of a rule naturally carry more weight, such as the rate, base, and blending, determining the rule’s true impact would often require evaluating all its parts instead of analyzing various components in isolation. Importantly, that approach would not require evaluating a particular taxpayer’s facts and circumstances or computing hypothetical tax liabilities. Thus, it would promote simplicity and administrability at the cost of some precision (although perhaps no adverse impact on achieving the policy objectives of pillar 2, if the loss of precision falls in an acceptable range).

Third, a list of compliant jurisdictions with a substantially similar IIR could be developed. That would require a survey comparing each jurisdiction’s IIR with the benchmark rule. The list could be developed multilaterally or unilaterally, although multilateral consensus is greatly preferable. Further action would need to be taken to address subsequent amendments in the laws.²⁸ It has been suggested that a working group similar to the Global Forum on Transparency and Exchange of Information for Tax Purposes could monitor international tax developments and keep the list up to date.²⁹ The process for developing a whitelist would be similar to the qualitative analysis described above, but it would be undertaken on an abstract basis by a particular tax authority or international body rather than a particularized basis by a given taxpayer. Doing it once on a multilateral and broadly applicable basis would promote simplicity and administrability, as well as introduce additional transparency, without sacrificing furtherance of the policy objectives of pillar 2. Indeed, pillar 2 might adopt a multilateral review process through which compliant IIRs are determined.³⁰

b. GILTI Design Aspects

Several design aspects of the U.S. GILTI regime could affect the substantial similarity determination.

i. Scope

Any U.S. shareholder is required to include GILTI in gross income.³¹ There are two ownership thresholds relevant for a GILTI calculation. First, only the income of a foreign corporation owned more than 50 percent by vote or value by U.S. persons that are 10 percent shareholders (a controlled foreign corporation) is subject to GILTI.³² Second, for a U.S. person to have a GILTI inclusion of a CFC’s income, that person itself must be a 10 percent shareholder — that is, it must own at least 10 percent of the vote or value of the CFC, and thus be a U.S. shareholder.³³

Accordingly, in a multinational group, if a U.S. corporation meets those ownership thresholds for a foreign corporation, regardless of whether it is the ultimate or an immediate shareholder, a GILTI inclusion may be required.

The GILTI rules do not include a group size threshold. As noted above, if pillar 2 adopts a substantially different relatedness threshold, the U.S. GILTI regime may be substantially different from other IIRs. But that difference likely can be isolated and addressed by applying other pillar 2 rules to subsidiaries that fall outside the scope of the GILTI regime. Further, if, for example, in pillar 2, the scope of covered entities is determined by consolidated financial statements and only entities that are consolidated on a line-by-line basis are within scope, the GILTI regime may be considered more comprehensive for some entities.

ii. Base

In many respects, the GILTI regime complements the U.S. CFC regime — that is, the subpart F regime — but it also departs from it in some important ways.³⁴ The tax base for the GILTI regime is calculated under detailed regulations promulgated under IRC section 951A. Most important, while the GILTI inclusion amount is calculated at the U.S. shareholder level, some items are initially determined at the CFC level and then aggregated for the U.S. shareholder. Also, a reallocation of the GILTI inclusion amount to CFCs is required for multiple reasons.³⁵

Accordingly, as a first step under GILTI, the income of each CFC must be calculated for the year of inclusion. For this purpose, a CFC’s tested income generally consists of the CFC’s gross income reduced by properly allocable deductions and taxes.³⁶ Specific items are excluded from tested income, including any income effectively connected with the conduct of a U.S. trade or business, any amount includable in shareholders’ gross income under the subpart F regime, intragroup dividends, and any foreign oil and gas extraction income.³⁷ Further, some income that is subject to high taxes (at least an 18.9 percent rate) in a foreign jurisdiction may, on election, be excluded from tested income.³⁸ If the tested income computation yields a loss (a negative number), the CFC has a tested loss for that inclusion year.³⁹

After determining each CFC’s tested loss and income, the U.S. shareholder’s pro rata share of the tested income and loss for each CFC is aggregated. The excess amount of aggregate tested income over aggregate tested loss constitutes the shareholder’s net CFC tested income.⁴⁰ The tax base for GILTI is that net CFC tested income amount reduced by the exemption amount for a routine return on tangible assets based on a formulaic calculation.⁴¹ The routine return exemption is generally equal to a 10 percent return on the tangible and depreciable property used outside the United States — that is, each CFC’s aggregate adjusted basis in that property⁴² — and reduced by the CFCs’ aggregate net interest expense.⁴³

In the context of U.S. consolidated groups, each member of a group that is a U.S. shareholder of a CFC must include its GILTI amount in income, and the rules generally follow the computation principles described above.⁴⁴ For each member, the aggregate tested income is the aggregate of the member’s pro rata share of the tested income of each CFC.⁴⁵ Computation of other tested items requires aggregation at the consolidated level followed by reallocation. For example, the aggregate tested loss of each member is further aggregated, resulting in the consolidated tested loss that is then reallocated to each member according to the ratio of that member’s aggregate tested income to the consolidated tested income.⁴⁶ Similar mechanisms are used for determining consolidated amounts for exemption for routine return amounts and interest expense.⁴⁷

The U.S. GILTI regime essentially adopts the home country’s tax laws for measuring its base. That is consistent with the policy objective of pillar 2 to ensure that an MNE’s global operations carried on through foreign subsidiaries are subject to a minimum level of taxation.

iii. Rate

The GILTI amount is included in the gross income of the U.S. shareholder, and therefore, in principle for corporate shareholders, is subject to the general corporate income tax rate.⁴⁸ There is, however, a 50 percent deduction for the GILTI inclusion amount, resulting in a 10.5 percent overall rate for GILTI inclusions.⁴⁹ Also, the GILTI regime allows a credit for foreign taxes paid or accrued for tested income but limits it to 80 percent of the relevant foreign income taxes.⁵⁰ As a result of that 20 percent haircut, under the GILTI rules, the effective tax rate (ETR) threshold increases to 13.125 percent. In other words, foreign income that is subject to a rate lower than 13.125 percent outside the United States is subject to residual U.S. tax under the GILTI regime. As discussed below, however, additional limitations on the U.S. FTC may result in additional U.S. tax costs, even when the foreign effective rate equals or exceeds 13.125 percent. It is unknown what minimum rate will produce multilateral consensus and whether the GILTI threshold will meet or exceed it.

The 50 percent deduction on the GILTI amount will decrease to 37.5 percent for tax years beginning after December 31, 2025.⁵¹ Accordingly, assuming the corporate rate is still 21 percent at the time of that scheduled change, the overall tax rate on GILTI inclusions will increase to 13.125 percent, while the ETR threshold will increase to approximately 16.4 percent.

iv. Blending

Because the tested income of each CFC is aggregated for the tax base regardless of the location of the CFC or where the income arises, the GILTI regime uses a global blending approach. The impact of that blending option on pillar 2 coordination is explored below.

v. Carveouts

As noted, the GILTI regime includes several carveouts. First, it provides a general exemption for routine returns on tangible assets. Second, any income taken into account for subpart F purposes, any intragroup dividends, and any foreign oil and gas extraction income are out of the scope. If the taxpayer elects to exclude eligible high-taxed tested income, that income is also excluded from the scope. Except for oil and gas extraction income, there are no industry-specific carveouts from the application of the GILTI rules. The effects of those carveouts on pillar 2 coordination are discussed below.

vi. Time-Varying Tax Base Volatility

The GILTI regime does not include any specific measures to address time-varying tax base volatility,⁵² although determining its base under U.S. tax principles incorporates some provisions relevant to time-varying tax base volatility.⁵³ That is inconsistent with other aspects of the U.S. international tax system⁵⁴ and with the direction of the pillar 2 IIR. Pillar 2 is expected to include measures addressing time-varying tax base volatility, including an effective credit mechanism for excess IIR taxes paid.⁵⁵ The lack of those measures makes GILTI more onerous than other IIRs, including that in pillar 2, but it does not appear to make the GILTI regime less effective at achieving the policy objectives of pillar 2. Consequently, the GILTI regime’s lack of specific measures to address time-varying tax base volatility should not render it substantially dissimilar from a benchmark IIR.

vii. Relevant Foreign Taxes

The GILTI regime does not require an ETR threshold calculation — that is, there is not an initial step to determine whether the GILTI regime applies.⁵⁶ Instead, the GILTI regime applies if conditions are met and provides for limited crediting for foreign income taxes paid or accrued that are properly attributable to the relevant tested income.

Only income and similar taxes imposed on income included in the GILTI tax base are taken into account for foreign tax crediting purposes. Once those tested foreign income taxes of each CFC are determined, they are aggregated and adjusted for the 80 percent limitation, tested losses, and the routine return exemption for crediting purposes.⁵⁷

For allocating and apportioning foreign income taxes, while the foreign tax base is calculated under foreign law, U.S tax principles determine the characterization of income included in the foreign tax base. Foreign taxes are allocated to an item of income if the income is included in the base on which the foreign income tax is imposed. A foreign withholding tax is allocated to the income from which it is withheld. If a tax is allocated to more than one separate category, including GILTI, the tax is apportioned among those categories.

U.S. tax principles include detailed rules for determining which foreign taxes are creditable. Particularly, credits are granted only for foreign income taxes and taxes that are in lieu of income, war profits, or excess profits taxes and generally imposed by a foreign country. The foreign tax must be a compulsory payment in accordance with the authority of a foreign country to levy taxes, have the predominant character of an income tax in the U.S. sense, and be considered paid or accrued. If a foreign tax depends on the availability of a credit against income tax liability to another country — that is, a soak-up tax — it is not an income tax in the U.S. sense.

The GILTI regime does not allow for the carryback and carryforward of excess credits. Thus, a taxpayer loses the ability to use any foreign taxes in excess of the limitation in a given year. That feature has been criticized for failing to address time-varying tax base volatility. Coupled with the manner in which timing differences arising from foreign tax redeterminations are addressed (an additional tax liability on foreign tax redetermination relates back to the year in which tax accrued), the unavailability of a carryback or carryforward mechanism may have a major impact on some taxpayers.⁵⁸

2. Scope of Compliance

a. In General

In addition to determining a standard for evaluating whether an IIR is pillar-2-compliant the extent to which a jurisdiction will yield application of its pillar 2 rules to entities or items of income subject to the compliant IIR — that is, the controlling jurisdiction’s IIR — must be determined. For example, if a pillar 2 regime (the yielding regime, or the jurisdiction that gives priority to the application of another jurisdiction’s IIR) yields to another that is pillar-2-compliant (the controlling regime, or the jurisdiction with priority to apply its IIR), but the controlling regime applies only to certain subsidiaries of a shareholder and not to others, it may be desirable for the yielding regime to apply to the subsidiaries not subject to the controlling regime. The same may be true if the controlling regime applies to a smaller portion of a subsidiary’s income than would the yielding regime, or if the controlling regime does not apply to every item of income to which the yielding regime would apply.

Determining the extent to which a regime yields to a controlling regime must take into account the purposes of yielding in the first instance: to promote simplicity, administrability, and transparency; minimize double taxation; and preserve the policy objectives of pillar 2. Taken to an extreme, a regime could yield to a controlling regime only if the controlling regime applies in the same manner and to the same extent as the yielding regime. In that case the yielding regime’s jurisdiction effectively would have embraced the identicality standard for determining pillar 2 compliance, which would entail a credit approach for coordinating pillar 2 regimes.⁵⁹ Thus, an exemption approach to coordination requires that there be instances in which a yielding regime allows a controlling regime not to apply to the same extent that it would apply.

The most effective ways to coordinate pillar 2 regimes involve one jurisdiction fully yielding to another. Thus, a yielding regime should yield to a controlling regime whenever possible. Consequently, if the standard for yielding is substantial similarity and variation is insubstantial, then the yielding regime arguably should not apply at all, including to the variation. If the variation is substantial and the yielding regime can be readily applied to the variation in isolation, arguably the yielding regime should be deemed as partially compliant and the application of the yielding regime may be limited to the noncompliant part. If the variation is substantial and the potential yielding regime cannot be readily applied to the variation in isolation, then, as discussed above, the potential controlling regime might not be substantially similar to the potential yielding regime, which may apply in its entirety.

b. Non-U.S. Shareholders

Perhaps the most obvious circumstance in which a regime should not yield to the U.S. GILTI regime is if an item of income is not included in a U.S. shareholder’s pro rata share of that item. For example, if a CFC is 60 percent owned by a U.S. shareholder and 40 percent owned by unrelated non-U.S. shareholders, presumably only 60 percent of the CFC’s income would be treated as subject to a pillar-2-compliant regime by reason of the application of the GILTI regime. The remaining portion might not be treated as subject to a pillar-2-compliant regime, although that portion can be readily isolated and addressed by other pillar 2 rules because it is a defined portion of each item of income. At a specific ownership threshold (such as 80 percent, 90 percent, or perhaps even higher), the difference may be viewed as so insubstantial as to warrant exempting all of the CFC’s income from further application of pillar 2 rules.

c. Treatment of Income Carveouts

As discussed above, exclusions from the GILTI tax base include income includable in a U.S. shareholder’s gross income under the U.S. CFC rules (that is, subpart F income), income subject to U.S. tax as effectively connected with the conduct of a U.S. trade or business, foreign oil and gas extraction income, and high-taxed subpart F income and tested income that are excluded via election. A yielding regime could treat that income as subject to a pillar-2-compliant IIR because the GILTI regime excludes it from tested income; alternatively, a yielding regime could consider that excluded income as not subject to a pillar-2-compliant IIR because the GILTI regime does not apply to it. The approach taken potentially would determine which income, if any, will be subject to another jurisdiction’s IIR or UTPR.

Determining whether income carved out of the U.S. GILTI regime should be considered subject to a pillar-2-compliant regime is closely related to the level of taxation to which the income is subject. A CFC’s income that has been taken into account under the subpart F regime or treated as effectively connected with the conduct of a U.S. trade or business is subject to the full corporate tax rate (21 percent), which is significantly higher than the effective rate for GILTI inclusions. The legislative history indicates that foreign oil and gas extraction income was excluded from GILTI because it is generally immobile and subject to high levels of foreign tax.⁶⁰ High-taxed income that is excluded from GILTI by way of election is subject to a rate of at least 18.9 percent. For purposes of the election to exclude high-taxed income, the GILTI rules require testing whether income is high taxed at the level of tested units — that is, units more granular than CFCs — and thus provide more restricted blending opportunities.⁶¹ Also, while intragroup dividends were excluded from GILTI to help implement the dividend exemption system,⁶² in the context of the IIR, one approach to address intercompany transactions is to disregard those transactions that offset each other and make adjustments for dividends and other distributions. Further, any underlying income that is distributed as a dividend is expected to be subject to the GILTI regime (albeit not taxed thereunder) as income of the distributing entity, so the rule that excludes intragroup dividends from GILTI’s scope may rather be seen as an income allocation rule. In sum, any income that is excluded from the GILTI regime is either typically subject to a higher level of tax than the GILTI ETR threshold or falls outside the scope of IIRs.

As discussed above, carveouts arguably should not result in substantial variations among IIRs, and thus should not warrant additional application of pillar 2 rules if the carveouts further the policy objective of pillar 2. The income carveouts under the U.S. GILTI regime do just that.

Also, although not an explicit carveout from the U.S. GILTI regime, consideration should be given to payments that are disregarded for U.S. tax purposes. Because the U.S. entity classification system allows taxpayers to elect to disregard many non-U.S. entities, payments between entities treated as fiscally opaque for non-U.S. tax purposes will frequently be disregarded for U.S. tax purposes. While the income resulting from a disregarded payment is not included in tested income or a GILTI inclusion, neither is the expense. In other words, the payment is netted like any other intercompany payment may be under a global blending option. Consequently, payments resulting from disregarded transactions do not reduce or increase a CFC’s tested income and in principle do not affect GILTI liability because of the netting effect. That result is consistent with a global blending approach to an IIR, and thus does not warrant additional application of other pillar 2 rules.

d. Loss Blending Under Overall Approach

As discussed, the tax base for GILTI is the pro rata share of the net CFC tested income (that is, aggregate tested income, reduced by the aggregate tested loss, of each CFC). Accordingly, by allowing the netting of tested income with tested loss in the group, some tested income is offset and the GILTI inclusion percentage of aggregate tested income is reduced. The blending of tested losses and tested income on a global scale is a natural result of the global blending approach adopted in the GILTI regime.

While that approach reduces volatility in the IIR’s tax base, the GILTI regime does not include other measures to do so. Specifically, the regime does not include any mechanism to carry back or forward foreign losses or otherwise take into account timing differences between U.S. and non-U.S. tax bases. For example, if a foreign jurisdiction allows losses to be carried forward, the local tax base is narrowed while the GILTI tax base is not, resulting in timing differences. Because in the absence of a specific rule that loss carryforward is not recognized for GILTI purposes, the ETR on that income would be reduced, potentially leading to higher GILTI liability.

As noted below, finding the appropriate mechanism to address volatility is an important design aspect of pillar 2. In other words, despite allowing for blending of losses within the group for a given tested year, by not carrying back or forward losses or providing other measures to address timing differences, the GILTI regime implements a mechanism to address tax base volatility in a given year but does not address time-varying tax base volatility from year to year.

Whether GILTI’s loss blending will be considered compliant with pillar 2 ultimately depends on the blending approach that will be taken in pillar 2 and the flexibility it will offer on the matter. As discussed above, however, differences in blending options arguably should be insubstantial because they all further the same primary objective of pillar 2 while emphasizing different aspects of that objective.

e. Treatment of Routine Return Exemption

Similar considerations exist for the routine return exemption, whose application further reduces the GILTI inclusion percentage. In particular, because the exemption for routine returns under GILTI applies only to tangible property, in most cases it will further the objective of pillar 2 to police only low-taxed income because tangible property is most frequently held and used in high-tax jurisdictions (as opposed to intangible property, whose ownership can be more easily moved to low-tax jurisdictions). Also, in practice, the GILTI exemption for routine returns has proven to be a relatively small piece of the overall regime, so variations resulting from it are likely to be insubstantial. Moreover, if pillar 2 provides a combined carveout for payroll and tangible assets or determines a higher percentage for routineness, the scope of the GILTI exemption may be considered relatively narrow.

Because the routine return exemption offsets the net tested income on a global calculation, and because of the lack of allocation rules, if the routine return exemption were deemed noncompliant with pillar 2, it remains unclear how the noncompliance should be addressed. The exemption could be applied pro rata to all tested income, resulting in a portion of each item of tested income being subject to additional pillar 2 rules, or the routine return exemption could be allocated to the entities that hold the assets that give rise to the exemption, in which case a greater portion of fewer items of tested income would be subject to additional pillar 2 rules. The second method would ensure the routine return exemption aligns with the income that includes an actual routine return on the tangible assets. Also, as opposed to the GILTI regime, pillar 2 may include a combined carveout for payroll and tangible assets, whose scope is expected to be greater than the routine return exemption in the GILTI regime.

V. Variations in Pillar 2 Implementation

As discussed above, coordinating pillar 2 rules with the U.S. GILTI regime may be implemented by way of grandfathering or treating the regime as a pillar-2-compliant IIR. But just as there are likely to be variations in the domestic implementations of the primary pillar 2 rules, so too may there be variations in the domestic coordination of pillar 2 rules.⁶³ Further, if the pillar 2 IIR is set as a minimum standard framework in lieu of a common rule, the variations between individual rules would gain even more significance.⁶⁴ Accordingly, it may be useful to consider variations in domestic implementations of pillar 2 and how those variations would affect coordination with other pillar 2 rules.

Evaluating whether a regime’s IIR complies with a benchmark IIR arguably should be by way of establishing substantial similarity between the tested rule and the benchmark rule. Several essential design choices will need to be evaluated to determine whether a tested IIR is substantially similar to a benchmark IIR. That is particularly important when it is acknowledged that some variation among the IIRs is inevitable and a framework to evaluate the degree to which divergence is inherently necessary.

Also, as a result of the potential application of multiple IIRs and other pillar 2 components on the same entity or income, coordination rules will be needed. If action grandfathering the U.S. GILTI regime or treating it as a pillar-2-compliant IIR is not taken on a multilateral or unilateral level, coordination rules would be expected to equally apply to the GILTI regime. Thus, this section also discusses potential coordination methods and attempts to illustrate the implications of certain ordering rules and the absence of ordering rules.

A. Coordination Methods

1. Exemption Approach

Many commentators have highlighted the need to prioritize the application of either income inclusion (and switchover) rules or undertaxed payments (and subject-to-tax) rules under pillar 2. Each of those rules is intended to serve as a final backstop against perceived undertaxation, but only one can be the ultimate backstop. The examples below illustrate the need for coordinating the different pillar 2 rules and the significant complexity that can arise in applying them.

Ultimately, the most effective alternatives for coordinating pillar 2 rules rely on determining whether various countries’ IIRs and UTPRs are pillar-2-compliant. As noted, pillar 2 is expected to include rules coordinating its components. Specifically, an IIR is expected to be given priority such that a UTPR is not expected to apply to payments made to an entity that is subject to an IIR. The examples below illustrate the implications of that rule, alternative ordering rules, and the absence of ordering rules.

Example 1: Parent (a Country A corporation) wholly owns Sub X (a Country X corporation) and Sub Y (a Country Y corporation), which wholly owns Sub Z (a Country Z corporation). Sub X makes a deductible payment to Sub Z (see Figure 1). Countries X, Y, and Z have all adopted IIRs and UTPRs, but Country A has not.

Country X can choose to apply its UTPR either solely by reference to Country Z tax imposed on the Sub X payment or by taking into account the application of Country Y’s IIR to Sub Z’s receipt of the Sub X payment. Similarly, Country Y can choose to apply its IIR either solely by reference to Country Z tax imposed on the Sub X payment or by taking into account the application of Country X’s UTPR to Sub Z’s receipt of the Sub X payment. But if each country chooses to apply its pillar 2 rules without regard to the other, Sub Z’s income could be treated as undertaxed from each country’s perspective and subject to additional tax in each.

Prioritizing Country Y’s IIR over Country X’s UTPR or vice versa could be viable, but the first appears to be more administrable for a few reasons. First, Country Y’s IIR likely would take into account all of Sub Z’s income and expense, not just the Sub X payment, and thus would consider the Country X tax rate holistically. However, if Country X’s UTPR deviated from the expected pillar 2 coordination rules and was imposed before the IIR rule, Country X’s UTPR might consider the tax imposed by Country Z only on the Sub X payment. Consequently, Sub Z’s expenses would need to be allocated to Sub Z’s various items of income to determine the net amount of the Sub X payment subject to tax.

Second, if Country Z imposed different rates on different items of income, Country Y’s IIR would be better suited to blend those tax rates (whether on a per-country, worldwide, or other basis) than Country X’s UTPR, which naturally would focus on a particular item of income (the Sub X payment) without any blending.

Finally, because Sub Z is a subsidiary of Sub Y, Country Y may already require a determination of Sub Z’s net income under its tax principles. In that case, the taxpayer may already have much of the information necessary to apply Country Y’s IIR, whereas applying Country Z’s UTPR may require additional information (and possibly at a more granular level than needed for other purposes) or computations.

While coordinating UTPRs with IIRs is critical to the successful implementation of pillar 2, it is not the only coordination necessary, as illustrated below.

Example 2: This example is the same as Example 1, except Sub Z also makes a deductible payment to Parent (see Figure 2).

Country X must consider how to coordinate its UTPR with both Country Y’s IIR and Country Z’s UTPR. As discussed, Country X can coordinate its UTPR with Country Y’s IIR by prioritizing IIRs over UTPRs or vice versa.

Separately, Country X can coordinate its UTPR with Country Z’s UTPR under one of three approaches⁶⁵: (1) prioritizing the last jurisdiction in a chain of ownership or payments that imposes a pillar-2-compliant rule to a given item (the ultimate jurisdiction approach); (2) prioritizing the first jurisdiction in a chain of ownership or payments that imposes a pillar-2-compliant rule to a given item of income (the immediate jurisdiction approach); or (3) reducing the extent to which an UTPR applies by the extent to which the UTPR of each preceding jurisdiction in a chain of ownership or payments applies to a given item of income (the credit approach).⁶⁶ In either the ultimate jurisdiction or immediate jurisdiction approach, a determination must be made regarding whether another country’s UTPR is pillar-2-compliant.⁶⁷

Also, Country Y again must choose to apply its IIR either solely by reference to Country Z tax imposed on the Sub X payment (net of the Sub Z payment) or by taking into account the application of Country X’s UTPR to the Sub X payment (and potentially the Sub Z payment). Country Y now also must consider how to take into account the application of Country Z’s UTPR to the Sub Z payment, although it seems Country Y ought to apply its IIR as if Country Z did not have an UTPR because Country Z’s UTPR polices the tax imposed on Parent’s income, which is outside the scope of Country Y’s IIR. If Country Y adopts that approach and countries X and Z adopt an ultimate jurisdiction approach to coordinating their UTPRs, prioritizing Country X’s UTPR over Country Y’s IIR would face an additional administrative difficulty of needing to allocate the Country X UTPR impact between the Sub X payment (subject to the Country Y IIR) and the Sub Z payment (not so subject).

The next example illustrates how the complexity described above magnifies as the organization grows and how that complexity can be managed effectively with coordination rules that prioritize the most common pillar-2-compliant regime.

Example 3: This example is the same as Example 2, except Parent is now Sub A and is wholly owned by USP, a U.S. corporation (see Figure 3).

In Example 3, it is obvious that the coordination of the country X, Y, and Z income inclusion and UTPRs can be greatly simplified by yielding to the U.S. IIR (the GILTI regime). To achieve that result, countries X, Y, and Z would all need to prioritize IIRs over UTPRs, adopt an ultimate jurisdiction approach to coordinating IIRs, and treat the GILTI regime as pillar-2-compliant.⁶⁸ Under that approach, (1) Country Z would not apply its UTPR to the Sub Z payment because Sub A is subject to a pillar-2-compliant IIR; (2) Country Y would not apply its IIR to Sub Z because Sub Z is subject to a higher-tier pillar-2-compliant IIR; and (3) Country X also would not apply its UTPR to the Sub X payment because (a) it has adopted an immediate jurisdiction approach to UTPRs, Sub Z’s income is subject to a pillar-2-compliant IIR, and the Sub Z payment is subject to a pillar-2-compliant UTPR, or (b) it has adopted an ultimate jurisdiction approach to UTPRs, Sub Z’s income is subject to a pillar-2-compliant IIR, and Sub A’s income is subject to a pillar-2-compliant IIR.

The examples are rather simplistic. MNEs commonly have dozens, if not hundreds, of subsidiaries organized in various countries that may be held through multiple tiers, depending particularly on how acquisitive the enterprise is. And those subsidiaries almost invariably make regular deductible payments to one another for intercompany services, licenses, and interest.⁶⁹ Failing to coordinate pillar 2 regimes, or doing so inconsistently, would result in extraordinary complexity and administrative difficulty, as well as risk double taxation, ultimately undermining the efficacy of, and public confidence in, pillar 2. Consequently, it seems appropriate to multilaterally adopt clear coordination rules that simplify the application of pillar 2 regimes. As noted, the simplest approaches, as expected to exist in pillar 2, will involve at least one country yielding to another’s pillar-2-compliant regime. Thus, it is critical that an IIR’s pillar 2 compliance can be evaluated.

2. Credit Approach

The majority of the discussion in this article has focused on the treatment of a pillar 2 regime as pillar-2-compliant, thus warranting a complete exemption under other pillar 2 regimes. However, if the U.S. GILTI regime is not considered pillar-2-compliant, or if a jurisdiction adopts a credit approach to coordinating pillar 2 regimes, it will be necessary to determine the U.S. tax imposed under GILTI. The peculiarities of the GILTI regime and its placement in the broader U.S. international tax system warrant brief additional discussion.

The residual GILTI tax is calculated under a credit system, and there are considerable limitations to the ability to use FTCs. Importantly, the allocation of shareholder-level expenses will reduce the shareholder’s GILTI category income for FTC limitation purposes, decreasing the shareholder’s available FTCs. That has the effect of denying the shareholder a deduction for the portion of its expenses related to the GILTI inclusion.

Also, the mechanism to reach a 10.5 percent overall rate for GILTI inclusions — the 50 percent deduction for the GILTI amount — is limited to 50 percent of the U.S. shareholder’s aggregate taxable income, and a taxpayer’s excess GILTI FTCs cannot be carried back or forward. As a result, GILTI inclusions that would be subject to a U.S. tax rate of 10.5 percent or lower (depending on the effective foreign tax rate) can be offset by shareholder-level losses that otherwise would be available to offset the shareholder’s income subject to a 21 percent U.S. tax rate.⁷⁰ These aspects of the regime may result in U.S. shareholders paying additional U.S. tax due to GILTI even if the foreign ETR is above the regime’s 13.125 percent threshold. That additional U.S. tax should also be taken into account in determining the impact of the GILTI regime if a potential yielding regime does not provide an exemption.

The allocation of that additional U.S. tax to different items of CFC income presents another challenge if no exemption is provided for the application of the GILTI regime. Although one method would be to allocate additional U.S. tax pro rata to the aggregate net CFC tested income of the U.S. shareholder, that method might not accurately align the additional U.S. tax with its source. Specifically, a U.S. shareholder can eliminate the additional U.S. tax on CFC income subject to a foreign ETR above 18.9 percent by making a high-tax election. Thus, if that election is not made, the additional U.S. tax arguably is allocable solely to CFC income subject to a foreign ETR at or below 18.9 percent. An alternative approach would be to calculate the U.S. tax resulting from the GILTI regime on a hypothetical per-jurisdiction basis and allocate the actual amount of that tax among jurisdictions in proportion to their shares.

B. Relevant Design Choices

There are many design choices to be made in determining an IIR’s tax base. First, policymakers must decide on the scope of the IIR, including determining the relatedness and group size thresholds. Second, the proper methods for determining the base and the applicable law or standard (those of the host or home state) must be chosen. If a financial accounting standard is to be applied, choices will need to be made regarding the required adjustments to the financial accounts, including for permanent and temporary differences. The minimum tax rate, which serves both as a threshold for the application of the IIR and the applicable rate for the tax liability under the IIR, must be considered. The blending options have important consequences for the rule’s policy impact and administration. Carveouts, depending on their extent, may considerably shape the rule’s overall reach. Other aspects to consider include measures to address volatility and determine relevant foreign taxes.

Each of those design aspects should affect the IIR’s overall reach and thus inform whether a particular IIR is substantially similar to a benchmark IIR. Further, some of them can give rise to variations among IIRs that, even if substantial, can be addressed in isolation by limiting the extent to which a yielding regime yields to a controlling regime.

1. Scope

An IIR applies to a shareholder’s proportionate share of the income of a foreign branch or subsidiary. It is necessary to decide what level of relatedness is sufficient to take into account a shareholder’s proportionate share of a subsidiary’s income. As a starting point, the income of any foreign subsidiary that is wholly owned by the shareholder would be subject to the IIR. While the inclusive framework’s consultation document on the digital economy suggested that the rule would apply to any shareholder with a significant direct or indirect ownership interest in that company and that a 25 percent ownership threshold could be appropriate,⁷¹ pillar 2 could instead follow consolidated financial statements for determining the entities that are included in the group. In that case, the threshold would not be lower than control. Pillar 2 is expected to adopt some definitions used in country-by-country reporting for defining the scope of the IIR.⁷²

Further, a €750 million revenue threshold adopted for CbC reporting and proposed for pillar 1 might be used for pillar 2.⁷³ It is also expected that a de minimis profit exclusion will be provided whereby jurisdictions in which an MNE group earns less than a set percentage of the group’s overall profit are excluded from the scope of pillar 2, including the IIR. Further, pillar 2 is expected to exclude from its scope some types of entities, such as investment funds and pension funds.⁷⁴

The scope of IIRs may be relatively consistent if multilateral consensus or agreement can be reached. Differences in IIR scope, particularly regarding relatedness and group size thresholds, could be substantial in that they create significant new administrative burdens, implicate vastly different business arrangements, or trigger different financial accounting standards. In that case, however, a yielding regime likely can apply to the variation in isolation because the gap in scope is limited to the income of at least one particular legal entity. Where these differences are substantial, however, a yielding regime likely can apply in isolation to the variation because the gap in scope is limited to the income of one or more particular legal entities. Differences may be more difficult to address in isolation depending on the level of blending used (for example, global blending versus per-country or per-entity blending).

2. Measuring the Tax Base

The tax base of the foreign minimum tax may be determined under the laws of the jurisdiction applying the tax (the parent’s jurisdiction or home state), the laws of the foreign jurisdiction to which the income is allocated (the subsidiary’s jurisdiction or the host state), or another set of rules or standards.

a. Using Financial Accounts

The inclusive framework has suggested that financial accounting statements may be used as a starting point for an IIR, followed by some adjustments to arrive at a proper measure of income. It seems likely that pillar 2 will adopt an approach like that.⁷⁵ There are many important questions on the administration and impact of the use of financial accounts.

i. Choice of Accounting Standard

There is not a single globally accepted accounting standard, so if an IIR uses financial accounts as a starting point, it is necessary to determine which accounting standard is to be used. Similar to the design choices available in choosing the applicable law, policymakers choose the accounting standard used by either the home or host state. The issue is also closely linked to the blending option chosen for the IIR in question.

If consolidated financial statements are already prepared by the shareholder in the jurisdiction that is applying the IIR, they could be a starting point under a global blending approach. If they are unavailable, it would be necessary to aggregate the figures in financial statements prepared by separate entities that may be using different standards.

Using financial statements may be particularly challenging for properly allocating income and taxes among jurisdictions. Financial statements do not necessarily follow the income and tax allocation rules in tax laws and may not present solutions to difficult fact patterns. That would be further exacerbated in a jurisdictional blending setting. The pillar 2 proposal might use the parent’s consolidated financial statements with adjustments, including for dividends, gains from stock dispositions, and stock-based compensation expense. Also, to align the use of financial statements with the jurisdictional blending approach, intercompany items in the same jurisdiction might be excluded.

ii. Adjustments

In applying an IIR, adjustments to the financial statements would have to be made. As an initial matter, depending on the threshold of ownership for the IIR, it would be necessary to determine and distinguish the income of foreign corporations where the shareholder’s ownership does not meet the threshold. Also, if an entity is not wholly owned by the group, financial statements may reflect the entirety of the assets, income, and liabilities of the entity with an equity adjustment. Adjustments to the financial accounts would need to be made to correctly represent the entity’s items. Depending on the complexity of the corporate structure, those adjustments could prove considerably burdensome and impair the ostensibly competitive advantage of using financial accounts.⁷⁶

The inclusive framework has focused on adjustments for permanent and temporary differences arising between the income computed under financial accounting standards and the tax rules. It has suggested three approaches for addressing temporary differences: carryforward of excess taxes and tax attributes, deferred tax accounting, and multiyear averaging.

Carrying forward excess taxes and tax attributes would entail carrying the taxes paid in excess of the minimum tax rate to future years, refunding or crediting foreign income taxes paid against another of the taxpayer’s tax liabilities when the entity pays foreign taxes over the minimum foreign rate, and a loss carryforward rule.

In deferred tax accounting, temporary differences are addressed by creating deferred tax assets and liabilities.

Multiyear averaging would provide a partial solution to the problem by calculating the annual ETR based on the total income and taxes of relevant subsidiaries. While the inclusive framework has indicated that deferred tax accounting would bring relatively fewer compliance burdens in the pillar 2 context,⁷⁷ it instead may choose to implement a mechanism to carry forward excess taxes.

iii. Financial Account Implications

There may be important implications of using financial accounts as the tax base for an IIR. It has been noted that it is questionable whether an accounting measure of profit is a better measure than taxable profit.⁷⁸

Arguments for aligning accounting and tax income generally revolve around simplicity and reduced compliance costs similar to those in the pillar 2 context,⁷⁹ as well as tackling tax avoidance.⁸⁰

There are also arguments against that practice. First, financial accounts and tax accounts have fundamentally different objectives. Financial accounts are prospective, meant to inform investors and other stakeholders about a company’s well-being and prospects, and involve judgment and valuation.⁸¹ They therefore present more than one figure to capture multiple aspects of a company’s financial status and focus on the economic position, not the legal form.⁸² Tax accounts, on the other hand, are retrospective and meant to depict a single accurate picture of the company at a specific point in time for revenue collection purposes.⁸³ There are concerns that aligning the two accounts runs the risk of distorting those objectives.⁸⁴ Further, using financial accounts may discourage host states from providing tax incentives⁸⁵ or implementing appropriate tax measures in response to local economic conditions.⁸⁶ Also, accounting standards are set by private bodies, which raises numerous questions, including on constitutionality and accountability.⁸⁷

Unless necessary adjustments are made, using financial accounts would impair the impact of tax incentives.⁸⁸ Whether preventing host states from providing appropriate incentives would result in increased nontax incentives such as investment funding or research and development funding is another issue to consider.⁸⁹ Because there are considerable differences between accounting standards, observers have also questioned whether using financial accounts may simply pave the way for another type of forum shopping.⁹⁰

The benefits of using financial accounts partly depend on the level of ownership the IIR applies to, as well as the blending option chosen. If the IIR applies at the level of the immediate, rather than ultimate, shareholder, the required financial reporting information might not be readily available because consolidated financial accounts do not distinguish items between intermediary entities.⁹¹ Financial statements might also exclude some information based on immateriality, which might not conform with IIR thresholds.⁹² It has been suggested that under an entity-based or jurisdictional blending approach, relying on national tax accounting standards does not represent a feasible option.⁹³

b. Applying Home-State Law

Using the home state’s laws would have clear benefits.⁹⁴ First, the jurisdiction applying the tax would be more familiar with its own laws, rendering the rule more administrable. Further, because some jurisdictions already have CFC rules, which the IIR would complement, experience in that area could be useful. Some calculations might already be required for the application of a CFC regime.

Using the home state’s laws, however, does not come without complications. If the home state’s laws are used to measure the base, differences may occur depending on the comparative scope of tax bases under the host and home state’s laws, potentially resulting in residual tax being imposed solely because of those differences. In other words, the use of the home state’s laws creates increased risk of over- or undertaxation solely because of timing differences, including the treatment of loss carryforwards, the recognition of income and expenses, and amortization and depreciation differences.⁹⁵

c. Effect on Pillar 2 Coordination

Unlike differences in the scope of an IIR, which often can be addressed in isolation, differences in measuring an IIR’s tax base cannot be readily isolated. The rules used and adjustments made to measure the tax base are fundamental to the application of an IIR. Thus, if there is a substantial difference in the base of two IIRs, it might be impossible for one jurisdiction to yield to another.

Even so, although differences in IIR tax bases are likely to be common, they often may be insubstantial. One IIR objective is to ensure the income of a home-country shareholder earned through branches and subsidiaries is subject to a minimum level of worldwide tax. The variations in measuring the base that falls in scope largely relate to promoting administrability and minimizing distortions, but in furtherance of that objective. Differences are most likely to be substantial when a host country’s laws or financial accounting standards, rather than those of the home country, are used, because the core objective shifts from policing the home country’s tax base to policing each host country’s tax base.

3. Rate

In imposing the rate for the IIR, multiple options are available: a fixed percentage rate; a percentage of the corporate rate applied by the jurisdiction applying the IIR; or a range of rates that takes into account other design elements. The inclusive framework has indicated its preference for a fixed percentage rate for the IIR, because using a percentage of a host jurisdiction’s rate would result in variations in the rates applied worldwide.⁹⁶ Presumably, a primary outcome of the inclusive framework’s work will be to establish multilateral consensus on an appropriate minimum rate to satisfy pillar 2 compliance.

4. Blending

The tax base of an IIR may be determined on a per-entity, per-jurisdiction, or overall basis. The inclusive framework has indicated that it is focused on blending at the jurisdictional and overall levels.⁹⁷ Each of those options targets separate policies. A per-jurisdiction tax base ensures a set level of tax is paid in each jurisdiction where a group operates, while an overall tax base ensures a group is exposed to a specific level of taxation on its overall foreign income.

a. Jurisdictional Blending

A jurisdictional blending approach requires a taxpayer to determine its foreign income in each jurisdiction.⁹⁸ Accordingly, under a per-jurisdiction approach, a taxpayer’s foreign income and taxes are assigned and allocated among relevant jurisdictions. If an item of income is subject to tax or taken into account in more than one jurisdiction, it must be determined to which jurisdiction the tax should be allocated. An item of income may give rise to tax in multiple jurisdictions in multiple contexts, such as dual resident companies or income subject to both residence and source state taxation.

In those cases, a tiebreaker rule is necessary to determine to which jurisdiction the income and related taxes would be allocated for IIR purposes. Alternatively, the income and tax may be allocated between the two jurisdictions. With MNEs, there could be more than two jurisdictions taxing a single item of income, further complicating the matter.

Here, how intercompany transactions are addressed becomes crucial. They could be regarded as requiring the application of the method chosen to assign the income — that is, a tiebreaker or allocation rule — to them (without considering whether they are regarded under local laws).⁹⁹ They could also be regarded depending on whether they are regarded under the local law.¹⁰⁰ In that case, distortions could occur because of how different laws treat the same transactions involving hybrid elements. Another way to address intercompany transactions under a jurisdictional approach would be to disregard them by allowing their income and expense legs in one jurisdiction to offset each other and make adjustments for dividends and other distributions (or transactions with one leg).¹⁰¹ In that case, there would be no income allocation to the residence state, and the taxes imposed in the residence state would be allocated to the source state. Tiebreaker or allocation rules would still be needed when two states subject the same income of the same entity to tax as source state income (such as when a foreign branch has income sourced in a third jurisdiction).

Allocating income and taxes between a resident and source state (and potentially between source states) is particularly important for branch and partnership income. There can be differences in how the head office and branch jurisdiction’s laws attribute profits to a branch. Thus, to resolve those differences, the rules would need to include principles to allocate that income, rely on the tax accounts held in the branch according to local laws, or rely on the attribution made at the head office jurisdiction. For partnership income, policymakers have similar choices: The IIR may include novel principles to allocate the income, rely on the tax accounts of the partnership, or rely on the income inclusion at the partner level under the laws of the partner’s jurisdiction. Other complexities could arise for income arising under transactions between a partnership and its partners. One group has suggested that for some complications with partnerships, it is implausible to create a standard rule, so a facts and circumstances type of analysis would be required.¹⁰²

For branches, partnerships, and fiscally transparent entities, an overarching consideration would be the income of hybrid entities, which are classified differently across jurisdictions. In those cases, the rules would need to choose to respect the classification either at the entity or owner level. In complex structures, that might mean that a jurisdiction applying the IIR would need to disregard the classification under its own laws unless those laws provide priority to the classification.

There is also the issue of income that is not taken into account in any jurisdiction (stateless income). The rules would need to address the allocation of that income (or disregard it and therefore implicitly exclude it from the scope of the IIR).

A jurisdictional approach would require a detailed allocation of income and tax and would thus entail meticulous rules addressing various fact patterns. Particularly, depending on how intercompany transactions are addressed, determining the tax base under a jurisdictional approach might require measuring the tax base on an entity level, even if all tax bases in a jurisdiction are later aggregated for determining whether the minimum level of taxation is met.

b. Entity Blending

Under an entity-level blending approach, a taxpayer must determine the income and taxes of its group’s foreign entities (and any of the entities’ branches) and whether the ETRs imposed on them are below the minimum rate. The challenges and design concerns of the jurisdictional blending approach generally also are relevant for the entity approach.

An entity blending approach also poses major administrative difficulties, especially for groups that operate in consolidated or similar groups in foreign jurisdictions. To alleviate the difficulty in singling out an entity’s income and taxes when income, losses, or other items are grouped under the foreign jurisdiction’s laws, the inclusive framework has suggested that local group blending could be allowed for entities that are members of a consolidated group or take advantage of group tax relief under foreign law.¹⁰³ Under that approach, entity-level blending comes close to jurisdictional blending for groups that operate in those regimes. Because tax consolidation or similar regimes are not available everywhere, allowing local blending only for jurisdictions with those features could result in having both the per-entity and per-jurisdiction base contingent on a jurisdiction’s permission of consolidation. Considering the tax consolidation or group tax relief regimes available in a host state jurisdiction has the primary effect of being able to use losses sooner.¹⁰⁴

c. Global Blending

Under a global blending approach, the tax base would be the group’s overall foreign income; thus, allocating income and taxes is scaled down to distinguishing foreign income and taxes from domestic income and taxes, rendering the allocation of income and taxes among foreign countries irrelevant. An IIR could include its own sourcing and allocation rules or rely on the rules of the jurisdiction applying it.

Because the sole distinction under a global blending approach is between foreign and domestic income and taxes, the noted allocation issues still exist but on a much smaller scale. Any income deemed domestic by a jurisdiction and thus subject to its general taxation regime should be outside the scope of the IIR by definition. In a global blending approach, allocation matters would be expected to be consistently resolved for general income tax purposes under the home jurisdiction’s laws. That approach would also provide a considerable administrability benefit. Particularly, global blending would provide a workable solution for tracking additional third-country taxes, such as withholding taxes on lower-tier subsidiaries imposed by an upper-tier subsidiary’s jurisdiction.

As discussed above, the choice of blending largely depends on the desired policy objective.¹⁰⁵ Global blending offers considerable advantage and flexibility in the administration of an IIR. Compared with jurisdictional and entity-based blending, global blending could put less pressure on the host country incentives and result in less volatility in the IIR’s tax base.¹⁰⁶

d. Impact on Pillar 2 Coordination

Blending options, like permutations of measuring the tax base, might create IIR variations that cannot be easily isolated and addressed by a yielding regime. But also like differences in measuring the tax base, those in blending options frequently may be insubstantial. Although they could create differences in the extent to which residual tax is imposed under an IIR, the substantive objective under any blending option remains the same: ensuring a taxpayer’s global income from its foreign branches and subsidiaries is subject to a minimum level of taxation. Different blending options will emphasize various drivers of the core objective differently. For example, global blending offers greater administrability and simplicity and reduces volatility in the tax base, while per-jurisdiction blending provides a more granular evaluation of the ETR for each item of income. At the same time, all the approaches still substantially achieve the same policy objective.

5. Addressing Time-Varying Base Volatility

As discussed, the inclusive framework has explored the carryforward of excess taxes and tax attributes and multiyear moving averages as potential methods to address the temporary differences arising from the use of financial accounts. In fact, regardless of the measure chosen to determine the IIR’s tax base, policymakers must choose whether and how to address volatility in tax bases arising over consecutive years. Design options include carrying forward and back losses, excess taxes, and tax attributes, or using a multiyear moving average tax base. It has been suggested that because some jurisdictions are unwilling to provide carryback and carryforward mechanisms for a sufficient time — and even when they do, obtaining refunds can be difficult — under a jurisdictional blending approach, those mechanisms would not effectively address volatility.¹⁰⁷

Tax base volatility may also occur retroactively as a result of carrying back losses or foreign tax redeterminations. In contrast with carrying back losses that mostly affect only a few preceding years, foreign tax redeterminations, which are increasingly common in the global tax landscape,¹⁰⁸ may alter the ETR in much earlier years. Further, resolving tax disputes could take an excessively long time. Hence, the options discussed here might be insufficient to satisfactorily address the impact of foreign tax redeterminations that extend over longer periods. Special rules may be needed to ensure that those redeterminations are taken into account when carryback or multiyear moving average mechanisms do not properly address their effects. At the very least, in those situations, taxpayer could be given the option to apply any additional tax arising from the foreign tax redetermination to the current year.

As with blending options, other measures addressing tax base volatility are difficult to extricate from an IIR and thus unlikely to be readily isolated, but their differences are also unlikely to be substantial relative to an IIR’s overall policy objective. IIRs with fewer measures to address tax base volatility likely would be both harsher and simpler, but they will further the same objective as more precise and complex IIRs.

6. Carveouts

The inclusive framework has indicated that it is exploring carveouts from the IIR. Potential carveouts may include substance-based carveouts, including carveouts for returns on tangible assets and CFCs with related-party transactions below a threshold, as well as sector- or industry-based carveouts.¹⁰⁹ One group has suggested a safe harbor rule carving out groups subject to tax at sufficiently high effective rates, as measured by consolidated financial statements.¹¹⁰

Several noteworthy concerns and observations have been submitted regarding carveouts. While a substance-based exception for cases in which enough activity occurs in a jurisdiction likely would diminish an IIR’s effectiveness in eliminating tax competition,¹¹¹ the lack of a general substance-based carveout would be a strong indication that the scope of pillar 2 departs from the rationale behind the BEPS project.¹¹² Reportedly, some countries particularly favor substance-based carveouts, which they view as necessary to ensure that pillar 2 focuses on remaining BEPS questions.¹¹³

One group has suggested that because specific industries are subject to special taxation regimes in some jurisdictions (such as shipping industry, extractive industries, and financial sector), carveouts are justified for those businesses.¹¹⁴ However, carving out certain industries or sectors while leaving others within the scope of the rule would result in additional administrative burden.¹¹⁵

By providing an exemption for returns on tangible assets, the IIR would exclude profits that do not exceed certain routine return levels and arguably do not constitute a problem area from a base erosion perspective. There are complexities in designing a carveout for returns on tangible assets. As an initial matter, it would be required to determine an acceptable or routine return. A universally applicable appropriate “routine” return appears unattainable because routineness would depend on a variety of factors specific to circumstances, such as industry, jurisdiction, market, and time period. An ideal carveout design could address those differentiations, either by different setting rates based on circumstances or by making adjustments, although that would lead to significant administrability challenges.¹¹⁶

Because a carveout for returns on tangible assets would favor asset-intensive businesses over capital-intensive or intangible-asset-intensive businesses, some observers have suggested it be implemented in tandem with other carveouts based on alternative measures of genuine economic activities to prevent ring-fencing.¹¹⁷ Providing a carveout for tangible assets naturally would require distinguishing tangible assets. Local laws might define those assets differently, and depending on the law to be applied to the IIR, those differences could distort the application of the carveout. The inclusive framework is reportedly considering a combined carveout from the pillar 2 IIR based on payroll and depreciation of tangible assets.¹¹⁸

An overarching comment on pillar 2 is that the IIR arguably should not eliminate the ability to provide incentives with nontax policy objectives.¹¹⁹ One group has said countries should be able to attract genuine investments in accordance with international development commitments, including the U.N. sustainable development goals.¹²⁰

Carveouts frequently can be addressed by a yielding jurisdiction in isolation if they apply to specific entities or items of income. In several cases, carveouts may be implemented in furtherance of the core policy objectives of pillar 2. For example, carveouts based on high levels of taxation directly align with the objective of ensuring a sufficient level of taxation is imposed. Carveouts based on thresholds might not achieve that objective but could be justified because they promote simplicity and administrability. On the other hand, some may argue that carveouts based on substance, routine returns, or income immobility might stray from the objectives of pillar 2, which arguably include ensuring a minimum level of tax is paid on all income, irrespective of whether it was shifted to a low-tax environment to achieve tax savings. When carveouts stray from those objectives but have a relatively small impact on the overall application of a regime, however, the variation may still be insubstantial.

7. Relevant Foreign Taxes

IIR design should specify which foreign taxes will be considered so one can determine whether the ETR is below the minimum tax rate, compute the amount of tax liability under the IIR, credit the foreign taxes against the foreign minimum tax,¹²¹ and compute excess taxes for carryback and carryforward rules. While it generally would be straightforward to identify the nature of a foreign tax as an income tax, it could sometimes be complex.

There is common agreement that not all payments made to a government are taxes. The criteria to distinguish taxes from other payments made to a government include the compulsoriness and unrequited nature of payment. While there is not a universally accepted definition of the terms “tax” or “income tax,” pillar 2 could provide a general definition for its purposes.

While there may well be many variations in the taxes relevant to different IIRs, they should be considered insubstantial so long as they reasonably attempt to capture income taxes and similar taxes. If an IIR were to take into account other taxes, such as value added taxes, net wealth taxes, or property taxes, however, the variation could be substantial because it would deviate from the primary objective of pillar 2, which appears focused on income and similar taxes. If a substantial variation exists, it is unlikely to be readily isolated because it would require a hypothetical application of the potential controlling regime without regard to specific taxes and a residual application of the potential yielding regime if the remaining taxes did not meet the minimum threshold.

VI. Conclusion

There are multiple options for coordinating the GILTI regime with pillar 2. Primarily, GILTI may be grandfathered. A grandfather would treat an MNE subject to the GILTI regime as exempt from further application of pillar 2 rules and eliminate the need for determining the extent of GILTI’s compliance with pillar 2. In the absence of grandfathering, it would be required to determine whether GILTI is a compliant regime. Given the vast variations in existing taxation regimes and practices around the globe, substantial similarity, not identicality, should be the standard for determining whether a regime is compliant with the pillar 2 IIR. Determining whether a particular regime, and GILTI, is substantially similar, in turn, should take into account only whether an appropriate balance is stricken between efficacy, simplicity, administrability, and transparency, as well as double taxation risk is minimized. Compared with the pillar 2 IIR, various design features of the GILTI regime are expected to be more onerous. With respect to other aspects of GILTI which could be viewed as relatively less onerous (primarily global blending), the variations should generally not result in substantial differences on a qualitative basis. And, even if differences were found to be substantial, arguably they could be readily isolated and thus resolved by way of limiting the application of other jurisdictions’ pillar 2 rules to those differences, instead of treating the GILTI regime as noncompliant in its entirety. If the GILTI regime is neither grandfathered nor treated as a pillar-2-compliant, in order to determine whether GILTI, just like any other tested IIR, is substantially similar to a benchmark IIR, a categorical evaluation of the essential design choices will be needed. In that case, the extent and efficacy of coordination rules will be of utmost importance.