If Pillar 1 Needs an MLI, Why Doesn’t Pillar 2?

On October 8, 2021, the OECD/G-20 inclusive framework on base erosion and profit shifting agreed to a two-pillar framework to address tax challenges arising from a digital economy and set forth the components of each (the “October statement”) as well as a detailed plan for implementation.¹ As of today, over 140 countries are inclusive framework members.

The inclusive framework recognized early on that some form of multilateral instrument would be needed to implement pillar 1 in addition to domestic law changes that would be required in many countries.² As discussed further below, an MLI was deemed necessary because profits of a taxpayer would be reallocated to countries where that taxpayer did not have a permanent establishment or other taxable nexus as required under international tax standards.

However, the inclusive framework envisions that pillar 2 can generally be implemented through domestic legislation notwithstanding similar reallocations of profit across multiple jurisdictions. This article concludes that an MLI is also necessary to implement pillar 2 and the development of a pillar 2 MLI must be included in the inclusive framework workplan if participating countries wish to faithfully implement pillar 2 in a manner consistent with their existing treaty network.

The Pillar 1 MLI

Pillar 1 reallocates profit to countries that would not otherwise have a right to tax such profits under international tax norms. Under international tax principles recognized around the world in domestic legislation, in bilateral income tax treaties, and in the OECD, U.N., and U.S. model treaties, a country may tax the business profits of a nonresident taxpayer only if such taxpayer has some form of taxable nexus with the country.

Pillar 1 creates a new taxing right enabling a market country to tax an allocated portion of profits of an in-scope taxpayer regardless of whether that taxpayer has a PE or other taxable nexus in the country. The reallocation of profit and resulting taxation arises simply because the taxpayer meets a minimum sales threshold in that country. Subject to some adjustments, this reallocation, referred to as amount A, reallocates 25 percent of the residual profits of an in-scope taxpayer above a 10 percent profit margin. Amount A is surrendered from the jurisdictions that earn the residual profits to market jurisdictions meeting a minimum annual sales threshold of €1 million and is then apportioned among those jurisdictions based on relative sales.³

The PE provisions in income tax treaties prevent a market jurisdiction from imposing amount A simply because the taxpayer sells in that country. As a result, the inclusive framework recognized that treaty changes would be necessary to implement pillar 1. Although the inclusive framework countries could have undertaken the heavy lift of amending their bilateral tax treaties to address this conflict, the inclusive framework determined an MLI would be a more efficient approach, similar to the approach taken regarding the BEPS action 15 MLI. Moreover, given the formulaic aspects of the allocation of amount A to market jurisdictions and the fact that an allocation to one market country will directly affect the allocation in another market country, an MLI was the only logical approach to pillar 1 implementation.

The Pillar 2 Architecture

As anticipated in the October statement, pillar 2 includes a hierarchical set of rules that shapes the OECD’s common approach to granting jurisdictions additional taxing rights: (1) the global anti-base-erosion (GLOBE) rules and (2) the subject-to-tax rule (STTR). The introduction of these taxing rights seeks to (i) ensure minimum taxation while avoiding double taxation or taxation in which there is no economic profit, (ii) cope with different tax system designs by jurisdictions as well as different operating models by businesses, (iii) ensure transparency and a level playing field, and (iv) minimize administrative and compliance costs.⁴

The GLOBE rules are the principal mechanism used by the inclusive framework to achieve the pillar 2 objectives. These rules consist of two interlocking rules that impose a top-up tax using an effective tax rate (ETR) test that is calculated on a jurisdictional basis⁵ and require an assignment of the income and taxes among the jurisdictions in which the multinational enterprise operates and to which it pays taxes.⁶

The first prong of the GLOBE rules is the income inclusion rule, which imposes a top-up tax on the ultimate parent entity (UPE) with respect to the low-taxed income of a constituent entity. The IIR allocates the top-up tax based on a top-down approach subject to a split-ownership rule for shareholdings below 80 percent.

The operation of the IIR is, in many respects, based on traditional controlled foreign corporation rules that many jurisdictions have already enacted in domestic laws. As such, the IIR operates by requiring a parent entity to bring into account as income its proportionate share of the income of each subsidiary located in a low-tax jurisdiction in which it owns an equity interest. That income is then taxed up to the GLOBE minimum rate, after crediting any covered taxes on that income.⁷

The second prong of the GLOBE rules is the UTPR. Although originally styled as an undertaxed payments rule, the UTPR can apply without regard to whether payments have been made, so it is more accurately referred to as an undertaxed profits rule. The UTPR acts as a backstop to the IIR by denying deductions or requiring a reallocation of profits to increase the ETR of the UPE group to offset income taxed below the global minimum rate.

Using the same ETR mechanics as the IIR, the UTPR first determines the top-up tax in a particular country by subtracting the ETRs of the members of the UPE group from the 15 percent minimum tax rate. After the total top-up tax is determined for all jurisdictions, it is then allocated to the countries based on a two-factor formula using the relative number of employees and book value of tangible assets. The UPE group members in each country are then denied deductions or are otherwise reallocated profits to increase their tax expense up to the amount of the top-up tax allocated to that country.

Pillar 2 and Treaties

The three components of pillar 2 directly implicate bilateral income tax treaties in various degrees. The IIR (like its model in the U.S. global intangible low-taxed income regime) is a CFC-driven mechanism. As such, it effectively imposes tax at the UPE level. Accordingly, the pillar 2 blueprint concludes that the IIR is compatible with treaties because of the saving clause in most bilateral income tax treaties that clause gives a treaty partner the right to tax its own residents notwithstanding a treaty obligation.⁸

There are two issues with hanging the IIR hat on the saving clause rung. First, the saving clause is not included in every treaty.⁹ Second, the application of the saving clause to CFC regimes has been questioned.¹⁰ A pillar 2 MLI would address these concerns.¹¹

The UTPR, however, is not a CFC-driven regime, so a saving clause would not protect it. The pillar 2 blueprint emphasizes that the UTPR does not violate article 9 of the OECD model because domestic law governs the deductibility of expenses allocated to a resident under the model or a PE.¹² The blueprint also concludes that a UTPR along the lines envisaged under the GLOBE rules is compatible with the obligation not to discriminate on the basis of the residence of the recipient of a deductible payment set out in article 24(4) of the OECD model.¹³

These arguments ignore the basic fact that profits are being reallocated to jurisdictions where the taxpayer does not have a PE.¹⁴ That is exactly what is happening in pillar 1 where there is universal agreement that a treaty violation is occurring.

Moreover, the pillar 2 blueprint arguments ignore how the UTPR top-up tax is allocated to the various countries. After the UTPR top-up tax is determined, it is allocated among the UTPR countries by referencing a two-factor formula: (i) relative number of employees and (ii) relative net book value of tangible assets.¹⁵ The model rules note that these factors were selected specifically to represent the relative substance in each country. The link to substance is important because that arguably means that the allocation of profits (after application of the UTPR) would be at arm’s length. However, this premise is based on the faulty assumption that the number of employees and net book value of tangible assets are, on their own, sufficient factors to determine an arm’s-length allocation of profit under article 7 or article 9.¹⁶ Although simple and easy to administer (to the extent anything in the UTPR can be characterized as simple), this version of formulary apportionment may not be consistent with the authorized OECD approach for profit allocation under articles 7 and 9. A pillar 2 MLI would address these concerns as well.

The almost forgotten (but equally important, at least for developing countries) STTR is itself a treaty-based rule that allows source jurisdictions to impose limited source taxation on specified related-party payments subject to tax below a minimum rate of 9 percent. The STTR is designed to complement the GLOBE rules by denying treaty benefits for some deductible intragroup payments made to jurisdictions where those payments are subject to no or low rates of nominal taxation.

The October statement notes that the STTR was an integral part of achieving the consensus of developing country inclusive framework members. Nevertheless, implementation of the STTR is left to separate bilateral treaty negotiations between developed and developing countries. Putting aside the sheer length of time this will take, most developing countries do not have a robust treaty network. Moreover, leaving implementation of the STTR to bilateral negotiations will result in inconsistency around the world and undermine confidence in the viability of the STTR. These issues were part of the impetus for the BEPS action 15 MLI. A pillar 2 MLI would address these concerns as well.

Domestic Tax Laws Are Not Enough

The October statement indicates that, except for the STTR, pillar 2 would be implemented through domestic legislation in each of the inclusive framework countries. However, it is by no means universally accepted that domestic tax legislation can override a treaty.¹⁷

In the United States, treaty overrides have been the subject of much debate. The later-in-time rule codified in 1988 through amendments to section 7852(d) generally provides that a treaty and the IRC have equal weight so that whichever expression comes last controls. In principle, this means that if Congress enacts pillar 2 legislation that violates an in-force treaty, the legislation will control. Nevertheless, scholars continue to debate the application of the later-in-time rule when U.S. tax legislation directly violates a treaty.¹⁸ At a minimum, any U.S. legislation to implement pillar 2 that fails to expressly indicate that Congress intends to override treaties could be at risk.¹⁹

The position is similar in the United Kingdom, where the Padmore decision²⁰ confirmed the ability of Parliament to intentionally override a treaty through subsequent legislation.

In the rest of the world, treaties have a more exalted position relative to domestic legislation. For example, section 49 of Singapore’s Income Tax Act provides that income tax treaties take precedence over domestic law. Accordingly, some form of treaty modification would be needed to implement pillar 2.

Similarly, in Brazil, article 98 of the National Tax Code provides that treaties revoke or modify domestic tax law. There has been much debate around the meaning of this article, but the traditional position among Brazilian scholars is that tax treaties prevail over domestic tax law. This position is also adopted, in several circumstances, by Brazilian case law, under which the superior and federal courts confirmed that domestic legislation cannot override treaty provisions (mainly because of the lex specialis derrogat generalis criterion).²¹

In Germany, the question of whether domestic law may override international treaty law has been debated for many years. In 2015 Germany’s highest court, the Federal Constitutional Court, decided that international treaty law does not generally rank higher than German domestic law and, therefore, can be changed by new German domestic law (under the lex posterior rule).²² However, two more Federal Constitutional Court proceedings regarding tax treaty overrides are pending.²³

The situation is equally complex in France. As a general principle, article 55 of the French Constitution provides that international law takes precedence over domestic law. However, tax treaties are subject to a subsidiarity principle according to which a tax treaty can never serve as a basis for taxation in France. In other words, double tax treaties are only applicable to the extent that French law first provides a basis for taxation.²⁴ As a result, domestic tax legislation would be necessary in France to implement pillar 2, but because that legislation would conflict with existing treaties, a pillar 2 MLI would also be necessary.

Concluding Observations

It is interesting to compare why the OECD concluded that an MLI was needed for pillar 1 but not pillar 2. The need for a pillar 1 MLI is obvious. Profits are being reallocated from one country to another without regard to traditional taxable nexus. Moreover, that reallocation (both the profit surrender and the profit capture) is being done on the basis of formulaic principles that have no basis under international tax norms. From the outset, pillar 1 has been described as a new taxing right, and as a result, there has been little debate that a pillar 1 MLI was required because the treaty inconsistency is clear.

As argued here, similar concerns relate to the operation of pillar 2. The UTPR reallocates profits between countries in a manner inconsistent with income tax treaties. The OECD goes to great lengths to argue that pillar 2 does not violate treaties, but there is substantial room for doubt. A pillar 2 MLI would address these concerns as well as solve the primacy issues in many countries regarding treaty overrides.

Moreover, let’s not forget the STTR, which has been shunted aside to be implemented through bilateral treaty negotiations. If operational, the STTR has priority application over the other pillar 2 rules. Because the STTR is itself a treaty mechanism, unless STTR implementation is delayed until all the inclusive framework countries have implemented the IIR and UTPR, how will the STTR be able to retain its priority status on a globally consistent basis while the 140 inclusive framework members are off writing IIR and UTPR legislation? A pillar 2 MLI would solve this.

The pillar 2 blueprint itself acknowledges the benefits of a pillar 2 MLI, noting that “although it is not a prerequisite, a multilateral convention would be the only means to enshrine rule co-ordination in a legally binding form.”²⁵ If serious legal concerns exist concerning the implementation of pillar 2, moving forward without an MLI will result in uncertainty, litigation, and instability, all of which is contrary to the OECD’s objectives. Agreeing to another MLI will be difficult because the underlying architecture for pillar 2 is providing the rules to determine which country gets the first bite at the tax apple. The challenge is how best to implement this work around the world. Adding a pillar 2 MLI to the process removes many of the obstacles blocking the inclusive framework’s intended results.