Tax Notes logo

A Tale of Two Subject-to-Tax Rules

Posted on Mar. 4, 2024
Bob Michel
Bob Michel
Jeffery M. Kadet
Jeffery M. Kadet
Sol Picciotto
Sol Picciotto

Sol Picciotto is emeritus professor of law at Lancaster University in the United Kingdom, coordinator of the BEPS Monitoring Group, and research associate of the International Centre for Tax and Development. Jeffery M. Kadet practiced international tax law for over 32 years with several major accounting firms and now periodically teaches international tax courses at the University of Washington School of Law in Seattle as an adjunct lecturer. Bob Michel is a lawyer, an international tax treaty and tax development policy adviser, and a comparative policy and legal analyst for the Tax Justice Network.

In this article, the authors analyze and compare the two proposals for a subject-to-tax rule to be included in tax treaties, one from the U.N. Tax Committee and the other from the G20/OECD inclusive framework on base erosion and profit shifting.

Copyright 2024 Sol Picciotto, Jeffery M. Kadet, and Bob Michel.
All rights reserved.

In this article, we analyze and compare two proposals for a new subject-to-tax rule (STTR) provision to be included in tax treaties, one from the U.N. Tax Committee and the other from the G20/OECD inclusive framework on base erosion and profit shifting. The U.N. proposal is broad, and would clarify that restrictions in tax treaties on taxation of income at the source where it is derived are conditional on that income being taxed at an agreed-upon minimum rate in the country where it is received. The inclusive framework version is much more limited, being confined to payments between connected entities and specific categories of income; it is also subject to markup and materiality thresholds, as well as a capped total tax rate (taking account of the tax in both the source and recipient countries) of 9 percent. These restrictions and conditions not only make the inclusive framework version complex and hard for under-resourced tax administrations to administer, but they also create a considerable scope for differing interpretations. However, the inclusive framework version is complete and ready for implementation (on a take-it-or-leave-it basis) while the U.N.’s will require negotiation over the applicable rate and other details, but provides more comprehensive protection for source taxation. In our view, all countries would benefit from and should strongly support the U.N.’s STTR, which should be regarded as a minimum standard, especially once it is included in the Fast Track Instrument also being developed by the U.N. Tax Committee.

I. A Problem and the Two Proposed Solutions

States generally claim the right to tax the income of their residents, as well as income from the activities of nonresidents within their territory, creating a potential jurisdictional overlap between states that tax income from cross-border business and international investment. Tax treaties aim to prevent this possible “double taxation” through reciprocal obligations on the state parties. First, they specify restrictions on rights to tax source income derived from activities in their territory by residents of the treaty-partner state. In exchange, the country of residence must provide either an exemption for the income or a credit for source tax paid in accordance with the treaty. Aside from this exemption or credit, the residence state remains free to decide whether and how to tax its residents. The aim of a subject-to-tax rule is to make the restrictions on source taxation conditional on a minimum level of taxation by the residence jurisdiction of this foreign-derived income.

This is mainly because some jurisdictions have opted to apply low or zero taxes on some residents or types of income, even if they are not taxed at source because of treaty restrictions. These preferential regimes encourage multinational enterprises to create affiliates that are treated as residents in those countries, particularly those with extensive treaty networks, which make them “investment hubs.” These affiliates act as conduits, receiving payments (like interest, royalties, and fees for services) that are both deductible from the business income of the payer and untaxed in the source country, reducing the source tax base, while generally flowing through the residence jurisdiction to third countries with low or zero taxation. Consequently, this income remains untaxed anywhere, creating “double non-taxation.”

The efforts made in the OECD/G20 project on BEPS to combat these practices, especially provisions against treaty abuse and “treaty shopping,” have proven unsatisfactory so far and have mainly benefited investment hubs.1

These practices have made countries wary of concluding treaties with jurisdictions that act as investment hubs; some have cancelled these treaties or have not ratified treaties they already signed.2 This may be the best solution. However, it is politically difficult to cancel existing tax treaties, given the belief that they help to attract foreign investment. This is especially true for treaties with OECD members, which are generally considered sources of investment capital, even though many of these countries are now recognized to be investment hubs, like Ireland, Luxembourg, The Netherlands, and Switzerland.3 It can also be difficult to apply the principal purpose test, because it is couched in broad terms giving wide scope for interpretation.

These concerns have led to measures and proposals to amend treaties to prevent their use for payments that erode the tax base at source, in the form of an STTR. This is based on a simple principle: When states accept an allocation of taxing rights through treaties that restrict their right to tax income at source, it is on the condition that such income will be taxed by the other state at least at an agreed-upon minimum level. This principle is essential to ensure that treaties do not enable double nontaxation, as is now accepted. It has already been made explicit regarding the residence country’s obligations.4 However, for taxing rights ceded by source countries, it has remained largely implicit in tax treaties until now,5 although a few have included a tailor-made STTR.6 The principle is squarely in line with the mandate for the BEPS project to ensure that MNEs can be taxed “where activities occur,” and to eliminate double nontaxation.

The BEPS project has, since 2016, been open to all willing countries through the inclusive framework, which reached agreement in 2021 on a “two-pillar solution.” Pillar 2 aims to ensure a minimum level of tax in all countries where MNEs declare profits under existing rules, especially through a global minimum effective corporate tax (global anti-base-erosion, or GLOBE). Although this would do much to put a floor under the competition to reduce corporate tax rates, it allocates rights to apply a top-up tax which grants priority to investment hubs, then to residence countries, with only a backup right for source countries.7 So it is once again the investment hubs that would be the main beneficiaries of the GLOBE rules.8

As a counterbalance, pillar 2 also included a proposal for an STTR, with a commitment from members of the inclusive framework to include it in their treaties with developing countries if requested.9 In October 2023 the OECD published a report containing the text of this STTR with an explanation that acts as a commentary,10 as well as a slightly different version of the STTR annexed to a multilateral instrument (MLI) enabling its rapid adoption by willing states.11

In the meantime, the U.N. Tax Committee, at the urging of its developing-country members, has reached agreement on a different STTR. This U.N. version will be included in the next update of the U.N. model convention, due in 2025. The committee is also working on a fast-track instrument (FTI),12 which would be a U.N. multilateral convention intended to facilitate rapid adoption of key provisions of the U.N. model, including this STTR.

In this report, we analyze and compare the two proposals, particularly from the perspective of developing countries. Both aim to insert a new provision to apply an STTR into existing tax treaties, but are different in their approaches and operational details. The U.N. model provision is a paragraph of less than a half-page, with an explanatory commentary of less than five pages.13 The OECD’s version is eight pages as a model provision in the report, released on October 11, 2023, with an additional 54 pages of commentary, and the MLI is accompanied by a 37-page explanatory statement to “clarify its operation.”

Overall, the pillar 2 STTR has a range of restrictions which greatly narrow its scope and results in significant complexity. The U.N. STTR is much simpler and broader in scope, giving the source state greater freedom to determine its application, but requires agreement with the residence state, particularly on the minimum tax rate applicable, which is predefined in the pillar 2 STTR.

In the next two sections we briefly analyze each provision, then compare their implementation arrangements, concluding with a comparative evaluation.

II. The Pillar 2 STTR

The pillar 2 STTR overrides the provisions in the covered tax treaty that limit source taxation rights. It applies if the tax rate applicable in the recipient’s state of residence is below 9 percent, either generally or on specified items of covered income, including because of a preferential adjustment allowed under the recipient country’s rules, even if introduced subsequently.

However, significant and detailed restrictions on the source country’s right to tax under pillar 2’s STTR remain. This STTR is limited to:

  • payments between connected legal persons, not including individuals (STTR paragraph 8.a & b,14 defined in paragraph 10, with a targeted antiavoidance rule in para. 11);

  • specified categories of payments or income (paragraph 4): interest; royalties; income arising in the jurisdiction for the provision of services; payments for distribution rights; insurance and reinsurance premiums; financial guarantees and other financing fees; and payments for the rights to use equipment;

  • income (other than interest and royalties) that is above a markup threshold (paragraph 9): the income must be higher than the direct and indirect costs incurred by the recipient plus 8.5 percent;

  • connected recipients with an aggregate annual total of covered income of at least €1 million in the jurisdiction, or €250,000 if either jurisdiction has a GDP less than €40 billion (the materiality threshold, paragraph 12);

  • excluding: (i) pension funds or schemes and defined investment entities; (ii) nonprofits and state entities; (iii) payments for the use of a ship for international passenger or cargo transportation, or income treated by the recipient state as subject to a tonnage tax regime (paragraphs 4b and 8);

  • a “specified” tax rate capped at 9 percent (paragraph 2), taking into account the tax paid by the recipient due to both tax payable in the source state under the relevant treaty (paragraph 3), and the tax payable by the recipient on that item of covered income in the residence jurisdiction (paragraph 5), including the effects of any “preferential adjustment” (a permanent reduction in the covered income taxable or the tax payable on it: paragraph 6), or by a permanent establishment in a third jurisdiction (paragraph 7).

The tax is chargeable each year following that in which it applies (STTR paragraph 14); because of its design, information would not be available when the payment is made for calculating the tax liability, since this is subject to several conditions (the markup threshold, the materiality threshold, and the preferential adjustment).

Largely because of these delimitations and restrictions, the measure is lengthy, detailed, and complex, making the provisions difficult to apply by under-resourced tax administrations. This creates uncertainty and potential disagreements of interpretation.

For example, uncertainties may arise over the categories of covered income. The report’s commentary provides four pages of clarification and discussion of potential issues. For example: (1) payment for the full transfer of distribution rights is not considered “for the use of, or the right to use” these rights; (2) a fee charged to an MNE group member by another member acting as the group’s treasury to broker a loan from an unconnected third-party lender would be covered income, though any other fees or interest paid to the lender would not be; (3) “equipment” refers to tangible assets employed in a business, but not if fixed and regarded as immovable property (e.g., pipelines and cables); and (4) payments for the capacity of assets (e.g., to transmit communications, or transport oil) that do not entail control of the equipment are not considered to be for “the use of, or the right to use” the equipment (but might come under the general category of payments for services). These categorization problems create complexity and administrative difficulties and will be a fertile ground for disagreements.

For interest and royalties the definitions in the OECD model apply, or the parties can apply the definitions in the covered agreement if specified. Importantly, there has been a long-running disagreement about whether “royalties” include payments for using computer programs or software. The OECD model commentary was amended in 1992 to exclude most of these payments, and this interpretation was also referenced in the U.N. model’s commentary. However, this view was not accepted by many non-OECD countries, and the U.N.’s commentary in 2021 inserted an alternative interpretation preferred by some states under which royalties do include payments for the use of software.15 MNEs would likely argue against this alternative interpretation, as well as asserting that such payments fall outside the other categories of covered income in the STTR (like consideration for the provision of services or for the right to use equipment). Excluding payments to unrelated parties is already a serious limitation of this STTR, and this would be even more consequential if all payments for the use of software were excluded.

Similarly odd is excluding payments for the use of a ship for international passenger or cargo transportation, as well as income treated by the recipient state as subject to a tonnage tax regime. There is an exclusion for international shipping income for the minimum tax under the GLOBE model rules, but it only covers income directly connected with international transportation of passengers or cargo.16 Tonnage tax regimes tend to cover a wider range of income, including, for example, income from offshore geo-exploration, pipe and cable laying, or dredging activities. Income from these activities is subject to GLOBE minimum tax, so it is difficult to see why it should be excluded from the STTR.

Significant uncertainty seemingly remains over the definition of “services”, particularly activities that are digitalized. The commentary states only that:

The term “services” should be interpreted in accordance with its ordinary meaning and should generally be interpreted to mean an action performed for the benefit of another person. The method of delivery is not relevant to the determination of whether income is received in consideration for the provision of a service.17

The unqualified use of the term “services” supports an interpretation that includes all kinds of services, including all automated digital services. However, the phrase “an action performed” may be argued to entail a human activity. This is the interpretation that has been given to tax treaty provisions allowing withholding taxes on fees for professional and technical services. These have been held to require human intervention, and consequently not to apply, for example, to digitalized advertising or streaming services, which are delivered automatically over the internet.18 While most automated digital services would also not be covered by the pillar 2 STTR because they are not between connected persons, some intragroup services can be automated. This can be seen in the OECD transfer pricing guidelines discussions of “low value-adding intragroup services,” in which it is evident that many of these can be supplied digitally with no human intervention.19 The clarification in the second sentence that the method of delivery is irrelevant would explicitly allow inclusion of human-to-human services delivered by digital means, but this may not include automated digitalized services, which may now be powered by artificial intelligence. This would be a significant limitation and necessitates further clarification whether “services” unquestionably include nonhuman actions.

III. The U.N. Model STTR

The U.N. STTR takes a very different approach.20 It would insert a simple general principle in article 1 of the treaty to reassert the right of states to tax income derived from their territory, notwithstanding other provisions in the treaty, if that income is not taxed by the treaty partner at the agreed minimum rate. Including this right for the source state would complement the long-standing principle — explicitly stated in article 1, paragraph 3 of both the OECD and U.N. treaty models since 2017 — that the treaty does not affect the right of the contracting states to tax their own residents (except under specified articles).

The provision states that the tax treaty concerned shall not affect taxation of income “arising in” a contracting party and “derived by” a resident of the other contracting party “if that income is subject to a low level of taxation” in the residence state, as defined in the article. Income is considered subject to a “low level of taxation” if (1) it is subject to a statutory tax rate at the agreed-upon minimum or less, or (2) although subject to a higher rate than the minimum, its beneficial owner is entitled to a “special exemption, exclusion or reduction” linked directly to the income or the receiving entity so that the tax rate paid is below the minimum.

The provision applies to all types of income for which the source state’s taxing rights are limited by the treaty, if “that income” is subject to a rate at or below the agreed minimum. Because there is no reference to specific categories of income, it applies to any “business profits,” avoiding the problem of defining types of income. This also avoids the difficulties caused by digitalization over the definitions of royalties and of services. Given the broad definition of “income” in tax treaties, it also covers capital gains, though countries may make this explicit if they wish.21

It also applies regardless of whether the income is paid to a related or unrelated entity.

There is no threshold or minimum level of income.

The tax can be charged at the time when the payments of income are made (e.g., by withholding); this is because the conditions of its applicability are designed so that they can usually be known at that time.

The minimum rate refers to each category of income and taxpayer concerned.22 Thus, the provision applies even when the general statutory rate is above the agreed-upon minimum if the state provides for a lower rate on either (i) certain categories of income, or (ii) taxpayers with certain characteristics or meeting certain conditions. For example, if a rate below the agreed-upon minimum is applied to income from exploiting intellectual property (e.g., because of a “patent box”), even if the recipient pays a higher rate on its total income, the provision could be applied by the source state to these payments.

The provision applies if the statutory rate is at or below the agreed-upon minimum,23 or the income is taxed below the minimum rate due to a “special exemption, exclusion or reduction” to which the beneficial owner is entitled.24 This condition is in lieu of any detailed rules for calculating an effective tax rate. It also applies prospectively to special regimes introduced after the treaty comes into effect.

Paragraph 3.b.ii entails a determination of what constitutes a “special exemption, exclusion, or reduction.”25 The commentary explains that this does not include deductions for “ordinary business expenses”26 that reduce the ETR, like deductions for interest, royalties, or salaries; also not included are deductions for depreciation or capital investments “in the relevant property,” even up to 100 percent in the first year. However, it does include extraordinary deductions, like allowances exceeding the actual expenditure (e.g., investment credits or deductions for depreciation that exceed the cost of a depreciable asset).

The scope for interpretation of what constitutes a “special exemption, exclusion or reduction” could lead to disagreements. The commentary suggests that treaty negotiators should discuss any aspects of their domestic law that provide such preferential treatment.27 Specific exclusions (or, presumably, inclusions) may be included by agreement in such cases. Some countries do not treat tax incentives implemented through treatment of the taxpayer’s costs or expenditures related to items of income as a special exemption and, in these cases, disagreements could be prevented by including specific clarifications. The OECD model’s commentary specifies in some detail the adjustments that should and should not be acceptable, and a mechanism for competent authorities to notify each other of relevant changes in domestic law.

Tax levied by the source state under the provision is “in accordance with” the treaty, so that the residence state must provide double taxation relief. The article does not apply if taxation below the agreed minimum on that item of income by the residence state results from a credit granted for tax levied under the treaty by the source state. The commentary provides language that could be included by agreement to cap the source country’s tax, ensuring that the combined tax imposed in the two countries (accounting for any foreign tax credit) does not exceed the rate envisaged by the withholding tax provided in the treaty.28 This reflects language that is explicitly included in the pillar 2 STTR.

The U.N. model STTR would grant a simple and broad right to the source state to tax income that the recipient country taxes at or below the agreed-upon minimum. It is left to the states concerned to agree on (1) the minimum rate applicable, and (2) any specific tax regimes that are not regarded as a “special exemption, exclusion, or reduction” and hence excluded from its scope. The commentary provides additional language that some countries may wish to include, and other variations may also be agreed to by negotiation.

IV. Implementation

Pillar 2’s and the U.N.’s STTRs have similar procedures for implementation: Both can be used as a model provision for bilateral negotiations, or may be adopted through a multilateral treaty for easy pairing with willing partner states. The principal difference between the two regarding implementation is the commitment made by those inclusive framework members that apply a rate of less than 9 percent to the categories of covered income to accept inclusion of its STTR if requested by a developing country. Joining the MLI would make this easy and automatic, without the need to ratify an amending protocol. However, countries should carefully consider whether by doing so they accept this much more limited provision as a new de facto global standard, rather than the broader and more comprehensive version developed by the U.N. committee, which should be important for both developed and developing countries alike.

A. The Pillar 2 STTR

For countries ratifying the MLI for pillar 2, the STTR would enter into effect for in-force treaties that have been notified to the MLI depositary as covered tax agreements by both contracting states.29 Participating states must identify and notify their covered tax agreements upon ratification or signature,30 and at least provide a provisional list on signature,31 though the list of notifications may be extended later.32

They must also notify any of the optional provisions in Annexes II-V that would apply to each covered agreement.33 Annex II specifies the method for calculating the tax rate on an item of covered income when the application of taxation is on a basis other than net income. Annex III specifies the method for calculating the tax rate on an item of covered income when a contracting jurisdiction imposes corporate income tax on the item at the point of profit distribution instead of when that income is earned. Annex IV provides a definition of “recognised pension fund” for the purposes of the exclusion of recognized pension funds from the STTR. Annex V provides a procedure for suspending the STTR if a contracting jurisdiction becomes classified as a high-income economy for five consecutive years; this implies that in the absence of such a notification the STTR continues in force.

In addition, there are obligations for the competent authorities of contracting states to inform each other in writing of:

(a) the source taxation provisions in a covered tax agreement that would be overridden (MLI article 10.6.a);

(b) items of covered income (interest, royalties, services, etc.) (MLI article 10.6.b);

(c) sourcing rules (MLI article 10.6.c);

(d) exemptions, deductions, or credits not subject to preferential adjustments (MLI article 10.6.d);

(e) certain income effectively connected or attributable to a PE (MLI article 10.6.e);

(f) double tax relief (MLI article 10.6.f);

(g) the materiality threshold (MLI article 10.7); and

(h) statutory tax rates on items of covered income, and provisions that may result in a preferential adjustment (STTR paragraph 1.5.c).

B. The U.N. STTR

The provision is designed as a model article which states can use as a basis when negotiating for inclusion in bilateral treaties. Because it is not prescriptive, negotiations can result in variations in the language and scope of the actual provision. It specifically leaves open for negotiation: (1) the minimum tax rate that triggers the source state’s right to tax, and (2) any exemptions that may be agreed to as applicable, like collective investment vehicles or pension funds.

However, it is also proposed to be included in the FTI being developed by the U.N. Tax Committee. The FTI will establish a matching procedure under which parties sign one or more of the protocols (including the STTR), to signify that they wish to include it in existing bilateral treaties and give the depositary a list of such treaties. If a protocol requires agreement on a tax rate or other terms (as for the STTR), states should also indicate in their notification list the rates (or a range of rates) or terms (or alternative terms) they may be willing to accept. The FTI’s secretariat will notify parties with matching notifications and assist and encourage them to conclude an amending agreement, which effectively amends the relevant treaty. An amending agreement may cover more than one protocol, and the draft FTI also provides for the possibility of a multilateral amending agreement if more than two parties agree to the same terms for one or more protocols. There is an optional enhanced procedure which allows parties to opt for automatic conclusion if the secretariat identifies matching notifications.

V. Evaluation

Although broadly aimed at the same issue, the two measures are clearly very different in both conception and operation. The pillar 2 version creates a measure that aims to be uniform, so it is defined strictly and in considerable detail. The U.N. model provision is very broad, with no limitations on scope, except those that may be specifically agreed to by the states concerned. The pillar 2 version is complete and ready for implementation, but on a take-it-or-leave-it basis, while the U.N.’s requires negotiation over the applicable rate (and perhaps other details) but provides more comprehensive protection for source taxation.

Both provisions authorize, subject to the conditions in each, source-state tax on income that would otherwise be restricted by the covered treaty concerned. This includes the general restriction on taxation of business income unless attributable to a PE (article 7 of the model treaties), as well as the more specific limitations for withholding taxes on interest, royalties, and services.

A major limitation of the pillar 2 STTR is that it is restricted to payments between connected persons. This severely limits its restrictions on profit shifting to payments within MNEs that already have a taxable presence in the source country. So, it does nothing to limit base erosion by avoiding the PE requirement, particularly through delivering services (digitalized or not) into a country from entities located offshore where the income is low-taxed.

However, it does include a targeted antiavoidance rule for payments routed through an intermediary that is subject to tax at a rate above 9 percent but that makes “all or substantially all” of these payments to a connected payee, including through “back-to-back” arrangements with a third party such as a bank. If these onward payments are deductible from the intermediary’s taxable income, there will be little taxable profit realized by the intermediary. The antiavoidance rule aims to ensure that the STTR applies based on the rate ultimately applicable to the payments in the hands of the connected payee. The commentary claims that this is a “mechanical rule,” but it includes the condition that it should be “reasonable to conclude that the intermediary would not have made the related payments in the absence of the original payment,” which entails a qualitative judgment. This rule may not be as effective as intended; for example, if the intermediary can be said to perform a function, like managing intellectual property rights, while retaining only a small share of the payments received as remuneration. However, it can be used in conjunction with the principal purpose test34 or a limitation on benefits provision if available. Countries that prefer the U.N. model STTR would also have to apply a principal purpose or limitation on benefits provision in these cases.

The Pillar 2 STTR is also restricted to specified categories of income, which raises significant doubts about its scope. Payments for the use of software may not be covered as royalties, and automated digital services may arguably fall outside the definition of services.

Both measures give source taxation rights that would take precedence over taxes applied by other states under the GLOBE rules.35 However, the pillar 2 STTR is designed to complement the GLOBE rules in a way that would preserve some of their priority for residence taxation.

The 9 percent rate in the pillar 2 STTR is both the minimum that the residence state should apply and the maximum that can be applied at source. It is set at 60 percent of the minimum rate specified by the GLOBE rules (15 percent). Hence, parent or conduit jurisdictions are entitled under the GLOBE rules to apply an additional top-up tax of up to 6 percent on profits that have been shifted out of the source country. The rationale some suggested for the STTR’s lower rate was that it applies to gross payments rather than net profits. However, the pillar 2 STTR does not apply unless the income exceeds the markup threshold based on local costs. This is similar to the substance-based income exclusion carved out in the GLOBE rules, which specify the “excess profits” on which a top-up tax can be applied. The markup threshold means that the STTR applies only if there is income exceeding a normal return on expenses, so it acts like a top-up tax on excess income, just as the GLOBE rules do. Yet the STTR’s cap of 9 percent on the gross payment can, in many circumstances, be significantly lower than the GLOBE rules’ minimum net rate of 15 percent on income exceeding the expenses attributable to the substance-based income exclusion.

The U.N. STTR’s broad base, and lack of either a “materiality threshold” or a “markup threshold” protects the source country’s right to tax income derived from that country much more effectively. It would be more beneficial even if the agreed-upon minimum rate were only 9 percent.

Clearly, the U.N. version is far more beneficial for taxation at source. It effectively implements the general principle that the ceding of taxing rights by source states — which is the core reason why countries sign tax treaties — is subject to the condition of sufficient tax in the residence state. It is also administratively much simpler, and applies to current payments rather than requiring the filing of a tax return, which could require some auditing, in the succeeding tax year. Its only disadvantage is that its adoption relies on the agreement of the states concerned. The pillar 2 provision has the advantage that those inclusive framework members applying a rate of less than 9 percent to the categories of covered income have made a commitment to agree to accept the provision in their treaties with developing countries. It also aims to resolve, through its complex and detailed rules, issues that might otherwise need to be dealt with by negotiations between treaty partners.

Keep in mind that most existing tax treaties of developing countries already provide for withholding tax rates as high or higher than 9 percent. An analysis by the IMF, done before the final version was published, says that:

the number of treaties that would be impacted by the STTR shrinks substantially when accounting for the nominal corporate tax rate in the recipient country. For interest and royalties, the top-up rate for the STTR is positive for only 26 and 27 treaties. For technical service fees, 101 treaties have room for additional STTR.36

There seems little point in a country joining the MLI if it has only a few treaties with low withholding tax rates — it would be better simply to renegotiate these treaties.

The main divergence is over the right to tax income from services. Pillar 2’s STTR would still deny the right to tax this income in the country in which services are paid for unless (1) the payments are to a connected person, and (2) the income exceeds the “markup threshold” of 8.5 percent over direct and indirect costs. As pointed out by the IMF, this issue is “of critical importance” to developing countries that are primarily importers of services.37 Taxation of cross-border services has been confirmed as a “specific priority issue” in the resolution on the promotion of inclusive and effective international cooperation on tax matters adopted on November 22, 2023, by a substantial majority of U.N. members in the Second Committee of the U.N. General Assembly.38

In our view, the pillar 2 STTR is unsuitable for developing countries because of its restricted scope and great complexity. Adopting this STTR would entail accepting continuing erosion of their tax base through payments that fall outside its scope, notably (1) payments to unrelated persons, (2) income that does not exceed the markup threshold, (3) income below the materiality threshold, and likely (4) all payments for the use of software. They should instead aim to safeguard their source taxation rights through (i) appropriate measures in domestic law and (ii) ensuring that all their treaties include the U.N.’s STTR, either through direct bilateral negotiations, or by joining the U.N.’s FTI when it becomes available (including its STTR protocol).

All countries should immediately review the operation of their existing treaties to identify any that are facilitating erosion of their tax base through low-tax payments in the recipient jurisdiction. These should be considered for renegotiation or cancellation. For developing countries that are expanding their tax treaty network, including the U.N. STTR should be a nonnegotiable element. Hopefully all countries, whether OECD members or not, will recognize the inherent fairness and equity that the U.N. STTR provides to all source countries, whether developing or not.

Countries should also consider introducing measures into their domestic law to deny deductions for payments that are made to jurisdictions where the income is low-taxed. Some countries have already taken these steps: for example, Australia’s measures tabled in June 2023 deny deductions for payments attributable to an intangible asset made to an associated entity if the income is subject to a low tax rate.39 Greater use could (and should) be made of antiabuse provisions in domestic law, in conjunction with the principal purpose test that should by now be included in all tax treaties, because it was a minimum commitment in phase one of the BEPS project. This justifies the denial of treaty benefits to entities that are not genuinely carrying out activities that generate the relevant income.40

FOOTNOTES

1 Including these provisions was agreed on as a minimum commitment under action 6 of the BEPS project and applied to members of the inclusive framework in 2016. It required, in particular, inclusion in treaties of a broad principle purpose test, optionally in combination with a simplified but still detailed limitation-of-benefits provision; or a detailed limitation-of-benefits provision supplemented by anti-conduit rules. Despite the action 6 measures, there is evidence that activity in conduits has scaled up, and any reduction in treaty shopping has been modest; see Antonia Hohmann, Valeria Merlo, and Nadine Riedel, “Multilateral Tax Treaty Revision to Combat Tax Avoidance: On the Merits and Limits of BEPS’s Multilateral Instrument,” University of Tübingen Research School of International Taxation Working Paper 10/22 (2022).

2 Notably Mauritius, which in 1992 adopted domestic legislation facilitating its use as a conduit, at the same time began negotiating treaties (particularly with African countries); a study for the World Bank estimated that this has caused losses of between 15 and 25 percent of corporate income tax revenues in those countries, with no counteracting benefits from increasing foreign direct investment; see Sebastian Beer and Jan Loeprick, “The Cost and Benefits of Tax Treaties With Investment Hubs: Findings from Sub-Saharan Africa,” International Monetary Fund WP/18/227 (2018). However, treaties with Mauritius have been cancelled by Senegal (2019) and Zambia (2020), and renegotiated by South Africa (2013), Rwanda (2013), and India (2016); Nigeria did not ratify the treaty signed in 2012, and Kenya also has not ratified the one it signed in 2019, largely because of pressures from civil society, including a court case. Additional examples are Mongolia’s termination of its treaty with the Netherlands in 2011 and the United States’ termination of its 1980 treaty with Hungary in July 2022 (see H. David Rosenbloom, “Hungary, Farewell,” Tax Notes Int’l, Feb. 12, 2024, p. 889).

3 See Felix Hugger et al., “Effective Tax Rates of MNEs: New Evidence on Global Low-Taxed Profit,” OECD Taxation Working Papers No. 67, at Table D.1 (2023).

4 Notably, article 23A of the model treaties conditions the obligation to grant an exemption on the source country exercising its right to tax.

5 See the draft commentary to the U.N. proposed STTR at para. 10, which refers to para. 20 of the introduction to the U.N. model, quoting para. 15.2 of the OECD model.

6 See the protocol to the 2012 Germany-Mauritius treaty, and that in the 2014 Denmark-Ghana treaty.

7 See our analysis in “The BEPS Proposals and Alternatives,” BEPS Monitoring Group (July 6, 2023); and Emmanuel Eze et al., “The GloBE Rules: Challenges for Developing Countries and Smart Policy Options to Protect Their Tax Base,” South Centre Tax Cooperation Policy Brief No. 35 (Aug. 18, 2023).

8 Estimates show that “in the likely scenario that no-tax and low-tax jurisdictions implement QDMTTs to capture the additional revenue domestically, investment hubs would retain $95 billion or 89 percent of total revenue gains,” and high-income countries 9 percent; F. Reitz, “Revenue Effects of the OECD Corporate Tax Reform — An Updated Impact Assessment of Pillar Two,” IFF-HSG Working Paper 2023-17, at 39 (2023). A recent study by the OECD, based on more granular data revealing pockets of low-taxed income in high-tax states (Hugger et al., “The Global Minimum Tax and the Taxation of MNE Profit,” OECD Taxation Working Papers No. 8 (2024)), suggests “broad gains across all jurisdictional groups, with higher gains for high income jurisdictions relative to lower and upper middle income” (para. 105; see also Annex B); however, all these categories include investment hubs, and they, particularly the high-income country hubs, would be the main beneficiaries of the GLOBE rules. While there may be some gains for low- and middle-income countries, better methods are available to them for protecting the source tax base; see Eze et al., supra note 7.

9 OECD, “Statement on a Two-Pillar Solution to Address the Tax Challenges Arising From the Digitalisation of the Economy” (Oct. 8, 2021). The commitment is made by members of the inclusive framework that apply nominal corporate income tax rates below the minimum rate of 9 percent to interest, royalties and the defined set of other payments. Developing countries are defined as those with a gross national income per capita, using the World Bank Atlas method, of USD 12,535 in 2019, to be regularly updated; this includes all but high-income countries. If a country later becomes a high-income country the STTR continues to apply unless the treaty partner has opted for Annex V of the MLI.

10 OECD/G20 Inclusive Framework on BEPS, “Tax Challenges Arising From the Digitalisation of the Economy — Subject to Tax Rule (Pillar Two)” (Oct. 11, 2023).

11 OECD, “Multilateral Convention to Facilitate the Implementation of the Pillar Two Subject to Tax Rule”.

12 The latest version is in the Committee of Experts on International Cooperation in Tax Matters, “Taxation Issues Related to the Digitalized and Globalized Economy,” Co-Coordinators Report (Annex 1), E/C.18/2023/CRP.40 (Oct. 2, 2023).

13 Committee of Experts on International Cooperation in Tax Matters, “Co-Coordinators’ Report: Proposal for the Inclusion of a General ‘Subject to Tax’ Rule in the United Nations Model Double Taxation Convention Between Developed and Developing Countries,” E/C.18/2023/CRP.12 (Mar. 10, 2023).

14 Citations are to the paragraphs of the model article in the Report of October 11, 2023, which correspond to the articles of Annex 1 of the MLI released on October 3, 2023.

15 It has now developed a revision of its model article to be included in the 2025 update, but the pillar 2 STTR will apply only to the version of the royalties article in existing treaties.

16 See the GLOBE model rules, at art. 3.3.

17 See OECD commentary, para. 33.

18 C.I.T., Delhi v. Bharti Cellular, Supreme Court of India, Aug. 12, 2010; Income Tax Officer, Kolkata v. Right Florists, I.T.A. No.: 1336/ Kol. / 2011, Apr. 12, 2013. Hence, the U.N. Model Convention now includes two separate provisions: Article 12A for Fees for Technical Services, and 12B for Automated Digital Services. An alternative version of 12A that covers “any service” is included in its Commentary (para. 26).

19 See OECD, Transfer Pricing Guidelines for Multinational Enterprises and Tax Administrations, at “Chapter VII: Intra-Group Services,” at para. 7.44 et seq. (2022).

20 See Commentary to the U.N. STTR, art. 1, para. 10.

21 Id., at para. 11.

22 Id., at para. 16.

23 See U.N. STTR, at para. 3.b.i.

24 Id., at para. 3.b.ii.

25 See Commentary to the U.N. proposed STTR, at para. 14.

26 Id., at para. 15.

27 Id., at para. 17.

28 Id., at para. 20.

29 MLI at art. 2.

30 Id. at art. 10.1.a.

31 Id. at art. 10.4.

32 Id. at art. 10.5.

33 Id. at art. 10.1.

34 U.N. model convention, at art. 29.9.

35 This is made explicit in the OECD’s STTR report, para. 43.

36 IMF, International Corporate Tax Reform 43 (2023).

37 Id. at 39.

38 United Nations General Assembly, “Nigeria: Revised Draft Resolution. Promotion of Inclusive and Effective International Tax Cooperation at the United Nations,” A/c.2/78/L.18/Rev. 1 (Nov. 15, 2023). Paragraph 5.e requests the ad hoc intergovernmental committee elaborating the draft terms of reference for a framework convention “to consider simultaneously developing early protocols . . . on specific priority issues, such as measures against tax-related illicit financial flows and the taxation of income derived from the provision of cross-border services in an increasingly digitalized and globalized economy.”

39 Australian government, “Multinational Tax Integrity — Denying Deductions for Payments Relating to Intangible Assets Connected With low Corporate Tax Jurisdictions,” Exposure Draft Legislation for Mar. 31 to Apr 28, 2023. It is limited to payments by “significant global entities,” though this limitation seems unnecessary and undesirable.

40 See Example G in the commentary to article 29.9 of the U.N. model (at 895 of the 2021 edition).

END FOOTNOTES

Copy RID