INSIGHT: Taxing the Digital Economy—Pillar One Is Not BEPS 2 (Part II)

The Unified Approach issued by the OECD Secretariat on Oct. 9, 2019, is the OECD’s most recent attempt to find international consensus on BEPS Action Item 1, “Taxing the Digital Economy.” Our assessment is that the Pillar One proposals in the Unified Approach suffer from several defects, the most important of which is their apparent abandonment of the arm’s-length principle. The overarching goal behind the 15 BEPS Action Items issued in 2013 was to strengthen the international tax system by removing egregious tax loopholes and ensuring that profits were taxed where economic activities occurred and value was created.

The Pillar One proposals, however, are not “BEPS 2”; they do not share the same agenda, do not build on international tax and transfer pricing principles, and will weaken not strengthen the current system. We believe that the current tax and transfer pricing rules, inclusive of BEPS 1 changes, can encompass the digital economy. We offer six policy recommendations designed to move the global economy onto the BEPS 2 path, a path appropriate for 21st century digital multinationals that will benefit both developed and developing countries. This is Part II of a two-part analysis of the Pillar One proposals. In Part I, we provided a summary and analysis of the proposals. In Part II, we examine their implications and provide some recommendations

A CRITIQUE OF PILLAR ONE AND OUR POLICY RECOMMENDATIONS

Pillar One Does Not Share the Same Agenda or Empirical Support as BEPS 1

Pillar One Is Not BEPS 2
A core question for stakeholders, in evaluating the Pillar One proposals, is to ask whether they are an integral part of the original BEPS reforms (i.e., BEPS 2 or BEPS 2.0) or if their underlying policy rationale is different. The earlier BEPS reforms, notably those reflected in the revised 2017 OECD Transfer Pricing Guidelines, were targeted at (1) eliminating tax loopholes, and (2) aligning profit allocation with value creation by improving (modernizing) the application of the arm’s-length principle (ALP).

Our brief history of the OECD’s attempt to build consensus on taxing the digital economy under BEPS Action Item 1 suggests that the general public and the media view Pillar One as a continuation of the earlier BEPS project and therefore as “BEPS 2.” The media portrays Pillar One as targeting the “tax dodging” behavior of digital MNEs (i.e., the first BEPS goal) and thus an extension of earlier BEPS reforms. (See for example Jim Tankersley, 2019, “Tech Giants Shift Profits to Avoid Taxes. There’s a Plan to Stop Them,” New York Times, October 9.) The Pillar One proposals also address the second BEPS goal, aligning profit allocation with value creation, which suggests a second reason for viewing Pillar One as BEPS 2.

We believe that Pillar One is not BEPS 2 and has a different agenda from the earlier BEPS reforms. Pillar One attempts to redefine (broaden) the conception of what constitutes value creation in relation to digital/non-physical business models and to design a new way to tax this “value creation.” The proposals essentially demand the inclusion of “external inputs” into the value creation attributable to a digital MNE, i.e., to tax currently “untaxed digital activity” such as remote sales of digital firms.

These new manifestations of value creation remain highly debatable. A key problem with the argument used to justify Pillar One, i.e., there are huge untaxed digital activities, is that the concept reflects a rather hazy and new understanding of what constitutes a “value contribution.” Rather, we view Pillar One as closer to “value capture disguised (masquerading) as value creation.” Whether there are untaxed digital activities depends on whether one accepts the underlying premises of the user participation argument, namely:

“[…] the idea that soliciting the sustained engagement and active participation of users is a critical component of value creation for certain highly digitalised businesses. The activities and participation of these users contribute to the creation of the brand, the generation of valuable data, and the development of a critical mass of users which helps to establish market power” (Public Consultation Document, para. 18).

The notion of “BEPS 2” therefore seems ill-conceived for describing the nature of Pillar One, which appears to be antithetical to the BEPS 1 value creation approach. Even if one subscribes to this new approach to value creation, it is important to recognize that the Pillar One proposals have moved a long way from the DEMPE concepts embedded in the 2017 OECD Transfer Pricing Guidelines. (See Lorraine Eden, Niraja Srinivasan and Srini Lalapet, Transfer Pricing Challenges in the Digital Economy: Hic Sunt Dracones? (Part I of II), 48 Tax Mgmt. Int’l J. 251 (June 14, 2019).)

Moreover, Amount A in the Pillar One proposals applies not only to digital MNEs but to all “consumer-facing” businesses (the scope of which still needs to be defined). BEPS Action Item 1 was about taxing the digital economy; Amount A, on the other hand, could tax almost all businesses. Amount B also is much broader, applying to all marketing and distribution activities in user/market jurisdictions. Not wanting to “ring fence” the digital economy appears to be the motivation behind Amounts A and B, but broadening the scope means that the proposals are no longer about “taxing the digital economy.”

Our view is different. The international tax and transfer pricing rules, once the 2017 BEPS reforms have been diffused and adopted through the Multilateral Instrument, are more than adequate as a general framework for taxing MNEs in the digital economy. If specific rules need to be written for a particular digital business model due to some unique facet of the business model, better to write specific rules to cover the exception (as the IRS is proposing for cloud computing), than to force the same rules on all or most businesses, which the Pillar One proposals recommend.

If Pillar One Is Not BEPS 2, What Is the Agenda behind Pillar One?
We argue that the political agenda for Pillar One had its birth in the combination of highly visible (and apparently highly profitable) digital MNEs, together with the desire of market/user jurisdictions to raise more tax revenues. This combination was the triggering factor behind, first, the rapid introduction of digital sales/service taxes (DSTs) and, second, the OECD/G20 attempts to craft a viable counter proposal to DSTs.

At the individual country level, the political agenda behind Pillar One varies across different constituencies. For example, the big emerging countries (e.g., India, China, Brazil) have huge actual and potential markets. While their per capita incomes are not high relative to developed economies, their numbers of users/consumers (actual and potential) are much larger. Thus, their “consumption power” offers a big tax base, one that they can tap with a general sales or value added tax—or a DST. Remote digital sales by foreign firms can be particularly attractive as a taxing opportunity to large market/user jurisdictions. The incentive to tax remote digital transactions should also be strongest in high-income market countries that attract remote digital sales but are not headquarters for digital MNEs—for example, European Union member countries. For them, DSTs look like an attractive income source and tit-for-tax retaliation from other governments is less likely.

Governments of such countries as the U.S. and Japan that are headquarters for digital MNEs, on the other hand, are unlikely to want DSTs, particularly when their own MNEs either have or are trying to build a dominant market position on a worldwide basis. Their goal is to contain the spread of other countries adopting DSTs that could target their digital MNEs. The Pillar One proposals offer a useful second-best alterative to DSTs, particularly if the policy focus on digital MNEs is removed. Thus, broadening Amount A to cover “consumer-facing” industries is an attractive option. Having a large, high-income market where there is little public support for introducing a federal general sales tax or VAT (e.g., the U.S.) also helps. Attempts to squash or at least limit the floor for Amount B on marketing and distribution activities in host jurisdictions would also benefit their MNEs.

Another constituency involves small and medium-sized developing countries that have neither large consumer bases nor are home to digital MNEs. Their ability to capture benefits from either the spread of DSTs or the Pillar One proposals (especially Amount A) appears limited. Moreover, small countries are price takers, not price makers, and thus bear the burden not only of their own taxes but also those levied by other countries. Thus, small, less resource-rich market jurisdictions are unlikely to win from either DSTs or the Pillar One proposals.

There is also an ideological agenda behind Pillar One that has been regularly voiced, primarily by academics and non-governmental organizations (NGOs), who want to replace the arm’s-length principle with global formulary apportionment (GFA). The Pillar One proposals have at their core a heavy emphasis on dividing the worldwide profits of MNEs into pieces that are then distributed among market jurisdictions (Amounts A and B) and host countries (Amount B) using elments of fractional apportionment methods. A second agenda behind Pillar One is therefore captured by the metaphor of the Trojan Horse: Pillar One provides a unique opportunity to get rid of the arm’s-length principle and bring in formulary apportionment through the back door.

In sum, we believe that a combination of politics and ideology is driving the Pillar One proposals. Pillar One is not BEPS 2 and has a different agenda. The OECD Secretariat needs to directly confront the question as to how and why Pillar One is different from BEPS 1. The political impetus to tax digital MNEs in market countries is understandable but it needs to be kept in a “reasonable” perspective. The political pressure faced by the OECD may also be somewhat self-inflicted. By not clearly delineating BEPS 1 effects from “untaxed digital activities”, the OECD Secretariat may have facilitated the general public’s perception that digital MNEs are engaged in tax avoidance in market jurisdictions.

The OECD Secretariat should clarify that Pillar One—at its core—has a completely different agenda: creating a “new taxing right” for user/market jurisdictions and opening the door to formulary apportionment. We are strongly against both items in this agenda, which we view as antithetical to the current international tax and transfer pricing system.

Pillar One Does Not Share the Same Empirical Support as BEPS 1
In addition, the lack of empirical foundation for the Pillar One proposals should demand a high level of caution. There simply are no advantages—and many dangers—to hastily constructing fundamentally new principles for taxing the digital economy, especially ones that are poorly targeted due to a lack of understanding of the nature of the challenges. We should therefore be careful, when evaluating whether BEPS 1 reforms such as DEMPE work in the context of day-to-day transfer pricing, to not apply an ill-defined “yardstick.” For example, to infer from BEPS 1 tax-gap estimates that taxing the digital economy presents a substantial challenge, at least at this stage of the digitalization of the economy, seems presumptuous and most likely erroneous.

It is also important to recognize that, at this point in time, there has been little empirical work evaluating the impact of the BEPS reforms implemented since 2017. It is simply too soon to know whether they will achieve their original goal of reducing/eliminating base erosion and profit shifting. Over time, as these studies become available and MNE strategies and structures adjust to the 2017 reforms, we will have a much better understanding of the remaining challenges (tax gaps) with respect to taxing MNEs and to taxing digital business models.

Amount A (New Taxing Right) Is Antithetical to Existing International Tax Principles

Amount A Introduces New and Untested International Tax Policies
Clearly the most revolutionary piece of the Pillar One proposals is Amount A. It involves three “game changing” policy changes—nexus, unitary allocation, and formulary apportionment—all of which are antithetical to the current tax/transfer pricing rules and with which countries have little to no practical experience at the subfederal level, much less at the international level.

• Nexus: The nexus rules under Article 5 (Permanent Establishment) of the OECD Model Tax Convention are to be modified using some (unspecified) new criteria. Under the new nexus rules, the parent firm P Co is deemed to have a new non-physical, taxable presence in Countries 2 and 3. A Virtual Permanent Establishment (VPE) is created in each market jurisdiction and its tax authority now has the right to tax the VPE’s income. In effect, Country 2 now has two permanent establishments, Q Co and the new VPE, which can be taxed separately or together (the latter should be more easily administered). Country 3 has now become a new “source” country by virtue of its VPE and can claim “first crack” source taxation rights for its VPE.
• Unitary Allocation: The profit allocation rules in Article 7 (Business Profits) and Article 9 (Associated Enterprises) of the OECD (and UN) Model Tax Conventions, which underpin the arm’s-length principle, are ignored. The ALP is replaced by “unitary allocation” where the MNE group is treated as a unitary enterprise and some portion of its worldwide profits are allocated to market jurisdictions. The MNE’s profits are aggregated in some manner, either as a whole or by business line, and then a series of calculations that are based not on the ALP but rather on arbitrary percentages are used to determine the aggregate residual profit to be allocated to market jurisdictions. Thus, Amount A contradicts the ALP in Articles 7 and 9.
• Formulary Apportionment: A formulary approach (to be determined) is used to allocate Amount A among the market jurisdictions.

Presumably the residence country (Country 1) must now make “tax room” for the new taxing right (Amount A) allocated to each VPE by providing a foreign tax credit or deduction for the new taxes being levied by Countries 2 and 3 on their VPEs. If Country 1 refuses, then double taxation will result.

Layering a new taxing right for user/market jurisdictions on top of the existing tax rights for source and residence jurisdictions, in effect, stacks one set of taxing rights on top of another. Adding a new taxing right, by definition, must be a zero-sum game since any new right must take away an old taxing right from another jurisdiction or jurisdictions. If the source or residence jurisdictions do not provide “tax room,” the clear result will be double taxation. For countries without extensive tax treaty networks, the message is clear: use the “new taxing right” and face double taxation in the short run and less inward FDI (foreign direct investment) and economic growth over the long run.

We believe that Amount A in the Unified Approach represents a clear and present danger to the ALP. Not only does Amount A include a new and undefined nexus rule, it also creates an allocation mechanism based on unitary taxation and formulary apportionment methods. If introduced, Amount A could be the Trojan Horse that spells the death of the arm’s-length principle and its replacement by global formulary apportionment. We therefore make the following two recommendations.

• Recommendation 1: The OECD Secretariat and Members of the Inclusive Framework should reaffirm their commitment to the current international tax rules (inclusive of earlier BEPS reforms) as the “bedrock” for taxing MNEs and their activities, both digital and non-digital.
• Recommendation 2: The OECD and Members of the Inclusive Framework should reaffirm their commitment to the arm’s-length principle under Articles 7 and 9 of the OECD Model Tax Convention and their rejection of formulary apportionment.

Amount A Ignores the Difference Between Exports and FDI
The corporate income tax (CIT) is a direct tax on the return to investments made by firms at home or abroad through FDI. The international tax rules are designed to tax MNE profits from engaging in FDI by setting up offshore branches or subsidiaries. Source and residence principles determine which country has the “first crack” at taxing foreign source income, together with rules to prevent double or under taxation.

There are other modes of entry that firms can used to go abroad instead of FDI, for example, by exporting or importing goods and services. However, when firms engage in international trade, their sales or purchases do not create a taxable entity in the trading partner country. The “tax menu” for the trading partner is a general sales/VAT and/or an export/import tax, not a corporate income tax levied on profits.

There may be cases, of course, where it is difficult to determine whether a firm’s offshore activities are export/import activities or rise to the level of FDI. To resolve this issue, the permanent establishment (PE) rules under Article 5 of the OECD Model Tax Convention are used to determine whether an entity has sufficient attachment or presence in a jurisdiction to create a taxable entity; in effect, whether the foreign activity is trade or FDI. If sufficient nexus exists, a PE is created and then taxed under Article 7 of the OECD (or UN) Model Tax Convention.

Amount A in the Unified Approach distorts the existing tax rules by creating a “new taxing right” for user/market jurisdictions. While the term “user/market jurisdiction” (the scope question) is unclear, the text suggests the new taxing right would apply to products imported from foreign firms, i.e., to “remote sales” or exports. Moreover, while the justification for the new taxing right is remote sales by digital businesses, the text of the proposal suggests a much wider scope that is “broadly focusing on consumer-facing businesses” (para. 15). Amount A therefore does not build on existing international tax principles. “Remote sales” are simply exports, not FDI, and should be taxed (if taxed at all) in the importing jurisdiction by a sales tax or VAT, not a corporate income tax. Moreover, basic international trade theory teaches that small countries (price takers) and immobile factors bear the burden of trade and tax policies. In effect, the tariff or tax is passed forward to consumers/users and not borne by the exporting firms. Thus, the incidence of the Amount A tax will be borne by small market jurisdictions and consumers, not by the digital MNEs.

We therefore agree with Schön’s (2019) assessment that the appropriate way to tax the exports of digital businesses is to determine whether their activities are simply export/importing activities or rise to the level of FDI. (See Wolfgang Schön, 2019, “One Answer to Why and How to Tax the Digitalized Economy”, Max Planck Institute for Tax Law and Public Finance Working Paper 2019-10, June.) If the former, there is no nexus and the CIT is not appropriate. Market jurisdictions that want to tax remote digital sales should do so with sales/VAT taxes.

If, on the other hand, the digital activities in the market jurisdiction rise to the level of FDI—the MNE is making remote investments in the local jurisdiction—there is the potential to find nexus under the post-BEPS Article 5 of the OECD Model Tax Convention (see below for details). ). If a VPE is created, its profits should be taxed according to the ALP as outlined in Article 7 of the Convention. In such cases, both types of taxes—a CIT on the VPE’s profits and a sales/VAT tax on its sales—can be levied by the market/source jurisdiction.

Amount A Should Be Replaced by Broadened Interpretations of Articles 5 and 7
Historically, Article 5 of the OECD (or UN) Model Tax Convention created a PE when there was a “fixed physical presence” tied to the export/import activity. Preparatory, ancillary, and warehouse activities (among others) were carved out and did not create a PE. The example in the Unified Approach is built on this “physical presence” assumption (see para. 41 on page 11).

The Article 5 PE rules were loosened under BEPS Action Item 7, by providing a broader range of export/import activities that could create a taxable PE. An anti-fragmentation rule in Article 5.4 was added to prevent MNEs from segmenting its activities in the host country into smaller and smaller functioning units in an attempt to avoid PE status. Article 5.5 added that if there is already a related party in the host country, for example, providing market services for the parent firm with respect to the parent’s sales in the host country, a separate PE could be created for the local sales. Article 5.6 added that even where there was no PE, the host country could look through and create a PE.

In 2018, the OECD provided Additional Guidance for applying these changes. Articles 5.5 and 5.6 are illustrated by Example 3 on page 19. The example deals with the remote sale of advertising on a website. This example is consistent with the Q Co case in the Unified Approach, which we have illustrated in Figures 1 and 2 in Part I. Based on the Additional Guidance provided in Example 3, as long as Q Co is regularly concluding the contracts with respect to the remote sales, then the market jurisdiction can create a separate PE for the sales activity. The activities of Q Co and the new PE are complementary and can be treated together as one entity for tax purposes (under the anti-fragmentation rule in Article 5.4). Example 3 then brings in the ALP to determine the price between the PE and the affiliate.

The text and examples in OECD (2018) suggest to us that the OECD could use the existing changes to Article 5 (Permanent Establishment) introduced under BEPS Action 7 as a blueprint for market jurisdictions. They could create a virtual PE when the remote digital business activities meet nexus requirements under Articles 5.4–5.7. Creating nexus in this manner, however, should be applied only where the remote activities of the digital MNE rise to the level of FDI in the market jurisdiction. Thus, the determination of nexus would remain a facts-and-circumstances matter that depends on the level of commitment of the digital MNE to the market jurisdiction.

Schön (2019, p. 28) recommends that nexus for remote digital activities should be determined using a two-prong test based on a minimum level of country-specific digital investment and the size of the local user base in the market jurisdiction. There are other possibilities also. See, for example:

• Economic presence test explored in Arthur Cockfield, 2002, “The Law and Economics of Digital Taxation: Challenges to Traditional Tax Laws and Principles”, International Bureau of Fiscal Documentation Bulletin, December, pp. 606-619; and Vishesh Dhuldhoya, 2018, “The Future of the Permanent Establishment Concept”, Bulletin for International Taxation¸ Vol.72, No. 4a (Special Issue)).
• Cloud computing—see the list of factors proposed by the Internal Revenue Service in August 2019 (see paragraph 1.861-19(c)) for determining whether cloud computing transactions are a lease of property or the provision of services;
• Factor presence tests for nexus (e.g., South Dakota v. Wayfair, Inc), explored in Monica Gianni, 2018, “OECD BEPS (in)Action 1: Factor Presence as a Solution to Tax Issues of the Digital Economy”, Tax Lawyer, Vol. 72, No. 1, pp. 255-298.

Once nexus has been established, the transfer pricing rules under Article 7 (Business Profits) of the Tax Convention should be used to determine the VPE’s taxable profits. We agree with Schön‘s (2019) assessment that the VPE’s profits for tax purposes should take account of the digital MNE’s return on its invested capital (i.e., the riskless rate of return, a risk premium, and any rents) in the market jurisdiction.

One concern that would need to be addressed is that under the AOA (Authorized OECD Approach) to Article 7, the PE is treated as a hypothetical separate entity and value creation is assigned based on where the significant people functions are located. The AOA would therefore create problems for VPEs, given that the digital activity is typically carried out from remote locations. However, most OECD countries have not adopted the AOA nor has the UN model. More work is therefore needed to determine how the ALP should apply in VPEs. We therefore argue that broadening interpretations of Articles 5 and 7 is the best way forward, one that builds on and extends existing international tax and transfer pricing principles to the digital economy.

• Recommendation 3: The OECD and Members of the Inclusive Framework should replace Amount A (New Taxing Right) with a proposal for a VPE, which covers digital business investments in market jurisdictions, builds on post-BEPS Articles 5 and 7 of the OECD Model Tax Convention, and expands the AOA approach to encompass VPEs.

Amount B Abandons the Arm’s-Length Principle in Favor of a Formulary Approach

The joint application of formulary apportionment and the ALP under the proposed Amounts B (based on the formulary approach) and C (based on the ALP) suggests a further “stacking” of taxation rights. Even if stakeholders were inclined to accept that the policy objective of allocating additional taxing rights to market jurisdictions is an inevitable political necessity, it is not at all clear that attaining this objective necessitates the potential abandonment of the ALP and its replacement by formulary apportionment.

The abandonment of a principled position on the ALP is most evident in the proposed calculation of Amount B, where a profit allocation mechanism is proposed that is alien and contradictory to the core tenets of the ALP. It appears that the OECD has succumbed (for no apparent reason) to the notion that the ALP is a failure and needs to be replaced—
“Moreover, there seems to be agreement that the arm’s-length principle is becoming an increasing source of complexity and that simplification would be desirable to contain the increasing administration and compliance costs of trying to apply it […therefore…] would introduce formula-based solutions in situations where tensions have increased—notably because of the digitalisation of the economy.” (Unified Approach, paras. 17 and 18)

If the OECD Secretariat, despite the contentious claims underlying this evaluation, officially believes that there is such a public agreement, it should take care to point out who is agreeing on what. In any case, such a vague and unsubstantiated claim or statement is hardly a sufficient justification for replacing the ALP with formulary methods, given the well-known theoretical and practical problems that plague formulary apportionment at the subfederal level and the lack of practical experience at the international level. (See, for example, the discussion in Lorraine Eden, 2019, “The Arm’s Length Standard Is Not the Problem”, Tax Management International Journal, October 2019.)

There is also some early evidence that the ALP can handle digital business models. Srinivasan, Lalapet, and Eden (2019), for example, provide a detailed analysis of one type of digital activity—the Internet of Things—and show that the existing transfer pricing rules under the ALP, while not perfect, can be applied to profit allocation within the MNE group. (See Niraja Srinivasan, Srini Lalapet and Lorraine Eden, “Transfer Pricing Challenges in the Digital Economy: A Case Study of the Internet of Things (Part II of II)”, Tax Management International Journal, 48, June 14, 2019.)

Moreover, with respect to Amount B, the question is whether the “tensions” relate to the application of post-BEPS ALP (including the DEMPE concept) or some “new and different” understanding of value creation. If the focus is on the latter, there is no reason to assume that the ALP would be unsustainable. The “new” concept of value creation (i.e., that value created “outside” of the MNE should be taxable in the market/user jurisdictions) is a purely “political” concept and is thus, in principle, subject to fostering consensus without diluting or abolishing the ALP.

The Unified Approach unfortunately fails to address this important aspect. Instead, the Secretariat assumes that, in addition to a change in nexus rules under Amount A, the rules on profit allocation for Amount B also need to be changed—
“[…] but also for those where there is [a nexus]. Otherwise, taxpayers could simply side-step the new rules by using alternative forms of an in-country presence (whether a local branch or related entity), making the new taxing right elective for taxpayers and creating an open invitation for tax planning” (Unified Approach, para. 16).

Where this notion of a new form of (presumably aggressive and not-to-be-countered-by-existing transfer-pricing-regulations) tax planning comes from is not clear. To consider that essentially the entire justification for Amount B hinges on vague and unsubstantiated notions is disheartening. The OECD should provide clearer arguments supporting this critical stance on the ALP. In its present shape and framing, the Unified Approach must be seen as a massive—and unnecessary—threat to the ALP.

• Recommendation 4: The OECD and Members of the Inclusive Framework should withdraw Amount B (which would therefore obviate the need for Amount C) and reaffirm their commitment to using the ALP for taxing the digital economy.

Adding Amount C on Top of Amount B Will Lead to More Tax Disputes

In the Unified Approach, the OECD Secretariat appears determined to put the worldwide residual profits of all consumer-facing businesses on the table. The MNE’s residual profits are subsequently to be allocated on a global basis (utilizing some form of formulary apportionment) away from residence to market jurisdictions in Amounts A and B. The promise to improve multilateral arbitration for Amount C is therefore likely to be little more than a placebo. Designing the Unified Approach in this manner, almost by definition, will result in double taxation (as any deviation from the baseline scenario of Amount B is likely to do) and thus create more interjurisdictional conflicts.

Whereas earlier international tax disputes were typically between a residence (home) country and a host (source) country, now three different types of jurisdictions will be vying to divvy up the MNE “profit pie”: residence, source, and market. Disputes among three parties are typically more difficult to solve than between two, simply because there are more opportunities for disputes.

First, there are likely to be disputes among the market jurisdictions. The totals resulting from Amounts A, B, and C clearly need to be calibrated in a way that the market jurisdictions (however defined) receive a high enough return to their VPEs to satisfy their demands for taxing the digital MNEs. If the return is not sufficiently high, the call for digital services taxes could again re-emerge. Moreover, the likelihood of disputes among the market jurisdictions as they attempt to influence the allocation key for apportioning the market jurisdiction return to the newly created VPEs in their jurisdictions seems very high. Even the identification of which jurisdictions are—and are not—market jurisdictions may be grounds for cross-border tax disputes.

Second are disputes between the residence jurisdiction and market jurisdictions. If the totals of A, B, and C are too high, the countries potentially surrendering their taxing rights will walk away from the table. While the simple example in the Appendix to the Unified Approach suggests that only one country (Country 1, the home country) loses taxing rights and taxable income, this will not necessarily be the case in practice. The likelihood that the MNE parent will declare the group’s returns from non-routine intangibles in the home country is not high. First, the income could have been earned outside of the home country so that other jurisdictions have “first crack” at taxing those profits. Second, as we have argued above, most residence countries have moved to territorial systems where they no longer tax foreign-source income so that MNEs have no incentive to declare intangible income in their home countries. Third, the mobility of intangibles together with holes in the international tax system created incentives for MNEs to park intangible income in tax havens, shifting taxable income and eroding tax bases. The BEPS project was designed to plug the most egregious of those loopholes. It is therefore highly probable that double taxation will emerge as market jurisdictions create VPEs and start taxing the virtual profits, but then cannot find a residence jurisdiction that will provide them with foreign tax credits or deductions.

A sneak preview of such distributional disputes may come from the reaction of the U.S. and its key investment locations to the Unified Approach. The top 10 countries where U.S. MNEs had direct investments abroad and declared non-U.S. profits in 2015 were the Netherlands, Ireland, the U.K., Luxembourg, Bermuda, Switzerland, the Caribbean Islands, Singapore, Canada, and China. To the extent that U.S. MNEs have shifted their intangible assets and income to these source countries, the new VPEs created by the Unified Approach should be looking for “tax room,” not only from the U.S. but from “primary source/secondary residence” jurisdictions for U.S. MNEs such as the Netherlands and Ireland, where the U.S. MNEs currently hold their intangibles and declare that income.

This could trigger disputes between the residence countries of the ultimate parent and the regional headquarters of an MNE group in terms of which jurisdiction gives up tax room. Many of the U.S. digital MNEs have their European headquarters in Ireland or the Netherlands. Their governments, particularly those in small countries, are unlikely to be in favor of shifting their tax base to market jurisdictions. It will be interesting to see the responses of Ireland and the Netherlands as well as small exporting countries and investment hubs, all of which seem likely candidates to walk away from the table if market jurisdictions begin creating VPEs and taxing formerly source-based income. (See, for example, Padraic Halpin, OECD Tax Overhaul Must Recognize Value Created in Ireland: Minister, Reuters Bus. News (Oct. 24, 2019), which quotes the Irish Minister of Finance as saying that “The significant and substantial value creating activity that happens in exporting countries like Ireland must continue to be recognized and rewarded.”)

Moreover, countries such as Germany, Japan, and South Korea may start to consider the potential long-term implications of an Amount B type of income allocation being applied beyond highly digitalized business models to their respective tax bases. Their domestic MNEs currently accumulate the bulk of residual profits at their headquarters, which is commensurate with the ALP. The prospect of having to share a substantial share of the profits realized by Daimler, Bayer, or Siemens with market jurisdictions may also be unwelcome. (See, for example, Lee Kyung-mi, OECD’s Digital Taxation May Expand to Target All Multinationals: South Korea’s Large Corporations Could Face Heavier Tax Burden, Hankyoreh (Nov.3, 2019).

In addition to these likely multi-country tax disputes, it is important to remember the institutional framework under which international tax laws apply. The international tax system is a network of double tax treaties, not trilateral or multilateral treaties. Many countries, especially developing countries, have very few double tax treaties. Treaty disputes often take years to resolve through existing dispute settlement procedures. Moreover, if some countries retain the ALP and others move to the Pillar One formulary apportionment approach for Amounts A and B, the resulting cross-border tax disputes are likely to lead to chaos and gridlock.

Our conclusion is that the way forward should not be through the Unified Approach—there are too many unknowns and too many likely pitfalls. Rather, the OECD should build on the BEPS 1 reforms. We make two proposals for moving forward, both of which build on existing international tax institutions and practices.

Our first proposal is to convene a working group to develop a chapter for the Transfer Pricing Guidelines on transfer pricing of digital transactions. The new chapter should parallel the existing chapters on goods, services, intangibles, and financial transactions. The chapter should also take account of differences between two-sided (buyer/user-seller/supplier) and multi-sided (platform) digital business models. In preparing this chapter, the OECD should work with economics and business faculty around the world who are studying these new digital businesses. In this way, the new field of digital economics can inform—and we believe strengthen and update—the ALP for 21st century MNEs. The new chapter should also reaffirm that the ALP remains the core principle guiding taxation for all related-party transactions. In addition, if areas are identified where the interpretation of the Transfer Pricing Guidelines is insufficient, treaty changes can also be recommended, but with a view to minimizing such changes.

• Recommendation 5: The OECD Secretariat and Members of the Inclusive Framework should prepare a new chapter for the OECD Transfer Pricing Guidelines on transfer pricing of digital transactions and businesses.

Our second proposal is to focus more on resolving disputes through the use of multilateral advance pricing agreements (APAs). APAs have long been used as the vehicle to bring the taxpayer and tax authority to a negotiated arm’s-length result, one that provides clarity and certainty for both parties. APAs not only build on existing international tax and transfer pricing principles but also offer flexibility and room for maneuver. They can also include old-fashioned bargaining tactics whereby a goverment that negotiates well can achieve new taxing rights for its jurisdiction. Multilateral APAs can bring willing parties to the table, eager to broker a win-win settlement. Bringing source, residence, and market jurisdictions to the table together with a digital MNE and hammering out a multilateral APA for the digital MNE’s business offers the opportunity for horse trading that can create both new taxing rights and negotiated arm’s-length results.

• Recommendation 6: The OECD Secretariat and Members of the Inclusive Framework should use multilateral rather than unilateral advance pricing agreements (APAs) as a dispute avoidance and settlement mechanism for taxing digital MNEs and activities.

CONCLUSION

From a conceptual perspective, the Unified Approach reflects an unsustainable “muddling through” approach to international tax reform, and the implied compromises are likely to have detrimental consequences for both the international tax system and the arm’s-length principle. Specifically, there is no conceivable scenario for limiting the scope of the “compromises” to the digital economy. The statement that the Unified Approach “[…] largely retains the current transfer pricing rules based on the arm’s-length principle but complements them with formula based solutions in areas where tensions in the current system are the highest” (Unified Approach, para. 15) amounts to wishful thinking.

The proposals also convey an overly pessimistic view of the ALP where the impact of BEPS reforms on MNE tax avoidance has been trivialized or underestimated. The OECD has apparently decided the ALP is not suitable (even post-BEPS) for ensuring that value creation and taxation are appropriately aligned. Even where a taxable nexus exists, this distrust of the ALP will inadvertently lead to second-guessing the efficiency of the ALP beyond digital businesses. Stakeholders should pause to consider whether the problems ascribed to the ALP (or the potential abuse of transfer pricing rules by MNEs) really necessitate such a broad re-shuffling of taxing rights and underlying paradigms.

If the Pillar One proposals are adopted, we believe the OECD will be trading a comparatively minor short-term gain (prevention of unilateral digital service taxes in market/user jurisdictions) for a potentially disastrous long-term consequence, namely, the erosion of the existing international tax system and the consensus on the ALP. The policy proposals—particularly Amount A with its changes to nexus, unitary allocation, and formulary apportionment—are ground-breaking but they are not based on international tax principles nor have they been tested as the international level.

We believe the proposals will lead to an over-politicization of international taxation, translating to more severe distributional conflicts between countries in the future. In such disputes it is the smaller, poorer countries and those without bilateral tax treaties that are most likely to suffer. Thus, we see few potential upside benefits and many possible downside risks to developing countries if they commit to these proposals.

We believe that the current international tax rules and the ALP are robust and flexible enough to handle 21st century multinationals. Our recommendations lay out a roadmap designed to move the global economy onto the BEPS 2 path, a path that we believe will benefit both developed and developing countries in their roles as source, residence, and market jurisdictions.

This column does not necessarily reflect the opinion of The Bureau of National Affairs, Inc. or its owners.

Author Information

Lorraine Eden is Professor Emerita of Management and Research Professor of Law at Texas A&M University (leden@tamu.edu). Oliver Treidler is CEO and founder of TP&C, a transfer pricing firm based in Berlin, Germany, and author of Transfer Pricing in One Lesson (2019, Springer) (oliver.treidler@tp-and-c.de). We thank William Byrnes, Niraja Srinivasan and others for helpful comments on earlier drafts; the responsibility for any remaining errors is ours. Last update: November 8, 2019. All rights reserved.