Transfer Pricing and the Arm’s-Length Principle After the Pillars

After years of discussion under the auspices of the OECD, 137 countries have now agreed to the outline of a package of important proposed changes to the architecture of the international tax system. The changes that have been agreed to in principle by the G-20, the OECD, and the inclusive framework are potentially complex and far-reaching. The two pillars of the agreement include: (1) the adoption of a form of minimum tax on global income; and (2) changes to the mechanism for allocating the income of multinational enterprises among taxing jurisdictions. Implementing the proposed changes will require a significant global undertaking over the coming months. That undertaking will provide opportunities for comprehensive analysis and comment once the technical details of the agreement are finalized and implementing instruments have been drafted. This article does not aspire to provide a comprehensive overview, critique, or analysis of the two-pillar proposal. Instead, it focuses more narrowly on the implications of the changes for the future application of the arm’s-length principle.

A key component of the proposed changes contained in the recently adopted agreement are modifications to the rules used to allocate income among the members of an MNE group and among taxing jurisdictions. For nearly a century, the arm’s-length principle has formed the basis for this allocation of income for tax purposes. The arm’s-length principle, which is enshrined in article 9 of the OECD model tax convention, requires related enterprises to transact among themselves using prices and other transactional conditions that would be applied by similarly situated unrelated parties — that is, the constituent entities of an MNE are required to transact on an “arm’s-length basis.” The new agreement does not do away with the arm’s-length principle. However, it supplements and modifies its application for the largest companies in the global economy, including the giants of the digital economy. In making these modifications to the operation of the arm’s-length principle, the agreement raises doubts about the continuing viability of, and the commitment of countries to, the arm’s-length principle as a foundation stone of the global tax system.

Our discussion in this article of the consequences of the recent agreement for the application of the arm’s-length principle will proceed in four steps. First, it will summarize pressures on the arm’s-length principle that have been building over recent decades. Second, it will consider the modifications to the arm’s-length principle established in the OECD’s base erosion and profit-shifting project as well as the perceived successes and failures of those changes. Third, it will summarize the reasons for the recent shifts in thinking about the arm’s-length principle that are evident in the pillar 1 agreement. Fourth and finally, it will consider the likely trajectory for the future use and application of the arm’s-length principle.

Pressures on the Arm’s-Length Principle

While the arm’s-length principle has been used since the 1920s, at least four distinct types of pressure on the traditional arm’s-length-principle-based income allocation system have been widely discussed in recent years. These pressures are not new. They have been considered, analyzed, and argued about over many years. In fact, these pressures were central to the concerns about the international tax system that gave rise to the BEPS project a decade ago. The BEPS project sought to address these pressures but, for a variety of reasons, did so with only mixed success. As a result, the same pressures continue to be relevant and have played an important role in motivating governments to move forward with the new two-pillar agreement.

Mobility

The first of the four identified pressures relates to mobility, and in particular, the mobility of income-producing assets, risks, and business functions of an MNE group.

The arm’s-length principle operates on a separate legal entity premise, which provides that income among separate related entities within the MNE group, and among taxing jurisdictions, is allocated based on the geographic location of the functions, assets, and business risks undertaken by the various members of the MNE group. Within many MNE groups, business functions can often be readily shifted from one legal entity to another and from one geographic location to another without a material detrimental impact on the MNE group’s overall business efficiency. Similarly, valuable income-producing intangible assets can often be shifted among group members with little tax cost or business disruption. Also, the incidence of some business risks can sometimes be shifted among members of the MNE group by, for example, carefully crafting contracts and other legal arrangements to assign the bearing of risk to locations where it may be most advantageous from a tax point of view.

With careful planning, the mobility of functions, assets, and risks, coupled with the separate-entity hypothesis that underlies the arm’s-length principle, often allows MNEs to shift substantial amounts of income. For example, MNEs can often locate in separate legal entities bearing low or limited risk important business functions that must be carried out in higher-tax jurisdictions geographically close to customers or supply sources. At the same time, mobile intangible assets and mobile risk exposure can be shifted to entities taxable only in low-tax jurisdictions. Such arrangements have the effect of separating business risk and intangibles from the high-tax entities performing business functions. Because risk bearing and intangible ownership should be compensated under arm’s-length principles, this separation results in group entities that are subject to tax in high-tax jurisdictions being able to limit their income to low or routine levels. Entities that are assigned a greater share of risk and the ownership of intangible assets have their operations and the income attributable to such intangibles and risks isolated in low-tax jurisdictions. The result can be a significant reduction in MNE group tax liability. While casual observers might perceive an MNE group as operating on a highly integrated basis across the legal boundaries separating individual group members, the diversion of mobile risks and intangibles into separate legal entities leads to a favorable allocation of taxable income and gives some MNEs the ability to argue that the arm’s-length principle properly allocates a high share of MNE group income to isolated, low-tax jurisdictions where little functional business activity takes place.

Heavy taxpayer reliance on such “limited risk/residual profit” arguments has been frustrating to many tax authorities. In the perception of many such tax authorities, the possibility of shifting large amounts of mobile income to low-tax jurisdictions has been affirmatively worsened by accommodating countries that are willing to structure their tax systems in ways that permit MNE income shifting in exchange for attracting employment or financial activity to that jurisdiction. While the relevant business strategies and arguments underlying this type of MNE income shifting are well understood, it has proven difficult to challenge common tax planning strategies involving income shifting because of the fixed separate-entity principles underlying the arm’s-length principle and the collaboration of countries pursuing tax competition strategies. This has, in turn, caused some observers to conclude that, in relation to income shifting, the arm’s-length principle may be more a part of the problem than the solution.

The Interaction of Nexus Rules

The second pressure relates to the interaction of the arm’s-length principle and nexus rules.

Within the international tax system, the separate-entity approach of arm’s-length-principle-based transfer pricing interacts closely with the notion that for an entity to be subject to tax in a jurisdiction on its business income, that entity must have sufficient contact with that jurisdiction. Under prevailing treaties, nexus determinations regarding business income are generally based on the entity having a physical presence in the jurisdiction in the form of an office or other fixed place of business.¹ When that type of physical presence does not exist, any business income allocated to the entity under the arm’s-length principle will not be subject to tax in the jurisdiction even if that income arises from sales to customers in that jurisdiction or from remote business activities directed at that jurisdiction.² Given these rules, tax arrangements are often structured to allocate income to entities that were specially designed to lack nexus with high-tax jurisdictions.

While nexus requirements based on physical presence in the jurisdiction may have seemed appropriate in the past, technological advances in the evolving economy have made a physical presence-based nexus standard seem less appropriate. Technological innovation in telecommunications, the expansion of the internet, and the expansion of rapid delivery capabilities have made it much easier to sell to customers in a particular country without the selling entity having a physical presence in the country. When combined with the mobility-related pressures discussed earlier, the prevailing nexus standards have made it possible for MNEs to locate large portions of their income in low- or zero-tax countries.

The resulting erosion of the tax base in the countries where products are marketed or produced has led some tax administrations to conclude that the mobility-related pressures on the tax system can be more easily addressed by expanding the reach of nexus rules rather than by investing enforcement resources in the arm’s-length-principle-based income attribution system. As a result, relatively arbitrary forms of taxation, including digital service taxes and diverted profits taxes, have become more commonplace as substitutes in attempts to fix the problematic results when applying the arm’s-length principle and the existing nexus rules.

Complexity

The third pressure relates to the complexity of applying the arm’s-length principle.

Regardless of one’s view of whether the arm’s-length principle represents an adequate policy response to the problems of income shifting, most observers readily admit that the application and administration of the arm’s-length principle is in practice very complicated. The arm’s-length principle is a fact-based system for allocating income. Its accurate application and enforcement require tax administrations and MNE compliance personnel to develop a detailed understanding of the business operations of the MNE and the many factors that may contribute to the generation of income by the MNE.

As the arm’s-length principle has evolved, increasing emphasis has been placed on the location of various business functions. For example, the nature and location of the management of risk can now play an important role in the geographical allocation of the income attributable to risk bearing. The location of management and decision-making functions regarding research and development activities can play an important role in the allocation of income related to intangibles. Finally, the determination of whether various functions are routine in nature or are instead high-value-adding functions, justifying commensurately higher prices, plays an important role in determining the amount of arm’s-length compensation for those functions.

All such determinations can be complex and fact intensive. They require detailed knowledge of the operations of the MNE, often across many countries. Moreover, because application of the arm’s-length principle usually requires a detailed comparison of the operations of the tested party with the operations of comparably situated independent enterprises, the detailed fact-finding must generally be extended beyond the taxpayer itself to include potential comparables. It is also often observed that, while examination of the operations of comparable enterprises is a theoretical necessity under the arm’s-length principle, it is often the case that true comparables are difficult to identify. Moreover, comparability adjustments are often essential, adding further complexity to arm’s-length principle analysis.³ Particularly in developing countries, it is often asserted that appropriate comparables for an arm’s-length principle analysis simply do not exist.

Tax administrators often observe that access to the facts necessary for a proper arm’s-length principle analysis is controlled by the taxpayer. It can, therefore, be difficult or impossible for the tax administration to gain access to the required facts. Moreover, taxpayer documentation of transfer pricing decisions may not satisfy the information requirements of tax administrations, thus making fact-finding and the development of required facts burdensome and complicated for both taxpayers and tax administrations.

Uncertainty

The fourth pressure relates to uncertainty.

Thus far, regarding the challenges of the arm’s-length principle, we have observed: the factual complexity of the arm’s-length-principle-based system; its failure to put an end to income shifting and resulting low tax rates of many MNEs; and the substantial pressures the system places on tax administrations to aggressively audit transfer pricing issues. The uncertainty surrounding the application of the arm’s-length principle places steep demands on what are often understaffed and inexperienced tax auditors, especially in developing countries. Such auditors may take positions in cases that can be difficult to justify. Moreover, the fact that audit adjustments often lead to claims that the tax base in other countries should be reduced or that taxpayers should face very large adjustments in the auditing country leads to extended tax controversy both between taxpayers and tax administrations and among interested countries.

Dispute resolution mechanisms are complex and demand extensive time and resources. Local country audits and ensuing litigation can drag on for many years. The competent authority process is often very time-consuming and costly. At all stages in the controversy, it can become clear that taxpayers, tax administrations, and different countries have very different ideas about how the arm’s-length principle should be applied and what the appropriate outcomes might be. Moreover, there is no guarantee that the outcomes of audits, litigation, and competent authority controversies will be satisfactory from the tax administration’s viewpoint. Indeed, because of their control of the relevant facts, their familiarity with the business, and the fact that, in high-stakes audits, taxpayers often are willing to commit extensive legal, economic, and expert resources to the controversy, taxpayers are often able to successfully resist audit adjustments.

This uncertainty of application and outcome can be problematic for tax officials who can often find themselves faced with the need to commit significant resources without assurance that success can be achieved. This uncertainty, in turn, can deter enforcement efforts and encourage taxpayers to take aggressive positions from the outset. Some tax administrations have concluded that arbitrary, but more readily applied, rules might be a better approach than the extended factual conflict baked into the arm’s-length principle system.

BEPS Successes and Failures

The pressures on the arm’s-length principle were well understood when the work on the OECD/G-20 BEPS project commenced. The BEPS action plan included work on measures it was hoped would address the issue of income mobility. These measures included specifically the ongoing work on the transfer pricing treatment of intangibles (including the consideration of measures to limit tax-motivated transfers of intangibles), transfer pricing treatment of risk, and nexus issues (including specifically the definition of permanent establishment under article 5 of the model convention).⁴ The action plan also called for the development of reporting and transparency measures, which were aimed at providing tax administrations with improved insight into taxpayers’ global allocations of income, thereby reducing the complexity of transfer pricing enforcement and the unpredictability of enforcement outcomes.

Importantly, the action plan authorized country delegates to consider specific measures going beyond the arm’s-length principle as they sought to develop measures to address the perceived income-shifting shortcomings of the arm’s-length-principle-based income attribution system.⁵ One key component of the BEPS work involved addressing the difficult income-shifting issues arising in connection with the digital economy.⁶ The intent was to include income attribution and nexus issues that were recognized to be especially difficult problems in a digital economy context.

While the objective of addressing these pressures on the arm’s-length principle in the BEPS project was evident in the action plan, the work that ensued in the following three years was only partially successful in finding effective responses to the recognized pressures on the transfer pricing system. The final BEPS reports did include new transfer pricing rules addressing the treatment of intangibles and the treatment of risk under the arm’s-length principle. This new guidance in the OECD transfer pricing guidelines resulted in more limitations on the ability of taxpayers to shift mobile intangibles and risks among MNE group members. The new rules emphasized the importance of examining the location of relevant business functions in determining where, within the MNE group, risks and intangibles should be located for transfer pricing purposes. Also, the development and adoption of country-by-country reporting rules improved factual transparency and the enforceability of transfer pricing principles. While the implementation of the CbC reporting system is still in its early stages, there is some reason to believe that improved transparency and better tax administration access to data will help address some of the uncertainty associated with the enforcement of the transfer pricing rules.

Notwithstanding these successes, the BEPS work also had several shortcomings. The changes to the transfer pricing guidelines substantially increased the complexity of transfer pricing analysis by placing much greater emphasis on detailed functional factors. Thus, the business functions associated with controlling risk became central to the determination of where within the MNE group various business risks should be deemed to reside. So-called DEMPE (development, enhancement, maintenance, protection, and exploitation) functions became critical in examining where in the group returns to intangible assets should reside. The changes in the guidelines failed to provide clear guidance as to where capital and the returns associated with capital investment should be attributed for transfer pricing purposes, and how arm’s-length returns to capital were to be measured. In particular, BEPS did not resolve, and indeed tended to highlight, glaring inconsistencies between the treatment of capital for MNE groups of corporations, on one hand, and groups operating primarily through branches, on the other. As a result, capital and returns to capital tended to follow key functions in the case of branches but did not necessarily follow functions or other mobile assets in the case of separate legal entity structures.

As a byproduct of the BEPS transfer pricing work, the term “value creation” took on added significance. But there was a less than adequate description of what the term means, and how the place of value creation is to be determined. Early attempts to apply the notion in audits and competent authority cases since the BEPS changes to the transfer pricing guidelines have reflected quite different understandings among country tax authorities as to the meaning and importance of value creation, as seen in individual cases.⁷

These shortcomings in the BEPS modifications of the transfer pricing rules were exacerbated by other shortcomings of the BEPS work. In particular, the BEPS work on controlled foreign corporation rules⁸ was generally perceived to have been ineffective in bringing about change. And importantly, while some improvements were made in the mechanisms for resolving disputes, those improvements fell substantially short of creating a system in which taxpayers and tax administrations could be confident that transfer pricing controversies would be resolved expeditiously and with confidence that consistent and principled outcomes could be achieved. In particular, the new guidance fell well short of a general institution of treaty-based arbitration.

Moreover, the BEPS outcomes left significant doubt whether the OECD countries were fully committed to the arm’s-length principle as the basis for income attribution in the international taxation framework. Some of the shortcomings in the transfer pricing work were perceived to have arisen because several countries had steadfastly affirmed their support for continued adherence to a separate-entity, transactional approach based on the arm’s-length principle. While the action plan had urged delegates to consider measures going beyond the arm’s-length principle to address the problem of income mobility and profit shifting, in the end there was little support for abandoning transactional arm’s-length-principle-based mechanisms and adopting special measures, even in narrow circumstances.

In other cases, however, solutions adopted in BEPS specifically rejected arm’s-length-principle-based approaches. The work on interest in action 4 rejected an arm’s-length principle approach in favor of a fixed ratio approach with little explanation (although noting the concerns of some countries about the effectiveness of the arm’s-length principle in preventing base erosion and profit shifting).⁹ In the harmful tax practices work on action 6 related to intellectual property boxes, a nexus approach was preferred, and an arm’s-length principle/transfer pricing approach was rejected.¹⁰ In the digital economy space, the final BEPS report addressing the digital economy failed to consider income allocation and related issues, leaving open the question of whether the arm’s-length principle was up to the task of attributing income among the constituent members of MNE groups engaged in digital commerce. Thus, there was a substantial question when the BEPS reports were issued whether the arm’s-length principle continued to have the full support of the countries participating in the BEPS work.

Changing Views Following BEPS

Following the completion of the BEPS reports, several developments led some countries to modify their views regarding the possibility of making changes to the architecture of the international tax system. Most importantly, work continued on the problem of the digital economy. It became clear through the ongoing OECD efforts that countries had very different views as to how the tax problems associated with the digital economy should be addressed. Some countries became more impatient with the slow pace of that work, repeatedly expressing the view that the work was urgent and that the low tax rates of some of the larger participants in the digital economy reflected a systemic failure of the international tax rules. There were strongly asserted views that, because of the mobility and nexus issues described above, so-called market countries were being denied an appropriate share of the global tax base.

As impatience grew, some countries began to adopt and implement unilateral DSTs.¹¹ Other countries, however, thought that the issues to be addressed were broader than just the large companies in the digital space, and that the scope of any revised measures narrowly targeting large digital enterprises would need to be broadened. Moreover, strong positions were articulated by some countries that DSTs and similar unilateral measures contravened existing trade and treaty obligations. Other countries expressed the view that the real problem in the tax system was that countries had never forged a unified commitment to deal with tax competition. The view was that, if an enforceable minimum tax regime could be developed, most other problems, including those arising in the context of the digital economy and transfer pricing, would be easily resolved.

Two other related events also had an impact on these discussions. First, the European Commission’s efforts to address shortcomings with transfer pricing enforcement under the auspices of state aid rules were not immediately successful.¹² Second, the U.S. Congress adopted in 2017 significantly modified rules for taxing the foreign income of U.S.-based MNEs, including the U.S.-based digital giants. While those rules were far from perfect, they gave the U.S. Treasury a sense that there might be useful avenues for protecting the United States’ right to tax digital economy giants and other large MNEs by means other than pure transfer pricing enforcement. This created an opening to discuss more openly rules that would limit profit shifting that went beyond the traditional view of the arm’s-length principle. At the same time, the U.S. tax reform legislation expressly considered, but then rejected, a pure destination-based taxing system.

Starting with informal discussions among some of the larger economies, countries began to consider whether meaningful change could be considered on a broad multilateral basis. It was clear to those involved in the discussions that any multilateral agreement would need to contain a variety of measures to satisfy all the interested countries. More specifically, the view was that an agreement would have to contain at least the following features:

• An agreement would have to grant greater taxing rights to market countries, particularly over companies operating in the digital economy, to address the politically sensitive issue of whether the digital giants were escaping market country taxation altogether. Notably, on this point, it quickly became evident that there would not be support for a general move to a pure destination-based tax system.

• An agreement would have to go beyond the digital economy so that large, profitable, non-U.S. companies in industries outside the digital economy (such as the pharmaceutical sector) would not be able to largely avoid market country taxes.

• An agreement would have to address tax competition and have the consent and participation of key low-tax countries.

• It was recognized that the tax system resulting from an agreement would have to retain, to a large degree, elements of the existing international tax architecture. In particular, there was a surprisingly strong view that the arm’s-length principle works in many cases and that it should be retained as an important building block of the system going forward. There was no enthusiasm for a wholesale conversion to a formula-based income attribution system. In addition, there was wide agreement that any new tax rules would need to adopt a “net basis” approach — that is, they would need to be based on the taxation of profits, not revenues.

• An agreement would have to be acceptable to a very large majority of countries and would have to obtain the support of smaller developing countries.

After numerous discussions among the larger economies, the likelihood of success seemed high enough, and the potential consequences of further widespread unilateral action seemed sufficiently dire, that it was deemed possible and appropriate to begin discussing the outlines of possible agreements with all interested countries through the auspices of the OECD, G-20, and inclusive framework.¹³ A two-pillar formula for change satisfying the aforementioned considerations was put on the table for discussion and, after more than two years of negotiation and modification, the general outlines of an acceptable approach were agreed.¹⁴

The Pillar 1 Agreement

The pillar 1 proposal, to which the inclusive framework has given its approval,¹⁵ is a complex and detailed set of rules.¹⁶ Considerations of space permit only a rudimentary summary of its core features here.

The pillar 1 measures are intended to apply to the largest MNE groups with global revenues of more than €20 billion. The market states in which these groups do business are intended to continue to receive their proper allocation of income based on the arm’s-length principle.

Pillar 1 introduces an additional allocation of taxable profits, which is 25 percent of MNE group profit over a profit threshold of 10 percent, determined by reference to consolidated group financial accounts. This measure of profit is known as amount A and, broadly, is allocated to market states in proportion to a sales factor that measures sales in any market state over total sales.¹⁷ The reallocated profit is taken away from the arm’s-length principle system through a mechanism for the elimination of double taxation. This is achieved by identifying from within the MNE group “paying companies” with residual profits and treating them as having paid the amount A tax levied by the relevant market states.

These new measures are not limited to MNE groups carrying on digital business but apply to MNE groups that satisfy the revenue threshold, and have an accounting profit margin (profit before tax and revenue) exceeding 10 percent. (There are exclusions for MNE groups whose business is concerned with natural resources and financial services.)

The pillar 1 measures also include streamlined dispute resolution mechanisms — both to prevent disputes under the new amount A system and to resolve disputes under the existing arm’s-length principle system. Separate measures are contemplated to simplify the application of the arm’s-length principle to marketing and distribution activities (this is known as amount B). States are required to remove DSTs as a mandatory part of the pillar 1 package.

Work is ongoing on several important features of pillar 1. In particular, important issues concerning the interaction of the new pillar 1 rules with the existing arm’s-length-principle-based system remain to be resolved. This includes work on the elimination of the double taxation mechanism, namely how and where pools of residual profit are to be recognized so that the relevant primary taxing rights over such profit can be granted to market states and taken away from the states to which such profit would be allocated under the arm’s-length principle rules.¹⁸ It also includes the workings of a potentially complex “marketing and distribution safe harbor.” The idea behind this feature of pillar 1 is that if there is already, under the arm’s-length principle rules, an allocation over a given level of profits to a market state, this should be reflected in a downward adjustment to amount A. The mechanism is likely to put further pressure on the interaction between the two systems.

In addition to pillar 1, there is also of course the pillar 2 package of measures, which means that pillar 1 is paired with a complex 15 percent minimum tax regime.¹⁹

From the foregoing brief description, it is evident that the new system will be a hybrid one of income allocation, with two different tax bases and two different allocation methods. The arm’s-length principle continues to apply to all companies and virtually all income, but the arm’s-length-principle-based allocation system is supplemented with a formulary override for a relatively small number of large, highly profitable companies. This makes the ongoing work to integrate the two systems of critical importance.

Implications for Now and the Future

We now turn to the question of what the matters addressed thus far mean for the current and future standing and operation of the arm’s-length principle. Some observations on the position of the arm’s-length principle may be readily drawn.

The arm’s-length-principle-based transfer pricing and PE attribution rules will remain the core mechanism for the allocation of corporate income. Indeed, for MNEs in the financial services and natural resources industries, and for all MNEs with consolidated accounting profit margins lower than 10 percent of revenue, the arm’s-length principle will be the exclusive income allocation mechanism. This means that all the current technical and practical problems of the arm’s-length principle remain with us, given that considerations of specific solutions relating to the arm’s-length principle have not been a feature of the OECD and inclusive framework’s work on digitalization and the two pillars.

It is harder to comment on the likely future direction of these measures. This difficulty exists for two main reasons. First, there is material uncertainty concerning how the two-pillar package will play out over the next two to three years. The uncertainty relates to questions such as the nature of the solutions to be found under the pillar 1 rules for difficult or contentious issues like the elimination of double tax, the marketing and distribution safe harbor, the tax certainty and dispute resolution package, and so forth. Also, there are broader questions relating to the two-pillar package, such as a host of questions concerning the process of adoption and implementation, whether pillar 2 might be decoupled from, and introduced earlier than, pillar 1, whether the U.S. Congress will approve the pillar 1 measures, and so forth. This means several other alternative future scenarios are possible.

A second reason for the difficulty in predicting future outcomes stems from the materially increased politicization of the process. The sharply increased political attention the OECD and inclusive framework’s reform efforts now garner can be seen through a comparison of the pre-BEPS position and the current one. Over this period, there has been a huge increase in political interest in the reform of the international tax system.²⁰ The practical impact of this increased political visibility is to make the policy and technical agenda less predictable, given the impact of political compromises and trade-offs that may affect the direction of the work.²¹ This in turn means that what might seem obvious (to the authors, at least) as the required policy or technical direction that needs to be pursued may be the last thing that is considered.

Immediate Priorities

Notwithstanding these difficulties, two immediate priorities arise from the current position, which is shaped by the agreement on pillars 1 and 2. These are: (1) the need to ensure the hybrid system can work effectively, especially in connection with the required interactions between the new destination system and the arm’s-length-principle-based system; and (2) ensuring that the arm’s-length principle can function properly within the new hybrid system.

Regarding the first point, there are various issues to consider but the most critical point will be the interactions between the two systems — and some of the most difficult issues yet to be resolved in the ongoing work on the detailed design of pillar 1 are central to this interaction issue. These include resolution of the elimination of the double tax issue (specifically, clarifying the rules for identifying the “paying entities” that will be deemed to have paid the amount A tax in the destination countries). Another outstanding design issue concerns whether and how to include the marketing and distribution safe harbor mechanism in the pillar 1 package. In particular, clear articulation of the circumstances in which amount A should be reduced because of an already adequate allocation under the arm’s-length principle will be essential. The design of this mechanism is important because it introduces a much more complex set of interaction issues between the new market state reallocation and the arm’s-length principle system. Both points raise several very difficult technical issues and extremely challenging political ones.

The second point, about ensuring that the arm’s-length principle-based rules can function properly within the new hybrid system, seems to require an appropriate “fixing” of the problems besetting the arm’s-length principle. Any such fixing seems to require the clearing of at least two hurdles. First, dealing credibly with known arm’s-length principle problem areas that undermine confidence in the transfer pricing or PE attribution rules (examples include matters such as the allocation of risk, the attribution and reward of capital, and some sort of reconciliation of the rules applicable to subsidiaries with those applicable to branches²²). Second, it would be necessary to deliver responses to these problems that are acceptable as an adequate solution to a significant majority of states. The arm’s-length principle is not fixed if a solution is proposed but a sizable part of the world disagrees with its viability or effectiveness.

The task of finding viable fixes to address the problems of the arm’s-length principle is immensely challenging. It requires several difficult issues to be confronted head on. These include issues such as how to deal with the allocation of risk within MNE groups and returns to capital, and the meaning and role to be assigned to the notion of value creation. In relation to risk, for example, there is a question about how to draw the line between, on one hand, taxpayer allocations of risk with some limited control over risk by the assigned party and, on the other hand, a system that simply looks at where risk is managed and controlled. These are two very different approaches, and countries obviously have different ideas about which is the right approach, notwithstanding the detailed guidance in the transfer pricing guidelines.²³

On capital, relevant questions include whether we have a system that allocates any residual income to the capital provider, or whether we should seriously try to sort out the questions about anticipated and unanticipated rates of return when applying a transactional approach to the arm’s-length principle, particularly in valuing intangible assets. Finally, regarding value creation, the following considerations are relevant: What do we really mean when we use the term “value creation,” and what is it meant to achieve, if anything? Is it meant to play a role in the operational application of the transfer pricing guidelines? Is it intended to refer to functional contributions? Or does it also cover the contributions from risk assumption and capital? Also, it would be important to understand how we geographically locate value creation? All these questions were heavily discussed in the BEPS project but not clearly or definitively resolved, as can be seen by varying country approaches to cases that have arisen in the last five years.

Some Broader Questions

The questions and priorities referred to above seem inevitable based on the current direction of the hybrid approach to the allocation of corporate income within the international tax system. However, the current position also leads to additional questions about the arm’s-length principle and its continuing role in the systems, including the following:

• If it is possible to fix the arm’s-length principle, our core income allocation system, why do we need the formulary overlay?

• However, if we cannot fix the arm’s-length principle, why are we keeping it as the core basis of income allocation for the international tax system?

• Is the pillar 1 solution intended to function as part of the fix of the arm’s-length principle (for example, as a supporting “prop” in areas where the arm’s-length principle needs additional supporting measures)?

• If so, how does pillar 1 deal with any identified problems, thus leaving the arm’s-length principle to function where it properly can work?

• If this is the intended approach, why is it limited to only approximately 100 companies?

• What is the position with non-covered entities (namely, less profitable MNE groups, MNE groups in the natural resources, extractives, and financial services sectors, and so forth)? How do we keep the arm’s-length principle system functioning for these entities? Or are they somehow less problematic? If so, why?

There are also questions about the role and impact of pillar 2, namely the following:

• Could the pillar 2 approach, combined with the arm’s-length principle, represent a fix for the problems of the arm’s-length principle?

• If so, could a successfully implemented pillar 2 reduce or remove the pressure for a partial reallocation of profit to the market?

The Agenda for Future Work

The questions posed above seem to be the right policy and technical questions that need to be addressed if the arm’s-length principle is to remain an important part of the international tax system. This suggests a direction for future work based on the policy and technical assessment of the current position in which the arm’s-length principle remains the core part of the income allocation system. The discussion also suggests in what direction resources, such as OECD working parties, and so forth, should prudently be directed. However, it seems naive to think the future progression will be so simple and ordered. Much more likely, we fear, is that not much attention will be given to the open questions relating to the arm’s-length principle until all the questions arising from the development of the two-pillar package — and from integrating the new and existing income allocation systems — are more fully elaborated upon and addressed. This period may extend for some time, such as to the point at which the two-pillar package has been fully implemented and bedded down.

Not only does this mean that the arm’s-length principle is maintained, but it also means that, for several years, limited headway will be made in addressing its problems. We think that situation will likely be the source of significant instability and will be a material threat to the implementation of adequate dispute resolution measures. The instability will be characterized by several friction-generating points in the system. These are likely to include all the arm’s-length principle issues and problems we already have, plus some new ones, such as a marked rise in arguments between tax authorities that boil down to what value creation means and what it requires.²⁴ We further expect more creative efforts to work around known transfer pricing areas of difficulty such as some states introducing new diverted profits taxes or other similar measures. Because functional dispute resolution rules (which are going to be even more necessary) are dependent on mutually understood underlying substantive income allocation rules, it will be difficult to progress work on dispute resolution while the substantive rules themselves are in flux.

Fixing the Arm’s-Length Principle

The foregoing discussion identifies the need to address head on the question of fixing the arm’s-length principle. But we have thus far offered no comment on whether, realistically, we think it will be possible to fix the arm’s-length principle in a manner that is acceptable to a wide body of states. We therefore now turn to that inevitably speculative point.

The key question is whether the arm’s-length principle can be fixed. On this issue, we note that the transfer pricing work on business restructurings from 2005-2010 sought to address some of the difficulties in applying the arm’s-length principle to counter MNE avoidance. Moreover, beginning in 2012, the BEPS project made several further significant changes to update and bolster the arm’s-length principle. A consideration of the arm’s-length principle, including its strengths and weaknesses, has also been a major part of the discussion at the OECD for the last three to four years, including discussion of the effectiveness of the BEPS changes and what further changes, if any, are needed. This is therefore by no means a new question.

In the context of the existing policy and technical debate, we question whether one can be confident that adequate fixes can be found for the arm’s-length principle. Referring to the two tests previously identified that would need to be satisfied to arrive at a suitable fix for the arm’s-length principle, we would observe that, concerning the need to devise suitable technical responses, developing solutions to core arm’s-length principle problems like mobility, capital, risk, and so forth, has certainly proved a lot more difficult than might be appreciated. In any event, even if such solutions were developed, they would not deal with some fundamental questions raised in the digitalization debate (such as the recognition of the demand side paradigm, nexus issues, related profit allocation issues, and so forth) and so would likely not be acceptable to many states. It seems unlikely that these arm’s-length principle issues can be addressed in a way that provides market countries comfort that they will get what they believe to be a fair share of global taxing rights. For these reasons, we are pessimistic about the prospects of being able to fix the arm’s-length principle system.

On the specific question whether pillar 2 might be regarded as some sort of fix for the arm’s-length principle (if pillar 1 is deferred or fails), neither do we see that as a realistic solution.

It is true that, at the time of this writing, there is some discussion about the possibility of decoupling pillar 2 from pillar 1. It is probably also true that pillar 2 could prove to be easier and quicker to implement. It is therefore a reasonable question whether a successful pillar 2 might reduce, or remove altogether, the momentum in favor of a reallocation of some taxing rights to market states. However, we think that outcome is very unlikely. Pillar 2 is not mandatory for states, and the extent to which the pillar 2 measures are actually adopted and applied by states remains an open question. Further, in isolation, the conceptual basis and direction of pillar 2 raises some difficult questions — most obviously, why bolster the arm’s-length principle system when it is evidently thought to be problematic?

It is also true that the pillar 2 approach does not provide any response to some of the core problems that have been central to the digitalization concerns, such as the problems from the existing nexus test and the functions-based transfer pricing rules that prevent the market state from having any ability to tax remote sellers, notwithstanding the perceived participation of remote sellers in the economy. (Stated differently, what justifies the default allocation of taxing rights to parent company states under the core income inclusion rule of pillar 2? And how can this, standing alone, satisfy market states?)

A pillar 2-only approach would also do nothing to ameliorate the operational problems of the arm’s-length principle — for example, the problem of inadequate or nonexistent comparables, the pervasive complexity, the proper meaning of the risk guidance, the treatment of capital, and so forth. Finally, if pillar 1 were to fail or be materially deferred, DSTs and other destination-based taxation of corporate income would be back on the table.²⁵ For all these reasons, we do not see the pillar 2 measures as bringing any kind of fix of the arm’s-length principle that might restore stability to the system.

Implications

The conclusion we draw from the foregoing discussion is that sooner or later further changes to the income allocation rules will be necessary. Such changes are arguably inevitable. This means that the inclusive framework agreement likely represents a way station on a potentially tangled path to somewhere else.

For the reasons provided earlier, we are unlikely to see a swift progression to the introduction of new approaches. This is mainly because the questions related to the integration of pillar 1 and the arm’s-length principle will likely take priority for the next few years and lead to a general disregard of key arm’s-length principle technical and policy issues. In what could be a rather messy interregnum, the likelihood is that some of the current pressures on the income allocation system will lead to developments such as a heavier reliance on profit-split approaches that are even further detached from results based on transactions between unrelated entities, an expansion of diverted profits and similar taxes, and a greater attempt by tax authorities to rely on safe harbors and formulas. We are also likely to see increased concern from MNEs about the complexity of the arm’s-length principle system and its interaction with the new system. Disagreement is likely given that the “gaming potential” upside from the arm’s-length principle is materially reduced (through the combined effect of the BEPS measures and the two-pillar package). Increased concern is also more likely considering that nothing has been done about the level of complexity in the existing system. In fact, more complexity was added in the form of the pillar 1 measures, especially their interaction with the arm’s-length principle system.

From a more general perspective, it seems hard to conclude that we are not on a path that leads away from the arm’s-length principle, at least to some degree. In this regard, many arguments have recently been deployed in support of such a direction. As previously discussed, there have also been many concerns voiced about the arm’s-length principle. As reflected in the pillar 1 measures, and as supported now by 137 states, the first steps have been taken to incorporate into the mainstream income allocation rules an element of a destination-based approach. This suggests there may be no going back, even if the two-pillar approach runs into difficulties. It also suggests the hegemony of the arm’s-length principle has already gone and that it will be impossible to restore.

The Last Word

The key issue concerning the income allocation rules of the international tax system is the route to reestablishing stability in that system. Ultimately, we do not think any stability is going to be achieved in the absence of an engagement with the fundamental policy and technical questions we have set out and those that concern the viability of the arm’s-length principle. Unfortunately, and for the reasons suggested earlier, we think it might be some time before there is a return to properly addressing these matters relating to the arm’s-length principle. We expect this to be the case even though, irrespective of pillar 1, the arm’s-length principle will remain the determining principle for the overwhelming majority of the global income allocation system.