An Analysis of the OECD's Discussion Draft on Profit Splits

The Tax Management Transfer Pricing Report ™ provides news and analysis on U.S. and international governments’ tax policies regarding intercompany transfer pricing.

The authors examine the OECD's base erosion and profit shifting discussion draft on use of the profit split method in transfer pricing, and analyze what they view as its primary weaknesses. They contend the discussion draft implies that a profit split could be applied when it may not be the most appropriate method, or based on subjective judgments regarding entities' relative contributions to returns on cross-border transactions.

Kenneth P. Christman Romita  Mukherjee

By Kenneth P. Christman and Romita Mukherjee, Ernst & Young LLP

Kenneth P. Christman is an executive director with Ernst & Young LLP in Washington, D.C. Romita Mukherjee is a manager with Ernst & Young LLP, San Jose, Calif.

The views expressed herein are those of the author and do not necessarily reflect the views of Ernst & Young LLP.

The Organization for Economic Cooperation and Development released its revised guidance on the profit split method on July 4, 2016, as part of its two-year project to combat tax base erosion and profit shifting. The OECD described the discussion draft as clarifying and strengthening the guidance on the transactional profit split method released the year before.

The discussion draft presents proposed revised guidance on selecting and applying a profit split method and includes questions for public comments that will be taken into account by Working Party No. 6 as it revises guidance in Chapter 2 of the OECD Transfer Pricing Guidelines for Multinational Enterprises and Tax Administrations . The views and proposals in the discussion draft didn't represent a consensus view of the OECD's Committee on Fiscal Affairs, but were released in draft form to provide an opportunity for public comments.

This article reviews the discussion draft and analyzes what the authors see as its primary weaknesses. In particular, the discussion draft fails to clearly articulate the considerations relevant to determining when a profit split method is the “most appropriate method.” It doesn't explain what evidence is necessary to reliably apply the profit split method. The authors also believe the discussion draft fails to give much guidance on how a profit split method should, in practice, be applied, nor does it address the factors that make the use of a profit split method challenging for taxpayers, including the common problem of finding sufficient evidence necessary to reliably apply the method.

For these reasons, the authors interpret the discussion draft as implying:

  •  that a profit split method could be applied in a situation where it isn't the most appropriate method; and
  •  that it could be applied based on subjective judgments about the relative contributions of different entities to the returns on cross-border transactions.

First Considerations: The Most Appropriate Method Rule

Fundamental transfer pricing principles mandate that any transfer pricing method be assessed against alternative available methods, taking into consideration the facts and circumstances of the transaction and the availability of comparable and reliable data in order to select the most appropriate method. Therefore, if a profit split method is selected as the most appropriate method, the taxpayer or tax administration applying it should be able to explain why a profit split method is the most appropriate method for the transaction at hand.

One might expect, therefore, that the discussion draft would begin with a discussion of the most appropriate method rule. This approach would reiterate the well-established principle that any transfer pricing method, including a profit split method, should be assessed against alternative transfer pricing methods, so taxing authorities aren't free to disregard a taxpayer's transfer pricing method unless they can demonstrate that another method is more reliable.

Consequently, one would also expect some discussion of the kind of empirical evidence available and necessary to apply the method with reasonable reliability. However, the discussion draft's guidance on the type of evidence necessary to reliably apply a profit split method is limited. The discussion draft does have some discussion of when a profit split method is likely to be useful, but provides little discussion of why a profit split method is particularly likely to be reliably applied in these situations or what evidence is necessary for a profit split method to be reliably applied.

In this context, the authors acknowledge that the most appropriate method rule is necessarily flexible and sensitive to the particular circumstances of each case. Therefore, the discussion draft couldn't reasonably be expected to provide an encyclopedic list of facts or circumstances that could bear on a profit split method's relative reliability. However, the dearth of such guidance might lead to inappropriate use of a profit split method with little or no consideration of relative reliability.

Profit Split as the Most Reliable Method

In the authors' experience, the application of a profit split method is rather infrequent. This is partly due to the fact that a profit split method is inherently a complicated method to apply and, in many cases, a simpler (probably comparable-based) method is available and can be shown to be applicable with reasonable reliability. In addition, as indicated previously, the most appropriate method rule necessarily emphasizes the role of empirical evidence in applying a transfer pricing method, including a profit split method. It is frequently the case that the empirical evidence necessary to apply a profit split method isn't available.

However, the discussion draft does identify two situations in which a profit split method is frequently applied:

  •  transactions that involve unique and valuable contributions; and
  •  transactions that involve highly interrelated functions being performed by entities in different jurisdictions.

In general, the authors agree that a profit split method is often applied in these two situations. However, the authors feel the discussion draft's elucidation of why the profit split method can be applied, even in these two situations, is inadequate. Further, the discussion draft provides insufficient guidance on what evidence will be necessary in order to apply a profit split method in these situations.

Unique and Valuable Contributions (And the Role of Asset Synergy)

Paragraph 23 of the discussion draft suggests that a profit split method is particularly appropriate when multiple parties make “unique and valuable” contributions.

A unique contribution is defined to mean a contribution in the form of a function performed, asset provided or risk assumed if the contribution is:

  •  not comparable to a contribution made by uncontrolled parties in comparable circumstances; and
  •  a key source of economic benefit.

A profit split method is often applied when dealing with such unique and valuable contributions. However, the existence of unique and valuable contributions isn't sufficient to make a profit split method the most appropriate method. As the discussion draft notes, a profit split method is more likely to be the most appropriate method when two parties make contributions that are so integrated that they can't be reliably evaluated in isolation but there is no discussion of when this might occur. Generally, this occurs when there is “asset synergy.”

To illustrate the concept of asset synergy, consider a transaction that involves two members of the same multinational group that are located in two different tax jurisdictions. Party A is located in Jurisdiction X and Party B is located in Jurisdiction Y. Party A has worldwide patents on Enzyme A. Party B has worldwide patents on Enzyme B. Neither Enzyme A nor Enzyme B is particularly valuable as a pharmaceutical because both are only marginally effective. However, combining Enzyme A and Enzyme B produces a drug that is very effective and successful in treating certain life-threatening diseases.

Assume that the worldwide value of Enzyme A is $20X and the worldwide value of Enzyme B is $40X but, when combined, Enzymes A and B have a worldwide value of $160X. Therefore, blending Enzyme A and Enzyme B results in an increase in value by $100X attributable to asset synergy.

In this example, both Enzyme A and Enzyme B might be considered unique and valuable contributions. However, it isn't at all clear that a profit split method would necessarily be the most appropriate method to apply if the enzymes weren't synergistic. Even accepting that a comparable-based method wouldn't be available, an income-based method might well prove relatively reliable in estimating the “stand-alone” arm's-length value for Enzyme A or Enzyme B (and since there is, by hypothesis, no asset synergy, that is all that is required). However, the presence of asset synergy means that the unique and valuable contributions can no longer be valued on a stand-alone basis.

To illustrate, assume in the above example that Party A transfers all its rights in Enzyme A to Party B. Determining the arm's-length price of such a transfer raises the issue of how much of the $100X increase in value, attributable to synergy, should be captured by Party A at arm's length and how much of the $100X will be captured by Party B. The issue now presented, in simplest terms, is how to “split” synergy value between two parties presumed to be bargaining at arm's length.

A profit split method may be well-suited to dealing with this issue (assuming the availability of appropriate evidence). By failing to acknowledge the role of asset synergy, the discussion draft misses an opportunity to delineate a particular circumstance in which a profit split method is particularly likely to be the most appropriate method.

Highly Integrated Operations

The discussion draft also notes that a profit split method should only be applied to a split of “actual profits” when there is a “high level of integration of activities.” It points out that in highly integrated transactions, the relevant parties each may perform similar or related functions, use core assets and assume risks that are integrated with each other and therefore can't be evaluated in isolation.

Further, the discussion draft does suggest that a “value chain” analysis might be used to identify whether a profit split method might be appropriate. The value chain analysis is described as determining where and how value is created in a business with emphasis on:

  •  identifying which entities perform which functions and how they are interrelated; and
  •  how the economic circumstances, such as the creation of unique intangibles, may allow the enterprise to capture profits in excess of the ordinary.

Nevertheless, the discussion draft notes that the mere fact that a multinational enterprise operates using a global supply chain isn't sufficient to establish that a profit split method is the most appropriate method.

The authors agree that a high degree of integration is a factor that may make a profit split method relatively more reliable compared to other transfer pricing methods. Additionally, if a profit split method is to be applied, a value chain analysis may well be helpful to its application. Finally, the mere existence of an integrated value chain isn't sufficient to establish that a profit split method is the most appropriate method. After all, in a global economy in which a supply chain for a particular product may be housed in multiple entities in different jurisdictions, some level of interdependence is ubiquitous.

However, that doesn't necessarily mean that a profit split method is necessary to determine appropriate returns to entities in a supply chain. Frequently, market benchmarks are available that permit determination of appropriate returns to particular entities using comparable-based methods. Similarly, appropriate returns to intangible assets can often be determined using income-based methods.

Unfortunately, the discussion draft provides little explanation of how a value chain analysis might be performed in general or how it (or some other approach) might be useful in determining if a profit split method is more or less appropriate. It doesn't provide guidance on the type of empirical evidence useful or necessary to applying the profit split method nor on the type of empirical evidence that may make other methods more reliable than a profit split method. Consequently, the discussion draft's observation that a profit split method might be applied if cross-border activities are integrated, provided without detailed guidance on how relative appropriateness is to be determined, provides few parameters for the application of a profit split method but implies that profit split methods can be very widely applied.

It is, however, possible to describe transactions that are so highly integrated that a profit split method is likely to be the most appropriate method. The highly integrated operations dealt with in the Internal Revenue Service's Notice 94-40 provides a good example of operations that are so highly integrated that a profit split method becomes the method of choice. It also provides a discussion of the evidence necessary to apply a profit split method with reasonable reliability in a particular situation.

The IRS notice observes that increasing globalization and technological developments have led to a high level of integration of the world's financial markets. Many financial intermediaries trade commodities and derivative financial products 24 hours a day by maintaining traders in offices around the world (referred to as “global trading”). Four general functions are most commonly observed in global trading operations—trading, sales, management and support.

Global trading operations of companies that are functionally fully integrated are characterized by centralized management of risk and personnel. The business is managed as one global position for purposes of risk management rather than several discrete businesses (which, of course, requires a sophisticated management information system function). A trading book isn't independently maintained for each trading location. Rather, one book is maintained and the trading authority for that book is passed from one trading location to the other at the close of each trading day for that trading location (the “global book”). To assist in the management of the risk, a central credit department monitors the credit-related exposure of the transactions entered into by the traders. In a fully integrated operation, each office contributes to the overall profitability of the global business.

The notice says that the advance pricing agreement (APA) process has proven to be a useful vehicle to allocate the income of such an integrated global trading business between taxing jurisdictions. In such APAs the IRS used a profit split method to allocate income of related operations between taxing jurisdictions. The notice observes that the IRS, treaty partners and taxpayers found that the use of a profit split method to allocate the income of these functionally fully integrated global trading businesses was appropriate because of the volume and nature of transactions involved.

Further, the notice says that a profit split method can be applied using certain factors that act as indexes of the relative contribution of each trading location to the profitability of the global book. Although the notice observes that different factors might be employed in different circumstances, they describe three factors that the IRS typically has used:

  •  a value factor measured by relative compensation of the traders in each trading location;
  •  a risk factor measured by the nominal amount of swap transactions (or open commodity positions) times the time to maturity of the transactions initiated at each trading location; and
  •  an activity factor measured by the relative compensation of support personnel in each trading location or the relative present value of cash flows of transactions entered into at each location.

As discussed above, an important characteristic of the arm's-length principle is that any transfer pricing method be implemented on the basis of empirical evidence. The reliability of a transfer pricing method requires that there be actual direct evidence available that supports estimation of relative contribution of the transacting parties to avoid subjectivity, so that the taxpayer can have confidence that the method is reliably applied. The notice is an example of guidance that attempts to provide detailed practical guidelines on what evidence will be necessary in order to apply a profit split method with reasonable reliability in a particular situation. As discussed below, this type of discussion is largely lacking in the discussion draft.

Factors Used to Split Profits

Having identified the circumstances in which a profit split method is likely to be applied, the discussion draft turns to the issue of how a profit split method could be applied to allocate synergy value among related parties. By and large, this discussion is largely unchanged from the current OECD guidelines.

The discussion draft suggests two possible approaches. First, profits may be split based on comparable uncontrolled transactions. However, it also notes that such comparable transactions are typically not available. Consequently, the discussion draft implies that the second approach is almost always the one applied. In this approach, income is split between related parties in proportion to their relative contribution to the creation of such income.

Assuming that these contributions create synergy value or are activities so highly integrated that it is impossible to identify separate returns to each contribution, the authors believe such an approach is reasonable. Of course, this raises the issue of how to measure relative contribution.

The discussion draft provides relatively little discussion of what indices might be used to measure relative contribution (and what discussion is provided is largely carried over from the current OECD guidelines). It suggests that indicia of relative contribution could be either based on costs incurred or assets provided.

The authors believe this is roughly correct and consistent with the activities of partners in partnerships (to the extent relative contributions can be assessed). However, they would also suggest that cost is generally likely to be less useful than relative value.

To return to the example illustrating asset synergy, one would expect that the party that would capture the greater part of the increase in value would be the party with the greater bargaining power. There may be a number of indicia of bargaining power that could be used but the most obvious would be the relative stand-alone value of Enzyme A and Enzyme B. Since Enzyme B is more valuable, one would expect that Party B would capture more of the increase in value attributable to synergy. Consideration of the time allocated to creating Enzyme A and Enzyme B is less useful precisely because it has no necessary relationship with the relative value of the two assets or the relative bargaining power of the two participants.

However, the discussion draft devotes very little attention (only two short paragraphs, both unchanged from current OECD guidelines) to the use of assets as indices of relative contribution. Neither paragraph discusses the uses of specific indices of relative value that might be used in particular circumstances (in contrast, for example, to Notice 94-40), nor are any examples included to illustrate such application. A similar observation applies to the use of costs incurred as a measure of relative contribution (although this is less of a criticism since, as noted, the authors believe use of costs as an index of relative contribution is likely to be less useful).

Simply put, the discussion draft gives little guidance on how a profit split method should, in practice, be applied. Furthermore, it fails to address the kind of evidence necessary to reliably apply a profit split method.

Failure to consider what evidence must be developed in order to reliably apply a profit split method is of significant concern. As noted, the existence of appropriate evidence is crucial in applying the most appropriate method rule. Ignoring the need to develop appropriate evidence implies that a profit split method may be more widely employed than permitted by the most appropriate method rule.

Applicability of profit split MethodTo Actual Versus Anticipated Income

The discussion draft also distinguishes between anticipated and actual profits and treats the distinction as significant. Indeed, the discussion draft states that it is primarily concerned with the application of a profit split method to actual profits. This would seem to suggest that application of a profit split method to actual profits is somehow preferred in achieving an arm's-length result.

However, there is no clear explanation of the rationale or significance of applying a profit split method to actual versus anticipated profits. Also, there is no discussion of how applying a profit split method to actual profits can be reconciled with the fact that the arm's-length standard involves estimating the result that would be achieved by two parties bargaining at arm's length and is, therefore, essentially an ex-ante standard. Indeed, there are no examples that illustrate how a profit split method would be applied to actual profits or how such application might differ from an application of a profit split method to anticipated profits.

Since no argument is made to justify applying a profit split method to actual profits, one is left to speculate as to why the discussion draft evinces a preference for such application. Perhaps the application of a profit split method to actual profits is preferred because it may allow tax authorities to make use of ex-post evidence, in effect doing an end run around the ex-ante nature of the arm's-length standard. If so, the effect would be to embed into a profit split method the use of ex-post information to reform transactions involving transfers of hard-to-value intangibles between related parties as envisioned by Section VI.D.4 of the OECD guidelines. The merits of reforming transactions based on ex-post evidence aside, it is clearly not intended to be applied widely (e.g., to all transactions to which a profit split method would be applied whether or not they involve hard-to-value intangibles).

Implications and Conclusions

To summarize, the discussion draft provides revised guidance on selecting and applying the profit split method and includes questions to practitioners. However, the discussion draft pays little attention to judging the relative reliability of a profit split method compared to alternative methods. There is little discussion of the complexities of a profit split method and relatively little discussion of the type of evidence that needs to be collected in order to apply a profit split method in a reliable manner.

This can be observed in the discussion draft's approach to asset synergies. The authors agree that a profit split method can often be used to deal with the problem of dividing synergistic returns between related entities. However, the discussion of the empirical evidence necessary to allocate synergistic returns is inadequate.

This lack of attention to empirical evidence and the most appropriate method rule result in the discussion draft providing, at best, an inconsistent guidance with respect to the scope of application of a profit split method. It specifically notes that the mere fact that a multinational enterprise operates using a global supply chain isn't sufficient to establish that a profit split method is the most appropriate method. However, the discussion on applying a profit split method when cross-border activities are integrated, provided without detailed guidance on how relative appropriateness is to be determined, provides few parameters for the application of a profit split method.

One could read the discussion draft as implying that a profit split method should be a common transfer pricing method since integrated supply chains in a globalized world are ubiquitous. A profit split method may indeed be the most appropriate method when dealing with certain types of highly integrated multinational businesses. However, as noted earlier, the mere existence of some integration isn't sufficient to establish that a profit split method must be the most appropriate method; market benchmarks are often available to determine appropriate returns to particular entities. Similarly, income-based methods can often give an appropriate return to intangibles.

Given the complexities of applying a profit split method and potential difficulties associated with finding indicia of relative contribution by particular entities to a multinational business, a profit split method is likely to be the most appropriate method only in limited circumstances (e.g., global dealing in financial derivatives).

Indeed, the lack of attention to empirical evidence demonstrating arm's-length pricing might be taken to imply that there is less need for empirical evidence. It might be taken to imply that a profit split method should be widely applied based on the subjective judgments of particular taxing authorities as to the appropriate returns to entities in their jurisdiction.

Lest this seem an unduly broad reading of the discussion draft, certain nongovernmental organizations in comments suggested that a profit split method should be considered the most appropriate transfer pricing method unless the taxpayer can establish that a one-sided method is more appropriate. They also suggest the application of a simplified a profit split method in which the need for detailed analysis would be replaced by a predefined set of allocation factors such as operating expense, the number of users or sales. To the extent there is international agreement on the use of predetermined allocation keys, this would effectively be a system of formulary apportionment. There is no reason to believe that international agreements on specific allocation keys and their relative weighting would be possible to obtain.

In the absence of international agreement, relying on each tax authority's subjective judgment on relative values and the tax jurisdiction's appropriate proportional share of a multinational's global profit is certain to lead to double taxation. In other words, transfer pricing will fail in its most fundamental purpose.

Finally, as noted above, it is unclear why the discussion draft emphasizes the application of a profit split method to actual profits over application to anticipated profits. However, it is possible that the preference represents an attempt to broaden the application of ex-post evidence, as envisioned by Section VI.D.4 of the OECD guidelines, to a far broader selection of transactions than previously envisioned.

Hear more on this topic in Paris March 27-28 at the 5th Annual Global Transfer Pricing Conference: https://www.bna.com/2017-global-transfer-pricing-paris/.

Copyright © 2017 Tax Management Inc. All Rights Reserved.

Request Transfer Pricing Report