Tax Treaties, Transfer Pricing and Financial Transactions Division
Centre for Tax Policy and Administration
Organisation for Economic Co-operation and Development

Thank you for the opportunity to provide comments on the OECD Public Discussion Draft on BEPS Action 10: Revised Guidance on Profit Splits dated June 22, 2017 (the Discussion Draft).

Keidanren has been concerned for some time about the increasingly broad ways in which the transactional profit split method is applied and believes that additional guidance is required for the method to be used appropriately. The Discussion Draft offers reorganized and expanded explanations on selecting the most appropriate method, provides more examples, and directly addresses issues involving profit split factors. These efforts are highly appreciated.

The Discussion Draft also includes many welcome statements, such as: "the selection of the 'most appropriate' method should take into account the relative appropriateness and reliability of the selected method" in paragraph 5; "an appropriate method using uncontrolled transactions that are comparable, but not identical to the controlled transaction is likely to be more reliable than an inappropriate use of the transactional profit split method" in paragraph 14; and "a lack of comparables alone is insufficient to warrant the use of a transactional profit split" in paragraph 28. We hope these statements will introduce new discipline to the application of the transactional profit split method.

On the other hand, the Discussion Draft may also be read as more broadly endorsing the application of the transactional profit split method when compared to the current OECD Transfer Pricing Guidelines for Multinational Enterprises and Tax Administrations (Transfer Pricing Guidelines) and the OECD Public Discussion Draft on BEPS Action 8-10: Revised Guidance on Profit Splits, dated July 4, 2016 (July 2016 Discussion Draft). For example, paragraph 11 states: "Transfer pricing methods are not necessarily intended to replicate arm's length behaviour, but rather to serve as a means of establishing and/or verifying arm's length outcomes for controlled transactions." This would seem to go beyond traditional guidance on the arm's length principle. We agree that transfer pricing methods do not necessarily replicate arm's length activities, but due respect should be given to the fact that the transactional profit split method is rarely considered when transaction prices are voluntarily negotiated at arm's length.

Another notable point in the Discussion Draft is the absence of references to sequential integration and parallel integration, which were included in the July 2016 Discussion Draft. In recent years, the businesses operated by multinational enterprises have become more complex and sophisticated. In this environment, these two concepts of integration are difficult to define. Nonetheless, we believed these concepts to be useful when examining the relationships between highly integrated business operations and the application of the transactional profit split method.

The transactional profit split method is inherently difficult to apply in practice. Where its use is increasingly accepted, or enforced arbitrarily by tax administrations, occurrences of double taxation are likely to increase, including cases that may not be resolvable via mutual agreement procedures. The final version of the guidance should therefore reiterate that the transactional profit split method is applicable only in very limited circumstances.

1. When is the Transactional Profit Split Method Likely to be the Most Appropriate Method?

The Discussion Draft includes three indicators that the transactional profit split method is likely to be most appropriate: the existence of unique and valuable contributions, highly integrated business operations, and shared assumption of economically significant risks. Of these, the existence of unique and valuable contributions is stated as perhaps the clearest indicator, in paragraph 13. We have commented on each of these indicators below.

(1) Unique and valuable contributions

In principle, when both parties to a transaction make unique and valuable contributions, we have no objection to concluding that the transactional profit split method is likely to be the most appropriate method (as already stated in the Transfer Pricing Guidelines). To be called "the clearest indicator," though, The Discussion Draft should clarify how unique and valuable contributions differ from other contributions (such as contributions that are not unique but are valuable, or contributions that are unique but not valuable). Otherwise, the subjective interpretation of this indicator by tax administrations may lead to the wider application of the transactional profit split method.

Keidanren's general view is that valuable contributions should not be identified unless the contribution of one party to the transaction has generated considerable value independently.

We would also like to emphasize that the existence of unique and valuable contributions alone does not always guarantee that the transactional profit split method is the most appropriate method. Cases exist that also require the shared assumption of economically significant risks; indeed, cases exist where much greater importance is attached to the shared assumption of risks than to unique and valuable contributions.

One example would be a value chain in which Company A, a parent company engaged in R&D activities enters into a contract with Company B, a subsidiary of Company A, according to which Company B distributes products developed by Company A. Company A also enters into a contract with Company C, another subsidiary of Company A, according to which Company C manufactures products for Company A on a contract basis. In this case, even if Company C's cost-cutting efforts at its factories or Company B's advertising and promotion campaigns are determined to be unique and valuable contributions, many of the risks assumed by the Company A will differ in nature and significance from the risks assumed by Company B or Company C. The economically significant risks are primarily assumed by Company A and neither Company B nor Company C would typically share in the assumption of those risks. Consequently, we would not consider it appropriate to apply the transactional profit split method in this case.

Even where both parties to the transaction develop valuable intangibles, these parties should not be considered to share the assumption of economically significant risks if the development risk is assumed by one party only. Hence, in this case it would not be appropriate to apply the transactional profit split method.

(2) Highly integrated business operations

As the business operations of multinational enterprises have become more or less integrated, the issue here continues to be how to distinguish business operations that are highly integrated from those that are not highly integrated. It is possible that where there is sequential integration of business operations, as defined in the July 2016 Discussion Draft, these business operations may now be determined to be highly integrated. Such operations include the intragroup supply of a broad range of parts, as well as the relationships, among Company A, Company B and Company C in the aforementioned example. In order to discourage arbitrary use of this indicator, the same consideration should be made as in the case of unique and valuable contributions — the shared assumption of economically significant risks should also be requisite for the transactional profit split method to be selected as the most appropriate method.

(3) Shared assumption of economically significant risks

Given that risk is an abstract concept, the key to considering this indicator in practice is presumably to determine whether the economic risks are significant. We are concerned about the lack of detail in the Discussion Draft regarding the "separate assumption of closely related economically significant risks," despite this being treated as equivalent to the shared assumption of economically significant risks. Example 3, to which we will later refer, includes some additional explanation but may too readily allow the risks assumed by the parties to a transaction at each stage of the value chain to be determined to be closely related.

Paragraph 24 refers to the treatment of cases "where a party contributes to the control of economically significant risk, but that risk is assumed by the other party to the transaction." We agree with the statement that "[t]he mere fact that an entity performs control functions in relation to a risk will not necessarily lead to the conclusion that the transactional profit split is the most appropriate method in the case." However, we ask that more detailed guidance and specific examples are provided regarding such cases.

2. Examples

We welcome the additional examples provided, including those where the use of the transactional profit split method is determined to be inappropriate. We hope that the finalized guidance will expand on the explanations accompanying these examples, in particular, regarding how unique and valuable contributions are determined to exist, whether the parties share in the assumption of economically significant risks, and the selection of appropriate profit splitting factors.

Example 1

This example in the pharmaceutical sector has a basic transaction structure resembling those of the similar examples presented in the Public Discussion Draft on BEPS Action 8: Implementation Guidance on Hard-to-Value Intangibles, dated May 23, 2017. Under the hard-to-value intangibles approach, a one-sided analysis is likely to be conducted as a rule. However, since Example 1 illustrates profit split methods, it may be helpful to explain how income should be calculated by applying both methods.

In addition, while Example 1 concludes that the transactional profit split method is likely to be the most appropriate method following the approval and commercialization of a new pharmaceutical product, an analysis of the product life cycle cannot be identified from the available information. Due attention should be given to the possibility that, depending on the timing of implementation, the analysis may fail to capture Company A's losses during the initial investment period by focusing only on the profit situation at the height of Company S's sales.

Furthermore, this example should explain whether Companies A and S share the assumption of economically significant risks.

Example 2

As with Example 1, our concern about Example 2 is that no mention is made of the shared assumption of economically significant risks and the conclusion that the transactional profit split method is the most appropriate method is based entirely on the basis of unique and valuable contributions.

Whereas Example 2 explains that B Co carries out extensive advertising campaigns that feature the tea grown and the unique blend developed by A Co, further explanation is required in order to clarify which of A Co or B Co is the principal. It seems unusual that these two companies would adopt a business model under which they divide the functions of tea-growing, blending, and advertising on equal terms. It is probably more natural that those functions are divided as one party being the principal and the other being the subcontractor, and that the principal assumes economically significant risks. If that is the case, a one-sided analysis of either party may suffice depending on the circumstances.

Examples 3 and 4

Examples 3 and 4 cover fact patterns based on which tax administrations and taxpayers are highly likely to have different views.

These examples offer important points in terms of comparison. Still, it is somewhat questionable whether the premise upon which Example 3 is based is realistic. Whereas paragraph 77 states that Company B undertakes "the global distribution of the goods," in reality a business model where Company B performs the global distribution function is implausible. The distribution function would more likely be divided among and performed by multiple distribution subsidiaries responsible for the Americas, Europe, Asia, and other regions, respectively, while the function of overseeing the formulation of global marketing strategies is performed by the parent that supervises the R&D and manufacturing of products (which corresponds to Company A in Examples 3 and 4). Many multinational enterprise groups do not have more than one center of value creation within the group, as a rule. Thus, actual cases to which Example 3 applies are probably limited in number.

In Example 4, we agree with the conclusion but request additional guidance on the following statement in paragraph 83: "The marketing activities performed by Company B are more limited and do not significantly enhance the goodwill or reputation associated with the trademark and its distribution activities are not a particular source of competitive advantage in its industry." On what grounds this determination is based should be explained. Similarly, it is unclear what kind of functional analysis would lead to the differing conclusions in Examples 3 and 4—that is, the risks assumed by Company B are determined to be economically significant for the business operations in the Example 3, but not in Example 4. It is a source of concern that this matter may tend to be viewed differently by tax administrations and taxpayers.

Example 5

This example should also explain whether WebCo and ScaleCo share the assumption of economically significant risks. We would also ask that the method used to price the transfer of the program is explained.

Example 6

We agree that the transactional profit split method is unlikely to be the most appropriate method, which is a reasonable judgement. In practice, however, the views of tax administrations and taxpayers are likely to differ when evaluating whether Company B "make[s] any unique and valuable contributions" as stated in paragraph 92. This matter thus calls for a deeper analysis.

Example 8

This example is useful in that it also refers to the splitting of losses. We also understand how the conclusion might be reached that the profits should be split among the three companies based on their respective development costs. Nonetheless, in reality, a new product development project within a multinational enterprise group would be normally established by one group entity acting as the principal, with participation by other group entities as subcontractors. This example would therefore seem to leave a number of issues unaddressed. For example, whether it is appropriate to split the profits generated from the joint development project based exclusively on the development costs, despite the project having already shifted from the development stage to the manufacturing and sales stage. The treatment of the principal's entrepreneurial profits also needs to be addressed.

In addition, while Example 8 may be considered a fairly typical example of the shared assumption of economically significant risks, we hope that additional examples, including those based on fact patterns other than joint development, will be provided in the finalized guidance.

Example 9

The question remains regarding the reasonableness of applying the transactional profit split method to the profits generated from the contributions of parent Company A and subsidiary Company B where the subsidiary's contributions are "innovative marketing activities" harnessing the intangibles licensed by the parent, and the parent's contributions include the know-how it has developed and the trademark whose value it has enhanced over the years. Although this example describes Company B's activities as innovative, in reality there would seem to be many cases where a subsidiary simply utilizes the know-how and trademark licensed by the parent in the subsidiary's country of residence. Accordingly, more detailed explanation of Company B's innovative marketing activities are essential to determining whether its contributions are unique and valuable.

In situations like Scenario 2 where Companies A and B share in the assumption of economically significant risks, it would seem logical to use actual profits when splitting relevant profits. However, in cases like Scenario 1 where Companies A and B do not share in the assumption of economically significant risks, it might first be advisable to question whether the transactional profit split method should be selected as the most appropriate method, rather than proceed directly to the discussion about splitting anticipated profits.

Example 10

This example relates to the automotive industry. In this industry, however, it is difficult to conceive of any company agreeing to enter into a commercial agreement like the one concluded between Companies A and B—namely, an agreement whereby the parties mutually buy and sell pieces, molds, and components to each other in apparently similar quantity. On top of that, this example does not sufficiently explain which party performs the R&D and product planning functions and what tasks these functions involve. Both pieces of information are important to intangibles analysis and need to be added. Paragraph 114 contains another point requiring further explanation, stating that "an asset-based splitting factor may be appropriate" if the analysis concludes that there is a strong correlation between the assets and the creation of value. An explanation of the analytical process leading to that conclusion would be helpful.

This example does not specify whether the assets referred to are tangible or intangible. If the assets include even limited tangible assets, it should be noted that the value of tangible assets, such as manufacturing facilities, does not necessarily correlate with the value of technologies integral to the tangible assets. In this case it might be prudent to consider other factors, such as the utilization rate of manufacturing facilities.

3. Responses to the OECD's Specific Questions

(1) Anticipated profits and actual profits

It is reasonable to a certain extent to regard the splitting of actual profits as the appropriate approach in cases where the parties share in the assumption of economically significant risks. However, given that it generally takes considerable time to identify relevant profits and make calculations based on profit splitting factors, companies may not be able to complete the splitting of actual profits before the tax return filing deadline if they operate in jurisdictions where the deadline is relatively earlier than in other jurisdictions.

If anticipated profits have been used, but actual profits turn out to be higher or lower than projected, care should be exercised to prevent tax administrations from enforcing the splitting of actual profits. Furthermore, irrespective of whether anticipated or actual profits are split, any calculations and adjustments related to the transactional profit split method must generally be determined "on the basis of information known or reasonably foreseeable" by the parties to the transaction at the time it was entered into, as clarified in paragraph 46.

(2) Profit splitting factors

1. (a) Capital or capital employed
   Although we do not deny the applicability of capital or capital employed as profit splitting factors at times, in our view capital reflects the outcomes of various economic activities, and as such is not appropriate when evaluating unique and valuable contributions. For example, where the parties to a transaction do not capitalize R&D expenses, splitting profits based on each party's tangible assets is inappropriate as it does not take into consideration the R&D activities that are a source of corporate value. At the very least, capital or capital employed need to be used in combination with another profit splitting factor.
   The OECD/G20 Base Erosion and Profit Shifting Project: Transfer Pricing Documentation and Country-by-Country Reporting, Action 13 - 2015 Final Report, dated October 5, 2015 indicates that the amount of a company's stated capital or tangible asset stated in its country-by-country report cannot be employed as a profit splitting factor, nor can it be used for formulary apportionment purposes, without conducting a detailed transfer pricing analysis. This also applies to the comments regarding head counts of employees below.

2. (b) Head counts of employees
   Levels of personnel expenses as well as of roles, motivation, and competitiveness of workers vary by country. Accordingly, overall head counts of employees are inappropriate for use as profit splitting factors. Nevertheless, headcounts of employees performing particular important functions in financial, investment management, and other industries could be used as one of several profit splitting factors after any required adjustments have been made.

3. (c) Adjustments for purchasing power parity
   In general, we believe it is appropriate to adjust profit splitting factors, such as assets and costs, for purchasing power parity when companies in countries with significantly different economic environments are compared or when exchange rates are rapidly and widely fluctuating. However, attention should also be paid to an increase in the compliance burden resulting from such adjustments. We hope that straightforward implementation guidelines will be provided in the future.

4. Other Comments

(1) Documentation

The Discussion Draft contains the following statement in paragraph 10: "It will therefore be particularly important to document how the transactional profit split method has been applied." We look forward to specific examples and guidelines, in particular the items that should be documented and the extent of the descriptions required.

(2) Uncontrolled transactions that are comparable but not identical

To prevent tax administrations from arbitrarily selecting transfer pricing methods, the following statement in paragraph 14 requires further detailed guidance: "An appropriate method using uncontrolled transactions that are comparable, but not identical to the controlled transaction is likely to be more reliable than an inappropriate use of the transactional profit split method." Guidance is needed especially in the form of examples of "uncontrolled transactions that are comparable, but not identical," as well as concerning to what extent such uncontrolled transactions are permitted to be used.

(3) Master files

Seemingly, the statement at paragraph 58 that the "Master File might be a useful source of information relevant to the determination of appropriate profit splitting factors" goes a little too far. Although a caveat is made at the end of the paragraph, which points out that the Master File is not required to contain granular or detailed information but intended only to provide a high-level overview of a multinational enterprise group, the statement in paragraph 8 should be worded more carefully.

(4) Residual profit split method

When splitting relevant profits in cases where both parties to the transaction make unique and valuable contributions, the residual profit split method should be used, as implied in paragraph 37. The residual profit split method is a conservative approach whereby profits are split based on a two-stage analysis, thus the residual profit split method is generally more reliable than the contribution profit split method.

(5) Cost-based profit splitting factors

In paragraph 67 of the section regarding cost-based profit splitting factors, the risk-weighting of costs is mentioned as an issue to be considered when each party contributes different intangibles. We look forward to examples illustrating how to weight costs by risk. Another point that requires detailed guidance is the following statement in the same paragraph: "[t]he manner in which independent parties would allocate retained location savings would need to be reflected in the profit split." This statement might be interpreted differently by tax administrations in different tax jurisdictions without further specific guidelines, in particular regarding how this relates to comparability analysis.

Sincerely,

Subcommittee on Taxation
KEIDANREN