SECTION 482 WHITE PAPER; SERVICE DEVELOPS NEW METHOD FOR FINDING COMPARABLE PRICE FOR INTANGIBLES.

CHAPTER 1 -- OVERVIEW

CHAPTER 2 -- TRANSFER PRICING LAW AND REGULATIONS BEFORE 1986

CHAPTER 3 -- RECENT SERVICE EXPERIENCE IN ADMINISTERING SECTION 482

CHAPTER 4 -- THE SEARCH FOR COMPARABLES

CHAPTER 5 -- FOURTH METHOD ANALYSIS UNDER SECTION 482

CHAPTER 6 -- THE COMMENSURATE WITH INCOME STANDARD

CHAPTER 7 -- COMPATIBILITY WITH INTERNATIONAL TRANSFER PRICING STANDARDS

CHAPTER 8 -- PERIODIC ADJUSTMENTS

CHAPTER 9 -- THE NEED FOR CERTAINTY: ARE SAFE HARBORS THE SOLUTION?

CHAPTER 10 -- ECONOMIC THEORIES CONCERNING THE IMPLEMENTATION OF SECTION 482

CHAPTER 11 -- ARM'S LENGTH METHODS FOR EVALUATING TRANSACTIONS INVOLVING INTANGIBLE PROPERTY

CHAPTER 12 -- HISTORY OF COST SHARING

CHAPTER 13 -- COST SHARING AFTER THE TAX REFORM ACT OF 1986

APPENDIX A -- ANALYSIS OF QUESTIONNAIRE RESPONSES

APPENDIX B -- SECTION 482 QUESTIONNAIRE -- INTERNATIONAL EXAMINERS

APPENDIX C -- TRANSFER PRICING LAW AND PRACTICES OF SELECTED U.S. TREATY PARTNERS

APPENDIX D -- AN EMPIRICAL ANALYSIS OF THE MARKETPLACE FOR INTANGIBLES

APPENDIX E -- EXAMPLES OF METHODS FOR VALUING TRANSFERS OF INTANGIBLES

I. OVERVIEW AND BACKGROUND

CHAPTER 1

OVERVIEW

A. INTRODUCTION

Section 482 of the Internal Revenue Code ¹ authorizes the Secretary of the Treasury to allocate income, deductions, and other tax items among related taxpayers to prevent evasion of taxes or to reflect their incomes clearly. The Tax Reform Act of 1986 (hereinafter 1986 Act] amended section 482 for the first time in many years by providing that the income from a transfer or license of intangible property must be commensurate with the income attributable to the intangible. The Conference Committee report stated:

The conferees are also aware that many important and difficult issues under section 482 are left unresolved by this legislation. The conferees believe that a comprehensive study of intercompany pricing rules by the Internal Revenue Service should be conducted and that careful consideration should be given to whether the existing regulations could be modified in any respect. ²

In response to this recommendation, the Internal Revenue Service and the Treasury Department have reexamined the theory and administration of section 482, with particular attention paid to transfers of intangible property. This study presents the findings and recommendations of the Service and Treasury.

The study is divided into four parts. Part I recounts the history of section 482 and the evolution of issues leading to the 1986 amendments. Part I also contains recommendations and suggestions for further consideration to assure both thoughtful analysis by taxpayers in setting transfer prices and disclosure of information to permit adequate development of transfer pricing issues on examination.

The problems that have been encountered in relation to transfers of intangible property are both legal and administrative. The 1986 Act clarifies the legal standard for determining arm's length pricing by stating that transfer prices for intangible property must be "commensurate with income." Part II discusses Congress' 1986 change to section 482 and explains that this standard requires periodic, and generally prospective, adjustments to transfer prices to reflect significant changes in the income attributable to intangible property. In any event, transfer prices must be determined on the basis of true comparables if they in fact exist. Part II concludes that the commensurate with income standard is fully consistent with the arm's length principle.

The primary administrative difficulty relating to transfers of intangible property is the failure of the regulations to specify a so-called fourth method of income allocation for situations in which comparable transactions do not exist. This problem has been particularly acute with respect to high profit intangibles. Part III of the study explores the economic theory underlying section 482 and proposes a methodology for allocating income, and thereby determining transfer prices in such cases, which draws upon various methods that have been used on an ad hoc basis by the Service, taxpayers, and the courts. The methodology would utilize functional analysis and allocate income by using comparable transactions when they exist, arm's length rates of return when comparables do not exist, and a profit split approach when neither comparables nor arm's length rates of return can be used to allocate all intangible income.

Part IV examines cost sharing arrangements and relevant implications arising from the 1986 legislation.

The specific chapters in each part and the appendices to the study are described below.

B. PART I: BACKGROUND

Chapter 2 reviews the history of particular transfer pricing legislation and regulations before 1986, including regulations promulgated in 1968, which are still in effect today.

Chapter 3 discusses administrative problems. This chapter is supplemented by a survey of selected International Examiners and Group Managers, summarized in Appendices A and B, which sought information about how the section 482 regulations work in practice. Significant problems include access to pricing information and difficulties in applying pricing methods to transfers of intangible assets. The chapter includes recommendations regarding the maintenance of transfer pricing information in the taxpayer's books and records, which would be required to be provided to the IRS immediately upon request in an examination, summary reporting of the taxpayer's transfer pricing methodology on Forms 5471 and 5472, and the assertion of appropriate penalties for failure to disclose information or for substantial understatements of income.

The regulations place strong emphasis on finding comparable unrelated party transactions as a guide for evaluating related party transactions. Chapter 4 discusses the search for comparables in the decided cases, and concludes that comparables are often either absent or misused when transfers of intangible property are at issue.

The regulations provide that, when comparables are unavailable, some other appropriate method of allocating income among related parties may be used. Chapter 5 examines the decided cases to see what other methods have been used, including profit splits, rates of return, income to expense ratios, and customs valuations.

C. PART II: SECTION 482 AFTER THE 1986 TAX REFORM ACT

Chapter 6 focuses on the "commensurate with income" standard incorporated into section 482 by the 1986 Act. After describing the legislative history, the chapter discusses limitations some have suggested on the scope of the standard, and explains its application to normal profit potential intangibles as well as to high profit potential intangibles.

Chapter 7 addresses an issue of major concern to the foreign trading partners of the United States: compatibility with international transfer pricing standards. The chapter concludes that the arm's length standard is the accepted international norm for making transfer pricing adjustments. The study reaffirms that Congress intended the commensurate with income standard to be consistent with the arm's length standard, and that it will be so interpreted and applied by the Internal Revenue Service and the Treasury Department.

Chapter 8 discusses the need under the commensurate with income standard to make periodic adjustments to intangible income allocations. Recommendations address the issues of the frequency of review, retroactivity, lump sum payments, and set-offs.

Taxpayers and practitioners have long advocated safe harbors as a solution to many of the problems arising under section 482. Chapter 9 discusses safe harbors in theory and analyzes some of the safe harbors that have been proposed. While the Service and Treasury do not categorically reject the possibility that some useful safe harbors might be developed, none of those currently proposed appears satisfactory.

D. PART III: METHODS FOR VALUING TRANSFERS OF INTANGIBLES

The current regulations adopt a market-based approach, distributing income among related parties the way a free market would distribute it among unrelated parties. Some critics have suggested that a unitary business approach, eliminating the fiction of arm's length dealing and accounting for economies of related party dealing through a formulary method, might be more theoretically sound. Chapter 10 examines these arguments and concludes that the market-based arm's length standard remains the better theoretical allocation method.

Chapter 11 discusses the formulation of a methodology for applying the arm's length standard to transfers of intangible property. Beginning with a discussion of the use of exact and inexact comparables, the chapter proposes as an additional method an arm's length return method that, with appropriate adjustments, could be used in a large percentage of cases. For cases involving intangibles in which comparables and the arm's length return method cannot account for all income to be allocated, a profit split addition to the arm's length return method is described.

E. PART IV: COST SHARING ARRANGEMENTS

Chapter 12 presents a description of cost sharing arrangements and describes the history of their tax treatment, comparing the detailed section 482 cost sharing regulations that were proposed in 1966 with the terse version actually promulgated in 1968. The chapter reviews foreign experience with cost sharing, a 1984 Congressional recommendation that the cost sharing rules be expanded, and the special rules governing cost sharing arrangements between possessions corporations and their domestic affiliates.

The legislative history regarding the change to section 482 in the 1986 Act states that Congress intended to permit bona fide cost sharing arrangements, but expected the economic results of such arrangements to be consistent with the commensurate with income standard. Chapter 13 identifies and discusses various issues related to the use of cost sharing arrangements after the 1986 Act.

F. APPENDICES

Appendices A and B to the study summarize the results of a survey of Service personnel about the administration of section 482. Appendix C analyzes the transfer pricing law and practices of selected jurisdictions. Appendix D describes the publicly available information about third party licensing practices. Appendix E contains 14 examples that illustrate how the principles explained in the study are applied in different factual contexts.

G. FUTURE AGENDA

This study reflects input from taxpayer groups, practitioners, and other concerned members of the public, as well as the combined experience and careful thought of those in the government charged with enforcing section 482. Nevertheless, it is only a beginning; it sets forth conclusions and recommendations in some areas, and describes the need for further study in others.

In the study, input is requested on specific issues from taxpayers and practitioners. More generally, however, readers are urged to provide any comments that would be useful in formulating a fair and workable system of administering a statute that has challenged taxpayers and the government alike. It is anticipated that comments will be taken into account in drafting proposed regulations and in examining additional issues not discussed in this study -- including such areas as the services portion of the section 482 regulations, the impact of currency fluctuations on transfer pricing, a more detailed review of functional analysis, and the proper methodology for valuing assets under the various "fourth method" approaches described in Chapter 11.

Comments should be forwarded in triplicate to the Office of Associate Chief Counsel (International), Branch 1, 950 L'Enfant Plaza South, S.W., Room 3319, Washington, D.C. 20024. Comments are requested to be filed by February 15, 1989.

CHAPTER 2

TRANSFER PRICING LAW AND REGULATIONS BEFORE 1986

A. EARLY HISTORY

The Commissioner was generally authorized to allocate income and deductions among affiliated corporations in 1917. ³ He could require related corporations to file consolidated returns "whenever necessary to more equitably determine the invested capital or taxable income. . . ." The earliest direct predecessor of section 482 dates to 1921, when legislation went beyond authority to require consolidated accounts and authorized the Commissioner to prepare consolidated returns for commonly controlled trades or businesses to compute their "correct" tax liability. ⁴ This legislation was passed partly because possessions corporations, ineligible to file consolidated returns with their domestic affiliates, offered opportunities for tax avoidance. ⁵ As early as 1921, Congress perceived the potential for abuse among related taxpayers engaged in multinational transactions.

When the predecessor to current section 482 was incorporated into the 1928 Revenue Act (as section 45), the provision was removed from the expiring consolidated return provisions and significantly expanded. ⁶ The Commissioner's authority to make an adjustment under section 45 was expressly predicated upon his duty to prevent tax avoidance and to ensure the clear reflection of the income of the related parties (to determine their "true tax liability," in the words of the legislative history). ⁷

B. REGULATIONS AND THE COURTS -- THROUGH THE EARLY 1960s

For many years, the small number of United States companies with multinational affiliates meant that section 482 had little impact in the international context. Prior to the early 1960s, the primary focus of the Service's enforcement efforts using section 482 was domestic. Regulations issued in 1935 ⁸ (under section 45) remained in effect substantially unchanged until 1968.

The regulations set forth the arm's length standard as the fundamental principle underlying section 482: "The standard to be applied in every case is that of an uncontrolled taxpayer dealing at arm's length with another uncontrolled taxpayer." ⁹ They did not, however, mandate the use of any particular allocation method.

The case law interpreting section 482 and its predecessors took a broad approach. The concepts of "evasion of taxes" ¹⁰ and "clear reflection of income" ¹¹ were developed into far-reaching weapons to attack a variety of tax abuses. The predecessors of section 482 were used to prevent recognition of a tax loss on securities following a tax-free transfer from the corporation that had incurred, but could not use, the loss, ¹² and to prevent the mismatching of the expenses incurred by one corporation in growing crops from the income artificially realized by another corporation from harvesting and selling those crops. ¹³

The courts applied a number of different standards for determining when transactions were conducted at arm's length. Transactions were scrutinized to determine if related parties received full, fair value, ¹⁴ a fair and reasonable price, ¹⁵ or a fair price including a reasonable profit. ¹⁶

Before 1964, it was generally understood that section 482 could not be used by the Service to place taxpayers, in effect, on a consolidated return basis. ¹⁷ In 1964, the Tax Court used section 482 to combine the incomes of two separate corporations that operated a downtown clothing store and its suburban branch store. ¹⁸ This case raised concerns among taxpayers over the use of section 482 in substance to ignore separate corporate entities.

C. DEVELOPMENTS IN THE 1960s

By the early 1960s, the business and regulatory climate in which U.S. and foreign multinationals operated changed substantially. In 1961, the Treasury Department urged that significant changes be made in the taxation of U.S. enterprises with foreign affiliates. In particular, Treasury contended that section 482 was not effectively protecting U.S. taxing jurisdiction. ¹⁹

In 1962, Congress considered how to stop U.S. companies from shifting U.S. income to their foreign subsidiaries. ²⁰ While the Ways and Means Committee observed that under existing law the Service could prevent this practice by allocating income under section 482, it proposed further legislation to minimize "the difficulties in determining a fair price," particularly in instances "where there are thousands of different transactions engaged in between a domestic company and its foreign subsidiary." ²¹

The Ways and Means Committee proposal as adopted by the House would have added to section 482 a new subsection dealing with sales of tangible property between U.S. corporations and their foreign corporate affiliates. ²² Unless the taxpayer could demonstrate its use of an arm's length price under the comparable uncontrolled price method, taxable income was to be apportioned between related parties under a formula based on their relative economic activities. In addition, no income was to be allocated to a "foreign organization whose assets, personnel, and office and other facilities which are not attributable to the United States are grossly inadequate for its activities outside the United States." ²³

The Senate version of the 1962 Revenue Bill, which prevailed in conference, omitted the House provision. Instead, the Finance Committee concluded that section 482 already provided ample regulatory authority to prevent improper multinational allocations. ²⁴ The Conference Committee endorsed this approach, stating:

The conferees on the part of both the House and the Senate believe that the objectives of section 6 of the bill as passed by the House can be accomplished by amendment of the regulations under present section 482.

Section 482 already contains broad authority to the Secretary of the Treasury or his delegate to allocate income and deductions. It is believed that the Treasury should explore the possibility of developing and promulgating regulations under this authority which would provide additional guidelines and formulas for the allocation of income and deductions in cases involving foreign income. ²⁵

D. THE CURRENT REGULATIONS

Treasury responded by promulgating regulations, issued in final form in 1968, that (with only a few changes) govern transfer pricing practices today. ²⁶ Those regulations reaffirmed the arm's length standard as the principal basis for transfer pricing adjustments but attempted, for the first time, to establish rules for specific kinds of intercompany transactions. The final regulations applied to the performance of services, the licensing or sale of intangible property, and the sale of tangible property. ²⁷

1. SERVICES. In determining an arm's length charge for services, section 1.482-2(b)(3) of the regulations provides:

For the purpose of this paragraph an arm's length charge for services rendered shall be the amount which was charged or would have been charged for the same or similar services in independent transactions with or between unrelated parties under similar circumstances considering all relevant facts.

The regulations do not provide any specific guidance for determining what the charge in independent transactions would have been in the absence of comparable transactions with independent parties.

2. INTANGIBLE PROPERTY. As to the licensing or sale of intangible property, section 1.482-2(d)(2)(ii) of the regulations provides:

In determining the amount of an arm's length consideration, the standard to be applied is the amount that would have been paid by an unrelated party for the same intangible property under the same circumstances. Where there have been transfers by the transferor to unrelated parties involving the same or similar intangible property under the same or similar circumstances the amount of the consideration for such transfers shall generally be the best indication of an arm's length consideration.

The intangible property portion of the regulations contemplate a failure to find appropriate comparables. Where they are unavailable, the regulations list 12 factors to be taken into account, including prevailing rates in the industry, offers of competitors, the uniqueness of the property and its legal protection, prospective profits to be generated by the intangible, and required investments necessary to utilize the intangible. ²⁸ The regulation offers little or no guidance, however, in determining how much relative importance particular factors are to be given.

3. TANGIBLE PROPERTY. Finally, the section 482 regulations set out detailed rules for determining the transfer prices of tangible personal property. Section 1.482-2(e)(2)(4) of the regulations describes three specific methods for determining an appropriate arm's length price: the comparable uncontrolled price method, the resale price method, and the cost plus method. All three rely on comparable transactions to determine an arm's length price, either directly or by reference to appropriate markups in comparable unrelated transactions. The regulations mandate that the three enumerated methods be used in the order set forth. They also authorize other unspecified methods, which have come to be known generically as "fourth methods":

Where none of the three methods of pricing . . . can reasonably be applied under the facts and circumstances as they exist in a particular case, SOME APPROPRIATE METHOD OF PRICING other than those described in subdivision (ii) of this subparagraph, or variations on such methods, can be used. [Emphasis supplied.] ²⁹

The specific transaction-oriented models described above for making transfer pricing determinations were adopted in lieu of "mechanical safe havens" based on profit margins, percentage mark-ups or mark-downs, and the like, which had been suggested by various taxpayers commenting on proposed regulations issued in 1965 and 1966. Such safe harbors were rejected for two reasons. First, because of the extraordinary range of returns earned at arm's length, even within a single industry or company, no principled and equitable basis for such safe harbors could be devised. Second, any effective safe harbor income allocation would inevitable serve as a "floor," applying only to those taxpayers not able to document a more advantageous fact pattern. ³⁰ As discussed in Chapter 9, infra, the concerns raised by safe harbors still have not been satisfactorily dispelled.

E. CONCLUSION

In general, the section 482 regulations relating to services, intangible property, and tangible property rely heavily on finding comparable transfer prices or comparable transactions. The regulations provide little guidance for determining transfer prices in the absence of comparables.

CHAPTER 3

RECENT SERVICE EXPERIENCE IN ADMINISTERING SECTION 482

In order to determine what difficulties International Examiners are encountering in administering the regulations, a questionnaire was prepared through the joint efforts of Treasury, Chief Counsel, and International Examination personnel. The questionnaire was sent to selected International Examiners (IEs) and IE Group Managers. In addition, selected IRS economists, IEs, Group Managers, and IRS trial attorneys were interviewed.

The results of the analysis of the questionnaires have been compiled and are set forth in Appendix A. Included in Appendix B is a completed questionnaire reflecting the aggregate data supplied by the respondents. ³¹ In general, the survey and interviews revealed no surprises. The two primary problems in administering section 482 have been the difficulty of obtaining pricing information from the taxpayers during an examination and the difficulty of valuing intangibles -- including the valuation of intangible property in connection with sales of tangible property. This chapter discusses these problems, makes several related suggestions regarding disclosure of information and penalties, and suggests that the early use of counsel and economic experts would alleviate these problems.

A. SERVICE'S ACCESS TO PRICING INFORMATION

A significant threshold problem in the examination of section 482 cases has been IRS access to relevant information to make pricing determinations. In some cases, relevant information is not furnished by the taxpayer to the examining agent. ³² In other cases, long delays are experienced by agents in receiving information, in most cases without explanation for the delays. In many cases, delays in responding to IE requests for information exceed one year. ³³ Because of the emphasis upon timely closing of large cases in the recent past, section 482 cases have been closed without receiving necessary information or without the opportunity for agents to follow up on information that has been provided. ³⁴

The experience of the agents has been that the vast majority of taxpayers, when asked, are unable to provide an explanation of how their intercompany pricing was established. ³⁵ This may account in large part for the denial of access to information and delays encountered by IEs.

In recent years the Service has placed an emphasis on examination of transactions with subsidiaries located in tax haven jurisdictions. ³⁶ Because of the financial and commercial secrecy laws that exist in tax haven jurisdictions, IRS access to third party data has been significantly hampered. Problems with access to information because of foreign secrecy laws have been to some extent alleviated by the enactment of section 982 ³⁷ and by a broad interpretation of the IRS administrative summons power by the courts. ³⁸ However, as will be subsequently discussed, agents have failed to use the section 982 and administrative summons procedures aggressively.

Because of the dramatic increase in recent years in direct foreign investment in the United States, ³⁹ the examination of transactions between foreign parents and their U.S. affiliates will become an increasingly more important part of the international examination program. A survey of rates of return on these companies based on IRS statistics of income ("SOI") data reveals a substantially lower than average profit in this country reported by these companies, which may involve transfer pricing policies. ⁴⁰

In practice, examinations of United States subsidiaries of foreign parents have developed into some of the Service's most difficult examinations. A primary reason for the difficulty is that agents are unable to obtain timely access to necessary data, which is typically in the hands of the parent company. In many cases, foreign parent companies refuse to produce this information upon request. An additional difficulty encountered by agents is that foreign parent corporations may not be subject to information reporting requirements similar to U.S. requirements. ⁴¹

Both the administrative summons procedures ⁴² and the formal document request procedures ⁴³ are tools that are available to IEs to compel production of information necessary to determine whether a section 482 adjustment is appropriate. Unfortunately, for a variety of reasons, IEs seldom serve administrative summonses or section 982 requests. ⁴⁴ The most common reason given for failing to use these procedures is the time delay necessary to follow them, which conflicts with the need to close the examination. Another reason given in many cases was the necessity of maintaining a good working relationship with the taxpayer, which IEs feared would be harmed if these procedures were used.

Although section 6001 contains a general requirement that a taxpayer maintain adequate books and records, the section 482 regulations are generally silent with regard to records and their accessibility to either support or to determine arm's length prices. ⁴⁵ Thus, the current regulations do not advise taxpayers specifically of the type of information that is necessary in order to determine compliance with section 482. Specific information on transactions between parent and subsidiary corporations is required on forms 5471 and 5472, which have been widely used by agents in planning and conducting section 482 examinations.

Service experience has been that many taxpayers do not rely upon any form of comparable transactions or other contemporaneous information either in planning or in defending intercompany transactions. ⁴⁶ Although the legislative history to the 1986 Act expresses concern that industry average royalty rates are used by taxpayers to justify royalties for high profit intangibles, ⁴⁷ the more serious problem has been that the taxpayer, not having structured the transaction with any comparable in mind, seeks to defend its position by finding whatever transaction or method gets closest to the transfer price initially chosen, whether that be an industry average rate of return or some other type of comparable.

Problems related to information and aggressive return positions would be alleviated if the regulations specifically set out a taxpayer's responsibility to document the methodology used in establishing intercompany transfer prices prior to filing the tax return and to require that such documentation be provided within a reasonable time after request. The documentation should include references to any comparable transactions, rates of return, profit splits, or other information or analyses used by the taxpayer in arriving at transfer prices. In general, a taxpayer making relatively minor investments would not be required to obtain information regarding comparable transactions outside of its own knowledge of its business affairs and those of its competitors, but to use information and analyses that generally would have been produced by the taxpayer in the course of developing its business plan. However, a taxpayer engaging in a major transaction or one involved in a complex profit split analysis involving significant high profit intangibles ⁴⁸ would be expected to gather and analyze the types of information illustrated by the examples in Appendix E which, once again, is information likely to be produced by the taxpayer in developing its business plan. In the absence of comparables, taxpayers should be required at a minimum to apply a rate of return analysis or profit split methodology that may be prescribed in regulations under which the taxpayer would identify assets and functions performed by it and its affiliates and identify the rate of return or profit split that the taxpayer believes should be assigned or allocated to each activity or function.

Furthermore, Forms 5471 and 5472 should be revised to include summary information describing how intercompany prices were determined and an attestation that the documentation required to be maintained under the section 482 regulations, as described above, was available at the time of preparation of the return and will be made available at the start of an IRS examination. Requiring information to be made available at the beginning of an audit would alleviate problems of receiving either too little or too much information near the expiration of the statute of limitations.

The Service and Treasury believe that taxpayer compliance in the transfer pricing area with respect both to disclosure of information and to conformity with the arm's length standard would be enhanced by the proper assertion of appropriate penalties. While the penalty imposed by section 6661 for substantial understatement of tax can apply, to date the Service has only infrequently imposed penalties in connection with making section 482 adjustments. ⁴⁹ The Internal Revenue Service is currently engaged in a comprehensive study of the role of civil tax penalties, ⁵⁰ as are many other interested parties. It, therefore, seems timely to focus now on the effectiveness of existing penalties in encouraging compliant taxpayer behavior and penalizing unjustified positions in the transfer pricing area. Consideration should be given to when the section 6661 penalty should be raised and whether it is adequate to deter instances where taxpayers do not make intercompany pricing decisions upon a reasonable basis, or whether a new penalty should be proposed.

The Service and Treasury are interested in recommendations in this area, including such specific comments as to the type and amount of penalties, and whether there should be certain transaction oriented thresholds that ought to apply before any penalty could be asserted. For example, a transaction specific penalty (similar to the overvaluation penalty of section 6659) may be an appropriate means of deterring substantial deviations from the commensurate with income standard. Specific consideration should be given to whether the applicable penalty provisions should be amended to apply if there is a substantial deviation from the appropriate commensurate with income payment regardless of whether there is disclosure on the tax return of the manner in which taxpayers computed transfer prices. Disclosure of the taxpayer's method of computing a transfer price can not adequately inform the Service as to whether such a transfer price substantially deviates from the appropriate section 482 transfer price absent a thorough audit. Consequently, such disclosure should not prevent the imposition of a penalty for substantial deviation from the correct section 482 transfer price. ⁵¹ Since it is possible to use the provisions of section 367(d) to deter abusive situations (see discussion of section 367(d) infra Chapter 6), it may also be appropriate to clarify how taxpayers may avoid imposition of penalties in the context of section 367 adjustments.

B. INTANGIBLES

A significant portion of section 482 adjustments proposed in recent years have involved an adjustment for pricing with respect to the licensing or other transfer of intangibles. ⁵² Because of the absence of comparables in many cases, intangible transfers generally are the most problematic of adjustments due to the inherent difficulty of valuing intangibles under the existing regulations. As previously noted in Chapter 2, the intangible property portion of the regulations contemplate a failure to find appropriate comparables and list 12 factors to be taken into account in valuing intangibles in the absence of comparables. No guidance is given, however, in determining the relative importance of particular factors.

In a significant number of cases, IEs relied upon sections of the regulations other than the intangibles portion to make a transfer pricing adjustment. ⁵³ Intangibles are often transferred by incorporation into tangible property that is sold or rented. In these types of cases, the taxpayers have not been required to isolate the value of the intangible. ⁵⁴ Incorporating a return on an intangible in a transfer price for tangible property does not alleviate, however, the difficulty of valuing the intangible.

A common example is the transfer of tangible property with a trademark, trade name, or recognizable logo attached. It is clear from the regulations that a trademark, trade name, or logo is an intangible. ⁵⁵ The regulations governing the sales of tangible property specify that, in applying the comparable uncontrolled price, resale price, and cost plus methods, adjustments must be made for sales with or without trademarks, provided there is a reasonably ascertainable effect on the price. ⁵⁶ In some cases adjustments for trademarks are relatively easy to make. The analysis, however, becomes much more complex if there are no similar products sold (with or without trademarks) on which to base a comparison. Setting a transfer price for a product in such a case involves the same difficult exercise as setting a royalty rate for a licensed intangible. One of the recent pharmaceutical cases presents an example of this latter situation since it involved the sale of unique pharmaceutical products. ⁵⁷

Intangibles may also be transferred in the form of services. In some circumstances, taxpayers have attempted to shift large amounts of income to tax haven subsidiaries by "loaning" a few key employees to a tax haven affiliate. By loaning employees, the parent company may simultaneously provide services and transfer valuable intangible know-how. In transactions which are structured as an intangibles transfer, it is difficult to value services rendered in connection with the transfer of intangible property, which may be necessary for purposes of determining the source of the income. ⁵⁸

A particularly difficult aspect of valuing intangibles has been determining what part of an intangible profit is due to manufacturing intangibles and what part is due to marketing intangibles. ⁵⁹ This problem has particular significance in section 936, since the possessions corporation is generally entitled to a return only on manufacturing intangibles when it elects the cost sharing method under section 936(h).

Problems with intangibles underlie the amendment made to section 482 by the 1986 Act, as discussed in Part II. The intangibles section of the section 482 regulations should be modified to provide a specific analysis to be used when comparable uncontrolled transactions do not exist. The method should provide for appropriate allocations of income when multiple intangibles (such as marketing and manufacturing intangibles) are present in the same set of transactions. Part III is devoted to the subject of an appropriate methodology for allocating intangible income.

C. APPLICATION OF PRICING METHODS FOR TRANSFERS OF TANGIBLE PROPERTY

When considering an adjustment with respect to the transfer price for tangible property, the regulations require both the taxpayer and the Service to follow a priority of pricing methods: first, the comparable uncontrolled price method must be attempted, then resale price method, then cost plus method, and, if none of them are applicable, some other method or combination of the prior methods. ⁶⁰ Five prior studies using data available from both the Service and multinational corporations have examined the frequency with which each of these methods has been used. The results of these surveys are set forth below:

                            Percentage of Cases in which Various

One justification given for the current priority of methods in the regulations is that both the taxpayer and the Service are thus directed to a common frame of analysis to avoid the problem of the Service using one method while the taxpayer uses another method. However, as currently structured, the regulations literally require that both the taxpayer and the agent attempt to apply the methods in priority order. Because the resale price method generally applies only to distributors of goods, while the cost plus method applies generally to manufacturers, there does not seem to be any reason in theory why the agent or taxpayer should attempt to apply the resale price method before applying the cost plus method. ⁶⁸ In practice, taxpayers and agents rely upon comparable uncontrolled prices or transactions, when they exist. When they do not exist, agents or taxpayers use whatever method they believe best reflects the economic realities of the transaction at issue. While there are valid theoretical reasons for retaining the priority of the comparable uncontrolled price method, ⁶⁹ there do not seem to be any valid reasons for preferring resale price over cost plus or another method, or for preferring resale price or cost plus over some other economically sound method. Rather, the method used should generally be the one for which the best data is available and for which the fewest number of adjustments are required.

One technique that is missing from the section 482 regulations that in practice is used extensively by the international examiners is functional analysis. This analysis focuses on the economic functions performed by the affiliated parties to a transaction and the economic risks borne by each of the parties. ⁷⁰ This technique is used by IEs and Service economists not as a method standing alone but rather as a means of verifying that prices or transactions are truly comparable to the situation under examination or as a basis for a fourth method.

As discussed in section B, intangibles are often transferred by incorporation into tangible property that is sold, and setting a transfer price for a product in such a case involves the same difficult exercise as setting a royalty rate for a licensed intangible. The difficulty of valuing intangibles is, therefore, as much a problem in the context of sales of property as in the case of licenses or other transfers of intangibles.

D. USE OF SPECIALISTS AND COUNSEL

The use of counsel and economic specialists at the examination level would ameliorate some of the problems, discussed above, of obtaining information and dealing with difficult intangible pricing cases. Legal assistance during examination is needed to assist in obtaining relevant information and in determining whether an appropriate legal basis exists for a proposed adjustment. Economists are needed in many cases to perform a functional analysis and to help evaluate the proper returns to be accorded to the related parties. Other experts may be required to analyze practices within the taxpayer's industry. The goal of the attorney, the economist, and other specialists should be to assist the IE in obtaining all relevant facts and to determine whether an adjustment may be sustained on appropriate legal and economic theories if the matter ever results in litigation.

For section 482 cases developed 10 years ago, it would have been normal for the IE to develop the case without the assistance of an economist or without the assistance of a Chief Counsel attorney. Authority and expertise in international tax matters were then split between the National Office Examination function and the Director, Foreign Operations District. Legal expertise in international tax matters was diffused among at least four national office divisions and was limited in field offices.

In May 1986 the Office of the Assistant Commissioner (International) was created to provide an emphasis upon, and a focal point for, development of international issues at the examination stage. The Office of Associate Chief Counsel (International) was created in March 1986 to provide a similar focal point for legal issues. In addition, a network of International Special Trial Attorneys and senior District Counsel attorneys has been created to litigate significant international tax cases, including section 482 cases. More importantly, these field attorneys and their National Office counterparts have been encouraged to assist the field in developing these cases, and IEs are encouraged to use their assistance. ⁷¹

Within the last several years, the Service has substantially increased the number of economists available to assist IEs and has decentralized those activities from the national office to three key District offices: Baltimore, New York, and Chicago. Use of economists in major section 482 examinations that do not involve safe harbors is now required. ⁷²

One criticism that has been made concerning the more extensive use of counsel and experts at the examination stage is that the time necessary to complete an examination (already lengthy) will be further extended. Service experience has been, however, that increased use of specialists has not unduly delayed disposition of the examination in the vast majority of the cases. ⁷³ Furthermore, the early use of specialists in some cases will prevent erroneous adjustments from ever being made, thus saving both taxpayers and the government substantial sums of time and money.

E. CONCLUSIONS & RECOMMENDATIONS

ACCESS TO PRICING INFORMATION

1. The failure of the taxpayer to document the methodology used to establish transfer prices under the section 482 regulations and delays or failure by taxpayers in supplying information to IEs are significant problems that hamper the IRS in its administration of section 482.

2. The section 482 regulations are deficient in not requiring taxpayers to document intercompany pricing policies and to supply information upon examination.

The section 482 regulations should be amended to require taxpayers to document the methodology used to establish transfer prices prior to filing the tax return and to provide such documentation during examination within a reasonable time after request. The documentation should include references to any comparable prices or transactions, rates of return, profit splits or other information or analysis used by the taxpayer in arriving at the transfer price.

3. Forms 5471 and 5472 should be revised to include: (a) summary information describing how intercompany prices were determined; and (b) an attestation that the documentation described in paragraph 2, supra, was available at the time of preparation of the return and will be made available at the start of an IRS examination.

4. IEs experiencing difficulties in obtaining transfer pricing information have failed to deal with non-compliant taxpayers through the issuance of section 982 requests and administrative summonses. The Service should more aggressively pursue noncompliant taxpayers that delay, without justification, in producing relevant pricing information by using the section 982 and administrative summons procedures.

5. The assertion of appropriate penalties is a necessary but often ignored element of transfer pricing compliance. In conjunction with the Service's broad-based review of penalties, the Government should determine whether existing penalties are sufficient to: (a) compel taxpayers to provide thorough and accurate information as set forth in paragraphs 2 and 3 supra; and (b) deter taxpayers from setting overly aggressive and unjustified transfer prices that are inconsistent with the commensurate with income standard. If it is felt that existing penalties are inadequate, legislative solutions should be pursued. The Service and Treasury encourage comments in this area, including the type of penalty, such as a transaction based penalty, that might be proposed.

INTANGIBLES

6. Establishing appropriate transfer prices for intangibles has been a significant problem because of the inherent difficulty of valuing intangibles -- particularly when intangibles are transferred simultaneously with the transfer of tangible property or the provision of services.

7. The intangibles section of the section 482 regulations should be modified to provide a specific method of analysis to be used when comparable uncontrolled transactions do not exist. This method should provide for appropriate allocation when multiple intangibles (such as marketing and manufacturing intangibles) are present in the same set of transactions. Part III is devoted to the subject of an appropriate methodology for allocating intangible income.

APPLICATION OF PRICING METHOD FOR TRANSFERS OF TANGIBLE PROPERTY

8. The current priority for the comparable uncontrolled price method should be retained, since such prices generally provide the best evidence of what unrelated parties would do in an arm's length transaction. There does not appear to be any reason to retain the current priority of the resale price method over the cost plus method, or for preferring resale price or cost plus over some other economically sound method. Rather, the method used should generally be the one for which the best data is available and for which the fewest number of adjustments are required.

9. Since intangibles are often incorporated into tangible property that is sold, the difficulty of valuing intangibles is as much of a problem in many transfers of tangible property as in the context of licenses or other transfers of intangible property.

USE OF SPECIALISTS AND COUNSEL

10. The use of counsel and economic specialists at the examination level would ameliorate the problems of obtaining information and dealing with the difficult intangible pricing cases. Chief Counsel attorneys familiar with transfer pricing issues should be involved in significant cases at an early stage to make sure that relevant information necessary for the examination is being obtained and that a technical basis for a potential adjustment exists. An economist needs to be involved at an early stage to perform a functional analysis and to evaluate the proper returns to be accorded to the related parties. The goal of the attorney, the economist, and other specialists should be to assist the IE in obtaining all relevant facts and to determine whether an adjustment may be sustained on appropriate legal and economic theories if the matter ever results in litigation.

CHAPTER 4

THE SEARCH FOR COMPARABLES

A. INTRODUCTION

As explained in Chapter 2, the section 482 regulations rely heavily on finding comparable goods, services, and intangibles to determine whether an arm's length price has been used. Where such comparables exist -- where arm's length transactions bearing a reasonable economic resemblance to those being examined have occurred in the free market -- application of the regulations is relatively straightforward. Where no comparables can be found, or where similar items are only distantly comparable, the regulations leave the Service, the taxpayers, and the courts with little guidance.

This chapter examines several recent cases decided under section 482 to assess the use of comparables by the parties and the courts, whether in the context of either sales of tangible property, the provision of services, or licenses or other transfers of intangible property. These cases show that comparables are often difficult to locate, and may be misused or misinterpreted even if they are found. In most of the cases discussed in this chapter, no comparables were available. The courts' resolution of the issues in the absence of comparables is discussed in Chapter 5.

B. SPECIFIC COMPARABLES

In recent years, transfer pricing cases involving highly profitable products -- which usually are associated with unique intangibles -- have severely tested the comparables approach of the present section 482 regulations. This problem is illustrated by the Lilly ⁷⁷ case. In Lilly, the U.S. parent corporation, Lilly U.S., transferred highly profitable manufacturing intangibles, including patents and know-how (primarily relating to the drugs Darvon and Darvon-N), to its newly-formed U.S. subsidiary in Puerto Rico, Lilly P.R., in a tax-free exchange for Lilly P.R. stock under section 351. The Service took the position that the income associated with those intangibles should be allocated to Lilly U.S., notwithstanding their tax-free transfer to Lilly P.R.

In preparation for trial, the government's experts surveyed the most successful U.S. pharmaceutical products. They discovered that the patents to such products were rarely transferred, except to a related party. The government argued that unrelated parties would not have transferred the Darvon intangibles and that, accordingly, there were no comparable marketplace transactions. While the Tax Court did not fully subscribe to the government's theory of the case, it nevertheless was not able to find appropriate comparables for the patented products in question for the first 2 years at issue, 1971 and 1972. ⁷⁸ The court proceeded to make its own allocations, basing its adjustments on the proposition that a distortion of income was created by the transfer of intangibles from Lilly U.S. to Lilly P.R. in exchange for Lilly P.R. stock. In the Tax Court's view, the distortion arose because it felt that Lilly would have demanded a stream of income from the transferred Darvon intangibles in order to fund a proportionate part of its ongoing general research and development efforts. The Tax Court also used a profit split approach to increase the return of Lilly U.S. on marketing expenditures and intangibles. ⁷⁹ On appeal, the Seventh Circuit rejected the Tax Court's allocation to support research and development, but affirmed its profit split methodology.

In Searle, ⁸⁰ the petitioner transferred the patents (or licenses) on its most successful pharmaceutical products to its U.S. subsidiary, SCO, operating in Puerto Rico. These intangibles represented products accounting for approximately 80 percent of the petitioner's profits and sales. As in Lilly, the government argued that a section 482 allocation from SCO to the petitioner was appropriate.

The petitioner, relying on section 1.482-2(d)(2)(ii) of the regulations, argued that, since it had originally acquired two of the transferred intangibles by licensing agreements carrying royalties of ten percent and eight percent of net sales, an unrelated party would not have paid more than a royalty in this range for the intangible property transferred to SCO. The court, however, found that the original licenses were not comparable; the products were licensed from European pharmaceutical firms prior to their approval by the FDA, and thus could not have been marketed in the United States at the time of the license. The court concluded that the intangibles to SCO were significantly more valuable than the "mere licensing agreements" upon which the taxpayer relied ⁸¹

Ultimately, the court found, despite the voluminous record, that "there is little hard evidence from which we can determine what consideration petitioner would have demanded had the transactions under scrutiny here taken place between unrelated parties dealing at arm's length." ⁸²

Problems with finding or applying comparables for valuable intangibles have not been limited to pharmaceutical companies. In Hospital Corporation of America, ⁸³ a U.S. hospital management company, HCA, entered into negotiations to recruit professional and non-professional staff to manage a state-of-the-art hospital in Saudi Arabia. It formed a Cayman Islands corporation, LTD, ostensibly to negotiate and perform the management contract. HCA performed services for LTD and made available at little cost all of its know-how, experience, management systems knowledge, and other intangibles. The parties offered no evidence of comparable transactions, and the court identified none. Nevertheless, the court allocated 25% of the income to LTD as compensation for its management service.

The Tax Court was also unable to find appropriate comparables in Ciba-Geigy Corp. v. Comm'r., ⁸⁴ where the Service sought to reduce royalties paid by a U.S. subsidiary to its foreign parent for the rights to manufacture and sell a herbicide. Unlike the approach taken by the courts in Lilly for 1973, where multiple adjustments were made to a third party transaction in order to determine a comparable price, the court in Ciba-Geigy rejected as comparables licenses of the same product to unrelated parties because of differences in geographic markets, years of the license, and differences in required purchases of raw materials. ⁸⁵ Instead the court relied upon testimony from an unrelated party about what his company would have been willing to pay in the form of a royalty for the same rights.

The comparability of third party resale price margins was at issue in E.I. DuPont de Nemours & Co. v. United States [DuPont]. ⁸⁶ In that case, the U.S. parent company incorporated DuPont International S.A., DISA, in Swizerland to serve as a super distributor of DuPont products in Europe. Internal DuPont memos indicated that DuPont planned to sell its goods to DISA at prices below fair market value, so that on resale most of the profits would be reported in a foreign country having much lower tax rates than the United States. ⁸⁷ Although for many products DISA performed no special services for either DuPont or its customers, DuPont structured its pricing to DISA anticipating that the latter would capture 75 percent of the total profits involved, although DISA actually realized less than this percentage. The Service reallocated much of this profit back to the parent.

The government introduced expert testimony at trial to the effect that, after the allocations, DISA's ratio of gross income to total operating costs was greater than that achieved by 32 specific firms that were functionally similar to DISA. Additionally, it was shown that, after the allocations, DISA's return on capital was greater than that of 96 percent of 1133 companies surveyed. ⁸⁸

The taxpayer, on the other hand, relied solely upon the resale price method. It contended that similar companies selling similar products experienced average markups of between 19.5 and 38 percent, comparing favorably with DISA's 26 percent gross profit margin. In rejecting the taxpayer's position the court made the following comments:

Taxpayer tells us that a group of 21 distributors, whose general functions were similar to DISA's, provides the proper base of comparison. Beyond the most general showing that this group, like DISA, distributed manufactured goods, there is nothing in the record showing the degree of similarity called for by the regulation. No data exist to establish similarity of products (with associated marketing costs), comparability of functions, or parallel geographic (and economic) market conditions. Rather, the record suggests significant differences. Defendant has introduced evidence that the six companies plaintiff identifies most closely with DISA all had average selling costs much higher than DISA. Because we agree with the trial judge and defendant's expert that, in general, what a business spends to provide services is a reasonable indication of the magnitude of those services, and because plaintiff has not rebutted that normal presumption in this case, we cannot view these six companies as having made resales similar to DISA's. They may have made gross profits comparable to DISA's but their selling costs, reflecting the greater scale of their services or efforts, were much higher in each instance. Moreover, the record shows that these companies dealt with quite different products (electronic and photographic equipment) and functioned in different markets (primarily the United States). ⁸⁹

Another case that raised questions of comparability is United States Steel v. Comm'r. ⁹⁰ There, the Service contended that Navios, the petitioner's wholly owned shipping company, was charging the petitioner more than an arm's length rate for shipping ore from Venezuela to United States ports. The government relied on evidence that, had U.S. Steel contracted with other shippers for the same tonnage per year, it would have paid considerably lower rates. The petitioner countered that, because Navios charged unrelated steel producers the same rate as the petitioner, a perfect comparable was available from which to determine an arm's length price. The government contended that the unrelated third party transactions were not comparable because they were few in number, they were not based on a continuing long-term relationship, and the volume shipped was much smaller than the ten million tons annually shipped by Navios for U.S. Steel.

The Tax Court did not decide the case on the basis of comparables. Instead, the court focused on constructed freight charges and on the profit that the tax haven subsidiary was projected by the taxpayer to earn on the activities it undertook.

On appeal, the Second Circuit held that, if appropriate comparables were available to support the petitioner's prices, no section 482 allocation would be sustained despite evidence tending to show that the activities resulted in a shifting of tax liability among controlled taxpayers. ⁹¹ The appellate court accepted the third party transactions as comparables and reversed the Tax Court on this issue, notwithstanding the substantial economic differences from the related party transactions.

C. INDUSTRY STATISTICS AS COMPARABLES

The Service and taxpayers have relied on industry statistics in several cases to justify or defend against section 482 allocations. Industry statistics have generally been offered as evidence of comparable uncontrolled prices or for markup percentages under the resale price or cost plus methods. The courts, however, have been reluctant to accept such statistics in the absence of a specific showing of comparability.

In the DuPont case, discussed supra, the taxpayer relied on gross profit margins of drug and chemical wholesalers contained in the Internal Revenue Service's Source Book of Statistics of Income for 1960 to support a gross profit margin of 26 percent. The gross profit margins of these companies averaged 21 percent and ranged from 9 to 33 percent. The court noted that in applying the resale price method it was necessary to find substantially comparable uncontrolled resellers. Because there was no indication from the Source Book that the necessary degree of comparability was present, the court rejected the taxpayer's industry statistics.

The government relied upon the Source Book of Statistics of Income in PPG Industries Inc. v. Comm'r, ⁹² to allocate a substantial portion of the income of a Swiss corporation to its U.S. parent. Rejecting this approach, the court found the Source Book evidence wanting because it could not be determined whether comparable transactions were involved.

In Ross Glove Co. v. Comm'r, ⁹³ the Service allocated income from a foreign glove manufacturer to its U.S. parent. The government relied upon expert testimony that the glove manufacturing industry was not a high profit industry, and that a typical glove manufacturer rarely had a year in which gross profits equaled three percent of sales. The court rejected this testimony because it did not relate to the rate of return earned by Philippine glove manufacturers, such as the taxpayer's subsidiary, whose profits generally were higher than those of U.S. manufacturers. Industry statistics were also rejected as unreliable in Edwards v. Comm'r ⁹⁴ and in Nissho Iwai American Corp. v. Comm'r. ⁹⁵

D. THE REGULATIONS IN THE ABSENCE OF COMPARABLES

The only detailed transfer pricing methods in the regulations rely in one way or another on comparables. The cases discussed in this chapter, in which comparables were generally unavailable, suggest that the regulations fail to resolve the most significant and potentially abusive fact patterns. This failure was noted both in the Court of Claims opinion and the trial judge's opinion in DuPont. Trial Judge Willi, after finding for the government, suggested that the current regulatory structure was wholly inadequate:

At least where the sale of tangible property is involved, the Commissioner's regulations seem to accommodate nothing short of a "pricing method" to determine the question of an arm's length price. Treas. Reg. section 1.482-2(3)(e)(1)(iii). Moreover, as plaintiff has correctly noted, the regulation approach seems to rule out net profit as a relevant consideration in the determination of an arm's length price, this despite Congress' encouragement to the contrary, as expressed in H. R. Rep. No. 2508, 87th Cong. 2d Sess. 18-19 (1962) (Conference Report).

As evidenced by the magnitude of the record compiled in this case, the resolution by trial of a reallocation controversy under section 482 can be a very burdensome, time-consuming and obviously expensive process -- especially if the stakes are high. A more manageable and expeditious means of resolution should be found. ⁹⁶

In difficult cases for which comparable products and transactions do not exist, the parties and the courts have been forced to devise ad hoc methods of their own -- so-called "fourth methods" -- to determine appropriate allocations of income. The next chapter describes the methods that the courts have used to resolve these issues.

E. CONCLUSION

The failure of the regulations to provide guidance in the absence of comparable products and transactions has created problems in cases involving sales of tangible property, the provision of services, and licenses or other transfers of intangible property. Taxpayers and the courts have been forced to devise ad hoc "fourth methods" to resolve such cases.

CHAPTER 5

FOURTH METHOD ANALYSIS UNDER SECTION 482

A. INTRODUCTION

Although the "other method" provision of section 1.482- 2(e)(1)(iii) (commonly known as the "fourth method") by its terms applies only to tangible property transfer pricing cases, the term "fourth method" has been used to describe any case resolved by using a method not specifically described in the regulations, typically when comparable uncontrolled transactions were unavailable. This chapter discusses the use of the "fourth method" approach in the decided cases, including cases involving the sale of tangible property, the licensing or other transfer of intangible property, and the provision of services.

B. PROFIT SPLITS

The most frequent alternative method used by the courts in the absence of comparables is the profit split approach. Under this approach, the court determines the total profits allocable to the transactions at issue and simply divides them between the related parties in some ratio deemed appropriate by the court. The validity of the method, of course, rests on the accurate determination of total profits and the reasonableness of the factors used to set the profit split ratio.

An illustration of the profit split method is found in Lilly. ⁹⁷ After rejecting the resale price and cost plus pricing methods advocated by the parties because of the absence of comparables, the Tax Court attempted to find an appropriate fourth method under section 1.482-2(e)(1)(iii) of the regulations. The court cited a number of studies and surveys indicating that fourth methods were used by the Service approximately one-third of the time, and determined that a profit split approach was permissible. ⁹⁸

In adopting the profit split approach, the Tax Court relied heavily on PPG Industries Inc. ⁹⁹ The court there, in considering the allocation of income between PPG and its foreign subsidiary, applied a profit split analysis (which produced a 55-45 profit split in favor of PPG) to buttress the court's primary analysis using the comparable uncontrolled price method.

The Tax Court in Lilly also found support in Lufkin Foundry & Machine Co. v. Comm'r, ¹⁰⁰ in which it had used a profit split method. On appeal, the Fifth Circuit rejected the Tax Court's profit split approach because the court had not attempted to apply the three specific pricing methods in the regulations and because, by itself, a profit split approach was not sufficient evidence of what parties would have done at arm's length. However, the court in Lilly distinguished the Fifth Circuit's reversal of the Tax Court's holding on the following grounds:

The three preferred pricing methods detailed in the regulations are clearly inapplicable due to a lack of comparable or similar uncontrolled transactions. Petitioner's evidence amply demonstrates that some fourth method not only is more appropriate, but is inescapable. ¹⁰¹

After providing for location savings, ¹⁰² manufacturing profit, marketing profit, and a charge for ongoing general research and development performed by the parent, the Tax Court in Lilly arrived at undivided profits of $25,489,000 for 1971 and $19,277,000 for 1972. ¹⁰³ It considered these amounts to be the profits from intangibles, consisting of manufacturing intangibles belonging to Lilly P.R. and marketing intangibles belonging to Lilly U.S. The court rejected the taxpayer's argument that its marketing intangibles were of little value and assigned 45 percent of the intangible income to Lilly U.S. as a marketing profit and 55 percent of the intangible income to Lilly P.R. as a manufacturing profit. The court did not explain how it arrived at the 45-55 split, other than stating that it used its best judgment and that it bore heavily against the taxpayer because it failed to prove the arm's length prices for Lilly P.R.'s products. ¹⁰⁴ On appeal, the Seventh Circuit affirmed the Tax Court's profit split. ¹⁰⁵

The Searle ¹⁰⁶ case was tried by the Tax Court shortly after Lilly. The primary facts that distinguished Searle from Lilly were that Searle transferred nearly all of its highly profitable manufacturing intangibles to its Puerto Rican subsidiary and that Searle did not purchase the products produced in Puerto Rico, but instead marketed them in the United States as an agent for its subsidiary. While the court could not technically apply a fourth method under the regulations governing sales of tangible property (since there were no intercompany sales), the court nevertheless imposed a profit split similar in result to the profit split imposed in Lilly.

In Searle, the Tax Court did not specifically determine the revenue that each of the parties should earn from manufacturing and marketing or which party should bear the expenses of research and development and administration. While suggesting that additional royalties were due Searle for the intangibles provided to SCO, the court stated that "whether our allocation herein is considered an additional payment for services or for intangibles that were not transferred or as a royalty payment for intangibles themselves, the result is the same." ¹⁰⁷

A profit split approach is also contained in section 936(h) of the Code, added by the Tax Equity and Fiscal Responsibility Act of 1982, effective for years beginning after December 31, 1982. In general, section 936(h) authorizes a profit split election under which the combined taxable income of the possessions affiliate and the U.S. affiliate, with respect to products produced in whole or in part in the possession, will be allocated 50 percent to the possessions affiliate and 50 percent to the U.S. affiliate. If a profit split election is made, section 482 is not available for any further allocation.

The section 936(h) 50-50 profit split does not, however, provide any logical support for 50-50 profit splits in cases not falling within the narrow scope of the section. Thus, even though the Tax Court and Congress have moved in the direction of 50-50 profit splits in some limited cases, it would appear that profit splits should only be used in the absence of appropriate comparables, and then only after a careful analysis of what functions each party has performed, what property they have employed, and what risks they have undertaken. When one affiliate's role in the transactions has been extremely limited, a 50-50 profit split may not be at all appropriate.

Such a lopsided division of relevant factors occurred in Hospital Corporation of America. ¹⁰⁸ The court's opinion recites in great detail the numerous services HCA provided for LTD in negotiating the management contract and in staffing and operating the hospital, as well as the numerous intangibles that HCA provided, such as its substantial experience, know-how, and management systems. Under these circumstances it would be extremely difficult to estimate accurately the arm's length value for the large volume of services and intangibles made available. It was certainly easier for the court to look at the relative value of the functions that each party performed, so that a profit split ratio could be developed. The court in HCA did just that, adopting a 75-25 (75 for HCA and 25 for LTD) split of the profits previously reported by LTD. ¹⁰⁹ Unfortunately, there is no discernible rationale contained in the opinion for such a split.

Hospital Corporation of America, like Searle, was not a transfer pricing case and therefore was not a fourth method case under the tangibles pricing regulation. However, both of these cases illustrate that, when highly profitable, unique intangibles are at issue, traditional methods of valuation will often fail because comparables are unavailable. In these circumstances a profit split approach appears reasonable as long as it is based on a careful functional analysis to determine each party's economic contribution to the combined profit.

C. RATE OF RETURN; INCOME TO EXPENSE RATIOS

Although profit splits are being used more frequently, the courts have used other methods as well to justify transfer pricing adjustments. Two of these methods are illustrated by the DuPont ¹¹⁰ case.

In defending the Service's section 482 allocations, the government used two different methods. The first method was computing the ratio of gross income to total operating costs (known as the "Berry ratio" because it was first used by the Government's expert witness, Dr. Charles Berry). DISA's Berry ratio before the allocation was 281.5 percent of operating expenses for 1959 and 397.1 percent for 1960. After the section 482 allocation, DISA's Berry ratio was 108.6 for 1959 and 179.3 for 1960. A survey of six management firms, five advertising firms, and 21 distributors (firms which were generally functionally similar to DISA) revealed average Berry ratios ranging from 108.3 to 129.3. Thus, DISA's combined Berry ratio for 1959 and 1960 before the allocation was about three times higher than the average for the other firms. As noted by the court, in over a hundred years of those companies' experience, none of them had ever achieved the ratios claimed by DISA. Even after the allocation, its Berry ratio was somewhat higher than that of the comparable firms. ¹¹¹

The second approach, developed by Dr. Irving Plotkin, was to compare DISA's rate of return on capital to that of 1133 companies that did not necessarily have functional similarities to DISA, but instead reflected a comprehensive selection from industry as a whole. Prior to the allocation, DISA had a rate of return of 450 percent in 1959 and 147.2 percent in 1960 -- rates higher than those of all 1133 other companies. Even after the allocation, DISA's rate of return exceeded that of 96 percent of the 1133 companies surveyed. ¹¹² Based on this evidence the court sustained the Service's allocations.

While the Berry ratio and the rate of return analysis found in DuPont are interesting, it should be kept in mind that the court may have looked favorably on this evidence partly because it indicated that even after the allocation DISA earned greater profits than almost any other corporation, whether comparable or not. These methods were not used directly to make a section 482 adjustment, but rather to support the reasonableness of the Service's allocation.

Evidence relating to rates of return was also presented in Lilly. ¹¹³ No general research and development costs for new drugs were being charged by the parent to the subsidiary. The Tax Court determined that a substantial adjustment should be made to the income of the Puerto Rican subsidiary to reflect a proportional payment by the subsidiary of the general research and development expense of the parent. ¹¹⁴ The difference between the rates of return to the two entities was not, however, due solely to the understating of the subsidiary's research and development expense (as determined by the court), but was also attributable to the presence of valuable intangibles that were not properly reflected in the transfer price. A rate of return analysis was used to identify what appeared to be excessive rates of return on assets, so that further inquiry could be made to determine if the returns were in fact excessive and, if so, why.

The rate of return analysis and other information contained in the report by Dr. Wheeler was as follows. ¹¹⁵

                                              1971     1972     1973

           Year           Net sales           Net income

                                             1974        1975

D. CUSTOMS VALUES

An additional approach to transfer pricing that has occasionally been used in litigation is that of adopting the values set by the United States Customs Service. For example, in Ross Glove Co., ¹²¹ the Tax Court accepted the taxpayer's use of the markup used by Customs in valuing gloves imported from the Philippines for purposes of applying the cost plus method. However, in Brittingham v. Commissioner,/122/ the Tax Court made it clear that it would not bind taxpayers to their own declared Customs' valuations where it could be shown that those values were erroneous. ¹²³

E. CONCLUSIONS AND RECOMMENDATIONS

1. Over the years the courts, and in particular the Tax Court, have used various fourth methods for determining appropriate arm's length prices for section 482 allocations. A profit split is appropriate in some cases to establish a transfer price on an arm's length basis because unrelated parties are concerned about the respective shares of potential profits when entering into a business arrangement. ¹²⁴ The problem with the profit split approach taken by the courts, however, is not that the courts have focused on the wrong elements of the transaction, but that they generally have failed to adopt a consistent and predictable methodology.

2. The rate of return on assets and costs to income ratio methods used in DuPont provide some reasonable basis for allocating income and determining transfer prices in the absence of comparables. However, these methods have not yet been sufficiently developed by the courts to fill the gap in analysis left by the section 482 regulations when comparable uncontrolled transactions cannot be located. A profit split or other method should be developed to determine transfer prices in the absence of comparables, which is the subject of Part III of this study.

II. SECTION 482 AFTER THE 1986 TAX REFORM ACT

Part I of the study described the history of section 482, its administration by the Service, and its interpretation by the courts. The lack of specific guidance in the tangible property, intangible property, and services provisions of the section 482 regulations to resolve cases for which appropriate comparables do not exist -- notably cases involving high profit intangibles -- has caused significant problems for taxpayers, the Service, and courts alike.

The amendment made by the 1986 Act to section 482 is Congress' response to the problem described in Part I of determining transfer pricing for high profit intangibles. Specifically, section 482 was amended to provide that income from a transfer or license of intangible property shall be commensurate with the income attributable to the intangible. This Part II discusses the scope of the commensurate with income standard and the requirement for periodic adjustments. The compatibility of these changes with the international norm for transfer pricing -- the arm's length principle -- is also discussed. Finally, this part explores the role of safe harbors for avoiding adjustments under section 482.

CHAPTER 6

THE COMMENSURATE WITH INCOME STANDARD

A. LEGISLATIVE HISTORY

The 1986 Act amended section 482 to require that payments to a related party with respect to a licensed or transferred intangible be "commensurate with the income" ¹²⁴ attributable to the intangible. The provision applies to both manufacturing and marketing intangibles. ¹²⁵ The legislative history clearly indicates Congressional concern that the arm's length standard as interpreted in case law has failed to allocate to U.S. related parties appropriate amounts of income derived from intangibles. ¹²⁶ The amendment is a clarification of prior law. Accordingly, it should not be assumed that the Service will cease taking positions that it may have taken under prior law.

The primary difficulty addressed by the legislation was the selective transfer of high profit intangibles to tax havens. Because these intangibles are so often unique and are typically not licensed of unrelated parties, it is difficult, if not impossible, to find comparables from which an arm's length transfer price can be derived. When justifying the compensation paid for such intangibles, however, taxpayers often used comparisons with industry averages, looked solely at the purportedly limited facts known at the time of the transfer, or did not consider the potential profitability of the transferred intangible (as demonstrated by post-agreement results). Taxpayers relied on intangibles used in vastly different product and geographic markets, compared short-term and long-term contracts, and drew analogies to transfers where the parties performed entirely different functions in deriving income from the intangible.

Congress determined that the existing regime, which depends heavily upon the use of comparables and provides little clear guidance in the absence of comparables, was not in all cases achieving the statutory goal of reflecting the true taxable income of related parties. Congress therefore decided that a refocused approach was necessary in the absence of true comparables. The amount of income derived from a transferred intangible should be the starting point of a section 482 analysis and should be given primary weight. ¹²⁷ Further, it is important to analyze the functions performed, and the economic costs and risks assumed by each party to the transaction, so that the allocation of income from the use of the intangible will be made in accordance with the relative economic contributions and risk taking of the parties. ¹²⁸ The application of the functional analysis approach to the actual profit experience from the exploitation of the intangible allocates to the parties profits that are commensurate with intangible income. Looking at the income related to the intangible and splitting it according to relative economic contributions is consistent with what unrelated parties do. The general goal of the commensurate with income standard is, therefore, to ensure that each party earns the income or return from the intangible that an unrelated party would earn in an arm's length transfer of the intangible.

In determining the income that forms the basis for application of the commensurate with income standard, what time frame should be used as a point of reference: the time of the transfer alone, or an annual or other periodic basis? The legislative history reflects Congressional concern that, by confining an analysis of an appropriate transfer price to the time a transfer was made, taxpayers could transfer a high profit potential intangible at an early stage and attempt to justify use of an inappropriate royalty rate by claiming that they did not know that the product would become successful. ¹²⁹ Accordingly, for these reasons, Congress determined that the actual profit experience should be used in determining the appropriate compensation for the intangible and that periodic adjustments should be made to the compensation to reflect substantial changes in intangible income as well as changes in the economic activities performed and economic costs and risks borne by the related parties in exploiting the intangibles. ¹³⁰ As discussed further below, this is consistent with what unrelated parties would do.

The legislative history indicates that the commensurate with income standard does not prescribe a specific, formulary approach for determining an intangible transfer price. For example, it does not automatically require that the transferor of the intangible receive all income attributable to the exploitation of the intangible. It does not prescribe (nor depend for its application upon) a specific legal form for transfers of intangible property. Thus, it applies to licenses of intangible property, sales of tangible property which incorporate valuable intangibles, and to transfers of intangibles through the provision of services. Nor does it mandate any specific treatment of the transferor or transferee. In particular, the provision does not mandate a "contract manufacturer" return for the licensee in all cases. ¹³¹

B. SCOPE OF APPLICATION

The scope of the commensurate with income standard is not discussed in the legislative history. Two proposals have been made for limiting the scope of the standard, one based on potential double taxation and one limiting the application of the provision to the types of cases that prompted the legislative change.

1. DOUBLE TAXATION AND RELATED ISSUES. Double taxation can occur when two countries have different rules of allocation; have the same rules but interpret or apply them differently in actual operation; have the same rules and interpret and apply them in the same way, but do not allow correlative adjustments; or permit correlative adjustments in theory but do not remove procedural barriers (seq., statutes of limitation on refund claims). ¹³²

Taxpayers and others have argued that the commensurate with income standard will necessarily increase the incidence of double taxation, and that therefore Congressional intent should not be fully implemented. As described more fully in the next chapter, the correct application of the commensurate with income standard is premised soundly on arm's length principles. The Service and Treasury therefore do not believe that the commensurate with income principle will increase the incidence of double taxation.

Indeed, in fairly common cases where the commensurate with income standard will be applied -- outbound transfers of intangibles from U.S. parents to foreign subsidiaries -- the issue of double taxation does not arise. In these situations, the foreign tax credit provisions of U.S. domestic law (including the foreign sourcing and characterization of royalties relating to intangibles used overseas) will normally prevent the double taxation of earnings. ¹³³ Furthermore, the outbound transfer patterns that were the subject of Congressional concern involve transfers to manufacturing affiliates located in tax havens, where there is no potential for double taxation. The question of whether the appropriate amount of income is attributed to foreign operations in these cases is, therefore, whether the correct amount of income is eligible for deferral from U.S. tax and whether it is properly characterized for foreign withholding tax purposes, rather than the issue of double taxation.

2. LEGISLATIVE IMPETUS. The commensurate with income standard was clearly intended to overcome problems encountered in applying the section 482 regulations to transfers of high profit potential intangibles, such as those at issue in Lilly and Searle. Because of its origin as a response to the problem of high profit intangibles, it has been suggested that the commensurate with income standard should be limited to transfers of high profit intangibles to affiliates in low tax jurisdictions. ¹³⁴ The statute, however, applies to all related party transfers of intangibles, both inbound and outbound, ¹³⁵ without quantitative or qualitative restrictions. Furthermore, the economic theory of arm's length dealing underlying the methods set forth in this study apply to all transfers of intangibles, regardless of the type of intangible or residence of the licensee. Consequently, the commensurate with income standard should apply to transfers of all related party intangibles, not just the high profit potential intangibles. The analysis set forth in Chapter 11 provides a framework for implementing the commensurate with income standard that can be applied to all intangible transfers, rather than merely to high profit potential intangibles.

C. APPLICATION OF COMMENSURATE WITH INCOME STANDARD TO NORMAL PROFIT AND HIGH PROFIT INTANGIBLES

1. NORMAL PROFIT INTANGIBLES. In related party transfers of normal profit intangibles, there are likely to be comparable third party licenses. Such licenses can produce evidence of arm's length dealings. The arm's length bargaining of the unrelated parties over the terms of the arrangement reflects each party's judgment about what its share of the combined income (or appropriate expense reimbursement) ought to be. Hence, each has made a judgment that the remuneration it expects to receive is commensurate with the income attributable to its exploitation of the intangible.

Application of the commensurate with income standard to normal profit intangibles will ordinarily produce results consistent with those obtained under pre-1986 law in those cases where economically appropriate comparables were used. For example, the licensing agreement for the formula to a particular brand of perfume is likely to have many "inexact" comparables. ¹³⁶ If appropriate comparables exist, they can be examined to determine an arm's length, or commensurate with income, return. Thus, in many cases the appropriate income allocation under both the existing regulations and the commensurate with income standard will be the same, provided that internal and external standards of comparability are met. ¹³⁷

2. HIGH PROFIT POTENTIAL INTANGIBLES. As described in Chapter 4, the difficulty in applying section 482 to high profit potential intangibles ¹³⁸ is that unrelated party licenses of comparable intangibles almost never exist. Consequently, if the appropriate related party transfer price for a high profit potential intangible is expressed in terms of a royalty, the result may not bear any resemblance to a third party license for a normal intangible. That is, owing to the intangible's enormous profitability, an allocation under the commensurate with income standard, if made solely through a royalty rate adjustment, might be so large compared to normal product royalty rates that it does not look like an arm's length royalty. Therefore, one might argue that an extraordinarily high rate could never be an arm's length royalty merely because third party royalties are never that high.

From an economic perspective, however, an unprecedented or "super-royalty" rate may be required to appropriately reflect a relatively minor economic contribution by the transferee and achieve a proper allocation of income. ¹³⁹ As discussed in Chapter 11, the commensurate with income standard, in requiring a "super-royalty" rate in order to achieve a proper allocation of income in such a case, does not mandate a rate in excess of arm's length rates. Nor does it permit taxpayers to set a "super-royalty" rate in excess of arm's length rates. For example, enactment of the commensurate with income standard would not justify royalty increases in excess of arm's length rates by U.S. affiliates of foreign parent corporations (or vice versa).

Rather than creating a new class of royalty arrangements, the enactment of the commensurate with income standard reflects the recognition that, for certain classes of intangibles (notably high profit potential intangibles for which comparables do not exist), the use of inappropriate comparables had failed to produce results consistent with the arm's length standard. Enactment of the commensurate with income standard was thus a directive to promulgate rules that would give primary weight to the income attributable to a transferred intangible in determining the proper division of that income among related parties. In the rare instance in which there is a true comparable for a high profit intangible, the royalty rate must be set on the basis of the comparable because that remains the best measure of how third parties would allocate intangible income.

D. SPECIAL ARRANGEMENTS

1. LUMP SUM SALES OR ROYALTIES. Some commentators have suggested that the commensurate with income standard should not prohibit the use of non-contingent, lump sum royalty or sale payments. While the Service and Treasury agree that parties are free to structure their transactions as either a sale or license, the economic consequences of a lump sum payment arrangement generally must resemble those under a periodic payment approach in order to satisfy the commensurate with income standard, unless the taxpayer can demonstrate, by clear and convincing evidence, that such treatment is inappropriate on the basis of arm's length arrangements, i.e., an exact or inexact comparable transaction. ¹⁴⁰ By its terms, the amendment to section 482 applies to ANY transfer of an intangible, which includes an outright transfer by sale or license for a non-contingent, lump sum amount. ¹⁴¹ Furthermore, exempting such arrangements from the commensurate with income standard would elevate form over substance and encourage non-arm's length lump sum arrangements designed to circumvent the new rules. Thus, periodic adjustments may be required under the commensurate with income standard even in the case of lump sum sale or royalty arrangements. ¹⁴²

2. INTERACTION WITH SECTION 367(d). Section 367(d), enacted as part of the 1984 Tax Reform Act, provides that when intangible property is transferred by a U.S. person to a foreign corporation in a transaction described in section 351 or 361, the transferor shall be treated as receiving annual payments, over the useful life of the property, contingent on productivity or use of the property, regardless of whether such payments are actually made. These payments are treated as U.S. source income. A subsequent disposition to an unrelated party of either the intangible property or the stock in the transferee triggers immediate gain recognition. The 1986 Act made the commensurate with income standard applicable in computing payments attributable to the transferor under section 367(d). The periodic adjustment of lump sum royalty or sale payments would merely achieve parity with section 367(d) transfers. ¹⁴³ Section 367(d) may also suggest that certain exceptions from the periodic payment approach may be appropriate -- e.g., transfers to corporations in which an unrelated corporation has a substantial enough interest that an objective valuation of the transferred intangible can be considered to be arm's length. ¹⁴⁴

Sales and licenses of intangibles are generally not subject to section 367(d), since they are not transactions described in section 351 or 361. The temporary regulations state that, when an actual license or sale has occurred, an adjustment to the consideration received by the transferor shall be made solely under section 482, without reference to section 367(d). ¹⁴⁵ However, if the purported sale or license to the related person is for no consideration ¹⁴⁶ or if the terms of the purported sale or license differ so greatly from the substance of an arm's length transfer that the transfer should be considered a sham, ¹⁴⁷ the transfer will be treated as falling within section 367(d).

In essence, the commensurate with income standard treats related party transfers of intangibles as if an intangible had been transferred for a license payment that reflects the intangible's value throughout its useful life, a result similar to section 367(d). Because the section 367(d) source of income rule can apply to certain transactions cast in the form of a sale or license, the temporary regulations could be amended to specify which sales or licenses are subject to both the commensurate with income standard and the U.S. source income characterization of section 367(d). Moreover, a license payment that is less than some specific percentage of the appropriate arm's length amount could be considered so devoid of economic substance that the arm's length charge should be subject to section 367(d). Thus, those related party transfers which deviate substantially from the proper commensurate with income payment would be subject to 367(d), even if cast in the form of a sale or license.

3. COST SHARING AGREEMENTS. The legislative history envisions the use of bona fide research and development cost sharing arrangements as an appropriate method of attributing the ownership of intangibles ab initio to the user of the intangible, thus avoiding section 482 transfer pricing issues related to the licensing or other transfer of intangibles. ¹⁴⁸ Use of cost sharing arrangements had previously been encouraged in connection with the enactment in 1984 of section 367(d). ¹⁴⁹ Cost sharing arrangements are discussed in detail in Chapters 12 and 13, infra.

E. CONCLUSIONS

1. Congress enacted the commensurate with income standard because application of existing rules had not focused appropriate attention upon the income generated by the transfer of an intangible in situations in which comparables do not exist.

2. Application of the commensurate with income standard requires the determination of the income from a transferred intangible, and a functional analysis of the economic activities performed and the economic costs and risks borne by the related parties in exploiting the intangible, so that the intangible income can be allocated on the basis of the relative economic contributions of the related parties. The commensurate with income standard does not mandate a "contract manufacturer." return for the licensee in all or even most cases.

3. The commensurate with income standard requires that intangible income be redetermined and reallocated periodically to reflect substantial changes in intangible income, or changes in the economic activities performed and economic costs and risks borne by the related parties.

4. The application of the functional analysis approach to the actual profit experience from the exploitation of intangibles is consistent with what unrelated parties would do and is, therefore, consistent with the arm's length principle.

5. Because the commensurate with income standard is consistent with arm's length principles, it should not increase the incidence of double taxation.

6. The commensurate with income standard applies to all types of intangible property transfers between related parties, not just high profit potential intangibles, including both inbound and outbound transfers of intangibles. In the cases of normal profit intangibles in which comparables normally exist, the new standard, like prior law, will ordinarily base the analysis on comparable transactions, with refinements in the definition of appropriate comparables. In any event, intangible income must be allocated on the basis of comparable transactions if comparables exist.

7. Lump sum sale and royalty payments for intangibles generally will be subject to the commensurate with income standard.

CHAPTER 7

COMPATIBILITY WITH INTERNATIONAL TRANSFER PRICING STANDARDS

A. INTRODUCTION

Shortly after passage of the 1986 Act, various U.S. taxpayers and representatives of foreign governments expressed concern that the enactment of the commensurate with income standard was inconsistent with the "arm's length" standard as embodied in tax treaties and adopted by many countries for transfer pricing matters. As a result, they argued, the application of the commensurate with income standard would lead to double taxation for which no remedy would exist under treaties, because of application of transfer pricing standards by the United States that would be inconsistent with those applied by various other foreign governments. ¹⁵⁰

To allay fears that Congress intended the commensurate with income standard to be implemented in a manner inconsistent with international transfer pricing norms and U.S. treaty obligations, Treasury officials publicly stated that Congress intended no departure from the arm's length standard, and that the Treasury Department would so interpret the new 1aw. ¹⁵¹ Treasury and the Service continue to adhere to that view, and believe that what is proposed in this study is consistent with that view.

B. THE ARM'S LENGTH STANDARD AS AN INTERNATIONAL NORM

The problem of double taxation arising from different transfer pricing methods has been addressed through intergovernmental negotiation and agreement, principally in bilateral tax treaties that specifically provide for certain adjustments by the treaty partners to the tax liability of any entity when its dealings with related entities differ from those that would have occurred between unrelated parties. For example, an OECD model income tax convention permits adjustments to the profits of an enterprise where, in dealing with related enterprises, "conditions are made or imposed between the two enterprises in their commercial or financial relations which differ from those which would be made between independent enterprises . . . ." ¹⁵² If the adjustment is consistent with that standard, the OECD Model Convention calls for the other contracting state to make an adjustment to the profits of the enterprise in its jurisdiction to take into account the first state's adjustments. ¹⁵³ If differences of opinion arise between the two states as to the proper application of this standard, the OECD Model Convention calls for the competent authorities of the respective jurisdictions to consult with one another. ¹⁵⁴ The other major model used by countries in negotiating their tax treaties, the United Nations Model Double Taxation Convention Between Developed and Developing Countries, contains an Article 9 entitled "Associated Enterprises" that is not materially different. ¹⁵⁵

In 1981, the Treasury Department released a model income tax treaty that it uses as a starting point for negotiating income tax treaties with other countries. ¹⁵⁶ Although this model has been revised in a number of particulars to account for the many changes in U.S. tax law since the time of its release, the provisions governing associated enterprises have not changed. The basic provision is virtually identical to the OECD Model Convention. ¹⁵⁷

The arm's length standard is embodied in all U.S. tax treaties; it is in each major model treaty, including the U.S. Model Convention; it is incorporated into most tax treaties to which the United States is not a party; it has been explicitly adopted by international organizations that have addressed themselves to transfer pricing issues; ¹⁵⁸ and virtually every major industrial nation takes the arm's length standard as its frame of reference in transfer pricing cases. ¹⁵⁹ This overwhelming evidence indicates that there in fact is an international norm for making transfer pricing adjustments and that the norm is the arm's length standard. ¹⁶⁰

It is equally clear as a policy matter that, in the interest of avoiding extreme positions by other jurisdictions and minimizing the incidence of disputes over primary taxing jurisdiction in international transactions, the United States should continue to adhere to the arm's length standard.

C. REFERENCE TO PROFITABILITY UNDER THE ARM'S LENGTH STANDARD

Because the arm's length standard is the international norm, a serious potential for disputes over primary taxing jurisdiction would exist if the United States were to implement the commensurate with income standard in a manner that violates arm's length principles. Does a system which, only in the absence of appropriate comparable transactions, places primary emphasis upon the income (or profits) related parties earn from exploiting an intangible violate the arm's length standard, as understood in the international context.?

Probably the most commonly referenced expression of the arm's length standard as understood by the nations that have adopted it is a report issued in 1979 by the OECD. ¹⁶¹ This report adopts the general principle of arm's length pricing for all transactions between related parties. Concerning transfers of intangible property, the report states:

The general principle to be taken as the basis for the evaluation for tax purposes of transfer prices between associated enterprises under contracts for licensing patents or know-how is that the prices should be those which would be paid between independent enterprises acting at arm's length. ¹⁶²

It is useful to refer to those methods that the report considers inconsistent with its arm's length concept to aid in defining such concept. These the report refers to as "global" methods for transfer pricing. They would include, for example, "allocating profits in some cases in proportion to the respective costs of the associated enterprises, sometimes in proportion to their respective turnovers or to their respective labour forces, or by some formula taking account of several such criteria." ¹⁶³ The report criticizes these methods as necessarily arbitrary. ¹⁶⁴

The report also notes that the effect of its arm's length approach, as distinguished from those it criticizes, is:

[T]o recognise the actual transactions as the starting point for the tax assessment and not, in other than exceptional cases, to disregard them or substitute other transactions for them. The aim in short is, for tax purposes, to adjust the price for the actual transaction to an arm's length price ¹⁶⁵

Nowhere, however, does the report suggest that the profits of the related enterprises are irrelevant to this determination. Indeed, there are several instances where the report specifically authorizes an inquiry into profits or profitability. For example, the report notes:

[Its criticism of global methods] is not to say, however, that in seeking to arrive at the arm's length price in a range of transactions, some regard to the total profits of the relevant [multinational enterprise] may not be helpful, as a check on the assessment of the arm's length price or in specific bilateral situations where other methods give rise to serious difficulties and the two countries concerned are able to adopt a common approach and the necessary information can be made available. ¹⁶⁶

In arriving at an arm's length price, the report specifically authorizes an analysis of economic functions performed by each related party in determining "when a profit is likely to arise and roughly what sort of profit it is likely to be." ¹⁶⁷

Other references to profits occur in the report. For example, in the section of the report relating to the sale of tangible goods entitled "Methods of Ascertaining an Arm's Length Price," methods are outlined that permit reference to comparable profits or returns on capital invested as a means of determining the appropriate transfer price. These methods are viewed by the report as a supplement to the traditional approach of looking to comparable transactions, but they are clearly suggested as appropriate tools for arriving at a proper transfer price. ¹⁶⁸

With regard to valuing transfers of intangible property, the report notes that, "[o]ne of the common approaches employed in practice is to make a pragmatic appraisal of the trend of an enterprise's profits over a long period in comparison with those of other unrelated parties engaged in the same or similar activities and operating in the same area." ¹⁶⁹ The report questions whether this approach is practical, because it would be difficult to isolate respective profits due to different accounting methods, and difficult to know how to apportion the overall profit between the two parties. No suggestion is made, however, that such a method could never be used in the absence of comparable transactions because it conflicts in principle with the arm's length standard. ¹⁷⁰

D. PERIODIC ADJUSTMENTS UNDER THE ARM'S LENGTH STANDARD

The next chapter describes an important element of the commensurate with income standard -- periodic adjustments must be made in appropriate cases to reflect actual profit experience under the license. As noted in that chapter, there are sound arm's length reasons to require such adjustments -- principally the rarity of long-term, fixed licenses negotiated at arm's length, particularly with respect to high profit potential intangibles, and the fact that actual profit experience under a license indicates in most cases anticipated profits that would have been considered by unrelated parties. Moreover, that chapter permits taxpayers to avoid adjustments over time if they can demonstrate on the basis of arm's length evidence that no such adjustments would have been made by unrelated parties. The Service and Treasury therefore believe that such periodic adjustments as will be made under the new standard will be consistent with the arm's length standard as embodied in U.S. double taxation treaties.

E. RESOLUTION OF BILATERAL ISSUES

The Service and Treasury recognize that implementation of the commensurate with income standard in all its particulars, including periodic adjustments, treatment of lump-sum payments ¹⁷¹ and access to information to perform the necessary analysis, may lead to differences with the competent authorities of our treaty partners and perhaps more general issues of treaty policy and interpretation. Recognizing this, the United States competent authority and the Treasury Department should be receptive to the concerns of foreign governments, and endeavor to seek bilateral solutions insofar as those concerns can be accommodated in a manner consistent with Congressional intent in enacting the commensurate with income standard.

F. CONCLUSIONS

1. The arm's length standard requires that each entity calculate its profits separately and that related party transactions be priced as if unrelated parties had entered into them. Reference to the profits (including the trend of those profits over time) of related parties to determine a royalty in a licensing transaction is intended to reflect what unrelated parties would do and, therefore, is consistent with the arm's length standard.

2. The arm's length standard as accepted by the international community does not preclude reference to profits of related parties to allocate income, but in fact encompasses such an approach as a supplement to the traditional approach of looking to comparable transactions. It is, therefore, reasonable to conclude that such an approach is consistent with international norms as applied to situations in which comparables do not exist.

3. The approach taken by Congress in enacting the commensurate with income standard and the approaches suggested in Chapters 8 and 11, infra, for implementing that standard, including the provision for periodic adjustments, are consistent with internationally recognized arm's length principles. Applied in a manner consistent with arm's length principles, the commensurate with income standard is not likely to increase international disputes over the right of primary taxing jurisdiction.

4. The United States competent authority and the Treasury Department should endeavor whenever possible to seek bilateral solutions to problems that may arise with our treaty partners in the interpretation and administration of the commensurate with income standard.

CHAPTER 8

PERIODIC ADJUSTMENTS

A. INTRODUCTION

As discussed in Chapter 6, an intangible transfer price that is commensurate with the income attributable to the intangible must reflect the "actual profit experience realized as a consequence of the transfer." ¹⁷² The "commensurate with income" language requires that changes be made to the transfer payments to reflect substantial changes in the income stream attributable to the intangible as well as substantial changes in the economic activities performed, assets employed, and economic costs and risks borne by related entities.

The Congressional directive to the Service to make adjustments to intangible returns that reflect the actual profit experience is in part a legislative rejection of R.T. French v. Comm'r. ¹⁷³ That case endorsed the view that a long-term, fixed rate royalty agreement could not be adjusted under section 482 based on subsequent events that were not known to the parties at the original contract date. Thus, underlying the directive is perhaps a view that contractual arrangements between unrelated parties -- particularly those involving high profit intangibles -- are not entered into on a long term basis without some mechanism for adjusting the arrangement if the profitability of the intangible is significantly higher or lower than anticipated. A very preliminary review of unrelated party licensing agreements obtained from the files of the Securities and Exchange Commission, discussed in Appendix D, and other input received to date, seems to support this view. Indeed, as a matter of long term business strategy, unrelated parties may renegotiate contractual arrangements even absent explicit renegotiation provisions to reflect revised expectations regarding an intangible's profitability. ¹⁷⁴

Aside from the empirical evidence of what unrelated parties seem to do, actual profit experience is generally the best indication available, absent comparables, of anticipated profit experience that arm's length parties would have taken into account at the outset of the arrangement. It is, therefore, perfectly consistent with the arm's length standard to treat related party license agreements generally as renegotiable arrangements and to require periodic adjustments to the transfer price to reflect substantial changes in the income stream attributable to the intangible. ¹⁷⁵

Intangible transfer prices will in any event be determined on the basis of comparables if they exist. If a particular taxpayer demonstrates that it has comparable long-term, non-renegotiable contractual arrangements with third parties, the arm's length standard will preclude periodic adjustments of the related person intangible transfer price. In that event, a comparable would exist by definition, which would determine the consideration for the related person transfer, both initially and over time. Comparables are always the best measure of arm's length prices. In the case of a high profit intangible, however, a third party transaction generally must be an exact comparable in order for the transaction to constitute a valid comparable. ¹⁷⁶

It may also be possible in certain other cases to exclude subsequent profit experience from consideration under the arm's length standard. To do so, the taxpayer would need to demonstrate each of the following to avoid an adjustment based on subsequent profit experience:

1. That events had occurred subsequent to the license agreement that caused the unanticipated profitability;

2. That the license contained no provision pursuant to which unrelated parties would have adjusted the license; and

3. That unrelated parties would not have included a provision to permit adjustment for the change that caused the unanticipated profitability.

For example, assume that there are twelve heart drugs that perform similar therapeutic functions, none of which has a dominant market share. Several of these drugs are licensed to unrelated parties under long-term arrangements which do not provide a mechanism for adjusting the royalty payments because of subsequent changes. The taxpayer's drug, which is licensed to a related party, uses an active ingredient which is different from the other products with which it competes. The competitors' drugs, however, lose all of their market share during the course of the license agreement because their products are found to cause serious side effects, and the licensed product's profits increase dramatically. In this case, if the taxpayer could prove the three factors above, the taxpayer could avoid an adjustment based on the increase in profitability.

As noted earlier, Congress was particularly concerned about taxpayers attempting to justify low-royalty transfers at an early stage based on the purported inability to predict subsequent product success. ¹⁷⁷ Because of this concern, it would be appropriate to impose a high standard of proof, such as a clear and convincing evidence standard, on taxpayers in order to demonstrate that subsequent profitability could not have been anticipated. In no event should this test be available to taxpayers if inexact comparable licenses with no provision for periodic adjustments cannot be found in the marketplace.

A substantial change in intangible income will not necessarily result in an adjustment. As discussed in Chapter 6 and described in Chapter 11, determining the intangible income is merely the first step in the analysis of allocating intangible income. The second step involves allocating income on the basis of the activities performed and economic costs and risks borne by the parties. If intangible income increases solely due to the efforts of the transferee, then the increase in intangible income will be allocated exclusively to the transferee, and no adjustment will be made to the income of the transferor.

B. PERIODIC REVIEW

Annual adjustments may not be required to reach the appropriate amount of income under the commensurate with income standard. Adjustments are not required for minor variations in intangible income, only for substantial changes in intangible income. ¹⁷⁸ Several issues are raised by this requirement. How often should the taxpayer review its transfer pricing structure to determine whether income is being properly reported and to avoid potential penalties? How often may the Service make adjustments in the course of examination? Should the regulations define substantiality? Should the adjustments be applied retroactively or prospectively? Should periodic adjustments be made in the case of a sale of intangibles and other situations involving lump sum payments? Should set-offs be permitted?

The frequency with which a taxpayer should review its related party intangible transfer agreements and how often the Service should be able to make adjustments are not questions that can be governed by inflexible rules. When the transferee experiences a substantial change in its profits from the intangible resulting from some particular event (whether anticipated or not), a review by the taxpayer is clearly warranted; further, an adjustment by the Service is warranted unless the taxpayer can demonstrate, by clear and convincing evidence, that the conditions discussed above for avoiding an adjustment based on subsequent profit experience are met. Even absent a clear-cut event, it is possible that gradual changes over time may create a substantial deviation from the parties' expectations at the time they entered into the contract.

In general, taxpayers should review transfer pricing arrangements relating to intangibles (especially high profit intangibles) as often as necessary to assure that their transfer prices are consistent with substantial changes in intangible income that may have occurred since the inception of the current transfer pricing arrangements. For industries that undergo rapid technological change or for products that have a relatively short life, this standard may dictate annual review. In short, the taxpayer should review its pricing structure relating to intangibles as often and thoroughly as necessary to assure that income is reported on its U.S. tax return in a manner that is consistent with the commensurate with income standard. Taxpayers that fail to do so risk the imposition of the substantial understatement or other appropriate penalty. ¹⁷⁹

On the other hand, the Service should be permitted to make a transfer pricing adjustment without necessarily having to demonstrate that its proposed adjustment is justified by identifiable changes in intangible income compared with a prior taxable year. In other words, if the adjustment can be supported on the basis of exact or inexact comparables, or on the basis of the rate of return analysis or such other methodology as is adopted by the Service for use in cases in which exact or inexact comparables do not exist, ¹⁸⁰ then the Service should not have to demonstrate that the adjustment specifically relates to identifiable changes in intangible income occurring since the last taxable year examined. An approach whereby the Service would be estopped from making an adjustment, absent clearly identifiable changes in intangible income, because of the Service's prior acceptance of some commensurate with income amount in a prior year would present problems of proof that are not necessarily relevant to the appropriateness of the adjustment. At most, consideration should be given only to requiring that the proposed adjustment to income be substantial in relation to the income reported by the taxpayer from the transaction -- and then only for audits subsequent to the first in-depth audit of transfer prices conducted for taxable years after 1986.

It may be advisable to publish in the Internal Revenue Manual a list of factors that, if one or more changes substantially, would indicate that there may be a substantial change in intangible income that may warrant an examination of the taxpayer's intangible transfer pricing. These factors might include: (a) the size and number of markets penetrated; (b) the product's market share; (c) the product's sales volume; (d) the product's sales revenue; (e) the number of uses for the technology; (f) improvements to the technology; (g) marketing expense; (h) production costs; (i) the services provided by each party in connection with the use of the intangible; and (j) the product's profit margin or the process' cost savings. ¹⁸¹

Any periodic adjustments that are made under the commensurate with income standard generally should be made prospectively -- i.e., for the taxable year under audit and subsequent taxable years (provided that there is no further substantial change in the intangible income and other relevant facts). Unless unrelated parties would have set a different royalty rate on the date of the transfer based upon expectations of future high profitability or other facts known on the date of transfer, the arm's length standard would require only that the transfer price be commensurate with actual income -- i.e., that the transfer price be changed only as the intangible income changes.

C. LUMP SUM PAYMENTS

As discussed in Chapter 6, the commensurate with income standard applies to transfers of intangibles by sale or license for noncontingent, lump sum amounts. Thus, periodic adjustments may be required under the commensurate with income standard in the case of lump sum sale or royalty arrangements as well as periodic royalty arrangements. In the case of a lump sum sale, how should the Service make a section 482 allocation if it is not apparent until many years after the sale that the lump sum payment was insufficient under the commensurate with income standard?

One possibility would be to recharacterize the sale as a license, thereby giving the Service the ability to require additional royalty payments sufficient to satisfy the commensurate with income standard. It is clear, however, that parties dealing at arm's length occasionally sell intangibles. Thus, failure to recognize sale arrangements for related party transactions could be viewed as a deviation from the arm's length standard.

Alternatively, the Service could recognize the transfer as a sale but make a section 482 allocation to increase the initial lump sum payment. Unless taxpayers using lump sum sale arrangements were required by regulation or statute to keep the statute of limitations open for the payment year, the statute of limitations could bar adjustments to a lump sum payment in closed taxabie years, contrary to Congressional intent. Moreover, other problems would exist in adjusting the lump sum even if the statute of limitations were open. For example, any mid-stream adjustment to the initial lump sum made before the statute of limitations expires on the year of sale would necessarily be based on a projection of future profits over the remaining life of the intangible that could be too high or too low. Furthermore, any mechanism, whether elective or mandatory, that would keep the statute of limitations for the year of transfer open for extended periods would disrupt the examination process by unduly delaying the closing of audits.

A lump sum sale arrangement should instead be treated as an open transaction to assure that the sale over time satisfies commensurate with income standard. This approach is the only approach which recognizes the transaction as a sale, allows for adjustments of the sale price under the commensurate with income standard, and minimizes the statute of limitation and other problems inherent in making adjustments to income in the year of the sale. Under this approach, a lump sum sale payment made in the year of the transfer would result in gain taxable in the year of transfer but would then be treated as a prepayment of the commensurate with income amounts. No section 482 allocation would be required until the aggregate commensurate with income amounts exceed the prepayments.

Under this method, the lump sum is treated as invested on the date of the lump sum payment in a hypothetical certificate of deposit ("C.D.") maturing on the last day of taxpayer's current tax year bearing interest at the appropriate federal funds rate based on the anticipated life of the intangible (for U.S. developed intangibles) or the appropriate rate in the development country. At the end of year one the balance of the C.D. investment would be computed. From this amount, the amount of the commensurate with income amount would be subtracted. The remaining balance would then be treated as invested in a C.D. maturing at the end of year two. At the end of each tax year a computation similar to that done at the end of year one would be made. When the C.D. balance is exhausted, the taxpayer would be required thereafter to include the entire commensurate with income amount in income each year.

Consider, for example, an intangible that is transferred for $1,000 and that would demand a commensurate with income amount in each of ten years as shown:

              Col. 1                Col. 2              Col. 3

Because section 482 may be applied only by the Service, no refunds could be allowed if an excessive lump sum was paid. However, to prevent abuse of outbound lump sum payments in inbound licensing arrangements, the Service would be allowed to adjust excessive lump sum payments that clearly exceed the commensurate with income standard.

D. SET-OFFS IN ROYALTY ARRANGEMENTS

It is possible that the initial royalty rate set by the parties could be too large in some years and too small in other years when analyzed under the commensurate with income standard. Under existing regulations, section 482 adjustments traditionally have been made on a year-by-year basis. Only intra-year set-offs of proposed adjustments against excessive income derived on related party transactions are authorized under section 1.482-1(d)(3) of the regulations. ¹⁸³ Thus, the Service could make adjustments in some years without making an allowance for excessive royalties paid in other years. Assume, for example, that a license generates a fixed royalty amount on an intangible that produces fluctuating income due to business cycles or changes in demand. Over time, the royalty may be an appropriate one on average, but in some years it may be too low and in others too high.

Because of the problems inherent in an open transaction approach, the current rule prohibiting multi-year set-offs should be retained. The potentially harsh effect of this rule will be mitigated by the fact that periodic adjustments generally should be made only in cases of substantial changes in circumstances and by the ability of taxpayers to adjust their own arrangements prospectively, reducing or increasing the royalty, to account for changed circumstances. It will also create an incentive for taxpayers to examine their arrangements periodically to see whether an adjustment favorable to them would be appropriate.

E. CONCLUSIONS

1. Periodic adjustments are necessary in order to reflect substantial changes in the income stream produced by a transferred intangible, taking into account the activities performed, assets employed, and economic costs and risks borne by the related parties.

2. Requiring periodic adjustments is consistent with the arm's length principle, since unrelated parties generally provide some mechanism to adjust for change in the profitability of transferred intangibles and since actual profit experience generally is the best indication available of the anticipated profit experience that unrelated parties would have taken into account at the outset of the arrangement.

3. Taxpayers should review transfer pricing arrangements relating to transferred intangibles as often and as thoroughly as necessary to assure that income is reported over time in a manner consistent with the commensurate with income standard.

4. Periodic adjustments made under the commensurate with income standard generally should be prospective unless a different royalty rate would have been set on the date of transfer based upon expectations of the parties and the facts known as of the date of transfer.

5. A lump sum sale of an intangible should be characterized as an open transaction whereby the lump sum sale payment results in gain at the time of transfer, but is then treated as a prepayment of the commensurate with income amounts. No section 482 allocation would be required until the aggregate commensurate with income amounts exceed the prepayment.

6. Multi-year set-offs of proposed adjustments against excessive related party income derived in other taxable years will not be permitted.

CHAPTER 9

THE NEED FOR CERTAINTY: ARE SAFE HARBORS THE SOLUTION?

A. INTRODUCTION

One of the most consistent criticisms of the section 482 regulations is that they do not provide taxpayers with enough certainty to establish intercompany prices that will satisfy the Service without overpaying taxes. Based on the government's experience in litigation, the current section 482 regulations also fail to provide the Service and the courts with sufficiently precise rules to make appropriate section 482 adjustments, especially when third party comparables are not available. ¹⁸⁴ One of the most common suggestions for solving these problems is to amend the section 482 regulations to adopt safe harbors, or simple, mechanical, bright- line tests that may be used in lieu of the fact-specific arm's length inquiry under section 482. ¹⁸⁵

B. GENERAL PROBLEMS WITH SAFE HARBORS

While numerous safe harbors have been proposed, they generally have taken two forms: (1) absolute safe harbors that grant the taxpayer total freedom from a section 482 adjustment once the criteria for the safe harbor are satisfied, and (2) conditional safe harbors that produce a rebuttable presumption or a shift in the burden of proof in the taxpayer's favor, but that may be overcome by the Service through evidence showing that a section 482 adjustment is necessary.

Although various types of safe harbors are available, they all have one common element that makes them both attractive to the taxpayer and potentially troublesome to the government: they generally would serve only to reduce tax liability. Taxpayers for which a safe harbor would produce a lower tax liability than the appropriate normative rule would use it. Those for which a safe harbor would produce a higher tax liability than the appropriate normative rule generally would not seek the protection of the safe harbor but would apply the normative rule. Reducing administrative costs, or the need for certainty, may encourage some taxpayers to use a safe harbor in marginal situations even if application of the normal rule would result in a tax savings. In general, however, the only benefit a safe harbor offers from the Service's perspective is a saving of administrative costs.

Ideally, safe harbor standards should be easy and inexpensive vehicles for selecting cases that warrant closer scrutiny. The perfect safe harbor would result in the elimination of all insignificant cases and the selection of cases for detailed analysis by taxpayers and further examination by the Service that would more likely produce sustainable, significant adjustments if analyzed incorrectly by the taxpayer. The question is whether there are any safe harbors that are capable of approaching these goals.

A look at the Service's experience with section 482 safe harbors is instructive. The best example is the safe harbor for interest rates found in section 1.482-2(a)(2)(iii). From the Service's point of view, results under this safe harbor have been mixed at best. The safe harbor was originally set between 4 and 6 percent. This was probably sufficient in 1968, but it soon became inappropriately low. However, the government was very slow to change the safe harbor range as interest rates rose. ¹⁸⁶ The safe harbor for interest now tracks the Federal rates required to be determined for purposes of the original issue discount rule under section 1274(d), which reflect market rates and are adjusted monthly. While this is probably a satisfactory solution, many taxpayers were able to gain a substantial windfall while the government made successive attempts to choose an appropriate safe harbor rate.

Another example of a safe harbor is found in section 1.482- 2(c)(2)(ii) of the regulations, which provides a safe harbor computation of an arm's length rental for the use of tangible property. Experience has demonstrated that this safe harbor was overly generous to taxpayers (i.e., requiring too little rent). It was repealed by regulations finalized this year. ¹⁸⁷ No substitute safe harbor has been provided to date. ¹⁸⁸

The government's experience in the section 482 area has been that safe harbors have generally treated amounts as arm's length that were usually different from market rates. This result is even more likely to occur in the transfer pricing area because of the inherent difficulty of constructing valuation safe harbors for the types of intangible and tangible property that have created transfer pricing problems under section 482. Furthermore, because of the complexity of the regulatory process and the difficulty in obtaining reliable data, adjustments or corrections to safe harbor standards would be slow. In any event, the fundamental deficiencies of safe harbors are not resolved by continually reviewing and revising the rates, or by intentionally setting the safe harbor on the conservative side for protection of the revenue. If safe harbors are set at non-market rates, they will be used only by taxpayers that will benefit by making or receiving payments at those rates.

C. SPECIFIC PROPOSALS

The following lists some of the safe harbors that have been proposed and includes a short explanation of some of the reasons why they have not been endorsed by the Service and Treasury.

1. PRICING BASED ON INDUSTRY NORMS. This approach is contrary to the legislative history of the section 482 changes in the 1986 Act. ¹⁸⁹ Industry norms generally do not reflect arm's length prices for highly profitable intangibles. Accordingly, any safe harbors based on industry norms or statistics would permit transfer prices that would be far different from the arm's length standard in the most significant cases.

2. PROFIT SPLIT -- MINIMUM U.S. PROFIT. This approach would guarantee that the United States would capture a certain minimum of the profit in transfer pricing cases, perhaps 50 percent. The commensurate with income standard is designed to divide the income involved between related parties to "reasonably reflect the relative economic activity undertaken by each." ¹⁹⁰ A safe harbor that splits profits a certain way in all cases would be inconsistent with the case-by-case factual determination that is necessary to measure the economic contribution made by each of the related parties. Furthermore, a fixed U.S. profit requirement would be objectionable to other countries when intangibles were developed outside the United States.

3. PROFIT SPLIT BASED ON TAXPAYER'S PROPORTIONATE SHARE OF COMBINED COSTS (ABA PROPOSAL). ¹⁹¹

The problem with this safe harbor is that it presumes that different types of expenses contribute equally to the combined profit. For example, expenses incurred for highly skilled technical services might contribute proportionately more to the combined profit than those incurred for unskilled services. Furthermore, it might be very difficult to determine what indirect expenses (including, for example, research and development expenses) are attributable to particular products, and taxpayers might be able to manipulate the profit split by shifting expenses from one product to another or one entity to another (such as by "loaning" employees).

4. PROFIT SPLIT BASED ON SHARE OF COMBINED COSTS AND ASSETS (ABA PROPOSAL). ¹⁹² The second ABA safe harbor is a modification of the first one. Instead of relying entirely on relative expenses, it relies 50 percent on expenses, and 50 percent on the fair market value of the assets used in the production of the property involved in the sale. However, at a minimum not less than 25 percent of the combined taxable income would be allocated to the buyer. If one of the parties employed its assets in a very inefficient manner, it would nevertheless be rewarded in the same manner as if it were highly efficient. Additionally, assets could be arbitrarily shifted from one entity to another; difficult questions of property valuation could arise; and property could be purchased just to tip the balance of profits. Because of those problems, a party could receive a substantial amount of the combined taxable income, yet be doing very little to earn the income.

5. INSUBSTANTIAL TAX BENEFIT TEST. This safe harbor would be available if the rate of tax in the foreign jurisdiction was at least 90 percent of the U.S. rate. The theory behind this safe harbor is that taxpayers will use arm's length pricing if no overall tax savings result from doing otherwise. While this approach may have some pragmatic appeal, there are still several problems with it. An adjustment under section 482 does not depend on an intent to avoid taxes. Even if the taxpayer is overpaying its worldwide tax liability, if U.S. income is being understated, an adjustment should be made. Furthermore, as a policy matter the United States will not cede its taxing jurisdiction to a foreign country other than by treaty. Accordingly, if a taxpayer intentionally or inadvertently shifts income to a high tax jurisdiction it should be subject to a section 482 adjustment without the benefit of a safe harbor. As a practical matter, however, the Service proposes relatively few adjustments between a U.S.-based parent company and its affiliates located in another jurisdiction whose effective tax rate is nearly the same or higher. Different problems are presented by foreign-based parents and their U.S. affiliates.

6. PROFIT DISTRIBUTION TEST. This safe harbor would be satisfied if at least 25 percent of the pre-royalty net profit of an affiliate was distributed to the parent. This test is directly contrary to the commensurate with income concept. If an affiliate is responsible for only 10 percent of the economic activity in question it should not be able to keep up to 75 percent of the profit involved.

7. PRIOR SETTLEMENT TEST. Under this proposal, when the Service had accepted a specific pricing method in a prior examination, the burden would be on the Service to show that the pricing method is unreasonable for the current year. This safe harbor is unacceptable because there could be any number of reasons why the Service had accepted a particular pricing method. To force the Service to demonstrate that the previously agreed upon method has become unreasonable could help perpetuate an error or make it more difficult to adjust to changing circumstances.

D. BURDEN-SHIFTING SAFE HARBORS

Some of the safe harbor proposals would operate by shifting the burden of proof to the Service. It has been proposed, for example, that a taxpayer's full disclosure of the method by which it determines its transfer prices would shift the burden of proof to the government. (See Chapter 3, supra, for a description of IEs' experiences in seeking information.) Section 6001 requires all taxpayers to maintain adequate books and records to substantiate positions taken on the tax return, including section 482 issues. Thus, a taxpayer could obtain a shift in the burden of proof merely by complying with the law.

The Service and Treasury do not believe that "burden-shifting" safe harbors are a viable approach. The critical issues presented in the section 482 area are almost always factual in nature, and taxpayers are almost always uniquely familiar with -- and in exclusive possession of -- the relevant facts. To place the burden of proof on the government in this situation would be unworkable.

E. CONCLUSIONS AND RECOMMENDATIONS

1. Historical experience with safe harbors indicates that they generally result in unwarranted windfalls for taxpayers, without significant benefits for the government.

2. In the highly factual section 482 context, no one safe harbor or combination of safe harbors has yet been proposed that would be useful but not potentially abusive.

3. While the possibility that useful safe harbors could be developed is not categorically rejected, additional section 482 safe harbors are not recommended at the present time.

III. METHODS FOR VALUING TRANSFERS OF INTANGIBLES

A significant reason for the enactment of the commensurate with income standard was the failure to properly take into account the income earned by related parties in exploiting intangibles. As detailed in earlier chapters, inappropriate comparables or ad hoc profit split approaches have been used to analyze cases involving related party transfers of unique intangibles.

This part of the study seeks to define the appropriate use of comparable transactions. It also proposes an alternative method of analysis that does not directly rely upon comparable transactions. Fourteen detailed examples applying the methods described in this part are set forth in Appendix E. These methods are generally consistent with various methods of income allocation used by the Service, taxpayers, and the courts under pre-1986 Act law and, if adopted, would appropriately be applicable to cases arising prior to the 1986 Act.

CHAPTER 10

ECONOMIC THEORIES CONCERNING THE IMPLEMENTATION OF SECTION 482

A. INTRODUCTION

The current section 482 regulations use a market-based approach to income allocation. The goal of this approach is to distribute income in the same way that the market would distribute the income; that is, related parties should earn the same returns that unrelated parties would earn under similar circumstances. This approach is implemented through separate accounting in which an individual transfer price is determined for each transaction. ¹⁹³

The argument for the market-based method to allocate income was articulated by Stanley Surrey, former Assistant Secretary (Tax Policy), who discussed the way that unrelated parties are taxed:

Tax administrators do not question transactions that are governed by the marketplace. If Company A sells goods to unrelated Company B at a certain price or furnishes services at a particular price, the income of both companies is determined by using that price. One company may be large and the other small; one may be a monopoly; one may be financially strong and the other in a weak condition. But these and other factors which may affect the price at which the transaction occurs are not the concern of the tax administrator. ¹⁹⁴

Having established the tax system's acceptance of the marketplace, he concludes:

Presumably, most transactions are governed by the general framework of the marketplace and hence it is appropriate to seek to put intra-group transactions under that general framework. Thus, use of the standard of arm's length, both to test the actual allocation of income and expense resulting under controlled intra-group arrangements and to adjust that allocation if it does not meet such standard, appears in theory to be a proper course. ¹⁹⁵

Recent criticism has questioned whether the market-based arm's length approach is flawed as a matter of general principle. An alternative approach might be based on the concept of an integrated business. Loosely defined, an integrated business consists of firms under common control and engaged in similar activities. Proponents of the alternative approach assert that one cannot assume that related parties conduct market-based transactions within the entity. They claim that, because an entity will not act as if its parts are unrelated, it does not make sense to try to account for individual transactions in the way that unrelated parties subject to market forces would account for similar transactions. Under this theory, the allocation of income can only be accomplished by applying some formula chosen by the government.

This chapter explores the tension between these alternative approaches, and suggests a way to apply the arm's length principle to an integrated business. It concludes that the market-based arm's length approach remains the best theoretical allocation method.

B. THE ARM'S LENGTH APPROACH IN AN INTEGRATED BUSINESS: THEORY

The goal of a market-based approach is to ensure that the return to an economic activity is allocated to the party performing the economic activity. Critics of the market-based approach argue that an arm's length price will not achieve this goal. Relevant practical issues are how close one can get to the right price and whether getting that price is costly relative to settling for an estimate.

One commentator has suggested that the difficulties encountered in administering the current regulations stem not from practical considerations, but rather from fundamental problems inherent in applying a market-based approach to transactions of integrated businesses. ¹⁹⁶ Specifically, it has been argued that the flaw in an arm's length approach is that it does not allow a return to the form of organization. That is, because an integrated enterprise is presumably more efficient, it will be able to execute an integrated economic activity at a lower cost than a series of independent firms whose joint efforts are necessary to execute the same series of transactions. This omission creates a "continuum price problem," a situation in which the sum of the returns for separate services rendered by independent parties is less than the actual return of the combined group. This argument grows out of the literature on the reasons for the existence of multinationals. ¹⁹⁷

That multinationals may exist because of integrated or "firm- specific" economies does not require a rejection of the arm's length principle. Transfer prices are supposed to reflect the contribution of the activity and assets utilized in each location to economic income. Therefore, each party should earn at least as much as it could have earned as an unrelated party under alternative arrangements.

Furthermore, an analysis of "alternative arrangements" need not be restricted to analyzing conventional arm's length transactions. Consider a U.S. firm that owns the worldwide rights to a unique patented drug that it wishes to sell into a new market (and assume that the drug or patent is valuable in its own right and that marketing activity, for example, is not an important factor). The firm could license the use of the patent to unrelated parties to produce the drug in the new market. Alternatively, the firm could enter into a joint venture by affiliating with a company that has the ability to produce the drug. If the pharmaceutical firm entered into a joint venture it would probably be able to negotiate a very large share of the profits given the value of the patent, because any number of local firms could provide the labor and capital necessary to mix and package the drug.

There are, therefore, two types of arm's length transactions to consider -- one in which the parties remain independent and another in which the two parties make an arm's length agreement to affiliate by merger, joint venture, acquisition, or simply through the hiring of local labor and capital within a subsidiary. Restricting attention to transactions between parties that remain unrelated can fail to accomplish the objective of allocating to each party its contribution to income, if such transactions do not accurately reflect the actual relations between the related parties.

Another way of describing the arm's length agreements that have to be considered is to say that they are the arrangements that would be made between unrelated parties if they could choose to have the costs of related parties -- i.e., to use the related party technology. In general, tax rules should distort business decisions as little as possible because rules that minimize such distortions will lead to the greatest possible production efficiency. Transfer pricing rules will allow the most efficient production technology to come to the fore if, holding the cost functions constant, they result in the same tax burdens whether or not the parties are related. In other words, if unrelated parties somehow had access to the technology available to related parties, their operations should not result in more or less total taxes than would be paid by a multinational using this technology. The difficulty, of course, is the practical application of this interpretation of the arm's length standard.

C. THE ARM'S LENGTH APPROACH IN AN INTEGRATED BUSINESS: PRACTICE

The arm's length approach can be more correctly applied to an integrated business by using certain tools of microeconomic theory. The continuum price problem arises when a vertically or horizontally integrated production technology that is available to multinational corporations results in lower costs than a non-vertically or horizontally integrated technology, which unrelated parties would have to use. How can an examination of unrelated party transactions lead to a satisfactory resolution of the transfer pricing problem?

As a first step, consider an industry in which there is no difference in costs between related party and unrelated party dealings; there is only one production technology, and it is available to the parties in both types of arrangements. There is thus no continuum price problem, and the arm's length standard, as traditionally interpreted, can be applied. It is likely that both unrelated party and related party transactions will actually occur in the marketplace, and it should be possible to observe prices from the former and use them to determine the incomes of each party in the latter.

This procedure satisfies the objective described in section B above of using information about unrelated parties operating at arm's length to determine the allocation of income in the related party setting. The related parties that sell intermediate goods will be given the same gross revenues as the corresponding unrelated parties; related parties that purchase them will have the same cost of goods sold as corresponding unrelated parties. Further, it has already been assumed that the two sets of parties operate in the same market and have the same cost structure; therefore, the external prices and internal costs will be equal. Thus, the related parties will have the same net taxable incomes as the corresponding unrelated parties. They should therefore have the same total tax burden as the unrelated parties with which they are competing.

Now return to the situation in which a vertically or horizontally-integrated technology, available only to multinational companies, is dominant. If multinational corporations are able to produce at lower cost, then in the long run it should be difficult for the smaller companies to continue in existence. Therefore, arm's length prices may be unavailable. An appropriate transfer pricing result will be achieved if each related party were assigned the income that the corresponding unrelated party would earn, if the latter were using the efficient cost structure.

Microeconomic theory leads to an unambiguous and natural statement of what the income of unrelated parties should be in these circumstances. As long as the industry under analysis is competitive and the factors of production are homogeneous and mobile between sectors, it is assumed that "economic," "excess," or "above-normal" profits will be zero in the long run. ¹⁹⁸ That is, each firm will earn just enough to be able to pay for the land, labor, capital, and other factors of production that it uses to produce its outputs.

The zero economic profit concept does not state that taxable income should be zero. If owners of the firm have supplied it with capital or other inputs, the firm should earn enough to be able to reward the shareholders for these factors; otherwise, the shareholders would be wise to find a better investment. Rather, the zero profit concept implies that in a competitive industry there should be an equality between the gross revenues of a firm and the summation of the market returns that are or could be earned by all of the factors of production that the firm employs. If gross revenues were higher than this amount, then the firm would be earning "above- normal" profits; the existence of "above-normal" profits would attract other firms to enter the industry until these "above-normal" profits disappeared through competition. If gross revenues were lower than this amount, a firm would not be able to earn enough to reward all of the factors it employs and, in the long run, would have to shrink or disappear.

This equality between revenue and the sum of returns to each factor of production may be used to determine the proper allocation of income among the related parties within the multinational. Specifically, subject to the discussion in section D infra regarding monopoly situations and intangibles, one should measure the factors of production used by each related party and compute the returns that each one would earn on its best alternative use in the marketplace. The sum of these amounts yields the total input returns that each related party would have to earn if it were an unrelated enterprise. The sum also equals the amount that the multinational enterprise would have to pay an unrelated party to get it to produce the same outputs (employing the same inputs and using the same technology) as the related party does. Attributing this gross income to each related party will result in its tax base being equal to the hypothetical unrelated party alternative; therefore, the tax burden will be equal. Thus, there will be no tax incentive or disincentive to related party transactions.

The theory discussed above implies that a competitive firm's gross revenue, which equals price times quantity of output, will be equal to the returns that the factors it employs could earn in the marketplace. The traditional arm's length approach looks at the gross revenue side of this equation; the alternative procedure outlined above looks at the input side. It starts by identifying the factors of production employed by the firm, determining the returns that these factors would earn in the marketplace, and computing the sum. In short, the traditional approach looks for the prices that the firm's outputs would command in the marketplace, whereas the alternative approach seeks to determine the returns that the firm's factors would earn in the marketplace. Both approaches are equally consistent with the basic goal of the arm's length principle, which is to use information about unrelated parties operating at arm's length to determine the allocation of income in a related party setting.

D. FURTHER PRACTICAL PROBLEMS

There are two further practical difficulties in applying the arm's length approach to integrated business economies: application of the approach to monopoly situations and valuation of intangibles. This section briefly describes these difficulties and suggests possible approaches to solving them.

1. MONOPOLY SITUATIONS. In a market that cannot be entered by more than one or a few firms, the existence of "above-normal" profits cannot be ruled out, because potential competitors will not be able to compete them away. The equality discussed above between gross revenue and the returns that factors could earn in the marketplace, and the results derived from it, cannot then be assumed to hold.

However, it may still be possible to apply the basic idea. For example, consider a situation in which a corporation has been granted worldwide patent rights to a unique product. The company still can choose between exploiting this patent through related party or unrelated party dealings, and it would be worthwhile for this decision to be made free of distortions that could be caused by transfer pricing rules. To get an unrelated corporation to provide a good or service, the company would have to pay the unrelated corporation the sum of the returns that would be earned by the factors it would employ. Therefore, it would still be proper to use the alternative procedure to determine the income of the related corporation. ¹⁹⁹

2. VALUATION OF INTANGIBLES. The starting point for the alternative application of the arm's length approach is to measure the factors of production employed by the related parties, and to determine the returns that they would earn in the marketplace. This procedure can be implemented in a straightforward fashion only if the factors can be identified and measured.

However, there is at least one factor of production, intangible assets, for which it is often difficult to assign a precise value. These assets are often unique and it is frequently difficult to decide what returns they would earn if separately employed in the marketplace.

One should not conclude that the presence of any intangible asset will make the alternative procedure impossible to implement. It may be that only one of the related parties employs intangible assets to any significant degree. In this situation it suffices to measure the factors of production employed by the party with measurable assets and to allocate the residual income to the related party employing significant intangible assets. If both parties employ intangible assets, valuation becomes more difficult and, in some cases, judgmental, but not impossible.

E. CONCLUSIONS

1. The arm's length standard remains the theoretically preferable approach to income allocation. Microeconomic theory can be utilized to apply the arm's length standard to an integrated operation.

2. In certain situations, production technologies may be such that unrelated parties operating at arm's length can be expected to coexist with vertically or horizontally integrated multinational corporations. In these cases, arm's length prices should exist and their application to related party transactions should lead to appropriate results. This is the traditional approach embodied in the section 482 regulations.

3. In other situations, vertically or horizontally integrated technologies available only in related party dealings may dominate. Third party prices will be difficult to find in these cases; moreover, the use of the rare third party prices that occur may be inappropriate. However, since information may exist as to the arm's length returns attributable to the factors of production employed by one or both of the related parties, this information can be used to modify the traditional approach and take account of the integrated businesses.

CHAPTER 11

ARM'S LENGTH METHODS FOR EVALUATING TRANSACTIONS INVOLVING INTANGIBLE PROPERTY

A. INTRODUCTION

This chapter discusses the methodology for implementing the arm's length principle for transactions involving intangible property. The goal of this chapter is to propose a theoretical framework for analyzing the situations that have caused the greatest amount of difficulty in the transfer pricing area in order to generate further consideration of the difficult issues involved.

B. ROLE OF COMPARABLE TRANSACTIONS

"Exact comparables," which are those involving the transfer of the same intangible property, supply the best evidence of what unrelated parties would do in a related party transaction. The weight to be given to evidence of "inexact" comparables, which generally are those involving different but economically similar intangible property, is not so clear. Nor is the resort to inexact comparables automatically justified by the arm's length principle. This section first outlines the standards for exact comparables. It then discusses the appropriate role for inexact ones.

1. EXACT COMPARABLES: TWO EXAMPLES

Exact comparables are most likely to occur in connection with the transfer of common products that embody intangibles that are widely available to producers, such as the once unique technology now employed in pocket calculators, digital watches, or microwave ovens. Comparability in such cases of widely available technology is usually easy to demonstrate.

The existence of an exact comparable for unique intangible property, however, is not inconceivable. Consider a multinational company that acquires an unrelated company whose only assets are a small amount of cash, equipment, and the rights to a valuable new invention. If, immediately after this acquisition, the multinational sells those rights to a subsidiary, there really can be no question as to the proper transfer price: it is the acquisition price minus the cash and value of the equipment. In fact, because this comparable is available, any other arrangement could be held suspect. Assume further that the subsidiary and the parent have no other transactions in the initial and following years. The subsidiary's income should include the return to the intangible in all future years, assuming that the subsidiary paid the arm's length consideration at the time of the initial transfer.

As a second example, consider a U.S. corporation that decides to exploit one of its intangibles. It sets up a Mexican subsidiary to serve the Latin American market, while it licenses the Asian rights to an unrelated Korean company. Assume further that the Asian and Latin American markets, and the parent's dealing between the Korean company and the Mexican subsidiary, are comparable in all important aspects. The Korean licensing arrangement should determine the Latin American subsidiary's allocation of income from the transfer of the intangible.

2. STANDARDS FOR EXACT COMPARABLES

The assumptions made in these examples raise the crucial question: How is one to know if a potential comparable is indeed exact? The first requirement is that the comparable transaction involve the same intangible property transferred under substantially similar circumstances. Thus, an exact comparable should involve the same patent, product design, process, trademark, or other intangible transferred to the related party.

However, licenses of intangibles are usually exclusive. Therefore, it is extremely unlikely that the same intangible would be licensed to two different parties for the same use and geographic market. The standards for exact comparables should not require these aspects to be identical.

Instead, two types of additional standards should be met. First, the comparable transaction and the related party arrangement must take place in substantially similar economic environments; these standards may be called "external" ones. Second, the transactions must contain substantially similar contractual features; they must satisfy "internal" standards of comparability.

No amount of general discussion of these standards is likely to turn them into objective tests. As in all matters concerning transfer pricing, facts and circumstances must determine the outcome of specific cases. The following observations, however, may suggest some useful guidelines.

a. "External" Standards

In examining external standards, the essential question is whether unrelated parties would regard the economic environment of the transaction under examination as similar to that of the proposed comparable transaction. In other words, would unrelated parties earn substantially similar profits from a substantially similar transaction? For example, the size and level of economic development of the markets should be substantially similar. ²⁰⁰ If one market is much larger, or if the product is already accepted in one market and not the other, one can presume that unrelated parties would not arrive at the same arrangements to license an intangible into these two markets. As another example, one market may contain many competitors, but in the other a licensee can expect to have a monopoly for a number of years. Again, it is reasonable to conclude that unrelated parties would come to different terms in negotiating licenses for these markets.

Another set of external standards concerns transactions between the licensor and licensee that are collateral to the transfer of the intangible in question. If the parties to one transaction have substantial dealings in the intangible with third parties (such as a cross-licensing arrangement) but the parties to the other set of transactions do not, external standards of comparability are not satisfied. There are clearly reasons why unrelated parties will reach different outcomes if they expect to have further dealings than if they do not. For example, an isolated exchange should not be taken as exactly comparable to a continuing transactional relationship. The comparable used in the U.S. Steel decision ²⁰¹ has been criticized on this basis.

Finally, the level of economic risks being assumed and the functions performed by each party must be similar. Clearly, it would be inappropriate to compare a related party transaction where the affiliate engages solely in manufacturing a product with a transaction in which the unrelated party not only manufactures but also must market the product.

b. "Internal" Standards

To meet this set of standards, the contractual aspects of the transactions being compared must be substantially similar in all important aspects. The most obvious ones include the amount and form of compensation for the transferred intangible. The most common compensation form is a royalty determined as a percentage of sales or quantity produced, but, as Appendix D discusses, other forms are sometimes used. If the comparable transaction contains accelerator or decelerator clauses under which the royalty increases or decreases as sales increase, for example, such clauses should appear in the related party transaction. Other elements of a transaction can have a significant effect on the income realized by unrelated parties to a license or similar agreement. These elements must be substantially similar in order for the unrelated party transaction to be an exact comparable. For example, if the unrelated party agreement provides for the licensee to receive a specified level of technical assistance and training, the related party transaction should contain similar rights. Similarly, if one agreement calls for the licensee to perform significant marketing or product development, while in the other the licensor performs the marketing, the agreements lack internal comparability.

B. EXACT COMPARABLES AND PERIODIC ADJUSTMENTS

A comparable that is exact at the outset of a transaction may lose its exactness over time. There should therefore be two requirements of continued use of the exact comparable over time. First, the arrangements must be consistent in their provision for options and other types of contingency clauses so that they provide for substantially the same types and amounts of adjustments for changing circumstances. Second, the comparables will not remain exact over time unless related parties perform these adjustments as unrelated parties do, under circumstances that are comparable.

Is the concept of an exact comparable so rigid that the results of related party and unrelated party agreements must be the same? Consider a U.S. company that licenses an intangible to two unrelated parties, one in Asia and one in Latin America. It is reasonable to predict that the arrangements will be similar if the economic environments are similar. However, it will probably not be the case that the U.S. company will realize the same income from the two transactions in every year. Business cycles, for example, vary across locations over time. The Asian licensee may have a very profitable experience when times are "lean" in Latin America, or vice versa.

The standards for exact comparables should not require year-by- year equality between the results of the unrelated party arrangement and of the related party one if it is reasonable to conclude that the long-term results will be comparable. Related parties should not be required to exercise rights they might have, if unrelated parties do not in fact exercise them.

4. THE ROLE OF INEXACT COMPARABLES

This section describes the appropriate role for unrelated party transactions that cannot satisfy one or more of the standards for exact comparables. Because of the unpredictable outcomes that inexact comparables have caused in the past, one might argue that they simply should not be used. However, the data presented in Appendix A suggests that some continued use of inexact comparables would be appropriate. The International Examiners reported that they made some use of comparables in making transfer pricing adjustments 75 percent of the time. ²⁰² Although the reported use of comparables for transfers of intangibles in general was lower, it was higher (76.5 percent) for marketing intangibles. ²⁰³ The IEs did not report making final determinations based solely on these comparables in all these cases, but that they made some use of them. Clearly, in practice, inexact comparable transactions provide significant information, even with respect to transfers of intangible property.

The problem, therefore, is not that inexact comparables are useless or misleading. Rather, either they have been given too much emphasis in many cases or inappropriate comparables have been used. The proper conclusion is that it is appropriate to make use of them, but that it is inappropriate to determine transfer prices solely on the basis of inexact comparables. This conclusion is fully consistent with the arm's length principle. The arm's length approach requires that exact comparables, when they are available, should determine transfer pricing allocations of income. However, it does not follow that the same is true of inexact comparables. That is, inexact comparables should be resorted to only when exact comparables are unavailable. Further, they should not be given priority over the alternative method outlined in section C of this chapter in all cases.

5. SELECTION OF APPROPRIATE INEXACT COMPARABLES

Once it is determined that an exact comparable does not exist, how should inexact comparables be selected? The most obvious point is that the external and internal standards discussed previously should parallel those in the transaction at issue as closely as possible. For example, if the unrelated parties in the potential comparable operate in a very different economic environment -- if the market is much smaller or the related parties carry on a much broader set of transactions -- then the comparable should not be used to justify the related party arrangement. Similarly, if the intangible in the unrelated party transaction is at a very different stage of development or concerns a dissimilar product or service, then its use as a comparable is inappropriate.

In more traditional terms, an unrelated party arrangement should be used as an inexact comparable if the differences between it and the related party transaction can be reflected by a reasonable number of adjustments that have definite and ascertainable effects on the terms of the arrangement. The current regulations for section 482, although silent on this issue in connection with transfers of intangible property, discuss it quite carefully in connection with transfers of tangible property. They mention that adjusting a sale for differences in transportation costs or minor physical modifications would probably be appropriate, but that an adjustment for the presence or absence of a trademark would not. ²⁰⁴

This approach should be extended to transfers of intangible property. For example, unrelated party arrangements frequently require the licensor to provide a specified amount of training or expert assistance to the licensee for a brief period. ²⁰⁵ It may be possible to adjust a comparable that includes such a provision by comparing it with an arrangement that does not, or that provides for less assistance.

At the other extreme, consider an attempt to compare an unrelated party license with a related party license when the unrelated party licensee performs different functions than the related party licensee. For example, the former may be responsible for substantial marketing, while the latter may not. It seems clear that the effect of this difference would not be definite and ascertainable. Therefore, an adjustment for it would be too speculative to be appropriate.

Similarly, intangibles differ in their fundamental profitability. Attempting to compare a low-profit intangible to a high-profit one by adjusting for this difference would clearly be too speculative to be appropriate. Comparable transactions involving intangibles that are likely to be of typical or average profitability are therefore appropriate inexact comparables only if the related party intangible under analysis is typical or average.

The current regulations contain a list of twelve factors which are essentially internal and external standards that might be examined in order to determine whether an unrelated party license is an appropriate inexact comparable. As many observers have pointed out, however, it is difficult to derive useful guidance from this list, because it does not discuss the relative weights to be placed on the factors in a given situation. For example, prospective profits to be realized from the intangible appears late in the list, but after the 1986 Tax Reform Act this factor must be given special consideration. In contrast, the first listed item cites prevailing industry rates, which should not be relied upon unless the intangible being transferred is demonstrably average, based on observable indicators of profitability.

Another approach that provides a framework for use of inexact comparables is "functional analysis." Although not explicitly mentioned in the regulations, this procedure is outlined in the IRS Manual ²⁰⁶ and has been found to be a useful place to start in transfer pricing situations. In essence, the goal of functional analysis is to identify the economic activities actually undertaken or to be undertaken by the parties in both the related party situation and unrelated party situation. The most appropriate comparables may be chosen by identifying the ones in which the unrelated parties carry on the same major economic activities as the related parties. Section C of this chapter examines functional analysis in the context of the arm's length rate of return method. Functional analysis is an equally valid approach for analyzing comparables on the basis of the similarity between the economic activities performed. ²⁰⁷

6. USE OF INEXACT COMPARABLES AND PERIODIC ADJUSTMENTS

It is inappropriate to use inexact comparables to justify a related party transaction merely by analyzing similarities at the time of the initial transfer. For example, there may be valid inexact comparables that justify the establishment of a related party agreement with a fifteen percent royalty rate. These comparables may further justify fixing this rate for two years. Even if no adjustment were required during the first two years, in year three the taxpayer may not continue to rely upon the prior inexact comparables unless, after re-examination, use of these comparables remains appropriate.

Suppose a significant change occurs during the term of a license agreement. For example, suppose a taxpayer licenses a product design to a related party. At the time of the transfer, the taxpayer makes a good faith estimate that the product will be a routine one, and will attain 10-25 percent of its market. Based on this fact, the taxpayer gathers information on comparable transactions (none of which can meet the standards for an exact comparable). The information indicates that a royalty rate of 10 percent of sales is appropriate. The comparables contain varying duration and contingency clauses. In year three, the product design becomes uniquely popular and garners 95 percent of the market. Given this set of facts, the inexact comparables previously used may no longer be used in year three and succeeding years to justify the 10% related party royalty rate. ²⁰⁸

Although unlikely, the taxpayer may find inexact comparables involving products with a 95 percent market share that demonstrate that the 10 percent royalty rate is still appropriate. More realistically, suppose that there are comparables that show that products in the taxpayer's industry with a 95 percent market share typically command a much higher royalty, or, as may be more likely, that there are no comparables for such a situation. In such case, the taxpayer must use the arm's length return method outlined in section C below either to justify the royalty rate previously set or to adjust the related party arrangement to bring it into compliance with the results of this new analysis.

C. AN ARM'S LENGTH RETURN METHOD

The previous chapter discusses why a method that looks to arm's length returns, as distinct from arm's length prices, is appropriate. This section discusses how such a method should operate. Although some of the terminology in this section may be new, most of the techniques discussed in it are not. One of the main arguments for the development of an arm's length return method, in fact, is that taxpayers, the Service, and especially the courts have found it necessary to use ad hoc and incompletely developed versions of such a method in the past and will undoubtedly continue to do so in the future. Therefore, the goal of this discussion is to lay a foundation for this approach so that it may be used to achieve more consistent and satisfactory results.

1. BASIC ARM'S LENGTH RETURN METHOD

a. GENERAL DESCRIPTION

Consider a U.S. company, Widgetco, that holds worldwide patent rights to the widget, a high-tech light gathering device that is expected to be a vital component in certain satellites and scientific instruments. Widgetco intends to exploit this patent in the following way. A foreign affiliate will manufacture the widgets, under license from Widgetco. Besides utilizing the license, Widgetco and the affiliate will engage in the following transactions. Widgetco will sell various types of microchips, seals and filters to the affiliate, which will also buy some of these products from unrelated suppliers. The affiliate will use these components to manufacture the widgets and sell them to Widgetco, which will market and distribute them. Widgetco maintains a research staff that developed the widget and will continue to try to improve it.

It would be very difficult to depend purely on comparable transactions, as traditionally defined, to establish the proper allocation of income in this sort of "round trip" transfer-pricing situation. There are three separate types of comparables that would have to be found. The first is a set of prices for the components the parent will sell to the affiliate. The second is the royalty that the affiliate should pay to the parent under the production license. The third is the transfer price for the finished widgets. Although it may be possible to find exact comparables for one of these types of transactions, such as the purchase price of some of the components, it will in general be impossible to find comparables for all three. Further, finding an answer to only one aspect of the problem provides little help in deciding the proper allocation of income between the related parties.

Another approach would be to try to find one or more comparable transactions in which a company contracts with an unrelated party to manufacture a product similar to the widget. It is likely that these transactions will not be good inexact comparables. In general, the form, risks, and extent of relationships in the related party case will at least appear to be quite different from those in a contract manufacturer transaction. Among other things, the foreign affiliate carries raw material, work-in-process, and finished goods inventories and should receive a normal market return on its activities that reflects its investment in such assets and the moderate risk that manufacturers using routine manufacturing processes bear with respect to their investment in manufacturing facilities and inventories. Using the terminology of the previous section, a contract manufacturer transaction is likely to fail both the external and internal standards for inexact comparables, because both the types of transactions between the parties and the terms of their agreements will differ. While comparables of this type should not be discarded from all consideration (because this type of comparable may provide useful information), it would be improper to base a resolution of the transfer pricing issue solely on such information.

An arm's length return approach would start from a different perspective. It would seek to identify the assets and other factors of production that will be used by the related parties in the relevant line of business and would try to assign market returns to them.

The first step in this process is to perform a functional analysis -- i.e., to break down each line of business into its component activities or functions. It should then be possible to identify which of the functions utilize only factors of production that can be measured and assigned market returns and those which do not. In most cases, identifying functions with measurable factors will lead to distinguishing between those functions that make significant use of preexisting intangible assets and those that do not. In Widgetco's case, Widgetco owns several types of assets that are difficult to measure, including the widget patent and other manufacturing intangibles, the ongoing enterprise value of the research staff, and the marketing intangibles. On the other hand, the foreign affiliate utilizes measurable factors of production, assuming that the manufacturing process is a routine one. Specifically, the affiliate employs labor, plant, equipment, working capital, and what might be called "routine" manufacturing intangibles -- i.e., know-how related to efficiency in routine manufacturing processes that most manufacturers develop through experience. Since it is easier to evaluate the factors of production to be used by the foreign affiliate, the market rate of return analysis will focus on the affiliate. As the theoretical analysis in the previous chapter demonstrates, focusing on the return of the affiliate in this manner is a valid extension of the arm's length principle.

Next, income should be assigned to each of the functions with measurable factors -- here the functions performed by the affiliate. The reason for identifying measurable factors (and, therefore, focusing on the affiliate) is that the functions that employ measurable factors will probably be carried on by a wide range of unrelated parties for which information will probably be available regarding market returns earned by them. A market return consistent with the returns of unrelated parties can be assigned to each of the affiliate's functions since they all employ measurable factors. Once returns are identified for all of the affiliate's functions, the residual income from the line of business is then allocated to Widgetco.

Assume, as stated, that the only function to be performed by Widgetco's foreign affiliate is manufacturing and that this function does not involve the significant use of intangible property developed by the affiliate or purchased by it from unrelated parties. Under the rate of return method, the assets of the foreign affiliate would be divided into liquid working capital and all other assets (i.e., the production assets). The actual return on the liquid working capital will be identified and allocated to the foreign affiliate. Rates of return on production assets used in similar manufacturing activities of similar risk must be identified or estimated. ²⁰⁹ Income will then be allocated to the affiliate for its manufacturing activity in an amount equal to the identified or estimated rate of return as applied to its production assets. This rate of return would include, by definition, a return on routine manufacturing intangibles that manufacturers commonly possess as well as a return for assuming normal business risks that manufacturers bear with respect to their investment in manufacturing facilities and inventories. ²¹⁰ The residual amount of income from the line of business is allocated to Widgetco.

The same allocation would be made to the foreign affiliate if it sold its output to a second affiliate and not back to Widgetco. In neither case would the foreign affiliate receive a return for marketing its product.

b. USE OF ARM'S LENGTH INFORMATION

There are two ways that arm's length information can be used to allocate income to the activities of the Widgetco manufacturing affiliate. The first method has been previously described -- to identify the unrelated parties' rates of return on assets utilized in a particular function, taking into account only the non-liquid assets relevant to the function in the line of business being examined. If satisfactory measures of the unrelated parties' assets are available, it should be possible to calculate an appropriate rate of return for each function and apply it to the related party's assets utilized in that function.

The second way to use the arm's length information is to measure it against a yardstick other than rates of return on assets. A common alternative is the ratio of income to operating costs. For example, in the DuPont case, ²¹¹ an expert witness, Dr. Charles Berry, computed the ratio of gross income before reduction by operating costs and interest to operating costs for DISA, DuPont's Swiss affiliate, and for a number of unrelated parties performing similar functions. This analysis is useful to measure returns on service activities and in other situations where assets are difficult to measure consistently or, more generally, where there is reason to believe that the relationship between income and costs is more stable or easier to measure than the relationship between income and assets. As is true with assets, it is important to consider the types of costs and their relationships to income earned, not just the totals. For example, some analysts have used the ratio of gross income to "above the line" costs. This approach is suspect if the unrelated parties incur proportionately larger amounts of "below the line" costs, such as advertising, than the related affiliate incurs.

The use of both types of unrelated party information is consistent with the fundamental goal of the basic arm's length return method, which is to use information about unrelated parties to determine the returns that would have been earned had the related parties' activities been undertaken at arm's length. Therefore, both approaches are potentially applicable depending upon the availability of either type of information and the appropriateness of using either type of information in the particular circumstances.

c. APPLICABILITY OF BASIC ARM'S LENGTH RETURN METHOD

The basic arm's length return method should have wide, but not universal, applicability in situations where exact or inexact comparable transactions are not available. It will not be sufficient alone, however, when both of the related parties own preexisting and significant intangible assets that are vital to the success of the project -- for example, if Widgetco's foreign affiliate actively markets the products it manufactures in a manner that utilizes significant self-developed marketing intangibles. In such cases, it will be difficult to find a set of unrelated parties that possess the same type and amount of intangible assets as the affiliate and are thus able to perform the same activities. It will be difficult, therefore, to obtain the arm's length information needed to assign a return to either affiliate's activities.

This discussion is not meant to imply, however, that the basic arm's length return method is to be avoided whenever an affiliate possesses any amount of intangible assets. It is unlikely, for example, that any manufacturing operation is so simple that it does not involve the use of some intangibles. An affiliate engaged only in manufacturing may employ a skilled labor force, and the efforts expended in recruiting and training it may create at least some amount of going concern type of intangible. Further, the affiliate's experience in producing the parent's designs may lead to the development of some amount of know-how. These facts alone, however, should not prevent the application of the basic method. The reason is that unrelated parties performing similar activities will, in general, possess these types of "routine" intangibles. Therefore, by measuring the return on assets that unrelated parties earn for performing similar activities and bearing similar risks, the basic method will automatically capture the returns earned by these "routine" intangibles and will properly attribute them to the affiliate. It is only when the affiliate owns some type of intangible that is of major importance to the enterprise, and which few unrelated parties possess, that the basic method is insufficient standing alone to resolve the issue.

The basic arm's length return method will probably be appropriate for most manufacturing affiliates. It should be possible to observe the rates of return on assets or ratios of income to costs that are earned by unrelated parties performing similar manufacturing activities involving similar risks and amounts of routine intangibles. It is possible to think of exceptions, however. Consider a corporation that has assembled a large and valuable team of engineers and skilled craftsmen within a European subsidiary in order to develop or perfect a complex manufacturing process. If no or few unrelated parties would be capable of performing this development activity, the basic arm's length return method will not be sufficient standing alone to resolve the case.

Similarly, the basic method will be applicable to many distribution and marketing affiliates. An affiliate that sets up and maintains a distribution network undoubtedly possesses a going concern intangible; an affiliate that markets products to industrial customers by participating in trade shows and maintaining a staff of salespersons undoubtedly possesses some amount of know-how. However, it seems likely that these sorts of activities are undertaken by unrelated parties that possess similar amounts of going concern value and know-how. Therefore, it should be possible to determine the arm's length returns on assets for these activities, which will include the appropriate returns to these "routine" intangibles. In other situations, however, the basic method alone will not suffice.

2. PROFIT SPLIT ADDITION TO THE BASIC ARM'S LENGTH RETURN METHOD

Although the basic arm's length return method should be widely applicable, there are situations in which its use alone will clearly be inadequate. A large multinational corporation may have foreign subsidiaries that have research, marketing, planning, manufacturing, and other divisions that are as large and active as those of all but the biggest U.S. companies. Therefore, these affiliates may perform complex functions, take significant risks and own significant intangible assets equal to those of the typical parent corporation. If so, the basic arm's length return method would be impossible to apply because exact or inexact comparables, or rates of returns, for these complex functions are generally unavailable.

Consider Teachem, a U.S. company that is a world leader in designing and producing educational toys. It serves its major overseas market, Western Europe, through a French sales affiliate, Enseignerem. Teachem is planning to license a new set of designs to Enseignerem, who will modify them in minor ways, such as translating the instructions and markings, and who will hire local contract manufacturers to produce them. Enseignerem will utilize its own trademark and be responsible for all aspects of marketing and distribution in Europe. It will decide which of the toys to include in its line, set its own advertising budget, and design campaigns to promote the new line and its Enseignerem trademark. It also maintains a sales force and distribution network. Teachem maintains a research and testing staff to develop new products. The affiliate does not.

Assume further that Teachem has a business policy of not licensing its designs to unrelated parties. Exact comparables will therefore be absent, and inexact comparables may be difficult to find. How would an arm's length return approach be applied to determine the appropriate royalty rate to be paid by Enseignerem for the use of Teachem's designs? The first step is to identify the functions performed by the parent and the subsidiary in the line of business in which the licensed designs are used (the sale of new toy designs in the European market). The parent should be allocated the returns on the basic product designs, while the subsidiary is entitled to the returns to be earned by its trademarks, marketing efforts, and distribution network, plus any intangibles related to the modifications.

The next step is to identify the functions that employ measurable factors -- i.e., activities that do not involve the use of significant intangible assets. These activities should be analyzed using the basic arm's length return method. Enseignerem's distribution and manufacturing activities may be examples. It should be possible to find unrelated parties that perform these types of activities and incur similar business risks. Thus, it should be possible to determine an arm's length rates of return on assets (or income-to-costs ratio) for each activity and apply the arm's length rate or ratio to the appropriate related party factors. The resulting income should then be allocated to the party performing the activity. In this case, income attributable to distribution and manufacturing activities would be allocated to Enseignerem. If the parent corporation performs or is to perform routine activities involving the line of business, they too should be analyzed using the basic method.

These two steps will leave a quantity of income not yet allocated and a set of activities involving significant intangible assets not yet accounted for. The goal of the first two steps is to isolate the income that is attributable to the significant intangible assets owned by the corporate group as a whole and used in the line of business in question -- primarily the designs owned by Teachem and the marketing intangibles (including the trademark) owned by Enseignerem.

The goal of the remaining step is to identify the intangible income attributable to the relevant line of business and then split that income according to the relative value that the marketplace would put on each party's significant intangible assets had they been employed by unrelated parties operating at arm's length. The intangible income is equal to the combined net income from the line of business less the income allocated under the prior steps to functions with measurable factors -- i.e., the residual combined net income determined after applying the basic rate of return method to activities with measurable factors of the parties. In splitting this residual amount between the related parties, it is not necessary to place a specific value on each party's intangible assets, only a relative value. Of course, it is easier to state this principle than to describe in detail how it is to be applied in practice. In many cases, there will be little or no unrelated party information that will be useful in determining how the split would be determined in an arm's length setting. Furthermore, the costs of developing intangibles, even if known, may bear no relationship to value, especially in the case of legally protected intangibles, and generally should not be used to assign relative values to the parties' intangible assets. Splitting the intangible income in such cases will largely be a matter of judgment. There are two possible sources of arm's length information, however.

First, it may be possible to find unrelated parties that engage in similar activities and that use similar intangibles. The unrelated party transactions must be economically similar, of course, including the level of economic risks assumed. It would be inappropriate to use a profit split derived from a situation in which the unrelated parties' intangibles were much less (or more) profitable than those owned by the related parties. Further, it would be inappropriate to compare the split derived from a transaction in which an unrelated party conducted only wholesale level marketing, for example, with a related party situation in which an affiliate markets products to consumers. These requirements resemble the standards discussed above for inexact comparables. The analysis of unrelated party profit splits should explain the relationship between the observed profit splits and the overall profitability of the significant intangibles involved with a reasonable degree of accuracy. It is also necessary to analyze the functions that the unrelated licensors and licensees perform and the risks that they bear. Comments in this area should focus on how such an analysis can be implemented.

Second, in some circumstances, a taxpayer may have arm's length evidence of the value of its own or its affiliate's intangibles. For example, a taxpayer may have recently purchased its affiliate and may have some basis for determining the value of the intangible assets using the purchase price.

3. ARM'S LENGTH RETURN METHOD AND PERIODIC ADJUSTMENTS

Issues involving periodic adjustments are easier to analyze for the arm's length return method than for the methods involving comparable transactions. Because the basic arm's length return method looks to the factors of production used by the parties, the income allocation should adjust as the factors change. Thus, as an affiliate's plant, equipment, and other measurable factors change from the projections in the initial analysis, the income allocated to them should change. Similarly, the profit split percentage is intended to reflect the relative values of significant intangible assets owned by the parties. When the value of intangibles belonging to one of the parties has changed, the percentage should be changed. For example, a dramatic increase in sales may be due to either a recent product improvement or an extensive marketing campaign, in which case proportionately more profit should be assigned to the developing or marketing affiliate, respectively.

These conclusions are not intended to imply, however, that there must be year-by-year equality between the related parties' incomes and the results of an ideal application of the method. The prior discussion of long-run versus year-by-year results is again relevant. For example, consider an independent firm that uses only plant and equipment. Although it should earn the market rate of return in the long run, it will, in general, experience lower returns or even losses in "lean" times and higher returns at the other end of the business cycle. Periodic adjustments will be required for significant changes in income. Therefore, changes (either negative or positive) that are less than significant may tend to even out in a long-term equilibrium. The absence of a requirement for an adjustment over time when insubstantial changes in income occur is a corollary of the rule that de minimis adjustments will not be made. Chapter 8 discusses this issue, including reasons why explicit inter-year set-offs would not be practicable, in more detail.

D. PRIORITY AND COORDINATION AMONG METHODS

The previous sections of this chapter discussed two broad approaches for analyzing transfers of intangible property: an approach based on comparable transactions and one based on arm's length returns to factors employed. Which is to be used in a given situation? The answer is clear in one case. If an exact comparable is present, it and only it should be used to determine the allocation of income from the transfer. It follows from the definition of exactness that there can be no better evidence of what unrelated parties operating at arm's length would have done.

Finding an exact comparable, however, can be extremely difficult. In the majority of cases, particularly contested ones, the allocation of income will come from either the inexact comparable method or the arm's length return method. The facts and circumstances of each case should determine which method -- or methods -- should be used.

There are four basic types of cases. In the first, the intangible for which a section 482 transfer price is being determined is comparable to those used by unrelated taxpayers and each of the related parties is expected to employ significant and complex intangibles. The inexact comparables method should be chosen.

In the second case, the section 482 intangible is unique, and the affiliate utilizing the intangible will engage in functions that use only measurable factors of production and routine amounts of intangibles. The basic arm's length return method should be used.

In the third case, the section 482 intangible has many competitors and the affiliate using the intangible will engage in simpler kinds of functions. BOTH of the methods are potentially applicable. Under the theory on which they are based, they should yield similar results. This situation should, in practice, be the easiest for the taxpayer to analyze and should engender the least amount of controversy.

In the final case, the section 482 intangible is unique and both of the related parties own one or more significant intangibles that will be used in exploiting it. This is the hardest case. The profit- split version of the arm's length return method must be used.

E. RISK-BEARING IN RELATED PARTY SITUATIONS

Economic environments are full of uncertainty, and this fact must be recognized in all methods of income allocation. In general, in a related party transaction, the market reward for taking risks must be allocated to the party truly at risk.

Companies take risks in all dealings in the marketplace, and are rewarded for doing so. Some of this risk disappears in related party transactions. The legislative history of the Tax Reform Act of 1986 noted:

In addition, a parent corporation that transfers potentially valuable property to its subsidiary is not faced with the same risks as if it were dealing with an unrelated party. Its equity interest assures it of the ability ultimately to obtain the benefit of future anticipated or unanticipated profits, without regard to the price it sets. ²¹²

How should risk be accounted for in related party transactions? The riskiness of true economic activities gives rise to greater returns in the marketplace; therefore, if one part of an enterprise is inherently more risky than another, more income should be allocated to it. This allocation should be based on the risks arising out of the true economic activities undertaken by the parts of the enterprise, not on mechanisms that merely shift risks within the group.

This conclusion has implications for the proper application of both the comparables approach and the arm's length return approach. First, in searching for appropriate comparables, one should look for situations in which an unrelated party contracted to perform an economic activity that is about equal in riskiness to the activity done by the affiliate; it would be inappropriate to rely on comparables in which the unrelated party undertook a significant risk of some kind not undertaken by the related party. Likewise, in applying the arm's length rate of return method, unrelated party rates of return should be used only if they reflect similar levels of risk. But merely stating that unrelated party transactions must bear the same level of risk as the related party begs the question of what risks the related party should be allowed to assume. It's necessary to decide first what risks may be appropriately assumed by the related parties, depending on the functions that each performs. Only then is it possible to identify what unrelated party arrangements are comparable so that comparable rates of return (or inexact comparable transactions) can be determined.

Returning to the Widgetco example, the affiliate, as the manufacturer, is at risk both with respect to its investment in plant and equipment and with respect to inventories. The risk with respect to plant and equipment will be significant only if the facilities cannot be used for other purposes without incurring significant additional costs. If there is risk that the product will not be successful because there is uncertainty that the product will perform as anticipated or have the usages anticipated, then that risk should be borne by Widgetco as owner of the manufacturing intangible (the patent) and should not be reflected in the affiliate's rate of return. The affiliate's return should reflect only the moderate level of risk borne by manufacturers of products that are reasonably expected to achieve market acceptance. Likewise, if there is uncertainty that the product will be marketed successfully, then that marketing risk also should not be borne by the affiliate but should probably be shared in some fashion by the owner of the manufacturing intangible and the marketer, depending upon the extent to which anticipated profits from the enterprise are attributable to the manufacturing intangibles or the marketing activities.

On the other hand, assume that the risk does not relate to undue uncertainty regarding the anticipated performance or usage of the product or the market acceptability of the product. Assume, instead, that it is highly uncertain whether the product can be produced cheaply enough to make the enterprise viable, or that it is uncertain whether the manufacturing process will produce the product with the same quality as prototypes produced in the laboratory. These risks are risks inherent in the manufacturing function and should probably be shared in some fashion between the owner of the manufacturing intangible and the manufacturing affiliate. If the manufacturing affiliate is itself developing an efficient production process that attempts to achieve low production costs or to assure consistency in quality of output, then the affiliate should be allocated a return that reflects a substantial portion of that risk being borne by the manufacturing affiliate. (In such case, the manufacturing affiliate would bring to bear significant manufacturing process intangibles which would necessitate the application of the profit split addition to the basic arm's length return method.) If, instead, the owner of the intangible has also developed the production process without significant contribution by the manufacturing affiliate, then a separate manufacturing intangible related to the production process has been created, and the owner of such intangible is entitled to an arm's length return. The manufacturing affiliate's return should not bear more than the moderate level of risk borne by manufacturers of products that are reasonably expected to achieve market acceptance.

F. COORDINATION WITH OTHER ASPECTS OF TRANSFER PRICING

The purpose of this chapter is to provide a framework for the development of new methods for allocating income from intangible property. Therefore, methods for allocation of income in situations involving provision of services ²¹³ or transfers of tangible property ²¹⁴ may appear at first glance to be outside the scope of the present discussion. However, the rules for intangible property must be coordinated with the rules for other types of transactions between related parties for obvious reasons. Transfers of tangible property and provision of services frequently accompany a transfer of intangible property; all three are often bundled into a single economic transaction. Further, if the rules relating to one type of transaction become more or less favorable to taxpayers, then they will easily be able to find ways to structure their transactions to take advantage of the disparities.

It may be helpful to establish a priority for single economic transactions that involve more than one type of transfer. For example, licensing agreements often contain clauses that require the licensor to provide training or other services to the licensee. Further, transfers of tangible property often involve intangibles, since the goods transferred often depend for their value on embodied trademarks or patents. In these cases, the basic allocation of income issue should be settled under the rules to be developed for intangible property.

G. CONCLUSIONS AND RECOMMENDATIONS

1. An approach incorporating two alternative methods for determining transfer prices for intangibles would achieve more appropriate allocations of income and greater consistency in result.

2. The first method uses exact or inexact comparables when they exist.

a. An exact comparable is the same intangible licensed to unrelated parties, when the circumstances surrounding it and the related party transfer are similar. The price derived from this method has priority over all others.

b. An inexact comparable is an intangible very similar to the intangible transferred to a related party, but not identical. Differences must be definite and ascertainable. If the other intangible, the contractual arrangements, or the economic circumstance are too different, it may not be used as a comparable.

3. The second method uses an arm's length return analysis instead of looking for a comparable transaction and adopting that transfer price as the section 482 transfer price for the related party transfer.

a. The basic arm's length return approach applies when one party to the transaction performs economic functions using measurable assets or other factors, but not using significant intangibles of its own. The first step is to break down the relevant line of business into its component activities or functions and measure the factors (generally assets) utilized by the party performing the simpler set of functions. Income attributable to those functions is determined by identifying rates of return to assets or other factors of unrelated entities performing similar economic activities and assuming similar economic risks, and applying a comparable rate of return to the assets or other factors of the related party. Any residual income is thereby effectively assigned to the other party. The royalty rate or other transfer price for intangibles utilized by the related party must be set to achieve the allocation of residual income to the other party.

b. When both parties perform complex economic functions, bear significant economic risks, and use significant self- developed intangibles, a profit split analysis must be added to the basic rate of return method. Under the profit split analysis, the combined net income from the line of business must first be determined. The profit split analysis assigns the residual net income, determined after applying the basic rate of return method to the measurable assets of the parties, between the parties based upon the relative values of the parties' unique intangibles.

4. While the standards for exact or inexact comparables do not require year-by-year equality between results of the unrelated party arrangements and related party arrangements, unrelated party arrangements can lose their comparability over time as the facts and circumstances relevant to the standards for comparability change. In such case an adjustment to allocations of income may be necessary. Under the arm's length return method, income allocations reflect the functions performed by the parties and -- i.e., they reflect the measurable factors of production and the value of significant intangibles employed by the parties in performing those functions. Therefore, income allocated to the related parties under the arm's length return method will change as the functions performed or the factors of production or value of intangibles employed by the parties change.

5. Other than in the case of exact comparables, there should be no priority among these methods. However, each is designed to be utilized under a specific set of facts, so the underlying fact pattern should determine the method or methods to be used.

6. Risk should be accounted for under all methods described in this chapter, since the market rewards risk takers. However, the risk premium should be attributed to the affiliate undertaking the economic function in which the risk inheres.

IV. COST SHARING ARRANGEMENTS Preceding parts of this study have analyzed the proper prices to be applied, or amount of income to be allocated, when an intangible is transferred between related parties. Cost sharing arrangements are an alternative method by which related parties can develop and exploit intangibles. The history of such arrangements, their acceptance for tax purposes, and an outline of rules that should be followed for post-1986 cost sharing arrangements are discussed in this part.

CHAPTER 12

HISTORY OF COST SHARING

A. INTRODUCTION

In general, a cost sharing arrangement is an agreement between two or more persons to share the costs and risks of research and development as they are incurred in exchange for a specified interest in any property that is developed. Because each participant "owns" specified rights to any intangibles developed under the arrangement, no royalties are paid by the participants for exploiting their rights to such intangibles. The Conference Report accompanying the 1986 Act indicates that Congress did not intend to preclude the use of bona fide research and development cost sharing arrangements. However, Congress expected the results produced under a bona fide cost sharing arrangement to be consistent with results under the commensurate with income standard. ²¹⁵

Cost sharing arrangements have long existed at arm's length between unrelated parties. Typically, unrelated parties pool their resources and expertise in a joint effort to develop a specified product in exchange for a share of potential profits. The Service has little experience with ordinary unrelated party cost sharing arrangements because they are at arm's length and normally do not have unusual tax consequences. ²¹⁶

In view of the limited information currently available on both related and unrelated cost sharing agreements, the Service and Treasury would appreciate receiving information from taxpayers regarding their contractual arrangements and experience with cost sharing.

B. 1966 PROPOSED SECTION 482 REGULATIONS

Proposed Treas. Reg. section 1.482-2(d)(4), published on August 2, 1966, provided extensive rules for cost sharing arrangements. ²¹⁷ The proposed regulations permitted any affiliate (other than one in the trade or business of producing intangible property) to participate in the cost sharing arrangement, provided that the intangible property was intended for use in connection with the active conduct of the affiliate's business. The regulations specifically authorized cost sharing arrangements for single projects, but did not disqualify multiple projects or continuing arrangements. The sharing of costs and risks was required to be proportional to the anticipated benefits to be received by each member from the arrangement. Cost sharing was required to be based on sales, profits, or other variable criteria.

The 1966 proposed regulations did not explicitly address the "buy-in" question -- that is, the compensation to be paid to the developer or owner of intangibles that are made available at the time the arrangement commences. They did, however, require that an arm's length amount be paid to the affiliate that provides intangibles that substantially contribute to the arrangement. Any required section 482 allocation for services provided by other affiliates was also to be included as a cost of the arrangement.

The proposed cost sharing regulations were ultimately abandoned in favor of the simpler general requirements presently contained in section 1.482-2(d)(4).

C. CURRENT REGULATIONS

The current rules in section 1.482-2(d)(4) state that a cost sharing agreement must be in writing and provide for the sharing of costs and risks of developing intangible property in return for a specified interest in the property that may be produced. A bona fide cost sharing arrangement must reflect an effort in good faith by the participants to bear their respective shares of all costs and risks on an arm's length basis. The terms and conditions must be comparable to those that would have been adopted by unrelated parties in similar circumstances. ²¹⁸

D. FOREIGN EXPERIENCE WITH COST SHARING AGREEMENTS

The 1979 OECD report on Transfer Pricing and Multinational Enterprises ²¹⁹ stated that, although international cost sharing agreements for research and development costs were not common, some had recently been entered into by large multinational enterprises. The OECD report indicated that, with the exception of the United States, none of its members had laws or regulations pertaining specifically to cost sharing arrangements. A major concern expressed by the OECD report was that the participants to the arrangement be in a position to benefit from any intangibles developed under the arrangement before the cost sharing payments would be allowed as deductible expenses. The OECD report stated that the United States did not require a profit mark-up for research and development activities performed. The OECD report reflected a consensus, however, that a profit mark-up would be appropriate when research was performed at the specific request of a member of the cost sharing group. There was also a consensus that withholding taxes should not apply to cost sharing payments when paid.

A few countries have specifically addressed cost sharing arrangements since publication of the OECD report. Germany has developed guidelines ²²⁰ for the use of cost sharing agreements in cases in which expenses for research and development can only be valued in the aggregate. Division of the costs must be based on the extent that each party actually benefits or expects to benefit from the arrangement. When determining costs incurred, no profit element is recognized for tax purposes. Appropriate costs to be shared may include a contribution to general and administrative costs.

E. DEFICIT REDUCTION ACT OF 1984

Section 367(d), enacted in 1984, provides that a transfer of intangibles to foreign corporations in an exchange described in section 351 or 361 is to be treated as a sale, with the transferor being treated as receiving amounts that reasonably reflect the amounts that would have been received under an agreement providing for annual payments contingent on productivity, use, or disposition of the property. Such payments are treated as reductions of the foreign entity's earnings and profits and as U.S. source income to the U.S. recipient. The "Blue Book" discussion of section 367(d) indicates that it is to have no application to bona fide cost sharing arrangements. ²²¹ The Blue Book further recognized that it may be appropriate for the Treasury Department to elaborate on the current cost sharing rules to address problems with cost sharing arrangements. ²²²

F. COST SHARING UNDER SECTION 936(h)

After 1982, the intangible income of a domestic corporation qualifying for the possessions tax credit must be included in the income of its U.S. shareholders, unless the possessions corporation either elects the cost sharing method or elects the 50% of combined taxable income method, both of which are contained in section 936(h). Under the section 936(h) cost sharing election, the possessions corporation must pay its share of the affiliated group's total research and development costs based on the ratio of sales by the affiliated group of products produced in the possession to total sales by the affiliated group of all products. The cost sharing payment must be computed with respect to "product areas" rather than single projects. "Product areas" are defined, in general, by reference to the three-digit Standard Industrial Classification codes (SIC codes) promulgated by the Commerce Department. Cost sharing payments made by the possessions corporation are not treated as income to the recipient but reduce otherwise allowable expenses.

A possessions corporation making the cost sharing election is treated as owning the manufacturing intangibles utilized in its business, and the income from such intangibles then becomes eligible for the possessions tax credit. Pricing of products between a possessions corporation electing the cost sharing method and its domestic U.S. affiliates must still meet the requirements of section 482, taking into account that the possessions corporation is treated as owning the manufacturing intangibles.

Pursuant to an amendment made by the 1986 Act, a possession corporation making the cost sharing election must pay the greater of 110% of the pre-1986 statutory cost sharing amount or the royalty required to be paid to the developer of the intangibles under the commensurate with income standard. ²²³ Given the special circumstances in which the section 936(h) cost sharing provisions apply and the 1986 Act changes, section 936(h) cost sharing arrangements do not provide much guidance with regard to the appropriate requirements for other cost sharing arrangements.

CHAPTER 13

COST SHARING AFTER THE TAX REFORM ACT OF 1986

A. INTRODUCTION

The Conference Report to the 1986 Act states that, while Congress intends to permit cost sharing agreements, ²²⁴ it expects cost sharing arrangements to produce results consistent with the purposes of the commensurate with income standard in section 482 -- i.e., that "the income allocated among the parties reasonably reflect the actual economic activity undertaken by each." The Committee Report also emphasized three potential problems that should be addressed in any revision of section 1.482-2(d)(4).

The first problem is selective inclusion in the arrangement of high profit intangibles. The Report states:

Under a bona fide cost sharing arrangement, the cost sharer would be expected to bear its portion of all research and development costs, on successful as well as unsuccessful products within an appropriate product area, and the cost of research and development at all relevant development stages would be included. ²²⁵

The second issue concerns the basis on which contributions are to be measured. The Report states:

In order for cost-sharing arrangements to produce results consistent with changes made by the Act to royalty arrangements, it is envisioned that the allocation of R&D cost-sharing arrangements generally be proportionate to profit as determined before research and development. ²²⁶

The third specific Congressional concern relates to the "buy-in" issue. The Report states:

In addition, to the extent, if any, that one party is actually contributing funds toward research and development at a significantly earlier point in time, or is otherwise effectively putting its funds at risk to a greater extent than the other, it would be expected that an appropriate return would be required to such party to effectively reflect its investment. ²²⁷

This chapter examines these and other issues that have arisen with regard to the requirements for a bona fide cost sharing arrangement after the Tax Reform Act of 1986. It should be made clear at the outset that, if an arrangement is not bona fide, any payments made under the cost sharing agreement will be considered as offsets to the arm's length price that should have been paid for the intangibles. While section 1.482-2(d)(4) limits adjustments by the Service in the context of cost sharing arrangements to an adjustment of contributions paid, this regulation presupposes that the arrangement is bona fide. If the arrangement is not bona fide, normal arm's length standards would apply, including the commensurate with income standard.

B. PRODUCTS COVERED

In section 936(h), product area research is defined generally by reference to the three-digit Standard Industrial Classification (SIC) codes, meaning that the section 936(h) cost sharing arrangement covers research and development costs over a very broad product area. ²²⁸ As described above, the legislative history to the 1986 Act contemplates that a section 482 cost sharing arrangement should cover all research and development costs within an "appropriate product area." The approach in section 936 and in the 1986 Act legislative history contrasts to the proposed 1966 regulations. Cost sharing arrangements described in the 1966 proposed regulations could cover a single project, although multi-project or product area cost sharing agreements were not prohibited.

As discussed in section C below, broad product area cost sharing arrangements raise the issue of whether the potential benefits are proportionate to the participants' cost sharing payments. This issue is of particular concern in cost sharing arrangements of foreign- owned multinational groups if U.S. persons are participants, since cost sharing payments made by U.S. participants are deductible for U.S. tax purposes. ²²⁹

On the other hand, single product arrangements present the potential that cost sharing may be employed solely for high profit potential intangibles, such that foreign affiliates of U.S. multinational groups acquire the rights to such intangibles without bearing the cost of research related to low profit potential intangibles and unsuccessful research. The incentive to include selectively only high profit potential intangibles in a cost sharing arrangement is most acute when tax haven entities are the primary or predominate participants in the arrangement. ²³⁰

Three-digit SIC code product areas would seem to be the appropriate scope of most cost sharing arrangements. Both the Service or the taxpayer should be permitted to demonstrate, however, that a narrower or broader agreement is more appropriate. Taxpayers choosing a narrower agreement would need to show that the agreement is not merely an attempt to shift profits from successful research areas while leaving expenses of unsuccessful or less successful areas to be absorbed by the U.S. or higher tax affiliates. For example, some members of a multinational food and beverage group might be interested in research and development to develop a multi-purpose artificial sweetener, yet their respective food and beverage product lines might be sufficiently diverse (or might be products for which research and development is not necessary) that a single product agreement would be appropriate. Taxpayers choosing a broader agreement would need to show that the agreement is not being used to charge U.S. affiliates or other participants for research and development without reasonable prospect of benefit. From the Service's perspective, a product area that is broader or narrower than three-digit SIC codes may be necessary to avoid these distortions.

C. COST SHARES AND BENEFITS

Underlying all of the problems discussed in the legislative history of the 1986 Act in relation to cost sharing arrangements is the fundamental principle that the costs borne by each of the participants should be proportionate to the reasonably anticipated benefits to be received over time by each participant from exploiting intangibles developed under the cost sharing arrangement. This cost share/benefit principle has several facets, including the appropriate product area to be covered (discussed in section B above), definition of costs to be covered (discussed in section D below), and the measurement of anticipated benefits and several other issues discussed below.

1. ASSIGNMENT OF EXCLUSIVE GEOGRAPHIC RIGHTS. In general, the computation of cost shares should reflect a good faith effort to measure reasonably anticipated benefits to be derived from the arrangement. While it is difficult under the best of circumstances to predict what benefits each of the participants will derive, it is virtually impossible to do so unless the participants are assigned specific exclusive geographic rights to intangibles developed under the arrangement. Specific assignment of rights could take the form of assigning the rights to manufacturing intangibles relating to products to be sold in the United States to a U.S. affiliate, rights related to European markets to an Irish affiliate, rights related to Middle Eastern and Pacific rim markets to a Singapore affiliate, etc. In such case, the U.S. affiliate would derive the income attributable to the manufacturing intangibles developed under the arrangement with respect to any sales in U.S. markets regardless of whether ultimately the U.S. affiliate is manufacturing the products sold in U.S. markets. (As discussed below, however, the participants must be those expected to exploit the intangibles by performing the manufacturing function themselves.)

Alternatively, exclusive worldwide rights to different types of intangibles developed under the arrangement could be assigned to particular participants. This latter type of arrangement would warrant special scrutiny to assure that the cost shares reflect reasonably anticipated benefits. ²³¹ Moreover, if research activities are not common to the various types of intangibles produced under the arrangement, then the research related to each type of intangible should be charged to the specific affiliate that will receive the rights to that type of intangible. This is particularly true of arrangements where one of the parties produces components. The Service and Treasury would welcome comments on this topic. In short, such research activities are not the proper subject of a cost sharing arrangement.

For various reasons, including consistency with longstanding section 367(a) policy, U.S. geographic rights should never be permitted to be assigned under a cost sharing arrangement to a foreign person if either: (1) the participants are part of a U.S. owned multinational group; (2) a significant portion of the research is performed in the United States; or (3) any U.S. person participates in the arrangement. Accordingly, U.S. rights could be acquired by a foreign person only in the case of a foreign-owned multinational group that conducts the research overseas and does not include any U.S. affiliates as a participant in the arrangement.

2. OVERLY BROAD AGREEMENTS. The principle that cost shares be proportionate over time to the reasonably anticipated benefits may affect the issue of whether cost sharing arrangements are overly broad in terms of the products covered or the affiliates participating in the agreement. For instance, a manufacturing conglomerate makes widgets and gadgets. An overall cost sharing agreement for research and development may be inappropriate if a particular affiliate does not make both widgets and gadgets. If a disproportionate amount of research and development relates to widgets but affiliate A manufactures only gadgets, affiliate A would be subsidizing the research for the widget manufacturers. Although every participant in a cost sharing agreement should not be required to benefit from every intangible that may be produced, cost shares should be proportionate to the reasonably anticipated benefits. It may be necessary, therefore, either to have separate cost sharing agreements for widget and gadget research, to adjust affiliate A's cost share to reflect the costs related to gadget research, or to exclude affiliate A from the cost sharing arrangement. ²³²

3. DIRECT EXPLOITATION OF INTANGIBLES BY PARTICIPANTS. The cost share/benefit principle may otherwise affect who may participate in a cost sharing arrangement. In general, the benefit to be derived under a cost sharing arrangement is the right to use a developed intangible in the manufacture of a product. Therefore, the participant must be in a position to exploit the intangible in the manufacture of products. ²³³ It is not necessary that all participants be capable of manufacture at the time costs are being incurred, so long as it is reasonably anticipated that the participants will be capable of manufacture once the intangibles are developed and will use intangibles developed under the arrangement in the manufacture of products. ²³⁴

4. MEASUREMENT OF ANTICIPATED BENEFITS. In order to determine whether cost shares are proportionate to reasonably anticipated benefits, it is necessary to measure the anticipated benefits. Obviously there has to be some prediction, rough though it may be, of the kinds of intangibles likely to be produced and the relative proportion of units that will be produced and sold under the rights of each participant. In many cases, estimated units of production may be an appropriate measure of benefit assuming that there is a uniform unit of production that can be used as a measuring device. If there is no uniform unit of production, then sales value may be an appropriate measure, if measured at the same level of production or distribution for all participants. As stated in section A above, however, the Conference Report anticipated that cost shares be proportionate to profit as determined before research and development. Given the legislative history, therefore, neither units of production nor sales would be an appropriate measure if it were apparent (or became apparent during the course of the agreement) that profitability differed substantially with respect to various participants' rights. This would be true, for example, in situations in which geographic markets differ significantly in terms of production costs, market barriers, or other factors that bear significantly on profitability. In such cases, estimated gross profit or net profit may have to be used, or some adjustments may have to be made to cost shares determined on the basis of units of production or sales.

It is not realistic, however, to expect taxpayers in most instances to be able to estimate gross or net profit margins from estimated sales. Even estimates of units produced or sales value probably would be imprecise. It may be that a cost sharing agreement should not be recognized if units of production or sales are not appropriate measures and gross or net margins are extremely difficult to estimate. In such cases, the relationship between cost shares and anticipated benefits may simply be too tenuous.

5. PERIODIC ADJUSTMENTS. The language in the legislative history that the results of cost sharing arrangements be consistent with changes made by the 1986 Act to royalty arrangements has one other obvious implication. The cost shares should be adjusted periodically, on a prospective basis, to reflect changes in the estimates of relative benefits, including a change in the measurement standard if that becomes appropriate. In any event there is always a risk that the cost sharing agreement could subsequently be rejected as a bona fide arrangement if the estimates of benefits derived under the arrangement proved to be so substantially disproportionate to the cost shares that the estimates of benefits cannot be considered to have been made in good faith. Periodic adjustments to the cost sharing arrangement would reduce that risk.

D. COSTS TO BE SHARED

In general, the costs to be shared should include all direct and indirect costs of the research and development undertaken as part of the arrangement. Direct costs would include expenses for salaries, research materials, and facilities. However, there should be a limitation on the annual inclusion of costs for depreciable assets that is consistent with U.S. tax accounting principles. Otherwise, deductions for outbound payments may be overstated. Indirect costs should include a portion of overall corporate management expense and overall interest expense that is allocated and apportioned to research and development activities in a manner consistent with U.S. expense allocation principles. The costs to be reimbursed should be net of any charges for research undertaken on specific request or for any government subsidies granted. ²³⁵

E. BUY-IN REQUIREMENTS

As previously stated, the legislative history to the 1986 Act states that a party to a cost sharing arrangement that has contributed funds or incurred risks for development of intangibles at an earlier stage must be appropriately compensated by the other participants -- hence the requirement for a buy-in payment. One of the primary reasons for adopting cost sharing provisions is to avoid the necessity of valuing intangibles. Yet, if there are intangibles that are not fully developed that relate to the research to be conducted under the cost sharing arrangement, it is necessary to value them in order to determine an appropriate buy-in payment.

There are three basic types of intangibles subject to the buy-in requirement. A participant may own preexisting intangibles at various stages of development that will become subject to the arrangement. A company may also conduct basic research not associated with any product. Finally, there may be a going concern value associated with a participant's research facilities and capabilities that will be utilized.

Fully developed intangibles command a royalty to the extent used by other participants and are generally not appropriately incorporated into a cost sharing arrangement. Thus, royalties for preexisting developed intangibles may not be included in the buy-in payment, but instead are subject to the general rules of the commensurate with income standard. Because a subsequent substantial deviation in the income stream from the intangible might require an adjustment, it is important to identify separately the income stream and royalties related to preexisting developed intangibles. In many situations, research is performed with respect to preexisting intangibles in order to improve the preexisting intangibles (improved software, for example) or to develop the next generation of intangibles. The requirement for an adjustment to the royalty paid for the intangible would not apply if the intangible is enhanced in value solely as a result of research undertaken after inception of the cost sharing arrangement.

The buy-in payment should reflect the full fair market value of all intangibles utilized in the arrangement and not merely costs incurred to date. To permit a buy-in based on cost would be inconsistent with the provisions of section 367(d) which effectively precluded the tax-free transfer of intangibles and, by implication, the transfer of intangibles at cost. ²³⁶ The buy-in payment could take the form of lump sum or periodic payments spread over the average life expectancy of contributed intangibles -- perhaps on a declining basis since intangibles generally have greater value in the earlier stages of their life cycle. Obviously, periodic payments should reflect the time value benefit of not making a lump sum payment at the inception of the agreement.

A "buy-out" occurs when a participant withdraws from a cost sharing arrangement. Having funded a portion of research and development prior to withdrawal, that person owns a share of whatever the agreement has borne to date and must be compensated by the other participants for the value of what the arrangement has produced to date and not merely reimbursed for costs incurred to date.

A "secondary buy-in" is required when new members are admitted after a cost sharing agreement is in place. If a new member is acquiring a portion of the geographic rights of one of the original participants, any buy-in amount should be paid to the affiliate whose geographic rights are being reduced. Once again, in order for the buy-in to be arm's length, any new member must compensate the original participants in a manner similar to the original buy-in computation, but based upon current values and not merely costs incurred to date.

F. MARKETING INTANGIBLES

The 1986 amendment to section 482 provides that the term "intangible property" shall have the same meaning as in section 936(h)(3)(B). The section 936 definition of intangible property includes marketing intangibles. This does not mean, however, that marketing intangibles are necessarily the proper subject of cost sharing rules developed with manufacturing intangibles in mind.

In general, marketing costs yield a present benefit (even if also a future benefit) and, therefore, are currently deductible expenses for which a charge must be made under the services provisions of the section 482 regulations if the benefit therefrom accrues to a related person. ²³⁷ (Research expenses related to manufacturing intangibles generally do not yield present benefits but, nevertheless, are currently deductible pursuant to the special provisions of section 174.) In the context of marketing expenses, the services regulations under section 482 presently serve the same function as would rules governing cost sharing arrangements in identifying the potential beneficiaries of a marketing expense and requiring an appropriate charge (albeit with a profit element in certain cases). There seems to be no need for an additional regime to deal with the tax treatment of cost sharing arrangements related to marketing expenses. Comment is requested, however, concerning any situations which are believed not to be covered by the section 482 services regulations or any perceived problems which arise under those regulations as they affect marketing expenses.

G. CHARACTER OF COST SHARING PAYMENTS

Under section 936(h), the amount of any required cost sharing payment is not treated as income of the recipient, but instead reduces the amount of deductions otherwise allowable. ²³⁸ More generally, when expenditures are made with the expectation of reimbursement, they are treated as loans, and therefore the reimbursement does not constitute gross income to the recipient. ²³⁹ Accordingly, cost sharing payments are not income to the recipient but, instead, reduce costs which are otherwise deductible in computing taxable income and earnings and profits. Since cost sharing payments are not gross income to the recipient, no U.S. withholding tax would be imposed on outbound cost sharing payments made by a U.S. person to a foreign person.

Characterizing cost sharing payments in this manner also reduces the amount of research and development expenses of the entity performing the research that are subject to allocation under the rules of section 1.861-8 and increases the amount subject to allocation by the participants making the cost sharing payments. ²⁴⁰ Furthermore, since the payments received by the entity performing the research will not constitute income, payments received by a U.S. entity from foreign affiliates are not foreign source income to the U.S. entity.

For purposes of calculating the credit allowable under section 41 for research expenditures, members of a commonly controlled group of corporations may disregard intercompany reimbursements for research expenditures. ²⁴¹ This rule treats a U.S. company that actually performs U.S. situs research as incurring 100 percent of the research expenses for purposes of calculating the research credit, even if the U.S. company is reimbursed for a portion of those expenses pursuant to a section 482 cost sharing arrangement.

One group of taxpayers has suggested that the regulations should allow a cost sharer to denominate its rights under a cost sharing arrangement as a geographically exclusive, no-royalty, perpetual license if a license is required to obtain local country tax benefits or if the parent would be in a better position to protect against infringement than the subsidiary. If under U.S. law, the participant is clearly the beneficial owner of intangibles developed under the cost sharing arrangement, then labeling its interest as a "license" will not change the characterization for U.S. tax purposes, even if legal title to the rights are held by its parent serving as nominee owner of such rights. Therefore, whatever label is applied to the arrangement for foreign law purposes generally would not affect its U.S. tax treatment unless the label affects substantive legal rights relating to the intangible.

H. POSSESSIONS CORPORATIONS

The cost sharing payment made by a possessions corporation pursuant to the special cost sharing election under section 936(h)(5)(I) must be determined under those rules and not under a contractual cost sharing arrangement that would otherwise govern the charges incurred by the participants. Indeed, the statute and regulations explicitly provide that the section 936(h) cost sharing payment shall not be reduced by a contractual cost sharing payment. ²⁴² Under section 936(h), the cost sharing payment by the possessions corporation must equal the greater of the amount required under the new commensurate with income standard or 110% of the pre- 1986 Act statutory cost sharing amount. Under the commensurate with income standard, the cost sharing amount must at least equal the fair market royalty which would have to be paid to the developer if the manufacturing intangibles had been licensed (even in cases in which the intangible had previously been transferred in a section 351 exchange).

The amount paid under section 936(h) entitles the possessions corporation to be treated as the owner of manufacturing intangibles previously developed by its U.S. affiliates. The fact that a possessions corporation has entered into a cost sharing arrangement for the development of future intangibles and is paying a lesser amount under that arrangement does not affect the amount required under the section 936(h) cost sharing election. Indeed, since the section 936(h) cost sharing payment is compensation for intangibles previously developed and the section 482 cost sharing payment made pursuant to the contractual cost sharing agreement is for the cost of developing new intangibles, both amounts appropriately must be paid initially -- one by statutory election and the second pursuant to the contractual arrangement. It might be argued that, once intangibles are developed under the section 482 cost sharing arrangement, the possessions corporation's section 936 cost sharing payment should be reduced so that the possessions corporation does not pay a second time for that intangible. The statute, however, precludes that result.

I. ADMINISTRATIVE REQUIREMENTS

Taxpayers seeking to enter into cost sharing arrangements should be required to make a formal election and to document the specifics of the agreement contemporaneously. Any U.S. participant should be required to include a copy of the agreement with its first return filed subsequent to the agreement's effective date. Taxpayers making the election would agree to produce, in English and in the United States, the records of foreign participants necessary to verify the computation and appropriateness of the respective cost shares within 60 days of a request by the Service. These records would include identification of the SIC code or other basis used to determine products covered by the agreement, and summary information concerning sales, gross margins, and net income derived with respect to such products. The House Report accompanying the 1986 Act suggests that the Service should require similar records to be kept and produced under the authority of section 6001 for section 936(h) cost sharing agreements. ²⁴³

J. TRANSITIONAL ISSUES FOR EXISTING COST SHARING AGREEMENTS

It is unlikely that there will be preexisting cost sharing agreements that will meet all of the standards described above. If such agreements are not recognized, the Service and taxpayers will encounter significant problems in determining ownership of preexisting intangibles and the treatment of the payments that have been made pursuant to the preexisting agreements. Some type of grandfather treatment would therefore appear to be appropriate. One possibility would be to permit any cost sharing agreement that conforms to the requirements of the existing regulations, and that has been in existence for more than 5 years prior to 1987, to be recognized fully if conformed within a certain period after the promulgation of the new rules with respect to matters other than the buy-ins that occurred prior to June 6, 1984 (the effective date of section 367(d)). If the cost sharing agreement has been in effect for less than 5 years and the agreement does not conform substantially to the new rules, then the old agreement would not be recognized. If a new agreement that conforms to the new rules is adopted, then all payments pursuant to the old agreement would be taken into account as an adjustment to any required buy-in payments relating to the new agreement. ²⁴⁴

K. CONCLUSIONS AND RECOMMENDATIONS

1. Congress intended to permit cost sharing arrangements if they produce results consistent with the commensurate with income standard in section 482.

2. Three-digit SIC code product areas seem to be the appropriate scope for most cost sharing arrangements. Both the Service and the taxpayer should be permitted to demonstrate, however, that a narrower or broader agreement is more appropriate.

3. The fundamental principle underlying the concerns identified in the legislative history of the 1986 Act with respect to cost sharing is that costs borne by each of the participants must be proportionate to the reasonably anticipated benefits to be received over time by the participants from exploiting intangibles developed under a cost sharing arrangement. This principle has several implications.

a. In order for taxpayers to make a good faith effort to predict anticipated benefits, it is essential that the participants be assigned specific and exclusive geographic rights to intangibles developed under the arrangement. U.S. geographic rights generally should not be permitted to be assigned to a foreign person.

b. Cost sharing arrangements may be overly broad in terms of products covered or affiliates participating in the agreement if some participants would utilize only a narrow range of intangibles developed under the agreement.

c. Since the benefit to be derived under a cost sharing arrangement is the right to use developed intangibles in the manufacture of a product, participants must generally be capable of manufacturing and using developed intangibles in the manufacture of products.

d. In estimating anticipated benefits, units of production or sales value generally would be an acceptable unit of measure unless profitability would reasonably be expected to differ significantly with respect to various participants' rights. In the latter instance, some adjustments must be made, or some other standard of measurement utilized, to reflect more accurately the reasonably anticipated benefits to be derived by the participants.

e. Cost shares should be adjusted periodically, on a prospective basis, to reflect changes in the estimates of relative benefits, including a change in the measurement standard if that becomes necessary.

4. The costs to be shared should include all direct and indirect costs determined in a manner consistent with U.S. tax accounting and expense allocation principles.

5. A party that contributes funds or incurs risks for development of intangibles at an earlier stage must be appropriately compensated by the other participants in the form of a buy-in for the value of preexisting intangibles (including basic research and the going concern value of research capability).

a. Fully developed intangibles command a royalty and should not be incorporated into a cost sharing agreement, with the result that the buy-in may not reflect compensation for fully developed intangibles.

b. A secondary buy-in is required whenever a participant withdraws from a cost sharing arrangement or a new participant enters the arrangement.

6. Expenses relating to marketing intangibles are presently governed by the services provisions of the section 482 regulations. There seems to be no need for marketing expenses to be subject to a cost sharing regime developed for manufacturing intangibles.

7. Cost sharing payments are not income to the recipient but, instead, reduce costs that are otherwise deductible for purposes of computing taxable income and earnings and profits. Consequently, outbound cost sharing payments are not subject to U.S. withholding tax, and inbound payments are not foreign source income.

8. Since a section 936(h) cost sharing payment is compensation for intangibles previously developed and a section 482 cost sharing payment is for the cost of developing new intangibles, both amounts appropriately must be paid initially if a possessions corporation making the section 936(h) cost sharing election enters into a section 482 cost sharing arrangement. Under the statute, the section 936(h) payment may in no event be reduced to reflect amounts paid under a section 482 cost sharing agreement.

9. Taxpayers should be required to make a formal election if they enter into a cost sharing arrangement, to file a copy of the agreement with their return, and produce records necessary to verify the computation and appropriateness of the respective cost shares.

10. Cost sharing agreements in existence for more than five years prior to 1987 should be grandfathered if they conform in certain respects with new rules to be promulgated. Other agreements will not be bona fide unless and until they substantially conform to the new rules.

² H.R. Conf. Rep. No. 841, 99th Cong., 2d Sess. II-638 (1986) (hereinafter 1986 Conf. Rep.].

³ Regulation 41, Articles 77-78, War Revenue Act of 1917, ch. 63, 40 Stat. 300 (1917).

⁴ Rev. Act of 1921, ch. 136, section 240(d), 42 Stat. 260 (1921).

⁵ S. Rep. No. 275, 67th Cong., 1st Sess. 20 (1921).

⁶ Rev. Act of 1928, ch. 852, section 45, 45 Stat. 806 (1928).

⁷ H.R. Rep. No. 2, 70th Cong., lst Sess. 16-17 (1928).

⁸ Treas. Reg. 86, section 45-1(b) (1935).

⁹ Id.

¹⁰ Asiatic Petroleum Co. v. Comm'r, 79 F.2d 234, 236 (2d Cir.), cert. denied, 296 U.S. 645 (1935) (concept of evasion for this purpose includes civil tax avoidance).

¹¹ Central Cuba Sugar Co. v. Comm'r, 198 F.2d 214, 215 (2d Cir.), cert. denied, 344 U.S. 874 (1952) (application of clear reflection standard does not require proof of tax avoidance motive).

¹² National Securities Corp. v. Comm'r, 137 F.2d 600 (3d Cir.), cert. denied, 320 U.S. 794 (1943).

¹³ Central Cuba Sugar, supra n. 11; Rooney v. U.S, 305 F.2d 681 (9th Cir. 1962).

¹⁴ Friedlander Corp. v. Comm'r, 25 T.C. 70, 77 (1955).

¹⁵ Polack's Frutal Works v. Comm'r, 21 T.C. 953, 975 (1954).

¹⁶ Grenada Industries v. Comm'r, 17 T.C. 231 (1951), aff'd, 202 F.2d 873 (5th Cir.), cert. denied, 346 U.S. 819 (1953).

¹⁷ Seminole Flavor Co. v. Comm'r, 4 T.C. 1215 (1945); cf. Moline Properties v. Comm'r, 319 U.S. 436 (1943). In extreme cases of income shifting, other legal theories such as assignment of income, substance over form, disregard of corporate entity, or treatment of corporate entity as an agent have been used by courts to attribute income to the appropriate person or corporate entity. These theories are beyond the scope of this paper, and they are generally not used by courts when section 482 is also applicable. See, e.g., Hospital Corporation of America v. Comm'r, 81 T.C. 520 (1983) (foreign affiliate not treated as a sham; section 482 applied for use of U.S. parent's intangibles).

¹⁸ Hamburgers York Road Inc. v. Comm'r, 41 T.C. 821 (1964); Aiken Drive-In Theatre Corp. v. U.S., 281 F.2d 7 (4th Cir. 1960) (the shifting of an abandonment loss from one corporation to another created an inaccurate picture of income, justifying use of section 482).

¹⁹ Hearings on the President's 1961 Tax Recommendations Before the Committee on Ways and Means, 87th Cong., 1st Sess., vol. 4, at 3549 (1961) (statement of M. Caplin, Commissioner of Internal Revenue, "Problems in the Administration of the Revenue Laws relating to the Taxation of Foreign Income").

²⁰ H.R. Rep. No. 1447, 87th Cong., 2d Sess. 28 (1962).

²¹ Id.

²² H.R. 10650 82d Cong. 2d Sess., section 6 (1962).

²³ Id.; see H.R. Rep. No. 1447, 87th Cong., 2d Sess. 537-38 (1962).

²⁴ See, e.g., Hearings on H.R. 10650 Before the Senate Committee on Finance, 87th Cong., 2d Sess., pt. 7, at 2913, 3011-3012 (1962) (statements by P. Seghers and D. N. Adams).

²⁵ H.R. Rep. No. 2508, 87th Cong., 2d Sess. 18-19 (1962).

²⁶ Proposed regulations were issued in 1965, were withdrawn and reproposed in 1966, and were issued in final form in 1968. Proposed Treas. Reg. sections 1.482-1(d) and 2, 30 Fed. Reg. 4256 (1965); Proposed Treas. Reg. sections 1.482-1(d) and 2, 31 Fed. Reg. 10394 (1966); and T.D. 6952, 1968-1 C.B. 218.

²⁷ In addition to the tangible and intangible property and the services regulations, there are safe harbors and other rules for interest rates on related party loans, Treas. Reg. section 1.482- 2(a), and rules similar to the services rules for related party leasing transactions, Treas. Reg. section 1.482-2(c). These rules are generally not discussed in this paper.

²⁸ Treas. Reg. section 1.482-2(d)(2)(iii).

²⁹ Treas. Reg. section 1.482-2(e)(1)(iii).

³⁰ Surrey, Treasury's Need to Curb Tax Avoidance in Foreign Business through the Use of Section 482, 28 J. Tax'n 75 (1968).

³¹ IEs and Group Managers were requested to complete one questionnaire for each of the three cases they considered to be their most important section 482 cases. In some instances respondents had not had experience with three important section 482 cases, so that fewer responses were made to some questions. In others, some respondents answered based on their general experience, rather than the particular case for which the questionnaire was completed. In many instances the categorization of particular issues entails a great deal of judgment. For example, a case such as Hospital Corporation of America, supra n. 17, could be viewed as a services case, an allocation of income case, a profit split case, or an intangibles case. For these reasons we have used the results of the questionnaire throughout this paper primarily for purposes of illustration. However, the results, where used, represent and correspond with the experiences of persons interviewed and others in the Service responsible for administering section 482.

³² Many of the requests for information that are not honored concern transactions with third parties that would provide comparables for analyzing a potential section 482 adjustment. Appendix B, infra, at Question 18C.

³³ Appendix B, infra, at Question 19.

³⁴ Appendix B, infra, at Question 13, and Appendix A, infra, at 4-5.

³⁵ Appendix B, infra, at Question 14.

³⁶ General Accounting Office, Report to the Chairman, Committee on Ways and Means, IRS Audit Coverage: Selection Procedures Same for Foreign and other U.S. Corporations 26-29 (1986) [hereinafter GAO, IRS Audit Coverage].

³⁷ Under section 982, a taxpayer which, without reasonable cause, fails to produce, within 90 days, foreign based documents sought by the agent during the course of the examination through the use of a formal document request may be precluded from introducing the documents sought in a subsequent court proceeding. A special court proceeding is established at which the taxpayer may show reasonable cause for failing to produce the requested documents or otherwise move to quash the formal document request.

³⁸ Vetco v. United States, 644 F.2d 1324 (9th Cir. 1981), cert. denied, 454 U.S. 1098 (1982) (summons for books and records of Swiss controlled foreign corporation enforced notwithstanding potential violations of the Swiss Penal Code).

³⁹ Foreign direct investment in the United States increased from about $34.6 billion in 1977 to about $100.50 billion in 1984. GAO IRS Audit Coverage, supra n. 36, at 10.

⁴⁰ Hobb, Foreign Investment and Activity in the United States through Corporations, 1983, SOI Bulletin 53-68 (Summer 1987); see BNA Daily Tax Report, April 1, 1987, at G2.

⁴¹ Wheeler, SEC Requires Less Disclosure from Foreign Corporations, Tax Notes, October 12, 1987, at 195-197.

⁴² Section 7602; United States v. Toyota Motor Corp., 561 F. Supp. 348 (C.D. Cal. 1983); United States v. Toyota Motor Corp., 569 F. Supp. 1158 (C.D. Cal. 1983).

⁴³ Section 982.

⁴⁴ In the survey conducted as part of this study, IEs reported using summonses and section 982 requests in approximately 5% and 4%, respectively, of the cases reported in the survey. Appendix B, infra, at Questions 21, 22.

⁴⁵ An exception in Treas. Reg. section 1.482-2(b)(7) requires adequate records to verify costs or deductions used in connection with a charge for services to an affiliate.

⁴⁶ Appendix B, infra, at Question 57.

⁴⁷ H.R. Rep. No. 426, 99th Cong., 1st Sess. 424 (1985) [hereinafter 1985 House Rep.]. The survey revealed that in approximately 41% of the cases in which taxpayers relied upon comparables, industry averages were used. Appendix A, infra.

⁴⁸ See discussion infra Chapter 11.

⁴⁹ Under Rev. Proc. 88-37, 1988-30 I.R.B. 31, a taxpayer that reports intercompany transactions, on Schedules G and M of Form 5471, may avoid the substantial understatement penalty. See Rev. Proc. 85-26, 1985-1 C.B. 580 (amended returns or statements made following commencement of a CEP examination may avoid assertion of the substantial understatement penalty).

⁵⁰ Commissioner's Penalty Study, A Philosophy of Civil Tax Penalties (discussion draft June 8, 1988).

⁵¹ See discussion of the commensurate with income standard and periodic adjustments infra Chapters 6 and 8.

⁵² In the survey conducted for the study, an adjustment was made under Treas. Reg. section 1.482-2(d) in about 50% of the reported cases. Appendix B, infra, at Question 68.

⁵³ In approximately 40% of the cases reported in the survey, IEs cited the inability to value an intangible as the reason why they failed to follow the intangibles section of the regulations. Appendix B, infra, at Question 72.

⁵⁴Rev. Rul. 75-254, 1975-1 C.B. 243.

⁵⁵ Treas. Reg. section 1.482-2(d)(3).

⁵⁶ See, e.g., Treas. Reg. sections 1.482-2(e)(2)(ii) and example (2), 1.482-2(e)(3)(ii) example (2), and 1.482- 2(e)(4)(iii)(c).

⁵⁷ Eli Lilly & Co. v. Comm'r, 84 T.C. 996 (1985), rev'd in part, aff'd in part and remanded, Nos. 86-2911 and 86-3116 (7th Cir. August 31, 1988) [Lilly]. See the discussion of Lilly, infra, Chapters 4 and 5.

⁵⁸ In certain circumstances, no separate allocation is required for services performed in connection with the transfer of intangible property. Treas. Reg. section 1.482-2(b)(8). Services are rendered in connection with the transfer of intangible property if they are merely ancillary or subsidiary to the transfer of the intangible property. The regulations give as an example of ancillary services start-up help given to a related entity in order for it to integrate a trade secret manufacturing process into its operations. The regulations then state that, should the transferor continue to render services after the process has been integrated into the manufacturing process, a separate allocation for services would be required under the regulations. The experience of the IEs is that the current regulations fail to give them specific guidance on how to determine when services rendered in connection with the transfer of an intangible require a separate allocation.

Appendix D discusses results of a preliminary survey of data available at the SEC. This data has the potential to determine when unrelated parties would extract a specific charge for services rendered in connection with the transfer of an intangible.

⁵⁹ In Lilly, supra n. 57, the Tax Court ultimately determined the partner company's marketing return based upon using its "best judgment." Lilly, 84 T.C. at 1167; See also G.D. Searle and Co. v. Comm'r [Searle], 88 T.C. 252, 376 (1987).

⁶⁰ Treas. Reg. sections 1.482-2(e)(1)(ii) and (iii).

⁶¹ Treasury Department News Release, Summary Study of International Cases Involving Section 482 of the Internal Revenue Code (Jan. 8, 1973), reprinted in 1973 Standard Federal Tax Report (CCH) par. 6419.

⁶² Tax Allocations and International Business: Corporate Experience with Section 482 of the Internal Revenue Code, Conference Board Report No. 555 (1972).

⁶³ Burns, How IRS Applies the Intercompany Pricing Rules of Section 482: A Corporate Survey, 54 J. Tax'n 308 (1980).

⁶⁴ General Accounting Office, Report by the Comptroller General to the Chairman, House Committee on Ways and Means, IRS Could Better Protect U.S. Tax Interests in Determining the Income of Multinational Corporations (1981) [hereinafter GAO, IRS Could Better Protect U.S. Tax Interests].

⁶⁵ IRS Publication No. 1243, IRS Examination Data Reveal an Effective Administration of Section 482 Regulations (1984).

⁶⁶ As stated earlier, the percentages from the 1987 survey do not represent a scientifically valid random sample. They are based upon responses to a questionnaire sent to selected groups of International Examiners who responded with respect to a small number of cases selected by them. Compared to the 1984 survey undertaken by the Assistant Commissioner (Examination), however, they suggest one significant trend: a substantial increase in the use of the cost plus method with a corresponding decrease in cases classified as either "comparable uncontrolled price" or "other." Such a trend would probably be due to an emphasis during the last several years on examining cases that involved manufacturing activities in tax haven jurisdictions. See GAO, IRS Audit Coverage, supra n. 36 at 26-29.

⁶⁷ Appendix B, infra, at Questions 62-64.

⁶⁸ In Lilly, supra n. 57, the IRS notice of deficiency was based upon the cost plus method while the taxpayer initially attempted to rely upon the resale price method. The Tax Court rejected application of the cost plus method and, also, the taxpayer's analysis under both the resale price and "fourth" methods. It ultimately adopted a profit split method for the first two years at issue and a CUP method for the final year. See discussion of Lilly infra Chapters 4 and 5.

⁶⁹ See discussion of this issue infra Chapter 11.

⁷⁰ I.R.M. section 4233(523.2). The Manual states that almost all cases can be analyzed using a functional analysis.

⁷¹ I.R.M. section 4233(524).

⁷² I.R.M. section 42(12)3.

⁷³ Appendix B, infra, at Question 32.

⁷⁷ Supra n. 57.

⁷⁸ For 1973, the Tax Court was able to use a comparable uncontrolled price approach because the Darvon patent had expired. However, numerous adjustments were made to reach a transfer price.

⁷⁹ See discussion of the Tax Court's profit split analysis infra Chapter 5.

⁸⁰ Supra n. 59.

⁸¹ Id at 375.

⁸² Id at 376.

⁸³ Supra n. 17.

⁸⁴ 85 T.C. 172 (1985).

⁸⁵ Id. at 225-26.

⁸⁶ 608 F.2d 445 (Ct. Cl. 1979).

⁸⁷ The facts in DuPont are similar to the abuse relating to the use of foreign base sales companies to defer the taxation of income in the United States that Congress sought to end through the Subpart F provisions enacted in the Revenue Act of 1962. H.R. Rep. No. 1447, 87th Cong., 2d Sess. 28 (1962).

⁸⁸ See discussion infra Chapter 5 regarding the income to costs ratio and return on capital methods used in DuPont.

⁸⁹ Supra n. 86.

⁹⁰ 617 F.2d 942 (2d Cir. 1980), rev'g T.C. Memo. 1977-140.

⁹¹ 617 F.2d at 951.

⁹² 55 T C. 928 (1970).

⁹³ 60 T.C. 569 (1973).

⁹⁴ 67 T.C. 224 (1976).

⁹⁵ T.C. Memo. 1985-578.

⁹⁶ 78-1 USTC para. 9374, at 83,910 (Ct. Cl. Trial Div. 1978).

⁹⁷ Supra n. 57. See discussion of Lilly supra Chapter 4.

⁹⁸ 84 T.C. at 1148-49. The results of these surveys and studies are referenced in Chapter 3, supra.

⁹⁹ Supra n. 92.

¹⁰⁰ T.C. Memo. 1971-101, rev'd, 468 F.2d 805 (5th Cir. 1972).

¹⁰¹ 84 T.C. at 1150-51.

¹⁰² "Location savings" were specifically authorized for certain Puerto Rican affiliates by Rev. Proc. 63-10, 1963-1 C.B. 490, 494. Location savings do not otherwise automatically accrue to an affiliate, but under the arm's length standard of section 482 are distributed as the marketplace would divide them.

¹⁰³ 84 T.C. at 1168, n. 102.

¹⁰⁴ 84 T.C. at 1167.

¹⁰⁵ See discussion supra Chapter 4.

¹⁰⁶ Supra n. 59. See discussion of Searle supra Chapter 4.

¹⁰⁷ 88 T.C. at 376.

¹⁰⁸ Supra n. 17. See discussion supra Chapter 4.

¹⁰⁹ 81 T.C. at 601.

¹¹⁰ Supra n. 86. See discussion supra Chapter 4.

¹¹¹ Id. at 456.

¹¹² Id.

¹¹³ 84 T.C. at 1157, 1161.

¹¹⁴ The court relied upon testimony by the Service's accounting expert, Dr. James Wheeler, to show that, if the taxpayer had transferred the rest of its successful products to Puerto Rico under terms similar to its transfer of Darvon, its return would have been insufficient to enable it to continue funding its R&D program, which the court characterized as the "life-blood" of a successful pharmaceutical company. 84 T.C. at 1160-1161. As noted previously in Chapter 4, supra, the Tax Court was reversed on this issue.

¹¹⁵ 84 T.C. at 1086-88, 1092-93. See Wheeler, An Academic Look at Transfer Pricing in a Global Economy, Tax Notes, July 4, 1988, at 91.

¹¹⁶ These assets must also have been recorded on Lilly's financial books of account; thus some intangible assets are not included. In a recent article, it was noted that Eli Lilly had a five year average return on shareholders equity of 23 percent (on an after-tax basis). Who's Where in Profitability, Forbes, January 11, 1988, at 216. Compare this consolidated return on assets with the return in excess of 100 percent earned by the Puerto Rican subsidiary during the years 1971-1973.

¹¹⁷ The adjusted taxable income for the subsidiary does not reflect the exclusion provided by section 931 of the Internal Revenue Code of 1954 and excludes interest income.

¹¹⁸ Supra n. 59.

¹¹⁹ Wheeler, supra n. 115, at 91.

¹²⁰ The Seventh Circuit in Lilly, supra n. 57, discounted this type of evidence because it called into question Lilly P.R.'s ownership of the intangibles at issue.

¹²¹ Supra n. 93. See discussion supra Chapter 4.

¹²² 66 T.C. 373 (1976), aff'd, 598 F.2d 1375 (5th Cir. 1979).

¹²³ Largely in response to the Brittingham case, Congress enacted section 1059A in 1986. This section generally forces an importer to use a value for income tax purposes no greater than the value declared for customs purposes.

¹²⁴ See J. Baranson, Technology and the Multinationals at 64 (1978).

¹²⁴ (e) Treatment of Certain Royalty Payments. --

(1) In General. -- Section 482 (relating to allocation of income and deductions among taxpayers) is amended by adding at the end thereof the following new sentence: "In the case of any transfer (or license) of intangible property (within the meaning of section 936(h)(3)(B)), the income with respect to such transfer or license shall be commensurate with the income attributable to the intangible."

(2) Technical Amendment. -- Subparagraph (A) of section 367(d)(2) (relating to transfers of intangibles treated as transfer pursuant to sale for contingent payments) is amended by adding at the end thereof the following new sentence: "The amounts taken into account under clause (ii) shall be commensurate with the income attributable to the intangible."

Sec. 1231(e)(1), Tax Reform Act of 1986, 100 Stat. 2085 (1986).

¹²⁵ For this purpose, intangibles are broadly defined by reference to section 936(h)(3)(B) under which intangible property includes any:

(i) patent, invention, formula, process, design, pattern, or know-how;

(ii) copyright, literary, musical, or artistic composition;

(iii) trademark, trade name, or brand name;

(iv) franchise, license, or contract;

(v) method, program, system, procedure, campaign, survey, study, forecast, estimate, customer list, or technical data; or

(vi) any similar item,

which has substantial value independent of the services of any individual. See also Treas. Reg. section 1.482-2(d)(3)(ii) and Rev. Rul. 64-56, 1964-1 C.B. 113, regarding the treatment of know-how as property in a section 351 transfer.

¹²⁶ 1985 House Rep. supra n. 47, at 420-427; 1986 Conf. Rep., supra n. 2, at II-637-638. Several commentators have suggested that the phrase "commensurate with income" derives from Nestle Co. Inc. v. Comm'r, T.C. Memo. 1963-14, where the Tax Court sanctioned a taxpayer's post-agreement increase in royalties paid by an affiliate for a very profitable intangible license. The opinion states that "[s]o long as the amount of the royalty paid was COMMENSURATE WITH THE VALUE OF THE BENEFITS RECEIVED AND WAS REASONABLE, we would not be inclined to, nor do we think we would be justified to, conclude that the increased royalty was something other than what it purported to be." (Emphasis supplied). There is, however, nothing in the legislative record to indicate that this is the case or to indicate Congressional approval or disapproval of the result in Nestle.

¹²⁷ 1985 House Rep., supra n. 47, at 426.

¹²⁸ 1986 Conf. Rep., supra n. 2, at II-637.

¹²⁹ 1985 House Rep., supra n. 47, at 424.

¹³⁰ Id. at 425-426.

¹³¹ Id. at 426.

¹³² International Fiscal Association, Cahiers de Droit Fiscal International (Studies on International Fiscal Law), Vol. LVI, at I-6 (1971).

¹³³ So long as the foreign affiliate ultimately pays out its residual earnings as a dividend and exhausts its remedies for obtaining an adjustment in the foreign jurisdiction, the total amount of foreign source income on the U.S. return in the relevant limitation category (and, therefore, the amount of limitation under section 904) will be the same no matter what the amount of royalty, and the taxes paid by the foreign affiliate will be deemed paid by the U.S. parent. The operation of the foreign tax credit will thus prevent any double taxation on those earnings irrespective of the amount of the royalty payment for U.S. tax purposes.

¹³⁴ Wright & Clowery, The Super-Royalty: A Suggested Regulatory Approach, Tax Notes, July 27, 1987, at 429-436.

¹³⁵ 1986 Conf. Rep., supra n. 2, at II-637.

¹³⁶ See discussion of the concepts of inexact and exact comparables infra Chapter 11.

¹³⁷ See discussion of the concepts of internal and external comparability infra Chapter 11.

¹³⁸ The term high profit potential intangibles refers to those products which generate profits far beyond the normal returns found in the industry. No specific definition or formula for determining whether an item is a high profit potential product is suggested herein. Nonetheless, hypothetical products such as an AIDS vaccine, a cure for the common cold, or a cheap substitute for gasoline would all fit into this concept because of the enormous consumer demand for such a product, the market protection provided by a patent, and the corresponding potential for enormous profitability. Similarly, a patented product that just happens to work better than others, or produces the same result with fewer side effects, may also qualify.

¹³⁹ The German tax authorities have faced a similar situation, and the imputation of very high royalty rates has led to the charge that the imputed royalties are not arm's length. See Jacob, The New "Super-Royalty" Provisions of Internal Revenue Code 1986: A German Perspective, 27 European Taxation 320 (1987).

¹⁴⁰ See infra Chapter 11.

¹⁴¹ 1985 House Rep., supra n. 47, at 425.

¹⁴² See discussion of the mechanism for making adjustments to lump sum payments infra Chapter 8.

¹⁴³ Staff of Joint Comm. on Taxation, General Explanation of the Revenue Provisions of the Deficit Reduction Act of 1984, 98th Cong., 2d Sess. 432-433 (1984) [hereinafter General Explanation of the DRA of 1984].

¹⁴⁴ The Service and Treasury invite comments as to whether this possible exception should be under a different standard than the concept of control under section 482.

¹⁴⁵ Treas. Reg. section 1.367(d)-1T(g)(4)(i).

¹⁴⁶ Id.

¹⁴⁷ Treas. Reg. section 1.367(d)-1T(g)(4)(ii).

¹⁴⁸ 1986 Conf. Rep. supra n. 2, at II-638.

¹⁴⁹ General Explanation of the DRA of 1984, supra n. 143, at 433.

¹⁵⁰ See discussion supra Chapter 6 regarding relief from double taxation pursuant to the foreign tax credit provisions and sourcing rules of United States internal law.

¹⁵¹ Letter from J. Roger Mentz, Assistant Secretary (Tax Policy) of the Department of Treasury to Representative Philip M. Crane (May 26, 1g87); Remarks of Stephen E. Shay, International Tax Counsel of the Department of Treasury before the International Fiscal Association (February 12, 1987). Appendix C to this study summarizes the legal and administrative approaches similar to those described throughout this study taken by some of our major treaty partners in dealing with transfer pricing issues.

¹⁵² Organization of Economic Cooperation and Development, Committee on Fiscal Affairs, Model Double Taxation Convention on Income and on Capital, Art. 9(1) ("Associated Enterprises") (1977) [hereinafter OECD Model Convention].

¹⁵³ Id. at Art. 9(2).

¹⁵⁴ Id.

¹⁵⁵ United Nations Model Double Taxation Convention Between Developed and Developing Countries, U.N. Doc. ST/ESA/102, at 27 (1980) [hereinafter U.N. Model Convention].

¹⁵⁶ U.S. Treasury Dept., Proposed Model Convention Between the United States of America and . . . . . for the Avoidance of Double Taxation and the Prevention of Fiscal Evasion with Respect to Taxes on Income and Capital (1981).

¹⁵⁷ The United States model adds a third paragraph to the OECD Model Convention Article 9 that reserves to each state the right to make adjustments under its internal law. The purpose of this paragraph is only to make explicit that the use of the word "profits" in the OECD model does not constrain either jurisdiction to make adjustments, consistent with the arm's length standard of paragraph 1, with respect to deductions, credits, or other allowances between related persons. This provision reads as follows:

3. The provisions of paragraph 1 shall not limit any provisions of the law of either Contracting State which permit the distribution, apportionment or allocation of income, deductions, credits, or allowances between persons, whether or not residents of a Contracting State, owned or controlled directly or indirectly by the same interests when necessary in order to prevent evasions of taxes or clearly to reflect the income of any of such persons. Id.

¹⁵⁸ U.N. Model Convention, supra n. 155, at 106; see generally Organization for Economic Cooperation and Development, Report of the Committee on Fiscal Affairs, Transfer Pricing and Multinational Enterprises (1979) [hereinafter OECD, Transfer Pricing and Multinational Enterprises].

¹⁵⁹ See e.g. Cross-Border Transactions Between Related Companies: A Summary of Tax Rules (W. R. Lawlor, ed. 1985) (discussion of transfer pricing practices of twenty-five different countries, most of which take the arm's length standard as their basic rule of transfer pricing).

¹⁶⁰ A recent article has suggested that the arm's length standard for transfer pricing should not limit the transfer pricing practices of governments. Langbein, The Unitary Method and the Myth of Arm's Length, Tax Notes, Feb. 17, 1986, at 625.

¹⁶¹ OECD, Transfer Pricing and Multinational Enterprises, supra n. 158.

¹⁶² Id. at 51.

¹⁶³ Id. at 14.

¹⁶⁴ Id. at 14-15.

¹⁶⁵ Id. at 19.

¹⁶⁶ Id at 15.

¹⁶⁷ Id. at 17.

¹⁶⁸ Id. at 42-43.

¹⁶⁹ Id. at 54.

¹⁷⁰ The objection raised in the report regarding the type of analysis advocated in this report is not that it violates the arm's length standard, but that it may call for more information than can be practically obtained and analyzed by the tax authorities. See id. at 15. As noted in Appendix D, the degree of detail and analysis that will be called for under the new methodology will depend in each case on the magnitude of the potential for income shifting. Further, in cases of transfers of routine intangibles, available comparable licenses will generally obviate the need for almost all of this information.

¹⁷¹ See the discussion infra Chapter 8.

¹⁷² 1985 House Rep., supra n. 47, at 425.

¹⁷³ 60 T.C. 836 (1973).

¹⁷⁴ Since related parties always have the ability to renegotiate contractual arrangements, explicit contractual provisions permitting renegotiation of related party arrangements would have little meaning and, therefore, should not be a prerequisite for making adjustments. Furthermore, related party contracts that contain these provisions will not necessarily lead to results that conform to the experience of unrelated parties operating under similar circumstances. If the contract proves more profitable than expected, the parties can refuse to renegotiate or adjust it, despite explicit provisions in the contract which permit or require them to do so. Thus, requiring that related party contracts mimic the terms of unrelated party contracts will not alone ensure that the results experienced by the related parties under those contracts will approximate arm's length dealing. 1985 House Rep., supra n. 47 at 425-426. Without the ability to make changes for adjustments over time, related party agreements will be observed when they suit the tax needs of the parties and amended or changed when they work to their detriment. Compare R.T. French Co. v. Comm'r, 60 T.C. 836 (1973), with Nestle Co. Inc. v. Comm'r, T.C. Memo. 1963-14.

¹⁷⁵ Periodic adjustments will also obviate the need for the often fruitless inquiry into the state of mind of the taxpayer and its affiliate at the outset.

¹⁷⁶ See the discussion infra Chapter 11, regarding the role of comparables in determining whether an adjustment over time is necessary.

¹⁷⁷ 1985 House Rep., supra n. 47, at 424.

¹⁷⁸ Id. at 426.

¹⁷⁹ As discussed supra in Chapter 3, the regulations or statute should be amended to ensure adequate disclosure of transfer pricing methodology and penalize unjustified substantial understatements of tax resulting from nonconformity to the arm's length standard.

¹⁸⁰ See infra Chapter 11.

¹⁸¹ See also the factors set forth in Treas. Reg. section 1.482-2(d)(2)(iii).

¹⁸² It has been suggested that the commensurate with income standard will result in all gains from the sale of intangibles being treated as royalties under the Internal Revenue Code and under our tax treaties, because of the provision in the Code and those treaties covering gains contingent on productivity, use or disposition of the relevant intangible. There is no intention by the Service or Treasury to eliminate the possibility of sale treatment for transfers of intangibles in appropriate cases, either for treaty purposes or for U.S. withholding tax purposes. Further, the Service and Treasury believe that the mere fact that subsequent profits may be taken into account in appropriate cases by U.S. tax authorities in determining transfer prices on audits, or that a lump sum is treated as a deposit on the appropriate section 482 transfer price in order to assure that a commensurate with income adjustment can be made notwithstanding the statue of limitations, does not have this effect. The terms of the transaction itself (i.e., whether it provides for contingent consideration based, e.g., on sales volume or units sold) will determine treatment under the royalty article. Further, even if the commensurate with income standard were incorporated by reference into the relevant sales document, there is no necessary relationship between productivity, use or disposition and a proper commensurate with income payment. For example, sales might increase dramatically in a given year, but the method called for may result in no increase in payments, or the taxpayer may have an arm's length basis for making no adjustment.

¹⁸³ But see Treas. Reg. section 1.482-2(d)(1)(ii)(d), Ex. 3, which appears to allow an inter-year set-off.

¹⁸⁴ GAO, IRS Could Better Protect U.S. Tax Interests, supra, n. 64, at 63; ABA Comm. on Affiliated and Related Corporations, Administrative Recommendation No. 8 (1986) [hereinafter ABA Admin. Rec.]; Langbein, supra n. 160, at 655.

¹⁸⁵ See GAO, IRS Could Better Protect U.S. Tax Interests, supra n. 64, at 48-50.

¹⁸⁶ The following changes have been made to the safe harbor interest rate: 6-8 percent effective January 1, 1976; 11-13 percent effective July 1, 1981; 100-130 percent of the applicable Federal rate effective May 9, 1986. Final regulations were published on June 13, 1988. T.D. 8204, 1988-24 I.R.B. 11.

¹⁸⁷ T.D. 8204, 1988-24 I.R.B. 11.

¹⁸⁸ See Treas. Reg. section 1.482-2(c)(2)(ii).

¹⁸⁹ 1985 House Rep., supra n. 47, at 424-25.

¹⁹⁰ Id.

¹⁹¹ ABA, Admin. Rec., supra n. 184, at 14. The ABA proposals are applicable only to the transfer of tangible property. Other proposals apply only to intangible property. Because many of the safe harbors would have the same advantages and disadvantages regardless of the type of property involved, this discussion does not address the different types of property separately.

¹⁹² Id. at 14-15.

¹⁹³ A variation of this approach retains the goal of a market-based allocation but claims that in some situations the target is best reached by an estimate, or that average prices can be used for certain transactions. The estimate can be provided by some type of formulary apportionment.

¹⁹⁴ Surrey Reflections on the Allocation of Income and Expenses Among National Tax Jurisdictions, 10 Law and Policy in International Business 409, 414 (1978).

¹⁹⁵ Id. at 414.

¹⁹⁶ Langbein, supra n. 160, at 627.

¹⁹⁷ Caves explains that many multinationals exist because of a failure in the market for intangibles. R. Caves, Multinational Enterprise and Economic Analysis (1982). In essence, intra-firm transactions can be more profitable than inter-firm transactions because of the expense of negotiating complete contracts or the inability of a firm to capture the full value of a piece of knowledge through contracts with unrelated parties without fully explaining the knowledge and thus eliminating its value.

¹⁹⁸ For a narrative explanation of the zero profit condition, see R. Lipsey and P. Steiner, Economics 229-231 (6th ed. 1981). For a mathematical presentation of the implications of this condition, see J. Henderson and R. Quandt, Microeconomic Theory 107-110 (3d ed. 1980).

¹⁹⁹ In more complicated situations, both the affiliated corporation and any potential unrelated party participant may each possess monopoly rights that allow them to earn above-normal profits. In deciding whether to use such an unrelated party, a corporation would have to consider what would happen if it attempted to bargain with it. There are analyses, relating to economic game theory, that attempt to predict what the range of outcomes would be in such a bilateral monopoly situation. If the outcome, specifically the income of the potential unrelated party, can be predicted, than it would be proper to use it to determine the income of the corporation's affiliate. This is so, to repeat, because this procedure would allow the corporation's choice between using an affiliate versus an unrelated party to be made free of tax distortions. To implement this procedure, however, one would need to analyze the theoretical models of bargaining situations in detail, and this analysis is beyond the scope of the present discussion.

²⁰⁰Rev. Rul. 87-71, 1987-2 C.B. 148.

²⁰¹ See discussion supra Chapter 4.

²⁰² Appendix A, infra.

²⁰³ Id.

²⁰⁴ Treas. Reg. section 1.482-2(e)(2)(ii).

²⁰⁵ See infra Appendix D for further discussion of unrelated party licenses.

²⁰⁶ I.R.M. section 600 et seq.

²⁰⁷ As Appendix D stresses, it is insufficient merely to replicate a royalty rate in order to achieve a comparable license. For example, the technological services provided by the licensor may have a large impact on the profitability of the license from the licensor's perspective.

²⁰⁸ There may be other cases where the taxpayer can demonstrate an arm's length basis for continuing to use inexact comparables. See discussion infra Chapter 8 regarding periodic adjustments.

²⁰⁹ Because manufacturing is a broad category, the function or activity would be defined more precisely. An example might be "medium instrumentation fabrication, assembly, and testing."

²¹⁰ See discussion of risk infra section E.

²¹¹ See discussion supra Chapter 4.

²¹² 1985 House Rep., supra n. 47, at 424 (1985). A line of court cases not directly relevant to section 482 has reached a similar conclusion. In Carnation and succeeding cases, members of an affiliated group of corporations were denied deductions for "insurance premium" payments to another member of the group that insured predominantly risks of the group. The courts decided that no true insurance was present, because there was no true risk shifting between the parent and its affiliates. Carnation Co. v. Comm'r, 71 T.C. 400 (1979), aff'd, 640 F.2d 410 (9th Cir. 1980), cert. denied, 459 U.S. 965 (1981).

²¹³ Treas. Reg. section 1.482-2(b).

²¹⁴ Treas. Reg. section 1.482-2(e).

²¹⁵ 1986 Conf. Rep., supra n. 2, at II-6.

²¹⁶ See generally Rev. Rul. 56-543, 1956-2 C.B. 327, revoked by Rev. Rul. 77-1, 1977-1 C.B. 161; see also Gen. Couns. Mem. 36,531 (December 29, 1975).

²¹⁷ 31 Fed. Reg. 10394 (1966).

²¹⁸ Whether a particular cost sharing agreement meets the requirements of section 482 is generally a factual question not appropriate for a private letter ruling. There have been private letter rulings regarding issues that are peripheral to the central question of whether a cost sharing agreement is bona fide. However, none of these rulings concerned the characteristics necessary for an agreement to be considered bona fide under the current regulations. See Priv. Ltr. Ruls. 8111103, 8002014, and 7704079940A.

²¹⁹ OECD, Transfer Pricing and Multinational Enterprises, supra n. 158, at 55-62.

/220 /The guidelines for cost sharing agreements are found in paragraph 7 of the German Transfer Pricing Guidelines. English and French translations of these guidelines are contained in Raedler- Jacob, German Transfer Pricing/Prix de Transfert en Allemagne (Kluwer Law and Taxation Publishers, Devender, Netherlands, and Metzner, Frankfurt, Germany 1984). The guidelines are also available in International Bureau of Fiscal Documentation, Tax Treatment of Transfer Pricing (Amsterdam, Netherlands 1987).

²²¹ General Explanation of the DRA of 1984, supra n. 143, at 433.

²²² Id.

²²³ Section 936(h)(5)(C)(i)(I), as amended by section 1231(a)(1)(A), Pub. L. No. 99-514, 100 Stat. 2085 (1986).

²²⁴ 1986 Conf. Rep., supra n. 2, at II-638.

²²⁵ Id.

²²⁶ Id.

²²⁷ Id.

²²⁸ Section 936(h)(5)(C)(i)(I).

²²⁹ Another potential abuse may arise in the context of a domestic affiliate that is a co-developer of the intangible, or otherwise participates in a de facto cost sharing arrangement. In such cases, the foreign entity may try to avoid the characterization of a cost sharing arrangement in order to extract royalties from the domestic affiliate, particularly where the withholding tax on the royalties is reduced by a tax treaty and the royalty income is not taxed or lightly taxed by the foreign jurisdiction.

²³⁰ See Lilly, supra n. 57, at 1150.

²³¹ As discussed in the next paragraph, rights to exploit an intangible in the U.S. must belong to a U.S. affiliate.

²³² The exclusion of affiliate A from the cost sharing arrangement raises the question of which of the other participants should pay for research related to intangibles that may be used by affiliate A. For various reasons, not all affiliates that anticipate using the intangibles developed under the cost sharing agreement may actually participate in the arrangement. For example, there may be reason to exclude a particular affiliate that manufactures only certain types of products and therefore will use only certain types of intangibles developed under the arrangement. Alternatively, the arrangement may not be recognized under foreign law for tax purposes, such that a deduction for cost sharing payments would be denied. In such cases, some affiliate must fund research for intangible rights to be used in manufacturing by the nonparticipants. While there is no clear answer, it seems appropriate that the affiliate that performs the research should fund and receive the residual rights.

²³³ As a roughly analogous requirement, the 1966 regulations required that participants use the intangibles developed under the cost sharing arrangement in the active conduct of their trade or business. Prop. Treas. Reg. section 1.482-2(d)(4)(ii)(b), 31 Fed. Reg. 10,394 (1966). The 1966 regulations would also have excluded as participants companies in the trade or business of performing research for others. This latter exclusion seems unnecessary so long as the affiliate that performs the research and development funds an appropriate portion of the research and development costs and is capable of using the intangible rights that it acquires under the agreement in the manufacture of products.

²³⁴ Expectations will not always prove true, and in some situations the participant that acquires certain rights to intangibles developed under the cost sharing agreement will not ultimately directly exploit those rights. For example, assume that a Dutch affiliate acquires the European rights to intangibles developed under the arrangement with the anticipation of manufacturing products for European markets in Ireland. It later appears that it will be necessary for various reasons to have a locally incorporated entity manufacture in Germany products to be sold in the German market. Unless the German rights to the intangible are transferred in a contribution to capital or other tax free transaction, the German rights would have to be licensed or sold to the German affiliate. In either case the intangible would be subject to section 482 and, generally, the subpart F provisions would treat the resulting royalty or sales income as foreign personal holding company income includible in subpart F income.

²³⁵ It is generally expected that there will not be a profit element in cost sharing agreements. A profit should be required, however, for research performed at the specific request of a group member or a for a person outside the arrangement group. OECD, Transfer Pricing and Multinational Enterprises, supra n. 158, at p. 119. In either case, the amount received should reduce the costs to be shared. One item that should not be included in the costs to be shared among the participants is the buy-in cost of transferring intangibles from the party by whom they were developed to the other participants. This general subject is discussed in section E below.

²³⁶ The legislative history of the 1984 Act states that the provisions of section 367(d) can be avoided by selling intangibles subject to the application of section 482 to the sale. S. Rep. No. 169, 98th Cong., 2nd Sess., vol. 1 at 368 (1984). The legislative history did not contemplate that a transfer at cost would avoid the application of section 367(d).

²³⁷ Treas. Reg. section 1.482-2(b)(2).

²³⁸ Section 936(h)(5)(C)(i)(IV)(a); Treas. Reg. section 1.936- 6(a)(5).

²³⁹ Boccardo v. U.S., 12 Cl. Ct. 184 (1987).

²⁴⁰ Section 1.861-8 sets out rules for allocating and apportioning deductions between U.S. and foreign source gross income. A special allocation rule gives companies the right to allocate fewer U.S. research and development expenses to foreign source income, even though the income generated by the expenses is foreign source. Treas. Reg. section 1.861-8(e)(3)(B)(ii). Under the 1986 Act, 50 percent of all amounts allowable as a deduction for qualified domestic research and experimental expenditures is apportioned to income from sources within the United States, with only the remaining 50 percent apportioned on the basis of gross sales or gross income of companies benefitting from the research and development. This special provision applies only to expenses incurred in tax years after August 1, 1986, and on or before August 1, 1987. Section 1216, P.L. 99-514, 100 Stat. 2085 (1986). Provisions similar in concept are currently under consideration in Congress.

²⁴¹ Treas. Reg. section 1.44F-6(e); Priv. Ltr. Rul. 8643006 (July 23, 1986).

²⁴² Section 936(h)(5)(c)(i)(I); Treas. Reg. section 1.936- 6(a)(3).

²⁴³ 1985 House Rep., supra n. 47, at 418-419.

²⁴⁴ This approach is generally consistent with the cost sharing regulations published in 1968, which permitted pre-existing cost sharing agreements to be modified within 90 days of publication of the section 482 regulations. Treas. Reg. section 1.482-2(d)(4).

APPENDIX A

ANALYSIS OF QUESTIONNAIRE RESPONSES

A. IRS ACCESS TO PRICING INFORMATION

- Significant section 482 issues were identified by IEs over 70% of the time and by the domestic agent about 10% of the time.

- Significant section 482 issues were initially identified using the following sources:

                Source              Percentage of Responses

- Almost 90% percent of respondents stated the time alloted to examine the section 482 issue was flexible.

- Factors affecting time allotment were:

                                    Percentage of Responses

- The portions of the Annual Report specifically considered for identification or development of section 482 issues were:

                                    Percentage of Responses

- In 29% of the reported cases, section 482 issues were identified and not pursued. In 20% of the reported cases, section 482 issues were identified and not changed.

- The taxpayer had no readily available basis to support its section 482 transaction in almost 75% of the cases.

- In over 50% of the reported cases, taxpayers failed to make timely and complete responses to questions asked in developing section 482 issues.

- More than 66% of the responses indicate that there was no reasonable explanation for any delay in responding to questions aimed at developing section 482 issues.

- Reasons given for delays in responding to IDRs:

- Tax department staffing.

- Records located overseas.

- Foreign parent refused to produce records.

- Extremely detailed requests for information from the foreign subsidiaries.

- Lack of cooperation existed between the parent and subsidiary.

- Unreasonable delays in responding to requests for information concerned:

- Control of affiliates -- 34.4% of responses.

- Existence of comparable transactions with third parties -- 48.5% of responses.

- Terms of comparable transactions with third parties

- 48.5% of responses.

- The average length of delay to responses was 12.2 months. The portion of the delay deemed as reasonable by the responding IEs averaged 2.2 months.

- Using a summons to obtain information was considered as follows:

                                    Percentage of Responses

- Formal summons discussed -- not used because case manager felt that the action would close the door to future cooperation in domestic audits.

- It was felt that the issuance of a summons for records would only delay the overall development and completion of the case.

- Summons considered due to delay in response to agent and economist. Not issued as taxpayer eventually did respond, although the responses were generally inadequate.

- Used as a threat to speed-up IDR response time. Generally not useful.

- Taxpayer complained that our request was overly broad. After discussion with Branch Chief, including the use of section 982 and summons, Taxpayer offered an alternative to books and records, under which most of the information requested was eventually received.

- Section 982 procedures arose to the following extent:

                                    Percentage of Responses

- The section 982 procedures were mentioned in opening conference.

- Taxpayer's practice was to furnish as little information as possible with approximately a 90-day turn around time. When subsequent IDRs needed to be issued, the same practice was followed.

- Taxpayer is well aware of our open year policy and planned closing dates. IE was of the opinion that taxpayer feels if they use delaying tactics the case will become "old" and will be closed out undeveloped. Taxpayer's delaying tactics prevented the issuance of follow up IDRs. Taxpayer refused to furnish its parent's cost data for the products that were at issue.

- Taxpayer was late in providing data after initial adjustments were prepared. Taxpayer's attorney's tactic was to continue indefinite discussion of the issue, including appeals to the National Office.

- Taxpayer's responses to IDRs took from 6 months to a year. The audit was stretched out to the point that the planned audit closing date became a problem.

- Section 482 issues were raised on the previous audit. The prior examiner received virtually no information from the taxpayer. Detailed information was submitted by the taxpayer in the Appeals protest. This information was used by the IE and the economist in subsequent years.

- Taxpayer clearly not responsive to IDRs that could hurt him. In three cycles, the key IDRs were not answered.

- District allows taxpayers excessive amount of time to respond to IDRs. A two year audit cycle takes five years to complete.

- According to responding IEs, the following adversely affected the development of section 482 issues:

                                    Number of Responses

- Only 3% of respondents claimed that competent authority considerations affected their decision to pursue any section 482 issue.

- Appeals settled 28% of the section 482 issues in the reported cases.

- 69% of respondents disagreed with the terms of the Appeals settlement.

- Counsel was involved in 26% of cases settled by Appeals.

- 76% of respondents did not receive a copy of the Appeals settlement.

- The section 482 issues were resolved at the examination level 43.4% of the time in the reported cases.

- The IE was appropriately consulted in 90% of the reported cases resolved by Examination.

- The section of the 482 regulations providing the basis for Examination's resolution was:

                                    Percentage of responses

B. APPLICATION OF PRICING METHODS

- IEs stated that the taxpayer used comparables as follows with regard to section 482 transactions:

                                    Percentage of responses

- Description of comparable(s) relied on by taxpayer in planning or defending its section 482 transaction:

- In performance of services, taxpayer tried to establish comparables based on charges to third parties.

- The taxpayer presented pricing data with an unrelated distributor of similar property in a different country.

- Sales invoices to third parties.

- Contracts between unrelated third parties.

- Taxpayer claimed it was charging the same royalties to all of its foreign subsidiaries.

- Taxpayer secured quote from third party in small quantity transaction.

- Weighted average of Canadian CFC's third party sales. Method required by Revenue Canada.

- Sales to 50% owned subsidiaries.

- Industry norms.

- Only 19% of those responding accepted the taxpayer's comparables.

- Examples of explanations why taxpayer's comparables were not accepted:

- The taxpayer was looking only at the services and not looking at the overall transaction, i.e. providing services, the transfer of technology and other intangibles.

- The comparables did not reflect true arm's length pricing because they ignored the fact that the parent performed substantial marketing, distribution, and trademarking functions, or the circumstances were otherwise different.

- The taxpayer's method generated approximately 185% of the combined profit to the low tax entity and a loss to the U.S. parent.

- The taxpayer's comparables included very small volumes.

- The taxpayer relied on comparables based on "industry norms" in 41% of the cases reported.

- Description of industry average "comparables" submitted by the taxpayer to support its assertions:

- Robert Morris Associates -- Annual Statistics by SIC Code of Gross Profit Margins for Wholesale Automotive Equipment Dealers.

- Taxpayer relied on published AFRA demurrage rates.

- Taxpayer used the average resale mark-ups for the industry.

- Comparables were used as a basis for adjustment in about 75% of the cases reported.

- Representative sources for finding comparables relied upon by IEs:

- The IRS Economist used industry (construction) comparables. The services that the offshore company performed were that of a construction manager. The economist determined that, based on comparables, the offshore company should receive a comparable profit. The remaining profit was allocated back to the taxpayer for services and intangibles.

- Economist used industry statistics from docketed cases and SEC reports of unrelated taxpayers.

- Taxpayer was requested and did provide comparable transactions of its manufacturer parent with its unrelated distributors.

- Information from third parties with respect to comparable transactions (a similar product under similar circumstances in a similar market).

- License agreement with related and unrelated parties.

- Analysis of industry, consulting with ISP, contacting other IEs examining similar companies.

- Third party agreements for similar services or intangibles with same taxpayer in same circumstances.

- Obtained Form 10K information from several U.S. entities and used to establish the arm's length price on a cost plus basis.

- Third party sales of taxpayer and compiled statistics from "Robert Morris & Associates -- Annual Statement Studies."

- The following are descriptions of significant problems encountered by IEs in developing a comparable:

- The information sought from third parties was old -- 5 to 6 years. In one instance the third party relocated and finding records was difficult. Records were not organized when obtained.

- Difficulty in acquiring information from third parties and obtaining permission to use data from the third party.

- Adjusting for differences in geographic markets.

- There are no comparables at this level of distribution. All manufacturing/sales companies in this industry are related.

- The products manufactured and sold by the Puerto Rican affiliates were the high volume, profitable products. The functions performed by the subsidiaries did not correspond to any third party situation. Consequently, the comparables identified were useless.

- Methods used in proposing tangible property adjustments by percentage of response:

           Comparable Uncontrolled Price      31%

- Market penetration was not considered as a factor when determining section 482 adjustment in about 75% of the cases reported.

- The taxpayer's documentation considered the priority of pricing methods in 50% of the cases reported where documentation existed.

- Excerpts of descriptions of "other methods" used by the taxpayer to justify its pricing policies.

- Taxpayer claimed intercompany price was arm's length because it was negotiated between the lower tier sub and its foreign parent.

- Taxpayer contended all income attributable to intangibles belonged to the Puerto Rican affiliate.

- Prior appellate settlements.

- Taxpayer attempted to identify other charges made by the parent to its subsidiaries that were equivalent to the royalty adjustment that was proposed.

- Taxpayer explained its method as being required by Revenue Canada.

- "Old fashioned business know-how".

- In over 75% of the cases reported, the taxpayer relied on a profit split to determine its transfer price.

- Descriptions of taxpayer's methods of computing profit split:

- Market penetration accounted for any difference in price.

- Resale price.

- Taxpayer computes revenues of products made in Puerto Rico then reduced them by: cost of sales P.R., an R&D cost sharing amount based on section 936(h), selling and administrative expenses based on a fractional calculation and other income or expenses using section 936(h) -- this gave CTI of which they considered the P.R. entity to possess half.

- Taxpayer used prior Appeals settlement profit split.

- Taxpayer allowed its domestic subsidiaries a profit of 6% on the cost incurred by such subsidiaries.

- Taxpayer used profit earned by the parent on other transactions with related parties. Taxpayer's contention is that the subsidiary's high profit is irrelevant as long as the parent made an adequate profit on the transaction.

- Taxpayer claimed that the marketing company should recover all of its marketing costs (11% of sales) plus derive a profit (4% of sales).

C. SERVICES

- The IEs proposed an adjustment for services in 32.8% of the reported cases.

- In 43% of the reported cases, difficulty in applying the services regulations was the primary reason for not making an adjustment.

- Difficulties reported by IEs in applying the service regulations included:

- Determining for whose benefit services were provided.

- Undue burden on the IE to: (1) isolate costs, (2) determine whether the service was an integral part of the business, and, (3) develop comparables to determine proper adjustment.

- Determining what services were rendered, by whom, the amount of time spent rendering the service, and the cost of the service.

- The service regulations do not allow a profit in the allocation.

- Examples of difficulties in deciding whether to propose an adjustment for services rendered or intangibles transferred included:

- Taxpayer only wanted to charge for services at the same rate they generally charged third parties. IE used a functional analysis to show that know-how was also transferred to related parties but not to third parties.

- Taxpayer does purchasing for a subsidiary and picks up a 5% profit. IE had no idea if the profit mark-up was appropriate.

- Taxpayer allocated a portion of cost based on time spent by officers. Because of the 25 percent rule, the IE was prevented from making an adjustment.

D. INTANGIBLES

- 50% of the cases reported involved a significant transfer of intangibles.

- Adjustments were made under Treas. Reg. section 1.482-2(d) in about 50% of the reported cases.

- Factors reported by IEs as affecting the decision to proceed under services or sales of tangible property regulations rather than under the intangibles section:

           a) Inability to separately identify

- Spell out that T/P's reputation is an intangible.

- Clarify that a CFC should not get a marketing profit if they don't do marketing.

- In reported cases involving the transfer of intangibles to a related party, the taxpayer acknowledged the transfer at the outset of the examination 48.6% of the time.

- Documentation produced by taxpayers with respect to the transfers of intangibles:

- Unrelated professional appraisal

- Corporate minutes and legal documents

- Licensing agreements

a. With related entities

b. With unrelated entities

- Section 351 transfer documents

- Section 367 ruling

- Marketing intangibles were involved in 25% of the intangible cases.

- Data relied upon or method of intangible valuation:

- Advertising and marketing expenses

- Trade name and trademark defense costs

- Distribution costs

- Market dominance -- 3rd party brokerage statements -- market studies -- royalties textbooks -- profit and loss comparisons -- patent infringement cases -- prevailing rates in the industry.

- Compared rates charged by taxpayer to unrelated parties.

- Used functional analysis to show that CFCs were not active in crude oil trading. Only administrative and communication services were performed. Economist determined an arm's length service fee due the CFC for services performed, then used the residual method to value the income to be allocated to the domestic subsidiary.

- Taxpayers used a cost sharing agreement with a related party in 17 1/2% of the cases reported.

- Description of cost sharing agreements:

- Parent billed Puerto Rican subsidiary for their share of R&D.

- R&D costs shared based on percentage of sales. Direct costs charged to entity deriving benefit.

- Reimbursed for R&D, marketing, and administration.

- Share in R&D and reimbursed marketing costs.

E. USE OF SPECIALISTS AND COUNSEL

- The following specialists were involved with reported cases:

                               Percentage of responses

- Economist performed a functional analysis of activities of CFCs and identified comparables. The economist was requested after the taxpayer prepared a section 482 study to refute the proposed adjustment. In order to be successful in Appeals or court, an economist was considered essential.

- An economist was requested to assist in developing the third party comparables found by the IE and to assist in assessing taxpayer's arguments.

- An economist was involved in the prior year and, accordingly, was requested for the current cycle. An engineer was needed to assess the electrical engineering function of the related companies.

- An economist was requested for a valuation of risk capital.

- An engineer was used to compare the CFC shop to unrelated shops. An economist found comparable mark-ups.

- An economist was used on an informal basis as to procedure and appropriate percentage of profit.

- An economist was used on the royalty expense issue and to evaluate a trademark transfer; engineer was used to evaluate a fee structure.

- The economist was requested to review our position with respect to imputation of royalty and technical service income. Looking beyond this, the economist suggested that a potential pricing issue existed. The economist was assigned late in the examination and was not granted the time needed to develop the pricing issues.

- The economist added support in the development of the transfer of intangibles issue. The economist was brought into the case after we recognized and began developing the issue.

- An economist was requested by IE -- used to establish arm's length pricing of foreign autos.

- An appraiser was brought in by the IE and the Case Manager to evaluate the sale of U.S. entity's stock at book value and to establish control elements.

- The economist performed a functional analysis on Puerto Rican operations. It was difficult to attack taxpayer's pricing as long as we accepted the function of the Puerto Rican subsidiary as a full-fledged manufacturer.

- IE received informal advice on a stewardship issue.

- According to the IEs, specialists were brought in at appropriate times in 91% of the reported cases.

- Specialists raised additional issues in 9% of the reported cases.

- 79% of respondents thought that specialists' reports were particularly useful in proposing issues.

- Brief descriptions of specialists' reports which were helpful:

- The economist report was very useful since it discussed, in depth, the functional analysis and comparables used in determining the arm's length rates for intangibles and services.

- Economist report supported the IE's conclusion that market penetration was not prevalent in the years under examination, which was the thrust of the taxpayer's argument.

- Economist report gave a basis for reasonable profit factor in pricing computation.

- The economist's report was useful in establishing the service fee for CFC and the function of taxpayer's worldwide trading activity.

- The report made a good case for treating the subsidiary as a contract manufacturer. Prior to that, taxpayer was maintaining its position that the resale method applied.

- The economist developed a method of joining data secured by means of a private survey with data from a public source. The economist revealed to the agent a number of other sources that are available for statistical analysis and comparisons.

- Specialist's reports were used:

                                    Percentage of Responses

- 14% of respondents stated that restrictions were placed on their use of a specialist.

- The taxpayer employed a specialist in 27% of the reported cases.

- The taxpayer's use of a specialist was as follows:

                                    Percentage of Responses

- 14% of respondents claimed that had counsel been involved, their cases would have been better developed.

- Counsel was involved at a timely stage of case development in 76% of reported cases.

- Persons who determined that counsel should become involved were:

                                    Percentage of Responses

APPENDIX B

SECTION 482 QUESTIONNAIRE -- INTERNATIONAL EXAMINERS

CASE NAME ___________________________ Years _________________

           PLEASE ATTACH A COPY OF YOUR EXAMINATION REPORT

     (A) if the listed section 482 issue was present in this case;

                               (A)     (B)     (C)        (D)

[ ] Please check here if this case involved section 936.

[ ] Please check here if you have answer question 100 on this questionnaire.

 1.   Who initially identified the significant 482 issues in this

TRANSFER PRICING LAW AND PRACTICES OF SELECTED U.S. TREATY PARTNERS

CANADA

The statutory basis for transfer pricing allocations is section 69(2) of the Income Tax Act ¹ that provides as follows:

Where a taxpayer has paid or agreed to pay to a nonresident person with whom he was not dealing at arm's length as to price, rental, royalty or other payment for or for the use or reproduction of any property or as consideration for the carriage of goods or passengers, or other services, an amount greater than the amount (in this subsection referred to as the "reasonable amount") which would have been reasonable in the circumstances if the non-resident and the taxpayer had been dealing at arm's length, the reasonable amount shall, for the purpose of computing the taxpayer's income under this Part, be deemed to have been the amount that was paid or payable therefor.

Section 69(3) of the Income Tax Act provides as follows:

Where a non-resident person has neither paid nor agreed to pay to a taxpayer with whom he was not dealing at arm's length as to price, rental, royalty or other payment for or for the use or reproduction of any property, or as consideration for the carriage of goods or passengers or for other services, the amount that would have been reasonable in the circumstances if the nonresident person and taxpayer had been dealing at arm's length, that amount shall, for the purposes of computing the taxpayer's income under this Part, be deemed to have been received or receivable by the taxpayer therefor.

Sections 69(2) and (3) apply to all taxpayers including individuals, unincorporated businesses, trusts, and corporations. However, section 69(2) applies only when the Canadian taxpayer has paid more than a reasonable amount and does not apply when the Canadian taxpayer has paid less than a reasonable amount. Similarly, section 69(3) applies only when the Canadian taxpayer has received less than a reasonable amount and does not apply when the Canadian taxpayer has received more than a reasonable amount.

Interpretation of this statute by the Canadian government has been provided in an Information Circular issued by the Department of National Revenue in 1987. ² In this Circular, Revenue Canada interprets the phrase "reasonable in the circumstances" to mean the price that would have prevailed if the parties to the transaction had been dealing at arm's length. In applying this arm's length standard, Revenue Canada has endorsed the methods enumerated in the 1979 OECD report on Transfer Pricing and Multinational Enterprises. Although the methods are not prioritized as to the order that they must be used, Revenue Canada has expressed a preference for the comparable uncontrolled price method and the following sequence of tests:

     1. Sales by the taxpayer to unrelated parties;

          a) prevailing rates in the industry;

FRANCE

The statutory basis for transfer pricing allocations is Article 57 of the French General Tax Code which is as follows:

In assessing income tax due by enterprises which are subordinated to or controlled by enterprises established outside France, the income to which is indirectly transferred to the latter, either by increasing or decreasing purchase or sale prices, or by any other means, shall be restored to the trading results shown in the account. The same procedure is followed with respect to undertakings which are controlled by an enterprise or a group of enterprises also controlling enterprises located outside France.

* * *

Should specific data not be available for making the adjustments provided for the preceding paragraph, the taxable profits are determined by comparison with those of similar undertakings run normally. ⁶

The tax administration has endorsed the 1979 OECD Report on Transfer Pricing and Multilateral Enterprises, although the OECD guidelines are not binding on the French authorities. ⁷ In enforcing Article 57, the authorities generally compare profit margins of similar entities to identify any abnormalities. ⁸

A similar approach is apparently taken with respect to the payment of royalties by a French taxpayer in that a deduction will be allowed for the payment only to the extent that the net income of the payee or subsidiary is at least equal to that realized by a French enterprise engaged in a similar activity. Furthermore, under the French exchange control law, all royalty agreements with and payments to nonresidents must be reported, prior to payment, to the National Institute of Industrial Property. ⁹

The experience of the Office of Assistant Commissioner (International) is that French companies filing consolidated returns that include foreign subsidiaries must agree to allow French tax authorities on-site inspection of the subsidiaries' books and records; that, as indicated above, profit experience is a very important factor in examinations; and that no guidelines have been developed for evaluation of royalty cases involving transfer of intangibles. When intangibles are transferred to a tax haven, payments received on account of the transfer are deemed to be unreasonable, and the burden is on the taxpayer to establish that the payments are reasonable.

Although cost sharing arrangements are permitted, ¹⁰ the French authorities do not have specific rules applicable to such arrangements.

GERMANY

The statutory basis for intercompany pricing adjustments includes Article 8(3) of the Corporation Tax Law, which states that hidden distributions of income do not reduce taxable income. ¹¹ Section 31 of the Corporation Tax Regulations interprets "hidden distributions" to include a benefit granted by a company to a related person, outside the normal statutory profit distributions, which an orderly and conscientious manager would not have granted to an unrelated party under comparable circumstances. ¹²

Similarly, Article 1(1) of the Foreign Tax Affairs Law states that where the income of a taxpayer resulting from his business relationship with a related person is reduced because the taxpayer, within his business relationship extending to a foreign country, has agreed on terms and conditions which deviate from those unrelated third parties would have agreed upon under the same or similar circumstances, then his income shall, notwithstanding other provisions, be determined as if the income had been earned under terms and conditions agreed upon between unrelated third parties. ¹³ Article 1(1) applies to all related or affiliated taxpayer, including individuals, partnerships, and corporations.

Paragraph 2.1.1 of the Transfer Pricing Guidelines of 1983 ¹⁴ (hereinafter referred to as Guidelines) requires that business dealings between related parties be evaluated on the principle of arm's length dealings between independent parties acting in a situation of free competition.

Paragraphs 2.2.2 through 2.2.4 of the Guidelines list the following as the standard methods for evaluating transfer prices: comparable uncontrolled price method, resale price method, and cost plus method. In contrast to section 1.482-2(e)(l)(ii), which requires that these methods be used in the order prescribed if the circumstances of the case permit, paragraph 2.4.1 of the Guidelines states that, "[t]here is no single correct sequence of standard methods for the examination of transfer prices which applies to all groups of cases." Also, paragraph 2.4.2 of the Guidelines, similar to the "fourth method" provision of section 1.482-2(e)(l)(iii), allows use of a combination of the three standard methods or of other methods.

In cases involving transfers of intangibles to offshore manufacturing affiliates, paragraph 5.1.1 of the Guidelines recognizes that a determination must be made whether the transferor has received an adequate consideration for the transfer of the intangible. Paragraph 5.2.2 of the Guidelines states that an arm's length price for transfer of an intangible is to be determined under "an appropriate method." Paragraph 5.2.3 of the Guidelines indicates that the preferable method is the comparable uncontrolled price method but, if the facts of a case will not support application of this method:

then the starting point for the examination is the consideration that a sound business manager of the licensee enterprise would only pay a royalty up to an amount which leaves the enterprise WITH AN ACCEPTABLE COMMERCIAL PROFIT from the licensed product. [Emphasis added.]

According to paragraph 5.2.4 of the Guidelines, the cost plus method may be used "in exceptional cases." One such exceptional case is described in paragraph 3.1.3 (Example 3) of the Transfer Pricing Guidelines of 1983, as follows:

An enterprise transfers particular manufacturing functions to a foreign subsidiary corporation. Production and marketing by the foreign corporation are closely tied in with the business of the domestic enterprise.

The articles produced are purchased by the parent corporation under a long-term arrangement. The subsidiary corporation with its limited range of production could not in the long run survive as an independent corporation. Between unrelated parties the production would have been carried out on a subcontract basis. . . . The transfer price can accordingly be determined using the cost plus method.

This approach is essentially the same as that of the IRS in the Lilly case, discussed supra, Chapters 4 and 5.

One commentator, Mr. Friedhelm Jacob, Counselor for Tax Affairs at the West German Embassy in Washington, D.C., has noted that, in contrast to the 1986 Tax Reform Act amendments to section 482, which require adjustments over time for substantial changes in circumstances, the German approach has been that the determination of fair market consideration for an intangible is made at the time of the transfer. ¹⁵

Paragraph 2.4.3 of the Guidelines recognizes that related entities may enter into bona fide cost sharing arrangements and that such arrangements can affect transfer prices. Under paragraph 7.1.1 of the Guidelines, cost sharing arrangements are to be taken into consideration in making transfer price income allocations between the entities involved in the arrangement. However, full costs, direct and indirect, must be calculated and included under a recognized accounting method. If the cost sharing arrangement is to be recognized, the services must be distinguishable and the aggregate of the costs must be separable as to the services. No profit is permitted in such an arrangement. Furthermore, a taxpayer seeking a deduction for a cost sharing payment must have "a specific right, definite both in nature and scope, to benefit from the activities of the central organization."

The experience of the Office of the Assistant Commissioner (International) has been that, if the comparable uncontrolled price method cannot be utilized, the German tax authority generally allows a price or royalty that leaves the enterprise with an acceptable commercial profit from the sale or license, although there are no published industrial safe harbor profit norms. With respect to the transfer of intangibles, the tax authority does not consider that the intangible property itself was used when a person acquires the goods or merchandise produced with the intangible.

JAPAN

Article 66-5 of Japan's Special Taxation Measures Law is effective for taxable years beginning on or after April 1, 1986. ¹⁶ This law applies only to corporations (and certain other legal entities recognized under Japanese law), but does not apply to individuals, unincorporated joint ventures, and similar entities. Furthermore, Art. 66-5 applies only to transactions between a foreign corporation and a corporation that is subject to Japanese tax and only when the two entities are related by at least a 50 percent direct or indirect ownership.

Article 66-5 is as follows:

(1) In the event that, during a business year commencing on or after April 1, 1986, a corporation ("Corporation A") has conducted sale or purchase of assets, provision of services or other transactions with a foreign affiliated corporation ("Corporation B") which has a relationship with Corporation A in which one of the corporations in question, directly or indirectly, owns a number of shares comprising 50 percent or more of the total number of issued shares of the other one or has any other special relationship with Corporation A ("Special Relationship") and, in connection with such above mentioned transaction ("Transaction"), if the amount of consideration of which payment was received by Corporation A from Corporation B was less than an arm's length price, or if the amount of consideration which Corporation A paid to Corporation B was greater than an arm's length price, then, for purposes of corporate income taxation of Corporation A for the said business year, the Transaction will be deemed to have been carried out at an arm's length price.

Paragraph (2) of Article 66-5 lists the methods by which the arm's length price is to be determined, although in contrast to section 1.482-2(e)(l)(ii) the methods are not prioritized as to the order in which they must be employed. In the case of sale or purchase of inventory assets, the permissible methods are comparable uncontrolled price, resale price, cost plus, and, if none of the first three methods may be applied, a method "similar to" the first three methods or "other methods prescribed by cabinet order." In the case of other transactions (i.e., that do not involve the sale or purchase of inventory assets), the arm's length price is determined by "a method which is equivalent to" the comparable uncontrolled price method, the resale price method, the cost plus method, and, if none of the first three methods may be applied, a method "equivalent to" or a method which is "similar to" one of the first three methods. ¹⁷

The comparable uncontrolled price method is described generally as the price that would have been paid between a buyer and a seller who are unrelated, where the sale or purchase of inventory is the same type of inventory as the inventory involved in the subject transaction, and the circumstances, including commercial level and transaction volume, are similar. It is permissible to use this method where the transactions are not precisely comparable, but it is possible to adjust for differences. ¹⁸

The resale price method is described as the price computed by deducting a normal amount of profit (meaning an amount computed by multiplying the resale price by a normal profit percentage) from the amount of consideration when a buyer of inventory assets involved in the subject transaction resells inventory assets to a person with which it has no special relationship. ¹⁹

The cost plus method is described as the price computed by adding a normal amount of profit (meaning an amount computed by multiplying the amount of costs by a normal profit percentage), to the amount of the costs of the seller to acquire, by purchase, manufacture, or other acts, the inventory assets involved in the subject transaction. ²⁰

The guidance given by the Japanese legislation for determinating an arm's length price for the transfer of an intangible asset is that methods similar to comparable uncontrolled price, resale, and cost plus methods can be used, and that, if none of these methods is applicable, a fourth method may be used.

A unique aspect of Japan's transfer pricing law is a pre- confirmation system whereby a Japanese parent or subsidiary may request pre-approval of a sale or purchase price from a foreign related entity from the tax administration. The purpose for this procedure is to reduce the number of transfer pricing cases and to eliminate uncertainties in international transactions. No such procedure is available under U.S. law, and the IRS would not grant such a ruling because of the factual nature of the issue. The Japanese experience to date is that taxpayers are taking a "wait and see" attitude towards the pre-confirmation procedure. ²¹

UNITED KINGDOM

The statutory basis under U.K. law for adjustments to transfer prices is section 770 of the Income and Corporation Taxes Acts of 1988 [ICTA]. This section provides that where any property is sold and:

(a) the buyer is a body of persons over whom the seller has control, or the seller is a body of persons over whom the buyer has control, or both the buyer and the seller are bodies of persons over whom the same person or persons has or have control; and

(b) the property is sold at a price ("the actual price") which is either --

(i) less than the price which it might have been expected to fetch if the parties to the transaction had been independent persons dealing at arm's length ("the arm's length price"), or

(ii) greater than the arm's length price,

then, in computing for tax purposes the income, profits or losses of the seller where the actual price was less than the arm's length price, and of the buyer where the actual price was greater than the arm's length price, the like consequences shall ensue as would have ensued if the property had been sold for the arm's length price.

This section applies to rentals, grants and transfers of rights, interests or licenses, loan interest, patent royalties, management fees, payments for services, and payments for goods.

Guidance with respect to application of section 770 of the ICTA is provided in Notes published by Inland Revenue. ²³

An Inland Revenue Note defines "arm's length price" as the price which might have been expected if

the parties to the transaction had been independent persons dealing at arm's length, i.e. dealing with each other in a normal commercial manner unaffected by any special relationship between them. ²⁴

The Note dealing with methods of arriving at arm's length prices is as follows:

In ascertaining an arm's length price the Inland Revenue will often look for evidence of prices in similar transactions between parties who are in fact operating at arm's length. They may however find it more useful in some circumstances to start with the resale price of the goods or services etc. and arrive at the relevant arm's length purchase price by deducting an appropriate mark-up. They may find it more convenient on the other hand to start with the cost of the goods or services and arrive at the arm's length price by adding an appropriate mark- up. But they will in practice use any method which seems likely to produce a satisfactory result. They will be guided in their search for an arm's length price by the considerations set out in the OECD Report on Multinationals and Transfer Pricing. (This Report examines the considerations which need to be taken into account in arriving at arm's length prices in general and also in particular in the context of sales of goods, the provision of intra-group services, the transfer of technology and rights to use trade marks within a group and the provision of intra-group loans). ²⁵

² Department of National Revenue Information Circular No. 87- 2, International Transfer Pricing and Other International Transactions (Feb. 27, 1987).

³ Id. at paras. 13-19.

⁴ Id. at para. 46.

⁵ Id. at para. 42.

⁶ Code General des Impots, art. 57.

⁷ See Instruction Administrative (May 4, 1973), published in Bulletin Official de la Direction Generale des Impots 4A-2-73.

⁸ BNA, No. 364-9253, France: Transfer Pricing Within Multinational Enterprises and Article 56 of the French General Tax Code 11 (1980).

⁹ Id. at 11.

¹⁰ Id. at 11.

¹¹ Koerperschaftsteuergesetz art. 8(3).

¹² Koerperschaftsteuerrichtlinien section 31.

¹³ Koerperschaftsteuergesetz art. 1(1).

¹⁴ See Intl. Bureau of Fiscal Documentation, The Tax Treatment of Transfer Pricing (1987) (English translation).

¹⁵ Jacob, The New "Super-Royalty" Provisions of Internal Revenue code 1986: A German Perspective, 27 European Taxation 320 (1987).

¹⁶ Japan Special Taxation Measures Law, art. 66-5 (March 31, 1986), an outline of which is contained in Appendix I, to the Speech of Toshiro Kiribuchi, Deputy Commissioner (International Affairs), National Tax Administration, at the International Tax Institute Seminar, New York, N.Y. (June 2, 1986).

/17 Special Taxation Measures Law, art. 66-5, section 2(i) and (ii).

¹⁸ Special Taxation Measures Law, art. 66-5, section 2(i)(a).

¹⁹ Special Taxation Measures Law, art. 66-5, section 2(i)(b).

²⁰ Special Taxation Measures Law, art. 66-5, section 2(i)(c).

²¹ Go Kawada, Director, Office of International Operations, National Tax Administration, Remarks at the International Tax Institute Seminar, New York, N.Y. (June 27, 1988).

²² Income Tax Acts of 1988, section 770.

²³ Thomas, Richard, Deputy to Assistant of Taxes, Board of Inland Revenue, Remarks at the International Tax Institute Seminar, New York, N.Y. (June 27, 1988).

²⁴ The Transfer Pricing Of Multinational Enterprises, Notes by the UK Inland Revenue (Jan. 26, 1981), at 1, reprinted in Int'l Bureau of Fiscal Documentation, Tax Treatment of Transfer Pricing (1987).

²⁵ Id. at 3.

APPENDIX D

AN EMPIRICAL ANALYSIS OF THE MARKETPLACE FOR INTANGIBLES

A. INTRODUCTION

One question that has been raised during the preparation of the paper is whether the requirement for periodic adjustments in certain situations is consistent with the manner in which unrelated parties structure arrangements involving transfers of intangibles. Additionally, the Congressional concern about related party use of inappropriate comparables raises questions about which characteristics of unrelated party arrangements should be included in related party arrangements. This appendix draws upon academic work that examines actual licensing experience by unrelated parties in an effort to address this issue. ¹ Although each of the papers examined had a different goal, the underlying data collected through surveys or interviews with licensees and licensors provides relevant information about what unrelated parties do.

In addition to synthesizing other authors' work on technology transfers, the Service and Treasury have collected a sample of license agreements drawn from the records of the Securities and Exchange Commission ("SEC"). The SEC requires that companies disclose license agreements that are "material" to the operation of the company. ² The disclosures to the SEC provide a potential data base of over five hundred agreements. Only sixty of these agreements have been analyzed for this report. The choice of agreements in the sample was largely dictated by the ease of discovery and the availability of documents within the SEC's files. The sample consists of forty agreements for the manufacturing industry, most of which are clustered in the Standard Industrial Classifications (SIC) for pharmaceutical preparations, toilet preparations, electronic computing equipment, semiconductors, surgical and medical equipment, and ophthalmic goods. The other twenty agreements are for the services industry, mostly within the SICs for computer programming, computer software, and research and development laboratories.

In this study, the SEC documents have been used primarily for further illustration of the points raised by other authors. Further examination of the SEC agreements will be conducted after publication of this study, with a view toward relating them to financial accounting and tax data of the firms involved, and publishing the results. An analysis of available data might give a more complete picture of the incomes realized by the two parties to the transaction and suggest criteria for determining a profit split in comparable cases.

B. GOAL OF LICENSING AGREEMENTS

Parties contemplating entering into a license agreement are ultimately concerned with the income they can expect to receive from the arrangement. The existence of proprietary knowledge suggests that production and sale of the product will result in profits that are greater than those necessary to provide a normal rate of return to the inputs provided by both parties. The actual division of these profits will depend on each party's forecast of the total profits, and on the relative bargaining strength of the two parties.

Some authors have formally diagrammed and discussed the negotiating range of unrelated parties. ³ The basic premise is that the parties are concerned with the split of the net income from the product. Baranson reports that Bendix officials "indicated that, if a broad cross-section of American industry were polled, one would find that the average goal is a return to the licensor equal to about one- third of the profit of a well-run, well-established licensee with a broad market." ⁴ Caves et al. find that the licensor will capture an average of forty per cent of the expected profits that remain after the deduction of a normal return on capital. According to their sample, the range of the split leaves one-third to one-half of the profits to the licensor. ⁵ Contractor's overview of the literature suggests that licensors "should settle for a 25 to 50 percent share of the licensee's incremental profit." ⁶

These averages can not necessarily be used to replicate an individual arm's length deal because they do not, for example, control for the specific economic activities undertaken by the parties. They do show that unrelated parties enter negotiations for a license agreement with expectations about the total profits that they think will be earned from the exploitation of the intangible, and about the share of the profit that they can expect to receive.

C. PAYMENT TERMS FOR A LICENSE

Although the goal is to capture a portion of the profits, the provisions in license agreements rarely specify that a percentage of the profits will be paid as compensation for the right to exploit the intangible. This may be because the licensor fears that the accounting for profits can be manipulated too easily by a licensee, who may be able to choose what costs are included. Instead, a royalty that is a percentage of the net selling price is typically chosen. ⁷ Fifty-eight percent of the agreements in the SEC sample used a royalty based on the net selling price. ⁸ This means of achieving the split of profits from the intangible leads to a more easily verifiable number for the licensor.

Of the SEC agreements that have royalties based on the net selling price, 40 percent have a constant royalty rate. Because costs vary over time, a flat royalty rate will lead to a different profit split on a year-by-year basis. Therefore, an examination of the returns over the lifetime of the agreement is necessary to determine the true division of the profits.

Some agreements apparently attempt to even out the returns earned by varying the royalty rate. A number of different structures are possible. Some agreements have a royalty rate that declines over time; others are structured so that the rate rises and then falls. Thirty per cent of the SEC agreements that have royalties based on the net selling price vary the royalty rate over the years of the agreement.

The remaining 30 percent of agreements with royalties based on the net selling price vary the royalty rate based on total sales of the product. This structure may be most clearly tied to the returns that the licensor requires. It may also provide incentives to the licensee; this aspect will be discussed below.

Not all compensation packages are based on a percentage of the net selling price. Eighteen percent of the agreements in the SEC sample require a royalty per unit. Once again, the royalty per unit may be constant or it may change as more units are sold. In the SEC sample, 60 percent of the royalties per unit were constant, and 40 percent declined per unit as the number of units increased. The licensor may prefer to base the royalty on units sold instead of on net selling price if the licensee is contemplating discounts or if the intangible may be sold as part of a larger package, such as when software is distributed free of charge to the buyer of a computer. In the SEC sample, 80 percent of the royalty per unit agreements occurred in licenses for computer software. Computer software would seem to be the type of product for which "free" samples may be provided or which may be part of a package deal. Indeed, 67 percent of the sample agreements in the SIC code for "Computer Programming and Software" contain royalties per unit. ⁹

Some agreements combine advance or lump sum fees with a royalty based on sales or units. Different factors might explain these payments in different settings. If the licensee's country of incorporation limits the allowable royalty rate, an initial lump-sum fee might be used to ensure that the licensor earns the required return. ¹⁰ Alternatively, the goal of a front-end fee may be to lower the licensor's risk by ensuring a minimum return. In the SEC sample, 16 percent of the agreements with per unit or net selling price royalties also have initial lump sum fees.

Another means of lowering the licensor's risk is for the licensee to prepay a nonrefundable amount of money against which future royalty obligations are credited. In addition, a minimum payment may be due each year. If the total royalties paid are less than this amount, the licensee must pay the difference to the licensor. Forty-seven percent of the agreements with per unit or net selling price royalties contain this type of arrangement.

Finally, a lump sum fee may provide the sole compensation for the use of an intangible for a certain number of years. Twenty percent of the SEC agreements used a one time, lump sum payment or annual lump sum payments. Such an arrangement fixes the return that the licensor will receive. This payment scheme leaves the licensee to absorb all the variance in the amount earned. Just as in the minimum payment scheme, if the product is much less popular than expected, the licensee will absorb the loss. However, unlike the minimum payment scheme, if the product is much more popular than expected, the licensee will reap all of the unexpected rewards. This type of arrangement could provide a strong incentive to the licensee to concentrate resources on selling the product.

Forcing the licensee to absorb the risk may not be the only reason that lump sum fees are chosen. The licensor may have patented a new technique or instrument that the licensee wishes to use to attempt to create another product. In this case there is no readily apparent unit to be produced, nor is anything being sold initially. Therefore, a lump sum fee may be the only practical means of compensating the licensor for the use of the patent. Lump sum fees may also be used to settle disputes over patent infringement claims.

D. PROVISIONS WHICH MAY AFFECT RETURNS

Other clauses in the agreement, which do not explicitly affect payment, may affect the returns earned by the licensor and the licensee. For example, the licensor may require the licensee to perform a certain amount of marketing. This clause can be very specific, and may require that a certain amount be spent or that a certain percentage of the licensee's marketing expenditures be devoted to the licensor's product. ¹¹ Alternatively, the marketing or advertising clause may be open-ended, requiring that the licensee use its "best efforts." ¹² Although such a clause does not readily translate into a dollar figure, it potentially gives the licensor the ability to terminate the agreement or to file suit if unsatisfied with the results.

One might expect to find these clauses in licenses for products in which advertising plays a pivotal role in determining the success of the product. Indeed, in the SEC sample, these marketing or advertising clauses seem to be particularly prevalent in the SIC code for toilet preparations. Seventy-three percent of the agreements in the SEC sample that contain advertising clauses are for licenses with respect to clothing articles or toilet preparations. Once again, certain features of agreements seem to be product specific. This suggests that a comprehensive analysis of the marketplace for intangibles should ideally focus on individual product groups.

In addition to lowering the licensor's risk, these marketing clauses imply that the licensee is engaging in a significant economic activity beyond the manufacture and distribution of the good that embodies the intangible. One would expect that the performance of this additional activity would affect the returns that each party anticipated.

A major factor affecting the licensor's return from licensing the intangible is the amount of technical support required as a condition of the license. The total expense of transferring a technology to a licensee will depend on the technology and on the licensee's level of expertise. Transfer costs include the physical transfer costs of plans, specifications, and designs, as well as the cost of training the licensee to make use of them. Since the licensor has typically already created the product being licensed, the cost of transferring the technology may be the biggest resource cost to the licensor. Indeed, Contractor finds that the most important factor in determining the licensor's return on an agreement is the amount of technical services provided. ¹³ By carefully limiting the amount of service automatically provided, the licensor can minimize uncertainty of return from the transfer.

Additional technical support is sometimes provided on a time and expense basis. ¹⁴ The split between "free" technical support and additional support for which the license is charged varies depending on the circumstances. Additional detail would be necessary to test hypotheses concerning how expectations about technical support affect technical assistance provisions and how these provisions affect the whole licensing package. The SEC data reveal a variety of solutions to the technical assistance question. Some set a specific time period for "free" technical support. ¹⁵ Others require that technical assistance be reimbursed at cost, ¹⁶ at a fixed rate, ¹⁷ or at the lowest rate charged by the licensor to third parties. ¹⁸ As is true of the marketing clauses in a licensing agreement, a licensor may need to balance the desire to push all of the technical costs onto the licensee with the need to ensure that the licensed intangible is used productively.

E. PREPARING FOR SURPRISES

An arm's length license agreement is shaped by each party's expectations about costs, sales, and the overall profit potential from the use of the intangible. The parties' expectations may differ, and they may differ markedly from the actual profit experience with the product. Therefore, even if both parties are pleased with the royalty rate to be paid, the level of technical services to be provided, and any marketing clauses or clauses on market restrictions, it is possible that future events will leave one or both of the parties dissatisfied with the arrangement.

There are two types of surprises from which the parties may desire protection. One surprise occurs if further development of the intangible significantly improves the product's profitability. The other occurs if several years of actual profit experience lead to a change in expectations about future profitability.

The first surprise is of particular concern to the licensor. In order to insure against this risk, many license agreements contain "grant back" or "technology flowback" clauses. These clauses specify that the licensor receives, free of charge, any enhancements to the technology that are developed by the licensee. These clauses serve to discourage the licensee from doing its own R&D and potentially competing with the licensor. They further protect the licensor from losing out on serendipitous discoveries by the licensee. Caves et al. found grant back clauses in 43 percent of the 257 agreements they studied. However, the clauses appeared 76 percent of the time in agreements involving "dominant product" licensors. ¹⁹

The second type of surprise, changes in the parties' expectations about future profitability, can create problems from which both licensors and licensees may want relief. Termination clauses provide one kind of protection in these circumstances. In one agreement, a license for the use of a trade name, the licensee was required to meet certain sales targets. This type of arrangement allows the licensor to exit from the deal or renegotiate if the volume is insufficient to realize the expected returns from the intangble.

In other cases, termination clauses allowed the parties to end the contract, without cause, after giving notice. This safety valve may not lead to actual termination but instead may offer an opportunity for renegotiation if one of the parties thinks that its returns are inadequate. The structure of termination clauses varies. Clauses may allow immediate termination, or they may require several years' notice. Manufacturers' distributors can typically be dropped by the manufacturer to whom they are under contract on 30 days' notice. ²⁰ At the other end of the spectrum, some terminations without cause are tied to the length of a patent. However, not all agreements involving patent life specify such a long period before renegotiation is considered. Thirty-four percent of the SEC agreements provide for termination without cause for the licensee, while 21 percent allow the licensor to terminate without cause after a given period of time.

Some agreements have no termination clause except for cause. There are several situations in which a license might be likely to lack a termination clause. If the licensee was required to make a substantial initial investment in order to make use of the intangible, one can hypothesize that the licensee would not enter into the agreement if the licensor could easily pull out of the deal. Another type of agreement without a termination clause might be one involving the license of products, such as computer software, that have a very short lifespan. In this case the relevant life is so short that termination is not a useful option; both parties must choose correctly the first time. Related to this group are agreements that are scheduled to end after a specific, and relatively short, length of time. These agreements will automatically be renegotiated if the parties wish to extend the license.

Fifty-five percent of the SEC sample agreements allow no termination except for cause. However, of this subset, 22 percent are agreements that have a specified length of less than three years, and another six percent are agreements of five or six years in duration. These agreements seem to fall into the category of licenses with relatively short lengths.

Twenty-eight percent of the SEC sample agreements without a termination clause (16 percent of the total sample) were agreements with a duration of ten years or more. More information on these licenses would need to be gathered in order to test the hypothesis that extended licenses tend to exist when the licensee is required to make a substantial initial investment. ²¹

Of the remaining agreements with no termination clause, 22 percent were for agreements with lump sum payments and 22 percent were for agreements of indeterminate length. The latter group typically specified that the agreement lasted until the patent expired; these patent expiration dates were not readily available. ²²

The existence of termination clauses shows that companies are concerned about their ability to predict the total profits from the exploitation of an intangible. Regardless of the existence of termination clauses, agreements do get renegotiated. The frequency of renegotiation would give information about the "surprises" that occurred and the companies' ability to predict the outcome of a license agreement. Although it is not possible to make general statements about the overall frequency of renegotiations or terminations based on the sample of agreements we examined, some examples of renegotiations were found. A license to manufacture and sell clothing under a trademark was renegotiated in the second year of a six year term. The amended agreement provided the licensor with a royalty rate, based on net selling price, one percentage point higher than the original rate. However, the new agreement also lowered the percentage of the net selling price to be spent by the licensee on advertising by one-half a percentage point.

Other agreements provide clear evidence that the parties contemplated the possibility that renegotiation might be necessary. One agreement, with no termination clause, provides for renegotiation of the royalty rate after three years. At that time, "[B]oth the royalty rate and the scope of the Patented Portions will be renegotiated . . . [to] aggregate to not less than 1.5 percent and not more than 2.5 percent of the Royalty Portion selling price of such Licensed Products." Another agreement provides for renegotiation after certain events occur, instead of after a certain time period. This is a 15 year agreement which the licensee can terminate on six months' notice. It says: "Licensor or Licensee are unable at the Effective Date of this Agreement to value on a proportionate basis the future worth to Licensee of the rights presently owned by Licensor in the Technological Field. . . ." The agreement goes on to provide for current royalty payments and for additional payments, depending on the outcome of certain events.

This discussion of terminations and renegotiations shows that there is no single way of dealing with uncertainty. Ultimately, much more detailed analysis would have to be undertaken to determine what circumstances lead unrelated parties to renegotiate or terminate contracts.

F. CONCLUSIONS AND RECOMMENDATIONS

The agreements filed with the SEC and those analyzed in the academic work provide a body of information concerning arm's length transactions involving intangible property. This body of information points out key factors that should be considered when determining the proper allocation of income in related party situations. The SEC data and other sources will be further analyzed in greater detail following publication of this paper. For example, patterns might be disclosed between royalty rates and specific levels of technical assistance, or marketing expenditures, either in general or by specific industry groups. Moreover, a study of the prevalence of, and circumstances that trigger, termination or renegotiation clauses (as well as the results following exercise of the clause) might be helpful in determining when unrelated parties exercise these rights.

² "Material" is an accounting concept that describes items about which a prudent investor ought reasonably to be informed. For an explanation of a "material contract," see Item 10 of the Instructions to the Exhibit Table for Form 10-K, 17 C.F.R. section 229.601.

³ See, e.g., F. Contractor, International Technology Licensing, at Chapter 3 (1981).

⁴ J. Baranson, Technology and the Multinational 64 (1978).

⁵ Caves, Crookell, and Killinger, The Imperfect Market for Technology Licenses, 45 Oxford Bull. of Economics and Statistics 249, 258 (1983).

⁶ Contractor, supra n. 3, at 125.

⁷ The net selling price is typically the price charged by the licensee to unrelated parties on an f.o.b. factory basis after deduction for state and local sales taxes and, sometimes, after deduction for quantity discounts.

⁸ Recall that the sample size is small and was not chosen randomly. All percentages provided are purely illustrative and should not be accorded the weight one would give to a larger, randomly selected sample.

⁹ This product specific use of a certain type of royalty base is the type of information that one would hope to find in more situations after a thorough examination of the SEC data.

¹⁰ If the licensee's country of incorporation intends to limit the compensation flowing out of the country, attempts to provide a lump sum payment may only serve to shift the analysis. In addition to limiting the royalty rate, the country may also challenge the lump sum payment.

¹¹ One agreement states that:

[I]n each License Year during the term of this Agreement, Licensee shall expend a sum equal to 2% of the Net Sales of Licensed Products . . . for trade and consumer advertising of Licensed Products under the Licensed Trademarks. All such advertising shall be in accordance with the provisions of this Agreement. Licensee shall furnish Licensor with copies of each such advertisement and with proof of such advertising expenditure.

The agreement goes on to define advertising and to require that "such advertising shall have been submitted to Licensor and received its prior written approval."

¹² One such clause reads: "[Licensee] shall use its best efforts to document, package, market, distribute, advertise and promote the Use of the Software. All advertising and promotion . . . shall be undertaken at [the licensee's] expense. . . ."

¹³ Contractor, supra n. 3, at 123, n. 6.

¹⁴ Baranson, supra n. 4, at 65, n. 4.

¹⁵ A license for the design, use, and sale of laser accessories with a per unit royalty provides:

Upon [licensee's] request, [licensor] shall give or shall cause to be given to [licensee] such technical assistance and shall give or shall cause to be given to [licensee's] employees such training for 6 months after the date hereof as [licensee] may reasonably require in connection with the transfer of technology provided in the preceding paragraph. . . .

¹⁶ A license to manufacture and sell clothing using the licensor's trademark with a royalty based on net sales notes:

Licensor agrees to furnish technical aid to the licensee (including information concerning advertising, packaging and customer lists) if requested in writing, provided that the licensee pays all of the expenses for such aid. . . .

¹⁷ A 1985 license agreement to modify and sell software where the license makes regular, fixed royalty payments says:

[Licensor] agrees to provide technical assistance concerning the Software to licensee, upon request by Licensee, in the development of the Modified Software; provided, however, that in addition to all other sums payable under this Agreement, Licensee agrees to pay [licensor] the sum of $100.00 per hour for all labor provided by [licensor], plus reimbursement for all expenses incurred by [licensor] in providing such technical assistance to Licensee.

¹⁸ A license for the use of a new type of laser where the licensee provides fixed annual royalty payments requires:

[Licensor] shall make available such technical assistance as [licensee] may reasonably request for understanding or exploiting the Proprietary Technology at [licensor's] standard rates, and under terms that are no less favorable than those extended to any of its other customers.

¹⁹ Caves, supra n. 5, at 261, n. 5. A dominant product is one that accounts for more than 60% of a firm's sales. This classification relates only to the level of the firm's diversification. It does not make any distinction with regard to the overall profitability of the product. Id. at 252.

²⁰ Galante, Quickie Divorce Curbs Sought By Manufacturer's Distributors, Wall Street Journal, July 13, 1987, at 25.

²¹ In addition representative information on all licenses, regardless of term, that require the licensee to make a substantial investment would be necessary in order to test the hypothesis that a large initial investment by the licensee leads to a license of long duration.

²² These agreements could be of short duration; several of the other license agreements in the sample were for patents with only two or three years left before expiration.

APPENDIX E

EXAMPLES OF METHODS FOR VALUING TRANSFERS OF INTANGIBLES

PREAMBLE TO EXAMPLES

The examples that follow illustrate the principles and methods described in Chapter 11. They are intended to set forth what the Service and Treasury believe is an ideal application of these principles and methods to specific factual situations. They are intended to serve as guidance for taxpayers in planning their pricing transactions as well as for both taxpayers and the Service on audit. In large part, the types of information set forth are based upon information used by IEs and economists on audit and used by taxpayers or outside economists for planning purposes.

In general, it is expected that the amount of information about comparable transactions, rates of returns, and costs for a taxpayer's industry and claimed comparable transactions will be the greatest following a full examination of the taxpayer's return. But, as Chapter 3 makes clear, taxpayers have a burden to document contemporaneously, and to justify, their transfer pricing policies and their return positions. The Service and Treasury recognize the practicalities involved in locating and analyzing the type of information set forth in certain of these examples when transactions are planned or returns filed. In general, the taxpayer making a relatively minor investment would not be expected to have gathered and analyzed data outside of its own knowledge of its business affairs and those of its competitors. As Chapter 3 indicates, however, a taxpayer engaging in a transaction involving high profit intangibles should arguably be expected to gather and analyze the type of information set forth in certain of the examples, to the extent that it is available, contemporaneously with the transaction.

EXAMPLE 1: EXACT COMPARABLE

Hydrangea, Inc. is a U.S. developer, producer, and marketer of business software for personal computers. It has developed a new line of specialized accounting software that it expects to sell mainly in foreign markets.

Hydrangea expects that this product will have a life cycle similar to most other products in its line. Thus, it expects that this software will have a peak productive life of three to five years. If it is moderately successful, there will be a small, declining market after that, as obsolescence sets in. If the product is very successful, Hydrangea may decide to develop an enhanced, substantially modified version after the peak period, which it would treat as a new product.

Hydrangea has decided to serve the market in country F for this software by licensing it to an unrelated country F corporation, Fleur, with which it has had satisfactory dealings in the past. Specifically, Hydrangea and Fleur have negotiated a licensing agreement with the following terms. Fleur receives an exclusive license to market Hydrangea's product in country F and agrees to pay Hydrangea 20 percent of the net selling price for each copy it sells. Fleur agrees not to market competing products while the license is in effect. Fleur will market the product under its brand name and will perform the necessary work to modify the product to integrate it into its own line of accounting software. Hydrangea agrees to provide Fleur, for free, with any corrected, revised, and enhanced versions of the software that it releases publicly during the first four years. (Fleur and Hydrangea understand that, after that period, the latter is permitted to develop an enhanced, substantially modified product and to call for new negotiations with Fleur, find another licensee, or market the new product itself.) The agreement grants Fleur a perpetual license to the current product. During the first four years, neither party can terminate without cause; after that period, Fleur can terminate with six months' notice.

Hydrangea will serve the market in country B through its wholly owned local subsidiary, Royal Hydrangea. The markets in countries F and B are substantially similar in size, sophistication, and ability to use business software intended for personal computers. Royal Hydrangea performs the same functions as Fleur relating to marketing and distribution of accounting software. Thus, it will modify the product as necessary for local requirements; it has been and will continue to be responsible for marketing its products and developing its trademark; and, it maintains a distribution network, including a sales staff. For these reasons, Hydrangea concludes the external standards for using the Fleur agreement as an exact comparable are satisfied.

Hydrangea satisfies the internal standards by including all important features of the Fleur agreement in its agreement with Royal Hydrangea. Thus, the royalty is set at 20 percent of net selling price. The provisions concerning corrections and revisions are also, therefore, included, as are the provisions concerning duration and termination.

Each year, Hydrangea reexamines its related party arrangement to determine if the exact comparable approach is still valid. Specifically, it determines whether the two markets are still similar, and whether Fleur and Royal Hydrangea still perform similar functions. If these aspects of the external standards have substantially changed, or if Fleur terminates its agreement, Hydrangea must reestablish the appropriateness of its related party transaction, which may require adjusting the royalty rate.

EXAMPLE 2: UNAVAILABILITY OF COMPARABLES

A U.S. company has developed a unique good that it believes will capture 80 percent of the relevant market. The U.S. company plans to produce the good in the United States and to license the rights to production and sale for the rest of the world to its foreign subsidiary. The U.S. company can find no examples of situations in which an unrelated party licensed the rights to production for a product that captured such a large share of the market. Therefore, the company should use the arm's length return method in order to set the appropriate royalty rate with its foreign subsidiary.

EXAMPLE 3: INEXACT COMPARABLE

Shampoo Inc., a well known hair-care products corporation in the United States, plans to set up a subsidiary in country Z in order to introduce its line of products in country Z. The subsidiary will manufacture, distribute, and market the products using the Shampoo trademark. When planning the appropriate transfer price for the license, Shampoo officials started with the knowledge that one of their competitors, Condition Corp., licensed a line of hair-care products to an unrelated party in country Z, Lotions, Inc.

The Condition license covers the formulas for all of Condition's hair-care products as well as the Condition trademark. The license gives Lotions the right to manufacture, distribute, and market the licensed products. Terms include an exclusive license in country Z for a term of four years paying a royalty of four percent of the net selling price. The licensor agrees to provide the licensee with product formulations, scientific data, manufacturing know-how, marketing, public relations, and related assistance. The licensee must adhere to strict quality control standards in manufacturing, distribution, and marketing. The licensor has the right to inspect operations of the licensee to verify such quality. The licensee is prohibited from manufacturing, importing, or marketing competing products in country Z.

From the terms of the agreement it is clear that Lotions performs the same functions that Shampoo's subsidiary in country Z will perform. It is also clear that Condition provides the same type of services and quality control for Lotion's operations that Shampoo will provide for its subsidiary in country Z. It is anticipated that Shampoo's subsidiary will have a volume of sales similar to Lotion's once its operations are fully developed. Finally, Shampoo knows that the gross margin, (net sales - cost of sales)/net sales, on sales of Shampoo's products in the United States is similar to the gross margin achieved by Condition in the United States, which indicates that their manufacturing processes and sales activities have comparable efficiencies. Therefore, it is appropriate for Shampoo to set a royalty rate of four percent of the net selling price for a four year term.

EXAMPLE 4: LIKELY USE OF INEXACT COMPARABLES

Computers Inc., a U.S. software company that specializes in games, plans to acquire the rights to manufacture, market, and distribute a new computer game, Gizmo, created by its European subsidiary. Gizmo will be an addition to Computers' existing line of games. Computers Inc. projects that the total number of Gizmo copies distributed will be close to the industry average. Numerous third party licenses for computer software are available. Computers examines these licenses for appropriate inexact comparables and should be able to determine the appropriate royalty amount and other terms for its agreement with its affiliate from the third party agreements.

EXAMPLE 5: BASIC ARM'S LENGTH RETURN METHOD: U.S. IMPORTER AND DISTRIBUTOR

TravelFun is a large publicly traded foreign corporation with a U.S. subsidiary, TravelUS. TravelFun produces a unique recreational product using sophisticated and highly sought-after production technology. TravelUS imports the assembled product and distributes it under the Travel name. TravelUS has the exclusive right to develop the Travel name in the United States. Because of the importance of the intangibles, TravelUS must apply the rules governing the transfer of intangible property.

TravelFun does not license the Travel name to unrelated parties, nor does it allow unrelated parties to distribute the product. Therefore, no exact comparables are available. Similar products exist that could potentially serve as inexact comparables; however, none of them are sold to unrelated distributors. Therefore, when TravelFun sets up its policy for transfer prices of units sold into the United States for 1989 it uses the rate of return method. In order to apply the method properly, the following information is necessary: 1) a general description of the functions that TravelUS performs, 2) financial information on companies performing similar functions, and 3) analyses of appropriate rates of return.

1) FUNCTIONS OF TRAVELUS. TravelUS imports the product from TravelFun and distributes it to retailers. TravelUS is responsible for developing the marketing strategy in the United States.

2) COMPANIES PERFORMING SIMILAR FUNCTIONS. Initially, TravelFun seeks data about publicly traded independent operators of wholesale distribution businesses. A search of the appropriate SIC categories yields a number of companies that differ in the following ways: 1) some are importers, while others acquire their products in the United States; 2) some are distributors of final products, while others distribute parts; and, 3) some apply intensive marketing, while others do not. Examination of these firms' published financial information indicates that the sample should be narrowed to 16 firms in order to reflect more clearly the functions performed by TravelUS. TravelUS is an importer that distributes final products and performs an important marketing function. Each firm in the final sample has some combination of these characteristics. Balance sheet and income data are then collected for the sample of 16 companies.

3) ANALYSIS OF APPROPRIATE RATES OF RETURNS. The company evaluates the available infomation in order to determine the appropriate ratios on which to base its comparisons. Comparable asset data are not available for all firms in the sample. Therefore, an attempt must be made to determine a rate of return for TravelUS based on available cost data for the sample of firms. A number of ratios can be considered as a means of determining an appropriate return on costs. Possibilities include the ratio of gross profit to operating expenses (the Berry ratio), the ratio of operating income to the cost of sales and operating expenses, and the ratio of net pre-tax income to total expenses. The choice of the appropriate ratio will depend on the composition of the sample and the stability of the ratios over time.

For the sample of 16 companies, all of the ratios lead to similar results. TravelUS retains the information that supports this claim, but upon examination presents only the analysis using the Berry ratio. As defined above, the Berry Ratio is the ratio of gross profit to operating expense:

(Net Sales - Cost of Sales) / (Operating Expenses)

Net sales are total revenue from sales less cash discounts to customers for payment within a specified time. Cost of sales is also referred to as cost of goods sold, including freight charges. Operating expenses include selling expenses such as sales salaries and commissions, advertising and marketing expenses, depreciation expenses, supplies, office salaries, and payroll taxes. The major expense not included in either cost of goods sold or operating expenses is interest expense.

For the 16 firms in the sample the average Berry Ratio is 1.40 with a standard deviation of .15. (The minimum ratio was 1.17 and the maximum was 1.61.) TravelUS uses the average ratio, 1.40, in order to determine the payment that should be made to the parent. Additional information that will be necessary includes net sales, operating expenses, and cost of sales that are not included in the payment to the parent for the product.

TravelUS projects sales of 20,000 units, net sales revenue for 1989 of $100 million, and operating expenses of $30 million. Cost of Sales are projected to be $2 million plus transfer payments to TravelFun. Plugging this information into the equation for the Berry ratio yields:

1.40 = ($100 mil. - [$2 mil + 20,000 x ]) / $30 mil

Therefore, x, the transfer price paid for each unit, is $2800. TravelUS will pay TravelFun $2800 per unit of import and projects that it will pay TravelFun a total of $56 million in 1989.

EXAMPLE 6: BASIC ARM'S LENGTH RETURN METHOD: FOREIGN SUBSIDIARY SERVING LOCAL MARKET

Counter Inc., a U.S. corporation that specializes in over-the- counter drugs, plans to set up a subsidiary in country X. The subsidiary will manufacture, distribute, and market Counter's products in country Z. The manufacturing process is not particularly complex. The subsidiary will set up its own distribution network, which will be of average size for the industry. Further, it will perform its own marketing; however, because the subsidiary will, in general, sell "generic" products that will sell under its customers' brand names and trademarks, its marketing activities will involve contacting drug stores and other selling concerns, and not the development of a unique, consumer-level marketing intangible.

Counter's search for unrelated party licenses for comparable products proves fruitless. The search does yield a number of licenses in which the functions performed by each party are similar. The products are not similar enough to over-the-counter drugs to be classified as inexact comparables; however, the level of manufacturing, the type of distribution network, and the type of marketing performed in each case are similar. Therefore, Counter searches for information about the returns earned by each of these companies. Analysis of the income statements and balance sheets of the firms in the sample yields an average rate of return earned. This average can be used by Counter to determine the royalty rate to be paid by its subsidiary in country Z.

EXAMPLE 7: BASIC ARM'S LENGTH RETURN METHOD: FOREIGN SUBSIDIARY PRODUCING FOR U.S. MARKET

A U.S. corporation has developed and patented the formula for a new heart drug that has fewer potential side effects than any drug in existence. The U.S. corporation plans to manufacture the drug in a foreign subsidiary to be located in country Y, which has very low labor costs. The completed product will be returned to the parent for sale in the United States. In addition, some of the manufactured drug will be shipped from the manufacturing subsidiary to a marketing subsidiary in country X for sale in Europe. The parent wishes to fashion the transaction so that a royalty will be paid by the subsidiary to the parent for the right to manufacture and sell the drug. The parent and the subsidiary in X will then pay the manufacturing subsidiary for the finished product.

The following information is known or projected:

1. The drug will sell for $2.00 per pill.

2. The volume of sales in the United States in 1989 will be approximately 900 million pills.

3. The volume of sales in Europe in 1989 will be approximately 600 million pills.

4. Marketing and distribution costs in the United States are estimated to be $14.4 million.

5. Marketing and distribution costs incurred by the country X subsidiary are estimated to be $9.6 million.

6. The manufacturing subsidiary's costs will be as follows:

     Cost of Chemicals $110 million

A sample of manufacturers in locations with low labor costs shows that manufacturers earn an average rate of return on their manufacturing assets of 12 percent. The subsidiary's total manufacturing assets, the factory, cost $360 million. Prices should be structured so that the manufacturing subsidiary earns profits of $44.2 million ($43.2 million return on the factory asset plus the known $1 million return on cash).

If all of the drug were resold to the parent, the split between the cost of the final pill and the license payment would be unimportant as long as the correct amount of income remained allocated to the subsidiary. However, in this case the final product is also being sold to the subsidiary in country X, so the correct split is important.

Information on marketers and distributors of drugs shows that they earn approximately cost plus 25 percent, both in the United States and in country X. Based on the costs outlined above, the parent should earn net income of $3.6 million on its distribution and marketing activities, and the Country X subsidiary should earn net income of $2.4 million. Revenue from the sale of the heart drug will be approximately $1800 million in the United States and $1200 million in Europe. Therefore, the manufacturing subsidiary should charge $1.98 per pill.

Total revenue received by the manufacturing subsidiary will be $2971 million ($2970 million from sales and $1 million from interest income.) The royalty to be paid to the parent will be a percentage of the net sale price of $1.98. The correct royalty rate, r, can be determined by the following equation, which shows the manufacturing subsidiaries revenues and costs:

     Net Income = Total Revenue - Cost of Chemicals

Therefore, the appropriate royalty rate is 92.1 percent of the net selling price of $1.98.

EXAMPLE 8: LIKELY USE OF BASIC ARM'S LENGTH RETURN METHOD

A U.S. company manufactures electronic equipment for sale in the United States. The U.S. company designs the equipment and licenses the designs to its foreign subsidiary. The subsidiary assembles the circuit boards and other components for the products and sells them to the parent. For transfer pricing purposes, the parent searches for the rate of return earned by independent computer assembly operations in order to determine the amount of income that should be attributed to its foreign subsidiary.

EXAMPLE 9: LIKELY USE OF EITHER INEXACT COMPARABLES OR BASIC ARM'S LENGTH RETURN METHOD

A U.S. company manufactures and markets a line of sportswear. The company plans to introduce the same line of clothing to Europe through its European subsidiary. The subsidiary will manufacture, market, and distribute the casual wear using the parent's trademark. The clothing is marketed toward middle income consumers and is projected to sell at prices and earn a market share similar to several other brands which are marketed to this group. The parent has two options when setting its transfer pricing policy. If unrelated party licenses of trademarks for clothing can be located, then these inexact comparables can be used to establish appropriate terms for the license agreement. If information is available on the returns earned by unrelated parties that perform functions similar to the European subsidiary, then the rate of return method can be employed.

EXAMPLE 10: PROFIT SPLIT METHOD USING SPLIT OBSERVED IN ARM'S LENGTH TRANSACTION

ABC is a U.S. corporation that produces advanced machine tools. It maintains a large artificial intelligence research lab, which has made significant advances in computer vision. Recently, this work has begun to yield marketable products. Specifically, ABC has developed a "sighted" numerically controlled machine tool (NCMT) that can be programmed to recognize the pieces on which it should perform its fabrication tasks. ABC expects this new device to be a significant advance over competing NCMTs because the pieces will not have to be precisely aligned before the fabrication operations can be performed; therefore, the "sighted NCMT" should be much easier to operate and integrate into an assembly line. The key element in this advance is the software that allows the NCMT to determine the precise position and orientation of a piece placed on its operating deck. ABC has obtained worldwide patent protection for this software. As is true of most of ABC's products, the device must be substantially modified for each customer's specific application, and ABC maintains a large and expert engineering staff to accomplish this.

ABC projects that devices based on the new technology will eventually become an important source of revenues and profits for the company. During the first three to five years, ABC expects to have no significant competitors, and plans to market the devices to the high price, high mark-up, low volume, most technologically advanced segment of the NCMT market. After that period, as the technology becomes more common, ABC expects that sighted NCMTs will, in general, replace other types of NCMTs and that its lead will enable it to capture a significant share, perhaps 50 percent or more, of the overall NCMT market.

ABC-Europe is a wholly owned subsidiary of ABC incorporated in country X. All of ABC's products currently sold in Europe are produced and marketed by this company. ABC-Europe maintains its own research and engineering staffs and manufactures all of the devices it sells. A majority of the products in its line involve technology licensed from ABC, but a significant fraction depend on technology developed through its own research efforts. ABC-Europe performs all of the marketing for Europe, and its engineering staff performs the necessary development work of the devices for each customer.

ABC plans to transfer the European rights to exploit the software and associated technology for sighted NCMTs to ABC-Europe. ABC has no plans to license the technology to an unrelated party; therefore, no exact comparable is available. ABC has conducted a search for inexact comparables; although the search does turn up unrelated party transactions involving licenses of machine tool devices and patents, ABC has concluded that none of them can meet the standards for the inexact comparable method. Specifically, none of the potential comparables are for devices which involve profit margins as high as the sighted NCMTs will have in the short run, nor market shares as large as ABC anticipates having in the long run.

ABC next considers the basic arm's length return method. It concludes that ABC-Europe's activities in exploiting the sighted NCMT technology can be split into four functions: (a) conducting research to search for uses in the European market, (b) marketing the devices, including participating in trade shows, conducting demonstrations, and providing technical assistance (mass-market retail-level advertising is not necessary in this industry), (c) designing the specific devices to meet the requirements of each customer's application, and (d) manufacturing and distributing the devices.

ABC concludes that some but not all of these functions can be analyzed under the basic arm's length return method. Once each customer's design has been set, manufacture of the devices will not be such more complicated than current NCMTs, and ABC is familiar with firms that manufacture current-generation NCMTs according to others' designs. Therefore, ABC concludes that function (d) can be analyzed in this way; specifically, it concludes that a rate of return to operating assets of 16 percent is the average for firms that perform this function. Marketing is not a major activity in this industry, because the customers are extremely knowledgeable. ABC deals with firms that perform marketing functions for it; based on its knowledge of these firms, ABC concludes that a 20 percent ratio of income to costs is a reasonable way to value the contribution of function (b).

Functions (a) and (c), however, cannot be analyzed in this way. ABC-Europe's staff of scientists and engineers, while smaller than ABC's, is still one of the largest and most expert in Europe. ABC knows of no independent firm in the machine tool industry, in the United States or Europe, that would be able to conduct research, development, or design work as satisfactorily as or on a scale comparable to ABC-Europe. Therefore, ABC concludes that it would be inappropriate to value the contributions that ABC-Europe's performance of these functions will make toward selling sighted NCMTs in Europe by multiplying the assets employed by a rate of return, or multiplying the costs incurred by an income-to-costs ratio. In short, a profit-split approach is necessary.

Although ABC's search for comparables did not turn up appropriate licenses in the machine tool industry, other transactions between unrelated parties were found. For example, ABC obtained information about the following transaction: a group of professors, in partnership with their university, established a consortium to patent and exploit a process through which a new product can be produced by a genetic engineering technique. The consortium bargained at arm's length with several large chemical companies, and negotiated a licensing agreement with one of them. The licensee manufactures the product, tailors it to meet the specific needs of various groups of farmers, and markets it. The product has no significant competitors and has achieved widespread use in certain important agricultural applications. The licensee pays the university consortium a royalty of $7 per pound.

ABC next gathers information about the chemical company and the industry in which it operates. It is able to determine that the chemical company maintains a large staff of scientists and engineers which performs functions concerning the new product that are comparable to the research and development activities that ABC-Europe will perform. The chemical company undertakes significantly more marketing activities than will ABC-Europe, and the manufacturing process for the product is not comparable. Further, ABC is able to determine the following information: (a) the product sells for $27 per pound; (b) production costs are ten dollars per pound; (c) independent firms that produce chemicals using similar production techniques earn profits equal to 20 percent of costs; (d) marketing and distribution expenses are three dollars per pound, and (e) independent firms that perform similar marketing and distribution activities earn 33 percent of expenses.

This information allows ABC to determine the profit split between the basic technology contributed by the university consortium, on the one hand, and the research, development, and application activities and know-how contributed by the chemical company, on the other. Specifically, the latter's profit per pound, net of royalty and expenses, is seven dollars ($7 = $27 - $7 - $10 - $3). Of this amount, two dollars should be attributed to the manufacturing activity and one dollar to the marketing and distribution ($2 = $10 x 0.20, and $1 = $3 x 0.33). This leaves four dollars per unit as the return to the chemical company's know-how and skills as to R&D and application of technology to its customers' needs. The university's income is the seven dollars royalty. Therefore, the profit split is 64 percent (64 percent = 7 / (7 + 4)) for the licensor's basic technology and 36 percent (36 percent = 4 / (7 + 4)) for the intangibles employed by the licensee. (Note that the licensor and licensee each earn 50 percent of the total profits, since they each earn seven dollars per pound; however, 50 percent vs. 50 percent is not the relevant profit split for this situation, because it does not distinguish between the profits for the manufacturing and marketing functions.)

Finally, ABC is able to determine the proper arrangement for its license to ABC-Europe. There are many ways ABC could structure the arrangement. One would simply be to specify that ABC-Europe (a) determine gross profits from sales of sighted NCMTs (with gross profits defined as sales receipts minus manufacturing and marketing costs), (b) subtract 16 percent of the value of assets used in manufacturing the devices, (c) subtract 20 percent of the marketing costs, (d) subtract 36 percent of the remainder, and, finally, (e) remit the remaining amount to ABC as a royalty.

Alternatively, ABC could use additional information about ABC- Europe's future activities to set a more traditional licensing arrangement. ABC's projections for sales of sighted NCMTs in Europe during the first three years of operations include the following figures. The devices will sell, on average, for $100,000 each. Cost of production will be $56,000 and will require $50,000 of production assets per device. Marketing costs will be $5,000 per device. These projections imply that gross profits, defined as sales receipts minus manufacturing and marketing costs, equal $39,000 per machine. Of this amount, ABC-Europe should be allocated $8,000 for the manufacturing function and $1,000 for marketing ($8,000 = $50,000 x 0.16 and $1,000 = $5,000 x 0.20). Remaining profits are thus $30,000 per device. This amount should be split 64 percent to ABC and 36 percent to ABC- Europe; thus, ABC should be allocated $19,200 per device and ABC- Europe the remaining $10,800. A more traditional licensing arrangement, therefore, would require that ABC-Europe pay ABC a royalty equal to 19.2 percent of sales (19.2 percent = $19,200 / $100,000).

If ABC chooses the former type of arrangement, periodic adjustments to it are less likely to be necessary, because the allocation of income between ABC and its affiliate will automatically adjust to a large extent. ABC should reconsider periodically, however, whether the manufacturing rate of return to assets, the marketing income to cost ratio, and the profit split percentages are still appropriate. If ABC chooses the latter type of licensing arrangement, many more periodic adjustments to it will likely be necessary. Specifically, in addition to considering the preceding factors, ABC must determine if actual experiences depart from the projections enough to imply significant changes in the appropriate allocation of income. If so, ABC will have to recalculate the sales based royalty rate by substituting the relevant actual figures for the projections in the preceding paragraph.

EXAMPLE 11: PROFIT SPLIT METHOD USING INFORMATION ABOUT RELATIVE VALUES OF PREEXISTING INTANGIBLES

Teachem is a U.S. corporation that designs, produces, and markets educational toys in the U.S. It maintains a staff of educational psychologists and engineers to develop and design the toys, which are perceived as uniquely high quality and sell at a premium. Enseignerem is a wholly owned affiliate of Teachem and is incorporated in country F. It is one of the largest toy companies in Europe. It was, and still is, the largest toy company in country F when it was acquired by Teachem a number of years ago. Enseignerem incurs large advertising and other marketing costs to develop its trademark and reputation as a producer of high quality educational toys. It is responsible for its own marketing strategies, which are different in important respects from Teachem's marketing efforts in the United States. For example, Enseignerem maintains a large sales force that calls on schools and other institutions, and institutional sales account for a much larger proportion of its revenues.

Teachem has recently developed a new line of electronic toys and intends to license the European rights to the designs to Enseignerem. Teachem does not plan to license them to any unrelated parties; therefore, an exact comparable is not available. Further, Teachem expects that Enseignerem will be able to capture its usual high market share, especially in the institutional market, and will be able to sell the toys for its usual significant premiums over its competitors. For these reasons, Teachem decides that suitable inexact comparables will probably not be available.

Teachem next considers the basic arm's length return method. Enseignerem will perform three functions with respect to the new line of toys. It will be responsible for manufacturing them; specifically, it will negotiate contracts and supervise independent contract manufacturers who will actually produce the toys. Second, it will distribute them. Third, Enseignerem will be responsible for all aspects of marketing them.

The first two functions can be analyzed under the basic arm's length return method. Teachem projects that the new toys will sell for $100 each in Europe. Payments to the contract manufacturers will be approximately $40 per toy. Enseignerem has found that distribution costs, including transportation and costs of holding inventories, are usually one-half of production costs, and expects that the new line will be typical in this regard. Therefore, Teachem projects that distribution costs will be $20 per toy. Finally, Enseignerem expects to incur costs of four dollars per toy relating to the supervision of the contract manufacturers. These costs include the salaries of engineers who will be assigned to visit and test the contractors, and premiums for liability insurance.

In some of its product lines, Enseignerem employs contract manufacturers who are willing to distribute, as well as produce, the items. By comparing these contracts with those calling for manufacturing only, Teachem concludes that the independent firms that perform distribution earn a return for it equal to 25 percent of the distribution costs. Teachem therefore allocates five dollars per toy to Enseignerem for the distribution function ($5 = $20 x 0.25). Teachem also decides that a 25 percent income-to-costs ratio is appropriate for the first function, responsibility for manufacturing. Thus, Teachem allocates one dollar per toy to the affiliate as the return for performing it ($1 = $4 x 0.25).

Teachem decides that the affiliate's final function, marketing, cannot be analyzed by the basic method. Enseignerem is not planning to incur any significant costs attributable solely to the new toys. In general, it focuses its advertising on promoting the Enseignerem reputation rather than displaying a single item, and does not plan to issue a separate catalog or set up a separate sales force for the new line. Therefore, Teachem decides that it is not possible to identify or measure the costs or assets that Enseignerem will devote to the new product line. However, it would clearly be wrong to conclude that Enseignerem deserves no return for the marketing function, because its preexisting reputation, sales force, and knowledge of its market are crucial to the success of the new product line in Europe. Therefore, a profit split is necessary.

To summarize the analysis to this point, the toys are projected to earn a gross profit of $36 each ($36 = $100 - $40 - $20 - $4). Of this amount, six dollars should be allocated to Enseignerem for the functions analyzed with the basic method. Thus, $30 per toy is left as the combined return to Teachem's product designs and Enseignerem's trademark, sales force, and other marketing intangibles. The next step is to split these profits in a way that reflects the relative economic values of these sets of intangibles.

Teachem concludes that the new line of toys is similar to other lines that the corporate group has introduced in the past few years in terms of the importance of the underlying design relative to marketing intangibles. Specifically, the designs involved a typical amount of research and development effort and the toys will be marketed in ways similar to, and with similar intensity as, other products. Teachem analyzes its own performance record and educational toy industry information on the relative importance of design and marketing intangibles therein. Based on a good faith analysis of this data, Teachem concludes that it is reasonable to assign a relative value of the design intangibles equal to one-half the value of marketing intangibles. Accordingly, it allocates ten dollars of the $30 to be split to itself and the remaining $20 to Enseignerem.

Teachem can structure the arrangement in any form that achieves the appropriate allocation of income, ten dollars per toy to the parent and $20 to Enseignerem. Specifically, it could establish an agreement in which Enseignerem pays Teachem a royalty for the European rights to the product designs at a rate of ten percent of sales. In future years, Teachem must reexamine its arrangement and, if any key element in the analysis described above changes significantly, must adjust the royalty rate accordingly.

EXAMPLE 12: LIKELY USE OF PROFIT SPLIT METHOD

The research staff of a European company that manufactures and markets food products has just created a chemical compound that will alter the way that the human digestive system reacts to sugar. The company believes that by adding the compound to its products, the products will pass through the human digestive system without being absorbed. The compound is unique because it leaves the taste of the product unchanged. No information is known about the possible side effects of this compound. The company wants to use this discovery to offer a whole line of diet products. The European company has a U.S. subsidiary that presently manufactures and markets existing products in the United States. The U.S. subsidiary also has a research staff. Because the prime market for this new product is the weight-conscious United States, the parent licenses the compound to the U.S. subsidiary for development and for the extensive and expensive testing that will be necessary in order to obtain approval from the Food and Drug Administration. Because the product is unique and because the subsidiary performs such complex functions, the profit split arm's length return method is probably most appropriate.

EXAMPLE 13: PERIODIC ADJUSTMENTS TO REFLECT CHANGES IN FUNCTIONS

A U.S. corporation produces and markets widgets in the United States and it has a subsidiary in country X that produces and markets widgets in Europe. The U.S. parent is in the early stages of developing a new super-widget. In 1988 it is clear that this could be a major breakthrough in widget technology; however, the manufacturing process is still cumbersome. It is unclear whether the process can be developed to the point that it would be possible to mass-produce the super-widget. The U.S. parent believes that the team of employees at its subsidiary in country X is best suited for the time-consuming and expensive job of developing the process to produce the super-widget.

In determining an appropriate transfer price for the license of the technology, the parent can find no inexact comparable for the super-widget. Similarly, the basic arm's length return method is not feasible because neither party is performing standardized functions. Therefore, the U.S. parent attempts a profit split analysis.

Based on the best information available in 1988, the U.S. corporation predicts that the development process should be completed by 1994. An increasing number of super-widgets will be produced between 1988 and 1994; however, only in 1994 will true assembly-line style production be feasible. Based on an analysis of the relative costs incurred by the parent and by the subsidiary, and on an analysis of the relative returns earned by unrelated parties when risky products are jointly developed, a 50-50 profit split on the returns of the design of the super-widget is adopted by the parent.

By the end of 1989, as the parent is filing its 1989 tax returns and is rechecking its transfer price policy for 1990, super-widgets are being successfully mass-produced at close to the volume predicted for 1994. Instead of requiring the extended development process predicted two years earlier, establishing production was more similar to the effort necessary when adjusting production lines for improved versions of products. Accordingly, the parent adjusts its transfer price policy to a basic arm's length return analysis for its subsidiary in country X. Specifically, the parent determines the average rate of return earned by independent companies that manufacture a product similar in complexity to super-widgets. Because the parent is particularly cautious and feels it would be difficult to sustain its profit split for 1989, it also modifies the 1989 policy to a rate of return analysis. While at the outset of this transaction it appeared that the subsidiary in country X would be required to use significant intangibles of its own to establish the production process, the actual experience of the parties was that no unique intangibles were contributed by the subsidiary. The decrease of five years in the time expected to develop production to the 1994 level constitutes a significant change that requires an adjustment.

EXAMPLE 14: PERIODIC ADJUSTMENTS TO REFLECT CHANGES IN INDICATORS OF PROFITABILITY

A U.S. pharmaceutical company has patented the formula for a new anti-arthritic drug with fewer side effects that those in existence. The U.S. parent's subsidiary in country Y will manufacture the drug and market it worldwide. There are numerous third party licenses for the existing anti-arthritic drugs. The parent decides that these products are comparable because it feels that its product will be a close competitor to them, and will sell for a similar price and capture a similar market share. Specifically, it believes that its drug will capture approximately 15 percent of the market, as do several of the existing products. The parent uses the eight percent royalty on net selling price that is found in those licenses and adopts other significant feature of such licenses as well. For example, the length of the agreement is for the length of the patent.

The U.S. parent reviews its license with the subsidiary at the start of year two and finds that its drug has only an eight percent market share. However, the market share seems to be continuing to grow. Indeed, at the beginning of year three its market share is 16 percent and at the beginning of year four its market share is 21 percent. In each of these years the U.S. parent decides that the inexact comparable is still appropriate.

By the end of year four the popularity of this drug has skyrocketed and it captures 50 percent of the market. Since this share of the market is far beyond that captured by any of the third party licenses, it can no longer be assumed that the level of overall profitability for the product licensed to the related party is similar to that for the products licensed to unrelated parties. Specifically, to the extent that market share is an indication of the mark-up that can be charged on a product, the related party product, which captures 50 percent of the market, is probably much more profitable than products that capture only 15 percent of the market. Therefore, the present inexact comparables are no longer valid. A search for other inexact comparables fails to produce a license involving a similar market share. Therefore, the parent turns to a basic arm's length return analysis to determine what its subsidiary should earn.