Attorney Recommends Changes to Proposed Cost-Sharing Regs

Dear Mr. Commissioner:

The Internal Revenue Service ("IRS") published a notice of proposed rulemaking in the Federal Register on August 29, 2005 concerning the treatment of research and development ("R&D") cost sharing under section 482 of the Internal Revenue Code of 1986, as amended (the "Code"). This letter responds to the request for comments in the notice of proposed rulemaking.

                       TABLE OF CONTENTS

Section 482 states that the Secretary may, in the case of commonly owned or controlled organizations:

In 1986, Congress amended section 482 by adding a second sentence intended to revise, in a narrow, limited way, the section's application to intangible property:

The second case decided under the 1986 amendment involves issues arising in tax years 1996-1999 concerning R&D cost sharing. The court's opinion holds that the amendment "supplements and supports" the arm's length standard, not supplants it.⁴ The IRS had argued in court that the amendment "replaced" the arm's length standard.

It was not until 1968 that the IRS promulgated regulations providing any guidance on cost sharing.⁵ The 1968 regulation on cost sharing reads as follows:

The August 2005 proposals have greater detail and more requirements than the regulation in force today, running over 100 pages of rules as delivered to the Federal Register. Notwithstanding, or perhaps because of, all this detail the proposed regulations essentially define only one set of circumstances in which cost sharing arrangements ("CSAs") qualify under section 482. Otherwise taxpayers must try to rely on the other, more general, transfer pricing regulations for guidance when attempting to enter into CSAs that qualify under section 482.

The preamble to the regulations proposed in August 2005 explains why the IRS and the Department of the Treasury ("Treasury") believe that additional proposed regulations are necessary at this time to address issues involving R&D cost sharing. It appears from the preamble that the updated guidance provided by the proposed regulations is needed from the perspective of the IRS and the Treasury, mainly "to respond to arguments that have been encountered in the examination experience of the IRS under the existing regulations." Also important, the preamble explains that in drafting the regulations the IRS and the Treasury want to "dispel the misconception that cost sharing is a safe harbor." The preamble and the proposed regulations assert that a cost sharing arrangement covered by the regulations "must produce results consistent with the arm's length standard." (Emphasis supplied.)

The August 2005 proposals also appear to be motivated by concerns articulated in the preamble to the regulations proposed by the Treasury and the IRS in 2003 for services and intangibles: i.e., a desire to stem untaxed migrations of intangibles from the United States.⁹

In 2003, the IRS and Treasury also stated their desire to harmonize the section 482 regulations for services and other transactions, particularly transfers of intangible property. According to the IRS and the Treasury, such harmonization is, as is explained in the 2003 preamble, necessary:

II. INFORMATION ABOUT THE COMPANIES COMMENTING

The comments in this letter reflect the consensus of a diverse group of 16 multinational corporations (the "Transfer Pricing Discussion Group" or "Group"), and thus provide an objective and balanced assessment of the proposed regulations. The diversity of the companies is illustrated as follows:

• Some of the corporations are U.S.-based multinationals while others are foreign-based.

• The corporations include both exporters and importers.

• The corporations conduct research and development ("R&D") in the United States as well as in other countries.

• The corporations use different approaches for the management, development, ownership and use of intangible property, including R&D cost sharing, contract R&D services, centralized ownership of intangible property, decentralized ownership of intangible property and the cross-licensing of intangible property.

• Some of the corporations own intangible property in the United States while others own it either outside of the United States or both inside and outside of the United States.

• The corporations conduct business in a variety of industries, including the following:

• Section 482 retain its commitment to the arm's length standard; i.e., what would unrelated parties charge each other in the marketplace under similar circumstances and conditions.

• Regulations promulgated under section 482 ease the burden of administration currently imposed on taxpayers (and the IRS). Thus, the regulations should provide as many safe harbors as is possible, even as "exceptions" to the arm's length standard.

• The regulations should also take steps to reduce rather than increase the areas for potential controversy with the IRS, such as by using as many "bright-line" tests as is possible.

• Finally, in implementing the arm's length standard, the United States should adopt positions that minimize the risk of conflicts with generally accepted international transfer pricing norms, such as those reflected in the OECD Guidelines.¹⁰ This is important to avoid exposing taxpayers to the danger and cost of international double taxation of income.

III. OVERVIEW OF COMMENTS

In one sense, the regulations proposed in August 2005 could be viewed narrowly. That is, they nominally affect only those taxpayers who use, or might consider using, R&D cost sharing. However, by their nature the regulations also have to be evaluated broadly. As noted in the preamble, however, the Treasury and the IRS believe that the principles espoused in the proposed regulations are based upon and are consistent with the arm's length standard. Thus, as interpretations of the arm's length standard these proposals have to be evaluated broadly for they have implications well beyond just R&D cost sharing.

The subjects that the IRS and the Treasury are addressing in these proposed regulations are, particularly when viewed as interpretations of the arm's length standard, important to the international tax aspects of the Code as well as to international trade.

The proposed regulations involve difficult, challenging issues. The Treasury and the IRS have had these issues under consideration for a substantial period of time. The Group understands that the IRS and Treasury are concerned about the positions taken by certain taxpayers using cost sharing. The Treasury and the IRS should be acknowledged for the efforts they expended on this complex matter.

This letter focuses its comments on four new sets of principles proposed in the regulations that, in particular, appear to be important to the arm's length standard generally (i.e., not necessarily just for R&D cost sharing):

• Realistic Alternatives Principles

• R & D Work Force Compensation/Valuation Principles

• "Investor Model" Valuations, Discount Rate Principles, Etc.

• Periodic Adjustments Principles

In general, the Transfer Pricing Discussion Group believes that the just-mentioned principles in the proposed regulations deviate significantly from the arm's length standard and should be eliminated or substantially modified in any further IRS proposals on R&D cost sharing, particularly if they state they are following the arm's length standard. Also, due in part to their deviation from the arm's length standard, these four proposed principles should not, in the view of the Group, be used in any other transfer pricing areas, such as determining the arm's length compensation for the provisions of services or the transfer of intangibles.

As explained in the conclusion, the Group recommends that the Treasury and the IRS consider a different approach to R&D cost sharing, one that contains alternative methods and safe harbors. A different approach should be able to address the Treasury's and IRS's concerns about certain practices, while maintaining the principles of the traditional arm's length standard.

IV. DETAILED COMMENTS

The basic premise of the RA provision, as explained in the preamble to the proposed regulations, is that a controlled party, "like any rational investor, would not enter into an investment when a better alternative is available." The RA provision is based in this regard on a mistaken assumption of equivalence between CSAs and other types of intangibles arrangements.

As already noted, the preamble to the proposed regulations claims that the proposed regulations "dispel the misconception that cost sharing is a safe harbor." The proposed regulations further assert that the results under a CSA are consistent with an arm's length result under the general provisions of the 482 regulations.¹¹

An inference to be drawn from these two points is as follows: if the use of a realistic alternative analysis helps to determine an arm's length result for a CSA, then why wouldn't the IRS use the proposed RA provision outside of the CSA context? In fact, IRS economists are already trying to impose the concept in other contexts. There is nothing explicit in the proposed regulations to prevent IRS agents from using the RA provisions outside of the CSA context.

As explained further below, the RA position is objectionable for a variety of reasons:

• The type of analysis required by the RA provision raises administrative concerns and creates uncertainty because it gives the IRS the ability to engage in "what if" type second-guessing, based on hypothetical transactions that never occurred.

• The RA provision is inconsistent with general tax principles, as well as well-settled case law concerning the arm's length standard, by undermining or overriding the taxpayer's choice of a structure for a transaction even when it has substance.

• The RA provision is also contrary to the arm's length standard in that it is premised on assumptions that are at odds with real world practice, and as such are not representative of what parties do in unrelated party transactions.

• The proposal is contrary to well settled jurisprudence and will invite challenges and controversy.

As currently drafted, the RA provision raises several concerns about administrative issues. First, the RA provision mandates that a taxpayer must consider "alternative arrangements realistically available". The examples then go on to suggest that there should be "reasonably reliable" estimates as to the terms of such alternatives. (Example 1.) Other than these two points, the RA provisions provide no direct guidance as to how to determine a "realistic alternative." Is the reliability of an alternative determined by reference to the same analysis used to determine if a comparable uncontrolled transaction ("CUT") exists? If the CUT rules do not control the analysis, which rules do? Moreover, the RA provisions do not clearly indicate which party carries the burden of showing whether an alternative is realistic. If a taxpayer makes a reasonably diligent effort to show that a reliable realistic alternative either does not exist, or does exist and takes that alternative into account in setting the terms of its arrangement, there is nothing in the regulations that requires the IRS to respect that determination.

In brief, as currently drafted, the RA provisions would support the IRS developing, after the fact, its own view of what constitutes a realistic alternative and using that view to make an adjustment notwithstanding any analysis done by the taxpayer.

The concerns about the vagueness of the RA definition are compounded by the fact that the regulations suggest that any single alternative may be sufficient. Thus, arguably the IRS need only find "a single" alternative to use the RA provisions to make an audit adjustment. This suggests that the IRS can engage in "what if" second guessing and restructuring, rather than limiting its analysis to a determination of whether the arrangement chosen by the taxpayer provides an arm's length return.

The section 1.482-7(g)(2)(iv)(A) definition of what constitutes a realistic alternative is broad, vague and leads to uncertainty in application. At the same time, the road map for implementation of that concept set out in the examples in section 1.482-7(g)(2)(iv)(B) seems to be results-oriented and to drive the analysis to favor U.S. income. The examples set up a two step analysis. Step one calls for a comparison of net present values ("NPVs") to the major intangible contributor (i.e., does the NPV of the chosen arrangement equal the NPV of "any" realistic alternative). Step two looks at the return to the non-IP contributing party and limits that return to the routine functions to be performed by that party.

The combination of these two steps and the lack of definition as to what constitutes a realistic alternative leaves the IRS with broad ability to adjust and effectively disregard the taxpayer's arrangement. That is, if the IRS does not like the result of the taxpayer's arrangement, the IRS has wide latitude under the RA provision to contest the taxpayer's assessment of what does or does not constitute a realistic alternative. The IRS may either find another alternative that would suggest that a greater NPV is due to the IP contributor or dispute the taxpayer's assertion that an alternative exists and move to step two of the analysis and essentially remove any intangible return from the non-contributing party beyond a discount rate of return based on a broad-based discount rate.

Moreover, it is well established that the IRS cannot recast the form of a transaction chosen by the taxpayer if the substance of the transaction is consistent with the form, even if an economically similar result could be obtained through use of a different structure.¹⁴

The principles of Gregory and other similar cases have been recognized and applied in the case law under section 482 long before its 1986 amendment.¹⁵ As is explained above, the 1986 amendment to section 482 was narrow in scope and, according to the Podd decision, only changed the then current state of the tax law by allowing the value of transferred intangible property to be determined by taking into account facts arising after the time of the transfer. The Tax Court held in Xilinx that the 1986 amendment did not replace the arm's length standard, notwithstanding the IRS' litigation arguments to the contrary.

The 1986 amendment to section 482 most certainly did not repeal Gregory type principles or make it improper to take into account tax planning in establishing international business structures and the terms of transactions between affiliates with a view towards certain transfer pricing methods or results. As the Claims Court said in Merck:

In Bausch & Lomb, Sundstrand and Seagate the IRS argued, unsuccessfully, that the taxpayers' licenses to foreign manufacturing affiliates and similar arrangements should be recast to treat the foreign affiliates as contract manufacturers.²² For example, the IRS wanted to ignore a separate license of intangibles by Bausch & Lomb U.S. to Bausch & Lomb Ireland and the purchases by Bausch & Lomb U.S. of contact lenses produced by Bausch & Lomb Ireland using technology that it licensed from Bausch & Lomb U.S. The court disagreed with the IRS and respected the separate nature of the transactions. The court used a comparable uncontrolled price ("CUP") approach to determine the transfer prices for the contact lenses and a financial return analysis to determine the arm's length royalty to be paid back to Bausch & Lomb U.S. (Using today's terminology, the court's methodology for determining the appropriate royalty might be characterized as "CPM.")

It is difficult to square the conclusions of the courts in Bausch & Lomb and the other cases mentioned above with an IRS "RA" argument that the taxpayer would not accept the (otherwise) arm's length return of its chosen transaction (i.e., R&D cost sharing) because a "better alternative", such as a contract manufacturing or a license relationship, was available.

It is certainly the case that a well managed company that discovers an idea may consider a range of possible alternatives to determine the best way to take advantage of it. However, the RA provision assumes, with no explanation, that various alternatives for dealing with research and development for intangibles -- in particular cost sharing and self-development/licensing -- are equivalent when in fact they obviously are not. At a minimum they involve differences in risk diversification and capital requirements. Thus, from a business perspective, these would not be viewed as equivalent options. Cost sharing involves a trade off; i.e., reducing upfront risks and costs in exchange for lower returns on commercialization if the research pays off. In contrast, licensing involves taking on the full research, and possibly development, costs and risks, in exchange for reducing the commercialization risk and return.

Along the same lines and as discussed further in section IV, C, the RA provision uses a discount rate as a measure of what a party that invests in research would expect to earn. In fact, discount rates (which are frequently referred to in transactional analyses as "hurdle rates") are used to determine the minimum return that companies would expect to earn before they agree to enter into an investment. Except for debt-like structures, parties dealing at arm's length do not typically use the hurdle rate to limit returns from a transaction.

When a transaction involves substantial funding of research before a commercially ready product exists, it is highly unlikely that the IRS could find an arm's length transaction in which the party that agreed to pick up a share of the research expected to earn only a discount rate of return based on a minimum or hurdle rate. Put another way, if a party runs the risk of putting money into research or building of intangibles, they expect to share in the full upside. The RA provision assumes that a party dealing at arm's length would strike a deal with terms that would limit the return to a hurdle rate. This over-simplification is not supported in the real world of arm's length negotiations.

As noted above, the RA provision allows the IRS to change the terms of an intercompany arrangement based on a single "realistic alternative." Once again, in real world business dealings, companies do not negotiate the terms of business transactions using other deals as determinative bench marks. Certainly, if a company is negotiating a license for a patent it will agree to a royalty and upfront payment structure that are consistent with other transactions with which it is aware. However, the company will not necessarily forego a transaction simply because it cannot get a royalty rate within the customary range. There may be valid business reasons why a company might agree to a royalty below the customary range (or pay a royalty above that range). Furthermore, if the company does negotiate a royalty rate within the range, it will not forego the transaction because there is one (or some) deals with rates outside that range. In contrast, the RA provision suggests that parties dealing at arm's length would take such a position.

As noted in the recent Xilinx decision, IRS "theories" of what should or might happen under the arm's length standard cannot prevail over what actually happens in market place transactions. Moreover, the IRS does not have the latitude to issue mandatory rules designed to overturn well-established jurisprudence.

In Treas. Reg. section 1.482-1(f)(2)(ii) it is stated, in connection with the IRS considering taxpayer "alternatives", that the IRS will adjust for material differences between the taxpayer's actual transaction and an alternative being considered. If adjustments for material differences are faithfully made under this rule by the IRS, then its effect is unclear. However, the proposed RA rules do not mention the need to make adjustments for differences and mandate that certain consequences flow from the alternatives chosen by the IRS. Thus, the RA rules appear to differ substantially, both in their purpose and effect, from what is in the current regulations.

In brief, the RA provision should not be part of any regulation, R&D cost sharing or otherwise, that wants to be consistent with the arm's length standard.

The Group submits that when researchers contract out their services they typically do so on a service fee basis (which does not include any intangible return) and without any additional upfront "commitment" payment. If the proposed regulations are intended to require uniformity that related parties engaged in cost sharing compensate each other for the mere existence of their research teams, then in the experience of the Group, this is a theoretical position not supported by market place evidence. Such a requirement, if intended, is therefore inconsistent with the arm's length principle and should be eliminated.

The preamble's and proposed regulations' attempt to fit PCTs involving assembled research teams into the existing section 482 regulations for purposes of valuation (see proposed Treas. Reg. section 1.482-7(a)(2)), unfortunately creates confusion where clarity is badly needed. The appropriate "reference transaction" type (e.g., transfer of intangible or tangible property, provision on of services) and thus the appropriate transfer pricing method to use to determine the value for such a PCT is left to the taxpayer. In Example 2 of proposed Treas. Reg. section 1.482-7(b)(3)(viii), the act of making an assembled research team available to other cost sharing participants is described alternatively as a transfer of intangible property (e.g., a "commitment", an "existing integration")²⁵ and as a provision of services.

The availability of a trained research team simply does not fit within the definition of "intangible" used when applying the arm's length standard. According to Treas. Reg. section 1.482-4(b), the definition of "intangible" includes patents, inventions, designs, copyrights, trademarks, licenses, methods and several other types of property having substantial value based on their intellectual content or other intangible properties (see also Code section 936(h)(3)(b)). Furthermore, the current regulations also provide that intangible property is "independent of the services of any individual." Given that the asserted value of the research team as a contribution to a CSA is founded precisely on the participants' "access" to the team members' services, it is difficult to envision applying proposed Treas. Reg. section 1.482-4 in this context.

Past taxpayer attempts at characterizing a trained workforce in place as a discrete intangible asset susceptible to valuation have met with little success. The Service successfully argued in Ithaca Industries, Inc. v. Commissioner, 17 F.3d 684 (1994), aff'g 97 T.C. 253 (1991), that a trained workforce in place was incapable of being valued separately from goodwill for tax purposes because there is no way to predict how long it would continue to exist.²⁶

The Ithaca Industries case was only one of many such tax disputes flowing through the courts at the time that concerned the identification and valuation of acquired intangible assets based on perceived "resources or capabilities". Eventually, these difficult, time-consuming disputes were ended by legislative fiat when the enactment of Code section 197 rendered most questions concerning intangibles identification and valuation for purposes of amortization irrelevant.

Under the arm's length standard, the scope of "intangibles" should remain limited to the traditional realm of legally protectable intangible property found in (among other provisions) the current cost sharing regulations. The proposed regulations' attempt to expand the reach of PCTs to include assembled research teams, and other "resources or capabilities" similarly difficult to identify and value reliably, should be withdrawn or there will be new rounds of unnecessary controversy about the meaning of the arm's length standard.

While these concepts are generally used as tools in business, the proposed regulations use the tools improperly and out of context. Thus, as proposed, the regulations develop artificial benchmarks that are not commonly used in finance or in third party dealings, and lead to uneconomic results. The regulations take a significant step away from the arm's length standard by applying the concepts in such a way as to ignore the actual intercompany transactions and how risks and functions were actually borne, and recast them in a way that distorts the outcome.

While these proposed regulations apply only to R&D cost sharing, it is clearly possible that the concepts and measures introduced here will be misapplied in other types of transactions, leading to more departures from the arm's length standard.

The regulatory investor model requires that the valuation of the amount charged in a buy-in payment must be consistent with the assumption that each controlled participant's aggregate net investment in developing cost shared intangibles pursuant to the cost-sharing agreement is reasonably anticipated to earn a rate of return equal to the appropriate discount rate over the entire period of developing and exploiting the cost shared intangibles. This benchmark is not arm's length as it constrains the taxpayer to an artificial level of profit even when there is evidence that independent arm's length parties earn more than the discount rate in similar arrangements. In effect, the model results in a mandatory formulary method not consistent with the arm's length standard.

Whenever possible, the section 482 regulations should first rely on third party evidence, if available, and only rely on formulary methods when such data cannot be found or as part of a taxpayer safe harbor. In contrast, these proposals foreclose the possibility of using arm's length comparables and instead compel the taxpayer into prescribed artificial returns.

The underlying assumption throughout the regulations appears to be that cost sharing is an artificial arrangement, particularly if one participant is actively conducting R&D and the other is providing primarily financial resources, and therefore no arm's length benchmarks could, in IRS's view, exist to aid in determining the appropriate amount of income. However, cost sharing in this form has a parallel in venture capital investments. The returns earned by venture capitalists are typically far greater than the discount rate suggested by the regulations. Gathering evidence on various types of venture capital arrangements and their returns would be a relatively easy exercise for a taxpayer that had an interest in risk sharing for R&D.

The proposed regulations' inappropriate reliance on fixed measures rather than on arm's length evidence could, unfortunately, be extended to other areas of transfer pricing beyond cost sharing. As noted above, these measures are unsuitable for analyzing cost sharing type arrangements and therefore should have no place in analyzing other types of intercompany transactions.

In determining whether to make periodic adjustments, the IRS may consider whether the adjusted outcome more reliably reflects an arm's length result under all the relevant facts and circumstances, including any information known as of the determination date. In this regard, the regulations provide that the determination date is the date of the relevant determination by the IRS.

This section of the proposed regulations raises several questions about their consistency with the arm's length standard. First of all, unrelated parties do not ordinarily agree to an arrangement that provides one party with exclusive rights to alter the distribution of benefits or to reset the terms, let alone to decide unilaterally on the time for such a determination. Asymmetrically distributed rights as provided by the proposed regulations are inconsistent with the arm's length standard.

Typically third party agreements decide on fixed terms based on ex-ante expectations. For example, arm's length agreements carefully consider, ex ante, the specific intended usage of each party's contributions, although third parties do not necessarily actually conduct valuations of their contributions. Unrelated parties also at times agree, up front, to specific mechanisms for modification and termination of agreements.

The preamble states in its discussion of the investor model: "ex post interpretations of ex ante expectations are inherently unreliable and susceptible to abuse." This is hardly consistent with the proposed rules on periodic adjustments, since these allow the IRS, and only the IRS, to take the ex post view of the facts and circumstances and to choose which facts to consider.

Lastly, the proposed language states that the IRS "may" consider facts and circumstances in order to arrive at an assessment of the appropriate pricing. The arm's length standard would, in contrast, require the IRS to consider all relevant facts in order to arrive at an assessment of the appropriate pricing.

V. CONCLUSION

The proposed cost sharing regulations do not, with all due respect, achieve a principal goal set out in the preamble; i.e., to strengthen the application of the arm's length principle with respect to cost sharing. The rules, if they became final, would introduce a great deal of uncertainty as they deviate from the arm's length standard, not only as it has been interpreted in the United States, but also internationally. For the taxpayer, this implies a risk of double taxation. The rules, therefore, would effectively reduce the level of interest of multinational companies in investing in research and development in the United States.

As indicated in the beginning of these comments, the Transfer Pricing Discussion Group has focused its comments primarily on the arm's length standard as interpreted by four parts of the proposed regulations. Although the preamble states that the proposed regulations are consistent with the arm's length standard, the Group disagrees. In particular, the four sets of principles discussed in these comments deviate significantly, in the view of the Group, from the arm's length standard as it has been applied over the years. These regulatory principles are inconsistent with market place behavior and the financial results of transactions entered into by parties that are not related to each other.

The Group fully understands that the IRS and the Treasury believe that the detailed, extensive regulations on R&D cost sharing initially finalized in 1995 have been misinterpreted or misapplied by taxpayers in some circumstances. The purpose of these comments is neither to defend the R&D cost sharing practices of all taxpayers, nor to challenge the views of the IRS and the Treasury on whether some of those practices were inappropriate.

Regardless of the outcome of that debate, there should be room for the IRS and the Treasury to propose regulations focusing on R&D cost sharing that are, in law and in fact, consistent with the arm's length standard. As a reminder, one paragraph of regulations, set forth in section I above, seemed to provide a reasonable balance of IRS and taxpayer needs from 1968 through 1995. Perhaps a lesson to be learned from the current cost sharing regulations is that additional length and detail do not necessarily work out well for both taxpayers and the IRS.

The proposed regulations, unfortunately, add even more length and detail, with numerous adverse consequences, including defining a very narrow set of circumstances that only a few R&D cost sharers can meet. For all other costs sharers, the proposals effectively provide no meaningful guidance and, thus, leave them (and IRS auditors), with considerable uncertainty.

The Group recommends that the IRS and the Treasury consider a new regulatory approach to the topic of R&D cost sharing. The Group believes that an approach different from that of the August 2005 proposals could meet Treasury and IRS objectives while remaining consistent with the arm's length standard and providing guidance that is effective. First of all, new proposed regulations on R&D cost sharing could and should present alternative means of establishing whether a cost sharing arrangement satisfies the arm's length standard; i.e., different "specified methods." For instance, why should R&D cost sharing regulations not explicitly have an analogue of the "CUP" or "CUT" method for R&D cost sharing for those cost sharing arrangements between related parties that contain financial principles like those found in cost sharing arrangements between third parties? A "CUP" or "CUT" type method for R&D cost sharing would mean evaluating cost sharing between related parties based on market place transactions and in a manner that is consistent with the arm's length standard, as opposed to imposing new theories of how unrelated parties based on market place transactions and "should" in theory behave. As is the case for regulations addressing other types of intercompany transactions, reproposed cost sharing regulations should provide for adjustments for any material differences between the unrelated-party and related-party R&D cost sharing arrangements.

Further, the IRS has acknowledged in regulations that the arm's length standard can be met for intercompany transfers of goods, services and intangibles by both specified and unspecified methods. Why shouldn't the same principle apply explicitly to R&D cost sharing arrangements? The Group believes that the cost sharing regulations should have a provision allowing unspecified methods to be used for cost sharing.

In addition to providing a defined set of methods for cost sharing, the IRS could specify one or more safe harbor methods with their own conditions and requirements. An R&D cost sharing "safe harbor" could be helpful to companies for whom certainty is paramount to achieve that result, even if that means complying with more detailed IRS rules and with bright line, reasonable tests. To avoid disturbing the arm's length standard and creating the other problems cited here, any conditions, requirements or "methods" used to create this type of certainty for taxpayers, and ease of administration for the IRS, should be clearly labeled as part of a "safe harbor." In general, the IRS should be looking for ways to create safe harbors to ease the administrative burdens for both taxpayers and the IRS on section 482 matters.

Thus, there are different ways of curing the problems that the IRS sees in today's final regulations without distorting or violating principles of the arm's length standard that have been accepted, appropriately, for many years.

The Transfer Pricing Discussion Group looks forward to the opportunity to discuss the contents of this letter with you. You may contact the undersigned with any questions or comments.

² Tax Reform Act of 1986, Pub. L. No. 99-514, § 1231(e)(1); 1986-3 C.B. 2.

³Podd v. Comm'r, T.C. Memo. 1998-231, at 45 n.8. (Emphasis supplied.)

⁴Xilinx Inc. v. Comm'r, 125 T.C. No. 4, p. 31 (August 30, 2005)

⁵ T.D. 6952 (1968), as amended.

⁶ T.D. 8632 (1995), as amended.

⁷ T.D. 8670 (1996), as amended.

⁸ T.D. 8930 (2000) and T.D. 9008 (2003; corrected 2004) made certain focused changes to the regulations.

⁹ Preamble to REG-146893-02; REG-115037-00.

¹⁰Transfer Pricing Guidelines for Multinational Enterprises and Tax Administration, July 1, 1995, of the Organization for Economic Co-Operation and Development.

¹¹ Prop. Treas. Reg. section 1.482-7(h).

¹² Proposed Treas. Reg. section 1.482-7(g).

¹³Gregory v. Helvering, 293 U.S. 465 (1935).

¹⁴See, e.g., Esmark, Inc. v. Comm'r, 90 T.C. 171 (1988), aff'd per unpublished order 886 F.2d 1318 (7th Cir. 1989)(IRS attempt to invent transactions that did not take place overruled where form chosen by taxpayer was consistent with substance); Grove v. Comm'r, 490 F.2d 241 (2d Cir. 1973)(court refused to recast a shareholder's gift of stock to charity followed by issuer's redemption of that stock as a redemption from the donor followed by gift of cash proceeds), Comm'r v. Estate of Gilmore, 130 F.2d 791 (3d Cir. 1942)(court refused to allow IRS to recast otherwise tax free merger as a taxable transaction where transaction as structured by the taxpayer was consistent with the "spirit and purpose" of the reorganization provisions).

¹⁵See, e.g., UPS of Amer. v. Comm'r, 254 F.3d 1014 (11th Cir. 2001); Merck & Co. v. United States, 24 Cl. Ct. 73 (1991); Procter& Gamble Co. v. Comm'r, 95 T.C. 323 (1990); G.D. Searle & Co. v. Comm'r, 88 T.C. 252 (1987); Eli Lilly & Co. v. Comm'r, 84 T.C. 996 (1985); Hosp. Corp. of Amer. v. Comm'r, 81 T.C. 520 (1983).

¹⁶Merck, 24 C1. Ct. at 85.

¹⁷Nat'l Westminster Bank, PLC v. United States, 58 Fed. Cl. 491 (2003).

¹⁸ The provision at issue is paragraph (2) of Article 7 (Business Profits) of the Convention for the Avoidance of Double Taxation between the United States and the United Kingdom signed on December 31, 1975.

¹⁹Nat'l Westminster Bank, 58 Fed. Cl. at p. 498 (Emphasis supplied.)

²⁰Id. at 504. (Emphasis supplied.)

²¹Id. at 498. (Emphasis supplied.)

²²Bausch & Lomb, Inc. v. Comm'r, 92 T.C. 525 (1989), aff'd, 933 F.2d 1084 (2d Cir. 1991); Sundstrand Corp. v. Comm'r, 96 T.C. 226 (1991); Seagate Tech., Inc. v. Comm'r, 102 T.C. 149 (1994).

²³Bankers Trust New York Corp. v. United States, 225 F.3d 1368 (Fed. Cir. 2000).

²⁴ Treas. Reg. section 1.482-1(d)(3)(iv)(H); Treas. Reg. section 1.482-1(f)(2)(ii).

²⁵ See also 2005-40 I.R.B. at 629, ("the Treasury and the IRS believe that a contribution of such an experienced team in place would result in the contribution of intangible property within the meaning of § 1.482-4(b). . . .").

²⁶ Interestingly, the taxpayer attempted to value its workforce on the date of its transfer (and thus amortize the asset for tax purposes) with the help of statistical evidence of the rate of change in the membership of the workforce following the transfer. The court rejected this strategy based in part on the conclusion that, due to the small size of the taxpayer's workforce (only 5,000 employees!) one cannot reliably predict what portion of the asset would remain four years after the transfer.