PHARMACEUTICAL COMPANIES URGE ELIMINATION OF PRIORITY GIVEN TO CPI METHOD.

Mossinghoff further writes that the comparable profit interval (CPI) is not usable in the pharmaceutical industry when the developer of the intangible and the manufacturing and marketing affiliates all have valuable intangibles. The CPI has no industry convergence, he writes, because the relative factors radically differ even within the same industry. Mossinghoff urges elimination of the priority given to the CPI methodology.

We also request that there be a public hearing on the proposed regulations.

Sincerely,

The following comments are submitted on behalf of the Pharmaceutical Manufacturers Association in response to the request for comments accompanying the filing of proposed regulations in the Federal Register on January 30, 1992 (57 Fed. Reg. 3517), implementing Congressional amendments to section 482. The Pharmaceutical Manufacturers Association is a nonprofit scientific and professional organization of more than 100 firms that discover, develop and produce prescription drugs and biological products in the United States. The Association's members produce most of the prescription drugs used in the United States and a substantial portion of the medicines used throughout the world.

The proposed revisions of the section 482 regulations are occasioned by a 1986 amendment to section 482, the IRS\Treasury review of inter-company pricing issues in the White Paper of 1988, intervening court decisions, and continuing Congressional and administrative interest in inter-company pricing. The cumulative result of this continued emphasis is a product calling for substantial change in existing methodologies. We believe that it is important to point out that the overwhelming number of inter-company pricing transactions that are audited by the IRS under existing law are accepted, or that agreement on an adjustment is reached, without resort to prolonged controversy or litigation. In our judgment, the proposed regulations will cause an unjustifiable increase in controversy and in litigated cases. Most intangible and tangible transfers do not require the detailed and burdensome analysis that would be required by the proposed regulations. We think that in revising existing regulations more recognition should be given to the fact that in most instances present law does work and that major controversies are the exception.

We also believe that the proposed regulations do not, in their present form, represent an improvement in the rules for administering inter-company pricing under the tax laws of the U.S. The proposed regulations, which focus on after the fact profitability, do not provide taxpayers guidance for planning or comfort that they will be in compliance with a future IRS hindsight evaluation of their pricing. Indeed, the limited methodologies that are available as a practical matter under the proposed regulations provide less guidance than existing regulations. As discussed below, the proposed rules concentrate upon a methodology for transferred intangibles that is not suitable for the worldwide production and distribution of prescription drugs produced by the pharmaceutical industry. We also believe that there are aspects of the proposed regulations that do not reflect what independent parties would do in transactions with each other and, therefore, do not represent an arm's length standard.

We are also concerned by a continued pattern of U.S. unilateral innovation in tax rules that directly affect tax relations with other countries without adequate consultation and development of a consensus among the countries. The result is to leave multinational corporations in the middle, facing potential controversy, litigation, and uncertainty. The concepts in the proposed regulations which are not used in establishing prices with third parties will be viewed by other countries as a departure from the internationally recognized arm's length standard and will be challenged. This will almost certainly expose companies to a higher risk of double taxation, of increased legal cost in the U.S. and abroad, and to retaliation from other taxing jurisdictions.

We have the following specific comments on the details of the proposed regulations:

1. THE COMPARABLE PROFIT INTERVAL

Among the most significant changes that should be made in the proposed regulations is to place the comparable profit interval (CPI) methodology in its proper perspective. For the pharmaceutical industry, the CPI, as proposed, is simply not usable where the developer of the intangibles and manufacturing and marketing affiliates all have valuable intangibles. The CPI has no industry convergence because the relative factors, e.g., proper identification of an industry or industry sub-group, product life cycles, product mix, varying functions performed, and other factors radically differ even within the same industry. Some pharmaceutical firms are vertically integrated whereby they produce all items in the chain of production while other well known firms contract for all or some production with third parties.

Discovering a new medical therapy and bringing it to market is very costly and time consuming. The industry estimates that it costs approximately $230 million and takes about 10 years to discover and market a new drug. The risks are great and a few successes must finance past failures and future efforts. There are few businesses that are comparable. Because of the specific economics of the industry and the functions performed by different entities, we do not believe that there are data on operations that most correspond to the "products and functions of uncontrolled taxpayers for which reliable data is available" and that this problem will not be solved, as the regulations suggest, by increasing the scope of the applicable business classification to "as broad a category of products as is necessary to obtain reliable data."

Attachment A illustrates the lack of convergence in profit level indicators among pharmaceutical companies. An individual product analysis would be likely to show even greater dispersion. A sampling distortion will be introduced if the sample consists of companies that are different in their line of business, functions, risks and economic conditions, including product life cycles. Comparisons in foreign markets will be difficult because of different accounting conventions. Even under U.S. accounting rules, the CPI can vastly differ when alternative acceptable accounting methods are selected (e.g., LIFO v. FIFO; accelerated v. straight line depreciation, elective capitalization of costs, election of differing lines for amortization, etc.). This means that alternative and more traditional section 482 methodologies to the CPI are more appropriate.

The fact is that arm's length inter-company pricing is ultimately a question of facts and circumstances and no single formula or method of economic measurement, such as a CPI analysis, will fit all businesses and all transactions. The CPI is simply one method for testing the results of inter-company pricing, suitable for certain situations, e.g., relatively simple structures where, for example, there are intangibles in only one party. CPI is not a method used by independent third parties in negotiating proposed transactions. Generally, this type of data is not publicly available. As such, parties will negotiate based on a gross profit basis. CPI information, if product specific, may represent an anti-trust violation.

We strongly urge that for the above and the following reasons the regulations be revised to eliminate the priority given to the CPI methodology and that the methodology itself be revised:

A. The CPI is not inherently superior to any other measure of comparability be it average industry royalty rates or the use of inexact comparable prices. As noted above, it is more suitable to testing certain factual patterns than to others.

B. The CPI test should not have priority over the Comparable Adjustable Transaction (CAT) method. The latter method can be highly probative of arm's length prices and should be allowed to stand on its own.

C. There is an inordinate burden imposed by requiring a search for comparisons of data on the profitability of third parties which simply do not exist in the public domain. Such information generally will not be able to be secured because of confidentiality and competitive reasons and would never exist on a prospective basis for use in filing tax returns. By moving away from traditional inter- company pricing the regulations add significant cost burdens for retention of new staff or outside consultants not otherwise required to conduct the business.

D. Even in the case of other products where a CPI might be more relevant for comparisons, the taxpayer should be able to demonstrate why the taxpayer's profitability indicators differed from averages. As presently cast, the CPI methodology does not permit that to be done. Yet there are readily apparent market factors that will result in different profitability levels. The proposed regulations do not address the lack of relevancy and comparability of the data. A CPI is determined by available data which by its very nature may be comprised of broad industry averages. These averages will fail to recognize the many intangible assets of a particularly successful company including good management and location savings, or the lack of such elements in a poorly run company.

E. The "tested party" mechanics, as set forth in the proposed regulations, present a serious problem for licensing intangibles by U.S. based multinationals if applied to test foreign affiliates on a company-by-company basis since the mechanical result may dictate different royalty rates for the same intangibles licensed into different countries. This would be a wholly unacceptable result for foreign tax administrations and highly disruptive of business arrangements.

2. PERIODIC ADJUSTMENTS

The proposed regulations assert the right of the IRS to require periodic adjustments in royalty payments based upon ex post facto developments. This is not an arm's length standard since it is not what independent parties normally provide for in their agreements and such U.S. initiated adjustments will not be recognized by foreign tax authorities. The result will be uncertainty for the pricing of U.S. products and increased double taxation for U.S. companies.

The proposed exceptions to this rule for periodic adjustments are so limited as to be without significance. The proposed regulations have not even attempted to narrow the potential application of this requirement but have left it as a rule applicable in all situations for all intangibles and all products. We strongly urge that any such rule be confined to situations where the results had a substantial degree of foreseeability. We note that to the extent any periodic adjustment requirement is retained, such adjustments must also be applicable where they have a favorable result for the taxpayer.

3. PROFIT SPLITS

We believe that where the parties to a transaction both have significant intangibles a profit split analysis will often be the most appropriate inter-company pricing method. The proposed regulations describe two types of profit splits: an Overall Profit Split and a Residual Profit Split. Regrettably, the data required by the proposed regulations to implement any profit split make it improbable that taxpayers can make such tests. The proposed regulations require data equivalent to unconsolidated income statements and balance sheets for uncontrolled parties whose functions and risks parallel that of the taxpayer's controlled group. Finding data and making an analysis that would isolate transactions of unrelated parties that, as required by the regulations, structure operations in the same form, serve the same market, and perform functions similar to those of the taxpayer would be a very rare prospect.

The regulations should recognize an internal profit split method. Profit split analysis is commonly done by many U.S. corporations for projecting and establishing transfer prices between related parties and is generally accepted by foreign governments in determining whether transfer prices are reasonable.

4. ADDITIONAL COMMENTS ON THE INTANGIBLE AND TANGIBLE PROPERTY RULES

A. We believe that the concept embodied in the proposed regulations of looking not merely at one year for testing inter- company pricing results is correct and should be retained. Revised regulations should contemplate a longer span of years.

B. We believe that the proposed regulations contain a very important recognition that there is a range of possible pricing outcomes rather than one "correct" price.

C. We support the decision to eliminate the priority of methodologies for pricing of tangible property. We would hope that the wisdom of eliminating hierarchial priorities would prevail throughout the regulations, including the CPI. We believe that the right methodology is the one most appropriate to the particular facts and circumstances. The U.S. was instrumental in the worldwide acceptance of methodologies implementing the arm's length standard, as evidenced in the guidelines issued by the OECD. Those useful principles, such as the resale price method, should not be abandoned in favor of an overriding unilaterally adopted concept. Diminishing the role of the resale price and other accepted pricing methods cannot be justified as carrying out legislative intent or arm's length pricing standards.

D. We believe that the proposed regulations unwisely restrict the utilization of inexact comparables which in many cases may be the most probative information available. For example, while it is observable that every chemical compound has unique properties, royalty rates within a range should be acceptable measures for many similar compounds, with only the anticipated "blockbuster" discovery being possibly outside such a range.

E. The proposed regulations have largely ignored the effect of exchange rate differences in using comparable economic measures.

F. We believe that clarification is required with respect to the pronouncement that the pricing concepts for intangibles are applicable wherever material intangible property is incorporated in a tangible product.

G. Taxpayers today are faced with very onerous section 482 penalties. Tests that are applied on a retrospective basis do not afford any planning assistance to taxpayers, who, in good faith, wish to avoid future adjustments.

5. COST SHARING

We believe that the proposed cost sharing regulations have generally responded constructively to concerns expressed about the cost sharing discussion in the 1988 IRS/Treasury Department White Paper. This is reflected in providing flexibility in the scope of research, participants, and geographical rights. We believe that the Treasury Department should try to make cost sharing agreements relatively easy to use as a way of reducing transfer pricing disputes.

The proposed regulations assign a priority to sharing of costs based upon profitability (similar to the emphasis on the CPI in other pricing contexts). Sharing costs on the basis of existing profitability data in many cases will be an appropriate basis for cost sharing. We concur with the regulations looking back at profitability over several years as providing a sensible mechanism for establishing prospective cost shares. However, requiring use of the current year's profitability could present a practical difficulty for timely information and should not be required.

The proposed regulations recognize -- but not sufficiently because of the ultimate testing role assigned to profitability -- that other methods can also be appropriate for establishing cost shares, such as actual and anticipated sales, units of production and any other measure that "reasonably predicts the benefits to be shared". For example, profitability is less useful for start up operations, projects where there is a narrower scope of research, and for any situation where there is difficulty in identifying and isolating specific profits of an entity to be the bench mark for a targeted cost shared research project. Pharmaceutical products may take a dozen years to bring to market, if they are successful at all. Cost sharing agreements which are entered into, therefore, by definition must be based on anticipated results from the product or products included in the cost sharing. Anticipated results can take into account experience in the industry as well as specific company profitability. Cost shares based on ratios supported by documentary evidence, including ratios based upon a particular company's experience or forecasted projections for a particular product, should be accepted. Once the cost share is set for a particular product, it should not be changed unless there is some change in circumstances or forecast for the market.

Depending upon the scope of the cost sharing agreement, cost shares could be different for each product or could be done on an aggregate basis if various products in research are included in the cost share. The important thing is that the cost sharing should be established up front by a reasonable method. If a methodology utilized by a taxpayer is reasonable in that objective, it should not be upset by an after the fact, actual profitability test.

While we agree that recent profitability data can be a very appropriate method reasonably to predict benefits to be shared, and therefore an appropriate basis for sharing costs, the current approach of using actual profitability to adjust the prior cost share and/or determine the cost sharing is not valid and does not work in this industry. Adjustments made after the product has been approved for research that was done 5, 10 or 15 years earlier make no sense, present difficult tracing problems, and will not be accepted by foreign governments. If the data from immediately prior periods (adjusted annually) used to fix the shares year-by-year meet the reasonable predictability test, the allocation of costs should stand. Independent parties would not retroactively adjust cost shares based on later profits. There are many reasons for subsequent changes in profitability that do not alter the reasonableness of using the original data for assigning cost shares.

With respect to other aspects of the proposed cost sharing rules:

A. We remain concerned that the "buy-in" costs continue to be a barrier to wider use of cost sharing agreements. As elsewhere under section 482, the regulations should state that third party comparables should be used as primary evidence of a proper buy-in.

B. We do not believe that if a party has borne its appropriate share of costs in return for its share of benefits that it should be precluded from transferring rights to independent parties provided that such potential exploitation was contemplated in projecting the benefits.

ATTACHMENT A

PROFIT LEVEL INDICATORS FOR INTEGRATED PHARMACEUTICAL COMPANIES AND GENERIC PHARMACEUTICAL COMPANIES

The attached tables were developed using Form 10-K for integrated and generic pharmaceutical companies. Where segmented information for the pharmaceutical segment was provided within the Form 10-K, only that information was used in constructing the profit level indicators (PLI).

The proposed regulations imply that there will be a convergence of constructive operating incomes when actual financial data of public companies is examined and that there will be an appropriate data cluster to test whether a particular taxpayer's pricing results differ from others performing similar functions and should be adjusted. The attached tables show that actual financial data may not converge.

Following the proposed regulations under section 482, two PLIs were developed and applied to the Tested Pharmaceutical Company's sales and assets in order to attain constructive operating incomes. The data and graphs on tables 1.1 and 2.1 clearly indicate that convergence could not be attained. There is much greater dispersion than in the hypothetical data used in the examples in the proposed regulations. Furthermore, constructive operating income vary widely; there is no single cluster to be used over other data. It is believed these examples, unlike those used within the proposed regulations, are a better reflection of what will happen when trying to apply the CPI method based upon information that would be available to most companies.

[charts & tables omitted]