I Introduction
A mature pioneer pharmaceutical product is one for which the primary exclusivityFootnote 1 covering the product has expired. That is, sellers of generic or follow-on versions of the product are able to obtain regulatory approval to sell bioequivalent products commercially. Even at the mature stage in the life cycle of the pioneer product, various forms of residual or secondary intellectual property are embodied in the product (“residual intellectual property”).
There are various commercial reasons for valuing the residual intellectual property including licensing, litigation, and tax considerations. For this paper, I focus on valuing the residual intellectual property to comply with tax regulations governing transactions between affiliates of a multinational company located in different countries (“controlled transactions”). The company risks tax authorities assessing penalties for inaccurate values. Consider controlled transactions for a multinational research-based pharmaceutical company (“Company”) in which its US affiliate (“Affiliate”) owns the relevant intellectual propertyFootnote 2 and its international affiliate (“International”), the aggregated entity consisting of all the Company’s entities involved with sales of mature products other than the Affiliate, manufactures the mature product outside of the US and sells it in countries throughout the world (“relevant controlled transactions”).Footnote 3 To report revenues and costs accurately from the US and foreign operations in the individual countries, International requires a licence at an arm’s length price from Affiliate to use the residual intellectual property.
My analysis follows the Internal Revenue Service Final Section 482 Regulations (TD 8552) for Intercompany Transfer Pricing, issued 1 July 1994 (“Section 482 regulations”).Footnote 4 The governing principle for valuing both tangible and intellectual property in the Section 482 regulations is the arm’s length standard; that is, the result that would have been realised had uncontrolled entities engaged in the same transaction under the same circumstances as in the controlled transaction.Footnote 5
To determine the arm’s length consideration due to the Affiliate for relevant intellectual property, it is most helpful to review internal Company documents and publicly available information.Footnote 6 Licence agreements negotiated between the Company and third parties are usually the most reliable and complete data available for determining arm’s length prices (eg, royalty rates). The Company’s agreements with third parties should be separated into three types of relevant intellectual property:
∙ trademarks or trade names (“trademarks”);
∙ intellectual property that was not a composition of matter patent (eg, process, method, or formulation patents; know-how; or technical information) (“non-composition of matter patent intellectual property”); and
∙ intellectual property related to generic products or embodied in products that are not covered by a valid patent claim in a given territory (“intellectual property related to generics or with no valid claim”).Footnote 7
Statistical tests such as Comparison-of-Means Tests should be applied to the pricing terms for the intellectual property in these various agreements.
The Company’s own behaviour in the marketplace for intellectual property generally represents the best data for measuring arm’s length prices. The comparable uncontrolled transaction (“CUT”) method as described in the Section 482 regulations is the best method for assessing the transfer of the relevant intellectual property. Applying the CUT method to determine ranges for the arm’s length transfer royalty rate yields:
∙ one range for trademark royalty rates; and
∙ another range for all other types of relevant intellectual property (ie, non-composition of matter intellectual property and intellectual property related to generics or with no valid claim).
Depending on the type(s) of relevant intellectual property embodied in a specific mature product, royalty rates in these ranges may be applied to determine an overall royalty rate to apply to International’s Net SalesFootnote 8 of the mature product. If a controlled transaction includes the transfer of both a trademark and another form of relevant intellectual property the arm’s length royalty rate range is additive.Footnote 9
Tax authorities do not accept central tendencies from general industry data as a reliable basis for assessing arm’s length prices. However, it is often helpful to compare the results obtained from the Company’s own behaviour with third parties in the marketplace for intellectual property with results of studies in the licensing literature on central tendencies for similar types of transfers in the pharmaceutical and chemical industries.
The remainder of this paper is organised as follows:
— Section II provides background information on the Company in the research-based pharmaceutical industry and the importance of intellectual property in the industry;
— Section III describes the pricing methods for intellectual property in the Section 482 regulations, discusses the best data, and explains the basis for concluding that the CUT method is the best method given the available data;
— Section IV discusses the statistical analysis of the best data and the application of the CUT method to determine the value of the residual intellectual property;
— Section V addresses royalty rates in the licensing literature;
— Section VI summarises the conclusions.
II Background
1 Company’s worldwide activities
The Company is a research-based, global pharmaceutical company focused on discovering, developing, manufacturing, marketing, and selling large and small molecule pharmaceutical products. The Company discovers and develops innovative products through research conducted in at various other locations around the world. The Company either identifies molecules at its own R&D facilities or licenses molecules for further development from third parties. During the development process, the Company evaluates the commercial or financial opportunity for a potential product and decides whether to continue development of the associated molecule.
If a molecule progresses through the development cycle, the company applies to a regulatory agency such as the US Food and Drug Administration (“FDA”) or the European Medicines Agency for approval to market the product in a country or region. When necessary, the Company undertakes planning for associated manufacturing and marketing activities. The Company manufactures bulk material and finished pharmaceutical products at plants throughout the world.
Prior to the date of expiration of the primary composition of matter patent (or other primary form of exclusivity), the Company develops plans to compete with potential generic versions of the product from new entrants who intend to enter the market upon the pioneer product’s loss of exclusivity. These plans may involve continuing to sell the brand version of the pioneer product and/or launching an authorised generic version of the brand product, or exiting from the marketplace altogether.
2 Pharmaceutical patent process
Title 35 of the US Code, known as the Patent Act, allows innovators to obtain patents on processes, machines, manufactures, and compositions of matter that are useful, novel, and non-obvious.Footnote 10 To receive a patent, an inventor must file a patent application including a specification about the invention that is sufficiently detailed. The patent application must also contain distinct, definite claims that set out the proprietary interest asserted by the inventor and establish no prior art. If the US Patent and Trademark Office (“PTO”) determines that the application fulfils these requirements, it will grant a patent. Granted patents give the patentee the right to exclude others from making, using, selling, offering to sell, or importing into the US the patented invention.Footnote 11 Parties who engage in those acts without the permission of the patent holder during the term of the patent can be held liable for infringement. Although issued patents enjoy a presumption of validity, accused infringers may assert that the patent is invalid or unenforceable for a variety of reasons.Footnote 12
Patent protection in the US lasts for 20 years from the date the application is filed.Footnote 13 This period of exclusivity allows innovators or their assignees (including research based pharmaceutical companies) to recover the costs of investing in risky R&D.Footnote 14 In most industries, a patent filing is delayed until the associated product is ready for commercial sale.Footnote 15 In the pharmaceutical industry, a company will often begin the patent process for a unique, novel, and non-obvious technology while the underlying pharmaceutical product is still in development. The patent application is typically filed at the PTO before clinical trials are completed.Footnote 16 Given the separate regulatory approval process for a pharmaceutical product, the window of time that a pharmaceutical product may be legally distributed for commercial sale prior to patent expiration (and potential generic entry) is typically much shorter than the statutory patent term.Footnote 17 After the composition of matter patent for a pioneer pharmaceutical product expires, other companies may obtain regulatory approval to commercialise a generic version of the product assuming they can document the bioequivalence of the generic product to the originator’s product without infringing any other patents the innovator lists in the Orange Book as covering the pioneer product that are still in effect.Footnote 18
3 Relevant intellectual property
Brand versions of pioneer pharmaceutical products are rarely protected by a single form of exclusivity on the underlying active ingredient (eg, composition of matter patent), but instead are covered by multiple patents. These other secondary patents cover innovative aspects associated with a product such as coating technologies, formulation technologies, or methods of use. Although secondary patents usually do not provide the same degree of protection as the primary form of exclusivity, they are often filed after the composition of matter patents. They will have patent terms that extend beyond the patent term for the composition of matter patent. As a result, they can provide some additional protection for the brand product.Footnote 19 In effect, a secondary patent potentially extends the commercial life of the brand product by making it more difficult for generic companies to obtain regulatory approval to commercialise a generic.Footnote 20
In some cases, the innovator of the brand product develops product specific know-how, trade secrets, or technical information on identifying sources of supply for active pharmaceutical ingredients used in the product, manufacturing techniques associated with scaling a manufacturing process to commercial quantities, uses of the product in certain patient populations, or other factors associated with the product that may not be available publicly. This form of intellectual property evolves from the efforts that the innovator expended in developing the pharmaceutical product. Even though an innovator does not always obtain a patent, the intellectual property in the form of a trade secret provides advantages to the innovator after the primary exclusivity has expired. The innovator should receive a consideration for developing this intellectual property.
After the primary exclusivity for a brand version of the pioneer pharmaceutical product expires and a generic pharmaceutical launches its product, the brand pharmaceutical product may still be prescribed by physicians and preferred by patients due to their familiarity with the product and the perceived reliability that comes with the first-hand knowledge of the brand pharmaceutical product and the performance over years of being prescribed, dispensed, reimbursed, and consumed. In such instances, the associated trademark provides some value because the familiarity and knowledge it conveys contribute to continued usage and sales. For these reasons, trademarks are secondary forms of intellectual property that still retain value after the primary form of exclusivity expires. The owner of the trademarks is due a consideration.Footnote 21
Generally, it is possible to identify several common types of secondary intellectual property from licence agreements internal to the Company and in the public domain. These include, but are not limited to know-how, trademarks, trade names, technology, processes, systems, techniques, technical information, methods, technical assistance, formulations, designs, information, data, and intermediate patents. I refer to these types of intellectual property as residual intellectual property. Given the diverse nature of these types of intellectual property, I placed them into three general categories:
∙ trademarks, including trademarks, trade names, and copyrights, which are defined by the PTO as a “word, name, symbol, or device that is used in trade with goods to indicate the source of the goods and to distinguish them from the goods of others”;Footnote 22
∙ non-composition of matter intellectual property including know-how, technology, systems, processes, formulations, designs, information, data, technical information, technical assistance, trade secrets, and methods; and
∙ intellectual property associated with pharmaceutical products sold in countries without patent protection and for generic pharmaceutical products, which I refer to as intellectual property related to generics or with no valid claim.Footnote 23
Generally, the second category includes all information that represents the accumulated knowledge, skills, and experiences that could assist in the manufacture of a product and confer competitive advantages.
III Identifying the best method
In this section, I describe the methods for valuing intellectual or intangible property identified in the Section 482 regulations. I discuss the data I reviewed and explain the basis for my conclusion that the CUT method is the best method.
1 The Section 482 regulations – intellectual property
The Section 482 regulations specify the following methods for determining the arm’s length result for the transfer of intellectual property:
— CUT method;
— comparable profits method (“CPM”); and
— comparable and residual profit split methods.
The regulations also permit the use of unspecified methods.
There is no preferred method among those specified. Rather, whether the result of a controlled transaction constitutes an arm’s length result must be evaluated using a method or methods selected under the “best method rule”.Footnote 24 The best method is determined based on the facts and circumstances of the controlled transaction including the degree of comparability between the controlled and uncontrolled transactions, the completeness and accuracy of the underlying data, the reliability of the assumptions, the sensitivity of the results to deficiencies in the data and assumptions, and the confirmation of the results by another method.Footnote 25
a Comparable uncontrolled transaction method
The CUT method “evaluates whether the amount charged for a controlled transfer of intangible property was arm’s length by reference to the amount charged in a comparable uncontrolled transaction”.Footnote 26 To be considered comparable under the CUT method, the intellectual property involved in the controlled and uncontrolled transactions must be used in connection with similar products or processes within the same general industry and must demonstrate similar profit potential.Footnote 27 Profit potential is most reliably measured by calculating the net present value (“NPV”) of the benefits to be realised as a result of the transfer.Footnote 28 The circumstances of the controlled and uncontrolled transactions must be comparable.Footnote 29
b Comparable profits method
The second method specified in the Section 482 regulations is the CPM, which establishes arm’s length operating profits by reference to the profitability of comparable third parties engaged in similar business activities under similar circumstances.Footnote 30 Under this method, an appropriate profit level indicatorFootnote 31 is calculated from comparable companies’ data and applied to the tested party.Footnote 32 The reliability of the method depends on the degree of comparability between the tested party and the uncontrolled parties.
c Profit split methods
The Section 482 regulations also allow a profit split approach, which takes the form of either a comparable profit split method (“CPSM”) or a residual profit split method (“RPSM”).Footnote 33 Under these approaches, an arm’s length result is determined by comparing the relative contributions made by each of the controlled participants to the overall operating income of a transaction or business activityFootnote 34 and allocating returns based on the relative value of these contributions. The relative value assigned to such contributions must take into account functions performed, risks assumed, and resources employed by the controlled party in relation to any activity involving the transfer of the intellectual property.Footnote 35
i Comparable Profit Split Method
The CPSM may be applied when reliable income and cost data exist for uncontrolled companies comparable to the controlled party in the relevant business activity.Footnote 36 Factors that could affect prices or profits must be evaluated in establishing the comparability of transactions under the CPSM. These factors are: functions performed, contractual terms, risks assumed, economic conditions, and property or services transferred.Footnote 37 Once comparable uncontrolled transactions are identified, an arm’s length result is determined by applying the relevant profit split for the uncontrolled transactions to the combined operating profits of the controlled transaction.
ii Residual Profit Split Method
The RPSM uses a two-step approach to determine an arm’s length allocation of operating profits or losses for both routine contributions and non-routine intellectual property. First, RPSM requires an allocation of “operating income to each party to the controlled transactions to provide a market return for its routine contributions to the relevant business activity”.Footnote 38 Second, the RPSM requires that the resulting residual profits attributable to non-routine intellectual property be allocated among the controlled parties based upon the relative value of contributions of non-routine intellectual property to the overall enterprise.Footnote 39
d Unspecified methods
If an unspecified method provides a more reliable arm’s length result under the best method rule than any of the three specified methods, it should be applied to determine the true taxable income of a controlled party. The unspecified method must meet the same comparability and quality of data standards discussed with regard to the specified methods.Footnote 40
2 Data available
Selecting among the methods described above to determine the best method to apply to a controlled transaction depends on the data available. To identify comparable uncontrolled transactions that transferred relevant intellectual property in the pharmaceutical industry, I suggest searching the following sources:
— internal company files;
— Windhover’s Pharmaceutical Strategic Alliances ® ;
— RoyaltySource ®;
— RoyaltyStat ® ;
— Compact Disclosure; and
— Edgar ® Online.
In searching these sources, focus on identifying arm’s length licence agreements or uncontrolled transactions with royalty rates expressed as a percentage of sales paid to the innovator or patent owner for exclusive rights to relevant intellectual property; that is, trademarks, non-composition of matter patent intellectual property, and/or intellectual property related to generics or with no valid claim. These data are the best data for the transfer pricing analysis of the relevant controlled transactions.
3 The best method
The CUT method is the best method for determining the arm’s length consideration due Affiliate given the available data. The comparability factors for the CUT include: functions performed, contractual terms, risks assumed, economic conditions, and property or services transferred.Footnote 41
a CUT method
To apply the CUT method, the Section 482 regulations require that I identify uncontrolled transactions that are comparable to the controlled transaction based on the nature of the intellectual property, profit potential, and circumstances.Footnote 42 The uncontrolled transactions I identified involve the transfer of intellectual property for pharmaceutical products, specifying the consideration due to the licensor during the late stages of the products’ lifecycles. The potential profits for pharmaceutical products are often similar once the primary patents covering the products have expired and generic versions of these products have been approved.Footnote 43 The Company’s agreements for the transfer of intellectual property embodied in mature pharmaceutical products are the most complete and accurate data available for determining the arm’s length royalty rates for the intellectual property embodied in a product at a mature stage. Given these data, the CUT method is the best method under the Section 482 regulations. In cases where the profit potential of a mature product is extraordinarily high or low or the circumstances surrounding the product are unique, a subset of the best data may be most appropriate. The data I identified and the CUT method are the best data and best method, respectively.
b CPM
The CPM “evaluates whether the amount charged in a controlled transaction is arm’s length based on objective measures of profitability (profit level indicators) derived from uncontrolled taxpayers that engage in similar business activities under similar circumstances”.Footnote 44 Whereas the CUT method relies on the results from comparable transactions between unrelated parties, the CPM is based on the profitability of uncontrolled companies that engage in similar business activities under similar circumstances.Footnote 45 Given that applying the CPM involves using profitability measures based on the operating profits of uncontrolled companies, the reliability of CPM as an arm’s length measure may be reduced by factors such as varying cost structures, differences in business experience, or differences in management efficiency.Footnote 46 The data underlying a CPM analysis are not as reliable as the data used in a CUT analysis.
In addition, the Section 482 regulations require adjustments to the uncontrolled companies’ operating profits for any such differences.Footnote 47 In general, the tested party for a CPM analysis is “the least complex of the controlled taxpayers and will not own valuable intangible property or unique assets that distinguish it from potential uncontrolled comparables”.Footnote 48 The complex nature of the parties involved in the Company’s intercompany transactions and the other third party uncontrolled transactions I identified would likely require that adjustments be made to the operating profits of the uncontrolled companies. Since the reliability of any method is reduced as the number of adjustments increases,Footnote 49 the CPM analysis would not be as reliable as the CUT method in this case.
c Profit split methods
i CPSM
The CPSM may be applied when there exist reliable income and cost data for uncontrolled companies with similar transactions, activities, and operating profits to those exhibited by the controlled taxpayer in the relevant business activity.Footnote 50 Factors that could affect prices or profits must be evaluated in establishing the comparability of transactions under the CPSM. These factors are: functions performed, contractual terms, risks assumed, economic conditions, and property or services transferred.Footnote 51 Once comparable uncontrolled transactions are identified, an arm’s length result is determined by applying the relevant profit split for the uncontrolled transactions to the combined operating profits of the controlled transaction. Given that revenue and expense data related to the uncontrolled transactions were not available, the CPSM is not the best method in this case.
ii RPSM
The first step in an RPSM analysis is to determine the appropriate market returns for routine contributions to relevant business functions by comparing the returns to an uncontrolled company.Footnote 52 This step is subject to the same data and comparability considerations as those involved in a CPM analysis. These data and comparability considerations make CPM less reliable than the CUT method. Since the same difficulty exists for the RPSM, it is also less reliable than the CUT method. The second step in the RPSM is to determine the amount of residual profits attributable to each of the controlled parties based on the relative value of non-routine intellectual property contributed by each party to the overall enterprise.Footnote 53 This step does not rely on uncontrolled benchmarks. Under the best method rule, “[d]ata based on the results of transactions between unrelated parties provides the most objective basis for determining whether the results of a controlled transaction are arm’s length”.Footnote 54 Given that I have identified reliable uncontrolled transactions, the RPSM is not the best method available for determining the arm’s length result in this case.Footnote 55
d Unspecified methods
Even though an unspecified method may be applied, the CUT method provides reliable and accurate results in this case given the available data. It is unnecessary to employ an unspecified method in this case.
IV Applying the best data and best method
1 Statistical analysis of the best data
a Descriptive statistics
Descriptive statistics based on the midpoint royalty rate for each of the observations of the Company’s agreements with third parties are usually helpful.Footnote 56 The range of royalty rates from the observations and the interquartile range of royalty rates are relevant. Prepare a histogram with the distribution of royalty rates in the database, which illustrates royalty rates for the relevant intellectual property. Change in the value of the relevant intellectual property over time is sometimes raised by tax authorities. To anticipate this question, construct a plot of the data by the year of the agreement with the associated trend line. In my experience, there does not appear to be a trend between royalty rates and time.
There are three main types of intellectual property embodied in mature pharmaceutical products:
— trademarks;
— non-composition of matter patent intellectual property; and
— intellectual property related to generics or with no valid claim.
Separate the observations of royalty rates applied to sales for exclusive rights to these categories of relevant intellectual property and provide the descriptive statistics for each category.
b Comparison-of-means
Investigate whether the differences in the mean royalty rates for certain subsets of relevant intellectual property are statistically different from each other. The variables to investigate included licensor and licensee type (eg, company, university, individual), source of the data, type of intellectual property, territory, as well as whether the agreement contained supply provisions,Footnote 57 separate compensation terms for a composition of matter patent(s),Footnote 58 or additional forms of compensation (eg, upfront payment or milestone payments).Footnote 59 Apply the Comparison-of-Means test to determine whether the royalty ratesFootnote 60 for the rights to relevant intellectual property were significantly impacted by these variables.Footnote 61 Based on the results of these analyses, determine whether the variables I considered were statistically insignificant.
Given that patents are distinct from trademarks from both legal and economic perspectives, it may be possible to test whether royalty rates for non-composition of matter patents and royalty rates for intellectual property related to generics or with no valid claim differ from the royalty rates in the database for trademarks. The null hypothesis is that there is no difference in the means of the royalty rates for residual patents and royalty rates for trademarks. Construct a regression model with two dummy variables, one for intellectual property in the residual patent agreements and one for trademarks. For the first dummy variable, assign a “1” to each observation containing royalties for rights to intellectual property for non-composition of matter patent intellectual property or royalty rates for intellectual property related to generics or with no valid claim and a “0” for each of the remaining observations. For the second dummy variable, assign a “1” to each observation containing royalties for exclusive rights to a trademark and a “0” for each of the remaining observations. Perform Ordinary Least Squares (“OLS”) regression analysis using a multivariate linear regression model where the dependent variable is the royalty rate and the independent variables are the dummy variables.Footnote 62 The equation is:
— Y equals Royalty Rate;
— β equals the constant term;
— α equals the slope parameter for the associated x variable;
— x i equals the dummy variable related to intellectual property for non-composition of matter patent intellectual property or royalty rates for intellectual property related to generics or with no valid claim;
— x ii equals the dummy variable for a trademark; and
— ε equals the error term.Footnote 63
Depending on the P-value for the Trademark dummy variable, accept or reject the null hypothesis at a standard confidence level. Based on my experience, the type of relevant intellectual property is a statistically significant explanatory variable for determining the royalty rate in this type of model and that the difference between the means of the samples of royalty rates is statistically different than zero. The mean royalty rate for trademarks is likely statistically different from the mean royalty rate for the combined data on non-composition of matter patents and intellectual property related to generics or with no valid claim. That is, the mean royalty rate for intellectual property related to a trademark is typically lower than the mean royalty rate for non-composition of matter intellectual property or the royalty rate for intellectual property related to generics or with no valid claim. Trademarks are less valuable forms of protection in arm’s length negotiations. In fact, a survey of pharmaceutical R&D managers reported that trademarks are less important than process patents for protecting the results of R&D.Footnote 64
2 Determining the arm’s length price based on the CUT method
Assume for the remainder of this analysis that the royalty rates for intellectual property for trademarks are statistically different from the royalty rates for non-composition of matter patent intellectual property and intellectual property related to generics or with no valid claim. Applying the CUT method to determine an arm’s length transfer price (royalty rate) for an exclusive licence to two types of intellectual property embodied in mature pharmaceutical products requires constructing two arm’s length ranges: one range for trademarks, and another range for non-composition of matter patent intellectual property and for generics or with no valid claim.
V Comparing the results to literature on licensing
To put the results using the CUT method in perspective, it is helpful to search licensing literature for information on royalty rates in the pharmaceutical and chemical industries for limited intellectual property protection. Sources such as les Nouvelles, a journal published by the Licensing Executives Society are relevant for this task. Although not generally accepted by tax authorities as the primary analysis, such studies provide benchmarks for assessing the results using accepted methods.
VI Conclusions
To determine a general transfer pricing methodology for intellectual property embodied in mature pharmaceuticals, review internal Company documents and publicly available data. Focus on licence agreements with distinct royalty terms for the relevant types of intellectual property embodied in mature pharmaceutical products. The observations identified from the Company’s agreements with third parties generally provide the most reliable and complete data available for determining an arm’s length range. The CUT method emerges as the best method for determining an arm’s length transfer price (ie, royalty rate) for the relevant intellectual property under the Section 482 regulations. Apply the CUT method to determine an arm’s length transfer prices (range of royalty rates) due Affiliate from International for exclusive licence separately for trademarks, and for non-composition of matter patent intellectual property and for generics or with no valid claim.
In some cases, the Affiliate may own a trademark and another form of relevant intellectual property associated with the mature pharmaceutical product. In such a controlled transaction, the rates identified above should be aggregated.
Tax authorities have generally been reluctant to accept transfer prices based on industry benchmarks as the primary data. However, it is helpful to compare the results obtained by the approach described in this paper with studies in the licensing literature on central tendencies for similar types of transfers in the pharmaceutical and chemical industries.
