Vol. 37.1 The Review of Litigation- The Brief Fall 2018
Seeking Comparable Transactions in Patent and Tax
by Susan C. Morse*
Most business firms do not go around licensing their crown jewel intellectual property to unrelated third parties. This presents a problem for both patent law and tax law. In patent litigation, setting damages for a reasonable royalty under Georgia Pacific invites the use of a benchmark royalty rate that would have been agreed to had the litigating parties negotiated a market rate in advance. This counterfactual analysis repeats in tax law when firms allocate taxable income among affiliates located in different tax jurisdictions. Transfer pricing rules similarly seek a price, such as a royalty, that would have been agreed to had the related affiliates negotiated a market rate as adverse, or “arm’s length,” parties.
In their article, Tax Solutions to Patent Damages, Jennifer Blouin and Melissa Wasserman argue that tax transfer prices can provide some of the data needed to set patent litigation damages. One could also ask the converse, which is whether patent litigation outcomes can provide some data that tax transfer pricing needs. If patent law looks to tax transfer prices, it sees the advantage that the tax transfer prices are set ex ante when IP developed by one affiliate was first used by another affiliate. This roughly aligns with patent law’s touchstone of a “hypothetical negotiation” that produces an “ex ante” license. If tax law looks to patent law, it sees the advantage that patent damages emerge from an adversarial process. Patent damages may be set ex post, but their validity is bolstered by the fact that they are contested.
Blouin and Wasserman argue that parties and courts should make use of the large body of tax transfer price information to help support reasonable royalty calculations in patent damages cases. Perhaps so. But transfer pricing data is messy. Using tax transfer prices sets for parties and courts the challenging task of understanding the prices in context. The risk exists that the analysis will fail because of the weight of its own complexity.
Case law contacts between patent litigation and transfer pricing
Existing case law shows that patent damages law and transfer pricing law already encounter each other. For instance, in Medtronic, a tax transfer pricing position taken by a taxpayer was upheld in part because it could be justified with a calculation that started with a patent damages award. The case illustrates various factors, such as exclusivity and scope, that might distinguish a transfer price from a patent damage calculation, even where both amounts can be expressed as a royalty based on a percentage of sales.
Courts in patent damages cases come across transfer pricing evidence, may permit its discovery and even admit it into evidence. But they have been reluctant to allow tax transfer prices to materially influence patent damages awards. Usually, courts refuse to rely at all on evidence from tax transfer prices. Even though the tax rule aims to ensure that these prices are "arm's length," the patent case law responds that the prices are not reliable, reasonable royalty evidence, because the agreements were not struck between “competitor[s].” Patent law is right to be skeptical.
Problems with transfer prices
Transfer pricing rules are not particularly good at ensuring that prices between related parties are the prices that would have been arrived at if the parties had not been related. In part, this is because no coherent theory underpins the instructions given in transfer pricing law. In part, it is because of doctrinal soft spots. In part, it is because of administrative limitations.
First, the theory problem. Transfer pricing is concerned with prices within controlled, integrated firms. The theory of the firm teaches that the reason that such firms keep intellectual property closely held within the firm is tha such a strategy is more profitable. Holding IP within a firm minimizes hold-up problems that arise when someone outside the firm, such as an individual who helped develop IP, claims a right to it. The net result is that a firm may earn more profit if it owns its own IP. Some of the profit earned within a controlled group may not exist in any unrelated comparable. Prices based on a narrow conception of “arm’s length” that proceeds from available unrelated party prices cannot explain how to allocate this residual profit.
Second, the doctrine problem. There are soft spots in transfer pricing law that result from practical problems like poor drafting and, perhaps, agency capture. I.R.C. § 482 anticipates that the commissioner will require royalties "commensurate with income" between related affiliates. This strong language appears to require related parties to pay higher royalties for more successful products and ex post adjustments to accomplish that aim. But one tax regulation allows “platform contribution transactions,” also known as buy-ins, where one affiliate purchases intellectual property from another rather than licensing it. Another tax regulation allows “cost-sharing agreements,” where an affiliate acquires rights in IP by paying a share of development costs rather than a share of profits.
Both of these regulations give firms a wide avenue to avoid the statutory “commensurate with income” requirement. For instance, the taxpayer’s incentive for a buy-in transaction is to set a low value for existing IP. The tax regulations allow the value to be set at the time the intellectual property rights are transferred.  If the valuation meets the “arm’s length” standard (as imperfectly understood by applicable law and practice) it need not be adjusted upward in the event of high profits.
Finally, the administrative problem. The analysis of a correct transfer price is a factbound question, and taxpayers have greater planning and litigation resources compared to the government. On the ground, transfer pricing law gives taxpayers wide (though not unlimited) latitude to set intercompany prices. Transfer pricing plays an important part in enabling multinational firms to dramatically reduce their worldwide tax bill relative to the statutory rate. In general, the government has not been up to the task of challenging taxpayer positions, which are protected by thick and expensive transfer pricing analyses. Recent taxpayer victories in transfer pricing cases include Veritas, Xilinx, Medtronic, and Amazon.
Blouin and Wasserman’s argument, featuring the Foxtrot example
And yet, despite the imperfection of tax transfer pricing data, Blouin and Wasserman correctly argue that tax transfer pricing data could be relevant for patent litigation. Even if transfer pricing rules are weak, they could sometimes provide a boundary for the ability of firms to choose intercompany prices. Transfer pricing at least offers evidence of the outside limit of what royalty (or other price) the taxpayer could set for its IP with a reasonably straight face. And as Blouin and Wasserman point out, transfer pricing may sometimes offer both an upper and lower bound.
Blouin and Wasserman’s strongest case involves opposing tax incentives contained within a single firm. For one intercompany price, the firm prefers a low royalty. For another intercompany price, the firm prefers a high royalty. Their paradigm “Foxtrot” structure considers a multinational consumer products company with a conventional three-box income tax planning structure. Foxtrot U.S. is the parent. It owns Foxtrot Bermuda, a low-tax holding company. Foxtrot Bermuda in turn owns non-U.S. operating subsidiaries including Foxtrot Netherlands.
Blouin and Wasserman point to two intercompany price incentives for IP in the Foxtrot structure. First, there are prices paid by Foxtrot Bermuda to Foxtrot U.S. Foxtrot has the incentive to set these payments as low as possible, because they will increase the taxable income of Foxtrot U.S., the parent company. These include ten annual “buy in” payment installments structured as “declining royalty” payments. These payments are called royalties, but they do not depend on actual profits supported by the Foxtrot IP. Instead, they are based on a valuation of existing IP when the structure is created. The payments from Bermuda to the U.S. also include “cost sharing” payments. These also do not depend on the actual profit earned by Foxtrot. Rather, they allow Foxtrot Bermuda to purchase ownership of non-U.S. IP from Foxtrot U.S. by paying a share of the development costs of the IP. 
The second set of payments for IP in the Foxtrot structure consists of a royalty paid by an operating subsidiary, Foxtrot Netherlands, to Foxtrot Bermuda. Foxtrot has the incentive to set these royalties as high as possible, because they will produce a tax deduction that decreases the taxable income of Foxtrot Netherlands. The government report that provides the Foxtrot example says that the royalty was set under an agreement with the Netherlands tax authority, that “guarantees … a certain level of taxable income” for Foxtrot Netherlands. That is, it was likely set based on actual results, so that the royalties allocated a substantial share of the profit to Foxtrot Bermuda and left a relatively small share in Foxtrot Netherlands.
In the Foxtrot structure described in the Joint Committee report, the buy-in payment for existing IP, made over ten years by Bermuda to Foxtrot U.S., was $5 billion. The average annual cost-sharing payment, a share of the cost of IP under ongoing development, was $1.9 billion. And, the royalty paid by Netherlands to Bermuda was $4.6 billion annually. Not surprisingly, the Netherlands-to-Bermuda annual payments are larger than the Bermuda-to-U.S. annual payments. This gap is necessary to the goal of leaving a large amount of profit in the low-tax Bermuda holding company. A price range is created by opposing tax incentives within the same multinational group. This is promising data for patent litigation, as Blouin and Wasserman argue.
Giving context to tax transfer prices
But these amounts are also apples, oranges, and pears. They cannot be directly compared. If they could, the higher Netherlands to Bermuda royalty would be evidence of a comparable transaction that would support a strong U.S. government case challenging the validity of the low Bermuda-U.S. stream of payments. Some possible differences are suggested by the comparability factors in the Medtronic case. These include exclusivity, coverage of more or fewer patents, coverage of future-developed or past-developed IP, and coverage of corollary rights such as knowhow.
The different intercompany agreements also take different approaches to valuing IP in advance versus valuing IP based on the profits actually experienced. The cost share does not divide profit. Instead it divides expense—the cost of creating future IP—between the U.S. and Bermuda. The royalty paid by the Netherlands to Bermuda is based on actual profit. Its goal is to leave a small, agreed-on amount of actual profit in the Netherlands. The buy-in is midstream. It is based on a lowball valuation of “pre-existing” IP that, together with future-developed IP, supports the success of the company’s products.
Another difference between the affiliates’ agreements is the different enforcement motivations and capacities of the involved governments. The Netherlands government may prioritize foreign direct investment. It may make the deliberate policy decision to allow generous deductions, and thus offer a low de facto tax rate, powered by an agreement that permits large royalty deductions in order to attract Foxtrot’s business. The U.S. government might have a more substantial enforcement incentive to challenge the low payments made by Foxtrot Bermuda for IP developed by Foxtrot U.S. Its rules certainly require Foxtrot Bermuda to pay something for this IP. And the U.S. government’s willingness to challenge some transfer pricing planning presumably helps to put a floor under these prices.
The prospect of admitting transfer pricing evidence in the Foxtrot structure
Adjusting for all of these differences—comparability factors, taxpayer incentives, and government enforcement—would be part and parcel of admitting tax transfer pricing evidence in patent litigation. Here is an example of the possible payoff of admitting transfer prices as evidence of a reasonable royalty in patent litigation cases.
Let us assume that Foxtrot is engaged in a patent infringement lawsuit. Foxtrot is the plaintiff, meaning that Foxtrot claims the defendant infringed Foxtrot’s patents. Foxtrot claims reasonable royalty damages. Foxtrot would prefer to introduce evidence of high royalties paid with respect to its patents.
If transfer prices were admitted into evidence, Foxtrot presumably would emphasize its intercompany royalty agreements between the Netherlands and Bermuda. It would argue that these agreements show that a higher royalty is appropriate. But Foxtrot is limited by the additional evidence in its structure about the arrangement between the U.S. and Bermuda. The Bermuda-to-U.S. payments likely suggest a lower value for the patent. These higher and lower limits might usefully inform the setting of patent damages.
A problem with this analysis is that it is not easy. It is so difficult that a court might not succeed at it, and bad law could result. Tax transfer prices would require adjustments for comparable terms. They might differ from the reasonable royalty sought by the patent case, because they involve bundled licenses, exclusive rights, nonpatent rights, different timing, the use of expenses rather than profits, and so forth. The prices might be less probative of proper patent damages, because of taxpayer incentives or uneven government enforcement. Patent litigants may have the capacity and incentive to manage and argue over this contextual evidence, but they and the courts will struggle to analyze efficiently the many layers of interactions between two complex regulatory schemes: patent and tax. In other words, transaction costs could exceed the benefit of any increase in the quality of patent damage awards that results from a consideration of tax transfer prices in patent litigation.
What about other costs that could increase as a result of Blouin and Wasserman’s proposal?
A different issue is the possibility that allowing tax transfer prices to influence patent damages awards might encourage taxpayers to take transfer pricing positions for reasons other than tax planning. Some patent holders may not know in advance whether they would prefer a higher or a lower royalty rate in litigation, which would decrease the possibility that transfer prices would anticipate litigation. Other patent holders might know in advance that they prefer a higher or a lower royalty rate in patent litigation, but even so the link between the two systems should not decrease the quality of tax transfer prices.
Say the patent holder knows that it would prefer a low royalty rate, because it anticipates that it will face litigation that challenges the validity of a patent. If this incentive aligns with the tax incentive (because, for example, an intercompany royalty is includible in a high-tax jurisdiction), then the incentive will have little effect because the patent holder’s tax planning already pushes the royalty rate as low as the tax law allows. On the other hand, if the patent incentive is opposed to the tax incentive (because, for example, an intercompany royalty is deductible in a high-tax jurisdiction), then the patent incentive could improve the quality of transfer prices.
It may be that the cost of transfer pricing planning could increase if transfer prices also mattered in patent litigation. Some of this cost might be deadweight loss, and some of it might impact firms differently, thus skewing the market in which the firms compete. But these are not the only relevant items in a cost-benefit analysis. One question is the core motivation for Blouin and Wasserman’s proposal—use of transfer pricing data in patent litigation could improve the quality of patent damage awards. Another consideration is the considerable transaction costs of analyzing price data in light of the complexities of two intricate legal systems.
Tax transfer prices are imperfect. They are motivated by the incentive to reduce tax, not by the incentive to get the prices right. Theory, doctrine, and constrained administrative resources limit the quality or truth of transfer prices. But this does not mean that tax transfer prices are irrelevant to the problem of patent damages. It means that the prices are contextual. If they are used, they should be used with attention to comparability of terms, taxpayer incentives, and government enforcement. Patent litigants may have ample incentive to engage with questions of comparability, but understanding the interaction between the complex tax system and the complex patent system as applied to transfer pricing data would not be easy. It could be so hard that the transaction costs would exceed the benefit of any increase in the quality of patent damages awards.
* Angus G. Wynne, Sr. Professor of Civil Jurisprudence, University of Texas School of Law.
 Ga. Pac. Corp. v. U.S. Plywood Corp., 318 F. Supp. 1116 (S.D.N.Y. 1970), modified and aff’d., 446 F.2d 295 (2d Cir. 1971), cert. denied, 404 U.S. 870 (1971).
 Treas. Reg. § 1.482-2(a) (2011).
 Jennifer L. Blouin & Melissa F. Wasserman, Tax Solutions to Patent Damages, 26 Tex. Intellectual Prop. L. J. 1 (2018).
 See, e.g., Lucent Techs., Inc. v. Gateway, Inc., 580 F.3d 1301, 1324-25 (Fed. Cir. 2009) (explaining that the Georgia Pacific standard imagines a “hypothetical negotiation” that produces an “ex ante” license). As Blouin and Wasserman note, the license price should be set for patent damages purposes when infringement begins, and a timing adjustment compared to the tax transfer pricing assumption may often be required. Id. manuscript at 22 (noting that transfer price may be useful information for reasonable royalty calculation when “timeframes … overlap”).
 Other issues include the application of judicial discretion, see John M. Golden, Discretion in Patent Damages Infringment, Rev. of Litig. (forthcoming 2018) and the problem of the confidentiality of licensing arrangements, see John M. Golden, Principles for Patent Remedies, 88 Tex. L. Rev. 505, 550 (2010) (explaining that little information about damages exists, in part, because of confidential settlements).
 See Medtronic, Inc. v. Comm’r, 2016-112 (T.C. 2016) (starting with a 7% royalty based on a patent litigation settlement and adding 7% for exclusivity, 3% for future technology, 7% for knowhow, 3.5% to account for higher-profit patents, 2.5% to account for additional products to arrive at 30% retail sales royalty).
 For instance, transfer pricing data arises in connection with the question of whether a parent corporation may recover damages based on its subsidiaries’ lost profits. Compare Mars, Inc. v. Coin Acceptors, Inc., 527 F.3d 1359, 1367 (Fed. Cir. 2008) (suggesting that only if subsidiary profits “flow inexorably up to the parent” might a lost profit theory permit their recovery) with St. Jude Med. v. Access Closure, 2010 WL 4968147, at *6 (W.D. Ark. Dec. 1, 2010) (holding that plaintiff could pursue damages theory based on idea that reduced profits at wholly-owned subsidiaries reduced market value of parent). See also Tucker Terhufen, Hidden Risks in Exploiting Intellectual Property to Avoid Corporate Taxes, 47 Ariz. St. L. J. 1425, 1437-38 (2015) (proposing use of a patent infringement plaintiff’s own low-royalty tax structure to support argument for lower patent damages); Lane M. Webster, Note, St. Jude Medical v. Access Closure: The Other Lost Profits for a Patentee’s Subsidiaries, 16 S.M.U. Sci. & Tech. L. Rev. 237, 238 (2013) (noting “new judicial trend acknowledging parent-subsidiary corporate structures”). Cf. WesternGeco LLC v. ION Geophysical Corp., 138 S.Ct. 2129, 2137-38 (2018) (allowing recovery for lost foreign profits under domestic supply/foreign assembly facts and § 271(f)(2) of the Patent Act).
 See, e.g., Inventio A.G. v. ThyssenKrupp Elevator Ams. Corp., 2010 WL 9546391, at *3-5 (D. Del. June 17, 2010) (requiring production of documents including an internal “Transfer Pricing Analysis” of “New IP Systems” and an internal memo regarding IP licensing written by a tax department employee).
 See, e.g., Coin Acceptors, Inc., 527 F.3d at 1373 (considering though rejecting royalty rate agreed for UK tax purposes); see also Warsaw Orthopedic, Inc. v. Nuvasive, Inc., 2015 WL 12034662, at *1 (S.D. Cal. Sept. 15, 2015) (admitting evidence of royalty payments between Warsaw and its affiliates; although these might not be used as “evidence of an established royalty,” they might not be “completely irrelevant”; however, evidence of “true-up” payments required by transfer pricing not admissible).
 See, e.g., Coin Acceptors, Inc., 527 F.3d at 1359 (upholding District Court’s refusal to rely on 4% intercompany license rate negotiated with U.K. tax authority) (“While Coinco may be correct that the United Kingdom taxing authorities requested that the license rate be one that simulates the rate that would have been reached in an arm’s-length negotiation between independent enterprises, there is no evidence that suggested that the 4% rate would have been the rate at which Mars would have licensed a competitor.”); See also Allen Archery Inc. v. Browning Mfg. Co., 898 F.2d 787 (Fed. Cir. 1990) (reversing district court’s use of intercompany price to set damages and substituting price charged customers, because of an “absence of arm’s length bargaining”).
 See R.H. Coase, The Nature of the Firm, 4 Economica 386, 390 (1937) (noting that firm ownership makes sense if it decreases transaction costs by more than it increases agency costs).
 See Oliver E. Williamson, The New Institutional Economics: Taking Stock, Looking Ahead, 38 J. Econ. Lit. 595, 603 (2000) (explaining hold-up problem).
 See Susan C. Morse, The Transfer Pricing Regs Need a Good Edit, 40 Pepp. L. Rev. 1415, 1421 (2013) (noting lack of comparables).
 Both U.S. law and OECD guidance acknowledge the sticky problem of residual profit, though neither provides an allocation method that enjoys widespread use. The U.S. tax regulations acknowledge this problem in providing a “residual profit split” transfer pricing method. This method first allocates a profit based on comparables to affiliates that provide more routine services, such as marketing. It attributes residual profit to high-value intangibles. See Treas. Reg. § 1.482-6(c)(3) (2009). The OECD has also provided soft guidance on the subject of allocating residual profit from intangibles. It recommends an allocation of IP profit not according to legal ownership, but rather based on the “functions performed, assets used, and risks assumed” by different entities within a multinational firm. See Org. for Econ. Cooperation and Dev., Base Erosion and Profit Shifting Project Action Items 8-10: Aligning Transfer Pricing Outcomes With Value Creation 73-74 (2015).
 I.R.C. § 482.
 See Treas. Reg. § 1.482-7(a)(2) (outlining rules for “platform contribution” transactions, also known as buy-ins). See also Treas. Reg. § 1.482-7(a)(1),7(b) (providing rules for “cost-sharing arrangements”).
 Typically, such “buy-in” transactions are licenses for purposes of intellectual property law, but not for purposes of tax law. For instance, they are often structured as exclusive licenses within a particular geographic area, which for purposes of tax law is treated as a purchase and sale for an immediate lump sum consideration rather than a license involving the continued payment of future royalties. See E.I. Du Pont De Nemours & Co. v. United States, 432 F.2d 1052 (3d Cir. 1970) (holding that a license of Brazilian patents relating to the manufacture of nylon qualified as a sale for tax purposes). See also Treas. Reg. § 1.482-7(c) (establishing the default rule that platform contribution transactions feature exclusivity).
 Compare Treas. Reg. § 1.482-7(b)(5) Example 3 (software platform contribution transaction priced at $200 million over four years meets arm’s length standard) with Treas. Reg. § 1.482-7(b)(5) Example 4 (similar transfer with $10 million value does not meet arm’s length standard, because taxpayer assumed a useful life of only 8 years contemporaneous with patent life).
 See, e.g., J. Clifton Fleming, Jr., Robert J. Peroni & Stephen E. Shay, Worse Than Exemption, 59 Emory L. J. 79, 84-85 (2009) (noting role of transfer pricing in reducing multinational firms’ income tax liability); Edward D. Kleinbard, Stateless Income, 11 Fla. Tax Reg. 699, 707-14 (2011) (same); Richard J. Vann, Taxing International Business Income: Hard-Boiled Wonderland at the End of the World, 2 World Tax J. 291 (2010) (same).
 Veritas Software Corp. v. Commissioner, 133 T.C. 297 (2009) (rejecting government argument that taxpayer’s buy-in valuation was too low by a factor of ten).
 Xilinx Inc. v. Commissioner, 598 F.3d 1191 (9th Cir. 2010) (rejecting government interpretation of regulations that required taxpayers to include stock option deductions in the cost base shared under cost sharing agreements). In 2018, the Ninth Circuit upheld a revised regulation that explicitly referred to stock options and required the sharing of related costs, but then withdrew the opinion for a procedural reason. See Altera Corp. v. Comm’r, 2018 WL 3542989, *1(9th Cir. 2018) (upholding regulation “under general administrative law principles”), rev’g Altera Corp. v. Comm’r, 145 (T.C. Jul 27, 2015), withdrawn, 2018 WL 3734216 (9th Cir. 2018).
 Medtronic (2016) (rejecting government argument for higher royalty rate, and reaching a decision consistent with an earlier negotiated agreement between government and taxpayer).
 Amazon.com, Inc. v. Commissioner, 148 ( T.C. No.6 2017) (rejecting government enterprise valuation theory for valuing technology transferred under buy-in).
 The Foxtrot example is drawn from a government report. See Joint Committee on Taxation, Present Law and Background Related to Possible Income Shifting and Transfer Pricing (July 20, 2010) at 93-102 (hereafter JCT Report) (giving Foxtrot example). It is presented in simplified form here. For instance, this presentation omits entities such as Foxtrot Hong Kong and intercompany arrangements related to contract manufacturing.
 See JCT Report at 95-96 (explaining buy-in).
 See JCT Report at 99 (explaining royalties). This is consistent with other evidence of agreements between U.S. multinationals and some non-U.S. governments. See Ruth Mason, Special Report on State Aid – Part 3: Apple, 154 Tax Notes 735, 738-39 (2017) (describing agreement between Apple and Ireland, which ensured that Apple’s deductions in Ireland would produce the result of $30 - $40 million of profit allocated to Ireland, which was the amount that Apple “would be prepared to accept”).
 JCT Report at 95.
 JCT Report at 96.
 JCT Report at 99.
 See Medtronic (2016) (starting with a 7% royalty based on a patent litigation settlement and adding 7% for exclusivity, 3% for future technology, 7% for knowhow, 3.5% to account for higher-profit patents, 2.5% to account for additional products to arrive at 30% retail sales royalty).
 See Rosanne Altshuler & Harry Grubert, The Three Parties in the Race to the Bottom: Host Countries, Home Countries and Multinational Corporations, 7 Florida Tax Rev. 137 (2005) (explaining host countries’ practice of reducing tax burdens to attract foreign direct investment from multinationals).
 Cf. Mark A. Lemley, Distinguishing Lost Profits from Reasonable Royalties, 51 Wm. & Mary L. Rev. 655, 663-664 (2009) (explaining that residual profit not attributable to the patent should remain with the defendant after payment of reasonable royalty damages to a nonpracticing patent holder).
 See Andrew Blair-Stanek, Intellectual Property Law Solutions to Tax Avoidance, 62 UCLA L. Rev. 2, 35-39 (2015) (arguing that low tax transfer prices could support lower patent damage awards).