United States: Amazon Makes The C.U.T. – An Important Taxpayer Win, A Reminder To Consider Transactional Evidence

INTRODUCTION

In finding for the taxpayer in a recent transfer pricing decision,¹ the U.S. Tax Court followed its own determination in Veritas² in valuing a buy-in payment made as compensation for the right to use pre-existing intangible assets in a related-party cost-sharing arrangement ("C.S.A."). This decision, like other major transfer pricing decisions, serves as a reminder of the fact-intensive nature of transfer pricing matters and of the importance of uncovering and properly analyzing transactional evidence from the controlled transaction in question and from uncontrolled transactions or dealings of the business.

This article comments on the lessons learned from this important taxpayer victory. For a full discussion of the tax treatment of intellectual property ("I.P.") at issue in the case, see "Tax Concerns on Outbound I.P. Transfers: Pitfalls & Planning in Light of I.R.S. Defeat in Amazon Case" in last month's edition of Insights.³

BACKGROUND

Amazon.com, Inc. ("Amazon") entered into a C.S.A. with Luxembourg subsidiary A.E.H.T. in 2005. The C.S.A. covered: (i) the software and other technology underlying the Amazon European domain-name websites, fulfillment centers, and related business activities; (ii) marketing intangibles, including trademarks, tradenames, and domain names used in Amazon's European business; and (iii) customer lists, customer data, and the Amazon tradename and mark. The right to use the pre-existing intangible assets in these three categories was priced at $254.5 million, payable over a seven-year period corresponding with the useful life of the intangible assets.

Using the income method and the same discounted cash flow approach rejected by the court in Veritas, the I.R.S. estimated the arm's length value of the buy-in payment to be $3.468 billion, effectively disregarding the C.S.A. and valuing the transfer of rights as a business that would exploit infinitely-lived intangibles in perpetuity. The I.R.S. also disputed the Amazon failure to classify certain technology and content expenses of Amazon.com as intangible development costs, thereby biasing downward the income from annual cost sharing payments received from A.E.H.T. over the term of the C.S.A.

The economic substance of the A.E.H.T. Luxembourg operations hub was not critical on its own. Local language requirements, local vendor relations, and European logistics considerations and customer tastes were all factors contributing to the need to carry on a business in Luxembourg, and the change in economic position for A.E.H.T. expected to result from the C.S.A. In rejecting the I.R.S. transfer pricing method, the court made clear that "A.E.H.T. was not an empty cash box." This determination contrasts sharply with the O.E.C.D. outcomes under the B.E.P.S. Action Plan that attack hypothetical "cash boxes" that are legal owners of rights but lack the decision-making and risk management capacity needed to allocate capital to investments with uncertain returns. The dispute in Amazon therefore centered on (i) the transfer pricing method, (ii) the assumptions made and analyses used to value the buy-in payment, and (iii) the correct treatment of the intangible development costs within the term of the C.S.A.

AMA ZON TRIAL

In deciding for the taxpayer, the court relied on the testimony and reports of 30 experts in computer science, marketing, economics, and transfer pricing economics. The opinions of the computer science experts on the state and viability of the Amazon software and websites served as a stable foundation upon which the transfer pricing economics experts for the taxpayer could anchor their assumptions. These assumptions were critical – as the technical constraints of the software system provided a reliable estimate of the lifespan of the software used to power the core operations of the Amazon websites and fulfillment business. The marketing experts helped the court decide on a proper method to estimate key variable values used in the marketing intangibles value calculation. They also assisted the court in determining how the intangibles allowed a team of engineers – for whom no technical challenge seemed too large – to overhaul the websites without causing them to crash during popular shopping seasons.

However, the star of the trial was a Treas. Reg. §1.482-7 transfer pricing method – the Comparable Uncontrolled Transaction method ("C.U.T."). Amazon used an unspecified transfer pricing method resembling in some respects the profit-split method to calculate the original buy-in payment, while the I.R.S. used an application of the income method. The I.R.S. income method calculated the present value of cash flows forecasted to result from A.E.H.T.'s European business, using cash flow and balance sheet forecasts as its only company data input. Both approaches neglected or devalued Amazon's outsourced web store programs, and thousands of Associates or Syndicated Stores programs that provided customer referrals to Amazon.

The website platform and referrals transactional data alone did not win the case for Amazon. Considerable expert testimony was required to establish reliable assumptions of discount rates, value decay rates, useful asset life, and trademark ownership. The company's own information, however, was a crucial element in winning the case. C.U.T.'s that involve transactions between the taxpayer and independent businesses (sometimes called internal C.U.T.'s), are highly persuasive given these fit well within the framework of the comparability requirements of Treas. Reg. §1.482(c)(1), which is critical to selecting a best method. C.U.T.'s are not abstract agreements between third parties. They must bear some resemblance to one of the controlled parties and its business.

One small levity allowed in the 207-page decision was that "one does not need a Ph.D. in economics to appreciate the essential similarity between the DCF methodology that Dr. Hatch employed in Veritas and the DCF methodology that Dr. Frisch employed here." Similarly, a Ph.D. is not required to present a well-selected and adjusted C.U.T. to the I.R.S. or a Tax Court judge. It seems unlikely in the case of Amazon's C.S.A. that the I.R.S. would have paid any attention to a C.U.T. at the examination level, given the strong motivation within the I.R.S. to re-litigate Veritas. Nonetheless, C.U.T.'s remain a valuable commodity to be mined and stockpiled for use in appropriate circumstances.

CONCLUSION

Not only was Amazon's transactional data important in building its case in favor of the buy-in payment value, its Code §41 credit cost detail proved useful in substantiating the company's claim that a significant class of expenses should not be classified as intangible development costs and shared with other C.S.A. participants. This is another good example of seeking the data required within the company's records before reinventing the wheel.

An open question in the case is the treatment of employee stock option costs in a C.S.A. This question will have to wait for the outcome of the I.R.S. appeal in Altera.⁴

Pending a successful outcome in Altera, two theories used by the I.R.S. to attack a technology company C.S.A. could be blunted. To the extent that I.R.S. estimates regarding the size of the tax gap rely on large income windfalls from litigating C.S.A. positions of high-tech companies, Amazon could prove to be an early indication that these estimates require  downward adjustment.

Footnotes

1 Amazon.com, Inc. & Subsidiaries v. Commr., T.C., 148 T.C. No. 8 Docket No. 31197-12.

2 Veritas Software Corp. v. Commr., T.C., 133 T.C. 297 (2009).

3 Philip Hirschfeld, "Tax Concerns on Outbound I.P. Transfers: Pitfalls & Planning in Light of I.R.S. Defeat in Amazon Case," Insights 4 (2017).

4 Altera Corp. v. Commr., T.C. 145 T.C. No. 3 (2015).

The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.