Amazon.com v. Commissioner: Veritas Redux?

Neal Kochman
Stafford Smiley[1]

On December 28, 2012, Amazon.com, Inc. (Amazon[2]) petitioned the United States Tax Court (Tax Court or simply Court) for a redetermination of a $234 million tax deficiency for its 2005 and 2006 taxable years[3]. The deficiency resulted from Internal Revenue Service (IRS) increases to Amazon European Holding Technologies SCS (AEHT) "buy-in" payments to Amazon for pre-existing intangibles that Amazon made available to a qualified cost sharing arrangement among AEHT and certain Amazon U.S. subsidiaries (the CSA). 

The purpose of the CSA was to develop technology for operating Amazon-branded websites. Under the CSA, AEHT, Amazon's Luxembourg subsidiary, obtained a non-exclusive right and license to covered intangibles (i.e., the technology developed under the CSA) for use in operating European websites, and Amazon obtained rights to the intangibles for use in operating non-European websites. Amazon engaged Deloitte Tax LLP (Deloitte) to conduct a transfer pricing study and value the pre-existing intangible property[4]. The transfer pricing study concluded that the present value of the amount that AEHT should pay for the intangibles was $217 million as of January 1, 2005. The method used by Amazon to value the intangibles is unclear from Amazon's Petition and the IRS Answer; it might have been a declining royalty or some sort of profit split. What is clear is that Amazon's valuation was based on a useful life for the pre-existing intangibles of no more than seven years, that it was assumed the value of the intangibles would decay over the seven-year useful life, and that AEHT funded and co-developed the covered intangibles beginning in 2005. Moreover, according to the IRS Answer, it appears that Amazon did not separately value the items of pre-existing intangible property subject to the buy-in, but rather valued the property in the aggregate, and that it used a 13 percent discount rate to determine the present value of the buy-in payment.  Based on the transfer pricing study, AEHT agreed to make buy-in payments with a cumulative present value of $217 million over a seven-year period beginning in 2005. The 2005 payment was $73 million and the 2006 payment was $83 million.

The IRS began auditing the CSA in July 2008, and in 2009 engaged Horst Frisch, Inc. (Horst Frisch) to value the buy-in.  Horst Frisch issued a valuation report in January 2011. Details of the Horst Frisch valuation are not available, but the Petition and Answer agree that Horst Frisch :

• Applied the discounted cash flow (DCF) method as an unspecified method under Treas. Reg. § 1.482-4(d),
• Used European website 2005 through 2011 operating profit projections that were identified by Deloitte in its transfer pricing study,
• Used a 3.8 percent terminal growth rate,
• Used a 18 percent discount rate, and
• Did not separately value the items of intangible property that were subject to the buy-in, but rather valued the intangibles in the aggregate.

Horst Frisch determined a value for the pre-existing intangibles as of January 1, 2005, of $3.6 billion. Converting that value to buy-in payments over a seven-year period, the IRS, in a notice of proposed adjustment (NOPA) issued May 9, 2011, proposed increases in the 2005 and 2006 AEHT buy-in payments of $1 billion and $1.2 billion, respectively.

There is no information on how the dispute proceeded between issuance of the NOPA and Amazon's Petition, but it does appear that Amazon decided to forego the IRS administrative appeal process. Normally, if there are unagreed issues at the end of an IRS examination, the IRS issues a "30-day letter," which  provides the taxpayer 30 days to either accept or request review of the IRS proposed adjustments. If the taxpayer requests review, jurisdiction shifts from IRS Examination to IRS Appeals, an independent function within the IRS that may resolve the dispute based on "hazards of litigation," i.e., the likelihood that the taxpayer would prevail if the dispute were litigated.  If a settlement is not reached in Appeals, the IRS issues a notice of deficiency under which a taxpayer has 90 days to either to pay the tax or to petition the Tax Court for a redetermination of the deficiency. It appears that the IRS did not issue a 30-day letter in the Amazon dispute, which typically would happen only if the period for the IRS to assess tax would expire before it could issue a notice of deficiency if the IRS followed the normal 30-day letter procedures. Since a taxpayer can always agree to extend that period, the fact that the IRS issued a notice of deficiency without first issuing a 30-day letter must mean that Amazon refused to sufficiently extend the period of limitations to allow for the normal 30-day letter process, and thus wanted to skip Appeals and to go straight to Tax Court.

It is somewhat curious that Amazon decided to bypass Appeals. For a buy-in dispute, the only decided case is Veritas[5], which was a victory for the taxpayer.  Rumors are that taxpayers with buy-in disputes have received fairly favorable settlements in Appeals post-Veritas, since the hazards appear to weigh in taxpayers' favor. It is possible the IRS took a hard-line position with Amazon, viewing the case as an opportunity to reverse the result in Veritas, and therefore Amazon saw no point in engaging with Appeals. However, the IRS Large Business and international division's position on the case would not necessarily have any bearing on an Appeals Officer operating independently of Examination and weighing the hazards of litigation.

Amazon's argument, as outlined in the Petition, tracks many of the findings in Veritas. The Petition claims that the Horst Frisch DCF method is the same or similar to the method used by the IRS testifying economic expert in Veritas, that Horst Frisch valued the pre-existing intangibles into perpetuity even though the intangibles had a limited useful life, that it performed a business enterprise valuation rather than a valuation of the pre-existing intangibles as required by Treas. Reg. § 1.482-7(g)(2)[6], and that income attributable to AEHT's ownership of cost-shared intangibles under Treas. Reg. § 1.482-7(a)(2) was improperly allocated to Amazon. The IRS Answer denies these claims.

Veritas

Veritas Software Corp. (Veritas) was a developer and manufacturer of data storage management software[7]. In the late 1990s it sold six primary products -- one focused on the commercial market and five focused on the enterprise market. It sold those products directly to customers, through distributors and resellers, and through licenses to original equipment manufacturers (OEMs). The OEMs either bundled the Veritas software with their operating systems or sold the Veritas software to their customers as an option. Veritas collected royalties from both the OEMs and the resellers based on their product sales.

The data storage software market was one of intense competition and rapid technological change.  According to the Tax Court opinion, products lost value quickly as competitors duplicated capabilities or introduced new technology[8]. To remain relevant, Veritas had to continuously innovate; in the late 1990s and early 2000s a Veritas product had an average useful life of four years. 

Although Veritas generally continued to develop and update its products on a regular basis, at times it would produce custom products for OEMs. It would not perform development work on those custom products following delivery to the OEMs, and the license agreements for those products would include a decay provision. Under such a provision, the royalty rate would steadily decline over time to account for obsolescence.

Prior to 2000, Veritas had limited reach outside the United States.  In the mid to late 1990s, Veritas began expanded through acquisitions, and extended its reach in the Europe, Middle East, and Africa (EMEA) and Asia Pacific (APAC) markets. It incorporated two Irish subsidiaries (collectively Veritas Ireland), and in late 1999 entered into an agreement to share research and development costs (the RDA) and a technology license agreement (the TLA) with Veritas Ireland.  The RDA was a qualified cost sharing arrangement under Treas. Reg. § 1.482-7. In the RDA, Veritas Ireland received the right to use the intangibles developed under the agreement in the EMEA and APAC regions.  Under the TLA, Veritas granted Veritas Ireland the right to use pre-existing intangibles that Veritas made available to the RDA, and Veritas Ireland agreed to make a buy-in payment. Veritas Ireland made a lump-sum buy in payment in 2000 of $166 million, which Veritas reported on its 2000 federal income tax return.  Subsequently, the payment was amended to $118 million.

Veritas' testifying economics expert used the comparable uncontrolled transaction (CUT) method to demonstrate that the buy-in payment was arm's length. For comparable transactions, he used the license agreements between Veritas and OEMS. From those agreements he derived a range of starting royalty rates for use of the pre-existing intangibles. He assumed a useful life of two to four years for the intangibles, and reduced the royalty rates over the buy-in period based on decay provisions in the Veritas OEM agreements. The resulting valuation range was $94 to $315 million, and the economist concluded that most values fell between $100 and $200 million.

The IRS audited Veritas' 2000 and 2001 taxable years, and in 2006 issued a notice of deficiency, increasing Veritas' 2000 and 2001 tax by a total of $758 million. The deficiency was based on an economist report that applied three separate approaches to value the buy-in — the foregone profits method (essentially DCF), the market capitalization method, and the acquisition method — to derive values ranging from $1.9 to $4 billion. The economist concluded that $2.5 billion was the appropriate value for the buy-in.  That value was used by the IRS to determine the deficiency.

Veritas petitioned the Tax Court for a redetermination of the deficiency.  In its statement of position to the Court, the IRS at first indicated it was not sure which transfer pricing method it would use at trial; it then stated it would use the foregone profits method but would not rely on the previous economist report.  Subsequently, the IRS submitted an expert report that concluded, based on a DCF analysis, that the lump sum buy-in should be $1.675 billion. According to the Tax Court opinion, in arriving at that value, the IRS assumed the pre-existing intangibles had a perpetual life, used a discount rate of 13.7 percent, determined that Veritas Ireland's revenue grew at a compound annual growth rate of 17.91 percent from 2001 through 2005, and, based on that growth rate, assumed 13 percent growth from 2007 through 2010 and 7 percent growth thereafter.

Reading the Tax Court opinion, it is clear the trial did not go well for the IRS.  Certain facts appear to have undermined its expert report.  For example, whereas the IRS expert relied on a 17.91 percent growth rate from 2001 through 2005, Veritas Ireland's actual growth rate from 2004 through 2006 was 3.75 percent. According to the opinion, a buy-in based on the IRS assumed growth rates would have required Veritas Ireland to pay all of its actual and projected operating income through 2009 to Veritas. Accordingly, Judge Foley concluded the IRS expert had employed unreasonable growth rates. He also concluded that the discount rate used was too low, and that the appropriate discount rate was 20.47 percent.

The IRS argued that the transfer of pre-existing intangibles to the RDA was "akin to a sale" and that it was appropriate to value the transferred intangibles in the aggregate rather than separately valuing each transferred asset, so as to capture synergistic effects. However, although Treas. Reg. § 1.482-1(f)(2)(i)(A) authorizes an aggregate analysis when such an approach provides the most reliable means for determining arm's length consideration, Judge Foley rejected the akin to a sale construct.  He found it treated short-lived intangibles as having perpetual life and took into account subsequently-developed intangibles. Moreover, Judge Foley found that the IRS trial position reflected concepts such as "platform contribution" and "workforce in place" that were included in the temporary cost sharing regulations promulgated in 2009, but that were not reflected in the cost sharing regulations in effect for the years at issue.

Most importantly, Judge Foley concluded that the CUT method, and not the DCF, was the best method for determining Veritas Ireland's buy-in payment. He rejected the IRS contention that the intangibles transferred under the TLA were substantially different from those addressed in the OEM agreements and therefore were not good CUTs. He did make certain adjustments to Veritas' application of the CUT method.  Specifically, he concluded that a 32 percent starting royalty was appropriate rather than the 20 to 25 percent used by Veritas, that the pre-existing intangibles had a four-year useful life, that the royalty rate should ramp down at 33 percent per year beginning in year 2, and that a 20.47 percent discount rate should be used to determine the present value of the buy-in payment. The decision does not show the impact of these changes on the buy-in payment.

The IRS Response to Veritas

The IRS did not appeal Veritas.  It did, however, issue an action on decision (AOD) in which it non acquiesced in the Tax Court holding because it believed both the factual findings and legal assertions were erroneous[9]. The major factual difference involved the nature of the intangibles to be valued.  As reflected in the 2009 cost sharing regulations, the IRS views a pre-existing intangible made available to a cost sharing arrangement as having two components of value:  ongoing research and development (R&D) value and value from the sale of existing products incorporating the intangible. The former reflects the value of an intangible as a foundation for the development of new intangibles that will be used in future products, while the latter reflects the make or sell rights.  There is little question that in a rapidly changing field such as software development, the make or sell rights have a fairly short useful life and the value of those rights degrades as the software becomes dated. That does not necessarily mean, however, that an intangible ceases to have value as a platform for future R&D. Judge Foley rejected the proposition that the Veritas software had ongoing value, concluding instead that the only value in the Veritas pre-existing intangibles was the make or sell rights. Having reached that conclusion and with evidence of arm's length payments for the make or sell rights, it is hard to argue against application of the CUT method.

In the AOD, the IRS notes that the Tax Court's factual findings removed the underpinnings of the IRS' valuation and supported Veritas' valuation, and therefore it was unnecessary for the Court to make the broad findings it did on the governing law. The major legal finding the IRS objected to is the meaning of "buy-in" in Treas. Reg. § 1.482-7(g)(2), which provides in part, "[i]f a controlled participant [in a CSA] makes pre-existing intangible property…available to other controlled participants for purposes of research in the intangible development area…then each other controlled participant must make a buy-in payment to the owner." According to the AOD, "[t]he court construes the buy-in to exclude any consideration of the future income value or value attributable to intangibles to be developed under a CSA apparently on the theory that such future income stream is already paid for through the participants' cost shares of ongoing R&D."  In the view of the IRS, the ongoing cost sharing payments only account for a portion of the value of the intangibles to be developed under the cost sharing arrangement. The balance of that value is attributable to the head start afforded by the pre-existing intangibles. The IRS contends the Tax Court's interpretation reads "for purposes of research in the intangible development area" out of the regulation. That is, by ignoring the contribution of pre-existing intangibles to the value of intangibles developed under a cost sharing arrangement, the Tax Court limits the value of pre-existing intangibles to their make or sell rights, and does not include any value related to R&D rights. The IRS argues that its interpretation that R&D rights must be compensated is anchored in the regulations in effect for the years at issue, not just in the 2009 cost sharing regulations.

Other findings the IRS objected to involved the value of workforce in place and the appropriateness of an aggregate analysis. The Tax Court expressed the view that R&D and marketing teams do not have substantial value independent of the services performed by individual employees. The IRS view is that an experienced team in place may contribute value in excess of the compensation paid to individual team members, and therefore may constitute an intangible within the meaning of section 936(h)(3)(B). As to aggregate analysis, the IRS attributed the Tax Court's rejection of its approach to the Court's belief that there was insufficient support in the factual record for such an approach. The IRS notes that it will continue to take an aggregate approach where such an approach "provides the most reliable measure of an arm's length result."

Prospects for the Amazon Case

It is difficult to speculate on how Amazon will fair in the Tax Court, particularly with the limited detail available on the Amazon and IRS valuation methods. Amazon's position, as outlined in its Petition, tracks many of the findings in the Veritas opinion. However, the outcome of a transfer pricing dispute generally turns on the facts. In Veritas the court appears to have been heavily influenced by the fact that the subject pre-existing intangibles, or similar intangibles, had been licensed to third parties and were therefore amenable to a CUT analysis.  Amazon may have used a declining royalty valuation approach similar to that used by Veritas, but it is hard to imagine Amazon coming up with CUTs for the pre-existing intangibles. The subject of the CSA was development of technology to operate Amazon websites, and the pre-existing intangibles subject to the buy-in involved existing website technology. That technology plays a significant role in Amazon's competitive advantage and therefore is unlikely to be licensed to third parties. If Amazon did use a declining royalty method, it probably would have had to develop the royalties through some other approach, such as the comparable profits method. 

On the surface, some of the shortcomings found by the Tax Court in the IRS Veritas DCF analysis do not appear to be present in the Horst Frisch DCF analysis. According to the IRS, Horst Frisch started with the same cash flow projections used in Deloitte's transfer pricing study. It then assumed a 3.8 percent terminal growth rate, as compared to the 7 percent assumed in the Veritas analysis. The IRS also claims that Deloitte did not separately value the individual items of pre-existing intangible property.  Assuming that information is correct, the cash flow data is much less likely to be an issue in the case, as are objections to an aggregate analysis. As to discount rates, Horst Frisch appears to have used a higher discount rate than Deloitte, so any movement towards Amazon's discount rate position would tend to increase the present value of the IRS valuation.

Whatever the outcome in the Amazon case, its impact in the cost sharing arena may be fairly limited. Although at one time the IRS had a huge inventory of buy-in cases under the 1995 cost sharing regulations, it appears that the bulk of those cases have been resolved through the IRS administrative appeals process. The new cost sharing regulations, which took effect in 2009, and are much more in line with the IRS position in Veritas and the AOD.  Both Veritas and Amazon should have limited bearing on the outcome of cases under the new regulations. Of more importance may be the extrapolation of Veritas and the outcome in Amazon to the valuation of intangibles in general under Treas. Reg. § 1.482-4. Although Veritas was about the appropriate value for the buy-in under the cost sharing regulation, ultimately that value turned on a determination of the best method for valuing the buy-in from among the alternative methods provided in the -4 regulations. The new cost sharing regulations provide specific methods for valuing buy-ins, but as yet those methods have not been extended to general intangible valuation.  Taxpayers therefore may attempt to apply principles from Veritas and Amazon (depending on how it turns out) to resolve future intangible valuation disputes.

[1] Mr. Kochman is a member of Caplin & Drysdale, Chartered, Washington, D.C.  Mr. Smiley is Professor, Graduate Tax Program, at Georgetown University Law School, and senior counsel to Caplin & Drysdale. 
[2] "Amazon" is used herein to refer to Amazon.Com, Inc. and its U.S. subsidiaries.
[3] Petition of Amazon.Com, Inc. & Subsidiaries, Amazon.com, Inc. & Subsidiaries v. Commissioner, Docket No. 31197-12 (Dec. 28, 2012) (the Petition).
[4] See Answer by the Commissioner of Internal Revenue, Docket No. 31197-12 (Mar. 8, 2013) (the Answer).
[5] Veritas Software Corp. & Subsidiaries , et. Al. v. Commissioner, 133 T.C. 297 (2009), nonacq., AOD 2010-005, 2010-49 I.R.B. (Dec. 6, 2010).
[6] Unless otherwise indicated, all references to Treas. Reg. § 1.482-7 are to the 1995 version of the regulation.
[7] Symantec Corp. acquired Veritas in 2005.
[8] The record in Veritas is sealed.  All factual information is drawn from Judge Foley's decision.
[9] AOD 2010-005, 2010-49 I.R.B. (Nov. 10, 2010).

"Amazon.com v. Commissioner: Veritas Redux?", published in Corporate Taxation, Volume 40, Issue 4, July/August 2013.  © 2013 by Thomson Reuters/Tax & Accounting.  Reprinted with permission.  All rights reserved. This information or any portion thereof may not be copied or disseminated in any form or by any means or stored in an electronic database or retrieval system without the express written consent of Thomson Reuters/Tax & Accounting.