TAX COURT REJECTS IRS DETERMINATION IN TRANSFER PRICING CASE.

Lloyd Silberzweig, James P. Thurston, Kimberley Peterson, David Rakonitz, Stephanie Profitt, Margaret Burow, and John Strate, for respondent.

                            CONTENTS

VERITAS US is in the business of developing, manufacturing, marketing, and selling advanced storage management software products. VERITAS US' products protect against data loss and file corruption, provide rapid recovery after disk or system failure, process large files efficiently, manage and back up systems without user interruption, and provide performance improvement and reliability enhancement features that are critical for many commercial applications.

In the mid to late 1990s VERITAS US expanded its business through corporate acquisitions and the establishment of foreign subsidiaries. On April 25, 1997, VERITAS US acquired and merged with OpenVision Technologies, Inc. (OpenVision). With the acquisition of OpenVision, VERITAS US obtained NetBackup;³ offices in the United Kingdom, Germany, and France; an engineering team; and skilled sales and marketing executives. By the end of 1997 VERITAS US had sales subsidiaries in Canada, Japan, the United Kingdom, Germany, France, Sweden, and the Netherlands. VERITAS US, on May 28, 1999, acquired Seagate Software Network and Storage Management Group, Inc. (NSMG). As a result of this acquisition, VERITAS US became the largest storage software company in the industry and obtained Backup Exec;⁴ a distribution channel in Europe, the Middle East, and Africa (EMEA); and a sales force that sold Backup Exec to customers in Europe. On July 2, 2005, VERITAS US was purchased by Symantec Corp. (Symantec) and became one of Symantec's wholly owned subsidiaries. References to petitioner are to VERITAS US, its subsidiaries, and Symantec (successor in interest to VERITAS US and subsidiaries).

Prior to 1999 only one application could access a data file at any given time. Thus, to back up data on secondary storage, it was first necessary to shut down all applications using the data. Most secondary storage was on magnetic tape and directly attached to a single server. After the CSA, there were important technological advances relating to the data storage software industry. In response to 24-hour Web sites, backup technology advanced significantly, enabling backups to run at any time. In addition, exponential increases in file size and data volume and the plummeting cost of disk storage spurred the use of disks as secondary storage. The switch to disks as the primary backup medium required the source code⁶ of backup products to be rewritten. The advent of storage area networks allowed storage to be shared by numerous computers, allowed more than one server to access a particular piece of data, and enabled applications to run continuously without interruption. Other technological advances dramatically increased storage capacity and also facilitated disaster recovery by allowing storage resources to be replicated several times in different data centers. These advances reduced the cost of physical storage and made it possible for many systems to share storage devices.

During the years in issue, VERITAS US had one primary commercial product (i.e., a product with a low price point and high-volume sales), Backup Exec, and five primary enterprise products (i.e., products with a high price point and low-volume sales): NetBackup, Volume Manager, File System, Cluster Server, and Foundation Suite.

Backup Exec, which was targeted to small businesses, was a data management product that provided backup, archive, and restore capabilities for a network's servers and workstations. NetBackup, Volume Manager, File System, Cluster Server, and Foundation Suite were purchased by businesses with large sophisticated information technology systems. NetBackup provided backup, archive, and restore capabilities for servers and workstations using complex UNIX, Windows, Linux, and NetWare operating systems. Volume Manager allowed an administrator to manage volumes (i.e., physical disks or hard drives that stored data) and also provided online disk storage management. File System was a journaling system that provided a directory index of files and made it easier to find and access files and data. File System also enabled fast system recovery from operating system failure or disruption. Cluster Server allowed multiple servers to be grouped together as a cluster and, if one server failed, another server was automatically activated to perform the functions of the failed server. Volume Manager had Cluster Server functionality by 1999 and File System had such functionality by 2000. Foundation Suite combined Volume Manager and File System to deliver a complete solution for online disk and file management functions. The consolidated product facilitated quicker and more efficient data transmission, storage, and backup. Foundation Suite was also sold in a high availability version. This version combined Foundation Suite and Cluster Server and ensured continuous uninterrupted operation in the event of system failure.

Many of VERITAS US' products were deemed "sticky" because after employing them it was difficult, costly, and time consuming for the user to change to a competing product. These products communicated with and controlled parts of the computer and its attached devices without support from standard application program interfaces (API)⁷ or device drivers. Consequently, the software code in these products included code inextricably tied to the most basic part of an operating system. In 1999 VERITAS US software products could run on systems and applications manufactured by Sun Microsystems, Inc. (Sun); Hewlett-Packard Co. (HP); Microsoft Corp. (Microsoft); International Business Machines Corp. (IBM); Red Hat, Inc.; Apple Inc.; Novell, Inc. (Novell); Oracle Corp. (Oracle); SAP AG; Sybase, Inc.; and VMware, Inc. After the CSA, VERITAS US released numerous versions of its aforementioned products. Each version contained new features. When new features were added to a product, the source code relating to these features was either added to existing files or placed in newly created files. While no one feature modification significantly altered the essential elements of the code, the cumulative effect of modifying hundreds of features typically resulted in significant code changes.

From November 1999 to 2006 OEM licensees paid VERITAS US $1.327 billion in royalties.⁸ The calculation of royalties was based on list price, revenues, or profits and the products were often sold at a discount off list price. VERITAS US generally received a one-time license fee upon entering into the agreement and additional license fees each time the OEM sold VERITAS US products bundled with an operating system. The royalty rates relating to VERITAS US' OEM licenses ranged from 10 to 40 percent for bundled products and 5 to 48 percent for unbundled products. Profit potential and sales volume were important factors in determining royalty rates. VERITAS US could not accurately predict the amount of its license revenue receipts attributable to OEM agreements because VERITAS US had no control over delivery dates or the number of VERITAS US products sold with OEM operating systems. This uncertainty led VERITAS US to explore other paths to market (i.e., distributors, resellers, and direct sales) for its products.

VERITAS US sold Backup Exec through distributors and resellers. The distributors sold Backup Exec to resellers and the resellers sold it to customers. VERITAS US sold NetBackup, Volume Manager, File System, Cluster Server, and Foundation Suite directly to customers and through resellers. Between 1997 and 2005 VERITAS US entered into reseller agreements with operating system, hardware, and database vendors including Compaq, Hitachi, Fujitsu, Ericsson, Dell, HP, NCR, Bull S.A., and EMC Corp. (EMC). The royalty rates relating to the reseller agreements ranged between 32.5 and 70 percent.

VERITAS US products competed with both comparable and free alternatives. The free alternatives included storage management products readily accessible on the Internet and those bundled with operating systems. Vendors sometimes incorporated storage management capabilities into their operating systems. Some customers preferred operating systems with built-in storage management software (i.e., integrated stacks).⁹ These stacks were less expensive and easier to deploy because purchasers were not required to acquire or install costly individual components. VERITAS US continuously sought to offer products that were faster and more efficient than comparable products or free alternatives. The performance advantages of VERITAS US products over free products decreased, however, as competitors improved their products' functionality.

Between 1996 and 2006 the primary competition for VERITAS US products was products sold by operating system, hardware, and database vendors such as Sun and Oracle. Sun and Oracle had a similar objective -- remove VERITAS US from their respective stacks and provide their respective customers with viable alternatives to VERITAS US products.

Sun, an operating system manufacturer and distributor, was one of VERITAS US' main OEM contractors and competitors. The relationship between VERITAS US and Sun evolved from a mutually beneficial partnership in 1997 to mutual tolerance in 1999 and, ultimately, to outright competition in 2006. Sun was committed to capturing the funds that its customers were spending on VERITAS US products. Sun upgraded its operating systems in an attempt to replicate the functionality of VERITAS US products and achieved this goal by adding to its operating system a supplanting product that was provided to customers at no cost. From 2000 through 2006 Sun released a series of operating systems that included software products that offered progressively more functionality. These products took market share from VERITAS US and closed the technology gap between Sun and VERITAS US.

Oracle, a software manufacturer known for its databases and applications, was as aggressive as Sun in competing with VERITAS US. Oracle offered software to directly, and successfully, compete with File System, Volume Manager, and Cluster Server. In an attempt to maximize revenues and customer loyalty, Oracle embarked on a strategy to build a complete stack and compete not just with VERITAS US, but also with operating system vendors. Oracle started with basic level technology and continued to innovate until it developed products similar, and ultimately, equal to VERITAS US products.

At the time of the CSA, VERITAS US products, on average, had a useful life of 4 years. In 2001 VERITAS US' board of directors realized that VERITAS US' primary products were approaching the end of their lifecycles and that the product pipeline was not capable of sustaining business growth. In 2002 and 2003 the board of directors recognized that revenues relating to Backup Exec and NetBackup had ceased to grow and that the revenues relating to Volume Manager and File System were declining. NetBackup's useful life came to an end in 2005 when a major overhaul was performed. Even as the products approached the end of their useful lives, they did not lose all of their value.

VERITAS US typically updated its products but, on occasion, an OEM would pay VERITAS US to build a custom item that would not be further developed. In these instances, the related OEM agreements contained a royalty degradation or technology aging discount provision to account for obsolescence and decay. Some agreements provided for the royalties to be decreased at a steady rate while others required royalty rate reductions that increased during the term of the agreement. Generally, the agreements did not provide a royalty rate reduction of more than 75 percent over a 4-year period.

In the EMEA storage management software market (i.e., in which VERITAS US sold Foundation Suite, NetBackup, and Backup Exec), VERITAS US' market shares in 1998 and 1999 were 8.9 percent and 13.2 percent, respectively. Computer Associates' ARCserve, Backup Exec's primary competition, dominated the EMEA market, holding more than 50 percent of the United Kingdom market and more than 60 percent of the French, Italian, and Spanish markets.

In 1999 the EMEA and APJ territories accounted for 92 percent of VERITAS US' international revenues and 22 percent of VERITAS US' total revenues (i.e., EMEA revenue totaled $110 million and APJ revenue was de minimis). VERITAS US' management recognized that geographic expansion in EMEA and APJ presented an opportunity to increase sales. After evaluating the cost of labor, employment laws, quality of workforce, and tax considerations, VERITAS US' management decided to headquarter its EMEA and APJ operations in Ireland.

Effective November 3, 1999, VERITAS US assigned to VERITAS Ireland all of VERITAS US' existing sales agreements with European-based sales subsidiaries (i.e., VERITAS UK, VERITAS Sweden, VERITAS Switzerland, VERITAS France, and VERITAS Germany). Also effective on that date, VERITAS US, VERITAS Operating Corp., NSMG, and VERITAS Ireland¹⁰ entered into the Agreement for Sharing Research and Development Costs (RDA), and VERITAS US and VERITAS Ireland¹¹ entered into the Technology License Agreement (TLA).¹²

Pursuant to the RDA, the signatories agreed to pool their respective resources and R&D efforts related to software products and software manufacturing processes. They also agreed to share the costs and risks of such R&D on a going-forward basis. The RDA provided VERITAS Ireland with:

In 1999 VERITAS Ireland paid VERITAS US $6.3 million and agreed to prepay VERITAS US, in 2000, the remaining consideration relating to the preexisting intangibles. In 2000 VERITAS Ireland made a $166 million lump-sum buy-in payment to VERITAS US, and in 2002 VERITAS Ireland and VERITAS US adjusted the payment to $118 million.

VERITAS Ireland developed the EMEA and APJ markets without significant input from VERITAS US. In 1999 VERITAS US' customer base had little or no value because of its minimal market share and limited presence in EMEA and APJ. At that time there were two offices in the United Kingdom: One in Chertsey and one in Reading. The Chertsey office was staffed by direct sales employees. The Reading office was staffed by two small teams (i.e., a distribution team and a reseller team) of inept workers. VERITAS Ireland focused on the basics of building a more extensive sales business and stronger distribution channels. In 2000 VERITAS Ireland hired a new distribution sales manager who was responsible for expanding its products' paths to market. VERITAS Ireland's new management totally changed the culture by continually upgrading VERITAS Ireland's sales resources; examining distributor and reseller reports; finding new customers; initiating interaction with the reseller base; providing sales incentives for distributors; training and educating the distributors' presales teams; and firing underperforming salesmen, distributors, and resellers. To further expand its sales presence, VERITAS Ireland accessed and leveraged its distribution partners' sales organizations and customer contacts.

VERITAS Ireland's operations and its presence in the EMEA territory grew substantially from 2000 to 2006. By 2001 the Ireland facility had increased from 12,000 to 40,000 square feet and the number of VERITAS Ireland employees had increased from 20 to more than 100. By 2002 VERITAS Ireland had over 25 new offices and subsidiaries in 19 countries, and by 2004 VERITAS Ireland had more than 1,500 employees in more than 30 countries. From 1999 to 2006 VERITAS Ireland spent $1.374 billion on sales and marketing expenses, $676 million on cost-sharing payments, $456 million on customer service expenses, $146 million on administrative expenses, and $124 million on buy-in payments. In 2000 VERITAS Ireland's first full year of operation, revenues were approximately $200 million. By 2003 VERITAS Ireland's license revenues had doubled, and by 2004 its annual revenues were five times higher than VERITAS US' 1999 revenues attributable to EMEA and APJ.

Respondent examined VERITAS US' 2000 and 2001 returns and concluded that the cost-sharing allocations reported did not clearly reflect VERITAS US' income. On March 29, 2006, respondent issued petitioner a notice of deficiency based on a report prepared by Brian Becker (Becker). In the notice, respondent stated:

On June 26, 2006, petitioner timely filed its petition with the Court seeking redetermination of the deficiencies and penalties set forth in the notice. On August 25, 2006, the Court filed respondent's answer, and on August 31, 2006, the Court filed respondent's amended answer. Respondent, in his statement of position filed September 6, 2007, stated: "In view of the fact that information is still being collected and analyzed, Respondent cannot state which transfer pricing method(s) he intends to utilize at trial." On October 11, 2007, respondent, in a supplement to his statement of position, notified the Court and petitioner that he was going to employ the forgone profits method, but was not going to rely on the market capitalization method or call Becker as a witness. Respondent, in the October 11, 2007, statement also stated:

On January 11, 2008, the Court filed petitioner's motion for partial summary judgment. In the motion, petitioner contended that respondent had abandoned the $2.5 billion allocation and the methodologies set forth in the notice; the notice was fundamentally defective; and respondent's determination was arbitrary, capricious, and unreasonable. Petitioner further contended that, pursuant to precedent governing the Court of Appeals for the Ninth Circuit (Ninth Circuit), the burden of proof shifts to respondent. The Court, on February 6, 2008, filed respondent's notice of objection to petitioner's motion for partial summary judgment.

On March 7, 2008, respondent submitted to the Court an expert report prepared by John Hatch (Hatch). Hatch, employing a discounted cashflow analysis, concluded that the requisite lump-sum buy-in payment was $1.675 billion, and calculated, as an alternative, a 22.2-percent perpetual annual royalty. In determining the best method to calculate the buy-in payment, Hatch rejected the comparable uncontrolled transaction method (CUT method)¹⁹ and the profit split method²⁰. He contended that prior to November 3, 1999, VERITAS US had made several acquisitions of software companies that offered complementary, and in some cases, competing products. Hatch opined that those acquisitions were comparable to the CSA because VERITAS US received rights pursuant to the acquisitions that were similar to those which VERITAS Ireland received pursuant to the CSA. On the basis of his findings, Hatch characterized the CSA as "akin" to a sale or geographic spinoff ("akin" to a sale theory) and employed the income method to determine the requisite buy-in payment.

Hatch defined the buy-in payment as "the present value of royalty obligations expected to be paid under arm's length royalty terms applicable to the rights conferred on a go-forward basis." He did not individually value any of the specific items that were allegedly transferred to VERITAS Ireland. Instead, he employed an "aggregate" valuation approach that was based on a three-step analysis. First, Hatch estimated the arm's-length royalty amounts that would be due in each period (i.e., each calendar year or portion thereof after November 3, 1999) of the CSA. Second, Hatch chose a discount rate to convert estimated future royalty payments into November 1999 dollars. Third, Hatch calculated the buy-in payment as equal to the present value of the royalty payments estimated in step 1, discounted at the rate determined in step 2. Hatch concluded that the requisite buy-in payment was $1.675 billion and that a 22.2-percent perpetual annual royalty was economically equivalent to the requisite $1.675 billion payment. In calculating the requisite buy-in payment, Hatch assumed that the preexisting intangibles have a perpetual useful life. In addition, he concluded that 13.7 percent was the appropriate discount rate and 17.91 percent was the appropriate compound annual growth rate.

On March 21, 2008, the Court filed respondent's motion for leave to file amendment to amended answer and lodged respondent's amendment to amended answer. In the proposed amendment, respondent alleged that the requisite buy-in payment was $1.675 billion, payable as either a lump-sum payment or a 22.2-percent perpetual royalty. In paragraph 9.f of the proposed amendment, respondent asserted an adjustment relating to a transfer of "certain other intangible rights." Respondent specifically alleged a transfer of access to VERITAS US' marketing team; access to VERITAS US' R&D team; and VERITAS US' trademarks, trade names, customer base, customer lists, distribution channels, and sales agreements (collectively, paragraph 9.f items). Petitioner, in its notice of objection to respondent's motion for leave to file amendment to amended answer filed April 10, 2008, contended that respondent's assertion of the paragraph 9.f items raised a new matter because the issue was not described in the notice of deficiency and required the presentation of new evidence.

On May 2, 2008, the Court held a hearing (May 2 hearing) relating to the aforementioned motions. In an order issued June 13, 2008 (June 13 order), we denied petitioner's motion for partial summary judgment and concluded that there was a genuine issue with respect to whether respondent had abandoned the theory and methodology set forth in the notice, petitioner had failed to establish that the notice was fundamentally defective, and petitioner had therefore failed to establish that the determination was arbitrary, capricious, or unreasonable. With respect to whether the burden of proof shifts to respondent, we concluded that it was premature to rule on the issue. We also granted respondent's motion for leave to file amendment to amended answer and concluded that the notice of deficiency was sufficiently broad to include the paragraph 9.f items and, therefore, respondent's amendment to amended answer did not raise a new matter. We stated:

In essence, respondent's determination began to unravel with the parties' pretrial stipulations of settled issues. After the parties' settlement relating to the arm's-length value of the RDA, as a practical and legal matter respondent was forced to justify the $1.675 billion allocation by reference only to the preexisting intangibles. As discussed herein, he simply could not. Respondent, in a futile attempt to escape this dilemma, ignored the parties' settlement relating to the RDA and disregarded section 1.482-7(g)(2), Income Tax Regs., which limits the buy-in payment to preexisting intangibles. In addition, respondent inflated the determination by valuing short-lived intangibles as if they have a perpetual useful life and taking into account income relating to future products created pursuant to the RDA.

After an extensive stipulation process, a lengthy trial, the receipt of more than 1,400 exhibits, and the testimony of a myriad of witnesses, our analysis of whether respondent's $1.675 billion allocation is arbitrary, capricious, or unreasonable hinges primarily on the testimony of Hatch. Put bluntly, his testimony was unsupported, unreliable, and thoroughly unconvincing. Indeed, the credible elements of his testimony were the numerous concessions and capitulations.

Respondent's predicament was primarily attributable to the implausibility of respondent's flimsy determination. In calculating the $1.675 billion allocation, Hatch used the wrong useful life for the products and the wrong discount rate and admittedly did not know precisely which items were valued. Furthermore, respondent's trial position reflected sections 1.482-1T through 1.482-9T, Temporary Income Tax Regs., 74 Fed. Reg. 349 (Jan. 5, 2009) -- regulations that were promulgated 10 years after the transaction and 5 months after trial.²¹ These regulations include specific examples involving "assembled workforce"²² and prescribe the income method as a specified method. In fact, after amending his amended answer, respondent began referring to the intangibles subject to the buy-in payment as "platform contribution" intangibles (i.e., the term used in sections 1.482-1T through 1.482-9T, Temporary Income Tax Regs., supra) rather than "pre-existing intangibles"²³ (i.e., the term used in the applicable regulations). We further note that the Administration, in 2009, proposed to change the law, expanding the section 482 definition of intangibles to include "workforce in place",²⁴ goodwill, and going-concern value.²⁵ See Department of the Treasury, General Explanations of the Administration's Fiscal Year 2010 Revenue Proposals 32 (May 2009). For the years in issue, however, there was no explicit authorization of respondent's "akin" to a sale theory or its inclusion of workforce in place, goodwill, or going-concern value. Taxpayers are merely required to be compliant, not prescient.

Pursuant to the law in effect at the time of the CSA, respondent's determination is arbitrary, capricious, and unreasonable, and VERITAS US' CUT method, with some adjustments, is the best method to determine the requisite buy-in payment.

Section 482 provides that in the case of any transfer of intangible property the income with respect to the transfer shall be commensurate with the income attributable to the intangible. In a qualified cost-sharing arrangement, controlled participants share the cost of developing one or more items of intangible property. See sec. 1.482-7(a)(1), Income Tax Regs. When a controlled participant makes preexisting intangible property available to a qualified cost-sharing arrangement, that participant is deemed to have transferred interests in the property to the other participant and the other participant must make a buy-in payment as consideration for the transferred intangibles. Sec. 1.482-7(g)(1) and (2), Income Tax Regs. The buy-in payment, which can be made in the form of a lump-sum payment, installment payments, or royalties, is the arm's-length charge for the use of the transferred intangibles. Sec. 1.482 7(g)(2), (7), Income Tax Regs.

Section 1.482-7(g)(2), Income Tax Regs., requires buy-in payments to be determined in accordance with sections 1.482-1 and 1.482-4 through 1.482-6, Income Tax Regs. Section 1.482-4(a), Income Tax Regs., provides:

Respondent's determination as set forth in the notice of deficiency is presumptively correct. Id. at 353. Respondent made two determinations with respect to the requisite buy-in payment, one set forth in the notice of deficiency and one set forth in the amendment to amended answer. Because we found in the June 13 order that the amendment to amended answer did not raise a new matter, the presumption of correctness that attached to the determination set forth in the notice carried forward to the revised determination set forth in the amendment to amended answer. See Shea v. Commissioner, 112 T.C. 183 (1999). Thus, we look to both the notice determination and the revised determination in the amendment to amended answer to decide whether respondent's section 482 allocation is arbitrary, capricious, or unreasonable.

A. Respondent's Notice Determination Is Arbitrary, Capricious, and Unreasonable

In the notice, respondent determined, using Becker's valuation, that the requisite buy-in payment was $2.5 billion. During trial respondent did not call Becker as a witness, place Becker's report in evidence, or present any evidence to support Becker's findings. Respondent, relying solely on the report prepared by Hatch, did not address Becker's $2.5 billion buy-in valuation but instead asserted a $1.675 billion buy-in valuation. The $825 million decrease in value with little explanation is just one of the factors we consider in evaluating the reasonableness of respondent's determination. There are other factors that collectively and convincingly establish that the notice determination was not only unreasonable but was also arbitrary and capricious. Using an income method, Becker and Hatch, respectively, employed a 12.8- and a 13.7-percent discount rate to calculate the requisite buy-in payment. Beta, a key component in the formula used to calculate the discount rate, is a measure of the tendency of a security's price to respond to swings in the market.²⁶ In calculating their discount rates, Becker and Hatch used essentially the same beta, 1.4 and 1.42, respectively. Petitioner's finance expert established that 1.935 was the correct beta. See infra, Discussion, sec. IV(D), The Appropriate Discount Rate. Hatch ultimately conceded that a 1.42 beta "could not, to a reasonable degree of economic certainty, be the correct beta." See infra, Discussion, sec. II(B)(4), Respondent Employed the Wrong Useful Life, Discount Rate, and Growth Rate. In essence, Hatch admitted that both he and Becker employed the wrong beta. Indeed, the beta Becker employed was even further removed from the correct beta.

In sum, respondent, without meaningful explanation, conceded $825 million of the buy-in amount set forth in the notice and at trial failed to offer even a token defense in response to petitioner's critique of Becker's conclusions. Moreover, respondent cannot convincingly contend that the notice allocations are reasonable while adopting the opinion of an expert who admits that a critical factor relating to the calculation of the allocation is incorrect. Accordingly, respondent's notice determination is arbitrary, capricious, and unreasonable.

B. Respondent's Determination in Amendment to Amended Answer Is Arbitrary, Capricious, and Unreasonable

Respondent's amendment to amended answer set forth a revised determination of the requisite buy-in payment. The revised determination, which is based on Hatch's report, takes into account certain items (i.e., the paragraph 9.f. items) that respondent alleges were intangibles transferred to VERITAS Ireland. Hatch's valuation was based on the theory that the collective effect of the RDA, TLA, and conduct of the parties was "akin" to a sale of VERITAS US' business. Respondent's determination is erroneous for several reasons.

Respondent contends that pursuant to section 1.482-1(f)(2)(i)(A), Income Tax Regs., he was authorized to aggregate the transactions and treat them as a sale. Transactions may be aggregated if an aggregated approach produces the "most reliable means of determining the arm's length consideration for the controlled transactions". Id. (emphasis added). Respondent's "akin" to a sale theory (i.e., a theory which encompasses short-lived intangibles valued as if they have a perpetual life²⁷ and takes into account intangibles that were subsequently developed rather than preexisting)²⁸ certainly does not produce the most reliable result. Thus, pursuant to section 1.482-1(f)(2)(i)(A), Income Tax Regs., respondent was not authorized to aggregate the transactions and treat them as a sale.²⁹

With respect to distribution channels, VERITAS US had relationships with distributors and resellers prior to the CSA, but those relationships were weak and had little value. In fact, it was not until VERITAS Ireland hired the channel manager from Computer Associates that the distribution channels were strengthened and maximized. Thus, to the extent VERITAS US' distribution channels were transferred to VERITAS Ireland, they had insignificant value. With respect to customer lists and customer base, Hatch agreed that, prior to the CSA, VERITAS US lacked the data systems needed to generate accurate and meaningful customer lists and that VERITAS US' customer base had no value given VERITAS US' marginal market share and limited presence in EMEA and APJ. Thus, to the extent VERITAS US' customer lists and customer base were transferred to VERITAS Ireland, they had insignificant value. With respect to "access to research and development team", Hatch testified that his valuation of the buy-in payment did not include access to R&D team and that access to R&D team "just was not on [his] radar screen or anything that [he] thought of." In addition, Hatch conceded that if he assumed that the agreement relating to the share of R&D expenses was arm's length, a fact that the parties stipulated, then access to the R&D team would have zero value. With respect to "access to marketing team", Hatch testified that he did not value VERITAS US' marketing team, did not know whether marketing support was provided by VERITAS US, and had no idea whether the alleged marketing intangibles existed or had been transferred. Hatch further testified:

Section 1.482-7(g)(2), Income Tax Regs., the regulatory authority requiring a buy-in payment, states:

In calculating his valuation of the buy-in payment, Hatch assumed a perpetual useful life for the transferred intangibles, yet acknowledged that "if you had 1999 products that you left untouched, that technology would age and eventually become obsolete" and that the preexisting product intangibles would "wither on the vine" within 2 to 4 years without ongoing R&D. The useful life of the preexisting product intangibles was, on average, 4 years, and certainly was not perpetual. Petitioner established that something, however, was perpetual -- VERITAS US was in a perpetual mode of innovation. Before and after the CSA VERITAS US released numerous versions of its products. Even with substantial ongoing R&D, VERITAS US products had finite lifecycles. By the time a new product became available for purchase, the next generation was already in development.

In determining the discount rate³² for the buy-in payment, Hatch used a weighted average cost of capital (WACC)³³ derived under the capital asset pricing model (CAPM).³⁴ Employing the CAPM,³⁵ Hatch used, as the risk-free rate, the yield on 20-year U.S. Treasury bonds as of March 31, 2000, without adjustments, and determined an equity risk premium of 5 percent. The equity risk premium is the expected long-term yield for the stock market less the risk-free rate. Hatch applied the 5-percent equity risk premium and an industry beta of 1.42³⁶ to calculate the applicable discount rate, which he concluded was 13.7 percent.

Petitioner contends that respondent employed the wrong beta, the wrong equity risk premium, and therefore the wrong discount rate. Hatch employed an industry beta to calculate the discount rate. He opined that using an industry, rather than a company specific, beta was preferred because, with respect to an individual company, a beta relating to an earlier period is a very poor predictor of the beta for subsequent periods. Hatch ultimately admitted, however, that "to a reasonable degree of economic certainty, the beta he used could not have been the correct beta for VERITAS US as of November 3, 1999."

Hatch's 5-percent equity risk premium was much lower than the 1926 through 1999 historic average of 8.1 percent which Hatch stated was reported by Ibbotson Associates (i.e., the recognized industry standard of historical capital markets data).³⁷ There are several problems with Hatch's analysis. First, in determining the equity risk premium, Hatch contended that employing the Ibbotson Associates' historic average equity risk premium, which was based on the expected long-term yield for the U.S. stock market, was not appropriate because the rights licensed to VERITAS Ireland were exploited in markets outside the United States. Rights licensed to VERITAS Ireland were indeed exploited outside the United States, but Hatch erroneously assumed that the long-term yield for the U.S. market was higher than the long-term yield for foreign markets. In fact, the literature upon which Hatch relied establishes that there was no difference between the observed risk premium in the U.S. market and the risk premium in foreign markets. See Brealey & Myers, Principles of Corporate Finance 159 (7th ed. 2003). Hatch's erroneous assumptions led to an underestimate of the appropriate equity risk premium relating to the buy-in payment.

Second, in determining the equity risk premium, Hatch applied the 20-year U.S. Treasury bond yield as the risk-free rate. Petitioner contends that the classic formulation of CAPM uses the 30-day U.S. Treasury bill rate as the risk-free rate, not the bond rate, and that if the bond rate is used, duration risk has to be taken into account. Ibbotson Associates' Cost of Capital 2000 Yearbook 34 states: "In all of the beta regressions, the total returns of the S&P 500 are used as the proxy for the market returns. The series used as a proxy for the risk-free asset is the yield on the 30-day T-bill." Furthermore, the text Hatch cites as support for his use of the U.S. Treasury bond rate states that "The risk-free rate could be defined as a long-term Treasury bond yield. If you do this, however, you should subtract the risk premium of Treasury bonds over bills". See Brealey & Myers, supra at 226 n. 8. Hatch, however, did not reduce the U.S. Treasury bond rate and, on cross-examination, acknowledged that he used the wrong risk-free rate. In sum, Hatch employed the wrong beta, the wrong equity risk premium, and thus the wrong discount rate to calculate the requisite buy-in payment.

Hatch also employed large and unrealistic growth rates into perpetuity. Hatch determined that from 2001 through 2005 VERITAS Ireland's compound annual growth rate was 17.91 percent. He projected that VERITAS Ireland's revenues would increase 13 percent each year from 2007 through 2010 and beginning January 1, 2011, would increase 7 percent each year into perpetuity. VERITAS Ireland's actual growth rate between 2004 and 2006 was 3.75 percent, 14.16 percentage points lower than the 17.91 percent growth rate Hatch employed for the same period. In calculating the buy-in payment, Hatch used VERITAS Ireland's actual income relating to 2004 through 2006 but opted not to use actual growth rates relating to those years. Moreover, he could not provide a plausible explanation for the growth rate he employed. Further, petitioner notes that a buy-in payment based on Hatch's growth rate would require VERITAS Ireland to allocate a buy-in payment equal to 100 percent of its actual and projected operating income to VERITAS US through 2009, resulting in $1.9 billion in losses over that period. Simply put, the growth rate Hatch employed was unreasonable.

In sum, VERITAS Ireland prospered, not because VERITAS US simply spun off a portion of an established business and transferred valuable intangibles, but because VERITAS Ireland employed aggressive salesmanship and savvy marketing, successfully developed the EMEA and APJ markets, and codeveloped new products that performed well in those markets. For the foregoing reasons, we conclude that respondent's allocations set forth in the amendment to amended answer and at trial are arbitrary, capricious, and unreasonable.

The CUT method evaluates whether the amount charged for a controlled transfer of intangible property is arm's length by referencing the amount charged in comparable uncontrolled transactions. If an uncontrolled transaction involves a transfer of the same intangible under the same, or substantially the same, circumstances as a controlled transaction, the results derived from applying the CUT method will generally be the most reliable measure of the arm's-length result. Sec. 1.482-4(c)(2)(ii), Income Tax Regs. If, however, uncontrolled transactions involving the same intangible under the same circumstances cannot be identified, uncontrolled transactions that involve the transfer of "comparable intangibles under comparable circumstances" may be used to apply the CUT method, but the reliability of the results is reduced. Id.

Respondent contends that the CUT method is not the best method and that petitioner has not presented comparable uncontrolled transactions to prove that its buy-in payment is arm's length. Specifically, respondent asserts that the rights licensed under agreements between VERITAS US and unrelated parties are not comparable because they involved either rights that are not comparable to those licensed under the CSA or licensees who are not comparable to VERITAS Ireland. Petitioner contends that the CUT method is appropriate and that the value determined by its expert, William Baumol (Baumol), was arm's length.

Baumol calculated, using the CUT method, a range of estimates for the value of the transferred intangibles and concluded that the lump-sum buy-in payment was within or exceeded the arm's-length range. Baumol used four parameters to estimate a value for the buy-in payment: The expected economic life of the intangibles, the annual rate at which the value of the intangibles declines as a function of time and new software replacements (i.e., the rate of obsolescence), the parameter value selected to determine the value of the licenses (e.g., royalty rates as a percent of revenues, list price, or profits), and the appropriate discount rate.³⁸

Baumol chose particular agreements (i.e., some involving bundled products and some involving unbundled products) between VERITAS US and seven OEMs (i.e., Sun, HP, Dell, Hitachi, NEC, Compaq, and Ericsson) to determine the appropriate starting royalty rate for the buy-in payment. Most of the product licenses that Baumol selected provide royalties as a percentage of list price (e.g., global list price, international list price, or U.S. list price). Based on his findings, Baumol derived a range of starting royalty rates of 20 to 25 percent of list price and opined that the low end of the range, 20 percent, was the appropriate starting royalty rate for the buy-in payment.

Baumol determined that the preexisting product intangibles had a useful life ranging from 2 to 4 years. Having determined both the starting royalty rate and the useful life, Baumol adjusted the royalty rate by ramping down (i.e., incrementally reducing) the rate over the buy-in period. Baumol analyzed royalty degradation and technology aging provisions in third-party agreements as evidence of the appropriate ramp-down rates. To confirm his ramp-down conclusions, Baumol relied on petitioner's source code expert, who opined that new lines of code noticeably increased after 1999 while the amounts of unchanged functional 1999 source code and files were virtually nonexistent within a period of 3 to 4 years.

Using the aforementioned findings, Baumol calculated a valuation range of $94 million to $315 million for the buy-in payment and concluded that "the preponderance of the values" fell between $100 million and $200 million.

A. Comparability of OEM Agreements

Use of the CUT method requires that the controlled and uncontrolled transactions involve the same or comparable intangible property. Sec. 1.482-4(c)(2)(iii)(A), Income Tax Regs. In order for intangibles involved in controlled and uncontrolled transactions to be comparable, "both intangibles must -- (i) Be used in connection with similar products or processes within the same general industry or market; and (ii) Have similar profit potential." Sec. 1.482-4(c)(2)(iii)(B)(1), Income Tax Regs.

In his CUT valuation, Baumol referenced, as comparables, agreements between VERITAS US and certain OEMs (i.e., Sun, HP, Dell, Hitachi, NEC, Compaq, and Ericsson). Respondent contends that the CSA involves the transfer of "platform contribution" intangibles and broad "make-sell rights"³⁹ with respect to VERITAS US' full range of products, while the OEM agreements did not. We note that the term "platform contribution intangibles" does not appear in the regulations applicable to the CSA but is set forth in section 1.482-7T, Temporary Income Tax Regs., 74 Fed. Reg. 352 (Jan. 5, 2009) -- regulations effective for transactions entered into on or after January 5, 2009. Thus, respondent's litigating position appears to mirror transfer pricing regulations promulgated 10 years after VERITAS US and VERITAS Ireland signed the CSA.⁴⁰ In essence, respondent contends that, pursuant to section 1.482-4(c)(2)(iii)(A), Income Tax Regs., the CUT method is not appropriate because the OEM agreements involve substantially different intangibles. We disagree.

VERITAS Ireland, pursuant to the TLA, received broad rights for the full range of VERITAS US products. The rights licensed under the OEM agreements referenced by Baumol involved Backup Exec, NetBackup, Volume Manager, File System, Cluster Sever, and Foundation Suite. While none of the individual OEM agreements evaluated by Baumol included a license for the full range of VERITAS US' product line, collectively the agreements did involve essentially the same intangibles that were transferred from VERITAS US to VERITAS Ireland. The OEM agreements Baumol selected do not, however, provide the most reliable measure for calculating the requisite buy-in payment.

B. Unbundled OEM Agreements Were Comparable to the Controlled Transaction

VERITAS US entered into numerous OEM agreements prior to and during the CSA. Baumol chose to use only a select few of those OEM agreements (i.e., some involving bundled products and some involving unbundled products) to calculate the requisite buy-in payment. His justification for rejecting particular agreements was simply: "I didn't find the numbers that I could use." Respondent contends that the OEM agreements Baumol selected are not comparable to the controlled transaction because the circumstances surrounding the selected OEM agreements and the circumstances surrounding the controlled transaction are different. We conclude that, collectively, the more than 90 unbundled OEM agreements the parties stipulated are sufficiently comparable to the controlled transaction.

When OEMs sold VERITAS US products bundled with the OEMs' operating systems, VERITAS US gained credibility and improved brand identity. The OEMs actively marketed the bundled products; listed the products on their Web sites; and provided equipment, technical support, and engineering assistance for those products. Because of these factors, OEMs paid a lower royalty rate with respect to bundled products. VERITAS Ireland, on the other hand, did not have a trade name as widely recognized as the trade names of the OEMs, guaranteed sales like the OEMs, or an operating system with which to bundle VERITAS US products. Therefore, VERITAS Ireland would not be entitled to similar royalty rates. In contrast to bundled products, unbundled products were not directly associated with the OEMs' products and the OEMs did not provide the same level of assistance (i.e., technical and engineering support). Thus, customers did not perceive unbundled products to be more reliable or of greater quality than other comparable products. The OEMs merely listed the unbundled products as an option (i.e., customers could purchase VERITAS US products or other products). Because such agreements are more comparable to the transaction between VERITAS US and VERITAS Ireland, use of the OEM agreements involving unbundled products provides a more reliable arm's-length result. Thus, we compare VERITAS US' unbundled OEM agreements with the controlled transaction.

The degree of comparability between controlled and uncontrolled transactions is determined by applying the comparability standards set forth in section 1.482-1(d), Income Tax Regs. Sec. 1.482-4(c)(2)(iii), Income Tax Regs. Section 1.482-1(d)(1), Income Tax Regs., provides that the following factors shall be considered in determining comparability between controlled and uncontrolled transactions: Functions, contractual terms, risks, economic conditions, and property or services. An analysis employing these factors confirms that VERITAS US' unbundled OEM agreements are sufficiently comparable to the controlled transaction.

The first factor, functional analysis, compares the economically significant activities undertaken, or to be undertaken, in the controlled transactions with the economically significant activities undertaken, or to be undertaken, in the uncontrolled transactions. Sec. 1.482-1(d)(3)(i), Income Tax Regs. VERITAS Ireland and the OEMs undertook similar activities (e.g., manufacturing and production, marketing and distribution, transportation and warehousing, etc.) and employed similar resources in conjunction with such activities. See section 1.482-1(d)(3)(i), Income Tax Regs., for a list of functional analysis comparability factors. Respondent contends, however, that the OEM agreements and the controlled transactions are not functionally comparable because R&D is a particularly significant function in the controlled transactions (i.e., VERITAS US and VERITAS Ireland agreed to share in ongoing R&D costs relating to the development of new software products), whereas the OEM agreements did not involve ongoing R&D activities. Respondent contends that the R&D function is important because VERITAS Ireland "received ownership interests in future generations of technology which germinated from the pre-existing technology." Respondent's functional analysis is misguided. Respondent is relying on rights involving subsequently developed intangibles to support his assertion that the OEM agreements are not comparable to the controlled transaction. As previously determined herein, VERITAS Ireland was required to make a buy-in payment with respect to the transfer of "pre-existing intangible property", not subsequently developed intangibles. See sec. 1.482-7(g)(2), Income Tax Regs. Thus, the focus of the buy-in payment analysis should be on transactions involving preexisting intangibles. For the products in existence on November 3, 1999, there are no significant differences in functionality.

The second factor is the comparability of contractual terms. Determining the degree of comparability between the controlled and uncontrolled transactions requires a comparison of the significant contractual terms that could affect the results of the transactions (e.g., the form of consideration; the sales volume; the scope and terms of warranties; the right to updates, revisions, or modifications; the duration of the agreement; etc.). Sec. 1.482-1(d)(3)(ii)(A), Income Tax Regs. Respondent contends that the contractual terms of the OEM agreements are not comparable to the controlled transaction for two reasons. First, respondent contends that the OEMs often provided VERITAS US with APIs, source code, or information about their hardware so VERITAS US could adapt VERITAS US products to the OEMs' hardware and operating systems, whereas VERITAS Ireland did not have an operating system, APIs, or source code. Some of VERITAS US' unbundled OEM agreements did contain contractual terms pursuant to which OEMs provided APIs and source code information to VERITAS US to assist with adaptation issues, but, unlike the contractual terms set forth in section 1.482-1(d)(3)(ii)(A), Income Tax Regs., the contractual terms relating to adaptability were not significant terms that affected the results of the transactions. The APIs and source code information did not change the essential functions of VERITAS US products but rather enabled VERITAS US products to run on the OEM's operating system. Second, respondent contends that the OEMs provided engineering assistance to VERITAS US in connection with the development of VERITAS US bundled products, whereas there is no evidence that VERITAS Ireland was in a position to provide engineering assistance to VERITAS US. While it is true that some OEMs did provide engineering support with respect to bundled products, the provision of engineering support was not a standard contractual term in OEM agreements relating to unbundled products. Indeed, the provision of engineering support was not a significant factor that affected the results of OEM agreements involving unbundled products. Thus, there are no significant differences in contractual terms.⁴¹

The third factor compares the significant risks borne by the parties that could affect the prices charged or the profit earned in the controlled and uncontrolled transactions. Sec. 1.482-1(d)(3)(iii), Income Tax Regs. The parties to the controlled and uncontrolled transactions bore similar market risks, similar risks associated with R&D activities, similar risks associated with fluctuations in foreign currency exchange rates and interest rates, similar credit and collection risks, and similar product liability risks. See section 1.482-1(d)(3)(iii)(A), Income Tax Regs., for a list of risk comparability factors. Respondent contends, however, that the risks borne by VERITAS Ireland and the risks borne by the OEMs are not comparable because the OEMs were subject to the risk that the version of technology they licensed would not do well in the market. VERITAS Ireland bore the same risk as the OEMs. In short, there are no significant differences in risks borne.

The fourth factor compares the significant economic conditions that could affect prices or profit in the controlled transaction to the significant economic conditions that could affect prices or profit in the uncontrolled transactions. Sec. 1.482-1(d)(3)(iv), Income Tax Regs. Respondent contends that the economic and market conditions affecting the OEM agreements are not comparable to those affecting the transaction between VERITAS US and VERITAS Ireland because, unlike VERITAS Ireland, the OEMs occupied significant positions in the market. Respondent further contends that the OEMs had established sales forces and relationships with resellers and distributors, whereas on the date of the transfer VERITAS Ireland was a startup with no customer relationships or other assets. We agree with respondent that the OEMs and VERITAS Ireland were at dramatically different stages of development and held different positions in the market. We note, however, that both the OEMs and VERITAS Ireland competed in similar geographic markets, incurred similar distribution costs, marketed products that faced similar competition, and were subject to similar economic conditions. See section 1.4821(d)(3)(iv), Income Tax. Regs., for a list of economic condition comparability factors. While certain economic conditions (e.g., interest rate fluctuations, general vicissitudes of the market, etc.) affect prices and profits for both startups and established businesses, the impact on a particular business may certainly depend on the business' economic stability and market position. Our analysis of this factor narrowly weighs against a finding of comparability.

The fifth factor compares the property or services provided in the controlled transaction to that provided in the uncontrolled transactions. Sec. 1.482-1(d)(3)(v), Income Tax Regs. Respondent contends that under the OEM agreements, VERITAS US generally contracted to provide only the development work necessary to ensure its products would work with the OEMs' products, whereas under the CSA, VERITAS US provided make-sell rights and preexisting intangibles for research to produce future generations of technology. Specifically, respondent contends that "VERITAS U.S. and VERITAS Ireland contracted to share all the costs of future R&D on future software generations and for each to hold separate exploitation rights. * * * Neither the property nor services were comparable." Once again, respondent's contention is misguided. Respondent is relying on rights involving subsequently developed intangibles to support his assertion that the OEM agreements are not comparable to the controlled transaction. As previously determined herein, pursuant to section 1.482-7(g)(2), Income Tax Regs., the requisite buy-in payment need not take into account subsequently developed intangibles. With respect to the controlled transaction involving the transfer of preexisting intangibles and the uncontrolled transactions involving VERITAS US' unbundled OEM agreements, there are no significant differences in property or services provided.

Although VERITAS US' unbundled OEM agreements are certainly not identical to the controlled transaction, an analysis of the comparability factors establishes that the unbundled OEM agreements are sufficiently comparable to the controlled transaction and that the CUT method is the best method to determine the requisite buy-in payment. There are, however, certain adjustments we must make to petitioner's CUT analysis to enhance its reliability.

Respondent contends that if the OEM agreements are comparable to the controlled transaction, petitioner's calculation of the starting royalty rate is nevertheless erroneous. In determining the requisite buy-in payment, Baumol used 20 percent as the starting royalty rate and acknowledged that he did not use any "sophisticated calculation" or "higher mathematics" to arrive at that rate. He based the 20-percent royalty rate on rates found in select OEM agreements involving bundled and unbundled products. As previously determined, OEM agreements involving unbundled products are the appropriate comparables. As petitioner did not use sufficiently comparable transactions in determining the starting royalty rate to calculate the requisite buy-in payment, and respondent has not provided a royalty rate other than one based on a perpetual royalty, the Court must determine the appropriate royalty rate.

The parties provided the Court with the royalty rates for more than 90 unbundled OEM agreements. Because each unbundled OEM agreement standing alone does not involve the full range of intangibles referenced in the TLA, the agreements must be looked at collectively. The royalty rates relating to VERITAS US unbundled products range between 25 and 40 percent. The mean (i.e., the average) royalty rate for VERITAS US' OEM agreements involving unbundled products is 32 percent of list price. Thus, we conclude that the starting royalty rate for the transferred product intangibles is 32 percent of list price.

B. The Appropriate Useful Life and Royalty Degradation Rate

The appropriate useful life of the preexisting product intangibles is 4 years. Indeed, as previously discussed, VERITAS US products, on average, had a useful life of that duration.⁴²

Licensing parties often agree to ramp down royalty rates to account for the gradual obsolescence of static technology.⁴³ Petitioner contends that the royalty rates for the preexisting product intangibles should be ramped down over the lives of the intangibles to account for obsolescence and decay of technology. The majority of VERITAS US' OEM agreements included provisions for updates and new versions, but the preexisting product intangibles transferred pursuant to the TLA did not. Thus, an adjustment must be made to the starting royalty rate to account for the static nature of the technology.⁴⁴ Consistent with VERITAS US' other agreements involving static technology,⁴⁵ the royalty rates for VERITAS US' preexisting product intangibles must be ramped down, starting in year 2, at a rate of 33 percent per year from the then-current percentage (i.e., 32 percent in year 1; 21 percent in year 2; 14 percent in year 3; and 10 percent in year 4).⁴⁶

C. Value of Trademark Intangibles and Sales Agreements

Petitioner contends that VERITAS US' trademarks, trade names, and brand names (trademark intangibles) lacked value because in 1999 "VERITAS" was registered in only a few foreign jurisdictions and was relatively unknown in the EMEA and APJ markets. Regardless of the number of foreign jurisdictions in which the "VERITAS" trademark was registered, the "VERITAS" trademark and the individual product names, especially "NetBackup" and "Backup Exec", were well known, respected, and valuable. Thus, pursuant to section 1.482-7(g), Income Tax Regs., VERITAS Ireland was required to pay VERITAS US a buy-in payment as consideration for those trademark intangibles.

Petitioner's trademark expert found that as of November 3, 1999, VERITAS US had trade names for Backup Exec, NetBackup, Volume Manager, File System, Cluster Server, and Foundation Suite, as well as certain other products. He believed that the value of the trademark intangibles was zero but nevertheless calculated another value for those intangibles. In calculating a value petitioner's trademark expert opined that the useful life of the trademark intangibles in VERITAS Ireland's territory should be no more than 7 years, selected a range of royalty rates from 0.5 to 1 percent of revenue, and concluded that before taxes the value for the trademark intangibles was between $1.7 and $3.4 million. He assumed that VERITAS Ireland was entitled to royalty-free use of the trademark intangibles for the duration of the TLA and concluded that the TLA, which did not have a termination date, had a term of November 1999 through October 2003. Thus, his initial valuation included a royalty for only 3 years (i.e., from November 2003 through the end of 2006). During trial, in response to Hatch's criticism of his findings petitioner's trademark expert revised his calculations to include a royalty that covered the entire 7-year useful life that he projected. He ultimately concluded that the revised upper-end value for the trademark intangibles was $9.6 million.

Petitioner's trademark expert was not convincing and when he was questioned regarding the calculation of his lower range of values, his response was incoherent. Respondent failed to estimate a value for these intangibles, and the paucity of credible evidence relating to this issue is disconcerting. Nevertheless, we conclude that petitioner's trademark expert's upper-end value of $9.6 million is the best available approximation of, and thus, the arm's-length value of the trademark intangibles.

The buy-in payment must also be adjusted to take into account the value of the sales agreements transferred from VERITAS US to VERITAS Ireland. We do not, however, have sufficient evidence to determine the value of those agreements. Thus, this matter must be addressed in the parties' Rule 155 computations.

D. The Appropriate Discount Rate

Petitioner's financial markets expert Burton Malkiel (Malkiel) applied the CAPM and concluded that 20.47 percent was a reasonable estimate of VERITAS US' WACC. There are two differences between Hatch's and Malkiel's applications of CAPM: The estimate of the beta and the equity risk premium. Malkiel, unlike Hatch, used reliable data to calculate both variables.

Malkiel had two reasons for employing a 1.935 company-specific beta, rather than a 1.42 industry beta, to calculate VERITAS US' WACC. First, the industry beta for VERITAS US' Standard Industrial Classification (SIC) code⁴⁷ was skewed because of the presence and size of Microsoft.⁴⁸ Microsoft dominated the personal computer operating system software market and had a stronger and more established business than VERITAS US. Thus, the risk level for VERITAS US' industry SIC group did not present a portfolio of comparable risk. Second, while betas for individual companies tend to be unstable, VERITAS US' betas were quite stable. Moreover, Malkiel used the historic average risk premium from 1926 to 1999 as reported by Ibbotson Associates (i.e., the best available data) to estimate the equity risk premium for 1999. See Brealey & Myers, supra at 157, 179. Accordingly, the appropriate discount rate is 20.47 percent.

Contentions we have not addressed are irrelevant, moot, or meritless. To reflect the foregoing,

Decision will be entered under Rule 155.

² Unless otherwise indicated, all section references are to the Internal Revenue Code of 1986, as amended and in effect for the years in issue, and all Rule references are to the Tax Court Rules of Practice and Procedure.

³ See infra, Background, sec. I, Storage Management Software Products, for a discussion of NetBackup.

⁴ See infra, Background, sec. I, Storage Management Software Products, for a discussion of Backup Exec.

⁵ "Backup" is simply making a copy and moving it to a safe location. "Restore" takes the copy from the safe location and makes it available to the end-user.

⁶ Source code is the human readable statement used to write computer programs and is commonly organized into files, which are composed of individual lines of code. Complex software, such as storage management software, often requires thousands of source code files and hundreds of thousands or millions of lines of code.

⁷ APIs, which provide software vendors with access to the operating system's features, allow an application written for one type of operating system to run on a different type of operating system.

⁸ From 1999 to 2006 Sun, VERITAS US' largest and most significant OEM partner, paid VERITAS US $657.4 million in royalties. During this period VERITAS US also received $292.9 million from HP, $181.6 million from Dell, $23.9 million from Hitachi, $7.9 million from NEC, and $780,000 from Compaq. In addition, from 2000 to 2008 VERITAS US received $23.9 million in royalties from Ericsson.

⁹ The operating system, the applications it supports, and the software it uses to manage devices attached to the computer are referred to as a "stack".

¹⁰ With respect to the RDA, "VERITAS Ireland" refers to VSHL and VSIL, the two parties who entered into the RDA with VERITAS US.

¹¹ With respect to the TLA, "VERITAS Ireland" refers to VSHL and VSIL, the two parties who entered into the TLA with VERITAS US.

¹² As previously stated, the CSA consisted of these agreements.

¹³ The RDA defined VERITAS Ireland's Territory as "Europe, Middle East, Africa, Asia, and the Asia Pacific."

¹⁴ The RDA defined Research Program as "all software research and development activity and process development activity".

¹⁵ Amendment No. 1 was effective as of Nov. 3, 1999; Amendment No. 2 was effective as of Aug. 1, 2001; and Amendment No. 3 was effective as of July 1, 2002.

¹⁶ In 2001 VERITAS Ireland outsourced the manufacture of Volume Manager, File System, Cluster Server, and Foundation Suite.

¹⁷ The parties stipulated the stock-based compensation costs at issue and agreed that the determination of whether such costs must be included in the cost-sharing pool would be "controlled by the final decision, within the meaning of section 7481 of the Internal Revenue Code (the 'Code'), in Xilinx, Inc. and Subsidiaries v. Commissioner, 125 T.C. 37 (2005), appeals docketed, No. 06-74246 and 06-74269 (9th Cir., Aug. 30 and Sept. 29, 2006)." In addition, respondent conceded adjustments relating to technical support services.

¹⁸ VERITAS Ireland's shares of reasonably anticipated benefits, pursuant to section 1.482-7(f)(3), Income Tax Regs., were 23.04 percent relating to 2000 and 28.47 percent relating to 2001.

¹⁹ See sec. 1.482-4(c)(1), Income Tax Regs. See also infra, Discussion, sec. III, Petitioner's CUT Analysis, With Some Adjustments, Is the Best Method, for a fuller discussion of the CUT method.

²⁰ See sec. 1.482-6, Income Tax Regs., for a discussion of the profit split method.

²¹ These regulations, promulgated in December 2008, are effective for transactions entered into on or after Jan. 5, 2009. See infra, Discussion, sec. III(A), Comparability of OEM Agreements, for a more in-depth discussion.

²² During the May 2 hearing, respondent referred to "access to R&D team" and "access to marketing team" as "assembled workforce".

²³ The term "pre-existing intangibles" is not used in secs. 1.482-1T through 1.482-9T, Temporary Income Tax Regs., 74 Fed. Reg. 349 (Jan. 5, 2009).

²⁴ During trial and on brief, respondent referred to "access to R&D team" and "access to marketing team" as "workforce in place".

²⁵ The Administration stated that the proposed change in law was simply a "clarification" yet estimated that this change, when combined with other "clarifications", would raise nearly $3 billion dollars over 10 years. See Department of the Treasury, General Explanations of the Administration's Fiscal Year 2010 Revenue Proposals, Table 1 (May 2009).

²⁶ A beta of 1 indicates that the security's price has tended to move in step with the market (i.e., a 1-percent increase in the market has led to a 1-percent increase for the security), a beta of less than 1 implies that the security is less volatile than the market, and a beta greater than 1 indicates that the security is more volatile than the market. See infra, Discussion, sec. II(B)(4), Respondent Employed the Wrong Useful Life, Discount Rate, and Growth Rate, for formula using beta.

²⁷ See infra, Discussion, sec. II(B)(4), Respondent Employed the Wrong Useful Life, Discount Rate, and Growth Rate.

²⁸ See infra, Discussion, sec. II(B)(3), Respondent's Allocation Took Into Account Subsequently Developed Intangibles.

²⁹ Even if respondent, pursuant to section 1.482-1(f)(2)(i)(A), Income Tax Regs., were authorized to aggregate the transactions, the "akin" to a sale theory may violate section 1.482-1(f)(2)(ii)(A), Income Tax Regs. This regulation provides that "The district director will evaluate the results of a transaction as actually structured by the taxpayer unless its structure lacks economic substance." The transaction at issue, which certainly had economic substance, was structured as a license of preexisting intangibles, not a sale of a business.

³⁰ The sales agreements were transferred, but the parties made no attempt to value them. See infra, Discussion, sec. IV(C), Value of Trademark Intangibles and Sales Agreements.

³¹ Even if such evidence existed, these items would not be taken into account in calculating the requisite buy-in payment because they do not have "substantial value independent of the services of any individual" and thus do not meet the requirements of sec. 936(h)(3)(B) or sec. 1.482-4(b), Income Tax Regs. "Access to research and development team" and "access to marketing team" are not set forth in sec. 936(h)(3)(B) or sec. 1.482-4(b), Income Tax Regs. Therefore, to be considered intangible property for sec. 482 purposes, each item must meet the definition of a "similar item" and have "substantial value independent of the services of any individual". Sec. 936(h)(3)(B); sec. 1.482-4(b), Income Tax Regs. The value, if any, of access to VERITAS US' R&D and marketing teams is based primarily on the services of individuals (i.e., the work, knowledge, and skills of team members). Nevertheless, respondent in support of his contention cites Newark Morning Ledger Co. v. United States, 507 U.S. 546 (1993), and Ithaca Indus., Inc. v. Commissioner, 97 T.C. 253 (1991), affd. 17 F.3d 684 (4th Cir. 1994). These cases, however, do not suggest that access to an R&D or marketing team has substantial value independent of the services of an individual, do not define intangibles for sec. 482 purposes, and do not even reference sec. 482. We note that in December 2008, the Secretary promulgated temporary regulations (i.e., secs. 1.482-1T through 1.482-9T, Temporary Income Tax Regs., supra) which reference "assembled workforce". In addition, the Administration, in 2009, proposed to change the law to include "workforce in place" in the sec. 482 definition of intangible.

³² The discount rate (i.e., the cost of capital) is an adjustment to a determined value to take into account the rate of inflation, the time value of money, and any attendant risk.

³³ The WACC provides the expected rate of return for a company on the basis of the average portion of debt and equity in the company's capital structure, the current required return on equity (i.e., cost of equity), and the company's cost of debt. The equation for calculating the WACC is: WACC = E(r_e) + D(r_d)(1-T), where D represents the company's average portion of debt, E represents the company's average portion of equity, r_e represents the company's cost of equity, r_d represents the company's cost of debt, and T represents the company's marginal tax rate.

³⁴ Estimating the WACC for a company requires estimating the company's cost of equity (r_e). The CAPM, which seeks to determine the rate of return for a specific security, is commonly used to estimate a company's cost of equity.

³⁵ The CAPM model uses the following equation to determine the cost of equity: r_e = r_f + $*(r_m - r_f), where $ (beta) is a measure of the volatility, or systematic risk, of a security or portfolio in comparison to the market as a whole, r_f is the yield to maturity for a U.S. Treasury bond (often referred to as the risk-free rate), and r_m is the expected long-term yield for the U.S. stock market as a whole.

³⁶See supra note 26 for a more detailed discussion of beta.

³⁷ A lower equity risk premium results in a lower cost of equity, lower WACC (i.e., discount rate), and larger buy-in payment.

³⁸ Using a value for the intangible license expressed in terms of revenues, Baumol employed the following formula to determine the requisite buy-in payment: V=E_t=0^AC(1-B)^t(1-D)^tP_t, where A is the expected economic life of the intangibles, B is the rate of obsolescence, C is the value of the parameter representing the ratio between the list price and the appropriate intangible license fee (i.e., the arm's-length license fee for the intangibles as a percent of revenue), D is the discount rate, and P is the revenue for products sold during the product's useful life, with the payment for the initial year being C(1B)(1-D)P₁.

³⁹ Make-sell rights are the licensed rights to manufacture and sell existing intangible property.

⁴⁰ Secs. 1.482-1T through 1.482-9T, Temporary Income Tax Regs., supra, provide "further guidance and clarification regarding methods under section 482 to determine taxable income in connection with a cost sharing arrangement in order to address issues that have arisen in administering the current regulations." 74 Fed. Reg. 340 (Jan. 5, 2009). These regulations include the income method and the price acquisition method and provide guidance on applying these methods for purposes of evaluating the arm's-length amount for platform contribution transactions (i.e., formerly referred to as transactions involving preexisting intangibles). The temporary regulations list the following specified methods: The CUT method, the income method, the price acquisition method, the market capitalization method, and the residual profit split method. The CUT method and the profit split method are the only two "specified methods" in the temporary regulations that were listed as "specified methods" in the regulations applicable to VERITAS US' transaction.

⁴¹ We recognize that one of the differences between the controlled and uncontrolled transactions is that, unlike the OEMs, VERITAS Ireland was not entitled to product updates, revisions, or modifications. We have concluded, however, that it was appropriate to make an adjustment to account for this difference. See infra, Discussion, sec. IV(B), The Appropriate Useful Life and Royalty Degradation Rate.

⁴² See supra, Discussion, sec. II(B)(4), Respondent Employed the Wrong Useful Life, Discount Rate, and Growth Rate, and supra, Background, sec. IV, Product Lifecycles and Useful Lives.

⁴³ While a static product may lose considerable value, the value of the product need not be zero in the final year of the product's useful life.

⁴⁴ In calculating ramp-down rates, Baumol relied on petitioner's source code expert. The source code expert's simplistic and mechanical analysis was not convincing.

⁴⁵ See supra, Background, sec. IV, Product Lifecycles and Useful Lives.

⁴⁶ The rates are rounded to the nearest percentage.

⁴⁷ The SIC is a U.S. Government statistical classification system that uses a four-digit numerical code to group businesses according to industry and subindustry groups. Businesses are grouped according to their primary economic activity (e.g., agriculture, fishing, manufacturing, transportation, communications, wholesale trade, etc.).

⁴⁸ Betas relating to industry portfolios typically reflect the capital structure of the companies included in the particular industry. Companies that have large market values (i.e., determined by multiplying the number of the company's shares of stock outstanding by the price of the shares) carry greater weight in the SIC group's portfolio.