INSIGHT: A Second Bite at the APA: Altera’s Rehearing and the Potential Invalidity of Cost-Sharing Regulations—Part Three

III. REMAINING QUESTIONS IN THE ALTERA-XILINX LINE OF CASES

Some crucial questions appear to emerge from this quagmire: Specifically, whether (1) the allegedly comparable uncontrolled cost-sharing transactions proffered during notice and comment were, in fact, comparable; (2) if so, to what extent was Treasury required to consider evidence of such transactions and make findings as to their comparability in order to comply with the Administrative Procedure Act (APA); and (3) can a court ignore the inherent, internal conflict between the arm’s-length regulation as reflected in Treas. Reg. 1.482-1(b)(1) and the stock-award regulation?

A. Were the Transactions Comparable?

A watershed question in the Altera-Xilinx line of cases is whether the supposedly comparable transactions presented by taxpayers, all of which were devoid of the sharing of stock-based compensation were, in fact, comparable. For example, transactions involving high-value intangibles in the years at issue might be distinguishable from older transactions involving other types of intangibles. If there were no comparable transactions, the taxpayer would have difficulty arguing against the adoption of an alternative method to determine cost allocations.

In actuality, it is possible that no comparable uncontrolled transactions would exist in the case of Altera, or perhaps even in any case relating to qualified cost-sharing arrangements (QCSAs). As Congress has pointed out,

“A fundamental problem is the fact that the relationship between related parties is different from that of unrelated parties… . The problems are particularly acute in the case of transfers of high-profit potential intangibles… . Industry norms for transfers to unrelated parties of less profitable intangibles frequently are not realistic comparables in these cases. Transfers between related parties do not involve the same risks as transfers to unrelated parties.” (H.R. Rep. No. 99-426, at 424–25.)

Treasury provided almost no explanation, however, of how it concluded that the allegedly comparable transactions presented during the notice and comment period were not in fact comparable to the Altera situation or other cases involving “high-value” intangibles. As described in II.C., Treasury disregarded the empirical evidence presented by commenters by making an unsupported assertion based on a hypothetical of its own creation that, notwithstanding any indication to the contrary, it still “believed” stock-based compensation would be shared by unrelated parties.

B. If the Transactions Presented Were in Fact Comparable, to What Extent Was Treasury Required to Consider Evidence of Such Transactions and Make Findings as to Their Comparability Under the APA?

If there are in fact comparable transactions, it is unclear whether Treasury has the authority to abandon the search for comparable transactions altogether in favor of a commensurate-with-income standard. The majority of the U.S. Court of Appeals for the Ninth Circuit panel seemed to think this was permissible, while the dissent seemed to think that it was not. As indicated in II.E.4., the dissent would have, at a minimum, required Treasury to affirmatively “say” that it was adopting a different approach in order to dispense with an arm’s-length method approach that would abandon the search for comparable transactions. According to the dissent, the legislative history of tax code Section 482 acknowledged that any standard adopted would have to comply with the arm’s-length standard, transfer pricing’s bedrock principle, while the majority in the Ninth Circuit seemed to take the position that it was free to abandon the principle completely in favor of more “flexible” approaches that “moved away” from the “traditional” arm’s-length standard.

If the transactions presented by commenting taxpayers are determined to in fact be comparable, to what extent does the APA require Treasury to consider them? Four possibilities seem apparent: (i) Treasury could opt to simply ignore the evidence provided in the comments because such comments were irrelevant to its ability to promulgate the regulations under the “relevant factor” and “outcome determinative” standards; that is, having an independent basis for promulgating the regulations under the commensurate-with-income standard meant that Treasury was free to completely ignore comparables if it so chose; (ii) Treasury was required to at least consider the evidence presented in the comments but was free to disregard after a merely perfunctory consideration; (iii) Treasury was required to consider the evidence and would need to articulate a concrete finding that such evidence was, in fact, not pertinent; or (iv) Treasury would have to present its own evidence as rebuttal. At least ostensibly, the dissent suggests that in order for the government to pass its burden under State Farm, the bright red line in making this determination may fall somewhere between the third and fourth possible outcomes described above. The dissent also suggests that Treasury could have bypassed this entire matrix by simply stating at the time of promulgation that it was affirmatively departing from the arm’s-length standard in deciding to issue Treas. Reg. 1.482-7A(d)(2).

Perhaps an even more interesting situation would arise if the government were to concede as a matter of law that the third-party transactions presented by the taxpayer were in fact comparable. Notably, in Xilinx, the government did not appear to question whether such comparable transactions existed, stating: “[t]he Commissioner does not dispute the court’s factual finding that unrelated parties would not share employee stock options (ESOs) as a cost. Instead, the Commissioner maintains ESOs are a cost that must be shared under [the regulation], even if unrelated parties would not share them.” Similarly, in the Altera Tax Court decision, the court held that Treasury, by virtue of ignoring the transactions presented by taxpayers during the Notice and Comment process, had ignored “relevant evidence.” Referring to this evidence as “relevant” seems to implicitly acknowledge that the court had concluded as a matter of law that the allegedly comparable transactions presented by commenters were in fact comparable. The government would have had to justify use of a method that, by its own admission, applies only in the absence of comparables. The dissent, in fact, seemed to think that this Xilinx question was the question the Ninth Circuit should have addressed in the first place. The dissent stated that it would have held that Xilinx controlled.

C. The Lingering Question From Xilinx: the Conflicting Regulations Issue

It is interesting to contemplate how the conflicting regulation problem that Xilinx focused on plays into this analysis. Recall that this question dealt with whether the amended 2003 “all-costs” regulation actually conflicts with the rule in Treas. Reg. 1.482-1(b)(1), which states that the arm’s-length standard governs in “every case.” Throughout the Ninth Circuit opinion, there is evidence of an inherent acceptance of Treasury’s approach to the regulations, which amounts to a claim that the commensurate-with-income standard comports with the arm’s-length standard simply because Treasury “says so.”

One wonders whether the tautological gymnastics required in order to square the stock-award regulation with the seemingly intractable arm’s-length standard in Treas. Reg. 1.482-1(b)(1) is the result of Treasury’s unwillingness to retract the latter regulation in the wake of what it likely considered unsatisfactory results, preferring instead to tie a “Gordian knot” around the cost-sharing regulations in order to drive the outcomes it hoped would have otherwise occurred naturally. However, if the Ninth Circuit truly believed this to be the case, there would be no reason in its opinion to either justify or explain a departure from the arm’s-length standard—indeed, the two standards would be one and the same. Instead, the opinion seems to implicitly acknowledge that this simply cannot be the case: It repeatedly points out that the arm’s-length standard was not required and how the legislative trend toward alternative methods justified a departure from it. The decision does little to clarify what the Ninth Circuit’s conclusion is, other than that it clearly intends to grant enormous leeway to Treasury.

Conveniently, the court pointed out that the regulatory conflict issue was not technically before the court and remained unaddressed after Xilinx. Xilinx, rather, had held that the two regulations could not be reconciled with one another, but, of course, Xilinx was looking at older regulations. Subsequent amendments to the regulations were issued, at least in part, to address this problem. The amendments included an explicit reference to the Treas. Reg. 1.482-1(b) arm’s-length standard. Nonetheless, it is possible that this question might again come to the forefront. Interestingly, in the Altera Ninth Circuit dissent, Judge O’Malley, who believed that Xilinx controlled the question at issue, presumably would have viewed the underlying issue in Xilinx and Altera as identical. That is to say, Judge O’Malley may have implicitly stated her view that the current Treas. Reg. 1.482-7A still conflicts with Treas. Reg. 1.482-1(b)(1), notwithstanding the 2003 amendments, and is therefore invalid under the reasoning articulated in Xilinx.

Notwithstanding the lack of clarity as to some of these substantive points, if nothing else is clear through the confusing web of case law, statutes, and regulations, it is that courts appear to be clearly communicating that government agencies must follow the APA in all cases. No special rule, i.e., no “tax exceptionalism,” applies to absolve Treasury from the APA requirements simply because the rules at bar are tax regulations.

IV. LOOKING FORWARD

While the final outcome of the case is anyone’s guess, we explore some of the possibilities below. The withdrawal of Altera at the Ninth Circuit means that the U.S. Tax Court decision invalidating the regulations once again retains the force of law, at least temporarily. The IRS announced on Jan. 12, 2018, in an LB&I memorandum, that while Altera is on appeal, no new examinations of stock-based compensation in cost-sharing arrangements will be initiated.

However, even if the Ninth Circuit, on rehearing, issues an opinion that mirrors the withdrawn opinion, such opinion would apply only to taxpayers in the Ninth Circuit. It is apparent that the majority of the high-tech companies with significant tax liability at stake in this issue are located in the Ninth Circuit, but taxpayers in other circuits may continue to treat the Tax Court decision as controlling unless and until their circuit reaches a contrary view. Ultimately, a circuit split on the issue could lead to review by the Supreme Court.

There may be reason to think that a reversal is forthcoming in Altera. The Ninth Circuit explicitly indicated that before his passing, Judge Reinhardt had participated in the decision, meaning that use of discretionary authority to revisit the case, particularly on an immediate basis and prior to a petition for rehearing, may signify a concern with the substantive outcome rather than the desire simply to correct a mistake in the procedural process. Moreover, it has been speculated by some that Judge Reinhardt may not have had an opportunity to consider Judge O’Malley’s dissent prior to his concurrence. Notably, Judge Reinhardt would have been permitted to change his vote at any time prior to issuance of the decision if he had been alive when the dissent was available. As a matter of common judicial practice, neither the reasoning nor the bottom-line result of an opinion such as Altera would be unchangeable until the final opinion issues and, even then, a panel would be permitted to decide to amend its opinion or grant panel rehearing and issue a new opinion.

Further, public outcry against the Ninth Circuit’s decision has been substantial. For example, many are pointing to the unanimous 15-0 decision by judges who are expert in the subject matter of taxes in the Tax Court and questioning whether the Ninth Circuit is overstepping its bounds by disagreeing with what the Tax Court had so clearly concluded. In addition, a parallel can be drawn to Xilinx v. Commissioner, in which an initial 2-1 panel decision in favor of the IRS was subsequently withdrawn and reversed, suggesting that history may repeat itself. Moreover, Judge O’Malley’s dissent is well-reasoned and raises plausible grounds for reversal.

The Chevron analysis as applied by the D.C. Circuit in Good Fortune Shipping, which appears to be almost identical to the reasoned decisionmaking standard of State Farm, provides additional hope for a reversal. Rather than simply considering whether the result of the regulation was reasonable, the court looked at whether the decisionmaking process itself was sufficiently well-founded. Application of this decidedly less deferential standard of review will undoubtedly be encouraging to taxpayers who hope to fight the validity of Treasury regulations in the future. Good Fortune Shipping makes clear that, in the D.C. Circuit at least, Treasury must carefully and adequately explain its reasoning in a way that satisfies APA standards at the time of promulgation, irrespective of whether the end result of the regulation is reasonable or whether some other form of reasoning would have also justified the regulation in retrospect. Certainly, Good Fortune Shipping is strong evidence that there are no special exceptions to the APA in the case of Treasury regulations.

However, notwithstanding the trend to invalidate regulations that Good Fortune Shipping may signify, there are also a few reasons to think that a reversal in Altera may not be forthcoming after all. For one, Judge Graber’s generally moderate judicial views have not previously evinced an inclination to invalidate regulations lightly under the APA and the Chevron standards, perhaps suggesting that to do so could only be the result of overwhelming rationale for a reversal. Second, the import of the decision, and the unusual circumstances associated with Judge Reinhardt’s untimely death, may simply suggest a purely procedural desire on the part of the court to ensure a valid opinion rather than any substantive concern with the decision. Finally, the very amount of the taxes at stake here may suggest that courts may feel inclined to read existing authority to favor the notion that stock-based compensation expenses should be included in the calculation of the cost-sharing payments. Some would suggest that Congress could not have intended to create an exception that would have so extensively reduced the tax on technology companies, even if there is no clear statutory or regulatory pathway to that conclusion.

V. CONCLUSION

The ultimate outcome of Altera appears, at the moment, to be effectively a coin toss. On the one hand, a decision in favor of the government would be viewed by many as a surprising departure from the arm’s-length principle, which has been thought of as the bedrock of U.S. transfer pricing standards and which long-standing practice would seem to dictate should be adhered to in all but the narrowest of circumstances. Not insignificantly, such an outcome would have a chilling effect on future procedural challenges to Treasury regulations under the arbitrary and capricious standard of the APA.

On the other hand, a decision for the taxpayer and the invalidation of the stock-award regulation may well have the opposite effect, opening the floodgates to invalidation of other regulations and creating an incentive for taxpayers to more closely examine the protocol that was (or was not) followed by agencies in the course of their promulgation. It bears noting based on Good Fortune Shipping that, given the possible willingness of courts to apply the more stringent State Farm reasoned decisionmaking standard (whether explicitly so or under the guise of Chevron step two) to all tax regulation challenges, Treasury may have its work cut out for it. Certainly, given the multitude of new statutory provisions fostered by the 2017 tax act that have taxpayers clamoring for much-needed regulatory guidance, the timing of this development may make Treasury’s mission more daunting than ever.

Nonetheless, while an attempt to invalidate Treasury regulations under the APA undoubtedly presents a significant challenge to taxpayers under both the State Farm and Chevron standards, Treasury’s approach in Altera may well have crossed the line. Notwithstanding the import of the commensurate-with-income standard and its role under Section 482, it is difficult to conclude based on any of the relevant legislative or regulatory history that this approach could be applied to the exclusion of, or at least the consideration of, empirical evidence of comparable uncontrolled transactions. Some commentators have suggested that a final decision will be out by the end of the year.

The full article was published originally in Bloomberg Tax’s Tax Management International Journal and is available here.

Peter J. Connors is a tax partner in the New York office of Orrick, Herrington & Sutcliffe LLP. Barbara S. de Marigny is a tax partner and Michael R. Rodgers is a senior tax associate in the Houston office of Orrick, Herrington & Sutcliffe LLP. The authors also extend special thanks to Professor William H. Byrnes of the Texas A&M University School of Law for his helpful thoughts, comments, and suggestions.