Pending Cert Petition in Altera: Tax Law in an Administrative Law Wrapper

In 2015, Altera won a unanimous decision in the Tax Court, which invalidated a 2003 regulation as arbitrary and capricious under State Farm. Then the government won in the Ninth Circuit before the original three-judge panel in 2018 (withdrawn because of the death of Judge Reinhardt), before a revised three-judge panel in 2019, and when the Ninth Circuit denied the taxpayer’s request for a rehearing en banc in 2019. We contributed amicus briefs [here (with coauthors Leandra Lederman and Clint Wallace), here and here] on behalf of the government before the Ninth Circuit, and have blogged previously about the case here and here. In February, the taxpayer submitted a petition for certiorari to the Supreme Court.

The tax issue in Altera involves a final Treasury regulation promulgated in 2003. The reg applies to qualified cost-sharing agreements, or QCSAs, made between U.S. firms and their offshore subsidiaries. A QCSA requires an offshore subsidiary to pay its share of the costs of developing IP. If QCSA requirements are met, the offshore subsidiary owns non-U.S. rights to intangible property developed by its U.S. parent company for tax purposes. Then the firm can shelter resulting offshore profit from U.S. tax. As relevant here, the 2003 regulation at issue in Altera conditions the favorable tax treatment available for QCSAs on the inclusion of stock-based compensation costs in the pool of shared costs.  

Technology and other multinational firms that use stock option compensation (and use strategies to shift profit from intellectual property across borders) have had an understandable and longstanding interest in this issue. An appendix to Altera’s cert petition lists 82 companies that noted the Altera issue in their public financial statements. One entry alone – that of Alphabet, Inc. – reports $4.4 billion at stake.

We think the regulation gets it right as a matter of tax policy. It properly prevents stock-based compensation deductions from reducing U.S. taxable income when these expenses support foreign profit. The regulation falls securely under the Commissioner’s statutory discretion (under I.R.C. Section 482) and responsibility to ensure clear reflection of income. It squares with modern financial accounting rules. And it aligns with OECD and other international efforts to combat base erosion and profit shifting to low-tax jurisdictions.

But the hook in the cert petition is not the tax issue. It is an administrative law issue. The taxpayer hopes to persuade four justices that Altera is an attractive opportunity to rein in an administrative agency’s power and further limit the case law that supports administrative agency discretion. Perhaps it appears particularly juicy because the administrative agency at issue is the Treasury, given the complicated history and relationship between Treasury regulations and administrative law. Indeed, the regulation in this case was promulgated well before the Supreme Court held, in its 2011 Mayo case, that Chevron deference (rather than National Muffler review) applies to tax regulations just as it applies to other federal regulations.

The taxpayer’s administrative procedure argument includes two main claims. The first is that Treasury did not provide a reasoned explanation for the regulation and that the regulation was therefore arbitrary and capricious under State Farm. The second is that the government engaged in post hoc rationalization to defend the regulation, in violation of Chenery I. (A third claim, derivative of the first two, asks whether, assuming a regulation is held procedurally defective, a court may nevertheless uphold it under Chevron.)

Five out of six filings submitted to the Supreme Court on behalf of the taxpayer – including the primary cert petition and four out of five amicus briefs – hang their respective hats on a single premise. This premise is that Treasury first suggested that comparability analysis was relevant under the stock-based compensation QCSA regulation, and then Treasury broke its word. The government’s brief takes this premise head-on and, we think, persuasively disproves it.

Altera’s petition claims that in 2002 and 2003, “the government never said it was … adopting a new approach to cost-sharing” (8) and that the rationale that the “commensurate with the income” language supported the new approach “appeared nowhere in the rulemaking record.” (10-11) Amicus briefs argue that the government advances “a new statutory interpretation” in litigation (Chamber of Commerce 16), describe the government’s allegedly “newfound litigation position that comparables are irrelevant” (Cisco 11), assert a “transparent post hoc rationalization” (National Association of Manufacturers 15) and claim that there would have been comments on “the applicability and scope of the arm’s length standard” in notice-and-comment if taxpayers had only been aware that the government meant to make comparability analysis irrelevant to the determination of an arm’s-length result for stock-based compensation costs in the QCSA context. (PricewaterhouseCoopers 16).

Interestingly, the fifth of five amicus contributions supporting Altera – a brief filed by a group of former foreign tax officials – paints a picture of continuity, rather than change, in arguments made by Treasury and the IRS. It acknowledges that both in 2002 and 2003 and also in litigation before the Ninth Circuit, the government “ignor[ed] … potentially comparable transactions” and simultaneously “claim[ed] that its approach comported with the arm’s length standard.” (9-10) 

The government argues as follows in its brief in opposition to Altera’s cert petition: The taxpayer’s arguments “conflate (i) the arm’s length standard … and (ii) the use of comparability analysis” and “misunderstan[d] the relationship between the two concepts.” (19) In its rulemaking, the government did not suggest that empirical analysis and comparability were relevant to the determination of an arm’s length result in this context. Rather, the internal method adopted by the regulation is “an alternative to comparability analysis as a means of achieving an arm’s length result,”(20) consistent with the statute, as “Section 482 does not require any analysis of identified comparable transactions between unrelated parties.” (21) Moreover, the rulemaking and litigation record shows a constant commitment to a method that is not based on evidence of comparables. The government’s rulemaking record, as well as its arguments in litigation, consistently references the “commensurate with income” statutory language added in 1986. (24) So the “commensurate with income” argument made in litigation was not new either.

The government’s narrative gets this right. As the government’s brief explains, the regulatory history – not to mention the plain language of the regulation describing an arm’s-length result in this context – makes clear that interested taxpayers and tax advisers knew that “the proposed regulation would make any evidence of comparable transactions irrelevant” in the context of QCSAs. (22) Taxpayers certainly understood the proposed regs’ departure from comparability analysis. They just didn’t agree with it. Indeed, the battle lines over comparability analysis in the context of stock-based compensation costs were already clearly drawn, well before Treasury issued its Notice of Proposed Rulemaking in 2002. As the Software Finance and Tax Executives Council explained during the 2002 notice and comment period: 

On audit, in Advance Pricing Agreement negotiations, in docketed Tax Court cases, in published field service advice, and in speeches by Service officials … the Service has taken the position that stock-based compensation … must be included in related parties’ cost sharing pools. … Taxpayers have steadfastly and vehemently disagreed[, … absent] any evidence that unrelated parties … share stock option “costs” in their own cost sharing pools. 

This disagreement between taxpayers and the government was a tax policy dispute over the role of comparables in transfer pricing between related parties. Taxpayers argued that the arm’s length principle required comparables, even in the specific case covered by the QCSA regulation. The government consistently took the opposite position, beginning well before 2002 and continuing through the present cert petition in Altera.

Taxpayers may still disagree with the government on the tax policy issue. But that ship has sailed. Indeed, there are other examples of transfer pricing methods that do not rely on comparable transactions. One is the 1994 promulgation of the residual profit split method, also contained in a final regulation issued under I.R.C. Section 482.  

The issue before the Supreme Court is an administrative law issue. A necessary premise of Altera’s argument is that Treasury started with, but then abandoned, a commitment to empirical comparables analysis for its rule covering stock-based compensation in QCSAs. And as the government explains, this premise does not hold up.