Menu
Tax Notes logo

Reviewing Transfer Pricing's Standard of Review After Coca-Cola

Posted on May 16, 2022

Every U.S. Tax Court opinion in an IRC section 482 case dutifully recites some variation of the necessary incantation: The IRS has broad discretion in applying section 482, and its determinations must be upheld unless the taxpayer can establish that the IRS abused its discretion by making adjustments that are “arbitrary, capricious, or unreasonable.” However, a substantive discussion of how that vague formulation should be applied in transfer pricing cases is much harder to find, and the Tax Court’s decades of manifest refusal to defer to the IRS’s section 482 interpretations compounds the mystery and confusion.

Before the Tax Court issued its opinion in Coca-Cola Co. v. Commissioner, 155 T.C. 145 (2020), the most substantive guidance on the matter arguably came from a 1935 case that predates section 482 and the Tax Court itself. In Asiatic Petroleum Co. v. Commissioner, 31 B.T.A. 1152, 1157 (1935), aff’d, 79 F.2d 234 (2d Cir. 1935), the U.S. Board of Tax Appeals held that the statutory text of section 45 of the Revenue Act of 1928 (which, regarding the grant of discretion, was substantially similar to what appears in the first sentence of section 482) contained a clear delegation of authority to the IRS. Section 45, like its modern successor, expressly granted the commissioner authority to distribute, apportion, or allocate income, deductions, or other items among commonly owned or controlled taxpayers “if he determines that such distribution, apportionment, or allocation is necessary in order to prevent evasion of taxes or clearly to reflect the income.”

Referring to the phrase “if he determines,” which it italicized, the board in Asiatic Petroleum says:

The italicized words made the allocation a matter of discretion with the commissioner. In matters [entrusted] to the discretion of administrative officers there is a heavy burden on him who claims error in its exercise — a burden undoubtedly heavier than that of overthrowing a purely factual determination upon which the ultimate determination must depend. [Internal quotations omitted.]

Asiatic Petroleum never spells out the precise weight of that heavy burden or the extent of its reach. Regardless, its acknowledgment of a burden on taxpayers that is undoubtedly heavier than that of disproving allegations of fact clearly implies that the IRS enjoys broad discretion and that its determinations are entitled to some significant (if largely undefined) form of deference by reviewing courts.

For decades, that deferential standard of review was dutifully observed — in name, if not in substance — with little meaningful elaboration. Tax Court opinions consistently remind us, often before summarily rejecting every IRS argument, that the agency’s section 482 allocations must be upheld unless they are arbitrary, capricious, or unreasonable. The odd pattern that emerged was for the Tax Court to overtly accept its obligation to review the IRS’s section 482 allocation with deference without observing anything that resembles deference in its rulings.

The Tax Court’s long-standing lip service to deference in transfer pricing cases may — at least partly — be a byproduct of its failure to establish what “abuse of discretion” means in section 482 litigation. Does it include deference to the factual findings that underpin the IRS’s position, or does it pertain to the IRS’s interpretation of the section 482 regulations as they apply in that case? Or is it something else?

The Coca-Cola Standard of Review

The first step in decades toward answering those questions came in 2020, when the Tax Court released its decision in Coca-Cola. That opinion stands out among the court’s section 482 decisions for its unusually lengthy discussion of the standard of judicial review, which is strikingly detailed given the issue’s limited practical importance in a case the IRS seemingly would have won under any standard.

Coca-Cola draws heavily on Asiatic Petroleum, noting that the scope of the IRS’s discretion (and the deferential standard of judicial review it requires) set out in the 1935 opinion “continues to apply today.” It cites the text of section 482 as the basis for the IRS’s discretion, borrowing language directly from Asiatic Petroleum:

Where a statute commits to an executive department of the government a duty requiring the exercise of administrative discretion, the decision of the executive department, as to such questions, is final and conclusive, unless it is clearly proven arbitrary or capricious, or fraudulent, or involving a mistake of law.

The Coca-Cola court then attempted to explain what the standard of review endorsed in Asiatic Petroleum means for modern section 482 cases, in part by describing how the methodology used by the IRS — that is, the transfer pricing method and the specific way it’s applied — factors into the abuse of discretion assessment. Although judicial review generally focuses on the reasonableness of the section 482 allocation and not the details of the method used to compute it, the review is necessarily method-oriented in cases that involve transfers of “unique and extremely valuable intangible property,” the court wrote. “To show that the commissioner has reached an unreasonable result in such a case, the taxpayer typically will need to establish that the commissioner employed an unreasonable methodology to reach his result,” it said.

According to the court, unreasonable methods typically come in two forms, the first consisting of those that entail “significant legal error.” Its examples of those kinds of methods include those at issue in Commissioner v. First Security Bank of Utah NA, 405 U.S. 394 (1972), which addressed a section 482 allocation of legally blocked income, and Amazon.com v. Commissioner, 148 T.C. 108 (2017), aff’d, 934 F.3d 976 (9th Cir. 2019), which found that the IRS’s valuation method improperly included the value of residual business assets that were non-compensable under pre-2018 law.

The second category consists of methods that were unreasonably applied and includes those that use incorrect data, make flawed assumptions, or contain internal inconsistencies. Examples cited in the opinion include the financial projections and discount rate used in the IRS’s discounted cash flow valuation in Veritas Software Corp. v. Commissioner, 133 T.C. 297 (2009), and the flawed variation of the cost-plus method applied by the IRS in Seagate Technology Inc. v. Commissioner, 102 T.C. 149 (1994).

By performing a detailed assessment of the IRS’s selection of the comparable profits method as the best method, the foreign supply points as the tested parties, independent bottling companies as comparables, and the return on assets (subject to adjustments) as the profit-level indicator, the Tax Court made clear that the standard of review doesn’t absolve the IRS of its obligation to make methodological determinations in accordance with its own regulations. Unsupportable interpretations of key regulatory concepts, such as the best-method rule and the standard of comparability, or failure to consider factors that the regulations identify as relevant, would presumably constitute what the court called “significant legal error.” And material miscalculations, flawed comparability adjustments, or inconsistent profitability comparisons could indicate that the IRS’s method was unreasonably applied.

However, the court’s detailed review of the IRS’s methodological determinations against the relevant regulatory standards doesn’t imply that it evaluated the parties’ positions on an equal footing. The signs of deference are subtle — likely because the court had no reason to explore the finer points of deferential review when (as the opinion suggests) it agreed with the IRS’s determinations outright — but they can be found on close inspection.

The Tax Court fully reviewed the reasonableness of the IRS’s method, concluded that the agency didn’t abuse its discretion, and stated that the primary section 482 adjustments would be sustained — all before even addressing the taxpayer’s methodological arguments. It accepted the IRS’s argument on every major issue, often based on fairly general justifications that — at least on the face of the opinion — weren’t subject to particularly intense scrutiny. And after acknowledging one theoretical weakness in the IRS’s CPM analysis, the court made clear that the taxpayer bore the burden of establishing the practical significance of that weakness.

Coca-Cola “emphasizes that the two sets of companies occupied different points in the company’s supply chain and did business at different ‘levels of the market’: Supply points sold concentrate to bottlers, and bottlers sold finished beverages to distributors and retailers,” the court said. “But [it] has failed to show how these distinctions affect the comparability of the functions the two sets of companies discharged or the operating profit they could earn.”

A New Standard?

Coca-Cola’s clarifications of the standard of review leave many questions unanswered, including the significance and materiality thresholds for legal error or flawed assumptions and the effect on the standard of review (if any) of the second and third sentences of section 482. Even so, they represent a clear improvement over the vague and hollow recitations that have typically appeared in the Tax Court’s section 482 opinions. By setting out the signature characteristics of transfer pricing methods that taint the adjustments they generate, Coca-Cola gives the abuse of discretion assessment some structure and focus that’s conspicuously lacking in earlier opinions.

If the Tax Court were to consistently adopt a more focused standard of review, that could have consequences for other high-profile transfer pricing cases. Medtronic Inc. v. Commissioner, T.C. Memo. 2016-112, vacated by 900 F.3d 610 (8th Cir. 2018), which involves competing claims regarding the best transfer pricing method and awaits a second opinion from the Tax Court, is a prime example. Evaluating the reliability of different transfer pricing methods under the best-method rule and the comparability of purported transactional comparables would, in the words of Asiatic Petroleum, seemingly qualify as matters “entrusted to the discretion of administrative officers” under section 482. The burden Medtronic must carry to successfully challenge the IRS’s selection of the CPM should therefore be “undoubtedly heavier than that of overthrowing a purely factual determination upon which the ultimate determination must depend.”

Coca-Cola makes it possible to identify with some specificity what Medtronic’s burden should entail: Medtronic would have to identify a legally erroneous premise, an internal inconsistency, or the use of incorrect data or false assumptions. That wouldn’t give the IRS carte blanche to claim that the regulations’ best-method rule or comparability standards mean whatever it says they do or to impose its chosen method without regard to the facts or the regulatory standards for applying it. It would, however, prevent the Tax Court from following its traditional practice of weighing the parties’ arguments equally and exercising total discretion to either accept one of the parties’ methods or devise a method of its own. Qualified deference to the IRS’s position, even if it falls somewhat short of the level of deference described in Coca-Cola, could conceivably tip the scales in cases like Medtronic, Amgen Inc. v. Commissioner, No. 16017-21 (T.C. 2021), and other cases that involve similar issues.

Applying the Coca-Cola standard of review could also be significant in Facebook Inc. v. Commissioner, No. 21959-16, which involves a dispute over the best method for pricing platform contributions to a cost-sharing arrangement. The IRS’s favored method in the case is the income method, which provides an aggregated valuation when a party makes multiple platform contributions. Facebook argues that its different platform contributions should be valued on a disaggregated basis using the comparable uncontrolled transaction method.

Under Coca-Cola, it seems unlikely that Facebook can establish that the IRS’s method was unreasonable on the basis of significant legal error. The cost-sharing regulations heavily favor the income method when (as was the case for Facebook’s cost-sharing arrangement) only one party makes platform contributions, so it would be difficult to identify some error in regulatory interpretation behind the IRS’s selection of method. To establish significant legal error, Facebook would likely have to persuade the court that restrictions associated with the term “pre-existing intangible property” somehow survived the 2009 temporary cost-sharing regs’ purge of that term, allegedly because of some statutory constraint hidden in section 482.

Assuming that argument fails, which the Ninth Circuit’s Amazon opinion (934 F.3d 976 (9th Cir. 2019), aff’g 148 T.C. 8 (2017)) and a perusal of the statutory and regulatory texts suggest it should, Facebook would have to establish that the IRS applied the income method unreasonably. As Veritas illustrates, financial projections, discount rates, and other assumption-based input variables necessary for applying methods like the income method involve a large degree of subjective judgment, so it’s possible that Facebook could successfully establish that the IRS applied the income method unreasonably. However, like the level-of-market comparability issue in Coca-Cola, Facebook would likely bear the burden of showing that the IRS’s data and assumptions were flawed, and that those flaws undermined the method’s results. Bearing that burden wouldn’t condemn Facebook to certain defeat, but it would narrow its path to victory.

If it were consistently applied, the standard of review described in Coca-Cola could extend well beyond today’s major transfer pricing cases. High-stakes transfer pricing disputes between the IRS and multinational taxpayers are bound to continue long after the Tax Court releases its second Medtronic opinion and decides the Amgen and Facebook cases, and many of those disputes will feature competing claims regarding the best method, the reliability of potential comparables, and the appropriateness of transactional aggregation. If the IRS can steer clear of the specific pitfalls associated with unreasonable transfer pricing methods — significant legal error, flawed assumptions, incorrect data, and internal inconsistencies — it should enjoy the upper hand under Coca-Cola.

Defending Deference

Although Coca-Cola adds some long-overdue guidance on what constitutes an abuse of discretion in a section 482 adjustment, it doesn’t resolve the broader debate on the merits of adopting the deferential standard of judicial review recognized in Asiatic Petroleum. There are legitimate reasons to question the wisdom of deferring to the IRS’s position in transfer pricing litigation, especially given the deference it already enjoys for questions of statutory and regulatory interpretation. Deferential review of transfer pricing adjustments is far from universal across jurisdictions, and many other national tax authorities seem perfectly able to administer transfer pricing regimes without it. Does the IRS, which makes extensive use of its regulatory authority to interpret and enforce section 482, really need yet another level of deference (on top of Chevron and what remains of Auer deference) once it enters the courtroom?

Some commentators don’t think so. Acknowledging that judicial deference to the IRS’s section 482 allocations is largely illusory, Matthew Frank of Steptoe & Johnson LLP has argued that there’s no great injustice in the current state of affairs. (Prior analysis: Tax Notes Int’l, Aug. 23, 2021, p. 1015.) Without denying the IRS’s broad discretion to enforce section 482, Frank argues that the agency has already exercised much of that discretion in the form of regulatory action. The statute “should not be applied by the government with the scales tipped more heavily in its favor than the regulations already do,” he said.

Comparing the regs to speed limits, Frank argues that more prescriptive rulemaking necessarily — and properly — narrows the scope of IRS discretion in its enforcement of section 482. “The shrinking scope for deference is not disturbing or surprising,” he said. “The IRS and Treasury have exercised much of their discretion in framing the terms of the transfer pricing debate in regulations that control the IRS and guide taxpayers.” It is “appropriate that the regulations should cabin further discretion and reduce the occasion for deference,” he added. “The litigants are appropriately left to compete on relatively level ground in the arena that Treasury and the IRS designed.”

Frank raises legitimate questions about the need for a deferential standard of judicial review of individual section 482 allocations after the IRS has already made extensive use of its discretion to enforce the statute through regulations. However, there are also valid reasons to favor a deferential review, notably including the statutory text. Section 482 provides that “the Secretary may distribute, apportion, or allocate gross income, deductions, credits, or allowances” among commonly owned or controlled organizations, trades, or businesses “if he determines that such distribution, apportionment, or allocation is necessary in order to prevent evasion of taxes or clearly to reflect the income of any of such organizations, trades, or businesses.”

The discretion that Treasury and the IRS have to interpret the statutes they administer and their authority to enforce their interpretations through regulatory action aren’t unique to section 482: They apply to every code section in equal measure. But as recognized in Asiatic Petroleum and confirmed in Coca-Cola, section 482 isn’t like every other code section: The statutory wording (“the Secretary may . . . if he determines”) reflects a grant of discretionary authority distinct from other tax statutes. Narrowing the specific grant of discretion Congress wrote into the statute whenever Treasury and the IRS exercise their generic discretion to issue regulations interpreting section 482 is arguably at odds with the statutory text and legislative intent.

It’s also debatable whether the section 482 regs really establish the kind of prescriptive guidance that a speed limit sign with a two-digit number provides. The regulatory concept of a reliable measure of an arm’s-length result may be more specific than the completely open-ended statutory concept of a clear reflection of income, but the regulations heavily rely on flexible concepts and principles — comparability, reliability, the best-method rule, and the arm’s-length standard in general — that more closely resemble a general ban on reckless driving than a speed limit. It’s unclear that replacing one broad statutory ambiguity with many narrower regulatory ambiguities necessarily reduces the scope of or need for agency discretion.

More broadly, deferential review of section 482 adjustments can arguably be justified by the same considerations that support deference to executive agencies in other contexts. A common premise underlying deference doctrines — whether they pertain to statutory interpretation under Chevron or regulatory interpretation under Auer and Kisor — is that executive agencies are generally better able than courts to resolve ambiguities in the complex statutory or regulatory schemes they administer. That’s because agency personnel typically have more specialized technical expertise in the relevant subject matter and, as servants of the executive branch, are subject to a greater degree of political accountability for their actions. According to Chevron, Kisor, and other U.S. Supreme Court precedent, those factors support the general inference that Congress would have favored some level of deference in judicial review of agency determinations.

Many of the same considerations apply to section 482 determinations. That section’s regs are unique in their heavy reliance on flexible economic and financial concepts and general avoidance of the kind of prescriptive rules that characterize other tax regs. Deciding whether one intangible asset has “similar profit potential” to another, or whether differences in risk exposure justify a particular difference between the discount rate applied to the cost-sharing alternative and that applied to the licensing alternative, requires specialized economic expertise that other areas of tax law don’t.

The unique status of transfer pricing as an autonomous domain within tax law is evident from the people and companies that provide transfer pricing services. Unlike other areas of international tax, transfer pricing practitioners often have advanced degrees in economics or finance instead of law. Many economic consulting firms that otherwise focus on nontax valuation issues offer transfer pricing services, and some firms — including at least one member of the Big Four — place their transfer pricing practice within their valuation group instead of their international tax group.

Consequently, there’s little reason to assume that a Tax Court judge’s expertise in other areas of tax law will better translate to a mastery of transfer pricing than a district court judge’s command of constitutional law would enable her to — borrowing an example from Kisor — scrutinize the FDA’s definition of an “active moiety.” The IRS’s transfer pricing experts enjoy a considerable advantage over judges, including Tax Court judges, in their understanding of the section 482 regulatory scheme, their grasp of economic concepts relevant to pricing and valuation, and the depth of their knowledge about how controlled parties transact. That, along with the broad grant of authority evident in the statute itself, suggests that legislative intent favors a deferential standard of review.

If there’s a sound case for deferential judicial review of the IRS’s section 482 position on the basis of relative expertise, then there’s an even stronger case for limiting Tax Court judges’ ability to cast aside the parties’ positions and conjure up their own ideas of an appropriate transfer pricing outcome. If the court finds both parties’ methods unreasonable, it has total discretion to devise a method of its own, free from any obligation to preserve the features of the IRS’s method that weren’t unreasonable. For example, if the IRS’s selection of transfer pricing method is reasonable but the data it uses to apply the method is wrong, the Tax Court has no responsibility to retain the reasonable method and simply correct the data. Tax Court judges generally have no particular expertise in economic theory or valuation principles, and vesting them with unchecked discretion to dictate the best transfer pricing method is questionable policy.

Finally, practical considerations support a deferential standard of review. If a jurisdiction’s tax administration routinely takes unprincipled transfer pricing enforcement positions, and the judiciary shows little interest in stopping it, a deferential standard of review may very well be harmful. Shifting the burden to force tax authorities to affirmatively establish mispricing by the taxpayer may even be warranted in those circumstances. However, the United States has a comparatively evenhanded tax administration, and its Tax Court has historically shown a troubling tendency to reflexively side with taxpayers in transfer pricing cases. If any tax authority needs the benefit of a deferential standard of review to effectively enforce a transfer pricing regime, it’s the IRS.

Copy RID