Are The EU’s State Aid Tax Challenges Of Flimsy APAs Doomed?

The European Commission’s ability to ensure that advance pricing agreements granted by member states comply with state aid rules is already under threat, but a recent advocate general opinion proposal would restrict it even further.

In her June 8 opinion in Commission v. Luxembourg, C-457/21 P, concerning state aid allegedly conferred on the Amazon group by a 2003 APA, Advocate General Juliane Kokott urged the Court of Justice of the European Union to affirm the General Court’s judgment in Luxembourg v. Commission and Amazon EU Sàrl and Amazon.com Inc. v. Commission, joined cases T-816/17 and T-318/18 (GCEU).

However, considering the judgment in Fiat Chrysler Finance Europe v. Commission and Ireland v. Commission, joined cases C-885/19 P and C-898/19 P (CJEU 2022), Kokott recommended that the court’s judgment regarding potential state aid granted to Amazon be affirmed on very different grounds.

According to case law interpreting article 107 of the Treaty for the Functioning of the European Union, member state tax measures that derogate from the reference tax system by favoring one or more businesses over others are prima facie selective.

Selective measures generally constitute illegal state aid, unless they can be justified by the inherent logic of the reference system. In its judgment under appeal, the General Court held that the commission had the authority to assess the selectivity of the 2003 APA granted to Amazon by reference to the arm’s-length principle, as set out in the OECD transfer pricing guidelines.

However, the General Court found that the commission had failed to demonstrate the APA’s inconsistency with the 1995 version of the OECD guidelines, which was the most recent version available in 2003.

The 2022 Fiat opinion upended this reasoning by insisting that national law, and only national law, can form the basis of the reference system used to assess selectivity. Accordingly, the commission’s state aid review of APAs is now strictly limited to each member state’s tax laws.

Although the Fiat judgment clarified that the OECD guidelines cannot be included in the reference system unless domestic law specifies otherwise, it was less clear on how the relevant standards apply to individual APAs. Kokott’s Amazon opinion attempts to fill that gap by proposing a structured analytical framework to determine when a member state’s approval of a taxpayer-friendly APA crosses the line between the exercise of national tax sovereignty and illegal state aid.

Although the proposed approach would clarify important aspects of state aid law post-Fiat, it may do so by making it prohibitively difficult for the commission to demonstrate that flawed APAs constitute state aid.

The Post-Fiat Void

The CJEU’s Fiat judgment provided an authoritative and unambiguous answer to a question of critical importance in Amazon (and other state aid cases) involving suspect unilateral APAs: Can the OECD transfer pricing guidelines form part of the reference framework when assessing the selectivity of an APA-endorsed transfer pricing method when they have no official standing under national law?

The answer, according to Fiat, is a hard no. While recognizing that member states’ domestic laws and the practices of national tax administrations often reflect the OECD’s interpretation of the arm’s-length principle, the court held that the selectivity of APAs can only be assessed by reference to domestic law:

“It is only the national provisions that are relevant for the purposes of analysing whether particular transactions must be examined in the light of the arm’s length principle and, if so, whether or not transfer prices, which form the basis of a taxpayer’s taxable income and its allocation among the States concerned, deviate from an arm’s length outcome. Parameters and rules external to the national tax system at issue cannot therefore be taken into account in the examination of the existence of a selective tax advantage. . . . unless that national tax system makes explicit reference to them.” [Emphasis added.]

The Fiat judgment expressly recognizes only one circumstance in which the commission can establish that an APA constitutes illegal state aid: The member state that granted the APA must apply the arm’s-length principle, and the applicable provisions of national law must inherently discriminate in favor of multinational group members. As explained in paragraph 122 of the judgment:

“After having observed that a Member State has chosen to apply the arm’s length principle in order to establish the transfer prices of integrated companies, the Commission must. . . . be able establish that the parameters laid down by national law are manifestly inconsistent with the objective of non-discriminatory taxation of all resident companies, whether integrated or not, pursued by the national tax system, by systematically leading to an undervaluation of the transfer prices applicable to integrated companies or to certain of them.” [Emphasis added.]

This is a reformulation of the standard established by Commission v. Gibraltar and United Kingdom, joined cases C-106/09 P and C-107/09 P (CJEU 2011), showing that the reference tax system itself constitutes illegal state aid.

Citing the principle that national regulatory technique cannot dictate selectivity, the CJEU held that, in that case, Gibraltar’s system for taxing all companies was designed in a way that inherently favored offshore companies.

The Fiat judgment’s cursory attempt to extend the holding in Gibraltar to concern the selectivity of an APA leaves some important questions unanswered. It offers little meaningful guidance for determining when national law applies the arm’s-length principle in a way that is “manifestly inconsistent” with the nondiscriminatory taxation of controlled and uncontrolled taxpayers.

Gibraltar involved a tax regime that allowed all domestic companies with no local employees or business premises to escape tax on their income entirely. It’s hard to imagine how national law could simultaneously apply the arm’s-length principle, or anything remotely like it, and categorically exempt offshore companies from taxation.

More importantly, the Fiat judgment only addresses the selectivity standard applicable to underlying national law. The risk that member states will discreetly offer state aid through individual APAs is surely greater than the risk that they will publicly announce a general policy of selectivity in a decree or circular.

However, the judgment fails to establish any framework to determine when national law has been selectively applied in any particular APA.

Explicit Language

Kokott’s Amazon opinion attempts to answer these questions, but the answers raise questions of their own. The opinion generally reinforces the Fiat judgment’s insistence that the OECD transfer pricing guidelines, as “parameters and rules external to the national tax system,” do not form part of the reference system unless domestic law explicitly refers to them.

The implication is that even OECD-inspired domestic laws, including those that use the same terminology and reflect the same principles, do not incorporate the guidelines unless domestic law directly and explicitly incorporates them by reference.

It’s unclear why reproducing the content of OECD guidelines and expressly referring to the OECD guidelines should be given completely different legal effects.

A tax law that incorporates principles drawn directly from the OECD guidelines, and uses terminology lifted verbatim from the guidelines to do so, is substantially the same as an otherwise identical law that makes explicit reference to the guidelines. The justification for strict exclusion in one case and inclusion in the other is not evident from the Fiat judgment or from Kokott’s opinion.

Whatever its justification, the explicit reference requirement is not the kind of straightforward, bright-line test that the Fiat judgment and Kokott’s Amazon opinion present it to be. Although explicit references come in different forms that can have different legal effects, neither the Fiat judgment nor the Amazon opinion sheds any light on the significance of these differences.

One form of explicit reference to the OECD transfer pricing guidelines is a consistency rule, which often appears in the transfer pricing legislation or interpretive guidance of OECD countries. A typical example is section 164 of the U.K.’s Taxation (International and Other Provisions) Act 2010, which requires that the country’s transfer pricing legislation “be read in such manner as best secures consistency” between it and the OECD transfer pricing guidelines.

Australia’s Income Tax Assessment Act 1997 contains a similar provision in section 815-135, and Canada is considering adding a consistency rule to subsection 247(2) of the Income Tax Act.

The references to the OECD transfer pricing guidelines that generally appear in consistency rules are undeniably explicit. However, they generally do not allow the OECD guidelines to override domestic law if there is an irreconcilable conflict between the two. A general consistency rule could thus have no domestic legal effect on any measure that is fundamentally at odds with the OECD guidelines, including potentially selective measures that have no plausible basis in the guidelines. This would effectively exclude the OECD guidelines from the reference system precisely when they would be of the greatest use in a state aid assessment.

It’s also unclear from Kokott’s Amazon opinion whether the effect of an explicit reference to the OECD guidelines varies based on the specific legal instrument in which it appears. As the consistency rule example illustrates, even explicit references that appear in the relevant legislative text could have different legal effects.

However, some EU member states instead recognize the OECD guidelines through country-specific forms of administrative or interpretive guidance, which may or may not have binding legal force under domestic law.

The Fiat judgment and Kokott’s Amazon opinion each emphasized that the OECD transfer pricing guidelines are not formally binding on EU member states, and it’s unclear why any nonbinding explanatory guidance issued by national tax administrations should be treated differently.

A handful of other EU countries recognize the relevance of the OECD transfer pricing guidelines through judicial precedent. Swedish and Finnish courts, for example, have accepted the OECD guidelines as interpretive guidance when deciding transfer pricing cases subject to domestic law.

However, judicial recognition of the interpretive value of the guidelines is like a legislative consistency rule. Because the OECD guidelines cannot serve as an interpretive guide for truly incompatible provisions of domestic law, it’s unclear whether general judicial recognition draws the OECD guidelines into the reference system when the relevant measure has no plausible basis in the guidelines.

Perhaps to soften the edges of the explicit reference rule established in Fiat, Kokott’s opinion proposes another way for the OECD guidelines to form part of the reference system. Although Fiat did not acknowledge this possibility, Kokott’s opinion argues that the OECD guidelines may form part of the reference system when the member state’s tax administration has a consistent practice of interpreting domestic law in accordance with the guidelines. But, as paragraph 72 of Kokott’s opinion explains, this possibility is an abstract one regarding the Amazon case:

“In my view, consistent administrative practice which serves to flesh out some feature of a legal definition (that of hidden profit distributions in the present case) could be sufficient to expand the reference system to encompass all or part of the OECD transfer pricing guidelines. However, for consistent administrative practice to be taken into account as a reference system, the commission would have to have investigated that administrative practice in the member state concerned and to have described that in its decision. That has not been done in the present case.”

The commission’s heavy reliance on the 2010 and 2017 versions of the OECD guidelines would have made any attempt futile in the Amazon case. It would have been very difficult for the commission to show that the consistent administrative practice of Luxembourg’s tax administration in 2003 was to follow guidance released in 2010 and 2017.

The most the commission could have achieved would have been to show that the reference system included the 1995 OECD guidelines, which provided no express basis for reallocating returns from LuxSCS to LuxOpCo based on control over risk or functional contributions to the development, enhancement, maintenance, protection, and exploitation of intangibles. Proving that Luxembourg’s consistent administrative practice was to follow the 1995 OECD guidelines would thus have been of little use in the Amazon case.

It’s unclear what the commission would have had to show to succeed with an argument it didn’t make, and Kokott’s opinion is short on detail regarding what conditions must be demonstrated.

The standard for establishing consistent administrative practice as a basis for incorporating the OECD guidelines into the reference system, assuming the CJEU approves Kokott’s recommendation, will thus likely remain murky for the foreseeable future.

Double Deference

Whether the OECD transfer pricing guidelines — or even a domestic variation of the arm’s-length principle — form part of the reference system has clearly been one of the most critical questions raised by the commission’s state aid investigations of APAs.

However, its prospective significance is unclear. At least since the conclusion of the BEPS project, most EU members refer explicitly to the OECD guidelines in some domestic legal instrument. Those that don’t still recognize the arm’s-length principle as the basis for their transfer pricing regimes.

Setting aside the distinction between the OECD’s interpretation of the arm’s-length principle and any interpretations endorsed by domestic law, the practical reality is that every EU member state endorses the arm’s-length principle in some form.

Once the arm’s-length principle has been properly identified as part of the reference system, the next step under Kokott’s proposed framework for assessing individual APAs is to apply what she refers to as a “modified standard of review.”

According to Kokott, respecting member states’ autonomy in direct taxation requires adaptation of the standard established in Gibraltar and confirmed in Commission v. Hungary, C-596/19 P (CJEU 2021), and Commission v. Poland, C-562/19 P (CJEU 2021). As explained in paragraph 94 of Kokott’s opinion:

“There is no apparent reason not to transpose that case-law to situations where the law is misapplied in favour of the taxpayer. It follows, then, that not every incorrect tax ruling but only those which are manifestly erroneous in favour of the taxpayer constitute a selective advantage. Derogations from the applicable national reference system are manifestly erroneous if they cannot be plausibly explained to a third party, such as the Commission or the Courts of the European Union, and are therefore equally evident to the taxpayer concerned. Such cases constitute circumventions of State aid law by means of a manifestly discriminatory application of the law.” [Emphasis in original.]

As Kokott’s opinion makes abundantly clear, this would bind the commission to a highly restrictive standard of review, even if the OECD guidelines form part of the reference system. The fundamental question of whether the transactional net margin method (TNMM) or the comparable uncontrolled price method was best for transactions covered by the 2003 APA granted to Amazon, and how the chosen method was applied, evidently lies outside the scope of Kokott’s recommended standard of review. As explained in paragraph 97:

“The Court of Justice must not examine whether the CUP method or, more correctly, the TNMM should actually have been applied. It only needs to decide — if the OECD Guidelines were the relevant reference system — whether the method applied by the Luxembourg tax authorities (the CUP method in the present case) was manifestly incorrect.”

The paragraphs that follow suggest that the mere recognition of both methods in the OECD guidelines and the commission’s acknowledgment that the CUP method “is relevant in the present case” were enough to establish that the Luxembourg tax administration’s acceptance of the CUP method wasn’t manifestly erroneous and discriminatory.

The standard of review proposed by Kokott is far more deferential than the standard adopted by the General Court in its Amazon judgment, which recognizes state aid when the variation between an approximation of an arm’s-length result and the results of applying the APA-approved method “go beyond the inaccuracies inherent in the methodology used to obtain that approximation.”

It is also distinct from, and in addition to, the highly restrictive standard of review for the underlying provisions of national law established in Fiat.

Adopting Kokott’s modified standard of review would thus stack one level of deference on top of another in a way that may make it almost impossible to prove that an APA constitutes state aid.

Although selecting one transfer pricing method or tested party over another can produce drastic swings in the allocation of income, paragraph 95 of the Fiat judgment implies that a member state’s rules concerning methods are generally immune from state aid scrutiny:

“Without harmonisation in that regard, any fixing of the methods and criteria for determining an ‘arm’s length’ outcome falls within the discretion of the Member States. Although the member States of the OECD recognise the merits of using the arm’s length principle in order to establish the correct allocation of company profits between different countries, there are significant differences between those States in the detailed application of transfer pricing methods.”

Therefore, the commission must accept a member state’s rules governing the selection of a transfer pricing method under Fiat unless they are “manifestly inconsistent with the objective of non-discriminatory taxation of all resident companies.” Then, at least according to Kokott’s opinion, the commission must accept any derogation from the rules adopted by a member state in the form of an APA unless it is “manifestly discriminatory” in favor of the taxpayer concerned.

In other words, Kokott’s modified standard of review for APAs would give member states, which can already choose virtually any method-related rules they wish under Fiat, the freedom to apply the method-related rules they’ve freely chosen in virtually any way they wish.

For reasons that aren’t entirely clear from her opinion, Kokott regards the 2003 APA’s ceiling on LuxOpCo’s returns of 0.55 percent of sales as having crossed her proposed manifestly erroneous and discriminatory threshold.

Although an arbitrarily plucked maximum return is undoubtedly suspect under the arm’s-length principle, Kokott’s explanation for why this feature — and only this feature — exceeds the sweeping autonomy implied by her proposed standard of review is superficial and unconvincing. Paragraph 116 explains:

“Since unconnected independent entities do not usually arrange royalty payments between themselves in such a way that the licence holder never has to pay tax on more than a certain amount (0.55 percent of annual turnover in the present case), that element — indeed in my view manifestly — does not satisfy the arm’s length principle.”

A cap on returns is not the same as a cap on tax liability, and its effect in the Amazon case is modest compared with the effect of selecting the TNMM and using LuxOpCo instead of LuxSCS as the tested party. Even this artificial feature did not support a finding of state aid because the commission failed to prove that the cap actually reduced LuxOpCo’s returns, according to Kokott.

The commission has been widely criticized for overstepping its bounds and using flawed arguments in its APA-related state aid investigations, and for good reason. But limiting the commission’s authority to review APAs in this way, and to this extent, seems to go beyond what’s necessary to protect member states’ tax sovereignty.

Giving member states the authority to allow and apply methods that potentially implement the arm’s-length principle in unreliable ways is questionable. However, once that authority is established, it’s hard to see how holding member states to the methodological rules they chose infringes on their autonomy in direct taxation.