INSIGHT: Transfer Pricing Controversy Takeaways of the Starbucks State Aid Case

On Sept. 24, 2019, the General Court of the European Union issued its judgment on the earlier European Commission decision that an APA issued by the Dutch tax authorities to Starbucks constituted state aid by The Netherlands to Starbucks.

An appeal was filed by both The Kingdom of The Netherlands (T-760/15) and by Starbucks (T-636/16). Per the request of Starbucks, the two cases were joined. The Court concluded that Starbuck’s Dutch APA regarding the arm’s-length remuneration of a Starbucks entity performing group functions in The Netherlands was held not to constitute state aid. While this outcome appears somewhat reassuring, it is not.

The decision makes clear that the European Commission can apply a concept of arm’s-length that is not necessarily consistent with the arm’s-length standard as defined under Article 9 of the respective Model Conventions for the Avoidance of Double Taxation or the OECD’s Transfer Pricing Guidelines. The decision also foreshadows much more scrutiny to transfer pricing matters by European Member State tax authorities and by the European Commission and the need for a much more elaborate and controversy defense-focused transfer pricing approach to withstand such scrutiny going forward.

In the following discussion, some aspects of the Starbucks decision are highlighted that may serve to forewarn taxpayers and aid in avoiding or handling future transfer pricing controversy.

State Aid

Article 107 of the Treaty on the Functioning of the European Union (TFEU) provides in relevant part that EU member states cannot provide aid to companies through state resources that would serve to distort competition and trade within the European Union by favoring certain companies or the production of certain goods. The criteria for the existence of state aid listed in Article 107(1) consist of four factors that need to be met:

1. There needs to be an intervention by the state or through state resources;

2. The intervention must be liable to affect trade between member states;

3. The intervention must confer a selective advantage on the recipient; and

4. The intervention must distort or threaten to distort competition.

Although all four aspects are relevant and each of them needs to be independently met, most gravitas in this case went to the selective advantage criterion. A ruling by the tax authorities that determines a company’s corporate taxes/taxable income, would constitute an intervention by the state and use of state resources (taxes), and as the company in issue is a multinational enterprise (MNE) with cross-border activities, and the ruling arguably lowered Starbucks tax liability in The Netherlands, it affects trade between member countries and therefore is also likely to threaten to distort competition. The aspect of conferring a selective advantage turns around whether the tax measure in issue derogates from the common or normal tax regime that applies to all taxpayers and whether there is a justification for that derogation.

To be able to conclude that there is a derogation of the normal tax regime, the position of the company receiving the challenged tax advantage needs to be compared with that of a company that did not receive such a tax advantage under the normal rules of taxation, according to the court.

Arm’s-Length Pricing Vis-à-vis Remuneration Based on Free Competition

The determination of what would have been the position of a company under normal rules of taxation is to be determined by market prices and market conditions. To conclude that there is state aid, it needs to be determined if the application of the arm’s-length principle (as defined under transfer pricing rules) to Starbucks Netherlands deviated from how stand-alone companies negotiating under free competition would have been remunerated. In other words, did the transfer pricing methodology applied in the APA and accepted by the Dutch tax authorities depart from a methodology that would give a reliable approximation of a market-based outcome (and if it did, was there any justification for that)? The Commission did not feel compelled to study the Dutch domestic transfer pricing rules interpreting the arm’s-length standard and the court unequivocally confirmed that all that the Commission needed to establish was whether there was a selective advantage under Article 107(1) TFEU. Does this mean that a perfect transfer pricing analysis, if such would exist, possibly could still violate EU state aid rules? The latter rules apply a concept of pricing/remuneration under free competition and are not necessarily restricted by transfer pricing methods and related (domestic) conventions. Arguably, yes, it seems that this indeed is the case at least in residual income analysis cases, as discussed hereafter.

Was the APA Derogating From a Normal Tax Regime?

While disputed in the filings, there should be little doubt among transfer pricing practitioners that the strategy for requesting the Starbucks APA was to determine the royalty payable by Starbucks Netherlands by way of a residual income analysis. In other words, Starbucks Netherlands received payments from related EMEA coffee shops for roasted beans and other products sold (food, tea, cups etc.) but not all that income was considered as properly “belonging” to Starbucks Netherlands at arm’s-length. The Dutch manufacturing/roasting company arguably only needed to be remunerated for its own roasting and selling functions.

The APA determined Starbucks Netherlands’ taxable income by applying the transactional net margin method (TNMM). The APA did not determine the royalty amount payable by Starbucks Netherlands to Alki UK by testing that royalty payment separately or by testing the functions/assets/risk of the IP-owning company, but consistent with traditional transfer pricing principles, the APA tested the “least complex” entity in the structure and benchmarked its functions, assets and risks to determine what a comparable entity would/should have earned. By default, all surplus (residual) income received by Starbucks Netherlands in excess of the TNMM-based return was considered to be allocable to the relevant intangibles that Starbucks Netherlands exploits and was paid out as royalty.

The Commission addressed this and maintained that:

• the APA did not consider the license agreement for which the royalty was paid as an important transaction and did not examine the royalty amount (although determining that residual royalty amount arguably was the purpose of entering into the APA);

• the comparable uncontrolled price (CUP) method should have been applied to determine the royalty rather than the TNMM to determine Starbucks Netherlands’ functions, considering the preference expressed in the OECD Transfer Pricing Guidelines for traditional transaction methods over transactional profit methods, and in addition; and

• the royalty should have been zero, because Starbucks Netherlands essentially did not exploit the IP rights.

The court emphasized that the European Commission has the burden of proof that a selective advantage and state aid was granted, and is required to conduct a diligent and impartial examination of the measures in issue. This means that when challenging the APA, the Commission is required to establish that the transfer pricing method applied in no way approximates a market-based outcome for Starbucks Netherlands.

The European Commission was held to not have met its burden of proof:

• While the Commission maintained that the APA—erroneously—did not examine the royalty, the APA indeed did not identify or analyze the transaction meriting the royalty payment from Starbucks Netherlands to Alki UK. The Commission was allocated the burden of proof to show that the methodological error that the royalty itself was not benchmarked led to the absence of a reliable approximation of an arm’s-length outcome. Pointing out the alleged error was not considered sufficient to establish that an advantage was conferred to Starbucks Netherlands that resulted in state aid.

• The Commission argued that the CUP method ought to have been used to determine the royalty amount and that such CUP transactions were available. It was confronted with the observation that it did not establish and present evidence that the CUP method would have led to a lower royalty amount. It was observed that the Netherlands tax authorities are obliged to commence analyzing a transfer pricing audit from the perspective of the taxpayer at the date of the transaction, and Starbucks Netherlands’ transfer pricing was analyzed using the TNMM.

• The Commission did not demonstrate that applying the TNMM analysis to Starbucks Netherlands led to an outcome that was not at arm’s-length, as it did not present an arm’s-length range for the royalty or that the royalty at arm’s-length had to be zero. Furthermore, the Commission had never denied that the coffee roasting IP represented something of value. The court concluded that the Commission did not establish that the APA derogated from a normal tax regime, and therefore it did not meet all the criteria to constitute state aid.

Some Relevant Observations

First of all, the decision makes clear that the European Commission is not bound by the definition of arm’s-length as defined by transfer pricing rules. If the Commission can drive home that the (government approved) pricing deviated from how stand-alone companies negotiating under free competition would have been remunerated and that such pricing led to a selective advantage, that can lead to the conclusion that state aid was granted.

Second, the Starbucks case was lost by the European Commission on the burden of proof. Importantly, the court did not address that by applying a residual income analysis the APA was (strategically) structured to make the royalty payment the default result. However, it is to be expected that in future cases, lessons learned, the European Commission will argue this matter more constructively and will conduct much more detailed transfer pricing analyses. It is likely to benchmark its findings and residual income in support of its positions, and obtain expert assistance in this respect. From a controversy perspective, the Starbucks case foreshadows that economic analyses and benchmark studies underlying intercompany transactions regarding intangibles need to be conducted with a keen eye for detail and simply must be prepared from a controversy defense perspective.

The case and decision raise the question whether substantiation of the arm’s-length nature of any surplus income that is flowing through operating companies and paid onwards intercompany will be required going forward. Chances are that advance approval by way of an APA with respect to a residual income analysis for intangibles will become more challenging to get.

In any case, the concern for state aid and its costly consequences is now squarely embedded in European transfer pricing. The transfer pricing cases that have been raised as state aid by the Commission largely focus on APAs/rulings, which involve explicit approvals by tax authorities. Could the state aid predicate possibly spread to other areas like domestic law safe harbors and tacit approvals after disclosure or horizontal monitoring interactions? The future will tell.

This column does not necessarily reflect the opinion of The Bureau of National Affairs, Inc. or its owners.

Author Information

Monique van Herksen is a tax lawyer and partner with Simmons & Simmons based in the Amsterdam office of the firm. She can be reached at monique.vanherksen@simmons-simmons.com.