Faculty of law blogs / UNIVERSITY OF OXFORD

EU State Aid Law and Transfer Pricing: A Critical Introduction to a New Saga


Liza Lovdahl Gormsen


Time to read

4 Minutes

The Court of Justice of the European Union (‘CJEU’) has made it clear from the early stages of its jurisprudence that State Aid rules are applicable to taxation measures. Thus, the fiscal nature of a measure does not in itself preclude the Commission’s scrutiny under Article 107(1) TFEU et seq. A number of cases have contributed to the development of the concept of fiscal aid, and the Commission has tried to provide further guidance through a series of soft law instruments. More recently, the Commission has been concerned with the application of State Aid control to instances of aggressive tax planning and, in particular, to cross-border transfer pricing rules. In a recent article, I explore the controversies surrounding the Commission’s investigations into advance pricing arrangements (‘APAs’) between national tax authorities and multinational companies. This issue has acquired particular prominence following the Commission’s recent decision in relation to Apple, which ordered the recovery of €13 billion of fiscal aid.

In 2001, the Commission initiated a number of investigations into transfer pricing schemes launched inter alia by Belgium, Luxembourg and the Netherlands. After a period of dormancy, the Commission launched a second wave of investigations into transfer pricing, but this time focusing on individual tax rulings issued by national authorities in relation to Apple, Starbucks, Fiat Finance and Trade, and Amazon. Moreover, the Commission has initiated an investigation in the McDonald’s case, involving a ruling on the interpretation of a Double Taxation Agreement. In my paper, I critically analyse the Commission’s interpretation and application of EU State Aid Law in these cases and I argue that this approach is not based on firm legal grounds.

First, the Commission does not properly distinguish between ‘selectivity’ and ‘economic advantage’, and in fact an examination of ‘selectivity’ was absent in the opening decisions cited above. The examination of selectivity was introduced in the final decisions of Starbucks and Fiat in the meantime. However, the point still applies to the extent that a number of Commission decisions and CJEU jurisprudence in the past linked economic advantage in fiscal aid with selectivity, arguing that an advantage implies prima facie selectivity – without appreciating the different meaning and purpose served by each element. Selectivity should be identified in every individual case, by means of the three-step test developed by the CJEU. Moreover, the benchmark used for the selectivity analysis should not be construed widely, and multinational companies that are in a comparable legal and factual situation should be recognised as a segregated category, and therefore merit a tailored treatment. Cross-border transfer pricing by definition can only apply to multinational companies, thus the appropriate benchmark cannot include standalone companies.

Second, the Commission puts forward the novel principle of a ‘prudent market operator’. This principle, strangely enough, does not purport to assess the State’s behaviour and whether it was in line with what an economic operator would have done. Rather it requires the beneficiary subject to a ruling on transfer pricing to only accept terms which a ‘prudent economy operator’ would have accepted in view of EU State Aid rules. While this author acknowledges the attractiveness, in principle, of an objective criterion to assess the existence of an economic advantage, this proposed test was rife with problems and moreover inconsistent with the OECD Model Tax Convention. As predicted, the prudent economy operator test was dropped in the final decisions of Starbucks and Fiat, and it can be reasonably expected that this will happen with the Apple and Amazon final decisions.

Third, my article uncovers the origin of the Arm’s Length Principle (‘ALP’) applied by the Commission to assess the existence of an economic advantage, that is, the OECD Model Tax Convention and the accompanying OECD Guidelines. It is argued that a degree of divergence can be observed between the interpretation of that principle by the OECD and the way in which the Commission applied it over the years. Arguably, the Commission’s interpretation and application of the ALP in the 2001 investigations was superficial and manifested an air of exploration by the Commission. In fact, the Court of Justice in Forum 187 did not even mention the principle by its name. In the recent opening decisions (and this is also clearly evident in the published final decisions), the Commission has embarked on an aggressive application of this principle as if it were an exact science which produces a precise result on which economic advantage can be determined. In its opening decision in the Apple case, for instance, the Commission substantiated the existence of both an economic advantage and selectivity on the alleged violation of the ALP. In the recent Press Release (30 August 2016), regarding the recovery of the aid allegedly granted by Ireland, the Commission reaffirmed that the method used for the allocation of profits between the so-called ‘head office’ and the Irish subsidiaries of Apple ‘has no factual or economic justification’. This departure from the ALP, constitutes the essence of the undue advantage which was allegedly granted to Apple. It will be interesting to read the much anticipated non-confidential version of the decision. In any case, the notion of the ALP articulated by the Commission is questionable: firstly, the OECD Guidelines acknowledge that either of five techniques may be used to assess whether the principle has been observed; secondly, while acknowledging that the choice of the methodology depends on a number of circumstances, the same Guidelines also note that each method will not generate a precise result, but likely an approximate range. More crucially, the Commission has imposed the ALP, as interpreted by it, on all Member States. However, in the absence of harmonisation on transfer pricing assessment in the EU, no such principle can exist. Hence any assessment must be based on national rules.

It is submitted that the Commission has taken a number of shortcuts (sometimes against its past practice and general principles of EU law) in an effort to develop the law and to expand its remit. This is conceptually and normatively problematic, but it seems that the Commission will maintain this approach, unless the CJEU says otherwise.

Liza Lovdahl Gormsen is Director of the Competition Law Forum & Senior Research Fellow in Competition Law at the British Institute of International and Comparative Law.


With the support of