Op-Ed: “The Apple Case: Who wins? What’s next?” by Adolfo Martín Jiménez and Jorge Piernas López
On July 15th, the General Court (‘GC’) rendered its much awaited judgment in cases Ireland v Commission (T-778/16), and Apple Sales International and Apple Operations Europe v Commission (T-892/16), annulling a Commission Decision on State aid granted to Apple by Ireland. The outcome was expected since much of the relevant principles in the judgment were already established in previous case law (notably but not exclusively, Starbucks and Fiat).
The Apple judgment can be read in different forms: from a tax policy (or, maybe, politics?) perspective, from a purely State aid standpoint, or from a strictly transfer pricing one. Before advancing some conclusions on these different fronts, it is useful to recall the main issue discussed in the case. Due to a well-known, consented to by both sides of the Atlantic, mismatch or gap between the United States and Ireland corporate tax legislation, it was perfectly possible for a US company with an Irish subsidiary managed from the US to obtain ‘ocean profits’, namely profits that were not taxable either in Ireland or in the US. The Commission decision in the Apple case wanted to correct that perfectly legal (although, arguably, ‘undesirable’ ) outcome by applying Article 107 TFEU in a ‘creative new form’: there was State aid (selective advantage) if the untaxed ‘ocean profits’ directly linked, in the specific case, with the IP licence controlled by the Irish subsidiaries were not taxed in Ireland. In fact, the Irish companies were treated as non-resident entities in Ireland, and the IP licence was managed and allocated to the head office in the US without the profits linked with the IP being taxed in the US or Ireland (the latter only taxed income effectively connected with functions performed in Irish territory).
From a policy perspective, the main effect of the GC judgment is to correct the innovative use of Article 107 TFEU by the Commission as a tool to tax ‘untaxed profits’ and close international tax planning schemes of multinationals (‘MNEs’). It is true that the judgment seems to start by (unsurprisingly) rejecting Apple’s and Ireland’s arguments that Member States and their tax policies are not subject to scrutiny under the State aid rules, and that sovereignty is not a valid exception to the application of Article 107 TFEU. However, the judgment makes very clear that the parameters to analyse whether there is State aid need to be defined by taking into account the tax system of the Member States, which severely limits the use by the Commission of Article 107 TFEU as a tool for disguised tax harmonisation or to close gaps facilitating international tax planning.
Paradoxically, however, the defeat for the Commission may also support other Commission initiatives. First, the GC seems to hint that if there was ‘State aid’ for Apple, it was granted by the US, where the relevant functions regarding the licensed IP were conducted and related profits untaxed, not by Ireland. Something should probably be done in a case where a third State is subsidising MNE groups operating in the internal market, and the Commission already has recently presented a proposal that, albeit indirectly, is strongly backed by the GC judgment (White Paper on levelling the playing field as regards foreign subsidies). Second, the victory of Apple may be taken by some as an (unfortunate) implicit support of another (populist?) measure promoted by the Commission and many Member States: digital service taxes (the problems of these taxes with the rule or law principle and international obligations or their difficult interaction with OECD’s current works are well-known and will trigger further conflicts).
From a State aid perspective, the judgment is not innovative but still further unfolds already established principles. First, it stresses the relevance, for the Commission, which has the burden of proof concerning the existence of a prima facie selective advantage, of correctly defining the right reference framework (‘normal taxation’) by paying strict attention to the domestic tax system of the specific country, including case law, before deciding that there is a selective advantage. For this purpose, the GC permits the use, as a benchmark to allocate profits to permanent establishments and branches, of some (not all) of the basic features of the OECD’s Authorized Approach (profits follow relevant functions, assets and risks carried on and controlled within a country). However, the GC does not seem to tolerate fictions –the ‘exclusion approach’ defended by the Commission as a first line of reasoning – leading to linking profits with a country where its tax system does not employ such a fiction or nexus rule. That is to say, if Ireland allocates profits to its companies and permanent establishments based on the functions they carry on within Ireland, the Commission’s view that profits linked with an IP licence should be allocated to Ireland when the functions connected with it are undertaken in the US does not hold ground, even if that means that such an income remains untaxed.
Second, the judgment confirms the soundness of the Commission’s joint analysis of the requirements of advantage and selectivity in the tax context, and its policy of closely scrutinizing tax rulings, although, like Starbucks, it places a notable high burden of proof on the Commission’s side to show that those requirements are met. It is not sufficient to identify that there are methodological errors or a defective analysis in the transfer pricing documentation and the ruling (as there were in the Apple and Ireland rulings), the Commission must go beyond that threshold and prove that these mistakes entail the concession of a selective advantage to the beneficiary of the ruling. The analysis proving that an advantage is granted also needs to be consistent with the facts of the specific case: the (subsidiary) analysis of the Commission reasoning in the Apple case (that the rulings did not produce an arm’s-length result) was permeated by incorrect assumptions and ‘side effects’ of its primary line of reasoning (allocation of profits of IP licence to the Irish branches to eliminate the advantage), with the consequences that many of its statements regarding the method chosen (‘TNMM’) to allocate profits to the Irish branches and its premises (tested party, choice of operating costs as profit level indicator and levels of return accepted) were flawed.
Third, the judgment also confirms, in line with previous case law, that, contrary to the Commission’s alternative line of reasoning in the case (the exercise of discretion in the rulings meant granting State aid), discretion of the public authorities as part of a ruling policy cannot be automatically identified with the finding of selectivity, although the reasoning is a bit difficult to follow because this conclusion was also directly connected with the fact that the Commission was unable to show that an advantage had been granted to Apple by Ireland, and, as mentioned, advantage and selectivity in this case were examined together.
Lastly, from a transfer pricing perspective, the Apple judgment makes clear that the Irish rulings and the transfer pricing policy they blessed rested on shaky grounds. This is a clear and loud warning for MNEs and Member States that offer ‘sweet deals’ in the form of tax rulings or advance pricing agreements. This time Apple was lucky because the Commission did not do its job properly, but there are sufficient signs in the judgment showing discontent with the outcome. There are also hints that a more refined job by the Commission would have tilted the balance against the MNE group, even if the resulting aid amount in that case would have probably been less spectacular and eye-catching for the public and mass media. The Commission is also given a template for how to do it right next time, so weak transfer pricing policies and tax rulings or advance pricing agreements should be reinforced and reconsidered by MNEs and tax authorities.
Whether the case will be appealed by the Commission is not known yet. At first sight, the law principles established by the GC are sound, although there are some unclear technical issues in part of the reasoning. For Apple it is a major success, one that may not last long, however, in view of the policy options this judgment will probably fuel. For the EU, the victory is not less significant: the State aid legal approach towards tax rulings has been validated even if corrected, the Commission has been shown how to proceed to assess the effects of tax rulings, the US arguments regarding discrimination against US companies have been dismantled and, more importantly, the rule of law has prevailed against attempts by the Commission to unduly expand the scope of Article 107 TFEU. Finally, the judgment stimulates new battlefields (for example digital sales taxes, application of the EU subsidy control mechanisms to third countries) where fights will be even harder than in this case.
Adolfo Martín Jiménez is Professor of Tax Law, University of Cádiz, Spain.
Jorge Piernas López is Associate Professor of International and EU Law, University of Murcia, Spain.
Members of the Editorial Board of EU Law Live