Lawyers across Europe holding their breath while awaiting the General Court’s ruling on the Belgian Excess Profits case were doubtlessly disappointed. On Valentine’s Day, the General Court reminded the European Commission that “tough love” is always a possibility, but the Commission’s defeat is no mortal blow. Its decision was annulled on more or less “technical” grounds, but its audacious doctrinal pillars remain unexamined and thus rest intact.
Why is this case important and why was it eagerly expected? The primary reason is that it is the first case from the most recent wave of fiscal state aid investigations to reach the General Court. Ever since June 11th 2014, when then Competition Commissioner Almunia opened in-depth investigations against Apple, Fiat and Starbucks, the Commission’s voice has dominated the relevant dialogue. What Member States, state aid lawyers and tax advisors keep “hearing” is the Commission’s viewpoint, whether in its decisions or its soft law Notice, on a fiercely disputed EU law issue.
In a nutshell, three questions lie at the heart of this debate. First, does an EU law arm’s length principle (ALP) stem from Article 107 (1) TFEU and is it derived from a ‘general principle of equal treatment in taxation’? Second, what exactly is this EU law ALP and where can more details on its substantive content be found? Third, does a tax arrangement (ruling or otherwise) which deviates from said ALP confer a ‘selective advantage’ on the relevant taxpayer?
These three questions are key to the Commission’s state aid analysis and it was hoped that the General Court would examine DG COMP’s assertions and rule on their validity. However, as will be seen, these hopes did not materialize, since the Court annulled the Commission’s decision on different grounds.
The national legal background and the alleged aid measure
The rules for the taxation of income in Belgium are laid down in its 1992 Income tax Code, Article 185 of which specifies that companies are to be taxed on the total amount of their profits. By law enacted in 2004, Belgium introduced new rules concerning the cross-border transactions of affiliated entities of a multinational group. This law provided for an adjustment of the profit subject to taxation in certain cases, an adjustment which could be either upward or downward (this is where the state aid plot thickens). The aim of the 2004 law, according to its explanatory memorandum, was to amend the tax code in order to explicitly include the arm’s length principle and also allow for the issuance of advance tax rulings to taxpayers on a case-by-case basis.
Zooming in on the epicentre of the Commission’s concerns, the downward adjustment allowed by the amended tax code was allegedly intended to avoid or undo potential double taxation. The methodology underlying this downward adjustment of Article 185 (2) (b) of the Belgian Tax Code comprised two steps. In the Court’s own words (paras 22-24), first, the arm’s length prices charged in transactions between the Belgian entity of a group and the companies with which it is associated were fixed based on a transfer pricing report provided by the taxpayer. Those transfer prices were determined by applying the transactional net margin method (TNMM). A residual or arm’s length profit was thus established, which, according to the Commission, corresponded to the profit actually recorded by the Belgian entity.
In the second step, on the basis of a second report submitted by the taxpayer, the Belgian entity’s adjusted arm’s length profit was established by determining the profit that a comparable standalone company would have made in comparable circumstances. The difference between the profit arrived at following the first and second steps (namely the residual profit minus the adjusted arm’s length profit) constituted the amount of excess profit which the Belgian tax authorities regarded as being the result of synergies or economies of scale arising from membership of a corporate group and which, accordingly, could not be attributed to the Belgian entity.
Under the scheme at issue, that excess profit was not taxed. According to the Commission, that non-taxation granted the beneficiaries of the scheme a selective advantage, particularly since the methodology for determining the excess profit departed from a methodology that leads to a reliable approximation of a market-based outcome and thus from the arm’s length principle.
This conclusion led the Commission to the adoption of a negative state aid decision with recovery in January 2016, the annulment of which was at stake before the General Court.
The applicants’ arguments
In their action for annulment, the Kingdom of Belgium and Magnetrol International (one of the alleged aid beneficiaries), essentially put forward four pleas in law. First, they asserted that the Commission had encroached upon the exclusive tax jurisdiction of Belgium and thus exceeded its powers. Second, they maintained that the Commission had committed an error in law by identifying a “state aid scheme” within the meaning of Article 1 (d) of Regulation 2015/1589. Third, they argued that the advance rulings did not amount to state aid since the Treaty’s conditions were not fulfilled. Finally, they accused the Commission of violating the principles of legality and of the protection of legitimate expectations by ordering the recovery of the alleged aid. The General Court rejected the first plea but upheld the second plea; it did not examine the third and fourth pleas. Let us now see how it all unraveled.
Exclusive tax jurisdiction versus…exclusive EU competence
The argument here is not novel. In fact, Member States have been employing various iterations of the same argument ever since the inception of the European Union. Its crux is as follows: direct taxation falls within the competence of Member States. It is one of the last vestiges of state sovereignty in an ever-evolving supranational project and it should, for numerous reasons, not be interfered with. Therefore, fiscal state aid scrutiny by the Commission is not welcome.
The EU judicature’s reply has always been fairly consistent. In what has become a jurisprudence constante (see e.g. Case 173/73), the CJEU has made it clear that Member States must exercise their direct tax competence consistently with EU law, and their interventions are not automatically shielded from Article 107 TFEU simply because they are of a fiscal nature. Belgium’s exclusive competence in direct tax matters is monitored by, inter alia, the EU’s exclusive competence in competition law matters, which include the state aid prohibition (Article 3 (1) (b) TFEU). The applicants’ first plea in law is thus summarily shot down. In the Court’s words (para 67): ‘since the Commission is competent to ensure compliance with Article 107 TFEU, it cannot be accused of having exceeded its powers by examining the measures comprising the alleged scheme at issue’.
However, before moving to the second plea in law, the General Court gives Belgium a rap on the knuckles and makes a remark that could well prove important in the future. Belgium had argued that it is exclusively competent to adopt measures that prevent double taxation, such as the alleged aid measure. Notwithstanding the fact that, as already mentioned, such an assertion, even if true, would not suffice to protect the measures from the Commission’s scrutiny, the General Court went a step further and noted that ‘it does not appear that the non-taxation of excess profit, as applied by the Belgian tax authorities, pursued the objective of avoiding double taxation’ (para 72). In essence, this means that, should the alleged aid measures’ classification as state aid make its way back before the Court and prima facie selectivity be established, the “avoidance of double taxation” justification will most probably not be available to them and they will be found to be conclusively selective.
The state aid scheme that never was
The applicants’ argument that was eventually upheld by the General Court and decided this case was that the Commission had incorrectly characterized the relevant Belgian measures as constituting an ‘aid scheme’ within the meaning of Regulation 2015/1589. Under Article 1 (d) of that Regulation, an ‘aid scheme’ means ‘any act on the basis of which, without further implementing measures being required, individual aid awards may be made to undertakings defined within the act in a general and abstract manner and any act on the basis of which aid which is not linked to a specific project may be awarded to one or several undertakings for an indefinite period of time and/or for an indefinite amount’.
Why is this classification important in practical terms? If the Belgian measures indeed constituted an aid scheme, the Commission would have been allowed to simply examine the characteristics of the scheme at issue to determine its state aid character; an analysis of the aid granted in individual cases under the scheme would have been unnecessary. This difference in the breadth of review is crucial, and thus the aforementioned classification bears practical significance. This is the reason why the Commission had devoted almost four pages on this issue in its final decision (paras 94-110). Alas, in the General Court’s view, its conclusion was erroneous.
As explained in the judgment itself, three cumulative conditions need to be fulfilled in order for a state aid scheme to be found to exist. Firstly, further implementing measures should not be required, which means that the essential elements of the aid scheme in question must necessarily emerge from the provisions identified as the basis of the scheme (para 86). In the Court’s view, this was not the case, since not all essential elements emerged from the Belgian acts identified by the Commission (paras 93-94 & 98). Secondly, national tax authorities cannot have any margin of discretion as regards the aforementioned essential elements of the aid and whether it should be awarded (para 87). Again, in the Court’s view, this condition was not fulfilled either, since the Belgian tax administration ‘enjoyed a genuine margin of discretion’ on whether to grant downward adjustments (para 104). Thirdly and finally, for an aid scheme to exist, the beneficiaries must be defined in a general and abstract manner (para 88). Unsurprisingly, the General Court concluded that they were not defined in such a manner and further implementing measures needed to be taken to define the alleged aid beneficiaries (paras 115-119).
The absence of just one of the three conditions analysed supra would have been enough to call the Commission’s conclusions into question, so the absence of all three was a devastating blow to its decision’s validity. The Court’s review was rigorous (see, for instance, paras 127-128) and its remarks were, at times, perhaps slightly mordacious (see e.g. its comment on the Commission’s apparent ambivalence in paragraph 82). The judges’ conclusion came naturally: the Commission ‘erroneously considered that the Belgian Excess Profit System […] constituted an aid scheme’ (para 135). The consequence was that the Commission’s decision had infringed Article 1 (d) of Regulation 2015/1589 and had to be annulled in its entirety.
As mentioned earlier, out of the four pleas raised by the applicants, only the first two were scrutinised by the Court, since the second plea was well-founded and thus sufficient to decide the case. This meant that the third plea, which concerned the fulfillment of the substantive state aid conditions and the existence of a ‘selective advantage’, did not receive any attention. This is regrettable, because it concerns the single most controversial assertion made by the Commission in its recent fiscal state aid crusade. As a result, we are none the wiser on whether an EU law arm’s length principle exists and ‘necessarily forms part of the Commission’s assessment under Article 107 (1) of the Treaty’ as maintained by the Commission in paragraph 150 of its now-annulled decision on the Belgian excess profits regime.
Still, we will hopefully receive an answer when the General Court rules on the validity of the Commission’s final decisions in the Fiat and Starbucks cases, where the same argument is put forward by the Union’s competition watchdog. The ongoing fiscal state aid wave started off with these two cases (plus the Apple case) in June 2014, and they were also the first ones to be decided by the Commission in October 2015. It is, perhaps, fitting that they will also be the first cases to allow us to discern where the Court stands on these questions.
Compared to how important this judgment could have been, had it touched upon the most contested issues in this area of law, the General Court’s ruling in Cases T‑131 & 263/16 could be said to be rather anticlimactic. So be it. In any case, both disappointment and celebrations by the parties are premature at this point in time; the real battle is to be fought on another day.