Some cases just have it all; the Apple case is one of them. First, size: at more than thirteen billion euros, the recovery order Ireland had to enforce dwarfed the previously biggest one (EDF, at around one billion euros). Second, international political implications: the case ignited transatlantic tensions between the EU and the USA, both under President Obama and, less elegantly, under President Trump. Third, domestic political implications: Ireland was stuck between a rock and a hard place, having to justify resisting the temptation of accepting the Apple “windfall” in order to defend the integrity of its tax system. Finally, legal complexity: the Apple case was the flagship of the Commission’s ongoing wave of fiscal state aid investigations targeting tax rulings, formally launched in June 2014.
This flagship has now been torpedoed. In a judgment handed down on July 15th, the General Court annulled the Commission’s 2016 decision in its entirety, concluding that ‘the Commission did not succeed in showing to the requisite legal standard that there was a selective advantage for the purposes of Article 107(1) TFEU’ (para 507).
In this post, the background to the dispute will initially be set out. It will then be shown how Ireland and Apple achieved this important legal victory. However, it will be argued that the Commission’s defeat, as a matter of legal principle, is not as devastating as it might initially seem. The post will conclude with some rumoured legislative developments and the next steps for the European Commission.
A bit of background
The Commission’s investigation into the tax affairs of Apple in Ireland was formally launched in June 2014 by Competition Commissioner Almunia, together with probes against Luxembourg (Fiat) and the Netherlands (Starbucks). It was concluded in August 2016 with a negative decision, ordering Ireland to recover 13 billion euros (plus interest) from Apple, an amount that was eventually placed into an escrow account. That decision, which had received extensive media coverage globally, is the decision that has now been annulled by the General Court, after almost four years.
The structure of the Apple group during the time period covered by the decision (and as described therein), was as follows. The Apple Group is composed of Apple Inc. and all companies controlled by Apple Inc., the latter being headquartered in the United States of America. Of the Apple Group companies incorporated in Ireland, but not tax resident in Ireland, Apple Sales International (ASI) is a fully owned subsidiary of Apple Operations Europe (AOE), which in turn is a fully owned subsidiary of Apple Operations International (AOI), which in turn is a fully owned subsidiary of Apple Inc.
We only need to focus on AOE and its subsidiary, ASI, which were the alleged aid beneficiaries and therefore the protagonists of this case. Both AOE and ASI operated in Ireland through branches.
According to the Commission, the taxable profits of Apple Sales International and Apple Operations Europe in Ireland were determined by a tax ruling granted by Ireland in 1991, which in 2007 was replaced by a similar second tax ruling. These tax rulings were the contested state aid measures. For the reasons briefly set out in the section that follows, the Commission’s conclusion in its 2016 decision was that these rulings amounted to state aid, since they “substantially and artificially” lowered the tax paid by Apple in Ireland by endorsing a method of establishing the taxable profits of AOE and ASI in a way that did not correspond to economic reality. This is because almost all sales profits recorded by the two companies were internally attributed not to the Irish branch of the companies but to a “head office”, were they remained untaxed. These head offices allegedly only existed on paper and could not have generated such profits.
Therefore, Ireland, by issuing these tax rulings, had “validated” an attribution of profits within AOE and ASI that unduly favoured said companies, thus granting them a very low tax effective tax rate and a corresponding state aid advantage.
The Commission’s case
In a nutshell, the Commission’s Apple decision is made up of a (more or less) primarily “state aid” part and a (more or less) primarily “tax” part. I hope the state aid cognoscenti will forgive this crude distinction, but it is easier to classify and explain the Commission’s complicated arguments on this basis.
The state aid part, on paper, is simple: the tax rulings amount to state aid under Article 107 TFEU if they fulfil all relevant conditions: advantage, selectivity, imputability to Ireland, use of state resources, distortion of competition and effect on trade. That’s it. In truth, only the “selective advantage” condition(s) seemed to pose a problem: the remaining conditions fell like dominoes.
The most audacious legal claim advanced by the Commission in the state aid part of its Apple case was similar to the one advanced in its Fiat and Starbucks decision, but with an added twist. First, the Commission interprets the ECJ’s 2006 Forum 187 case in a way that means that ‘a tax measure which results in an integrated group company charging transfer prices that do not reflect those which would be charged in conditions of free competition, that is prices negotiated by independent undertakings negotiating under comparable circumstances at arm’s length, confers a selective advantage on that company, in so far as it results in a reduction of its taxable base and thus its tax liability as determined under the ordinary rules of taxation of corporate profit’ (para 251, Apple decision). The added twist is the Commission’s argument that the ‘same principle applies to the internal dealings of different parts of the same integrated company, such as a branch that transacts with other parts of the company to which it belongs’ (para 253, Apple decision).
Therefore, the crux of the Commission’s state aid case was that ‘if it can be shown that the profit allocation methods endorsed by Irish Revenue in the contested tax rulings result in a taxable profit for ASI and AOE in Ireland that departs from a reliable approximation of a market-based outcome in line with the arm’s length principle, those rulings should be considered to confer a selective advantage on those companies for the purposes of Article 107(1) of the Treaty’ (para 258, Apple decision).
In other words: by proving a deviation from the arm’s length principle, which in practice would mainly be done by relying on OECD guidelines, the Commission would prove the conferral of a selective advantage on AOE and ASI: case closed.
The Commission decision then left the state aid “realm” and moved to the tax part from para 258 onwards, attempting to demonstrate exactly how such a deviation had, in fact, taken place. To do so, the Commission relied on primary (section 22.214.171.124) and subsidiary (section 126.96.36.199) lines of reasoning, as well as on two alternative lines of reasoning (sections 188.8.131.52 and 184.108.40.206).
For reasons of brevity, I will only refer to these lines of reasoning by way of “headlines”. The reader is free to delve deeper. In its primary line of reasoning, where the reference framework for the selectivity exercise was deemed to be the ordinary rules of taxation, including Section 25 of Ireland’s TCA 1997 about non-resident companies such as AOE and ASI, DG COMP asserted that Ireland had wrongly accepted the unsubstantiated assumption that the Apple IP licenses held by ASI and AOE should be allocated outside of Ireland. In its subsidiary line of reasoning, it maintained that AOE’s and ASI’s taxable profits had been undervalued due to the inappropriate methodological choices underlying the one-sided profit allocation methods endorsed by the contested tax rulings.
In its first alternative line of reasoning, DG COMP contended that, even if the appropriate reference framework for the selectivity comparison could be deemed to be the more specific Section 25 of Ireland’s TCA 1997 which only concerns non-resident companies, the contested tax rulings granted ASI and AOE a selective advantage in the form of a reduction of their taxable profit upon which corporation tax is levied under that provision, since the arm’s length principle is inherent in the application of Section 25 TCA 1997. Finally, in its second alternative line of reasoning, DG COMP argued that, even if it were not inherent, the contested tax rulings should still be considered to confer a selective advantage on ASI and AOE, since they would then be the result of discretion exercised by Irish Revenue in the absence of objective criteria related to the tax system.
The crux of the General Court’s reasoning in the Apple Judgment
When one reads the entire judgment (which, at 92 pages, is quite long), it becomes evident that the General Court was very methodical in its assessment of the Commission’s case and the applicants’ pleas against it. The author will obviously not endeavour to summarise the entire ruling in this blog post, but will only focus on the Court’s conclusions in relation to the Commission’s primary, subsidiary and alternative lines of reasoning.
As regards the Commission’s primary line of reasoning, as summarized by the Court in paragraphs 246 et seq. of its ruling, it was found that the Commission made errors concerning the application of, first, section 25 of the TCA 97, second, the arm’s length principle, and third, the Authorised OECD Approach, i.e. the 2010 OECD Profit Attribution Report which determines a common approach to the allocation of profits to permanent establishments of non-resident companies. Moreover, as stated in para 310 of its ruling, ‘the Commission has not succeeded in showing that, in the light, first, of the activities and functions actually performed by the Irish branches of ASI and AOE and, second, of the strategic decisions taken and implemented outside of those branches, the Apple Group’s IP licences should have been allocated to those Irish branches when determining the annual chargeable profits of ASI and AOE in Ireland’. The Commission’s primary line of reasoning was, therefore, found not to be convincing.
As regards the Commission’s subsidiary line of reasoning, as summarized by the Court in paragraph 480 of its ruling, the Commission did not succeed in demonstrating that the methodological errors to which it had referred with regard to the profit allocation methods endorsed by the contested tax rulings actually led to a reduction in ASI and AOE’s chargeable profits in Ireland. These alleged methodological errors were, in summary: the choice of the Irish branches as tested parties, the choice of the operating costs as the profit level indicator, and the actual levels of return accepted by the contested tax rulings. The Commission’s subsidiary line of reasoning was also, therefore, found not to be convincing.
As regards the Commission’s first and second alternative lines of reasoning, as summarized by the Court in paragraphs 487 and 493 of its ruling, the Court simply noted that it was based on assertions made in its primary and subsidiary lines of reasoning, which it had in fact been unable to prove. As regards DG COMP’s claim that the Irish tax authorities had exercised an overly broad discretion, the Court found that it had been unable to show that this had actually been the case (para 495).
Naturally, following the above, the General Court found (para 505) that the pleas in law relied on by Ireland and ASI and AOE against the assessments made by the Commission in connection with its primary, subsidiary and alternative lines of reasoning had to be upheld, thus making the annulment of the Commission’s 2016 decision inevitable.
DG COMP’s defeat is not complete, even though it might seem so at first glance. First, the General Court, as shown in para 246, accepted the “ordinary rules of taxation” in Ireland as the appropriate reference framework in the selectivity exercise. This is key since the reference framework is so broad that a derogation can easily be identified. Second, as expected after the Starbucks and Fiat judgments – which were handed down by the same Chamber in September 2019 – the General Court continues to support the Commission’s right to scrutinise national tax rulings based on an EU law-derived (but de facto OECD-inspired) arm’s length principle. Going one step further, in its Apple judgment the Court openly approved (para 247) of the Commission’s right to use the same tool to judge the allocation of profits within the same company (legal person), and not just between related companies (different legal persons of the same group), like in the Fiat and Starbucks cases.
On a different note, in an unusual quasi “obiter dictum”, which was also repeated in its press release, the General Court stated that it ‘regrets the incomplete and occasionally inconsistent nature of the contested tax rulings’ (para 479), but could not possibly hand the victory to the Commission for that reason alone, since it did not successfully carry the onus of proof that encumbered it.
It is difficult to predict whether the Commission will appeal to the ECJ on points of law, since it mainly lost on points of fact and due to matters relating to the burden of proof. Sure, an appeal may be lodged, but the Commission’s defeat is relatively similar to its defeat in the Starbucks case, where it won on matters of legal principle, but lost for being unable to demonstrate the conferral of a selective advantage “to the requisite legal standard”. For reasons I have explained elsewhere, the Commission ultimately decided not to appeal the Starbucks judgment of the General Court. Some of these reasons also apply to the Apple case, but perhaps the sheer size of the case, coupled with its unprecedented media coverage and the symbolism it carries, will force the Commission’s hand and lead to an appeal to the Union’s highest court regardless of the actual chances of success.
It is not inconceivable that DG COMP might decide to launch a new investigation against Apple and then adopt a new decision. Perhaps the latter will incorporate the lessons learned from the Starbucks and Apple litigation, but especially from the way in which it won the Fiat case before the General Court.
It’s also possible that the Commission goes down both routes simultaneously. As stated in a previous post, when the Commission lost the Belgian Excess Profit rulings case before the General Court in February 2019, not only did it appeal the judgment before the ECJ, but also opened 39 individual state aid investigations as a kind of “insurance policy” in case the annulment of its first decision is eventually upheld by the ECJ.
For the time being, the immediate reaction of Vice-President Vestager follows the familiar line: the Commission will ‘carefully study the judgment and reflect on possible next steps’. But the conclusion of her statement is more interesting: ‘[s]tate aid enforcement needs to go hand in hand with a change in corporate philosophies and the right legislation to address loopholes and ensure transparency.’
The “right legislation” most probably refers to the, so far neglected, Article 116 TFEU, which stipulates that when there is a distortion to the ‘conditions of competition in the internal market and […] the resultant distortion needs to be eliminated’, the Union can adopt legislation on the basis of qualified majority voting, therefore bypassing the unanimity requirement in tax matters. This provision has never so far been used as a legal basis for secondary legislation, but reports have been surfacing (see e.g. here by the FT) that the Commission is seriously considering “dusting off” Article 116 TFEU. In an opinion piece for the Financial Times, Commissioner Gentiloni stressed that ‘we must stand ready to activate all existing policy levers to protect our single market’, including ‘EU treaty provisions that let taxation proposals be adopted not by unanimity but by qualified majority’. One wonders: for whom the bell tolls?
The Apple case is, from a symbolic standpoint and in terms of size, the most important fiscal state aid case of the past decade. In that sense, the Court’s judgment is a blow to the European Commission, especially when it is the third of its four tax ruling decisions to have met a similar fate. Still, the judgment’s legal significance, as explained supra, is much more specific and should not be exaggerated. For the Commission, and VP Vestager in particular, this is surely a setback, but not the end of the road.