The General Court’s awaited ruling in the Apple[1] case contains some surprising parts, and it is not easily reconciled with case law from this Court and from the Court of Justice. But as in the Fiat[2] and Starbucks[3] cases, it reveals an impressive effort to analyse the issues at depth.

The facts are described in the decision of the European Commission and the ruling of the General Court. Very shortly, two companies, ASI and AOE, were registered in Ireland but were not resident there for tax purposes. The two companies had branches in Ireland, and the question concerns the attribution of profits to these branches. Through two rulings issued in 1991 and 2007, the Irish tax administration accepted certain methodologies to determine the remuneration of the branches. The European Commission found that the rulings implied the grant of illegal State aid, and ordered the recovery of such aids.[4] The Commission relied on three main types of arguments: the non-allocation of the Apple Group’s IP licences to the Irish branches of ASI and AOE (primary argument), the inappropriate choice of methods for allocating profits to those Irish branches (subsidiary argument), and the discretionary manner with which the rulings had been granted (alternative argument). In this blog post only a few aspects of the case are discussed.

The General Court ruled, firstly, on the right of the Commission to examine tax rulings in the light of the State aid rules. The “principle of fiscal autonomy” was found not to prevent State aid control, something that is justified by the well-known formulation according to which “while direct taxation, as EU law currently stands, falls within the competence of the Member States, they must nonetheless exercise that competence consistently with EU law”.[5] The opposite view would have been surprising, given the breadth of the tax measures that have already been found in breach of the State aid rules.

The General Court further considered, rightly in my view, that in cases such as tax rulings that may deviate from normal taxation, “the examination of advantage overlaps with the examination of selectivity”:[6] a deviation from the reference system implies necessarily a difference in treatment for the undertakings that are subject or not to the normal level of taxation. That the criteria overlap does not save the Commission from having to prove the benefit from a tax relief and the fact that such a tax relief is not available to all undertakings, as emphasised by the outcome of this case. It also supposes that the undertakings subject to different levels of taxes are in a comparable situation.

The General Court then moved on to the selectivity analysis. The Court analysed certain theoretical questions connected to the selectivity criterion, and endorsed the view according to which selectivity does not necessarily require a departure from the reference system, a difference in treatment between comparable undertakings being enough.[7] This view is not new, and has already been accepted by the Court of Justice, e.g. in the Asociación Española de la Industria Eléctrica case.[8] The General Court, however, applied the traditional method of analysis, starting with the determination of the reference system. The judges followed the Commission and found that the reference system should be broad enough, and determined both by the measure that is being examined and by the objective of the tax system.[9] This means that the rules for the attribution of profits to permanent establishments could not, as such, constitute the reference system. Were that to be the case, it would generally be allowed to treat residents and non-residents differently. In my view this finding is justified,[10] as it enables through State aid control the examination of whether or not taxes have been levied in a manner that is consistent with the objective of a tax system, without being dependent on the regulatory method used by the lawmaker. The downside might be the difficulty to correctly determine the objective of the tax system, which is an issue inherent to the selectivity analysis.

As a result, “the contested tax rulings form part of the general Irish corporation tax regime, the objective of which is to tax the chargeable profits of companies carrying on activities in Ireland, be they resident or non-resident, integrated or stand-alone”.[11] For this finding to be effective, the General Court found, consistently with the Fiat and Starbucks cases, that these categories of companies are in a comparable situation in the light of the objective of the corporate income tax system. This is an important point, which in my view the Court analysed correctly: it makes sense to consider these categories of undertakings as being comparable with regard to the objective of the tax system, to enable through the State aid rules an assessment of the general coherence with which a tax systems treats different types of undertakings.[12] At the same time, it does not either mean that all these undertakings must, eventually, be treated equally: differences in treatment might be justified by the logic of the system.

While winning these theoretical points, the Commission did not convince the General Court that a selective advantage had been granted. The Commission was deemed to have misapplied Irish domestic law through relying on an “exclusion approach”: according to this approach, since the companies where lacking employees outside of Ireland, it is considered that the relevant functions must have been performed in Ireland, and thus the profits should be taxed there. This approach was found in breach of Irish domestic law, under which no such presumption can be made: instead, one should prove that the control of assets is exercised in Ireland.[13]

Here, an important takeaway from a State aid perspective concerns the determination of the material content of the reference system. The Starbucks and Fiat cases seemed to support the theory of the Commission according to which the arm’s length principle “necessarily forms part of the Commission’s assessment of tax measures granted to group companies under Article 107(1) of the Treaty, independently of whether a Member State has incorporated this principle into its national legal system and in what form”.[14] The General Court has not supported this approach in the Apple case, contrary to what seemed to appear from its rulings in the Starbucks and Fiat cases. The General Court made clear that the reference system stems from domestic law, not from Article 107(1) of the TFEU alone. It is true that the General Court referred to the Forum 187 case,[15] but the interpretation that the Court does of this case points to a domestic obligation to treat similarly associated and independent enterprises, not one that would stem from State aid law. Although the arm’s length principle is – as in the Starbucks and Fiat cases – found to constitute a “tool”[16] or a ”benchmark”[17] to assess whether a favourable treatment has been granted, this assessment is conducted in the light of the domestic law.

The General Court considered that Irish domestic law actually contained the arm’s length principle, despite not explicitely referring to it in the domestic tax provisions. The arm’s length principle would mainly stem from case law, as illustrated by a case from 1984 in which the Irish High Court was deemed to make transfer pricing adjustments “equivalent to those proposed on the basis of the arm’s length principle, in particular in the OECD Transfer Pricing Guidelines”.[18] From a theoretical perspective I believe that this approach is correct in at least two respects: it enforces the principle of legality by relying on domestic law to determine the material content of the reference system, and the reference system is not only determined by the sole text of the law but it also takes into account the case law issued in the relevant country.[19] Consistently with the view according to which the reference system is made of the domestic law, the General Court quite logically held that there is no “freestanding obligation to apply the arm’s length principle arising from Article 107 TFEU obliging Member States to apply [the arm’s length principle] horizontally and in all areas of their national tax law”.[20] The General Court held a similar view in the Fiat case, considering that “the Commission does not, at this stage of the development of EU law, have the power autonomously to define the ‘normal’ taxation of an integrated undertaking, disregarding national tax rules”.[21]

The General Court also found that Irish tax law and the Authorized OECD Approach “overlap”.[22] This finding is rather weakly justified. The Authorized OECD Approach is to be applied in certain situations covered by a tax treaty, and this case was primarily one of domestic law. Further, one cannot assume that a country fully applies the Authorized OECD Approach by the simple observation that, for the purpose of the attribution of profits to permanent establishments, a country performs a functional analysis. There has historically been large differences in how countries deal with this issue, and in the OECD commentaries on article 7 different types of operations or payments have been treated differently: some are recognized, some are ignored, and when being recognized internal dealings were not always subject to a profit element until the adoption of the Authorized OECD Approach. The OECD commentaries have evolved, and the full Authorized OECD Approach is to be applied only to tax treaties concluded on the basis of the 2010 version of article 7 of the OECD Model Tax Convention. Certain parts of the 2008 OECD report on the attribution of profits to permanent establishments are in conflict with the previous commentary on article 7 and cannot be applied to tax treaties concluded on the basis of the old article 7.[23] If indeed the Authorized OECD Approach were applicable in this case, which the Commission did not really claim, the exclusion approach used by the Commission would not necessarily appear as fully unreasonable: since the Authorized OECD Approach relies largely on the substance existing in a permanent establishment and in the head office through the performance of significant people functions, the lack of substance outside of Ireland would hardly result in the attribution of profits to the head offices. But on the other hand, Ireland and Apple argued that there actually was some substance outside of Ireland through the management bodies of ASI and AOE,[24] something that the Commission did not deny.[25] Of course, it can also be argued that a large part of the profits belong to the US, but that is a matter of transfer pricing between the Irish companies and their US counterparts, thus potentially affecting the total profits of ASI and AOE, not the method to attribute profits between the head offices and the permanent establishments.

When it comes to the application of the arm’s length principle, the General Court ruled consistently with the Fiat and Starbucks cases. It correctly emphasised “the inaccuracies inherent in the methodology” used to obtain the approximation of an arm’s length price.[26] In other words, transfer pricing is not exact science. This implies that a margin of error should be accepted before considering that illegal State aid has been granted, and that prices or margins may be determined on the basis of ranges. This does not help determining what range is acceptable, and what is not, from a State aid perspective. As to the choice of a transfer pricing method, the General Court correctly found that “the mere non-observance of methodological requirements, in particular in connection with the OECD Transfer Pricing Guidelines, is not a sufficient ground for concluding that the calculated profit is not a reliable approximation of a market-based outcome”.[27] Indeed, what matters is the actual transfer price or profit margin, not the method that has been used to determine it, or even to test it. This finding is consistent with the view held in the Starbucks case.[28]

Eventually, despite a number of lacks and “methodological defects” identified in the way Ireland had granted the rulings, the General Court found that “the Commission did not conduct its analysis in such a way as to demonstrate that, as a result of that calculation, the tax actually paid by ASI and AOE on the basis of the contested tax rulings was less than that which should have been paid under the normal rules of taxation, had the contested tax rulings not been issued”.[29] The French version of the case employs an even harsher formulation, emphasising that the Commission “n’a pas mené à bout son analyse”, because it did not bring evidence of the existence of a selective advantage.

This is a key aspect of the case: should the Commission necessarily bear the whole burden of proof when it examines an arrangement between a Member State and an undertaking? Is this burden of proof reasonable when the Commission may lack information, and when it acts several years after the facts? Is the burden of proof expected from the General Court in line with the case law of the Court of Justice? Indeed, the Court of Justice has not always put such a high burden of proof when it comes to the evidence of the selectivity of a tax measure. The Gibraltar case provides an example where the tax system never came into effect, and thus did not effectively provide any advantage to any undertaking; the tax systemt was nevertheless found selective, on the basis of its potential effects.

Finally, the Apple case evidences the difficulty of challenging tax planning structures on the basis of the State aid rules. This calls for two observations. On the one hand, if the selectivity assessment is made in the light of the domestic law including the objective of the tax system, one may wonder if the determination of corporate residency could have been analysed from a State aid perspective: is a Member State granting State aid by accepting the non-residence of certain companies despite their presence and strong involvement in their territory? In the end, undertakings may escape worldwide taxation, and even the attribution of profits to permanent establishments on an arm’s length basis, by an operation that is more based on form (e.g. the lack of registration, or the lack of physical meetings of the board) than on substance. On the other hand, the case supports the efforts to meet the challenges arising from digitalization. At the same time, the State aid rules might constitute an obstacle to differentiated tax regimes depending on the level of digitalization of different companies. Should the Pillar one project be implemented in the Member States’ domestic laws, a portion of the residual profits of certain multinational enterprises would be reallocated to the market jurisdictions (i.e. the so-called “Amount A”). However, such a move is far from affecting all multinational enterprises in a similar manner.[30] As I emphasised in an earlier blog contribution,[31] a relief from tax in the country of origin would per definition not apply to independent enterprises or to domestic groups, and it would not affect all multinational enterprises in the same way. Therefore, it can be wondered if the envisaged changes do not amount to a selective advantage in the country of origin to the benefit of certain multinational enterprises, especially those that export goods or services, or have users abroad. Such a tax relief would hardly be motivated by the logic of the tax system of the country of origin, because the problem that is being tackled lies in other countries, where a multinational enterprise might be lacking a tax liability. In the country of destination, the new nexus rules would only target certain business models or certain multinational enterprises. This may result in a selective advantage to the benefit of undertakings that do not have a taxable presence, or for which little or no income will be allocated to the newly created taxable presence. There may also be a conflict with the fundamental freedoms, especially in the market jurisdiction where the tax base would increase, since the amendments to the arm’s length principle would not apply in domestic situations, and might not be justified by the need to prevent tax avoidance or safeguard a balanced allocation of the power to impose tax.[32] Consequently, the initiatives to amend transfer pricing rules in order to reallocate part of the corporate tax base to the countries of destination probably need to be completed by an analysis of the EU law implications of such reforms. The future judgment of the Court of Justice in the Fiat case[33], and perhaps also in the Apple case should the Commission appeal, will provide much needed imput on the reach of the State aid rules.

[1] T-778/16 and T-892/16, Ireland and Others v European Commission (hereafter “the Apple case”). This blog contribution builds partly on an article that the author will publish in the next issue of the journal Fiscalité Internationale.

[2] T-755/15 and T-759/15.

[3] T-760/15 and T-636/15.

[4] C(2016) 5605 final, case SA.38373.

[5] T-778/16 and T-892/16, para 105.

[6] T-778/16 and T-892/16, para 136.

[7] T-778/16 and T-892/16, para 147.

[8] C-105/18–C-113/18, para 60.

[9] T-778/16 and T-892/16, paras 150 and 152.

[10] Jérôme Monsenego, Selectivity in State Aid Law and the Methods for the Allocation of the Corporate Tax Base (Wolters Kluwer 2018), pp. 23-56.

[11] T-778/16 and T-892/16, para 155.

[12] Jérôme Monsenego, Selectivity in State Aid Law and the Methods for the Allocation of the Corporate Tax Base (Wolters Kluwer 2018), pp. 75-102.

[13] T-778/16 and T-892/16, para 181.

[14] Commission Notice on the notion of State aid as referred to in Article 107(1) of the Treaty on the Functioning of the European Union (2016/C 262/01), para 172.

[15] C-182/03 and C-217/03.

[16] T-778/16 and T-892/16, para 214.

[17] T-778/16 and T-892/16, para 215.

[18] T-778/16 and T-892/16, para 219.

[19] For a similar way of reasoning see Jérôme Monsenego, Selectivity in State Aid Law and the Methods for the Allocation of the Corporate Tax Base (Wolters Kluwer 2018), p. 53.

[20] T-778/16 and T-892/16, para 221.

[21] T-755/15 and T-759/15, para 112.

[22] T-778/16 and T-892/16, para 239.

[23] For example, the concept of significant people functions is not mentioned in the 2008 commentary on article 7. Also, the 2008 commentaries on articles 7 and 13 do not endorse the attribution of economic ownership of intangible assets to a permanent establishment.

[24] T-778/16 and T-892/16, para 254.

[25] T-778/16 and T-892/16, para 266.

[26] T-778/16 and T-892/16, para 216.

[27] T-778/16 and T-892/16, para 319.

[28] T-760/15 and T-636/15, paras 202, 209 and 211.

[29] T-778/16 and T-892/16, para 434.

[30] (see especially Annex B for a summary of the different tests to determine the liability to amount A).


[32] See e.g. C-382/16.

[33] C-885/19 P, Fiat Chrysler Finance Europe v Commission.


To make sure you do not miss out on regular updates from the Kluwer Competition Law Blog, please subscribe here.


Leave a Reply

Your email address will not be published. Required fields are marked *