A Bad Apple Ruling

13 September 2024       

Richard Collier (Associate Fellow, Oxford University Centre for Business Taxation)

The judgement of the European Court of Justice (ECJ) in the Apple Ireland case was released on 10 September 2024. The case is concerned with the question of whether two tax rulings (from 1991 and 2007) obtained by Apple from the Irish tax authorities involve the receipt of advantages in the form of state aid from Ireland. 

The judgement raises several fundamental – and highly problematic - legal issues which go well beyond what might be discussed within the confines of a single blog. However, one of the core issues in the case is a question that is of critical importance to the taxation of branches (technically, “permanent establishments” or “PEs”) under international tax principles and it is this point that is pursued here.

The facts concern two Irish-incorporated companies that were non-resident in Ireland for tax purposes, meaning they were to be treated as branches of non-resident companies for Irish tax purposes. The Irish branches carried on certain relatively routine
procurement, sales and distribution functions and the manufacture and assembly of computer, etc. products in Ireland. Under the rulings, the branches were allocated a (taxable) return commensurate with this activity. 

The disputed profits in this case were generated from certain IP licences. Pursuant to an international tax planning arrangement, those profits were booked in the two Irish companies, but they were created/ generated by the acts of employees of other non-Irish tax paying Apple group companies, notably of Apple Inc., the top company in the group (a significant number of directors of the boards of the two Irish companies were also directors of Apple Inc and they worked in Cupertino (US)). Nonetheless, those profits were not taxed in the US under US tax rules. The question at issue was whether the profits from the IP licences should for tax purposes be attributed to, and so taxed in the hands of, the Irish branches. The Irish tax authorities did not treat those profits as taxable in Ireland because they were not generated by the functions and activities of the Irish branches.  The ECJ decision on the other hand sides with the interpretation adopted by the European Commission which holds that the profits booked in each of the Irish entities should be allocated by reference only to the activity in each of the legal entities of which they are part (i.e., taking account of the activity in the head office and any branches), without regard to activity in any other legal entities in the group (and specifically without regard to the activities outside those two Irish companies that generated the profits). Since there was minimal activity at the respective head offices of the two Irish companies, the ECJ interpretation holds that the profits should be taxed in the Irish branches where there was the routine distribution and assembly activity referred to above. The effect of the approach adopted by the ECJ is that the profits are attributed by default to the Irish branches.

The European Commission, as supported by the ECJ, analyse the question of what profits are attributable to the Irish branches by reference to the landmark OECD guidance on how profits are allocated to PEs, namely the guidance in the OECD’s 2010 Report on the Attribution of Profits to Permanent Establishments. That Report overhauled the guidance in this area and introduced a new approach to the branch profit attribution exercise, commonly referred to as the “authorised OECD approach” or, more usually as simply the “AOA”. The AOA is intended to deliver PE profit attribution results that are aligned with the arm’s length principle. 

Given that the AOA was introduced some two decades after the first disputed ruling in the Apple case (and 3 years after the second ruling was given), the use of the AOA in this case raises significant difficulties. There are also some major questions on how the use of the OECD guidance can in any event be squared with ECJ’s own relatively recent decision in the Fiat case, which determined that the analysis in these state aid cases should be restricted to the domestic law system of the state concerned. Unfortunately, space considerations here mean that all those issues will need to be addressed in a separate blog.  

However, even if it were considered right to analyse the technical position by reference to the AOA, it cannot be concluded – as the decision does - that activities of other entities in generating the profits should not be taken account of in the analysis of what profits belong to the branches. 

This follows from the core requirement of the provisions in tax treaties (specifically provisions that correspond with Article 7 (2) of the OECD Model Tax Treaty) which the AOA was designed to clarify. This requires a focus on the activities in the branch/ PE in answering the question of, broadly, what profits a hypothetical distinct and separate enterprise would be expected to derive if it carried on the same activities (and under the same conditions) as those actually carried on in the branch. The more detailed guidance in the AOA adopts the same approach based on focusing on the actual functions and activities in the branch to determine what profit is properly attributable to the branch. The OECD guidance also includes instances where the contribution of separate entities is taken into account. It is therefore clear that the required OECD approach starts by analysing the activities in the branch to determine, by reference to those activities, what profits should be taxed in the hands of the branch. Where appropriate, that approach takes account of all other activity in the entity and in the group that contributed to the earning of profits. This is the opposite of what the ECJ judgement does – it starts not with the activity in the branch but with the profits in the entity.  It then allocates those profits taking account only of the activities within the legal entity (so ignoring the contribution of other group entities). The result leads to a non-arm’s length attribution of profits to the branches. This follows from the use of the “default” allocation of profits which are derived from activities that are outside both branches. 

The fact that the ECJ approach is misconceived is further illustrated by the point that, as the AOA principles themselves set out, “Profits may ... be attributed to a permanent establishment even though the enterprise as a whole has never made profits. Conversely, Article 7 may result in no profits being attributed to a permanent establishment even though the enterprise as a whole has made profits.” Again, this makes clear that the focus of the exercise should be on determining the attributable profits by reference to the activities in the branch/ PE, not by reference to the profits booked in the legal entity. 

If the contrary ECJ position were correct, it would lead to some absurd consequences. For example, if the activities conducted in Apple Inc that generated the profits, but which were ignored by the ECJ, were characterised as conducted in the US through a dependent agent PE of the Irish companies, then all contested profits would on the ECJ/ EC theory presumably be attributable to that activity (because they would now be located in a (US) PE of the Irish companies, not in a separate group company). This would mean those profits would be no longer taxable in Ireland, notwithstanding no change whatsoever in the activity in Ireland. This would mean that whether the relevant profit is taxed in Ireland depends on whether a tax inspector in another state (here, the US) makes such a PE determination. This is contrary to the historical framing of the branch profits approach, and it is hard to see any basis on which it makes sense or delivers a workable system of PE profit attribution.

Notwithstanding all the issues discussed above, it might possibly be argued that the ECJ decision in the Apple Ireland case serves as an appropriate response to those MNEs seeking to game the international tax rules through aggressive international tax avoidance structures. However, one of the enduring lessons of the 2015 BEPS reforms – and this is especially relevant in relation to the income allocation rules – is that changes to the rules motivated by an overriding anti-avoidance goal can have profound and pervasive impacts on the routine operation of the tax system, creating all sorts of difficulties to the daily application of the rules that were never considered when the original avoidance mischief was being targeted. The concern is that just this result will be one of the depressing legacies of the ECJ decision.

In summary, the short point made here is that the use in this decision of the AOA is highly questionable and that its interpretation by the Court is in any event misconceived. Given the numbers involved in the case, the result is to deliver a colossal failure in complying with the requirements of the arm’s length principle.

As a matter of transparency, it is noted that the author provided opinions in this case which reflected the views set out in this blog.