Those connected to the transfer pricing world have already learned how the iCase turned out. (here our LinkedIN post).

Four years ago (here, TPS commentary, August 2016), the European Commission imposed Ireland to recover 13 billion euros plus over a billion interests from the Apple Group, as illegal State aid granted through selective tax breaks.

At that time, the Commission seemed very convincing in having a case (with evidence in documents and minutes, like that Cost Share Agreement (here, commentary based on the Commission decision, March 2017 - RO).

Meanwhile, the evidence proved rather assumptions. And the EU General Court finds that “the Commission was wrong” (here, press release, and here, the judgement), after examining the pleas in law raised by Ireland and ASI and AOE (Apple’s divisions involved), after amending assumptions and speculations (i), but most of all, after carefully reading the “authorized approach in transfer pricing”, the OECD’  Guidelines.

This is the perspective we would like to emphasize today. 

And the winner is (for now) … 

As expected in a case involving such heavy players, with enormous economic and political implications, the 15 July decision is and will be open to interpretation. Yes, Apple won, but the giant already accepts that things are changed, in Europe but also at home. Yes, Ireland won, but the case was already seen as a lose-lose situation (on one side, 13-14 billion would have made a difference in times like this, on the other, Dublin know they cannot protect anymore a low tax regime that fantastically worked in attracting new technology stars). And yes, the EU Executive lost another technical battle (ii), but not the political war - the new taxation normal is already in place, the direct tax harmonization project is going on, and … the  Commissioner for Competition from 2016 is today the Executive Vice-President, setting the strategic direction of the political priority "Europe Fit for the Digital Age". And taxation has for sure allocated a pretty special place in this fitting.

Plus - do not forget Poland in this tableau! What about Poland? Good question. While it’s not unusual to see member states which support each other in their legal battles with the Commission on state aid hot issues (iii), this time we surprisingly have a member state sending its cavalry to "contends, in essence, that the Court should dismiss the action in Case T‑778/16, in accordance with the form of order sought by the Commission”. What are Polish stakes in this World Tax Fight? Until now, they get an order to bear their own costs in this judgement.

But, beyond all of this, we see this judgement rather a triumph of the Transfer Pricing Guidelines, that extraordinary platform, both stable and flexible in its structure, "seeking to achieve the balance between the interest of taxpayers and tax administrators in a way that is fair to all parties”, as the July 2010 Edition states. By the way, Happy Anniversary, dear partner!

Let’s recognize that into an EU having special relationship with the arm’s length principle (iv), the Authorized OECD Approach honored again its duty.

From this perspective, we proposed you a briefly review on this landmark case in a way that could be useful even for the not so special businesses.

Forget, for the moment, about Apple brand (considerate it a kind of fruit company, as Forrest Gump would say), forget about those 13 billion, forget about the political stakes. Just call it quite an usual transfer pricing case requiring, according to the Authorised OECD Approach, a more detailed  analysis of the functions actually performed within the permanent establishment, only that the tax administration is presuming that the functions had been performed by the permanent establishment when those functions could not be allocated to the head office of the company itself. (based on recital 242, the judgement of 15 July 2020)

So, please make you comfortable and pay attention:

  1. Not in dispute

 In the present instance, it is not in dispute that (recital 173):

–        ASI and AOE are companies (within the Apple Group) that are incorporated in Ireland, but which are not considered to be tax resident in Ireland.

–        The Irish Taxes Consolidation Act contains provisions that apply specifically to non-resident companies under which, where a non-resident company carries on a trade in Ireland through a branch, that company is to be taxed, inter alia, on all of its trading income arising directly or indirectly from the branch;

–        the non-resident companies ASI and AOE carried on a trade in Ireland through their respective branches.

  1. In dispute

The dispute in the present instance centres around the taxation of companies that are not tax resident in Ireland and which carry on a trade in that State through their Irish branches. The issue therefore lies in determining what profits must be allocated to those branches for corporation tax purposes as part of ‘normal’ taxation, taking into account the normal rules of taxation applicable in the present instance (204).

  1. A simple fact

In its primary line of reasoning the Commission considered, in essence, that the profits of ASI and AOE relating to the Apple Group’s IP (which, according to the Commission’s line of argument, represented a very significant part of the total profit of those two companies) had to be allocated to the Irish branches in so far as ASI and AOE had no employees capable of managing that IP outside those branches, without, however, establishing that the Irish branches had performed those management functions. (243)

       4. … it’s not an evidence

In order to substantiate its assessment, the Commission relies to the cost-sharing agreement concerning all of the relevant functions relating to intangible property subject to the agreement in question and the related risks. Each of those functions and risks has an ‘x’ next to it in the columns for, respectively, Apple Inc. (identified as ‘Apple’) and ASI and AOE (identified collectively as ‘International Participant’), except for IP Registration and Defence, which is solely associated with Apple Inc. (260)

      5. A simple list …

As Apple Inc. claimed in the administrative procedure and as ASI and AOE have argued before the Court, it is apparent from the exhibit in question that it lists the functions that the parties to the cost-sharing agreement were authorised to perform and the associated risks that they might have been required to assume. However, the Commission provided no evidence to show that ASI or AOE, let alone their Irish branches, had actually performed any of those functions. (263)

  1. ... doesn’t necessary make things “clear”!

In addition, with regard to those functions and risks, the Commission contends that it is ‘clear’ that, with no employees outside their Irish branches, ASI and AOE would not have been able to monitor such risks. However, the Commission provides no evidence that demonstrates that the staff of the branches in question actually performed those functions and managed those risks. (264)

  1. It is needed a detailed analysis of the type of activities…

After analysing the functions and activities performed by ASI’s Irish branch which the Commission had identified as justifying allocating the Apple Group IP licences held by ASI to that branch, it must be concluded that these are support activities for implementing policies and strategies designed and adopted outside of that branch, in particular with regard to the research, development and marketing of Apple-branded products. (283)

  1. … to justify the need of valuable or unique assets allocation

In those circumstances, it must be concluded, as was argued by Ireland and ASI and AOE, that the Commission erred when it considered that the functions and activities performed by ASI’s Irish branch justified allocating the Apple Group’s IP licences and the income arising from those licences to that branch. (284)

  1. Test only where there are reliable data!

(In the context of TNMM), the OECD Transfer Pricing Guidelines do not state which party to the transaction must be chosen, but recommend choosing the undertaking for which reliable data regarding the most closely comparable transactions can be found. It is then specified that this often means choosing the associated undertaking which is the least complex of the undertakings concerned by the transaction and which does not have any valuable intangibles or unique assets. It follows that those guidelines do not necessary require that the least complex entity be chosen, but that they simply advise choosing the entity for which the greatest amount of reliable data exists (335).

  1. Wondering which is the proper profit level indicator?

It follows from paragraph 2.87 of TP Guidelines that the profit level indicator [sales or operating costs] must be focused on the value of the functions of the tested party, taking account of its assets and its risks. Therefore, according to those guidelines, the choice of profit level indicator is not fixed for any type of function, provided that that indicator reflects the value of the function in question. (357)

  1. Again, a detailed analysis is the key

Such a fact [ASI’s sales increased exponentially during the reference period whereas the operating costs of its Irish branch remained stable] is not sufficient, in itself, to call in question the choice of the operating costs as the profit level indicator. Indeed, the Commission established its line of reasoning without indicating the reason why an increase in ASI’s sales would have necessarily involved an increase in the profits to be allocated to its Irish branch. (389)

  1. It’s not enough to claim “methodological errors”

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, consisting in the choice of the Irish branches as tested parties, the choice of the operating costs as the profit level indicator, and the levels of return accepted by the contested tax rulings  had led to a reduction in ASI and AOE’s chargeable profits in Ireland. Accordingly, it did not succeed in demonstrating that those rulings had granted those companies an advantage. (480)

  1. Reasonable and reliable! Rest my case.

Analysing transfer pricing is not an exact science and it is not possible to check for exact results as to what is considered to be an arm’s length level. In that regard, it is necessary to recall paragraph 1.13 of the OECD Transfer Pricing Guidelines, which states that the objective of determining transfer pricing is ‘to find a reasonable estimate of an arm’s length outcome based on reliable information’ and that ‘transfer pricing is not an exact science but does require the exercise of judgment on the part of both the tax administration and taxpayer’. (455)

    + 1 allusion to the (foreseen) new EU stage of development

The fact remains that, at the current stage of development of EU law, the Commission does not have the power independently to determine what constitutes the ‘normal’ taxation of an integrated undertaking while disregarding the national rules of taxation. (223)

(i) Just two examples:
The Commission argues, in essence, that ASI’s Irish branch performed all of those functions and assumed all of those risks relating to the Apple Group’s activities outside North and South America without providing evidence of the actual performance of those functions or assumption of those risks by the branch in question. Given the extent of the Apple Group’s activities outside North and South America, which represent approximately 60% of the group’s turnover, the Commission’s assertion in that regard is not reasonable” (267)
In any event, the mere allusion, during the exchanges between the Irish tax authorities and the Apple Group preceding the 1991 tax ruling, to the fact that the Apple Group was one of the largest employers in the region where the Irish branches of ASI and AOE were established does not prove that ASI and AOE’s chargeable profits were determined on the basis of employment-related issues (…) Thus, in the absence of other evidence, the Commission cannot argue that the tax ruling in question was issued as consideration for the potential creation of jobs in the region” (440+441).
 (ii) In 2019, the EU Court annulled Commission Decision 2017/502 on State aid implemented by the Netherlands to Starbucks, and Commission Decision 2016/1699 on the excess profit exemption, State aid scheme implemented by Belgium; In the same time, the Court uphold Commission Decision 2016/2326 on State aid which Luxembourg granted to Fiat.  
 (iii) For example, Ireland supported Belgium (Cases T‑131/16), Luxembourg (Cases T‑755/15), and Netherlands (T‑760/15), and received support from Luxembourg (T‑778/16)
(iv) “The arm’s length principle, as described by the Commission, is thus a tool enabling the Commission to make that determination in the exercise of its powers under Article 107(1) TFEU. Moreover, the Commission rightly noted, in recital 256 of the contested decision, that the arm’s length principle served as a ‘benchmark’ for verifying whether the chargeable profit of a branch of a non-resident company was determined, for the purposes of corporation tax, in a manner that ensured that non-resident companies operating through a branch in Ireland were not granted favourable treatment as compared with resident stand-alone companies whose chargeable profits reflected prices negotiated at arm’s length on the market”. (214+215)