The Tax Ruling Investigations: Initial Analysis

The Commission’s investigations into tax rulings by various EU Member States have attracted much publicity – understandably so given the size and importance of the companies involved and the general public concern about tax avoidance by multinational companies.

The Commission has now published three opening decisions under Article 108(2) of the Treaty on the Functioning of the EU: those relate to the Apple/Ireland OJ C 369, 17.10.2014, p.22, Amazon/Luxembourg OJ C 44, 6.2.2015, p.13, and Starbucks/Netherlands OJ C 460, 19.12.2014, p. 11 cases. This note attempts to identify some common themes arising out of those decisions.


Before looking at the opening decisions themselves, three important background points should be made.

First, although opening decisions set out the Commission’s concerns at that stage, it is important to remember that they are opening salvos rather than definitive statements of position. The Commission’s position can (and often does) change a lot between opening decision and final conclusion. And the opening decision may well make points which the Commission is not sure about but wants to test: there can be a “Devil’s advocate” flavour about some of the points made.

Second, an elementary State aid point often forgotten in the excitement is that these opening decisions do not – and could not – challenge low corporation or other business tax rates set generally by the Member States concerned. That point is worth making because both Luxembourg and Ireland are known for operating low general business tax rates. But a low general rate of corporation tax is not a matter for the State aid rules: State aid law is not concerned with differences in tax treatment between States but with differences of taxes within States: or, more formally, the frame of reference in looking at allegedly favourable tax treatment by a State is that State’s own tax system and not the tax system of another State.

Third, all these three cases are about the taxation of profits of group companies that supply goods and services to other group companies. So the amount of profit critically depends on the value attributed to sales made to other group companies. That means that all three cases are about transfer pricing – an issue of notorious difficulty and uncertainty even for experienced tax practitioners. The opening decisions all refer extensively to the OECD’s guidelines on transfer pricing, which emphasise the arm’s length principle (i.e. that prices of goods and services supplied between group companies should be priced by reference to what would have been charges as between arm’s length businesses). However, it is notoriously difficult to work out what the arm’s length price is of an intra-group service that may well have unique characteristics, and not have any traded equivalent. The OECD therefore accepts that in some cases one has to fall back on looking at the net profits of the entities involved, and its guidelines refer to the transactional net margin method or TNMM approach. Put as simply as possible, the TNMM approach aims to work out what profit margin – usually profit margin over costs – would have been made by the supplying company to an independent company and then takes the correct transfer price to be the price that achieves that margin of profit over costs.


Against that background, what are the Commission’s concerns in these three cases?

1. Duration of tax rulings

The three cases all concern advance tax rulings, by which the tax authorities agreed, in advance, a particular approach to transfer pricing. One concern identified is that those rulings lasted too long without revisions: in Apple a 1992 ruling lasted 15 years with no revision, and in Amazon a 2003 ruling was still operative in 2014. The Commission notes (and it is hard to argue with this) that both companies changed quite considerably over that period

2. No sufficient basis for method adopted

The Commission’s view is that use of the TNMM method cannot be justified merely on the basis that it is mentioned in the OECD guidelines. The Commission considers that use of the TNMM (or any other method) has to “approximate a market outcome in a correct way” and produce a result that can be regarded as “market conform[ing] remuneration” to a hypothetical independent supplier reflecting “normal conditions of competition“. It notes that in Starbucks it was felt necessary to adjust the TNMM by what it considers to have been questionable adjustments without a coherent basis – a sign, in its view, that TNMM was probably the wrong method. In Apple it is concerned that no justification for using the TNMM appeared in the papers, and in Amazon it is concerned both that the method adopted was not listed by the OECD (though it appeared to be a from of TNMM) and resulted in a very low rate of remuneration (only 4-6%).

  1. Understatement of cost base

In a method, such as the TNMM, that tries to calculate taxable profit by reference to a margin over cost, understatement of cost leads to understatement of profit. In that context, the Commission is concerned that that, in Starbucks, risk was not correctly allocated to the company (in particular, that risk was regarded as reduced because of the agreed tax treatment, which the Commission regards as circular) and costs to it were wrongly treated as pass-through costs on which private operator would not demand a profit element. In Apple the Commission states that it does not understand how the cost base on which profit was calculated increased only 10-20% over a three year period in which sales increased 412%.


  1. Failures by tax authorities to investigate

The Commission is critical of the tax authorities for not adequately examining the basis on which the case for the tax treatment at issue was based. In Starbucks, it states that the Dutch authorities did not look at the contracts underlying Starbucks’ assertion that it bore a limited costs risk. In Apple and Amazon it notes the absence of a transfer pricing report. And in Amazon it notes that the Luxembourg authorities granted the requested tax ruling only 11 days after the request: moreover, the request for a ruling provided insufficient detail of which group companies did what and failed to contain relevant information unearthed by House of Commons Public Accounts Committee in its investigation into alleged tax avoidance by Amazon.


  1. Royalty payments

 In Starbucks, the Commission was unimpressed by a calculation of royalty payments that was based on the desired accounting profit and on occasion resulted in negative payments. The Commission states that such a method could not relate to the economic value of the relevant intellectual property. In Amazon it was also unimpressed bya method that “expressed” the royalty payment as a percentage of revenues but which in fact based the calculation on desired profits.


  1. Evidence of improper considerations being taken into account

In Apple, the Commission quotes extensively from a memorandum which records Apple to have commenced its request for a tax ruling by referring to the number of employees it had in Cork, stating that it was reviewing its worldwide operations, and that it was “prepared to accept a profit of $30-40 million” while accepting that “there was no scientific basis for that figure“. The Commission is plainly concerned that the tax authorities reached an unprincipled and favourable tax deal in order to retain Apple in Cork.


The investigations are causing concerns in the tax world, for the following reasons.

  •  The Commission’s rhetoric on these cases casts itself as the protector of the ordinary taxpayer against deals made at their expense. However, protection of tax revenue is the task of national parliaments (in the UK, the Public Accounts Committee) and not obviously of the Commission. DG Comp is about competition, and it might be better if the competition justification for intervening was given more weight, at least in the Commission’s publicity.
  •  There is also concern that focusing on tax rulings will encourage companies not to seek rulings but simply to declare what tax they think due and rely on national authorities’ passive acceptance of their returns.
  •  Further, any tax ruling or settlement of a tax dispute involves compromise between different parties in areas where the tax position is not clear. Such compromise may be messy and unprincipled. Is any such compromise open to re-examination under the State aid rules on the basis that it does not “approximate a market outcome in a correct way“?
  • Finally, the consequences for the taxpayer if a tax ruling or settlement is declared to be a State aid are that the State has to recover the underpaid: but the obligation of recovery under the State aid rules is far more onerous than any obligation in national tax rules to pay underdeclared tax (the capping period is usually well under the State aid limitation period of 10 years, and is usually much less easily interrupted).

The Commission will need to be sensitive to the concern that a wide application of the State aid rules to (inherently contestable) tax settlements will cause significant difficulties. However, the opening decisions do suggest that the Commission is basing its position on more than a mere disagreement with the tax rulings but on a number of what appear at first sight to be serious doubts about whether those rulings could be sustainable as a reasonable settlement of tax liabilities. It is only when the final decisions are published that we shall see the extent to which there are particular features of these cases that differentiate them from the run of cases where a tax ruling is given that could, possibly, have been more unfavourable to the taxpayer.


This entry was posted in Commission decision. Bookmark the permalink.