Menu
Tax Notes logo

How Did McDonald’s Get Off the EU State Aid Hook?

Posted on Feb. 4, 2019
Barry Larking
Barry Larking

Barry Larking is an international tax consultant helping organizations manage technical and policy developments worldwide (www. barrylarking.com).

In this article, the author discusses the European Commission’s state aid case against McDonald’s.

Big Mac fans with a healthy interest in tax matters no doubt breathed a big sigh of relief in September when the European Commission announced it would not be pursuing its claim for alleged infringement of EU state aid rules against the McDonald’s group. Although this is a state aid case, it raises some interesting issues on tax treaty interpretation, as well as on the state aid process itself. It also addresses a state aid defense we might see more of in the future. The tax treaty issues have been intensely debated both in the context of the McDonald’s case and generally,1 but they remain difficult to grasp — as evidenced by the commission’s attempts to deal with them in the present case. This article may not resolve the debate, but it will hopefully shed more light on the issues.

Background

The case is one of a string of challenges brought by the European Commission against several EU member states alleging that the states had breached EU state aid rules by granting selective tax benefits to specific taxpayers.2 The European Commission is responsible for ensuring compliance with the state aid rules. If it has serious doubts about a measure’s compatibility with these rules, it opens a formal investigation setting out its analysis in a decision, giving interested parties the opportunity to submit comments. This is generally followed up by a final decision (which in the McDonald’s case was announced in September and released in December3). The commission’s decisions may be reviewed by the Court of Justice of the European Union. Although state aid procedures are initiated against the EU state that is alleged to have granted the aid, a finding of illegal state aid results in the benefit represented by the aid having to be recovered by that state from the taxpayer(s) in question. McDonald’s clearly had something to lose.

State Aid Rules

Restrictions on state aid go back to the very beginnings of the EU, and act essentially as a tool to regulate competition between member states. Tax is not directly mentioned in the primary state aid rules,4 but it was highlighted by the commission in the late 1990s as a potential source of state aid5 and became a focus (particularly tax rulings) for state aid investigations by the commission in 2014. The key aspect of the state aid rules in the McDonald’s case is the “selectivity” criterion, a complex concept that has been the subject of many state aid cases. In the McDonald’s case, selectivity means that the taxpayer should have benefited from an advantage (in that case, a tax reduction) not available to other businesses that were in similar circumstances and otherwise subject to the same tax system. The tax system in this case was taken by the commission to be the Luxembourg corporate income tax system, including its double tax treaties.

Why Was the McDonald’s Ruling Challenged?

In 2009 the U.S. parent of the McDonald’s group — the McDonald’s Corporation — transferred franchising rights to a Luxembourg group company, McDonald’s Europe Franchising. That company allocated those rights to a branch in the United States to which franchisees in various countries outside the United States paid royalties for the right to use the McDonald’s brand. The Luxembourg tax authorities issued a ruling confirming that those royalties would not be exempt from Luxembourg corporate income tax. The original ruling was conditioned on the royalties being “declared and subject to tax” in the United States, but this condition was dropped in a revised ruling issued a short time later.

The commission’s initial decision — Commission Decision C(2015) 8343 final (Dec. 3, 2015) — concluded that the revised ruling constituted state aid. The Luxembourg tax authorities had granted a selective advantage to McDonald’s by confirming in that ruling that the U.S. branch income would be exempt from tax. The essence of the commission’s argument was that, absent the Luxembourg-U.S. tax treaty, Luxembourg would have taxed that income under its domestic rules, and there was nothing in the U.S. treaty that required Luxembourg to exempt the income from taxation. The advantage was therefore selective and constituted state aid.

This argument seems to have been based on what the commission evidently thought was the plain meaning of the treaty text. The exemption provision in article 25 (following the wording in article 23A of the OECD model convention) provided that the residence state (Luxembourg) was only required to exempt income that “in accordance with the provisions of” the treaty “may be taxed” in the United States. If that condition was not satisfied, Luxembourg would be free to tax. In paragraph 89 of the December 3, 2015, decision, the commission said this condition was not satisfied because there was “no possibility that those profits ‘may be taxed’ in the U.S.”

The initial decision does not make entirely clear whether the “no possibility” to tax the branch in the United States resulted solely from U.S. domestic law, or also from the U.S. view of its treaty article 7 right to tax the branch. It is notable that the word “may” has at least two different meanings: in a permissive sense, and in the sense of something being possible. Why this is important will become apparent below. At any rate, there are indications (for example, the reference in paragraph 89 to the tax adviser’s explanation of the U.S. domestic law concept of “trade or business”) that the commission’s view was solely based on its understanding of U.S. domestic law. Apart from repeated references to a U.S. “PE,” which suggests the treaty term, there is little to suggest that the commission was referring to the U.S. treaty view.6 In its final decision,7 the commission clarifies its position, suggesting that there was also no U.S. right to tax the branch under the treaty.8

The commission backed up its initial decision with a reference to paragraph 32.6 of the OECD commentary to article 23A and B of the model. This paragraph deals with the scenario in which the source state, based on its domestic law, takes the view that it does not have the right to tax income, while the residence state takes the opposite view and so exempts the income, resulting in double nontaxation.

Paragraph 32.6 does not actually mention domestic tax, but a related passage that deals with double taxation does, and so does a later paragraph that paraphrases paragraph 32.6.9 It is therefore clear that this passage refers only to domestic law differences. However, it is not referring simply to differences between the domestic laws of the treaty states that lead to double nontaxation (this is dealt with separately in paragraphs 34 and 56.2 of the commentary), but to different applications of the treaty resulting from these differences.10

The most common situation in which this might occur is when a domestic law concept is used to interpret an undefined treaty term under the equivalent of article 3(2) of the OECD model. In such situations, the OECD’s solution is for the residence state to abandon its own view and follow that of the source state. This approach would therefore only be relevant to the McDonald’s case if, because of its domestic law, the United States lacked the treaty right to tax — for example, when domestic law was used to interpret undefined treaty terms. Whether the United States could tax the branch under its domestic law was not directly relevant; what was relevant was whether U.S. domestic law could be used to interpret the treaty such that the United States did not have the right to tax. Luxembourg would then have been right to follow the U.S. approach and would not have been required to exempt the income. It is remarkable that the commission did not explicitly address this.11

Says Who?

The relevance of the OECD commentary to the model convention and the 1999 partnership report from which many of the relevant parts of the commentary are derived may be, and has been, questioned in many respects. Notably, it has been questioned regarding its status as a treaty interpretation tool in general, and specifically, given the timing of the report and relevant passages from the commentary in relation to the Luxembourg-U.S. treaty.12

However, if, as some argue, the OECD’s above-mentioned approach simply reflects the proper interpretation of the treaty,13 these objections lose much of their force. That may also be why the initial decision first stated the commission’s own position and then backed it up with what the commission understood to be the OECD view. In fact, those objections would otherwise be particularly pertinent in light of a 2008 change to that part of the OECD commentary the commission had used to support its initial decision — that is, long after the conclusion of the Luxembourg-U.S. treaty. Before the change, the OECD’s approach also required that the source state (in this case, the U.S.) would have taxed the income if the treaty had permitted this.14 If the United States was indeed unable to tax the branch under its domestic law, this alone would have meant the commission could not rely on the OECD approach.

More Support for the Commission

It appears that another commission argument supporting its initial decision is that, in view of the wording of article 25, Luxembourg should only have exempted the income if double taxation would otherwise have resulted. The wording in question provides that “[i]n Luxembourg double taxation shall be eliminated as follows.”15 Again, assuming that the United States could not tax the branch under its domestic law, there was no risk of double taxation, so Luxembourg did not need to exempt it. Although the OECD does not seem to regard double taxation as a condition for residence state exemption in all cases,16 it has been persuasively argued that the specific wording of the Luxembourg-U.S. treaty should be followed in this case.17 In any event, the commission did not take this point further.

Why Did the Commission Change Its Mind?

In its final decision,18 the commission reversed its initial decision and concluded that there had been no grant of state aid by Luxembourg. Luxembourg was under a treaty obligation to exempt the U.S. branch. The commission reached this conclusion based on two (apparently alternate) grounds. The first concerned the treaty interpretation issue discussed above, while the second was a state aid issue.

The Treaty Issue

While the initial decision only referred to the OECD’s commentary as confirmation of the commission’s analysis, the reasoning of the final decision is based directly on the commentary — but, crucially, a different part.19

The commentary references in both the initial and final decisions assume that the two treaty states take different views regarding the source state’s right under the treaty to tax the income. The essential difference appears to be whether those different views are caused by variations in their respective domestic laws.

If the different views are caused by domestic laws (the OECD’s first approach, referred to in the initial decision), and the source state thinks it cannot tax, the residence state is free to do so. Likewise, if the source state thinks it can tax, the residence state must exempt.

If the different views are caused by a disagreement that is not caused by domestic laws (the OECD’s second approach, referred to in the final decision), the residence state should follow its own interpretation. This means it would have to exempt the income if, based on its own interpretation, it thinks the source state can tax.

Unfortunately, the OECD itself does not make this explicit. It contrasts divergences based on “different interpretations of the provisions of the Convention” with divergences based on “different provisions of domestic law.”20 This is confusing since different treaty interpretations can be based on different provisions of domestic law — for example, when treaty states apply domestic law to interpret undefined terms in accordance with provisions equivalent to article 3(2) of the model convention. However, the idea would seem to be to contrast that situation with different “autonomous” interpretations of treaty terms — that is, where a single “treaty” meaning is intended, without reference to domestic law. This analysis has implicit support from the commentary and commentators.21

While this analysis may seem a practical solution, it is not without its difficulties.22 For example, what if the source state applies its domestic law to interpret an undefined term under the equivalent of article 3(2), but the residence state takes the position that it should not have done so because the context “otherwise requires”?23 Or what if the residence state considers that the source state has applied its domestic law incorrectly?24 At any rate, the commission’s final decision was that the OECD’s second approach should be applied, and that Luxembourg was correct in following its own interpretation and exempting the income.

The commission’s final decision may be questioned for many reasons. In the first place, it applies the OECD’s second approach to a problem caused by the application of different domestic laws: “different interpretations . . . under Luxembourg and U.S. tax law have . . . led Luxembourg and the U.S. to . . . interpret . . . Article 5 of the . . . treaty, differently.”25 Attaching significance to differences in the two states’ respective tax laws contradicts not only what the OECD and other commentators appear to think this approach involves, but also the commission’s own understanding.26

Another questionable aspect of the commission’s analysis is that it considers that the OECD’s first approach — the one based on domestic law — only applies when the two states apply different provisions of the treaty. Accordingly, since it said that “different interpretations . . . have not led Luxembourg and the U.S. to apply different provisions of the double taxation treaty, but to interpret the same provision, i.e. Article 5 of the . . . treaty, differently” (emphasis added),27 it concluded that the first approach did not apply. Luxembourg was therefore bound to follow its own interpretation in accordance with the OECD’s second approach and exempt the U.S. branch income.

The commission seems to believe that only the first approach, based on domestic law, covers the situation in which the two states end up applying different provisions of the treaty. The OECD’s commentary illustrates this with an example involving the transfer of a foreign partnership interest that one state treats as the disposal of a share in a company, taxable only in the residence state under article 13(5) of the model convention, while the other state treats it as the disposal of the partnership’s business assets, taxable in the source state if the assets are attributable to a PE under article 13(2) of the model convention.28 However, the example illustrating the OECD’s second approach (the autonomous interpretation) also results in each state applying different parts of article 13. This does not therefore seem to be a valid distinction between the two approaches.

Looking at the McDonald’s case, the final decision suggests that a difference in domestic laws led Luxembourg and the United States to have different views regarding whether the United States had the right under the treaty to tax the branch income, in particular with reference to the term “business” in articles 5 and 7.29 In other words, applying the first part of the business profits article of the treaty, the United States considered that only Luxembourg had a right to tax, while Luxembourg considered that under the second part of that article the United States could tax the branch as a PE.30

Whether, as the commission suggests, this application of different provisions of the treaty is a requirement of the OECD’s first approach does not seem material: The United States as the source state ends up taking a different, domestic law-based view of its right to tax income under the treaty than that taken by the state of residence, Luxembourg. It is hard to see why, as a matter of principle, this should not be covered by the OECD’s first approach.31

The State Aid Issue

The commission finishes its analysis with a single paragraph stating that “the contested tax rulings do not depart from tax rulings obtained by other taxpayers in line with this interpretation and application of the double taxation treaty by Luxembourg.”32 The commission appears to be agreeing with comments made by McDonald’s on the initial decision that the benefit of the Luxembourg exemption was not selective because other taxpayers received a similar benefit (the benefit was of “general application”), and therefore the Luxembourg tax treatment did not constitute state aid. The reference to the other rulings being in line with this interpretation and application of the tax treaty indicates that the commission is referring to its analysis in the final decision based on the OECD’s second approach of an “autonomous” treaty interpretation. This is odd, since according to that analysis there is no taxpayer benefit, because Luxembourg was required under the tax treaty to grant an exemption. However, there seems no reason in principle why this “general application” argument could not be used to support the “no state aid” conclusion, in case the commission’s tax treaty analysis proved incorrect.

This raises several questions:

  • Why did the commission not elaborate on the relevance of this statement, which appears to have been added almost as an afterthought?

  • Why did the commission base this on evidence put forward by McDonald’s rather than on its own or an independent investigation?33

  • Why did the commission not investigate the reasons for the exemptions under the other rulings to see if they were indeed based on a similar difference of interpretation to that in the final decision in the McDonald’s case?34

  • Why didn’t the commission explain why it accepted this “general application” argument here, but not in other state aid cases (for example, in the Engie decision, C(2018) 3839 final, paragraph 229)?

Is the Commission the Right Forum?

One is left with the feeling that the European Commission may not have been the best forum in which to adjudicate these issues, or that it should at least have better equipped itself for dealing with them by, for example, obtaining independent expert evidence. There is no indication that the commission at any stage carried out an independent investigation into U.S. domestic tax law or treaty practice. The commission’s assumptions on these issues appear to have been based almost wholly on statements made by tax advisers for McDonald’s, McDonald’s itself, or the Luxembourg tax administration — not the most impartial of parties.35 The same applies to conclusions drawn on rulings granted to other taxpayers, which appear to have been based simply on assertions made by the taxpayer.36

A cynic might say the commission realized it had bitten off more than it could chew regarding the technical arguments and therefore decided to make a tactful retreat. As the commission points out in its final decision, “the burden of proof of the existence of a selective advantage lies with the Commission,”37 so even if it did not entirely abandon its position, the commission may have felt that the complexity and uncertainty did not make it worth pursuing the case, particularly in light of the sensitivity of state aid investigations involving U.S. multinationals.38

Meanwhile

Meanwhile, the tax world is moving on. The 2017 update to the OECD model convention, paragraph 12, explicitly states that treaties are not intended to lead to double nontaxation. The OECD’s multilateral instrument takes this further by including specific wording that should in the future prevent the final McDonald’s result.39 The commission’s initial decision seems to have wrong-footed both Luxembourg and McDonald’s: Luxembourg in June 2018 announced plans to amend its domestic law to prevent this kind of double nontaxation from arising in the future,40 and McDonald’s restructured by moving its Luxembourg operations to the United Kingdom.41 In 2016 the U.S. Treasury Department reported ongoing negotiations to amend the Luxembourg treaty in line with planned changes to Luxembourg domestic law that would prevent double nontaxation of this sort. It appears the treaty negotiations are still underway but, on this point, may anyway have been overtaken by the above-mentioned Luxembourg law change.42 Finally, there is the EU’s second antiavoidance directive, Directive 2017/952 (2017). The directive specifically targets this kind of “disregarded PE” — although the relevant parts are not due to kick in until 2020, and, according to (new) article 9(5), even then only if the directive doesn’t conflict with an existing non-EU country tax treaty.

FOOTNOTES

1 For more recent comments, see Fadi Shaheen, “Tax Treaty Aspects of the McDonald’s State Aid Investigation,” Tax Notes Int’l, Apr. 24, 2017, p. 331; Oliver Hoor and Keith O’Donnell, “McDonald’s State Aid Investigation: What the European Commission Got Wrong,” Tax Notes Int’l, Sept. 12, 2016, p. 975; Michael Lang, “2008 OECD Model: Conflicts of Qualification and Double Non-Taxation,” Bulletin for International Taxation 204 (June 2009); Jacques Sasseville, “Klaus Vogel Lecture – Tax Treaties and Schrödinger’s Cat,” Bulletin for International Taxation 45 (Feb. 2009); Klaus Vogel, “Conflicts of Qualification: The Discussion Is Not Finished,” Bulletin for International Taxation 41 (Feb. 2003); and Alexander Rust, “The New Approach to Qualification Conflicts Has Its Limits,” Bulletin for International Taxation 43 (Feb. 2003).

2 For a recent overview, see European Commission, IP/19/322 (2019).

3 Commission Decision C(2018) 6076 final (Sept. 19, 2018).

4 Article 107(1) of the Treaty on the Functioning of the European Union.

5 “Commission Notice on the Application of the State Aid Rules to Measures Relating to Direct Business Taxation,” 98/C 384/03 (1998). Now updated, see “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 (2016).

6 The commission confuses matters by combining references to a U.S. permanent establishment with domestic law references, such as a “PE subject to tax in the United States” (para. 88) and “a PE for U.S. tax purposes” (para. 91).

7 Supra note 3, at para. 121.

8  It may well be that this was simply based on the tax advisor’s statements in the revised ruling, noted at para. 41 of the final decision. For a technical explanation as to why the commission might have been right on this, see Shaheen, supra note 1, at 340-342.

9 Commentary to article 23 A and B of the OECD model convention, paras. 32.3 and 56.3.

10  See Shaheen, supra note 1, at 339.

11 It may be that the commission considered that the U.S. treaty view of its taxing rights followed from its domestic law analysis. In this context, see Shaheen, supra note 1, at 337.

12 See generally, e.g., Frank A. Engelen and Frank P.G. Pötgens, “Report on the Application of the OECD Model Tax Convention to Partnerships and the Interpretation of Tax Treaties,” European Taxation 265 (July 2000). It has been argued that these objections do not apply to the Luxembourg-U.S. treaty: see Shaheen, supra note 1, at 342.

13 In support of this view, see, e.g., Shaheen, supra note 1, at 342; contra, see, e.g., Lang, supra note 1, at 205.

14 OECD, “Draft Contents of the 2008 Update to the Model Tax Convention,” Part 1 (Apr. 21, 2008).

15 Article 25(2)(a) of the Luxembourg-U.S. treaty. This argument in support of the commission’s initial decision was only mentioned in the final decision (para. 117), perhaps as an afterthought.

16 See, e.g., supra note 9, paras. 56.1 and 56.2.

17 See Shaheen, supra note 1, at 340.

18 Supra note 3.

19 See supra note 3, para. 121. In fact, the commission only refers to a passage in the OECD’s 1999 partnership report (para. 108), which was transposed into para. 32.5 of the 2000 OECD commentary. Presumably they also had in mind related passages from the report and commentary.

20 See supra note 9, para. 32.5.

21 Contrast, e.g., supra note 9, para. 56.1 with 56.3. See also, in support, Lang, supra note 1, at 205; and Shaheen, supra note 1, at 339.

22 For a general critical discussion of this issue, see Engelen and Pötgens, supra note 12.

23 See Lang, supra note 1, at 206.

24 Although it has been suggested that it should be presumed that a country correctly applies its domestic laws (see Shaheen, supra note 1, at 343 (footnote 47)).

25 Supra note 3, para. 121.

26 See supra note 3, para. 120, where the commission describes the OECD’s second approach as involving differences of interpretation “unrelated to domestic law.”

27 Supra note 3, para. 121.

28 Supra note 9, para 32.4.

29 Supra note 3, para. 121.

30 Article 7(1) of the Luxembourg-U.S. treaty reflects the standard wording of the OECD model whereby, under the first sentence, business profits are only taxable in the residence state if there is no PE and, under the second sentence, they may (also) be taxed in the other state if there is a PE. See, suggesting this possibility, Sasseville, supra note 1, at 50.

31 See Shaheen, supra note 1, at 345, where it is argued that this approach should not be limited to “conflicts of classification.”

32 See supra note 3, para. 123.

33 See supra note 3, footnote 77.

34 For example, if in these other cases both treaty states took the view that there was a PE for treaty purposes, but the source state could not and did not tax purely as a matter of domestic law, the residence state would in principle still have to exempt but on different grounds (compare OECD commentary to article 23, paras. 56.1 and 56.2).

35 It appears from the final decision that there was some evidence that the U.S. tax administration confirmed that the U.S. branch was not taxable in the United States, but the reason is not stated. See supra note 3, para. 80.

36 See Shaheen, supra note 1, at 340-342, for the kind of analysis of U.S. and Luxembourg tax law that might have been expected from the commission.

37 Supra note 3, para. 108.

38 See, e.g., U.S. Treasury, “The European Commission’s Recent State Aid Investigations of Transfer Pricing Rulings” (2016).

39 Article 5(2) MLI. An equivalent article was added to the OECD model in 2000 (article 23A(4)), but was notably absent from the Luxembourg-U.S. treaty.

40 European Commission, IP/18/5831 (2018). The law was recently approved; see Law No. 1164 (Dec. 21, 2018).

41 European Public Service Union et al., “Unhappier Meal: €1 Billion in Tax Avoidance on the Menu at McDonald’s” (May 2018), at 6.

42 U.S. Department of the Treasury, Office of Tax Policy (June 22, 2016).

END FOOTNOTES

Copy RID