On 6 June 2016, the European Commission finally released its decision in the McDonald’s State Aid case. After the clarifications recently provided on the Commission’s position concerning transfer pricing cases and the arm’s length principle (see especially its decision in the Belgian Excess Profits Exemption Scheme, §§ 145–150, clarifying the Commission’s reliance on an independent EU-law principle of arm’s length based taxation), this decision now provides a clearer insight into the Commission’s perspective on cases where the primary issue appears to be double non-taxation (I have previously dealt with both types of cases here). In today’s blog post, I will comment on the reasoning applied by the Commission in its McDonald’s decision.

As a preliminary remark, it is worth acknowledging that the Commission is not arguing that double non-taxation per se violates state aid rules – something that had not been as clear from the original press release on its decision to open a formal investigation published in December last year, quoting Commissioner Vestager as saying “The purpose of Double Taxation treaties between countries is to avoid double taxation – not to justify double non-taxation”. However, the Commission’s argument effectively results in imposing on Member States an interpretation of a tax treaty that prevents double non-taxation. The difference might appear small, but is of course exactly what proper application of the law requires.

The McDonald’s case: Facts

As a quick reminder of the facts, the following graphic (taken from the published Commission decision) should illustrate the case:

McDonald's graphic

The main element of the structure at issue is that royalties received by the Luxembourg company are allocated to its US Branch under Luxembourg’s interpretation of the LU-US DTC. At the same time, the US did not tax these royalties, considering them not to be “effectively connected with a US trade or business”, a condition for taxability of non-resident taxpayers’ income in accordance with § 871(b) US IRC. In light of this non-taxation in the US, the Commission thinks the granting of exemption in Luxembourg to be based on an incorrect interpretation of its treaty with the US, claiming that the correct interpretation of Article 23 of that treaty would allow Luxembourg to tax income it allocates to a US permanent establishment if the US does not exercise its taxing right. I will return to this highly doubtful position below, but first it is worthwhile understanding why such erroneous interpretation of its tax treaty would amount to state aid.

Why “voluntary” exemption could be state aid

As noted above, the Commission does not rely on any international principle of single taxation to denounce a double non-taxation result as achieved by McDonald’s. Instead, its reference system is Luxembourg’s own corporate tax law (see § 69 – referring to Paint Graphos as a source to determine that system’s objective: somewhat surprisingly, considering that that case concerned Italy and not Luxembourg), which has as a general principle world-wide taxation of resident companies, unless a tax treaty provides otherwise. Consequently, refraining from such world-wide taxation without it being mandated by a tax treaty creates an advantage in light of that general principle. The Commission then relies on the Court’s judgment in the MOL case to refrain from a more detailed selectivity analysis: according to its reading of that judgment, such analysis is unnecessary whenever an advantage is granted by way of an individual decision. (I consider this interpretation of the Court’s judgment and its application to the case of an individual ruling questionable, but will have to reserve that discussion for another time).

The concrete circumstances of the granting of the exemption – Luxembourg had first issued a ruling granting the exemption under the condition that the US would tax the income, which was then replaced by a second ruling without such requirement – seemingly help the Commission’s argument, but should be irrelevant: if the tax treaty in its correct interpretation indeed does not require taxation in the US (see next paragraph), the first ruling was plainly wrong and therefore had to be replaced. The undesirability of the outcome from a policy perspective cannot be relevant in this legal assessment.

On the Commission’s assessment of the LU-US DTC

At the heart of the matter thus lies the correct interpretation of the LU-US DTC, notably Art. 5, Art. 7 and Art. 25/2 (= Art. 23A OECD MC). Interestingly, the Commission does not address Art. 5 at all, which defines a PE for treaty purposes, although it acknowledges that the US’s right to tax depends on the existence of a PE (§87). It focuses, rather, on the interpretation of Art. 25/2 and its requirement that the relevant income “may be taxed”. The Commission acknowledges that no actual taxation is required, but merely a right to tax for the US. It subsequently concludes that “the profits … cannot be taxed in the United States … [s]ince the US Franchise Branch does not constitute a permanent establishment for US tax purposes” (§89). It is unclear how it came to that conclusion without an analysis of Art. 5 DTC. It seems likely based on McDonald’s tax advisor’s explanation that no PE existed for “US tax purposes” (§78). That would seem to be a reference to US domestic law rather than to the DTC. Merely because there is not PE for “US tax purposes” does not mean that the US would consider no PE to exist for purposes of the LU-US DTC. This would remain true even if McDonald’s argued differently before the US authorities: in §55, the Commission refers to its US tax return claiming the absence of any PEs for DTC purposes.

To support its conclusion, the Commission also points to the OECD Commentary on Art. 23A, which addresses possible cases of double non-taxation. Before looking further into the Commission’s own interpretation of the DTC, a few observations concerning the use of the OECD Commentary:

  • First, the LU-US DTC explicitly only refers to the avoidance of double taxation and prevention of tax evasion, but not to the avoidance of double non-taxation
  • Second, the LU-US DTC has been concluded in 1996, years before a change of the OECD Commentary that lead to the first inclusion of the cited paragraph
  • Third, according to case law of the Supreme Administrative Court of Luxembourg, substantive changes to OECD Commentary cannot affect the interpretation of pre-existing tax treaties (in contrast to “clarifications”)
  • Fourth, the Commission’s result, appears to be based on a misunderstanding of the OECD Commentary: The rule cited by the Commission, para. 32.6[1], addresses conflicts of interpretation. It allows the residence state to tax income if the source state “considers that the provisions of the Convention preclude it from taxing an item of income or capital which it would otherwise have had the right to tax”. But it seems that the US did not consider itself precluded from taxing the income of the US Franchise Branch by the LU-US DTC; it simply chose not to exercise its taxing right (see §46 and §54, the latter of which appears to explicitly acknowledge that the US actually had a taxing right under the tax treaty: “whether the US actually exercises its taxing right under the DTT, is therefore irrelevant”).

Did the Commission get it right?

Following the Commission’s argument outlined above, the key question to be answered remains whether the US had a right to tax the royalties attributed to the US Franchise Branch under the DTC. The Commission’s assessment is ambiguous on that question: it cannot answer it without analysing whether a permanent establishment existed under Art. 5 as interpreted by the US and whether the income could be allocated under Art. 7 as interpreted by the US.

Despite the criticism above, it is not impossible that the Commission got the correct result: If the US interprets “income attributable to the permanent establishment” (the DTC term) in the same way as “effectively connected with trade or business” (the US IRC term), it may consider itself barred from taxing the royalties in question. One could also argue, however, that in this case the DTC does not really limit its taxing rights, as it would not be blocked from taxing if it changed its domestic law in such a way that it wanted to tax foreign income.

The Commission might have considered more detailed information about US tax law than is visible in the decision: §§52-53, which are heavily redacted, hint at information stemming from McDonald’s US tax returns. The decision makes it also possible, however, that the Commission was somewhat confused about the difference between a PE under domestic law and a PE under the DTC. Merely because there is no PE for “US tax purposes” does not mean that the US would consider no PE to exist for purposes of the LU-US DTC.

Conclusion

The case requires an intricate analysis of a tax treaty. It is questionable whether the Commission got this right. The Commission’s decision is only to open a formal investigation, so more complete enquiry will surely follow. This post aims to point to certain elements that will have to be looked at in more detail before a final decision.

 

[1] “The phrase ‘in accordance with the provisions of this Convention, may be taxed’ must also be interpreted in relation to possible cases of double non-taxation that can arise under Article 23 A. Where the Source State considers that the provisions of the Convention preclude it from taxing an item of income or capital which it would otherwise have had the right to tax, the State of residence should, for purposes of applying paragraph 1 of Article 23 A, consider that the item of income may not be taxed by the State of source in accordance with the provisions of the Convention, even though the State of residence would have applied the Convention differently so as to have the right to tax that income if it had been in the position of the State of source.”

 


________________________

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


Kluwer Arbitration
This page as PDF

8 comments

  1. Dear Werner,

    thanks for your interesting remarks. Allow me to make one remark and raise one question on that:

    1) REMARK

    Article 25(2)(a) US-Lux Treaty corresponds to Article 23 A(1) OECD MC. The Commission apparently relies on the so-called “new approach”, according to which this article has to be interpreted as to avoid conflicts of qualification. Such conflicts may arise due to disparities between the domestic tax systems of the contracting states. According to Article 3(2) OECD MC, for example, when a contracting state applies a double taxation convention, “any term not defined therein shall, unless the context otherwise requires, have the meaning that it has at that time under the law of that State for the purposes of the taxes to which the Convention applies, any meaning under the applicable tax laws of that State prevailing over a meaning given to the term under other laws of that State”. If the meaning of an undefined treaty term has to be determined by reverting to the domestic law definition, it is possible that the contracting states ultimately apply the double taxation convention differently. After all, the definitions of the contracting states’ domestic laws do not necessarily coincide. This may lead to double taxation or double non-taxation.
    Since the Partnership Report was implemented into the Commentary on the OECD MC, such results should be avoided by a certain interpretation of Article 23 A(1) OECD MC (= Article 25(2)(a) US-Lux Treaty). The term “may be taxed in the other Contracting State […] in accordance with the provisions of this Convention” is crucial in this respect: If the contracting states do not apply different allocation rules because of a misapplication of the provisions of the double taxation convention by one of the contracting states, but this outcome is rather due to deviating domestic laws, which are decisive in order to determine the applicable allocation rule (i.e. Article 3(2) OECD MC), both contracting states’ result is in accordance with the double taxation convention. Let me provide for an example to show the conclusions with regard to situations of double taxation:
    – State A is the residence state and in order to apply the DTC concluded with State B it reverts to its own domestic law. Thus, State A comes to the conclusion that the residence State (= State A) has the exclusive taxing right.
    – State B also reverts to its own domestic law. Since it deviates from State A’s domestic law, State B comes to the conclusion that it (= State B) (also) has a taxing right.
    – In such a situation, both states would tax, because – due to their respective domestic law – they conclude that they have a taxing right. This would result in double taxation.
    – However, if either contracting states’ result is in accordance with the double taxation convention, it may be argued that the residence state has to exempt the income (Article 23 A(1) OECD MC), because the icome “may be taxed in State B”, if State B’s domestic law is taken into account. It is irrelevant that State A would come to a different result due to its own domestic law. Since the OECD MC considers both domestic tax systems to be likewise decisive when determining the applicable allocation rule, both solutions are “in accordance with the provisions of this Convention”.
    So far so good – according to the OECD, this should also work in situations of double non-taxation. However, I have some doubts, whether this is actually true: Let me again provide for an example to illustrate my concerns:
    – State A is the residence state and in order to apply the DTC concluded with State B it reverts to its own domestic law. Thus, State A comes to the conclusion that the other contracting State (= State B) has a taxing right and State A has to apply the exemption method (Article 23 A(1) OECD MC).
    – State B also reverts to its own domestic law. Since it deviates from State A’s domestic law, State B comes to the conclusion that it (= State B) has no taxing right.
    – In such a situation, both states would refrain from taxaxation, because – due to their respective domestic law – they conclude that the respective other contracting state has the taxing right. This would result in double non-taxation.
    – According to the OECD, “[t]he phrase “in accordance with the provisions of this Convention, may be taxed” must also be interpreted in relation to possible cases of double non-taxation that can arise under Article 23 A. Where the State of source considers that the provisions of the Convention preclude it from taxing an item of income or capital which it would otherwise have had the right to tax, the State of residence should, for purposes of applying paragraph 1 of Article 23 A, consider that the item of income may not be taxed by the State of source in accordance with the provisoins of the Convention, even though the State of residence woul dhave applied the Convention differently so as to have the right to tax that income if it had been in the position of the State of source” (see Commentary on Article 23 A and B, para. 32.6.)
    – I have some doubts regarding this paragraph: If both conclusions are equally in accordance with the double taxation convention, because both domestic tax systems are likewise decisive when determining the applicable allocation rule (Article 3(2) OECD MC), this should also be taken into account when applying Article 23 A(1) OECD MC. If, by reverting to the residence state’s domestic law, an item of income or capital may be taxed in the other contracting state, the wording of Article 23 A(1) OECD MC suggests that the residence state has to grant an exemption, irrespective of whether the other contracting state’s domestic law would lead to a different conclusion. The application of the exemption method only requires that either the residence state’s or the other contracting state’s or – ideally – both contracting states’ domestic laws lead(s) to the conclusion that an item of income “may be taxed in the other Contracting State, […] in accordance with the provisions of this Convention”. The wording neither requires that both contracting states come to the same conclusion, nor stipulates that the conclusion of one of the contracting states is binding on the other one. However, according to the OECD, “[the residence state] should nonetheless consider that, given the way that the provisions of the Convention apply in relation to the domestic law of [the other contracting state], that State may not tax the income in accordance with the provisions of the Convention; [the residence state] is thus under no obligation to exempt the income” (OECD Partnership Report, at para. 110) Since there is no danger of double taxation, this interpretation is considered to be “consistent with the basic function of Article 23 which is to eliminate double taxation” (see Commentary on Article 23 A and B, para. 32.6.).
    So even if the Commentary on the OECD MC states that double non-taxation arising due to a conflict of qualification can be solved by a certain interpretation of Article 23 A(1) OECD MC, I am not sure whether such an interpretation is actually in line with the wording of this provision. As there is also not (yet) a specific aim of double taxation conventions to avoid double non-taxation, I am moreover not sure whether teleological considerations would necessarily require a different understanding.

    2) QUESTION

    Even if the doubts above are left aside, I wonder whether the McDonald’s case is actually about the OECD’s interpretation of Article 23 A(1) OECD MC. After all, the so-called “new approach” can only be applied, if the contracting states have to rely on their domestic law in order to determine how to interpret and apply the double taxation convention. But the McDonald’s case is about the question whether there is a PE according to the US-Lux treaty or not. I have always assumed that the term “permanent establishment” as well as the phrase “attributable to the permanent establishment” can be interpreted autonomously, without there being a need to look into the respective domestic law of the contracting states. If this is also the case with regard to the US-Lux treaty, one cannot rely on the OECD’s interpretation of Article 23 A(1) OECD MC anyway. It does only apply to conflicts of qualification that arise due to differences in the domestic law. Other situations of double non-taxation, which may arise due to “different interpretations of the provisions of the Convention or of the facts” (see OECD Partnership Report, at para. 113), can only be avoided by means of Article 23 A(4) OECD MC. Since the US-Lux treaty does not contain such a provision, I wonder whether the Commission has understood the Commentary on the OECD MC correctly?

    Best regards,

    Christoph

    1. Christoph,
      thanks for your very detailed comment. I think I fundamentally agree with your analysis and your view. Yet I am not sure whether I can entirely agree with your point concerning the application of the “new approach” only to cases in which contracting states “rely on domestic law for the interpretation of treaty terms”. You are of course correct that the OECD limits its proposed solution for instances of double taxation in the partnership report and para 32.3 of the Commentary to these cases. The reason is, of course, that the residence state should not be required to provide relief in cases of a genuine disagreement as to the content of the treaty. But the issues addressed in para 32.3 and 32.6 differ: double taxation = requiring relief from the residence state vs. double non-taxation = not requiring relief from the residence state. I think that the OECD’s solution to conflicts is thus also different when it comes to double non-taxation and also allows the residence state not to provide relief where it would consider that the source state had a right to tax, but the source state comes to a different conclusion (without any recourse to its domestic law).
      Quite different – and there I agree with you again (as I would have hoped becomes clear from my post) – is the case where both contracting states agree that the source state has a right to tax, but the latter chooses not to exercise it. I also think that this is what actually happened in the McDonald’s case and this is clearly not addressed by para 32.6 of the OECD Commentary. (As I wrote above, “the Commission’s result appears to be based on a misunderstanding of the OECD Commentary: … it seems that the US did not consider itself precluded from taxing the income of the US Franchise Branch by the LU-US DTC; it simply chose not to exercise its taxing right”)
      Werner

      1. Dear Werner,
        thanks for your response. Although our views seem to generally coincide, there is one issue where we seem to have differen opinions: If I understood you correctly, you would apply the “new approach” to all situations of double non-taxation that result from the contracting states applying the tax treaty – for whatever reason – differently. I am not sure, though, whether this is also the view of the OECD. There would have been (except for situations of “low taxation”) no reason to include Article 23 A(4) into the OECD MC, if double non-taxation could already be solved by interpreting Article 23 A(1) OECD MC in the way of the “new approach”. Moreover, at least from a policy perspective, it would be criticizable that the “new approach” regarding Article 23 A(1) OECD MC can solve ALL situations of double non-taxation, but only SOME situations of double taxation.
        Christoph

        1. You are right for cases where the tax treaty includes Article 23(4); indeed, the OECD makes that delimitation of cases of double non-taxation quite explicit in para 56.3 of its Commentary on Art 23. However, the wording of its paragraph 32.6 is broader and thus would allow a more extensive interpretation in cases where no Article 23(4) has been included. I agree that this solution is not entirely satisfactory – but it is not entirely impossible to argue, and might thus be a sort of life-line for the Commission, as they could apply the “New Approach” to argue against double non-taxation even though it would – in their own view – stem from a different interpretation of the treaty term “PE”, which is defined in the treaty. This is of course entirely academic, if, as seems most likely, the real reason for the non-taxation is the US not picking up its taxing right. I would certainly not countenance extending the New Approach to that situation (which is not even covered by Art 23(4)).

  2. Well done Werner. This is perhaps the most important treaty case in the EU at the present time. Its implications will be far reaching.
    The treaty needs to be interpreted in accordance with the principles of treaty interpretation set out in the Vienna Convention on the Law of Treaties (See my blog on this of 7 July 2015https://kluwertaxblog.com/2015/07/07/through-the-smokey-looking-glass-opaque-or-transparent/. The purpose of a treaty is not itself an interpretative tool. It is simply part of the general rule in VCLT Art 31(1) which requires the ordinary meaning of terms in context and in light of the object and purpose of the treaty. The OECD Commentary is supplementary material the use of which is discretionary and to be used as specified in VCLT Art 32.
    The second and more important constitutional issue is who should interpret a bilateral treaty to which a member state is a party? In Case C 128/08 Damseaux v Belgium, where the fundamental freedoms were involved, the ECJ ruled that this is a question for national courts. Armed with a decision of the national courts, the CJEU is then in a position to consider the relevant EU law issues.

    1. Jonathan, many thanks for your comment. I completely agree with you: this case has indeed far-reaching implications for the relationship of EU law and tax treaty law. It will be interesting to see whether the Commission continues on its current path to take its own autonomous view on the interpretation of tax treaties and whether the Court of Justice is ultimately going to accept this in light of the division of powers and competences among the European judiciaries.

  3. Thanks for the article, very interesting. I agree with your third-to-last paragraph; if the US rules of “connected income” are similar to the OECD meaning of “attributed income” and the income was neither connected nor attributable to the US PE, the exemption was unduly applied. However, if both concepts are different there would be a mismatch where the technical solution (OECD attribution but no US connection) leads to double non-taxation.

    I have suffered headaches in the past with similar mismatches. For instance, the UK distinct treatment for “trading” and “investing” non-residents in not always coherent with the tradictional tax treaty PE definitions.

    Thanks again!

Leave a Reply

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