Spanish Supreme Court Gores Taxpayer After Business Restructuring

By Gary D. Sprague, Esq.  

Baker & McKenzie LLP, Palo Alto, CA

The Supreme Court of Spain recently entered the ring to provide its interpretation of some of the permanent establishment (PE) principles that have been at the center of much discussion lately, including in the current Organisation for Economic Co-Operation and Development (OECD) project to provide clarifications to the OECD Commentary on the PE definition in Article 5 (Permanent Establishment) of the OECD Model Tax Convention on Income and on Capital (OECD Model).  This case1 shows that there is an extreme need for clarifications.

The taxpayer, Roche Vitamins Europe, a Swiss member of the Hoffman-La Roche group, and its Spanish affiliate Roche Vitaminas, S.A., perhaps got off on the wrong foot with the local tax inspector when the group engaged in a business restructuring of the Spanish operations. Prior to the restructuring, the Spanish entity ("SpainCo") had functioned as a manufacturer and distributor of products into the Spanish market. After the restructuring, SpainCo operated as a contract manufacturer for a related party established in Switzerland ("SwissCo"). SpainCo provided sales promotion services to SwissCo, although it did not hold the authority to conclude sales contracts in the name of SwissCo. Finally, SpainCo also rented to SwissCo some space for the storage of products. The case report disclosed the group's transfer pricing policies for these arrangements: cost plus 3.3% on the manufacturing side, and a 2% margin on the sales it promoted on the sales side.

The local tax inspectors were quick to challenge the new structure. The years covered by the case were the two years immediately following the restructuring.

The Court addressed the contentious issues of when a PE can be deemed to exist under the dependent agent provision of the treaty and of the measure of profits attributable to a deemed PE. In addressing the first question, the Court also waded into the issue of whether the OECD Commentary should be regarded as ambulatory, i.e., whether Commentary language adopted after a particular treaty has been ratified by the Contracting States can be used to interpret that treaty.

The Court's decision is consistent in one regard: its analysis is erroneous on all three points. The challenge for taxpayers will be how to properly plan their affairs in a legal environment where startling interpretations of the treaty can become the law of the land.

As a preliminary matter, the Court properly noted that the Spain-Switzerland Income Tax Treaty ("Treaty") would prevail over domestic law definitions of tax nexus, and that under the Treaty two independent bases exist for creating a PE. The applicable treaty was a fairly old one, having entered into force in 1966. Nevertheless, its relevant terms are essentially the same as those of the current OECD Model, with the important exception noted below. Article 5(1) of the Treaty defines a PE to mean "a fixed place of business in which the business of the enterprise is wholly or partly carried on." Article 5(4) of the Treaty sets forth the dependent person deemed PE rule as follows:A person acting in a Contracting State on behalf of an enterprise of the other Contracting State - other than an agent of an independent status to whom paragraph 5 applies - shall be deemed to be a permanent establishment in the first-mentioned State if he has, and habitually exercises in that State, an authority to conclude contracts in the name of the enterprise, unless his activities are limited to the purchase of goods or merchandise for the enterprise. 

This text is identical to the corresponding text of the 1963 OECD Model.

The Court began its analysis by examining each of the three commercial relationships between SwissCo and SpainCo to determine whether any of them created a PE. First was the lease of warehouse space. Adopting the reasoning of the lower court, the Court concluded that no PE existed under the "fixed place of business" test, on the basis that the activities conducted by SwissCo at the leased warehouse space were covered by the preparatory or auxiliary exceptions for the use of facilities and the maintenance of a stock of goods or merchandise solely for the purpose of storage, display, or delivery of goods.

Next up was the sales promotion contract. The Court again adopted the lower court analysis and conclusion, finding that no dependent agent deemed PE existed due to the fact that SpainCo did not have the authority to conclude contracts in the name of the principal. So far, so good.

The analysis went off the rails, however, when the Court turned to the contract manufacturing relationship. Even acknowledging that SwissCo did not operate through a fixed place of business in Spain, the Court nevertheless concluded that SwissCo conducted its business in Spain through a permanent dependent agent, and that this relationship created a deemed PE under the dependent person article of the Treaty.

The Court supported the conclusion that SwissCo was operating through a dependent person by reference to the two contracts between the companies, "one for manufacturing and one for promotion,"2 apparently concluding that the combination of those activities provided the factual basis for the conclusion. As further factual support for the conclusion, the Court referred to the lower court's statements that SpainCo manufactured product only for SwissCo, that it operated under SwissCo's guidelines, and that it was at risk only for production errors which resulted from not following SwissCo's quality parameters, while SwissCo bore the other manufacturing risks.

One certainly can debate whether these facts would justify a conclusion that SpainCo was a "dependent" person within the meaning of the Treaty. Further, Article 5(4) of the Treaty describes only those dependent persons that are acting "on behalf of" the nonresident enterprise, which could describe the acts of an agent with the power of representation but normally not a supplier of contract manufacturing services. Those points aside, most observers would have thought that the relationship between SpainCo and SwissCo could not create a deemed PE, since Article 5(4) of the Treaty on its face calls such a PE into existence only in cases where the dependent agent "has, and habitually exercises in that State, an authority to conclude contracts in the name of the enterprise."

The Court escaped this apparently clear requirement by a remarkable misunderstanding of the OECD Model, and an equally remarkable application of the doctrine of the ambulatory Commentary.

The Court clearly concluded as a matter of result that the contract manufacturing activities, in combination with the sales promotion activities, constituted a dependent agent deemed PE.  While the case report is not entirely clear on the legal analysis to support that conclusion, it appears that the Court justified this interpretation by concluding that changes to the OECD Model in 1977 supported the view that the dependent agent PE did not have as a requirement that the agent have and exercise the authority to enter into contracts in the name of the principal.

As noted above, the Treaty came into effect in 1966. Its Article 5(4) was identical to the text of the 1963 OECD Model. In 1977, the OECD released a new version of the OECD Model, and elaborated the dependent agent paragraph slightly. The renumbered Article 5(5) in the 1977 Model (which is the same as the current version) reads as follows, with the italicized language being the text which differs from the 1963 Model:Notwithstanding the provisions of paragraphs 1 and 2, where a person - other than an agent of an independent status to whom paragraph 6 applies - is acting on behalf of an enterprise and has, and habitually exercises, in a Contracting State an authority to conclude contracts in the name of the enterprise, that enterprise shall be deemed to have a permanent establishment in that State in respect of any activities which that person undertakes for the enterprise, unless the activities of such person are limited to those mentioned in paragraph 4 which, if exercised through a fixed place of business, would not make this fixed place of business a permanent establishment under the provisions of that paragraph.  

The Court apparently interpreted this last clause to mean that if the dependent person conducts activity with respect to the business of the principal which is not preparatory or auxiliary activity, then the deemed PE provision of the Treaty (or presumably of Article 5(5) of the OECD Model) can exist. This interpretation ignores the clear requirement in the Treaty and in Article 5(5) of the OECD Model that a dependent agent PE exists only if the agent has, and habitually exercises, the requisite authority to conclude contracts in the name of the principal.

The taxpayer, quite properly, argued that a change to the OECD Model in 1977 cannot change the plain meaning of the text of the treaty as actually entered into between the Contracting States 11 years before. The Court justified relying on the revised OECD Model by asserting that "the OECD recommended in 1997 that tax administrations in member States follow the criteria established in the new Model Tax Treaty to the extent possible for the purpose of applying and interpreting the network of Tax Treaties." This reference apparently is to the statement which now appears in paragraph 33 of the Introduction to the OECD Model, relating to the relationship between the 1977 OECD Model and prior versions, which provides as follows:When drafting the 1977 Model Convention, the Committee on Fiscal Affairs examined the problems of conflicts of interpretation that might arise as a result of changes in the Articles and Commentaries of the 1963 Draft Convention. At that time, the Committee considered that existing conventions should, as far as possible, be interpreted in the spirit of the revised Commentaries, even though the provisions of these conventions did not yet include the more precise wording of the 1977 Model Convention. It was also indicated that Member countries wishing to clarify their positions in this respect could do so by means of an exchange of letters between competent authorities in accordance with the mutual agreement procedure and that, even in the absence of such an exchange of letters, these authorities could use mutual agreement procedures to confirm this interpretation in particular cases. 

This paragraph is an expression of the doctrine of the "ambulatory Commentary," namely the view that changes to the Commentary can be taken into account to interpret treaty articles which were entered into prior to such Commentary changes. The exact scope of when this doctrine will be applied by courts around the world is uncertain. In the United States, the Tax Court referred to subsequently released Commentary to interpret the dependent agent clause of the applicable treaty where the text of such treaty did not differ from the OECD Model, Taisei Fire & Marine Insurance Co. Ltd., 104 T.C. 535 (1995), but the Court of Federal Claims concluded in National Westminster Bank, PLC v. U.S., 58 Fed. Cl. 491 (2003), aff'd, 512 F.3d 1347 (Fed. Cir. 2008), that the OECD Discussion Draft on Attribution of Profits to Permanent Establishments released after the ratification of the applicable treaty was irrelevant to interpret Article 7 (Business Profits) of such treaty.

The remarkable application of the doctrine here is that the Court looked not to later Commentaries interpreting unchanged treaty language, but to changes to the OECD Model itself. The Court thus looked for guidance in language of a subsequent OECD Model that was different from the actual treaty text being interpreted.

Once it is decided that a PE exists, the next step is to determine the profits attributable to the PE. Given the willingness of the Court to look at post-ratification materials, one would have thought that perhaps the recent OECD work on attributing profits to a PE, including revisions to Article 7 (Business Profits) and modifications to the Article 7 Commentary, might have been in play. Unfortunately for the taxpayer, the new Authorized OECD Approach of treating the PE as a separate and distinct entity, then determining the deemed dealings with the remainder of the enterprise to provide the foundation for attributing profits to the PE, was never mentioned.  Instead, the Court concluded that the profits of the PE included those arising from both manufacturing and selling activities, relying on paragraph 34 of the Commentary on Article 5, which states that once an agent creates a PE, a PE of the enterprise exists "to the extent that the person acts for the latter, i.e. not only to the extent that such a person exercises the authority to conclude contracts in the name of the enterprise."

While the text of the decision is not clear, it appears that the result of this conclusion was that all of SwissCo's income from sales to Spanish customers was attracted to the Spanish PE. This result is inappropriate on its face, as it gives no recognition to the business activities in fact performed by SwissCo with respect to the Spanish market. In any event, the only remaining profit attribution step addressed by the Court was the allocation of expenses to that income. The Court agreed that in principle the Swiss expenses would be properly allocable to the Spanish income. After dangling that possibility of a semi-reasonable result in front of the taxpayer, the Court then delivered the death thrust by concluding that no expenses could be allocated to the PE's gross income on the basis that the taxpayer had not sufficiently proved up the exact amount of allocable expenses, apparently leaving the taxpayer to pay income tax on its gross income from the Spanish market.

Time will tell which of the various messages being delivered by the Court will be the lasting lessons of the Roche case. It certainly is a shot across the bow for groups that have engaged in business restructurings. It is a very troublesome decision regarding the approach to profit attribution to a Spanish PE, and perhaps a warning to taxpayers that they should prove up allocable expenses early in the controversy in order to anchor that second line of defense should the PE barrier be breached. The conclusion that the combination of contract manufacturing and sales promotion in a single entity creates a dependent agent PE perhaps argues for bifurcating those activities into separate legal entities, at least in Spain, even though such a step should not be necessary under a proper interpretation of the Treaty.

It remains to be seen whether the Spanish authorities will continue to assert a deemed PE in cases involving a related-party contract manufacturer subject to the normal management typically supplied by the principal. The Discussion Draft issued by the OECD in its current project to clarify the application of the PE rules includes an example which clearly points to the opposite result; hopefully, the OECD will publish a final report that unambiguously states that typical contract manufacturing relationships do not create PEs, on either a "fixed place of business" or a "dependent agent" basis. Finally, from the perspective of the effect of the "ambulatory" Commentary, the Roche case perhaps gives some comfort to those who advocate that the Commentary should indeed provide guidance even for the interpretation of previously negotiated treaties, but the Court's failure to understand the difference between the OECD Model and the Commentary makes the Roche decision not a very credible exemplar for that view.

This commentary also will appear in the May 2012 issue of the  Tax Management International Journal. For more information, in the Tax Management Portfolios, see Katz, Plambeck, and Ring, 908 T.M., U.S. Income Taxation of Foreign Corporations, and Cole, Kawano, and Schlaman, 940 T.M., U.S. Income Tax Treaties-U.S. Competent Authority Functions and Procedures, and in Tax Practice Series, see ¶7130, Foreign Persons-Effectively Connected Income, and ¶7160, U.S. Income Tax Treaties.

 1 Supreme Court Judgment (Section 2) N 1626/2008 on January 12, 2012. 

 2 All translations in this commentary are informal.