Problems of Interpretation of Article 5(3) OECD Model

by Adolfo Martín Jiménez*

In appearance, the one-sentence wording of article 5(3) of the OECD Model seems very simple: “A building site or construction or installation project constitutes a permanent establishment only if it lasts more than twelve months.”

It is, however, a rather complex clause if account is taken that (i) it very much reflects all the problems connected with the permanent establishment (PE) concept and its evolution along the years, (i) due to the “mobility” of construction works, article 5(3) of the OECD Model was the precursor of very relevant changes to the PE concept and the Commentary on Article 5 of the OECD Model in its different versions from 1977, (iii) the same clause in the UN Model seems to have a different nature, (iv) construction PEs have been affected by the OECD base erosion and profit shifting (BEPS) works in a very relevant form, (v) despite clarifications to the OECD Commentary on Article 5(3) in 2017 very relevant interpretative doubts remain unresolved, and, last but not least, (vi) construction projects and contracts can be intricate and affect a number of different types of taxpayers and activities.

Complexity makes interpretation difficult and may lead countries to hold divergent positions on the application of article 5(3) of the OECD Model. Therefore, this chapter will explore the problems of interpretation of such an article. For that purpose, it will start in section 2 with a historical study in order to explain the nature of the construction PE, the contradictions in this concept that are evident in the evolution of the construction PE and the geographical and commercial coherence tests that were born in the context of article 5(3) of the 2017 OECD Model to explain and, if possible, avoid some of those contradictions. As will be shown, the commercial and geographical coherence tests, if interpreted differently by the source and residence states, which is not difficult, may cause asymmetries and different views on whether or not there is a PE with outcomes that can be detrimental either for states or for the taxpayers, depending on the situation.

Second, the UN version of the construction PE (article 5(3)(a) of the UN Model) represents a different PE test for building activities. This version can be explained in part as a reaction to how the same test has evolved in the OECD context, but it could not completely get rid of its links with the OECD version, which makes this provision even more difficult to interpret than article 5(3) of the OECD Model. The different nature of article 5(3)(a) of the UN Model (in comparison with article 5(3) of the OECD Model), its contradictions as well as its limited relevance and effect in the current UN Model (2017) will be explained in section 3.

The BEPS-derived modifications to the construction PE will be the object of section 4, where the new anti-splitting up of contracts clause and its problems will be studied, as well as the alternative anti-abuse approach to fight against the artificial splitting up of contracts. In the context of article 5(3) of the OECD Model, the limited success and problems of the anti-fragmentation clause/anti-abuse approach proposed by BEPS Action 7, the Multilateral Convention to Implement Tax Treaty Related Measures to Prevent Base Erosion and Profit Shifting (MLI) and the 2017 OECD Model will be explained in that section.

Contrary to what it may seem, not all the problems of interpretation of the objective and subjective scope of article 5(3) of the OECD Model have been resolved in 2017, even if the 2017 version of the OECD Model has added some relevant new paragraphs in the Commentary on Article 5 and clarified previous interpretative doubts. Therefore, section 5 will concentrate on the issues still open on the objective scope of article 5(3) of the OECD Model (the meaning of “building, construction or installation” and the problems with on-site planning and supervision). Section 6 will deal with the subjective scope of article 5(3) of the OECD Model and how the article applies to contractors and subcontractors, partnerships or joint ventures. In turn, section 7 will concentrate on the issues connected with the computation of the time threshold of the construction PE (12 months in the OECD Model, 6 months in the UN Model, with different options also in other tax treaties).

Finally, section 8 will briefly deal with attribution of profit issues linked with construction PEs only to note that this is a highly controversial area and that very different approaches can be adopted by source and residence states, in part as a consequences of the different models on attribution of profits that can be found internationally and domestically in different states (due in part to the failure and scarce widespread acceptance of the OECD authorizes approach to attribution of profits to PEs).

* This paper is published as chapter 5 in G. Maisto (ed.), New Trends in the Definition of Permanent Establishment, Amsterdam: IBFD, 2019. Full text: