Article 7(1) of the OECD model treaty is perhaps the most important rule regulating the international taxation of business. It sets out the fundamental basis on which businesses are taxed, that is, the state of residence has the primary right to tax with source state entitlement restricted to taxing the profits of permanent establishments. Source states are also entitled to tax, certain specific items from sources located within the state as authorised by the remaining distributive provisions of the model treaty.

The existence or nonexistence of a permanent establishment is thus of critical importance. The PE definition has remained largely unamended for over a century despite dramatic changes in the way business is conducted across borders and the amount of international business has exploded. In the OECD context, article 5 reads today as it did in 1977. Undoubtedly it needs to be re-examined.

The current re-examination in the context of BEPS is hurried and fails to consider this fundamental issue in the context of the overall balance of treaties, and a viable meaning that will provide stability and a fair allocation of taxing rights. Viewing the definition through the filter of preventing abuse is bound to produce a distorted result. That is exactly what the proposals in the revised discussion draft on Action 7 released on 15 May 2015 offer.

Where is the abuse?

Despite the stated objective of preventing the artificial avoidance of PE status, the revised wording proposed for article 5 makes no distinction between cases where PE status is artificially avoided, and cases where the proposed amended definition applies.

There is no indication whether any consideration has been given to the interaction between Action 7, and Action 6, which deals with treaty abuse generally. If, for example, the proposed principal purpose test (PPT) or treaty GAAR serves its purpose, surely it would solve those abusive cases where PE is artificially avoided.

This is particularly troubling as key decisions of some of the most influential supreme courts in the world , such as the French Supreme Administrative Court in Zimmer Ltd (10 March 2010), have rejected arguments of abuse in precisely the circumstances the OECD suggests are abusive. Although the case for including commissionaires as agency PEs is justified by reference to Zimmer (Revised Action 7 discussion draft of 15 May 2015, paragraph 10), the OECD assiduously avoids mentioning that the court rejected the contention of abuse.

A more transparent approach would be to recognise that, under the existing language, a commissionaire is not, in principle, an agency PE, but some tax administrations want commissionaires to constitute PE and, therefore, propose amendments to that effect. Clothing the proposal in the robes of anti-abuse stifles debate on what the appropriate balance is in allocating taxing rights.

Agency PE

Demotion of the expression “in the name of the enterprise” to only one of the three categories is a welcome improvement that terminates debate about whether undisclosed agents are PEs. The real effect of other amendments is far from clear. Agency PE now explicitly includes not only the conclusion of contracts but also the “negotiation of material elements of a contract”. This goes further than the previous OECD view in paragraph 33 of the Commentary to Article 5 that required a person to “negotiate all elements and details of a contract in a way binding on the enterprise”. This amendment addresses the evidentiary issues that arose in determining who concludes the contract in some situations rather than abuse of the rule. Negotiation is now sufficient.

The bigger question is, if a person who habitually negotiates but does not conclude contracts is a PE, how much profit can be attributed to such a PE. Where contracting is of the essence of the business, such as in the sale of goods, functional analysis would suggest that little value is added by simple negotiation.


The same concern arises with the proposal to extend agency PE to contracts for the provision of services. It is far from clear what abuse is countered by this proposal (It is not even mentioned in either the initial Action 7 discussion draft of 31 October 2014 or the revised draft of 15 May 2015). States that are concerned about the provision of services by non-resident enterprises without a fixed place of business in their territory, have the option to include service PEs as suggested in the UN Model Article 5(3)(b). The UN Model has included this since 1980.

In the case of services, the source of profit is the performance of the service. Again functional analysis would suggest that no real value is added by conclusion or negotiation of such contracts.

We may have more PEs as a result, but what is their point, if no real profit is attributable to them?


Another casualty of the haste is the drafting of revised treaty provisions. The proposed modification of the definition of independent agency in Article 5(6)(b) is a case study on how not to draft an amendment to a treaty.

It is recognised in the revised discussion draft that excluding all associated enterprises from the possibility of qualifying as independent agents is too wide. Instead, it is proposed that a 50% equity participation or actual control should disqualify agents from independent status. The choice of 50% rather than a majority participation is curious itself and unexplained in either discussion draft. Apart from any policy questions, it certainly complicates the drafting.

The drafters of this proposal have ignored existing terminology in the Model and invented new expressions which are bound to give rise to questions of interpretation: Article 9 (Associated Enterprises) expresses association by reference to participation in the “management control or capital” of an enterprise. Consistent with that, Article 10(2)(a) (Dividends: substantial participations) refers to “the capital” (of the company paying the dividends).

In contrast, the nine lines of the revised proposal for Article 6(b) refers variously to “beneficial interests”, “vote and value of the company’s shares”, “beneficial equity interest” and “control”. Given the years that the OECD has struggled with the meaning of “beneficial ownership”, these choices are hard to understand.

A simple clear rule, using existing language of the Model is what is needed. This would eliminate the need for the lengthy Commentary and avoid disputes. One possibility might read:
“In this article, a person is connected to an enterprise if one has more than one half of the rights conferring participation in the management control or capital of the other, or both are under the control of the same persons or enterprises who hold more than one half of such rights.”

Whatever the ultimate shape of the PE definition, it is too important to rush through changes that are not fully thought through both from a policy and drafting perspective.


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

Kluwer International Tax Law

The 2022 Future Ready Lawyer survey showed that 78% of lawyers think that the emphasis for 2023 needs to be on improved efficiency and productivity. Kluwer International Tax Law is an intuitive research platform for Tax Professionals leveraging Wolters Kluwer’s top international content and practical tools to provide answers. You can easily access the tool from every preferred location. Are you, as a Tax professional, ready for the future?

Learn how Kluwer International Tax Law can support you.

Kluwer International Tax Law
This page as PDF


  1. Dear Jonathan, clearly and nicely written; Y very much enjoyed reading your contribution to Kluwer International tax blog.

    Additionally what still bother me it is the fact that the PE definition remains unchanged and unable to catch at source the various manifestations of the digital economy, as if we still were in the “brick & mortar” era; thus, unavoidably shifting the tax jurisdiction on it to residence countries. By the way, I have always read the 1923 Experts Report to the Society of Nations slightly different: The primary right to tax business income lays on the source countries to the extent the business enterprise has a PE therein. If my reading is correct and BEPS is not aimed at altering the fundamental paradigms of international taxation (including the embodied inter-nation equity) my point would be to provide a solution to the digital economy that preserve inter-nation balance.

    In that sense, and despide the current trend under the BEPS Project, I wonder whether for digital economy outcomes It would not be preferable to turn to a substantial presence test, or even mature domestic standards under anglo-saxon legislation such as “trade or business” or “doing business in” (as opposed to doing business with), which are the existing equivalents to the former.

    If, for these purposes, the digital economy outcomes are not treated differently from the traditional forms of business, sooner or later we should reconsider the issue beyond the PE context; e.g., resorting to some sort of formulary apportionment or destination based corporate taxation, and, as you might agree, beyond the soundness of these systems’ conceptual base, if that happens outside the BEPS Project (by adoption of isolated jurisdictions or groups of countries), the result may be chaotic for the international taxation of corporate businesses.

  2. Thanks Guillermo. I agree with your concerns. It is unfortunate that the allocation of taxing jurisdiction is approached piecemeal without any overall framework (except the distortionary abuse filter). Permanent establishment is, in principle, a bricks and mortar concept. Instead of thinking about the concept, it seems to be gradually eroding into a mere label for nexus. The danger signs are in the revised discussion draft – Action points on PE profit attribution. The next step will no doubt be, e.g., to come up with some strained theories to allocate profit to agents contracting for services – making it up as they go along.

    In practice, some digital economy issues will be addressed by the PE changes, particularly on preparatory or auxiliary activities and the anti-fragmentation rules.

    In the end, indirect taxes such as VAT (which are inherently destination based) are more appropriate to address the cross-border digital economy than profit taxes (which are inherently origin based).

Leave a Reply

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