George Orwell’s parable, Animal Farm, offers insights on political and governmental institutions that bear revisiting from time to time. BEPS Action 6 with the broad title “Prevent treaty abuse” has among its objectives “developing model treaty provisions to prevent the granting of treaty benefits in inappropriate circumstances” and “to clarify that tax treaties are not intended to be used to generate double non-taxation”. If Orwell’s Snowball was charged with formulating the commandment on the proper operation of tax treaties, it might have read “preventing double taxation good, generating double non-taxation bad.”
However international tax is not that simple. The final report on BEPS Action 6 recognises one situation where double non-taxation is acceptable, that is where pension funds are concerned. The final report confirms that additional work will ensure that a pension fund should be considered to be a resident of the State in which it is constituted regardless of whether that pension fund benefits from a limited or complete exemption from taxation in that State.
This is not a new issue. Questions over whether a pension fund is a person for purposes of Article 1 within the definition in Article 3(1)(a), and whether it is “liable to tax” within Article 4(1) to qualify as a resident of a contracting state of the OECD Model Tax Convention have challenged taxpayers and tax administrators alike. The result of traditional analysis will be different from country to country and from one pension fund to another. Amendments to the Commentary to Article 18 of the OECD Model Tax Convention addressed some of the complex issues relating to retirement savings and payments. This included a model definition of pension scheme in paragraph 37. Other versions have been developed including in Article 3(1) of the US Model.
A Discussion Draft on Changes to the OECD Model Tax Convention Concerning the Treaty Residence of Pension Funds was published by the OECD on 29 February 2016. It changes to Articles 3 and 4 of the OECD Model Tax Convention, and to the Commentary as promised in the BEPS Action 6 Final Report. These changes comprise a definition to be included in Article 3(1) and an explicit reference to qualifying pension funds as automatic residents of a state in Article 4(1).
The definition is thus to be elevated from a paragraph in the Commentary to the text of the Model itself. While the stated intention is that a pension fund should be considered to be a resident of the State in which it is constituted, the language adopted in a proposed definition of “recognised pension fund”.
Rule of law
The proposal is a significant improvement on the draft definition in paragraph 37 of the OECD Commentary to Article 18. The earlier version was required to be “accepted by the competent authority” of a state in order to be recognised. This administrative decision-making power is replaced by a legal rule, capable of interpretation and adjudication by the courts. The scope of the definition is limited to pension funds regulated by the state in which it is established. The term “regulated” is likely to come under some scrutiny. Does this mean that only states that have relatively sophisticated statutory regimes for pensions will ever qualify? Would a pension scheme regulated by, for example, trust law, be “regulated”?
It is a requirement for recognition. Under the proposal, the pension fund must be treated as a separate person for purposes of the tax laws of the state in which the fund is established. To some extent this defeats the purpose of a special definition. A pension fund which is treated as a person under the tax laws of a state will in any event meet the requirement to be a person by reason of the definition in article 3(1)(a). A useful definition would treat pension funds that lack legal personality, but share tax attributes of those that do have legal personality in the same way.
No real explanation is given why part of the test is by reference to regulatory provisions while another is by reference to tax provisions. The drafters of the proposal have, no doubt, been influenced by the pension regimes of their own countries. Whether this will serve any wider interests is unclear.
US model treaty
The US model treaty has, since 1996 contained a definition of pension scheme. The current version has been in place since 2006 and appears in many treaties. The interpretation of that definition was examined by the UK Upper Tribunal in Macklin v HMRC  UKUT39 (TCC). See my blog of 8 April 2015. It is surprising that a definition that has been in use for 10 years is not adopted or its utility or otherwise, considered in the consultation document.
There is a complete absence of discussion on the policy issues surrounding these proposals. No particular reason is given for the special treatment proposed for pension funds, nor explaining why the tests proposed are adopted. Consistent with the BEPS ethos, the definition is narrowly and carefully drawn. It is clear that well drafted treaty provisions will confer benefits appropriately and ensure that those benefits are not given in inappropriate cases.
The absence of any discussion on the policy issues is of some concern when the rules of international taxation are in the process of fundamental change. Ad hoc definitions without explanation of the objectives or guiding principle, give no guidance as to when treaty benefits are appropriate or inappropriate.