MLI: Hot or not?

UK 5 – Netherlands 12. That may sound like a rugby score but actually it’s the number of optional anti-avoidance provisions that each country has signed up to under the OECD’s multilateral convention on tax treaties. Another telling score is UK 8 – Netherlands 1, being the number of such provisions each country has effectively rejected. It is tempting to draw some conclusions from these scores. For example, that the Netherlands has had enough of being been publicly branded in numerous high profile tax cases and international reports as one of the world’s biggest tax havens; and on the other hand, that the UK is keeping its tax haven options open in a post-Brexit world.

MLI: why? To better appreciate the significance of these scores, we should recall what the Multilateral Convention (or ‘Instrument’ as it is also known, hence, ‘MLI’) was created to do. The MLI was spawned from the OECD’s initiative consisting of 15 actions to combat tax base erosion and profit shifting (‘BEPS’) – a.k.a. international tax avoidance. The MLI was one of these actions and was designed as an efficient method of getting a whole raft of BEPS measures quickly incorporated into existing bilateral tax treaties (the rest of the measures being left to be implemented in domestic law). Tax treaties have traditionally been concluded in order to remove tax obstacles to cross-border business, but have increasingly been seen as vehicles that promote tax avoidance. The function of the MLI is to minimise the opportunities for abusing treaties while retaining their traditional role of helping international business. That’s why the MLI also contains some provisions aimed at helping taxpayers, in particular those for resolving disputes with tax administrations.

Big data So how successful has the project been to date? The OECD made a big thing of the numbers involved when it was signed last June: more than 2,350 ‘listed’ treaties, more than 100 tax jurisdictions involved in the negotiations, with 71 signed up and more in the pipeline. Even if one takes into account that there are more than 3,000 bilateral tax treaties that could have been included, 2,350 still sounds like a good score. But actually it’s a bit misleading since it’s basically the number of treaties (with a few exceptions) concluded by the participating countries, and so includes treaties with countries that haven’t actually signed up.

The Tinder test The more relevant number is the (more than) 1,100 treaties that the OECD says have been ‘matched’. For obvious reasons a bilateral treaty will only be modified if both parties have indicated that they want it to be, in which case it is ‘matched’. While that also sounds like a lot of treaties, even that number is misleading since, with the exception of a few obligatory provisions, a matched treaty will generally only be modified if both parties agree to the particular MLI provision in question. In this respect the MLI works a bit like the Tinder dating app: if both parties like the provision it will be included and otherwise not. However, unlike Tinder, most provisions are ‘liked’ by default, in other words, provided neither party objects (technically, enters a reservation) the provision is accepted. In other cases it works more like Tinder in that both parties must positively ‘like’ the same provision, i.e. opt for it, to have it included. However, the Tinder analogy only goes so far. For example, occasionally a provision will apply even though it has only been ‘liked’ by just one of the parties. How it then applies depends on the provision in question. And while ideally reservations are supposed to apply across the board, there is generally an option to limit a reservation to treaties that already contain a relevant provision, or conversely to those treaties that do not. In some cases reservations can be limited to specific treaties. In the end it’s a bit more complicated than Tinder! So to take the UK-Netherlands treaty by way of example, notwithstanding the 12 Dutch ‘likes’ (out of a possible 13*), mainly because of the more cautious UK approach, only four of the provisions in question have actually got a match.

MLI: hot or not? It’s tempting to conclude from the above that the MLI has not been the unqualified success its creators hoped it would be. However, before doing so one needs to look more closely into the possible reasons why a specific MLI provision was rejected by a particular country. There can be all sorts of reasons for this, of which some can sound very convincing. For example, the UK did not choose any of the options under the provision that is aimed at preventing abuse of the exemption method for relieving double taxation under treaties. There’s a very good reason for this, namely that the UK doesn’t apply the exemption method under its treaties. It could also be that a country considers the potential tax revenue to be gained – for example from some of the provisions on permanent establishments – insufficient to justify the additional compliance burden on the taxpayer: this may reflect a sound reason but equally it could mask a deliberate policy choice to forego revenue by, for example, not countering tax avoidance in favour of attracting foreign investment. Some provisions may be considered unnecessary in that they are already covered by another provision: a good example is the anti-splitting provision for construction PE’s which can be seen as an alternative to the (obligatory) ‘principal purpose’ anti-avoidance test. It could also be that some of the MLI’s provisions are considered too complex, again in relation to the potential revenue benefits involved: that appears to be the reason for the Netherlands to have entered a reservation against the MLI’s ‘saving clause’.

Seen in the light of the above, it is less obvious that either of my initial two conclusions is correct, and both the UK and the Netherlands may well be closer in terms of their positions on anti-avoidance than might at first appear. On the other hand, a country such as Cyprus, that has managed to reject every single provision that it was allowed to, may still have some explaining to do. Of course it could be that they are just still trying to work out what it all means and will withdraw some of the reservations once they get that far. Maybe. Well, I’m here if they want some help! Cyprus 0 – Netherlands 12?

*Transparent entities (Art. 3(1) and Art. 3(2)); Dual resident entities (Art. 4(1)); Elimination of double taxation (Art. 5 (options) and Art. 5 (unilateral application); Dividend transfers (Art. 8(1)); Capital gains from real estate shares (Art. 9(1) or Art. 9(4)); Third jurisdiction PEs (Art. 10(1-3)); Saving clause (Art. 11(1)); Avoidance of PE through commissionaires etc. (Art. 12(1-2); PE specific activity exemptions (Art. 13 (options) and Art. 13(4) (anti-fragmentation)); Splitting up contracts (construction PEs) (Art. 14(1)).