Non-cooperative tax regimes

Posted on

There's always been much debate about what a tax haven might be. It's one reason why I defined the term secrecy jurisdiction*. But secrecy is not the only criteria, although it is important. So too is the operation of the tax regime important. And this morning new research by Alex Cobham and Wouter Lips provides new insight into the combination of these factors and which countries are both secretive and have non-cooperative tax regimes.

This is important. The EU is issuing a list of non-cooperative regimes on 5 December but such black-lists have been notoriously unproductive because of political bias inherent within them - including a refusal to list 'friendly' states. That's why this new research is important. The authors say of it:

On 5 December 2017, the European Union is scheduled to release its ‘blacklist’ of non-cooperative tax jurisdictions, or tax havens. With pressure building after the release of the ‘Paradise Papers’ for decisive political action, media leaks indicate 53 jurisdictions have been given advance notification that they will be blacklisted unless they promise to change their tax rules. Historically, such tax haven lists have been highly political with no transparent criteria for inclusion. As a result, they have tended to identify the smallest jurisdictions, while ignoring the biggest players (Cobham, Janský & Meinzer, 2015). But the European Council published its criteria in advance, and so while their discussions in the Code of Conduct on Business Taxation group have remained typically opaque, it is possible to construct an assessment of which jurisdictions should be listed. We do this here, and provide the full dataset for each jurisdiction. We also go beyond the Council’s assessment in providing an evaluation of European Union member states on the same basis. In addition, we publish the full dataset openly so that others can make their own assessments.

In total, we identify 60 non-EU jurisdictions which fail to meet the criteria of which we think 41 should be listed, and an additional six EU member states: Cyprus, Ireland, Luxembourg, Malta, Netherlands and the United Kingdom. Full results are in Appendix Table A1.

I recommend taking a look. There are surprises: Guernsey and the Isle of Man do not make the list whilst Jersey and the UK do, for example. The authors add:

Of the Council’s three criteria, two are based on OECD criteria. We do not list low and lower-middle income countries on those criteria alone, since these countries had no effective voice in OECD processes.

Another important flaw is that the second criterion is impossible to assess externally. By construction, it depends on data provided only to the Code of Conduct group. Our evaluation here is necessarily subjective (although robust to criticism, as we argue). If we exclude jurisdictions that are only blacklisted against this criterion, in order to take the most conservative position possible, our results indicate blacklisting for a minimum of 28 non-EU jurisdictions (47 fail to meet the other two criteria). Finally, we highlight a further weakness of the criteria. As they are drawn, it is possible for jurisdictions to avoid blacklisting by providing very little cooperation indeed — as, for example, the United States will manage. A more rigorous approach would yield criteria that are more meaningful both in protecting EU members and in driving policy improvement internationally.

In other words, the EU blacklist will very much be a work in progress and not the finished article. But if it's as rigorous as Alex Cobham and Wouter Lips have been then it might just be useful.

* I define secrecy jurisdictions as places that intentionally create regulation for the primary benefit and use of those not resident in their geographical domain with that regulation being designed to undermine the legislation or regulation of another jurisdiction and with the secrecy jurisdictions also creating a deliberate, legally backed veil of secrecy that ensures that those from outside the jurisdiction making use of its regulation cannot be identified to be doing so.