Transparency International official responds to our FSI

We've had a lot -- a whole lot -- of media coverage for our Financial Secrecy Index, and we'll bring you a bunch of new links in the coming days. It's generally been excellent coverage, with the predictable furious responses from those near the top of our index. Geoff Cook of Jersey Finance, for instance, is outraged, and makes a comment that demonstrates that he cannot have looked at it in any detail:

The report is presented as though it is based on accepted international standards.

No it isn't, Mr. Cook. The report demonstrates, explicitly, repeatedly, prominently and in great detail, how appallingly weak his "accepted international standards" are. A little fact-checking from Jersey Finance might have helped his credibility.

There wasn't too much else out there, in the way of criticism, but we thought we would highlight probably the most thoughtful one which came from Christian Humborg at Transparency International. First of all, he welcomes our study:

"When a police officer is bribed, usually two people are involved: the bribe-taker and the bribe-payer. In larger corruption cases, a third person is needed: the bribe-hider.
. . .
Tax Justice Network’s FSI nicely complements our Corruption Perceptions Index, which looks at the demand side (or bribe-taking) of corruption, and Bribe Payers Index which looks at the supply side. For example, it highlights the fact that always stood out for me: while Singapore is one of the countries where public sector corruption is perceived to be the lowest in the CPI, it is one of the most secretive financial centres world wide.

With the BPI being published next month and the CPI in two months, it will be the first time that all three dimensions of corruption are covered in one year."

Indeed. Now for the criticisms. First, he notes in passing that he's surprised to see the UK with such a low secrecy score, and Germany so high up the ranking. Well, we were surprised too, and perhaps even a little uncomfortable to see such results - but these are what the numbers tell us. Then, Humbold says this (and to understand his comment, you need to understand that the FSI is created out of two elements - a 'secrecy score' and a global scale weighting for each jurisdiction):

"My doubt of the FSI’s methodology, from an anti-corruption perspective, is the use of a global weighting. If a financial centre would be able to attract tons of money while having a low secrecy score it might still be at the upper part of the list. If there are reliable estimates of the size of illicit financial flows into a country, this might be an alternative number to consider."

It's a worthwhile point to make, but we would respond as follows.

First, we'd say that if we had combined, say, a secrecy score with a measure of the size of illicit financial flows, we'd probably be guilty of creating some circularity in our arguments: since illicit financial flows are a function of secrecy (though the relationship isn't straightforward, of course), we'd be kind of measuring the same thing twice.

To put this another way, combining the secrecy score with a weighting is basically a rather indirect way of creating a measure of illicit financial flows. It complements other measures of these flows that are out there - but by splitting it into two components it provides a whole lot more detail, and the secrecy score in particular, with its 15 components, provides a very sharp and focused series of tools for action on transparency.

And with respect to his (and our initial) discomfort on, say, the position of Germany - we think that the secrecy score, which is plainly verifiable, demonstrates our case. Its score of 57 puts it only just below the United States. Germany is simply not particularly transparent. If only it were.

And, as we explain elsewhere, the big change from our 2009 index is that we mathematically de-emphasised the scale weightings and enhanced the impact of the secrecy score on the final rankings. We felt, based on feedback from our previous index in 2009, that the previous ranking gave too much importance to the scale weighting. This new set-up allows each elements to have a large impact on our results.

For example, look at Japan, ranked 8th, and Spain, ranked way, way down the index, in 53rd place. They have almost identical global scale weights - 0.018 and 0.016 respectively (which means they have 1.8% and 1.6% of the global market) and yet the secrecy scores diverge very significantly: 64 versus 34 (making Spain the cleanest on our list.)

We think this helps demonstrate that we probably chose the the right formula for combining the two elements.