The criteria for stress testing

In my previous blog I have explained why publication of the stress testing results in the EU are not likely to have any effect. Below, I want to explain shortly why the criteria for the EU stress testing were set such as they were. Furthermore, I would like to add some additional comments on the stress testing criteria and its effects.

Previously, I stated that extreme scenarios were excluded from the stress tests because they would lead to too many defaults. However, there is more to it than that. First of all, including more extreme scenarios seems politically impossible; politicians do not consider sovereign defaults a likely scenario. The ECB and such will do whatever is in their power to prevent sovereign default. Market interventions play a role here: they will prevent such extreme scenarios to be tested. But the aim of stress testing is to check which banks will require capital injections should a sovereign default occur. Again, this is what makes the stress tests not credible. Another argument for not including extreme scenarios is that it is not that hard to imagine what would happen if some Southern European member states’ governments would default. But what is then the purpose of a stress test?

IMF research on stress tests, taking Iceland as an example, shows that stress tests do not have to be incredibly sophisticated, as long as the assumptions and scenarios (or the weight of the credit and market risks analysed) are appropriate. Too bad that this is ignored by the CEBS (Committee of European Banking Supervisors). Or actually the CEBS is not to blame; with its mandate being “giving advice to the EC on policy and regulatory issues related to banking supervision, promoting cooperation and convergence of supervisory practice across the EU, and contributing to consistent implementation of guidelines and recommendations” it has no regulatory power to require needed criteria for the stress tests.

My advice is to make to CEBR independent from national supervisors and Central Banks and give it the needed regulatory power to set more strict criteria for the stress tests, including:

1. Scenarios thought likely by financial markets;

2. Defining banks’ capital not as Tier I capital, but as the more stringent core equity (which would be consistent with the new Basel rules);

3. Require uniform disclosure of the results, and;

4. Require banks and governments to draw up contingency plans should markets panic after publication.

One last thing: short after release of the U.S. stress tests’ results last year, markets restored. We do not know whether this was because of the publication of the results in itself, or the required raising of additional bank capital (the U.S. contingency plan).