29 January 2011

EU-wide Bank Stress Testing Exercise: 2011 vs 2010

On 23 July 2010 CEBS released its summary report on the results of the 2010 EU-wide bank stress testing exercise, which covered 91 banks in Europe with total assets of EUR 28 032bn as of end of 2009, and represented approximately 65% of the EU banking system and at least 50% of the national banking sectors in each EU Member State.

Besides and perhaps even instead of the stated objective (“...to increase the level of aggregate information among policy makers in assessing the resilience of the European financial system as a whole”), the aim of the exercise definitely was to convince markets that EU banking sector is resilient to the “extreme assumption” of sovereign shock.

Out of the tested 91 banks, only seven not-too-big-to-save ones failed the test. Despite of critical voices about the severity of the scenario, and concerns about analysis scope and assumptions, for a brief happy moment it seemed that tensions in Europe eased. Unfortunately, barely two months later on 30 September 2010, the Irish Government announced plans to use its National Pensions Reserve to inject EUR 3.7bn of capital into Allied Irish Banks (AIB), becoming the majority shareholder and effectively nationalizing the bank. This was despite that AIB had passed the test and the adverse scenario had not even materialised.

Indeed, the case of AIB looks ridiculous. According to the 2010 EU-wide stress test scenario including sovereign shock, the AIB’s Tier 1 capital ratio would have fallen from 7% on 31 December 2009 to 6.5% by the end of 2011, and the bank would have a capital surplus of EUR 352m (see the stress test result for AIB on the website of Ireland’s banking authority). In reality, by the end of the second quarter of 2010 (i.e. already before the stress test results were announced) the AIB’s Tier 1 capital ratio had fallen to 6% (the respective Q2 results for AIB can be found from the bank’s website) which was the critical threshold of passing or not passing the stress testing exercise. The assumed disposals that were also taken into account in stress test just did not save the bank for needing more tax payers’ money.

Now, when sovereign debt crisis is spreading from one EU country to another and trustworthiness of European banking regulators rightly questioned, we could expect 2011 EU-wide bank stress testing exercise to be taken much more seriously than the previous one. Soaring government bond yields and CDS spreads on the peripheral euro countries show that jokes are not tolerated any more.

It’s not sure if EBA (The European Banking Authority, who co-ordinates the EU-wide stress test this year) fully gets it: in its 13 January announcement it says that the objective of the stress test is to assess the resilience of the EU banking system to hypothetical stress events under certain restrictive conditions. No, this time the stress event shouldn’t be that hypothetical. Instead, for not being a pointless exercise, the stress test should rely on realistic view of solving the debt problems in euro area, which is more than likely to involve restructuring of at least some of the sovereign debts (or some other mean that effectively reduces the amount that heavily indebted nations need to pay back). The reason is that so far, there is no real solution to the euro area problems, and concerns about the exposures of EU banks to PIIGS countries are of not less importance in delaying tactics, that is giving more money and providing guarantees, and not knowing or even believing that debts will ever be paid back. Only after this conservative base case scenario is tested, we could speak about more hypothetical (although still likely events in a bit longer perspective) events like bursting of China’s real estate bubble within the coming let’s say two or three years horizon with all its consequences to the world economy and banking, incl. Europe.

The expected differences between 2011 and 2010 stress tests include the following:
* In 2011 we should assume at least somewhat tougher stress testing methodology covering both trading and banking books, and more stringent criteria for passing the test (e.g. capital definition close to core Tier 1 rather than less strict Tier 1). This is due to the criticism of 2010 stress test exercise (ShanghaiDaily.com even places European stress test into quotation marks), but also because the implementation of new Basel III rules is closer now. Among others, the EU Internal Market Commissioner Michel Barnier has been cited by Irishtimes.com saying: "We need to learn all the lessons of the first two rounds of stress tests and we need to make these more robust and more credible."
* The macro scenario will look like harsher. In 2010 exercise the adverse scenario GDP, cumulated over the current year and the coming year (2010-11) was close to three percentage points lower than the benchmark (base case) one for the EU and for the euro area. This time the difference between the cumulative GDP decline for the coming two years in the adverse scenario and the cumulative GDP decline for base case will be larger. The reason is simple: starting point in 2010 is stronger than it was in 2009.
* Liquidity will also be tested. Previously, Germany has resisted it, but now German Finance Minister Wolfgang Schaeuble has accepted the idea (although he added that it was not yet clear whether the findings from liquidity assessment would be made public).
* There will be at least some analysis of system-wide risks. According to Reuters, this is what European Central Bank intends to do. Broad risks will be reviewed by European Systemic Risk Board (ESRB).

There are those who believe that the results of 2011 stress test will not be much more reliable than the ones of previous stress tests. This is a very logical thinking: how could it possibly be that European authorities admit their own failure to contain banks and countries? For a reliable result the stress testing exercise should be performed by someone independent, that is by someone outside the system – but outsiders may have too little data and information. What they can do, is to specify and propose the scenario and criteria for passing the test.

A minimum set of “should be” improvements for 2011 is summarised below. This is in addition to the above-described expected changes.
* Firstly, the scenario simulation horizon should be longer. Only current year plus one more year leads to short-termism.
* Secondly, the scenario should be consistent and include a realistic view on how EU sovereign debt problems are going to develop in a given scenario without any further aid to Greece, Ireland, Portugal or any other European country. The assumption of sovereign defaults shouldn’t be excluded from the analysis. Furthermore, the reasoning of the choice of one or another scenario should be provided.
* Thirdly, uncertain assumptions regarding future disposals of a bank’s businesses or other measures planned for increasing the capital should not be included in the main analysis and in deciding of whether the bank is adequately capitalised or not.
* Fourthly, publication of the results, including the results of liquidity stress test is still expected.
* Fifthly, in addition to capital ratios and liquidity ratios, about each bank there should be provided an estimate of how large sovereign shock it would survive without any further help by tax payers, EU or IMF. This would be a valuable input for policy makers while helping them to decide on sovereign debt issues.
As you see, nothing more is asked than common sense would suggest in the current situation.

Stress testing may be valuable but it also may lead to false sense of security. Why not make it valuable this time?

No comments:

Post a Comment