European Banking Authority turns to disclosure to save latest stress tests

The Wall Street Journal ran an article on the latest European bank stress tests.  In an attempt to make the results credible, data on banks' credit and sovereign exposures will be published.  The hope is that market participants will find this data sufficient to run their own analysis and confirm the findings of the stress test exercise.

As this blog has said repeatedly, what ultimately instills confidence in the financial markets is disclosure that allows market participants to run their own analysis.

The idea that the regulators should run stress tests as a means of instilling market confidence is fundamentally flawed.  Regulators should run them for their own internal purposes, but never for publication.

Why?  Because publishing the results of the stress tests has the potential to make it look like regulators do not know what they are doing.  Look what happened with the last round of European stress tests (shortly after completion, banks that passed the test in Ireland were nationalized).

The idea behind limited disclosure of banks' credit and sovereign exposure is also flawed.  Once the decision has been made to disclose these exposures, there is no logical stopping point before asset-level disclosure.  All market participants know that data summaries mask a multitude of sins.

It is far better for the regulators to give up their information monopoly and require the banks to disclose their current asset and liability-level data to all market participants (including competitors and credit and equity market analysts).  This way, market participants can run stress tests on the banks and the regulators can piggy-back off of their efforts.

The European Banking Authority said Friday that it will publish the results of this year's round of stress tests on European Union banks July 15. ... A full, bank-by-bank breakdown of the results, complete with data on banks' credit and sovereign exposures, will be published. 

... As reported, some 91 banks and banking groups have been subjected to the tests this year, and have to show that they can keep their level of core tier 1 capital above 5% of risk-weighted assets, even in the event of a severe economic slowdown and renewed market stress over the next two years. 

... Public attention to the stress test process this year has again concentrated on the simple "pass" or "fail" process. However, authorities have stressed that the greater importance of the exercise lies in providing transparency on the risks and potential weaknesses of the banks taking the test, allowing the market to draw its own conclusions as to their long-term stability.

The Financial Times carried an interesting article that provides far more color on the latest European bank stress tests.

As bridge designers have learnt, the best way to prevent disaster is to implement effective stress-testing that takes into account not just individual actions but also herd behaviour. That is just what the European Banking Authority, the European Union’s new regulator, aims to accomplish. 

Next week, it is set to publish the long-awaited results of an ambitious exercise to stretch the balance sheets of 91 banks in 23 countries almost to breaking point. 

The results are being closely watched – not only by banks themselves but also by investors and policymakers. How many institutions will fail? How many will just squeak by the pass mark that requires them to maintain top-quality “core tier one” capital equal to 5 per cent of their assets, even after a disaster scenario? How much capital will those banks be forced to raise? 

The answers, and the capital and risk analysis that comes with them, could have a profound effect both on share prices and governments’ attitudes to their banks. 

Among the most scrutinised information will be unprecedented disclosure of loan portfolios and holdings of sovereign debt, including maturity profiles – particularly timely given that Greece is teetering on the brink of default and that any sovereign debt restructuring could have big ramifications for the French and German banks that are the leading holders of Greek government bonds. 

Regulators say the exercise is vital to force banks to be honest about potential risks. “It’s about overcoming disaster myopia in some firms,” says one. 

For years, banks have run their own internal tests that model for isolated stresses: how the investment portfolio would withstand a 40 per cent slump in share prices, say; or what a 2 per cent rise in unemployment would do to the likelihood of homebuyers defaulting on their mortgage repayments. 

As computer modelling improved, regulators began to use their own scenarios to evaluate soundness. But only since the financial crisis has stress-testing become a vital part of the regulatory arsenal. 

Attempts by increasingly sophisticated tests to model the effects of simultaneous, interdependent changes reflect the big lesson of the financial crisis – that systemic risk resulting from the interconnectedness of an increasingly complex financial sector can be economically disastrous. 

21st century information technology is not limited to the regulators, but exists in abundance with market participants.  If there were current asset and liability-level disclosure, market participants would regularly stress test counter-parties and adjust the pricing and amount of their exposure to the counter-parties based on the results.

As the US showed two years ago, tough transparent testing can also reassure jittery investors, particularly important with the troubles of the eurozone’s peripheral nations continuing to dominate the news.

As previously discussed, the "success" of the US stress tests was a function of Treasury pledging the full faith and credit of the US to support the banks, not anything reported as a result of the stress test.

This pledge is not an option for Europe as the size of the national banking sectors exceeds the capacity of each country to borrow enough money to bailout the banks headquartered in that country.

The EBA’s aims are twofold: first, to strengthen the system by pushing banks that are thinly capitalised relative to their underlying risks into raising fresh equity: and second, to convince the world about that strength – helping investors, particularly from outside Europe, to differentiate among eurozone banks and stop shunning them en masse.

Said another way, to answer the question of who is solvent and who is not.

... Few are convinced that the EBA can dispel such fears. Last year’s debut test in Europe, administered by the Committee of European Banking Supervisors, the EBA’s more flimsy predecessor, was discredited within months of completion. It gave a clean bill of health to all but seven of the 91 banks tested, identifying an aggregate capital shortfall of only €3.5bn ($4.7bn), way below the €30bn-€40bn estimates touted by bank analysts. 

The results buoyed market sentiment initially. But by year-end Irish banks that passed had to be rescued as part of a massive EU and International Monetary Fund bail-out. As the bridge engineers might say, it was the “wrong sort of testing”. 

This year’s test has also raised concern. At first, it seemed it would make similar mistakes to its predecessor. When the scenarios emerged in March, several were actually more benign than a year ago – including a 15 per cent equity markets fall, rather than the 20 per cent used in 2010, and a more benign economic environment. 

The timing is ominous, too. With Greece at risk of sovereign default, any test result that underplays that likelihood will be dismissed by the markets as puff. “If there is even a selective default [of certain Greek bonds], the test result could end up looking horribly like the Irish scenario last year,” admits one European government adviser.

As discussed above, this is why the regulators should make sure current asset and liability-level data is disclosed to all market participants and not be putting their credibility at risk by disclosing the results of stress tests they run.

... A few months ago, however, hopes of a credible outcome looked slim. In April, when the first round of results began trickling in to the EBA’s modest headquarters in a London skyscraper, Mr Enria’s heart sank. With the process of testing delegated to national regulators and then in turn to the banks themselves, he had clearly left them too much scope to make optimistic assumptions and exploit loopholes. Crucially, most had assumed there was no possibility that Greece would default. 

The EBA decided it had to come down hard to retain credibility. It sent out new guidance in June requiring all the banks to run their data again to address what Mr Enria has tactfully referred to in public as “inconsistencies and excessive optimism”. 

The letter made clear that strategic changes other than capital raising made after December 31 2010 could not be used to sweeten the results. The EBA also warned banks to rethink the impact of a sudden rise in interest rates – many had assumed they would benefit from bigger lending margins, rather than suffer as a result of market turmoil.

It was on sovereign debt, however, that the EBA’s letter announced its most significant change. Banks were told to assume specified potential downgrades in sovereign credit ratings, spelling out, for example, that debt rated triple C – the lowest possible before default, and the rating now applied to Greece – had a 36 per cent chance of defaulting and that they should model for losses of 40 per cent if that happened. 

The clampdown was Mr Enria’s way of addressing the biggest criticism of the tests – that political concerns in EU institutions meant the EBA could not be seen to countenance a default of a eurozone sovereign borrower such as Greece, despite the growing possibility of just such an event. 

By insisting on the rating agency methodology, the EBA was able to maintain that it was not breaching the taboo against imposing a formal “haircut” – or enforced loss – on the value of Greek sovereign bonds. But the tactic essentially had the same effect – to the dismay of Greek banks, several of which now look more likely to fail. Markets, on the contrary, are heartened by this more stringent approach. 

The challenge now is to force capital raisings on banks that fail or come close to doing so. European governments have already pledged that by the time of publication next week they will have concrete state-backed mechanisms in place to forcibly recapitalise banks within six months. That is another vast improvement on 2010, when the test’s limited bark came with no bite at all. 

Exactly how credible will a state-backed mechanism for Greek banks be?  Yet another reason to opt for disclosure and keeping the results of regulator performed stress tests to the regulators.