EU stress tests: for banks or governments?
- Laurence Copeland is a professor of finance at Cardiff Business School. The opinions expressed are his own.-
Worries about Europe‚Äôs banking system go back at least to 2007, but whereas the U.S. (and UK) banks appear to have weathered the storm, there are fears that for European banks the worst may lie ahead.¬† Concerns centre on four areas.
First, there are obvious worries about Greece and the other small countries facing debt problems, notably Portugal and Ireland, where the local banks have lent heavily to their governments and in addition may need to make provision for a substantial build-up in the level of bad debts in their respective corporate sectors as their economies struggle through the recession.
Second, there are worries about the small-to-medium banking sector in Germany, where some of the first signs of the oncoming crisis appeared early in 2007. It is hard to tell how seriously we should treat these concerns, because the Landesbanken are closely linked to their regional (‚ÄúLand‚ÄĚ) governments, so the question is unusually sensitive.
Third, there are worries about the European giants, especially the big French and German banks.
Not only is it still unclear (to me, at least) how badly hit they were by Lehman and its aftermath, it is still a matter of conjecture how much sovereign debt they are holding.
Fourth, there is the enigma of Spain, worth a blog on its own. The bald facts about Spain are frightening ‚Äď 20 percent unemployment (and nearly as much even before the credit crunch), the economy most dependent on construction of any in Europe, a large budget deficit, tourism suffering from the strong Euro.
Yet the big Spanish banks have emerged as European champions, sweeping up a couple of basket cases so as to establish a presence on the British high street and expanding their balance sheets dramatically on a number of fronts.
No doubt this is in large part a rare, almost unique case of regulatory success ‚Äď Spanish banks were never allowed to use off-balance sheet vehicles so as to enter the deadly U.S. mortgage market.
Nonetheless I suspect I am not alone in wondering what happened to all the loans that financed the property bonanza along Spain‚Äôs costas, by some measures the biggest real estate boom of all, and one of the first to go bust.
This must have been largely a story of local cajas, the regional savings banks with their political ties and their overblown ambitions, but if they were indeed the only culprits, many of the loans on their books must be of dubious value by now.
These are the background concerns which the EU stress tests are intended to address ‚Äď but don‚Äôt hold your breath.
The intention is to calm the markets in the same way that the Fed‚Äôs tests on the U.S. banks appear to have done in 2009, but they have been bedevilled from the outset by two problems which the Americans did not have to face.
First, there appears to have been some kind of conflict between Brussels and national regulators which has been settled by an unsatisfactory compromise ‚Äď a muddle, just when clarity is most needed.
It seems that, instead of a normal publication process, the results presented on July 23rd will be for 91 banking groups (‚Äúconsolidated results‚ÄĚ), with detailed deconsolidated results covering their subsidiaries to be released at the discretion of national regulators two weeks later.
This bizarre two-stage procedure itself inspires little confidence in the transparency of the exercise.
Second, and far more serious, is the problem of how to handle sovereign debt. As I have pointed out before, Germany was forced into bailing Greece out because its own banks had lent it so much that a default might have triggered a new European, or even global, banking crisis.
The situation arose in the first place because, ever since the Eurozone was set up, European regulators have refused to recognise any quality distinction between the debt of different member countries.
The question is: will they now finally bite the bullet? And if they do, how will they price sovereign debt on bank balance sheets, given that face values are no longer realistic?
One possibility would be to present balance sheets with government bonds marked to market i.e. priced off currently quoted prices or yields, or even credit default swaps.
At first blush, this approach is attractive, but on closer consideration it is highly problematic, since many of the relevant markets are incomplete and sometimes illiquid, and in any case the stress tests work by comparing a number of scenarios: what would happen to a bank‚Äôs balance sheet if GDP fell by 1 percent, 3 percent . . . and/or if the price level fell by 2 percent . . . ? and so on.
These are questions which cannot be answered simply by looking at current market prices, which in any case — we should not forget — are based not only on existing levels of support for sovereign debt from the European Central Bank and the IMF, but also on the assumption of some indeterminate level of support in the future.
That makes them singularly unsuitable for an exercise that seeks to estimate the true ability of banks to stand on their own two feet.
The only alternative seems to be model-based pricing i.e. using prices derived from an economic model which can be cranked up to generate scenarios of the kind which appear likely — or at least possible — over the next year or two.
Taking this approach means the results of the stress tests are only as good as the models used to generate the scenarios.¬† How good is that?
Having had decades of dispiriting experience of modelling financial markets, and of reading the depressing results achieved even by the superstars of the economics profession, I can hardly be optimistic about the prospects.
After all, as someone observed a long time ago, economic models are like sausages ‚Äď you‚Äôre best not to look too closely at what goes into making them.