According to the version of financial history favoured by America's Treasury and the Federal Reserve, the "stress tests" they did on big American banks in the first half of 2009 were a turning-point in the crisis. By independently examining banks' books, estimating losses, making public the results and forcing banks to raise capital, they restored market confidence.
Reality is more complex than that — the bailouts of Citigroup and Bank of America in early 2009 had just a bit to do with restoring confidence, too. Europeans tend to view the tests as a dazzling piece of PR, followed by brazen trumpet-blowing. One banker says the American experience was both "fantastic and a little depressing". A regulator agrees: "People have a naive view about what stress tests can deliver."
Yet with funding markets almost closed to some of their banks, many Europeans have changed their minds. The difficulties of Spanish banks are well known. Portugal's banks have also been tapping the European Central Bank (ECB) at a furious pace. There is talk that the malaise has spread to Italy. Americans are particularly reluctant to lend. Huw van Steenis at Morgan Stanley says that more banks now accept that an American-style stress test may be needed to restore confidence. Spain has become a passionate advocate and others have signed up, too. On June 17 the European Union's leaders agreed that stress tests should be made public in late July.
The good news is that a process already exists for doing stress tests. The bad news is that it involves the kind of European group-hug that inspires derision, not confidence, in hedge-fund managers. America's tests were run in military style by the Fed, with the Treasury writing the cheques. In Europe the economic scenario is being set by the ECB and the European Commission, which is then interpreted by the Committee of European Banking Supervisors (CEBS), a quango of regulators run with a skeletal staff, which in turn liaises with national supervisors in each country.
Kept private
CEBS did organise some tests last year, on 22 banks, but the results were kept private. It has just finished gathering data on 25 banks in what was meant to be just another routine and confidential test. These data will form the foundation of the souped-up public tests the politicians want, although more banks will be included. Yet given the stakes and compared with the mighty Fed, CEBS is fathoms out of its depth.
A lot depends, then, on national regulators, which do have real muscle and resources. But it is hard to avoid the impression of inconsistency. A supervisor from a medium-sized country says it has played a "pretty passive" role in the CEBS tests, with the banks interpreting what the broad macroeconomic scenario might mean. In another big country, the regulator says it just plugged the CEBS scenario into its model and sent them the results. This is concerning. Daniel Tarullo, a Fed governor, recently emphasised that some American banks weren't much good at doing stress tests on their own and that supervisors' judgment, not just models, was vital. The tests are being taken seriously by regulators in France and Spain, at least, which could pressure others to follow suit.
That still leaves the delicate issue of how to factor in Europe's sovereign-debt jitters. The tests are unlikely to assume a sovereign collapse, with the possible exception of Greece. Imagining anything worse would be tricky politically; it also goes beyond what most officials view as a plausible worst-case scenario.
"I don't need to run a stress test to understand what a default of Italy means," says one. Instead the tests will probably assume slower growth as a result of bunged-up debt markets and get banks to recognise the market price of sovereign instruments in their trading books (but not in their larger loan books). As well as this middle-of-the-road "worst case", the tests will also define banks' capital generously, using Tier-1 capital rather than the more stringent measure of core equity, which the Americans used and new Basel rules endorse.
Marco Annunziata, the chief economist at UniCredit, says a "stress-test lite" won't be much use. Two things would help. The first is a really serious kick of the tyres in Spain. If investors are convinced that all bad debts in Spain's banking system have been revealed and if Spain's government shows it can raise the cash to recapitalise weak lenders, then fears about the country's solvency should recede.
That in turn would make a "no sovereign default" assumption in other countries' tests more plausible. So far Spain's bail-out fund has raised 12 billion euros (Dh55 billion) and faces drawdowns of 10 billion euros. Providing the total bill stays below, say, 50 billion euros, its government should easily be able to raise the cash, even in nasty debt markets.
Hide and seek
Second, the stress tests should ideally provide uniform disclosure on individual banks' exposure to weaker countries, so investors can make their own minds up.
Simon Samuels of Barclays Capital says disclosure has been patchy to date. Furthermore, what firms say is hard to reconcile with the overall estimates. This has led to a game of "who's hiding dodgy debt?", something that is familiar to students of the subprime debacle. A recent study of confidential submissions from over 30 big European banks by Jean-François Tremblay of Moody's, a ratings agency, concluded they had exposure to Greece of 100 billion euros. Based on data from the Bank for International Settlements that leaves another 90-odd billion euros sitting in Europe's banks, presumably smaller ones.
It took a decade before Japan's supervisors lost patience and published their estimates of banks' aggregate bad debts in 2002 and 2003 (prompting banks to admit to more losses). As in Japan, the banks and regulators have lost the market's full confidence. Unlike Japan's banks, European banks rely heavily on wholesale borrowing that needs to be refinanced now. Europe will have to bite the bullet soon.
