Just a Flesh Wound
'Doctor, how am I? Tell me the truth.'
'Well, you have a mild case of cardiac arrhythmia, your cholesterol is about thrice of what it should be, your blood pressure is off the scales, and if I'm not mistaken, there's a spot of beginning, how shall I put it? Kidney and liver failure. Alas, unless your heart actually stops beating, I think you'll be fine. Of course that brain tumor might get you as well, but a committee of doctors is currently busy solving that particular problem, so we can safely ignore it for the time being. As causes of death go, it's too improbable anyway, right? I therefore pronounce thee to be in ruddy health. Take two aspirin and call me tomorrow.'
We're not quite sure what the EBA's (European Banking Authority) stress test of European banks was supposed to achieve. As far as public relations exercises go, the effort is a monumental failure. Recall that the first European bank stress test was immediately revealed as a sham when a number of Irish banks that had passed it collapsed a mere few weeks later. The results of the first test had of course been received with widespread incredulity, which the insolvency of Ireland's banks soon vindicated. This time the authorities attempted to deliver a more rigorous test, one that would produce enough failures to be credible, but not so many as to induce a panic. As we noted in 'Hurrah, We're Saved Again' in late June, these goals were formulated well in advance, with anonymous sources from the ECB quoted by Reuters noting that:
In the drive for credibility, the EBA, which runs the tests and the ECB, which sets the economic scenarios, have pushed for more banks to fail than last year's seven.
"How many do we expect to fail? I would say 10 to 15," said one senior euro zone central banking source.
The EBA wants the number of banks that do not pass the tests to be around that level to show the examinations were serious, said a second source, adding the authority did not want to push for more, for fear it could spark panic.
"In order to demonstrate that it is credible, the EBA would need to show that the number of bank failures is significant, without being substantial," said the source. "A number in the teens is about right."
Given such advance comments, the desired credibility was likely already lost before the publication of the stress test results. The actual results that have now been delivered quite possibly will serve to damage it further. Of 90 banks tested, only eight failed to pass the test's main requirement of keeping a core tier one capital ratio of 5% intact under the proposed 'stress' scenario. An additional 16 banks fell in the 5%-6% range, which is to say, they almost failed to pass.
We should interpose here that the stress test was certainly not an entirely useless exercise, even though it will likely fail to boost confidence. Its usefulness consists mostly of the fact that it forced the banks concerned to reveal their various exposures in great detail, which makes it easier for outsiders to analyze their situation (we are sure we will have occasion to refer back to these data).
The immediately obvious conclusion is that even by the not very onerous assumptions of the stress test, some 24 out of 90 banks, or nearly 27%, must be regarded as being in danger of failing if economic conditions worsen markedly.
Let us suppose that the worst case comes to pass and these 24 banks do in fact fail. Would the stress test results then still be applicable to the banks that passed? After all, one of the main problems the banking system faces is its interconnectedness. There is e.g. the ever present 'Herstatt risk', or more generally, settlement risk (the term stems from the 1974 failure of the German Herstatt bank, which was a big player in the forex markets and left its counterparties dangling when its banking license was one day withdrawn after the market close in Germany). When banks stopped trusting each other in the 2008 crisis and interbank lending rates exploded, it was precisely because of a perceived increase in this counterparty risk. As has been seen in the period August 2007 to March 2009, central banks will pump 100ds of billions into the system to avert a scenario of cascading cross-defaults when the payments system becomes paralyzed. As money supply data from this period prove out, it is not possible for these interventions to be neutral in their effect on money supply growth.
In addition, it is not possible in a fractionally reserved banking system for banks to pay all or even a large fraction of their extant demand deposit liabilities to depositors in case of a bank run. The legal privilege that allows banks to treat money that has been warehoused with them as a 'loan to the bank' which they can use for their own business activities means that depositors are exposed to a risk they have in reality not contracted for. Government guaranteed deposit insurance schemes have been set up everywhere, enticing depositors to ignore this risk and enabling commercial banks to engage in nearly unchecked expansion of credit and deposit money. It goes without saying that when governments themselves are threatened with going bankrupt, many of these deposit insurance schemes are not worth the paper they are printed on.
Why The Stress Test Is A Farce
The biggest problem with the stress test is that it glossed over what is currently the by far biggest problem for the European banking system – namely the collapse in peripheral sovereign bonds and the very real risk of one or even several sovereign defaults occurring in the not-too-distant future.
In short, the number one problem was treated as though it didn't exist, or as though it were only of marginal importance. There was no choice but to treat the problem in this manner if the test was to hew to the current 'party line' of the eurocracy, which holds that no sovereign defaults will be 'allowed'. This is considered an important plank in keeping up confidence in the fractionally reserved banking system and the mountain of government debt that is deeply intertwined with it.
In the stress test, regulators differentiated between two sets of books – the 'trading book' and the 'banking book' of the tested banks. Not surprisingly, the vast bulk of government bonds held by the banks resides in the 'banking book', where it is presumed that they are held to maturity. As long as a sovereign debtor is not officially in default, it is further assumed that these bonds will be redeemed at par. The haircut applied to this exposure was therefore zero. After all, in the official version of the future, there simply won't be a sovereign default (the EBA does demand provisioning against risk in the banking book, but as noted further below, these provisions are of negligible size).
Haircuts were however applied to the banks' trading books. As an example of how far removed from reality these haircuts are, consider that the biggest discount – applied to Greek government debt held in trading books – was 25%. Greek government debt meanwhile is trading at 50% and more below par in the markets. In other words, not even the sovereign bonds held in trading books were properly marked to market.
Of € 377 billion in loan loss provisions held by the 90 banks in total, only € 11.5 billion thus relate to their sovereign debt exposure. As a reminder, the total exposure of European banks to Italy's government debt alone amounts to € 1.2 trillion. € 11.5 billion represents not even one percent of this amount, basically it is what the banks could lose in the first ten to fifteen minutes of a bad hair day in the bond markets.
Provisions for loan losses per the EBA stress test, for various types of assets held by the tested banks. Only slightly below € 11.5 billion have been reserved for sovereign credit risk. This seems bizarre given the current state of the sovereign debt crisis.
