It was another of those late-night agreements, which are as legally sophisticated as they are financially innumerate. Euro zone finance ministers agreed last week that the European Stability Mechanism could devote €60 billion of its €500 billion total lending ceiling to the recapitalisation of euro zone banks. If that is not enough, the rest of the money for the recapitalisation of the euro zone’s banks will have to come from national governments, or through bail-ins of investors and depositors.

So how much recapitalisation can we expect? The French newspaper Les Echos last week produced an estimate of the total assets of the bad banks that have been set up to absorb losses from the housing crash and the US credit crisis. That estimate alone is more than €1 trillion – although it includes the UK. How much of these assets are ultimately underwater is anybody's guess. But you could safely assume that quite a lot of this stuff will ultimately be worthless. This estimate only relates to the bad banks. You have to add actual and hidden losses from the rest of the banking system. We do not know how big these are since hidden losses are, by definition, hidden. To disguise a loss, banks use tricks such as "pretend and extend": lenders can decide to renew a non-performing loan that technically becomes good again the moment the new loan is struck.

Hidden losses
The reason I believe the amount of hidden losses in bank balance sheets is ultimately quite large is the sheer number and scale of the accumulated crises during which European banks managed to lose money in recent years: the US subprime crisis, a euro zone housing bubble, the Greek debt restructuring, the Cypriot bank failures and the short and sharp 2009 recession followed by the Great Recession of 2011-13, with no end in sight for southern Europe.

One would hope that an asset quality review by the European Central Bank, envisaged for next year, would provide clarity. But I am doubtful. In the past, bank transparency exercises were undertaken with the intention of hiding the truth. Remember the stress tests of 2011? Or the apparently independent audit of the Spanish banking system, which concluded that Spanish banks only needed a teeny weeny bit of new capital?

What makes me specifically doubtful about the ECB’s exercise is that I cannot see the central bank conceivably coming up with a number that is larger than the available capital.

So instead of waiting for these estimates, here is a quick and dirty, back-of-the-envelope calculation. It has an error margin approximately the size of the Italian economy, but it nevertheless produces an order of magnitude in which to think about this problem.

The total balance sheet of the monetary and financial sector in the euro zone stood at €26.7 trillion in April this year. How much of this is underwater? In Ireland, the 10 largest banks accounted for losses of 10 per cent of total banking assets in that country. The total loss will be higher. In Greece, the losses have been 24 per cent of total assets. The central bank of Slovenia recently estimated that losses stood at 18.3 per cent. In Spain and Portugal, the recognised losses are already more than 10 per cent, but the numbers will almost certainly be higher. Non-performing loans are also rising rapidly in Italy.

Germany is an interesting case. The German banking system appears healthy at first sight. It certainly fulfils its function of providing the private sector with credit at low interest rates. But I still find it hard to believe that the German banking system as a whole is solvent.

On their own, the euro zone’s bad banks constitute about 5 per cent of euro zone banking assets. If you add another 5 per cent from hidden losses, the losses still being generated by the double-dip recession, and future losses through the bail-in of investors, you arrive at €2.6 trillion. Not all of these losses will have to be made good through a recapitalisation. Some banks may have some capital reserves. Other banks may be closed. But that just distributes the losses from one end of the banking sector to another.

Pretend and extend
Assume now that my estimate is wildly wrong, and deduct the size of the Italian economy from that back-of-the-envelope number. You still end up with €1 trillion. With this order of magnitude it mattered relatively little whether the ESM could contribute €60 billion, €80 billion or zero. Europe's national governments are clearly incapable and unwilling to fill the gap. And without the money for bank resolution, it barely matters whether the European Commission will become the resolution authority that does not do the job or whether someone else does not do it.

That leaves a long period of regulatory forbearance as the most likely outcome – a policy version of pretend and extend. They pretend not to see the losses, and extend the crisis.

munchau@eurointelligence.com