Hipóteses para a banca grega

Free Lunch: An alternative for Greece, por Martin Sandbu (Financial Times):

How to stay in the euro without more loans (...)

 Rather than the end of the world, deposit flight is a problem that has solutions. Three different ones, in fact.

First, if banks are solvent, it should be straightforward for the central bank to refinance any amount of deposits in a bank run. That's its job and carrying it out is the best way to make deposits return. Guntram Wolff has explained well that in principle, the ECB (or rather the Greek central bank through emergency liquidity assistance) can refinance Greek deposits in their entirety.(...)

The second option, then, is to introduce capital controls: legal limits on how much money can be taken out of the Greek banking system. (This and the previous solution are obviously substitutes: the less can be taken out of the system, the less central bank liquidity is needed in a run. In extremis, if nobody can get any money out, no liquidity is needed.) (...)

This would make the third solution more attractive: putting most Greek banks into resolution, to emerge with incontrovertibly solvent restructured banks. Sigrún Davíðsdóttir and Thórólfur Matthíasson have written an Icelandic lesson in bank stabilisation for Greece that deserves wide circulation among European policy makers. (So does Willem Buiter's 2009 "good bank" restructuring proposal - although in theory, bail-in is now official EU policy, Buiter's is the clearest explanation of why it's a relatively easy thing to do.)

The restructuring would split affected Greek banks in two. Collateralised assets would be handed over to secured creditors (the central banks) to extinguish their claims, or transferred to the new "good" entities in agreement with them. The main bad asset to deal with is not banks' exposure to the Greek government (which is small), but non-performing private loans. Go back to the balance sheets: these loans amount to €212bn, about one-third of which the ECB assessed as non-performing in October, or about €70bn worth of loans. Suppose they lost half their value: that's a €35bn loss. Suppose things get worse and they lose their entire value, or equivalently that the NPL ratio doubles. That's a loss of €70bn. How would the Greek banking system fare under those extreme assumptions? Well, it has already set aside €39bn in reserves for losses and it has another €34bn in capital on top of that. You could write down huge losses even before talking about a Cyprus-style bail-in of depositors. 

You'd be left with new "good" banks with deposits backed by a smaller and rock-solid balance sheet and an asset management company working out the bad assets. Even if you needed to find another €10bn or so for recapitalising the new (much smaller) balance sheet, that would involve a smaller bail-in than Cypriot depositors suffered.