The missing link between the Greek solvency problems and its membership of the eurozone is called liquidity. Absent liquidity problems, for example under 100% reserve banking, the two issues would be entirely disconnected, since the banking system would never need emergency liquidity assistance from the ECB.
I believe the doctrine of the ECB underlying ELA, is that ELA deals with a technical (as opposed to political) issue (The way ELA works is nicely described in this piece by The Economist). And this technical issue is what we call multiple equilibria.
According to this view, there exists a bad equilibrium where expectations of exit from the Eurozone and bank runs mutually reinforce each other. Under normal circumstances, banks can get fresh liquidity from the ECB at the going interest rate (which is very low), in exchange for collateral. This collateral has to be “good”. Assuming Greek banks hold Greek debt, the collateral becomes “worse”, the less solvent the Greek government. To the extent that Greek bonds are denominated in euros and owned by non-Greek residents, Grexit would either trigger a redenomination of those bonds in a depreciated drachma, or make it more difficult for Greece to pay them back, since it will presumably experience a substantial real depreciation. Therefore, expectation of Grexit makes Greek sovereign debt worse and Greekn banks less solvent. As depositors fear that their own bank could be liquidated, they run to the bank to get their cash back. Running to the bank delivers not a double dividend, nor a triple dividend, but a quadruple one:
First I am protecting myself against the bank’s bankruptcy. True, my deposit may be insured, but if the bank run is general the insurance system is undercapitalized and I may not get my money back.
Second I am converting the virtual euros of my bank account, that may be converted into drachmas any time, into actual euros that are physically the same as those circulating in Germany, and are much more difficult to convert except by extreme coercion.
Third, I am certain to have cash when I need it, instead of being rationed by bank holidays and controls on withdrawals.
Fourth, I am protecting my savings against a surprise capital levy that could be imposed by the troika on deposits in order to pay back creditors.
When faced with the bank run, the ECB has to do something. If it refuses to provide liquidity, say because the collateral is no longer acceptable, then the banking system in Greece collapses. And in order to provide liquidity to its banks, the Government will have no choice but to exit the Euro, making Grexit a self-fulfilling prophecy. The bank run equilibrium is a Grexit equilibrium, indeed expected Grexit acts as a catalyst for the bank run by reducing the market value of bank assets.
But it makes no sense for the ECB to value the collateral of Greek banks at its price under the Grexit equilibrium, if instead a better equilibrium can be enforced: that is, an equilibrium where the bank run does not occur. ELA is an attempt to enforce the good equilibrium by setting interest rates and haircuts at levels consistent with the good, not bad, equilibrium. If the good equilibrium is indeed an equilibrium, sufficient liquidity should be provided under those terms so as to “tame” the bank run and keep Greece within the Euro area.
The problem with this fairy tale is twofold.
First, technicality is not so easily distinguishable from politics. We see that the ELA tap is being shut by the ECB, because the Greeks are not accepting the conditions set by the troika. The banking crisis in Greece, initiated by the ECB, may bring the government down. There is no way Tsipras can decide, as the democratically elected ruler of Greece, to provide liquidity to banks despite the ECB veto, unless he exits the Eurozone immediately. Voters on Sunday may support the Yes vote because they believe banks will reopen sooner than if No wins. That is, the ECB can bring down a democratically elected government by stopping providing liquidity to its banks. In that respect, the similarity between Alexis Tsipras and Salvador Allende is striking; in both cases we have radical leftists heading a minority government. The difference stops in that instead of having a military coup (Pinochet and the CIA) we have a monetary coup (Draghi and the Troika/Eurogroup), and instead of the trucks blocking the roads, we have the banks blocking the transactions.
While closing the liquidity tap inevitably is a political coup against the Tsipras regime, keeping it open is another forced transfer from European taxpayers to Greek residents. In other words, regardless of the decision, the non-elected ECB cannot avoid doing politics.
Given the fiscal situation in Greece, the terms under which the ECB is accepting Greek bonds as collateral under ELA are probably too generous. Under ELA the ECB lends to Greece at 1.5 % and the haircuts are secret (See here and here on haircuts). Under the assumption that the ECB is successful at killing the bad equilibrium, the terms under which a Greek bank can borrow from the ECB should be the same as those in the market. Since Greece is tapping ELA it means that these terms are better, hence that there is a transfer. The transfer is larger, the larger the gap between the market prices the ECB is trying to achieve and actual market prices. If the ECB were fully credible in extending an infinite credit line to the Greek banking system, this gap would disappear. If the good equilibrium exists, the fiscal transfer to the Greeks is the mirror image of the ECB’s limited credibility and of the market view that the probability of getting to the good equilibrium is lower than 100%. The transfer, and the likelihood of the bad equilibrium, could then be eliminated by making the credit line infinite, killing any doubt by markets about the viability of Greek banks.
But, if the good equilibrium does not exist, Hans-Werner Sinn is right: The Greeks will take advantage of ELA to sell all the Greek debt to the ECB at an inflated price, converting their deposits into fresh and good euros, and leaving the ECB, that is, the taxpayers from solvent countries, with dubious claims on the Greek banks’ balance sheet. An unlimited credit line can no longer kill the bad equilibrium since it is the only one. It would simply raise the amount of money transferred to the Greeks by other euro area residents.
That the bank of Greece, not the ECB, is supposed to bear all the risk of ELA, is a cosmetic provision: ultimately, the bank of Greece accounts are implicitly consolidated with those of the bankrupt Greek government.
And it is hard to believe that the good equilibrium exists. That is, if the Euro were made as big a taboo as incest, the quadruple dividend from getting your money back would just become a triple dividend. Think about it. The benefits of having your money in a bank deposit rather than cash are quite small. The reason we keep it there is because it is the default option for most wage earners (they call it libertarian paternalism). So you only need a small cost of having your money deposited in a bank for it to be rational to withdraw your money. The asset side of the Greek banking system provides ample reasons to justify a bank run.
(The reason why the “good equilibrium” has existed for so long is that banks have been regularly provided with fresh cash by central banks, bailout money by governments, and deposits have been insured. But we may ask whether this form of financial intermediary makes sense. Perhaps the financial crises tell us that it is an evolutionary dead end and that banks should be eliminated. By saving them each time, public authorities are fighting against economic evolution and sowing the seeds of the following crisis.)
Post Scriptum — Here is an interesting piece by Dirk Niepelt. Clearly, the reason why governments would like to eliminate cash has everything to do with exercising discretionary control upon private savings — so as to implement a capital levy when needed, as advocated by Mrs Lagarde — and very little to do with “combating crime”. Virtual cash held at the monetary institute is also vulnerable to capital levy. In the long run, suppression of cash by heavily indebted government will eventually harm them, as private means of payments (like bitcoin) will arise.