With the Cypriot government still 'undecided' about what to 'take' and the European leaders very much 'decided' about what to 'give', the fact of the matter is, as JPMorgan explains in this excellent summary of the state of affairs in Europe, that because ELA funding facility is limited by the availability of collateral (and the haircuts applied to those by the central bank), and cutting the Cypriot banking system completely from ELA access is equivalent to cutting it from the Eurosystem making an exit from the euro a matter of time. This makes it inevitable that capital controls and a capital freeze will be imposed, in their view, but it is not only bank deposits that are at risk. A broader retrenchment in funding markets is possible given the confusion and inconsistency last weekend's decision created for investors relative to previous policy decisions. Add to this the move by Spain, which announced this week a tax or bank levy (probably 0.2%) to be imposed on bank deposits, without details on which deposits will be affected or timing, and the chance of sparking much broader deposit outflows across the union are rising quickly.
Capital Control Risks
What was widely viewed as an ill-conceived Cyprus deal last weekend renewed fears of a re-escalation of the euro debt crisis. The original proposal to hit insured depositors below €100k caused a bank run and set a new precedent in the course of the Euro area debt crisis, with potential negative consequences for bank deposits not only in Cyprus but also in other peripheral countries. Once again, as it happened with the Greek crisis last May, the Cyprus crisis exposes the fragmentation of the deposit guarantee schemes in the Euro area and its inconsistency with a monetary union.
Even if the original deal is eventually revised and the guarantee for depositors with less than €100k is respected, the damage from the original proposal will be difficult to undo, in our view.
Cypriot banks are relying on ECB’s Emergency Liquidity Assistance (ELA) to avert a collapse once they open next week. ELA reflects collateralized borrowing from the national central bank rather than the ECB directly, not only at a more punitive interest rate relative to refi rate but more importantly with much larger collateral haircuts. The ECB is still on the hook under ELA because the national central bank borrows these funds from the ECB, i.e. it generates a liability against the Eurosystem. The ECB’s provision of liquidity via ELA is admittedly not a given but it will be provided to Cypriot banks for as long as Cyprus is looking to finalize its revised bailout plan, the so called Plan B.
Although the ECB always states that it provides liquidity to only solvent and well-capitalized institutions, past experience with Irish and Greek banks and even with Cypriot banks shows that the ECB has tolerated long periods of liquidity provision to undercapitalized institutions. Greece is the most characteristic case. Greek banks had access to ELA even when the bank recapitalization was pending between April and December 2012. And Greek banks had access to ELA in-between the two Greek elections when it was not even clear whether Greece would stay in the euro. Cutting the Cypriot banking system completely from ELA access is equivalent to cutting it from the Eurosystem making an exit from the euro a matter of time. This is a political decision rather than a decision that the ECB can take alone. This would effectively cut the Central Bank of Cyprus off from TARGET2 and force it along with the Cypriot government to eventually issue its own money.
But even assuming that a new deal is agreed between Cyprus and the Eurogroup and ELA continues for the Cypriot banking system after Monday, this does not mean that this ELA is unlimited. ELA is limited by the availability of collateral and the haircuts that the central bank applies to this collateral. The Greek case is the most characteristic example of how punitive haircuts on ELA collateral can be. As of the end of January Greek banks used €122bn of collateral to borrow €31bn via ELA, i.e. an implied haircut of 75%. In contrast, they borrowed €76bn via normal ECB operations posting collateral of €97bn, i.e. the implied haircut on their normal ECB borrowing was 22%. The higher haircut on ELA collateral i.e. is mostly the result of the lower quality of this collateral, typically credit claims, vs. that accepted in normal ECB operations, typically securities. But it perhaps also reflects the higher riskiness the ECB sees with its counterparty, i.e. the national central bank and eventually the sovereign, when a country's banking system has to resort to ELA.
Because of the recapitalization issue which has been pending since last April, post the Greek PSI, Cypriot banks had been steadily losing access to normal ECB operations and had been increasing their reliance on ELA steadily since then. By November 2012 Cypriot banks had access to ELA only. This ELA borrowing peaked at €10bn last November and stood at €9bn as of the end of January.
What is the maximum ELA borrowing for Cypriot banks? Looking at their assets, Cypriot banks had €72bn of loans to non MFIs as of the end of January, roughly equal to total non-MFI deposits of €68bn. Assuming that all these loans are acceptable as ELA collateral with the same average haircut as in the case of Greek ELA, i.e. 75%, results to only €18bn of total ELA. Given that Cypriot banks have already €9bn via ELA, this leaves them with another €9bn of potential additional ELA. Of course the ECB could be more lenient with its ELA haircuts with Cypriot banks relative to Greek banks, and indeed even in the case of Greek banks, ELA haircuts appear to have been as low as 50% at certain points of time during 2012. But we doubt that total ELA could exceed €30bn, which represents more than 40% of the loan assets of Cypriot banks. In the case of Greek banks ECB reliance never exceeded 40% of total loan and security assets. So further liquidity support from the ECB seems limited, and not enough to offset the €21bn of non-euro area deposits with Cypriot banks, largely Russian (80%) and British (20%) and the €5bn of deposits with other euro area residents outside Cyprus.
This makes it inevitable that capital controls and a capital freeze will be imposed, in our view, even if a deal is reached by the end of the week, to prevent depositors, especially non-domestic depositors, fleeing the country. Article 63 of the Treaty on the Functioning of the European Union prohibits “all restrictions” on the movement of capital between Member States and between Member States and third countries. But there would be certain exceptions for measures justified on grounds of public policy or security, see Article 65 of the Treaty on the Functioning of the European Union. But even if allowed in exceptional circumstances, these capital controls and capital freezes are contagious and appear inconsistent with a monetary union.
The obvious risk is the impact that these capital controls will have on deposits in other peripheral countries. Large deposits, above €100k, and uninsured deposits are mostly at risk as these are the ones to be likely frozen in the Cypriot case. While a modest deposit tax might be acceptable to large depositors, a freeze of deposits for an un identifiable time period would likely be unacceptable to most large depositors such as corporations and institutional investors.
There are no recent data of how big this universe of large deposits is. Data from the European Commission suggest that in 2007 large deposits of above €100k and uninsured deposits comprised more than half of all deposits in peripheral countries. See Figure 1. The current shares are perhaps different from those reported in Figure 1 for 2007, but most likely the share of large or uninsured deposits is likely to be close to half of total deposits.
What cushions other peripheral countries relative to Cyprus is that these large deposits are mostly domestic. As explained in the next section, the share of non--domestic deposits in peripheral banks is rather modest at 7% as of the end of 2012.
But it is not only bank deposits that are at risk. A broader retrenchment in funding markets is possible given the confusion and inconsistency last weekend's decision created for investors relative to previous policy decisions:
1) In the case of Cypriot banks, depositors are hit while senior bond holders are spared, so seniority is not respected.
2) Deposits of foreign branches are protected while deposits of domestic branches are hit. This is the opposite of what happened to Iceland.
3) In the case if Ireland which also had a big banking system relative to the size of its economy, only sub debt holders, accounting for a very small portion of total creditors, were hit. No depositors were hit, in either domestic or foreign branches.
4) In the case of SNS sub debt holders were wiped out and reports suggest that the Dutch government came close to imposing losses on senior bond holders and was only prevented from doing so because of unsecured intergroup loans between SNS bank and Reaal insurance that would be subjected to the same losses as senior bond holders.
But beyond the confusion and inconsistency, all these trends and the case of Cyprus in particular, are not only showing bailout fatigue on the part of creditor nations, especially in Netherlands where economic conditions have been deteriorating rapidly, but they are also pointing to a shift towards bailing in private creditors in future sovereign bailouts or bank resolutions to avoid using taxpayers’ money.
Which funding markets do we need to track going forward? In our view, the excess cash in the Euro area banking system is the most important metric to track on a high frequency, daily, basis. This metric reflects the amount euro area banks borrow from the ECB in excess of their normal liquidity needs due to reserve requirements or autonomous factors. A loss in deposits or a loss in funding in wholesale markets forces banks to either access ELA or the Marginal Lending Facility at any time or, in less urgent situations, to access the standard weekly Main Refinancing Operation (MRO) every Tuesday. So euro area banks can borrow from the ECB and the excess cash in the euro area banking system can rise at any day of the week and not only with Tuesday's MRO. Any potential increase in ELA, such as from Cypriot banks, is reflected in the excess cash in the Euro area banking system via a decrease in autonomous factors rather than an increase in outstanding operations. The excess cash in the euro area banking system actually declined this week, with a decrease in outstanding operations and an increase in autonomous factors, indicating no signs of broad contagion yet.
In terms of the impact on wholesale bank funding markets, we can also track peripheral bank debt issuance directly. This week peripheral banks issued only €600m of bonds vs. €4bn in the previous two weeks. The represents a marked slowing, suggesting that Cyprus might be having some impact on peripheral wholesale funding markets.
On a lower frequency basis, we need to track the monthly Target2 balances for peripheral countries, which typically become available during the first two weeks of the following month, and the ECB data on MFI balance sheets which are published at the end of thee following month.
In what we view as another ill-conceived and ill-timed move, the Spanish Minister of Finance & Public Administration announced this week a tax or bank levy (probably 0.2%) to be imposed on bank deposits, without details on which deposits will be affected or timing.
This is adding to the Cypriot crisis in sparking deposit outflow risks.