Three European countries with oversized banking sectors have suffered major bank failures. Two of them are in the Euro Zone (Ireland and Cyprus) and one has its own currency (Iceland). Iceland and Cyprus imposed temporary capital controls, while Ireland did not. Iceland imposed losses on the foreign depositors in its large, failed banks while Ireland, under EU pressure bailed out everyone (even bond holders) except the shareholders.
The jargon used to describe much of this—“bail outs,” “bail ins,” “haircuts,” “good bank bad bank splits,” etc.—can be confusing. In this note I attempt to clarify the key concepts and their importance via the examples of Iceland, Ireland and Cyprus.
Market discipline vs. supervision and regulation
Incentives always matter. Banks, like any other business, are in business to make money. But the amount of risk they take (more risk more return—ON AVERAGE) depends on who regulates their behavior. Fundamentally, the market can regulate bank risk taking—by the willingness of investors to lend to banks and of depositors to place their money there—or the government can.
The last century has seen a steady shift away from market regulation toward government regulation. Deposit insurance is an important factor contributing to that shift by removing any concern by smaller depositors of the condition of their bank. Thus deposit insurance requires a substitution of the due diligence that used to be performed by small depositors with increased government regulation of bank risk taking. In the United States, the Federal Deposit Insurance Corporation (FDIC) provides much of that supervision and regulation.
However, increasingly countries became unwilling to allow banks to fail. While shareholders might be wiped out when a bank became insolvent (i.e., when the value of its assets fell below that of its deposits and other liabilities), country after country have “bailed out” all other bank creditors, including uninsured depositors. Bailing out depositors and other creditors means giving taxpayers’ money to the bank to make up for its losses and thus cover its liabilities (other than shareholders). For large, “systemically important” banks (meaning banks whose failure could cause fatal losses in other banks or firms), most countries are not willing to let them fail at all, thus bailing out shareholders as well in order to allow the banks to continue to operate. Hence the problem of banks that are “too big to fail.” Bailing out uninsured depositors made deposit insurance redundant and pointless. Market discipline was pushed aside all together. The safety and soundness of banks came to rest almost completely on the adequacy of regulations and the skills of supervisors. Bank owners, the only ones who care any more, now have a financial incentive to take big risks for potential big gains. If they lose, as they do from time to time, the government, i.e., tax payer, will pick up the bill.
It is desirable to shift more of the discipline of bank risk taking back to the market by convincingly putting bondholders and large, uninsured depositors at risk of loss if their bank becomes insolvent. They have a financial incentive to get it right that supervisors do not.
Resolution of insolvent banks
Best practice when a bank becomes insolvent is to resolve it quickly and fully and to put a large part of the cost of its losses on uninsured creditors (shareholders, bond holders and uninsured depositors in that order). Normal company bankruptcy can take the form of shutting down, locking the doors, and selling off anything of value (normally taking a few years) and distributing the proceeds to the creditors in the order of the legal priority of their claims. It is a transparent and objective, but slow process. In many instances the highest value for a failing company is obtained by selling it whole or in part to another company that is able to run it more efficiently. The recent bankruptcy of Sara Lee and sale of its best products to other companies is an example.
The bankruptcy and resolution of an insolvent bank is more challenging because of the ease with which depositors can run when they sense trouble. Thus the weekend sale of such banks in whole or in part to another bank is the norm for small or medium-sized banks in the U.S. The good bank bad bank split, as occurred recently in Cyprus, is a recent example. Laiki became the bad bank that was closed and is being liquidated and the Bank of Cyprus became the good bank. After wiping out its shareholders and bondholders and administering a large haircut to the uninsured depositors, it acquired the insured deposits of Laiki and an equivalent value of good Laiki assets. Such bank resolutions, which freeze depositors’ funds only for very short periods (a few days), require special bankruptcy laws for tailored for banks. As the surviving good bank must continue to operate with little to no interruption, more judgment and uncertainty is involved in valuing the assets that it acquires from the bad bank.
It is instructive to look more closely at the resolution process used in Cyprus. First, the two major banks in Cyprus, Laiki and Bank of Cyprus, incurred large losses on their holdings of Greek sovereign debt when all banks were required to “voluntarily” write off about 75% of its value. The magnitude of this loss was clear and well-known from October 2011. The only issue was who would pay for it, the Cypriot government, the EU, or the creditors (depositors) of these banks. Depositor’s obviously thought that they would be bailed out (i.e. that the Cypriot government or the EU would pay for the losses of Laiki and Bank of Cyprus) as had been all depositors in Europe before them, though the deposit liabilities of the Bank of Cyprus fell from 37.1 billion Euros at the end of 2010 to 32.1 billion at the end of 2011 to 28 billion at the end of September 2012 (the latest available).
After a terrible false start in which the Cyprus government attempted to pay for the losses by levying a wealth tax on all depositors (of good and bad banks), Cyprus choose to impose the entire loss on the respective banks’ owners and creditors, and to undertake the good bank bad bank split briefly described above (see my earlier blog on the subject: https://wcoats.wordpress.com/2013/03/27/the-cyprus-game-changer/). This was a dramatic change in approach that shifted the risk of bank behavior back to uninsured depositors. Many were shocked.
This approach is relatively easy for known losses and should have been undertaken a year and a half earlier when the Greek debt write off occurred. But many of the losses a bank has or is incurring are less clear. Of the currently delinquent mortgage loans, for example, how many will actually default and what will be the market value of the mortgage collateral. The recapitalization of insolvent Irish banks suffered from underestimation of the ultimate losses resulting in three separate injections of state money to recapitalize them, which weakened market confidence in the process. In part to deal with this uncertainty but to restore market confidence in the solvency of the surviving good bank (Bank of Cyprus), known losses were totally written off while the additional but uncertain further losses were covered by replacing an equivalent amount of deposits with equity claims on the BOC (shares). If losses turn out to be smaller than was provided for, these claims will have value and will thus reduce the size of the initial haircuts to deposits.
So “bailing out” a bank refers to covering its losses with someone else’s money (tax payers somewhere) and “bailing in” a bank’s creditors refers to covering its losses (after its capital is used up) with bondholders and uninsured depositors’ money via “haircuts” (writing off part of their value). The former “socializes” losses while leaving any gains from successful bets to the private owners and creates a serious moral hazard leading to excessive risk taking by banks. The latter makes depositors financially responsible for excessive bank losses and restores the market’s discipline of bank risk taking. This is very desirable as market discipline is more effective than regulatory discipline, but the dramatic change in the implicit rules in Cyprus was very large and abrupt.
As part of their respective bank resolutions, both Iceland and Cyprus imposed temporary capital controls, which, however, served very different purposes. Iceland has its own currency while Cyprus is part of the Euro zone.
At the time of Iceland’s banking crisis in 2008 its three largest banks had assets 11 times the total annual output of the economy. About half of their assets (largely loans) and their funding were outside of Iceland. Landsbanki, for example, funding its lending with roughly the same amount of borrowing and deposits (a highly risky strategy). When the borrowed funding of these three banks dried up, their size made it impossible for the Icelandic Central Bank (ICB) to provide their needed liquidity (much of which was in the Euro, a foreign currency), resulting in the failure of all three banks in the second week of October 2008.
Iceland honored all insured deposits domestically and abroad but moved all domestic deposits into newly established “good” banks from the three now bad banks, while leaving their overseas, uninsured deposits in these three banks in receivership. To the extent that these banks failed because of illiquidity (the cut off of their borrowed funding), the receivership should be able to recover all losses to depositors from the liquidation of the banks’ remaining assets.
The UK and Netherland’s objected to the unequal treatment of the uninsured deposits of Icelanders and of foreigners. While Iceland’s decision to bail out all of its domestic depositors may be questioned because of the moral hazard it perpetuated, they had no legal obligation to do the same for Euro deposits by foreigners. The UK and the Netherlands stepped in and followed the same policy adopted by Iceland by guaranteeing the deposits of their citizens. They then tried to collect the cost of these guarantees from Iceland, a very questionable claim.
As the three new “good” banks were fully capitalized, they should have been able to withstand any level of deposit withdrawal as long as the ICB was able to provide any liquidity needed against the good but illiquid assets of these banks. The return of depositor confidence to the banks invariably takes time and some depositors wanted to withdraw their funds. However, because Iceland has its own currency, nervous Icelandic depositors wanting to move their bank deposits abroad, would need first to convert them into Euros or U.S. dollars, which would have depreciated the international value (exchange rate) of the Icelandic króna, and depleted ICB’s international reserves. A depreciation of the króna would raise the cost of imports and reduce the standard of living in Iceland. To protect the exchange rate from excessive devaluation, the ICB imposed temporary limits on the amount of money its residents could move out of the country. These capital controls are still in effect.
Lucky Cyprus is in the Euro zone. After recapitalizing its banks, in part by writing down their deposit liabilities, they should have sufficient assets to cover all of their deposit liabilities and thus to cover any deposit withdrawals. The only issue would be whether the BOC’s assets were sufficiently liquid to cover the withdrawals. Within the Euro zone payments outside the country are made via the Target Payment System. A transfer of deposits from the BOC in Cyprus to a bank in any other Euro zone country is made by debiting the BOC’s clearing balance with the Central Bank of Cyprus (CBC) and crediting the recipient bank’s clearing account with its central bank via Target. If the BOC does not have sufficient funds in its clearing account with the CBC and is unable to sell sufficient assets to increase that balance, it can borrow the funds from the CBC using its good but illiquid assets as collateral. The CBC is able to do the same by borrowing from the European Central Bank (ECB), which is prepared to lend unlimited amounts against good collateral now that Cyprus has undertaken the measures required for the troika’s financial support (i.e., from the EU/ECB/IMF). There is no exchange rate issue or concern. It is purely a matter of the solvency and liquidity of Cypriot banks.
However, establishing sufficient liquidity to fund large deposit withdrawals may take a few weeks or months and thus Cyprus has imposed temporary capital controls that limit the amount of money that may be withdrawn each day as cash or by transfer. If the arrangements enjoy sufficient public confidence in the soundness and viability of the surviving Bank of Cyprus, the deposit withdrawals should be modest. The period of limits on withdrawals should be measured in weeks rather than months or years.
The resolution of Cyprus’s insolvent banks ultimately, after a false start, was achieved by bailing in its creditors. The resolution was relatively quick and seems complete. While Cyprus’s economy is likely to suffer its abrupt adjustment for some time, its banks should now be sound. The dramatic shift of the responsibility of regulating the risk taking of banks to their uninsured depositors, should, if it is maintained throughout Europe despite nervous claims that it is one-off and not a model, restrain excessive risk taking by banks and lead over time to a stronger banking system. In the interim, there may be some disruptive deposit shifts as previously reckless banks are forced by the market to clean up their acts.