The Need for an Emergency Bank Debt Insurance Mechanism

Javier Suarez

27 March 2008



The so-called subprime crisis that started in the Summer of 2007 has created an unprecedented situation in global money markets. Essentially all market means for the short and medium term financing of banks (from traditional interbank deposits to the most ingenious forms of securitization) exhibit large spreads and shrunk volumes of transactions.

Central banks have orchestrated lending of last resort (LLR) operations in an attempt to help many banks to sort out their liquidity problems. Newer forms of LLR schemes (such as the Fed’s Term Auction Facility) have been introduced and kept in place for several months now, but money markets do not seem to be returning to normality. Meanwhile, many investors are believed to hold on to their excess liquidity. But very much in parallel to what happens in a text-book self-fulfilling bank run, investors seem no longer willing to invest in short term bank liabilities.

In CEPR Policy Insight No. 19, I provide a more detailed diagnosis of this situation and its implications for the world economy, comment on the pros and cons of current LLR operations, and propose the creation of an Emergency Bank Debt Insurance Mechanism (EBDIM) as an alternative. The main argument goes as follows.

The current crisis has elements that respond to the combined logic of various well-known theories of adverse selection, bank panics, and credit cycles developed by economists over the last thirty years. It all started with a significant shock (the change in the cyclical position of the US housing market) and its impact on the oversized US subprime mortgage market. Securitization excesses (high leverage, large reliance on short term borrowings, large complexity, little transparency, and inadequate assessments of systemic risks) converted the global shock into a collection of asymmetric, potentially fatal, and hardly observable, shocks to the financial health of systematically important financial institutions. Money markets collapsed because of the fear of some not-yet-identified institution being the next to go under. Given the huge money-market refinancing needs implied by the prevailing model of banking, a situation similar to a bank panic of the old days emerged.

Coordinated LLR operations by major central banks since the start of the crisis are aimed at breaking the vicious circle whereby banks’ liquidity and solvency problems reinforce each other. If they do not succeed, the major threat for the world economy is that the weakening of bank balance sheets leads to an overall restriction in the supply of bank credit. Unfortunately, if one is to judge from the pessimism among bankers and bank stock traders, it is not clear that LLR interventions have succeeded in keeping banks’ marginal short-term refinancing cost at normal, non-crisis levels.

The central point of CEPR Policy Insight No. 19 is to advocate that the goals of systematic and prolonged LLR support during a crisis like the current one could be more effectively attained by an EBDIM. Both arrangements share the rationale of providing banks with sufficient liquidity in case of crises, such as the current one. The EBDIM would short-circuit the panic logic in money markets by temporarily providing full coverage to any short-term lending explicitly supported by the insurer. Under this coverage, money markets should recover their normal functioning and banks would be able to meet their refinancing needs as in normal times. This would buying time for both the banks and their supervisors to clear up the mess, disclose the losses, and proceed with the recapitalization, intervention or closure of the capital-impaired institutions.

The EBDIM and LLR also share some drawbacks such as the temporary removal of market discipline, the socialization of the losses due to private risk-taking decisions, and the potential creation of moral hazard. Under the EBDIM, market discipline would only be relevant insofar as banks wish to access uninsured sources of bank financing (such us core equity and subordinated debt issued in order to comply with capital regulation or to voluntarily enhance banks’ solvency). With low market discipline, supervisory discipline becomes essential and, thus, the providers of EBDIM should act in close cooperation with (or act as) the banks’ main supervisors. These supervisors should exert their discipline through the usual means (capital regulation and prompt corrective action), as well as by establishing the total borrowing under coverage (that might be made a function of a bank’s book-value capital position, perhaps after some supervisory-defined adjustment for the risk of the bank’s assets).

The EBDIM should be backed with government support and, in case of a bank’s default, the guaranteed borrowings should be repaid immediately (so the EBDIM should enjoy expeditious access to central bank liquidity). In order to facilitate its management, the EBDIM could limit its coverage to a small number of standardized instruments for short and medium term borrowing. This is most likely to produce a crowding-out effect on the uninsured instruments, which would not necessarily bad, especially in a transitory phase in which it would be most valuable for supervisors and the managers of the EBDIM to have real-time information on banks’ cross exposures.

The EBDIM should build on existing retail deposit insurance practices for the allocation of responsibilities for cross-border transactions. According to the “home country principle,” the EBDIM could be decentralized at country level, so that the various national EBDIMs would guarantee, subject to the conditions set by the national supervisors, the money-market borrowings of the banks under their jurisdiction.

Given their emergency nature, the EBDIMs should not be expected to build up funds with which to cover their potential losses, but they could compensate some of their costs by charging periodic fees (perhaps following the models used for the pricing of credit default swaps). Even if these fees resulted in charges equivalent to a spread of 20-30 basis points, insured borrowing would still be much cheaper than uninsured borrowing under the spreads observed during the current crisis.

As argued in the more detailed proposal, an EBDIM designed along the above lines would have five major advantages with respect to massive LLR operations:

  1. It is more flexible and informative. Money markets should work very similarly to “normal” times, liquidity surpluses would continue to be directly transferred from their holders to the banks with liquidity needs, and supervisors would have daily access to the information revealed by these transactions.
  2. It produces smaller interference with monetary policy implementation since it eliminates the need to sterilize the liquidity injections due to massive LLR operations.
  3. It is more explicit about the potential costs of the bank safety net to taxpayers and their cross-border allocation, and allows for the funding of its operating expenses with explicit assessments on the insured borrowing.
  4. It reduces conflicts of interest by explicitly assigning its potential losses to the authorities in charge of imposing supervisory discipline on the benefited banks.
  5. It could be introduced on a country-by-country basis, using the “home country principle,” although some degree of coordination between the adopting countries would facilitate its success in restoring the normal functioning of international money markets and minimize criticism on its potential for “unleveling the playing field” in international banking.




Topics:  Financial markets

Tags:  liquidity, financial crises, short-term lending, interbank deposits, EBDIM

Professor of Finance at CEMFI, Madrid and CEPR Research Fellow