Comments on Part III

Adrián Armas, Eduardo Levy Yeyati, and Alain Ize
Published Date:
July 2006
  • ShareShare
Show Summary Details
Philip Turner

The organizers are to be congratulated on bringing together three analyses which go to the heart of the prudential challenges raised by dollarization.

Chapter 7, by de Brun and Licandro, is a fascinating account of the Uruguay crisis. The interaction between exchange rate collapse, bank solvency and public debt is particularly well presented. This crisis, and many others too, demonstrates that the existence of a long-term market for government debt denominated in local currency makes it easier to contain a banking crisis. With such a market, the public would know that the government has the capacity to make good the banks’ obligations, and this reduces the risk of a run on the banking system. Moreover, any flight from bank deposits that does take place would not necessarily translate into a flight of capital from the country. The cost of servicing government debt would rise of course (as expectations of currency depreciation would drive down the value of local currency bonds) and the fiscal costs could be heavy, but outright default by the banking system could be avoided.

The two other chapters deal with two key aspects of prudential oversight under dollarization:

  • Credit risk. The simplest diagnostic of credit risk (debt service to income) is undermined in a highly dollarized economy because dollar debt service is lower than local currency debt service. In addition, the value of local currency collateral can be eroded by large exchange rate changes – which is often just when the collateral is really needed.

  • Liquidity risk. It is harder to borrow foreign currency than domestic currency in the event of a crisis.

Chapter 8 reviews various measures taken by supervisors in seventeen countries to address credit risk from mismatches of the borrowers (as distinct from those of the bank). The chapter represents a very thorough analysis of how supervisory tools could be used to address currency mismatches. It deserves to be widely read. One might quibble with some of the specific suggestions (e.g., differential capital requirements for business in different currencies). But the central message that better risk management practices and more ‘currency-mismatch aware’ supervision are both needed is surely right. Their finding that only a few dollarized countries have adopted measures to control the credit risks stemming from borrowers’ currency mismatches shows that more effective supervisory action is still needed.

The framework proposed by Basel II is better suited to this task than was Basel I. One key aim of Basel II is to encourage banks to develop a quantitative risk management culture. Risk assessment is based on default probabilities that are derived from actual history. If, for instance, the default probability of a foreign currency mortgage is x per cent, and that of domestic currency mortgages is y per cent, then the capital charge and risk management of the lending bank should reflect that.

In the near term, admittedly, there may well be a need to rely on some use of regulatory ratios (for instance, following the example just given, by requiring higher provisioning against foreign currency mortgage lending) in countries with a recent history of instability, because the information content of economic and financial variables in such countries is rather low. In the longer-term, however, developing this quantitative risk management culture is essential. Rigorous quantification on the basis of past history might help banks and bank supervisors to demonstrate to clients and to the public that their credit risk charges are not arbitrary. This might help shield supervisors from political criticism. In this respect, as Chapter 8 argues, disclosure is important to help market discipline work effectively.

The job of bank supervision is to ensure that each bank has in place the risk management procedures that are appropriate for its particular circumstances. Because the circumstances of individual banks do differ, attempting to impose one-size-fits-all regulatory ratios could well make it harder to implement the effective supervision of the risks of currency mismatches.

Having said all that, it has to be recognized that exchange rate-related credit risk is a complex concept to make operational. In principle, the lending bank needs to know not only the currency of the loan under consideration, but also about the currency of denomination of the total portfolio of a customer. Perhaps credit registers summarizing data from all banks could incorporate such information. A second complication arises from collateral: in the case of a default after a massive currency devaluation, recovery values might perversely be higher for those who have lent dollars – because dollar-denominated claims rise relative to local-currency claims.

Chapter 9 provides a very subtle discussion of liquidity. Liquidity is difficult to define because it reflects judgments of responses to not-yet-defined shocks (how would Bank X respond to some not-yet-known financial shock?).

The first responsibility for managing liquidity should be with banks – if they get it wrong, they should pay a penalty. The authorities could reinforce this by imposing liquidity buffers – requiring, for example, that short-term foreign currency liabilities incur higher liquid asset requirements than local currency liabilities. The fact is that the central bank in highly dollarized economies may not be able to supply foreign currency as readily as domestic currency in times of stress – and this needs to be reflected in some form of liquidity ratio imposed on banks. Chapter 8 usefully surveys some practices in this area. In addition, the central bank could also aggregate the liquidity gap analyses conducted by individual banks. It could also stress test the system for liquidity shortfalls.

If the banks and the supervisors nevertheless get it wrong and find themselves swamped by capital flight out of the banking system in a crisis, the consequences can be catastrophic. This is the very real problem that Chapter 9 addresses. It considers a pre-programmed suspension of deposit convertibility – a ‘circuit breaker’ – to keep the payments system working.

This idea merits careful consideration. There are, however, two problems that are reminiscent of the debate about a LOLR. The first is that an announcement in advance that circuit breakers can be activated could well create damaging expectations that legal contracts in general would not be respected during a crisis. The second is that governments, in practice, might not limit themselves to preannounced steps in the early stages of a crisis: people would suspect that behind even modest or very temporary measures lie more draconian measures.

Would it not be better to tighten liquidity rules on banks when aggregate liquidity begins to look doubtful? In my view, this is often not done because of fears it would encourage dollar deposits to move from local banks to banks based abroad. It would cause the local supply of dollar intermediation to contract – not popular with local banks and perhaps not popular with the country’s owner of the foreign exchange reserves. Nevertheless, it should never be forgotten that it is sometimes better to allow local banking business in dollars to contract than to allow things to deteriorate so far that contracts have to be abandoned.

    Other Resources Citing This Publication