The key to successful banking crisis management is coordination with the overall macroeconomic framework. Perhaps the most critical episode is the containment stage when, in addition to implementing the steps to address the immediate aspects of the crisis, macroeconomic policies may need to be tightened for confidence in the banking system and the currency to be restored. The appropriate mix of macroeconomic policies will depend on the nature of macroeconomic imbalances and the state of the banking system.
While this paper mainly concentrates on banking system restructuring, this section touches upon selected issues pertaining to the broader policy context within which bank restructuring must take place. Bank restructuring needs to be implemented in conjunction with supporting macroeconomic policies, taking into account existing macroeconomic constraints. Moreover, measures to contain the crisis and restructure banks may have macroeconomic consequences that need to be taken into account in policy design. Given this breadth of concerns in implementing a financial sector restructuring strategy, banking crisis management requires coordination of the macroeconomic policy framework, the reform of policy instruments, and the design of microeconomic initiatives.
This section discusses four specific issues. First, the interaction of monetary and exchange rate policies and bank restructuring, with a focus on the initial stages of banking crises. Second, existing evidence on the path of economic growth during and after a systemic crisis. Third, lessons from country experiences on reintermediation after a banking crisis has occurred. Fourth, the interaction between sovereign debt restructuring and bank restructuring—an issue that has emerged forcefully in the recent Latin American crises.
Issues in Monetary and Exchange Rate Policies
In the initial stages of a crisis, monetary policy will need to walk a fine line between conflicting goals. Monetary management should accommodate liquidity demands by banks because failure to provide liquidity could result in a collapse in the payment system and an acceleration in the deterioration of economic conditions. If liquidity support causes serious pressures on prices and the exchange rate, however, monetary policy will have to be tightened to attempt to reabsorb the excess liquidity promptly. In walking this fine line between preserving the payment system and losing monetary control, monetary policy targets—inflation, net domestic assets, or money base—will have to be flexible.
From the policy point of view, the central bank is faced with the task of mopping up the excess liquidity created by its emergency support operations. Accordingly, the interest rates on central bank securities will have to rise. Illiquid banks will be able to pass on the high rates to depositors, thereby raising the funds that will be used to purchase the policy instruments. As normal monetary operations resume, rates should be brought down as quickly as feasible, because high rates for protracted periods will severely damage borrowers’ repayment capacity, lead to deterioration in banks’ loan portfolios, and draw political criticism.39
In order to handle the crisis without losing monetary control, monetary instruments may need to be reformed or new ones introduced. As the initial stages of a crisis are likely to require substantial injections of emergency liquidity assistance, the policy instruments used by the central bank must be sufficiently flexible and available in sufficient abundance to absorb excess liquidity in the system. Policy instruments may be lacking or weak or a secondary market may be underdeveloped, resulting in the need to upgrade instruments and infrastructure. For instance, the relative merits of government and central bank securities for the conduct of monetary policy may need to be assessed, which in some cases might be a function of the sustainability (and related value) of sovereign debt.40 Interest rate liberalization could also be an essential aspect of making monetary policy more effective.41
Experience shows that rigid exchange rate regimes or policies are often abandoned in systemic banking crisis. If creditors and depositors have doubts about the sustainability of the exchange rate and macroeconomic policies, the run on the banks combines with a run on the currency. At the same time, capital outflows may put pressure on the exchange rate, even if it was flexible before the crisis, which will only stabilize when domestic policies take hold and confidence begins to return. In these cases, the run on banks will only subside once the exchange rate is credibly stabilized, as the recent experience of Argentina indicates.
A number of countries in recent years have imposed a broad range of controls on capital outflows to reduce pressure on the exchange rate in the context of banking crises.42 Malaysia and Thailand, for instance, reimposed capital controls during the Asian crisis of 1997–99; more recently in 2002, Argentina introduced a broad range of exchange restrictions in response to runs on the banks and the currency. The effectiveness of these controls in realizing their intended objectives has been mixed, with the highest success in stabilizing the exchange rate achieved in countries that introduced the controls in conjunction with other macroeconomic and financial policies.
Systemic Crises and Economic Growth
Systemic banking crises are almost always associated with severe recessions. Crises may be caused by the very same problems that led to recession. In addition, a crisis may induce a recession due to several factors, including inter alia curtailment of financial intermediation, increased economic uncertainty, negative wealth effects from depressed asset values, and delays in restoring creditworthiness of insolvent firms and financial intermediaries.
A decline in bank credit is expected in the wake of a systemic crisis and is typically a result of both demand and supply factors. Credit demand falls as economic growth declines or stock adjustments from major asset and credit bubbles occur. Some of the best borrowers may start borrowing directly in bond markets domestically or from various sources abroad. Meanwhile, the supply of credit may be reduced as banks begin to apply stricter lending criteria and stop new lending to a growing number of nonperforming borrowers, and as some banks exit the market. If loan recovery is in doubt, banks will seek risk-free investments and will prefer to stay liquid as a precautionary stance. Data also need to be interpreted carefully because the decline in bank credit may reflect increased loan loss provisions (if credit aggregates are reported on a “net” basis); the impact of bank closures; or the sales of loans to an asset management company.
A decline in credit may reflect a desirable market adjustment. The combination of tight credit and banks awash in liquidity is likely to lead to political pressures for quantitative credit targets, special lending schemes for the most affected sectors, artificially low interest rates, and even across-the-board reductions or “haircuts” of loan balances. Such pressures should be resisted because they undermine bank profitability and efficiency, delaying recovery and the restoration of solvency. Banks should be cautious and lend only to viable companies.
Interest rates are likely to rise in the wake of a crisis outbreak. Economic agents are likely to remain risk averse for some time with a strong liquidity preference, and there may be a considerable delay before depositors and other creditors regain sufficient confidence to entrust their financial assets to the banking system. Under these conditions, interest rates will rise, with rates increasing further the greater the expectations of a sharp increase in inflation and exchange rate depreciation.
Experience also indicates that in several cases systemic banking crises have resulted in permanent losses in aggregate output (Figure 3). After the crisis is over, the growth rate may return to its potential level, but in several of the cases examined there was no catch-up to the previous trend level of output. In some cases (i.e., Ecuador, Indonesia, and Thailand), the losses have been significant, while in others the catch-up was quicker (i.e., Chile and Korea). The savings and loan association crisis in the United States apparently had no impact on aggregate output, or its negative impact at that time was more than offset by other positive developments in the U.S. economy. Nonetheless, care must be taken in interpreting these graphs, as several factors were at work in the economies.43
Estimated Output Losses in Selected Crisis Countries1
Source: IMF, International Financial Statistics.1 The graphs show the actual level of GDP versus the level that would have resulted if a simple measure of the trend growth rate had been maintained throughout the period. Trend growth rates are calculated using a Hodrick-Prescott filter for the 10-year period before the crisis began.Estimated Output Losses in Selected Crisis Countries1
Source: IMF, International Financial Statistics.1 The graphs show the actual level of GDP versus the level that would have resulted if a simple measure of the trend growth rate had been maintained throughout the period. Trend growth rates are calculated using a Hodrick-Prescott filter for the 10-year period before the crisis began.Estimated Output Losses in Selected Crisis Countries1
Source: IMF, International Financial Statistics.1 The graphs show the actual level of GDP versus the level that would have resulted if a simple measure of the trend growth rate had been maintained throughout the period. Trend growth rates are calculated using a Hodrick-Prescott filter for the 10-year period before the crisis began.Estimated Output Losses in Selected Crisis Countries1
Source: IMF, International Financial Statistics.1 The graphs show the actual level of GDP versus the level that would have resulted if a simple measure of the trend growth rate had been maintained throughout the period. Trend growth rates are calculated using a Hodrick-Prescott filter for the 10-year period before the crisis began.Estimated Output Losses in Selected Crisis Countries1
Source: IMF, International Financial Statistics.1 The graphs show the actual level of GDP versus the level that would have resulted if a simple measure of the trend growth rate had been maintained throughout the period. Trend growth rates are calculated using a Hodrick-Prescott filter for the 10-year period before the crisis began.The return of economic growth to its long-term trend level requires appropriate macroeconomic policies in combination with swift restructuring policies.44 In general, in countries where the policy response was strong—that is, well coordinated, well communicated, and credible—the notional economic costs were lower. For example, in Sweden and Korea the growth rate rapidly returned to trend, while the impact was greater in Ecuador, where credibility was constrained by a high public sector debt burden and political turmoil, and in Indonesia, where the central bank lacked instruments to absorb liquidity and therefore to control inflation and capital flight.
A final consideration is whether banking crises themselves may change the potential growth rate of the economy. A more favorable long-term outcome could result if the crisis prompts reforms that promote a more stable macroeconomic environment, or if strengthened banking regulation increases the efficiency of financial intermediation. Alternatively, and particularly if the crisis is mishandled, value embodied in firms or banks that become insolvent and are closed may be destroyed, despite underlying efficiency. A permanent exodus of capital may result, and valuable knowledge, such as that contained in relationships between lenders and borrowers, may be lost.
Postcrisis Reintermediation
Reactions to crisis episodes by stakeholders—depositors, other creditors, and bank shareholders—have varied across countries and appear related to the origins of the crisis; the credibility of institutional arrangements, in particular deposit insurance schemes; and modalities of crisis management. Figure 4 shows the evolution in deposits and loans of deposit money banks in real terms for selected crisis countries. There is a fairly consistent overall pattern: rises in both variables in the precrisis period, declines during the crisis, followed by stability with little or lower growth afterward. There is, however, considerable variation in the magnitude and duration of these effects, depending on the economic conditions and policy choices in the different countries.
Reintermediation requires correcting the factors that triggered (or worsened) the crisis. Successful reintermediation would include sustainable macroeconomic adjustment, sound fiscal and public debt management, a stable currency in which to denominate financial contracts, sound banks and supervisory frameworks, and credible liquidity arrangements. The speed of reintermediation is swift when credibility in macroeconomic and banking policies is recovered rapidly. When confidence in the local currency has been fundamentally undermined, rein-termediation may be accelerated through indexation or a higher degree of dollarization (see Ecuador, for example, although the delay in the passthrough of depreciation to consumer prices exaggerates the stability of real loans and deposits during the crisis period). Indexation arrangements (particularly to the dollar) can be a source of longer-term risks, however, and financial instability and partial dollarization tend to complicate the conduct of monetary policy.
Repairing the fabric of bank intermediation will also take time when banks are weak and relations with their debtors affected by an unresolved debt overhang or a poor economic or political environment: the former effects were at work in Indonesia and Mexico, for example, the latter in Indonesia, República Bolivariana de Venezuela, and Ecuador). For banks whose capital and profits have been affected by the crisis, their ability to accept deposits at desired rates postcrisis will be affected by investment opportunities.
The introduction of foreign banks to the domestic system may also provide a positive contribution to the process of reintermediation on the liability side of banks’ balance sheets. Perceptions of quality, parent backing, and the introduction of new products have the potential to attract depositors back to the banking system. The evidence on the contribution of foreign banks to reintermediation on the asset side of banks’ balance sheets, however, is less clear and needs further analysis. Often, the gradual entry of foreign banks into the system has been accompanied by recovery and stability in real deposits but no consistent recovery in loans. While more savings may be available to the financial system for investment, foreign banks may be more risk averse than domestic banks because the latter may have a better knowledge of the local market.
The Impact of Sovereign Debt Restructuring
Unsustainable fiscal or debt policies create an additional set of complications and constraints for crisis management. An unsustainable fiscal position requires a strong fiscal adjustment and may require not only that shareholders take losses, but that depositors or other creditors also share in the losses. Limited fiscal resources may also constrain the authorities’ efforts to contain the crisis and assist in recapitalizing banks. More private financial support will be needed and, if such support is not available, more banks will need to exit the system.45
Only limited experience is available about sovereign debt restructuring and its impact on bank restructuring. Countries that have engaged in sovereign debt restructuring include Ecuador, Pakistan, Russia, and Ukraine. These countries had very different macroeconomic and institutional conditions, and few firm conclusions can be drawn from these cases. The impact of debt restructuring will depend, however, on a large number of issues, including the initial conditions, the size of banks’ holdings of sovereign debt, the currency of denomination, the law governing the debt, the size of the needed restructuring, the restructuring alternative used, and the associated macroeconomic policy mix.46
The impact of sovereign debt restructuring on the banking system will depend on the broad framework under which the restructuring is taking place. If the solution to the government’s financing problems takes place under a market-based framework through a debt swap operation, banks may achieve higher interest earnings and cash flow but at the cost of exposure to higher country risk. Banks may also participate as creditors in a collective negotiation, allowing them to influence the terms of the rescheduling and ensure that their medium-term viability is not jeopardized.
The unilateral restructuring of sovereign debt can cause deep insolvencies in the banking sector and disrupt the effectiveness of standard resolution tools. When banks hold a significant portion of government bonds, restructuring can both cause a banking crisis and make the resolution of a banking crisis more difficult. The size of losses will depend on the terms and modalities of the sovereign debt restructuring. Large restructuring may result in immediate undercapitalization or even banking system insolvency. At the same time, the range and effectiveness of resolution tools may be limited. The government may not be in a position to play the roles of creditor, guarantor, or owner of last resort as in a conventional banking crisis.
For these reasons, any strategy for sovereign debt restructuring must explicitly consider the impact on the banking system. If measures to achieve debt sustainability result in banking system insolvency, the final costs to the government and the economy may well overshadow any gains from debt reduction. Decisions on overall strategies, therefore, must be made in terms of the least cost, with costs being broadly defined in terms of immediate fiscal costs, costs to the banking system, and the impact on economic growth and employment. Policy measures should include sufficient fiscal adjustment so that a viable banking system remains at the end of the adjustment process.
Sovereign debt restructuring may be achieved by extending maturities, reducing interest rates, and/or imposing nominal reductions (haircuts) on the debt stock. In all cases, banks will suffer losses because both cash flow and the present value of their assets will be reduced. The impact of these losses on banks will depend on how the measures are implemented.
Maturity extensions and interest rate reductions will lower the cash flow of the banks, causing an economic loss. According to international accounting standards, the present value losses from restructuring must be disclosed, but there is no international best practice for the treatment of such losses during a crisis.47 Bank supervisors may require full provisioning, immediately affecting bank equity. Alternatively, regulations might permit banks not to make full provisions for impaired sovereign bonds, provided that the banks show they are viable notwithstanding the cash flow impact of the restructuring.
The least disruptive outcome for the banking system is when the restructuring of government debt does not cause nominal reductions in bank assets (i.e., without provisioning for economic loss or a haircut). In this case, actions to preserve bank viability may require reductions in interest rates or extensions of deposit maturities through withdrawal restrictions or securitization. Depositors will react negatively to any change in their contracts by drastically withdrawing available funds from the banking system (Box 13). However, as discussed above, measures may be taken to mitigate the runs and protect the operation of the payment system (see Section III on administrative measures). Moreover, the impact of restructuring may be eased by the adoption of a credible and comprehensive macroeconomic program in parallel with the restructuring.
The most disruptive outcome for the banking system occurs when nominal losses are imposed. Bank capital will be written down and the imposition of remaining losses on depositors may become unavoidable. If banks’ exposure to the government is relatively small in terms of total assets, shareholders or the government may still be able to recapitalize the bank to a minimum capital level and gradually return it to full prudential compliance.48 A more difficult situation occurs when banks hold a large portion of their assets in government debt and/or a majority of the stock of government debt. In this case, the sovereign debt restructuring will require a write-down of bank capital and then imposition of the remaining losses on depositors.
Haircuts on deposits are likely to be far more disruptive than deposit restructuring because value is permanently lost. Because depositors appear to have a strong preference for retaining the nominal value of their deposits, a nominal deposit loss and the accompanying elimination of any explicit deposit insurance would likely cause a long-lasting collapse in investor confidence. The effects will permeate the economy and jeopardize the medium-term viability of the banking system.49 The authorities will have few tools to address the crisis, and additional administrative measures may become necessary. In this option, depositors will face a nominal reduction of their deposits and may not be able to avoid the additional costs resulting from the administrative measures. Under such drastic conditions, the possibility of revitalizing financial intermediation will be small, with corresponding effects on future economic growth and the economic welfare of the population.
Depositor Reactions to Contract Changes
The experience of several countries with changing the contracted terms of bank deposits suggests that the risks of deposit runs and large drops in bank intermediation are significant.
Mexico (1982), Bolivia (1982), and Peru (1985) dedollarized bank liabilities. With little credibility in the domestic currencies, capital flight—despite capital controls—and bank disintermediation quickly followed. In Mexico, deposits-to-GDP ratios declined to 6 percent in 1988 from 26 percent in 1981. Similarly, in Peru they fell to 10 percent in 1989 from 21 percent in 1983. In Bolivia, the ratio of M2 to GDP declined to 11 percent in 1986 from 26 percent in 1982.
In Argentina, the Bonex plan of 1989 halved the stock of liquid assets by converting time deposits to 10-year bonds, and intermediation rates fell to 7 percent in 1991 from 21 percent. During the 2002 crisis, forced pesification of deposits resulted, despite strict withdrawal restrictions, in a leakage of 25 percent of deposits in the first 8 months of the year.
In Ecuador (1999), deposits were initially frozen, but when the freeze was lifted massive deposit flight led to currency pressures that were eventually resolved through full dollarization.
Pakistan (1998) froze deposits of banks as a temporary response to the collapse of the private banking system following sovereign debt default. Pakistan’s deposit freeze was confined to foreign currency deposits of residents and nonresidents to contain capital flight. The frozen deposits could subsequently be converted to free domestic currency deposits or special U.S. dollar-indexed bonds. The effect on intermediation was contained, in part due to the limited nature of the operation.
Uruguay (2002) reprogrammed dollar time deposits in public banks over a maximum period of three years. Dollar sight and saving deposits and all peso deposits in the domestic banking system were left unfrozen, as were all deposits in foreign banks.
Typically, contract modifications were introduced in times of large macroeconomic imbalances. When macroeconomic imbalance was the major factor leading to the episodes, strong credible stabilization programs have been critical to reintermediation. In Mexico, depositors remained offshore until 1988, when a stabilization program led to a massive return of flight capital. The experience was similar in Peru following the reform program in 1991–93, with intermediation ratios rising by 8 percentage points in the 1991–96 period. In Argentina, following the introduction of the Convertibility Plan in 1991, deposits-to-GDP ratios rose by 14 percentage points between 1990 and 1994. Ruble intermediation has not recovered in Russia and, conversely, foreign currency–deposits-to-GDP ratios are now 40 percent above precrisis levels. Pakistan too has suffered, but here the declines in intermediation are more associated with the decline in official financing following sanctions.
Reflecting the destabilizing impact of depositor and creditor haircuts, countries have rarely written down the contractual value of deposits.50 Given concerns about financial sector stability, losses on depositors and other creditors are typically imposed through a compulsory change in the contractual terms, such as a suspension or reduction in interest rates, or a conversion of depositor claims into long-term bonds or equity. Experience has shown that, to be successful, such measures must be accompanied by a coherent and comprehensive macroeconomic adjustment program. In addition, there will often be a need to strengthen debt management practices in order to rebuild the market for government debt.
