Overview of Strategy
Irrespective of its origin, a systemic banking crisis first emerges as a liquidity problem in some or all of the banks. Large creditors, both foreign and domestic, are generally the first to leave the banking system.15 As the outflow becomes known, smaller depositors follow quickly. Very large amounts can move in hours and, if not stopped, such runs may stall the operation of the payment system.
Liquidity and deposit withdrawals are symptoms of underlying problems but do not necessarily signal a systemic banking crisis. The authorities must form a quick judgment whether the deposit runs reflect concerns about the solvency of individual institutions, the stability of the banking system, or overall economic management. This judgment is often difficult because reliable data are scarce and quickly outdated. Moreover, wishful thinking and denial may affect the authorities’ judgment. Available supervisory, central bank, and market data will give some idea of the order of magnitude of the total losses in the banking system. Such data will also indicate whether deposit withdrawals reflect flight to quality within the banking system or a more general flight from the banking system. As a banking crisis becomes systemic, creditors are no longer able to distinguish viable from nonviable banks, and confidence in the overall stability of the system is jeopardized.
A top priority in the early stages is to stabilize banking system liabilities by stopping depositor and creditor runs. Irrespective of the causes of the crisis, the first step should be to provide sufficient liquidity to the banking system to protect the payment system and give the authorities time to determine the causes of the crisis and design an appropriate policy response. Options could include some combination of protection for creditors, particularly depositors; up-front closure of clearly insolvent banks; and adoption of a comprehensive macroeconomic stabilization package. If these measures are unsuccessful, or if emergency liquidity facilities are limited by a high degree of dollarization or the inability to sterilize liquidity injections, the authorities may be forced to consider administrative measures to stop the deposit runs, such as a reduction in deposit liquidity, a full deposit freeze, or capital controls. These administrative measures are likely to cause major economic disruption and, therefore, must be viewed only as a last resort if all other measures fail.
While liquidity support is needed to protect the payment system, the authorities must seek to limit the impact of this support on prices and the exchange rate. Liquidity support can cause serious pressures on prices and the exchange rate. Accordingly, the central bank must use available monetary instruments to sterilize the growth in base money. The tightening of monetary policy is likely to result in increases in domestic interest rates. These rates must be brought down as quickly as feasible because high rates for protracted periods will severely damage bank borrowers and banks, and draw political criticism, as discussed in Section VI.
An often-neglected aspect of crisis management is the importance of strengthening market confidence by announcing the authorities’ strategy. Notwithstanding the need to deal with an unfolding crisis and take a multitude of quick actions, attention should be given to making public announcements to explain the authorities’ understanding of the crisis and actions the government is taking. Such announcements should be consistent and authoritative and ideally be given by a single official spokesman throughout the containment and restructuring phases. To the extent possible, a political consensus for a crisis management strategy should be forged, and the working arrangement among different government authorities and agencies defined. Clear and consistent information will help contain rumors and misinformation, avoid new panics, and prevent further erosion of private sector confidence.
Policies for Stabilization
Emergency Liquidity Support
Central bank support provides the first line of defense against liquidity shortages in banks. Illiquid banks should be provided necessary resources, while the authorities identify the causes of the crisis, develop an appropriate response, and begin implementing actions to deal with insolvent banks. Support can be given in a variety of ways, including uniform reductions in reserve requirements, access to overdraft facilities and discount windows, open market operations, and instruments such as repos and reverse repos (see Table 1 and Appendix II for an overview of measures taken during the containment stage). The criteria for providing such support must be uniformly applied to all banks in the system, and the rules for providing such support should be transparent. Differentiating among banks will only cause uncertainties and could worsen depositor confidence. In highly dollarized economies, the central bank’s ability to provide liquidity support is constrained by the level of international reserves, and aggressive measures may be needed on bank intervention and resolution.
Liquidity Support During Selected Systemic Banking Crises
At peak.
Liquidity Support During Selected Systemic Banking Crises
| Country | Stock of Support1 | Form | Notes |
|---|---|---|---|
| Ecuador | 13 percent of GDP in 1998–99 | Central Bank of Ecuador loans and rediscounting of bonds issued by the Deposit Guarantee Agency (AGD) | Most loans not repaid; the central bank forclosed on some fixed assets used as collateral |
| Indonesia | Rp 156 trillion (16 percent of GDP) in August 1998 | Bank of Indonesia overdrafts | … |
| Korea | W 11.3 trillion + US$23.3 billion (2.5 percent of GDP) in December 1997 | Bank of Korea deposits and loans | All loans repaid |
| Malaysia | RM 35 billion (13 percent of GDP) at end-January 1998 | Bank Negara deposits | Most loans repaid by end-1998 |
| Mexico | MEX$38 billion + US$3.9 billion (2+1.3 percent of GDP) in April 1995 | Loans and capital injection from Banking Fund for the Protection of Savings (FOBAPROA) borrowed from the Bank of Mexico | All outstanding foreign currency loans repaid by early September 1995 |
| Russia | Rub 105–120 billion (4 percent of GDP) between August and October 1998 | Central Bank of Russia loans to 13 banks for a term of up to one year | Rub 9.3 billion repaid by end-1998 |
| Thailand | B 1,037 billion (22 percent of GDP) in early 1999 | Loans and capital injection from the Financial Institutions Development Fund (FIDF), a subentity of the Bank of Thailand | FIDF claim on financial institutions declined to B 227 billion by end-1999 |
| Turkey | TL 6 quadrillion (3.3 percent of GDP) in September 2001 | One-week repos by Savings Deposit Insurance Fund (SDIF) and state banks with the Central Bank of Turkey | One-week repos rolled over into longer-term instruments |
At peak.
Liquidity Support During Selected Systemic Banking Crises
| Country | Stock of Support1 | Form | Notes |
|---|---|---|---|
| Ecuador | 13 percent of GDP in 1998–99 | Central Bank of Ecuador loans and rediscounting of bonds issued by the Deposit Guarantee Agency (AGD) | Most loans not repaid; the central bank forclosed on some fixed assets used as collateral |
| Indonesia | Rp 156 trillion (16 percent of GDP) in August 1998 | Bank of Indonesia overdrafts | … |
| Korea | W 11.3 trillion + US$23.3 billion (2.5 percent of GDP) in December 1997 | Bank of Korea deposits and loans | All loans repaid |
| Malaysia | RM 35 billion (13 percent of GDP) at end-January 1998 | Bank Negara deposits | Most loans repaid by end-1998 |
| Mexico | MEX$38 billion + US$3.9 billion (2+1.3 percent of GDP) in April 1995 | Loans and capital injection from Banking Fund for the Protection of Savings (FOBAPROA) borrowed from the Bank of Mexico | All outstanding foreign currency loans repaid by early September 1995 |
| Russia | Rub 105–120 billion (4 percent of GDP) between August and October 1998 | Central Bank of Russia loans to 13 banks for a term of up to one year | Rub 9.3 billion repaid by end-1998 |
| Thailand | B 1,037 billion (22 percent of GDP) in early 1999 | Loans and capital injection from the Financial Institutions Development Fund (FIDF), a subentity of the Bank of Thailand | FIDF claim on financial institutions declined to B 227 billion by end-1999 |
| Turkey | TL 6 quadrillion (3.3 percent of GDP) in September 2001 | One-week repos by Savings Deposit Insurance Fund (SDIF) and state banks with the Central Bank of Turkey | One-week repos rolled over into longer-term instruments |
At peak.
Conditions for central bank liquidity support in normal times, such as collateral and penalty interest rates, may need to be eased. In a systemic crisis, the value and quality of collateral may not be known, while excessive penalty rates may add to banks’ distress. At the same time, the central bank must maintain its intervention rates at levels sufficiently high to make it a source of last (rather than first) resort for liquidity. Absent traditional central bank safeguards, borrowing banks should be subject instead to restrictions on their activities and heightened supervision.16
In principle, liquidity support should be provided only to solvent banks, but the differentiation between solvent and insolvent banks is often difficult in systemic crises because data are poor or outdated. Ideally, insolvent banks should be expeditiously intervened. Ceasing liquidity support to selected banks may not be feasible if this interrupts the payment system and risks triggering a wider crisis. Stopping liquidity support to the entire system would be even less feasible. Given these uncertainties, central bank support in response to a systemic crisis should be explicitly guaranteed by the government, in the understanding that any losses accrued will ultimately be carried by the government.
Special factors must be considered in highly dollarized economies. The absence of a lender of last resort may make dollarized systems more prone to runs and make runs more difficult to stop. Policymakers have tried to tap dollar resources with a variety of means, including high liquidity requirements and prearranged credit lines, and from various sources, such as the foreign bank sector. These measures may be beneficial in the case of small banking sector problems, but insufficient in the case of a systemic run.
In highly dollarized economies, the constraints on liquidity support and on depositor protection will make administrative measures more likely. These measures include securitization of bank liabilities or restrictions on deposit withdrawals, often preceded by a short bank holiday to design and prepare implementation of the measures. Nominal losses to depositors should be considered only as a last resort if all other options cannot be implemented or have failed, because nominal losses will make any reintermediation in the financial system even more difficult to achieve (Box 4).
Blanket Guarantee
A blanket guarantee can help to stop runs on banks caused by loss of depositor confidence in the overall banking system. A blanket guarantee typically consists of an announcement by the government that it will ensure that all bank liabilities except capital and subordinated debt are honored. If a limited deposit insurance system is in effect, it would have to be suspended until the blanket guarantee is removed. A blanket guarantee gives the government some breathing space to develop and start implementing a comprehensive restructuring strategy.17
A blanket guarantee alone will not contain a liquidity crisis. If implemented, it needs to be part of a package of credible stabilization measures. Credibility depends in the first place on the government’s perceived ability to honor the guarantee. While the government need not have sufficient resources to redeem the entire stock of bank liabilities, fiscal capacity should be seen as sufficient to meet any expected calls on the guarantee. A blanket guarantee also requires strong political commitment. The government must be perceived to stand fully behind the guarantee. Public disagreements or expressions of doubt would undermine confidence in a guarantee. Legislation may be required to make it fully effective. The details of the guarantee must be clearly explained to the public. A guarantee will always be tested. When depositors and creditors see that their claims are protected even when banks are intervened, they tend to stay in the system.
A blanket guarantee may not be credible in the face of unsustainable public debt dynamics or in highly dollarized banking systems. If political uncertainties or the government’s financial capacity to deal with the crisis is in doubt, a blanket guarantee will not be credible.18 If dollarization is moderate, a guarantee could cover foreign currency liabilities in their domestic currency equivalent at market exchange rates. If dollarization is high, however, and creditors and depositors are fleeing the currency and the banking system, the government will need to show both its capacity (access to foreign reserves and lines of external credit) and willingness to handle foreign exchange withdrawals for the guarantee to be credible.
The coverage of a blanket guarantee must be clearly spelled out (see Table 2 and Appendix II for country experiences). While shareholders and subordinated debt holders should not be covered, depositors and other creditors of locally incorporated banks (including branches of foreign banks) are typically included regardless of residency criteria and currency denomination.19 Off-balance-sheet items become covered when they enter the balance sheet.20 Insider claims are not typically covered.21 Important groups of near bank deposit-taking institutions may also be included.
Blanket Guarantees: Overview of Their Implementation
Blanket Guarantees: Overview of Their Implementation
| Country | Date of Introduction | Date of Removal | Removal Deadline | Previous Deposit Insurance Arrangements | Arrangements Agency | Fiscal Resources (US$billion) |
|---|---|---|---|---|---|---|
| Finland | February 2, 1993 | December 8, 1998 | … | Explicit | Government Guarantee Fund | … |
| Indonesia | January 27, 1998 | In place | Not yet announced | Implicit | Bank of Indonesia | 40.0 |
| Jamaica | January 1997 | August 1998 | … | Implicit | Financial Sector Adjustment Company | 1.8 |
| Japan | June 1996 | In place | April 2002 partial April 2003 | Explicit | Deposit Insurance Corporation | 500.0 |
| Korea | November 1997 | December 2000 | December 2000 | Explicit | Korea Deposit Insurance Corporation | 22.0 |
| Malaysia | January 1998 | In place | Not yet announced | Implicit | Danamodal | 7.1 |
| Mexico | April 1995 | In place | To be phased out by 2004 | Explicit | FOBAPROA | … |
| Bank Savings Institute (IPAB) | ||||||
| Temporary Capitalization Program | ||||||
| Sweden | December 18, 1992 | July 1, 1996 | … | Implicit | Bank Support Agency | … |
| Thailand | August 1997 | In place | Not yet announced | Implicit | Financial Institutions Development Fund | 34.0 |
| Turkey | December 2000 | In place | Not yet announced | Implicit | Savings Deposit Insurance Fund | … |
Blanket Guarantees: Overview of Their Implementation
| Country | Date of Introduction | Date of Removal | Removal Deadline | Previous Deposit Insurance Arrangements | Arrangements Agency | Fiscal Resources (US$billion) |
|---|---|---|---|---|---|---|
| Finland | February 2, 1993 | December 8, 1998 | … | Explicit | Government Guarantee Fund | … |
| Indonesia | January 27, 1998 | In place | Not yet announced | Implicit | Bank of Indonesia | 40.0 |
| Jamaica | January 1997 | August 1998 | … | Implicit | Financial Sector Adjustment Company | 1.8 |
| Japan | June 1996 | In place | April 2002 partial April 2003 | Explicit | Deposit Insurance Corporation | 500.0 |
| Korea | November 1997 | December 2000 | December 2000 | Explicit | Korea Deposit Insurance Corporation | 22.0 |
| Malaysia | January 1998 | In place | Not yet announced | Implicit | Danamodal | 7.1 |
| Mexico | April 1995 | In place | To be phased out by 2004 | Explicit | FOBAPROA | … |
| Bank Savings Institute (IPAB) | ||||||
| Temporary Capitalization Program | ||||||
| Sweden | December 18, 1992 | July 1, 1996 | … | Implicit | Bank Support Agency | … |
| Thailand | August 1997 | In place | Not yet announced | Implicit | Financial Institutions Development Fund | 34.0 |
| Turkey | December 2000 | In place | Not yet announced | Implicit | Savings Deposit Insurance Fund | … |
Any guarantee involves moral hazard, which can be contained if the guarantee is properly designed and managed. A blanket guarantee is a temporary assurance to stop runs. A guarantee does not imply that banks will not require intervention but rather provides a necessary depositor and/or creditor protection for undertaking far-reaching bank restructuring without ditsurbing market confidence. As long as shareholders understand that they will lose their investments if the bank is improperly run, moral hazard can be contained. At the same time, various restrictions may be used to limit reckless operations by banks, such as eliminating coverage of deposits on which excessively high interest rates are paid, and imposing a fee for the guarantee.
Dealing with Bank Runs in Dollarized Economies
A high degree of dollarization can have a significant impact on the policy options available to address bank runs because it imposes limits on the authorities’ ability to recognize and address emerging banking sector problems. This issue emerged forcefully during the recent Latin American crises but had also been a concern in earlier episodes of distress. Specific constraints imposed by dollarization include the absence of an effective lender of last resort or a blanket guarantee to respond to bank runs, which in turn can contribute to the size of the problem by increasing the speed and depth of deposit withdrawals.
Policy Environment with Dollarization
Dollarization complicates the task of addressing deposit runs, with partial dollarization creating the most challenges. For example, with dollarization there may be considerably more uncertainty about the extent of the crisis. In particular, in a partially dollarized system, with currency mismatches on the asset and liabilities side of bank balance sheets, rapid changes in the exchange rate may cause sudden changes in banks’ net worth that are difficult to detect promptly. In addition, dollarization may have an impact on credit risk if companies or individuals that borrowed in dollars are unable to generate dollar earnings.
More important, dollarization severely limits the availability of policy tools. It imposes limits on the possible extent of liquidity support—there is no unlimited lender-of-last-resort facility in foreign currency. The knowledge of rigid limits on possible support payments may make bank runs in highly dollarized economies more likely to occur and, once under way, more self-sustaining. Similarly, if the banking system is sizeable, the government may well be unable to provide a credible blanket guarantee for foreign currency deposits. In partially dollarized economies, while lender-of-last-resort facilities and a blanket guarantee may be provided in domestic currency, the relatively smaller domestic monetary base (compared to nondollarized countries) might lead to a very fast and high domestic liquidity expansion and/or an even faster collapse of the exchange rate. In addition, the nature of the guarantee is inconsistent with depositors’ currency preferences, and it is likely to be less effective than a guarantee in foreign currency would have been.
Partially or fully dollarized economies also face important constraints on supporting macroeconomic policies. Dollarization makes monetary policy less effective, and the need to maintain convertibility at par between local dollars and U.S. dollars removes the exchange rate as an equilibrating mechanism. Dollarization complicates fiscal support for the bank restructuring process because funds needed for any public recapitalization program may have to be committed in foreign currency to avoid asset/liability currency mismatches, and may not be consistent with a sustainable path of foreign debt.
Finally, it would appear that alternatives to lender of last resort and other safeguards—for example, the presence of foreign banks, prearranged credit lines, or sizeable earmarked reserves—are beneficial mainly in the case of smaller banking sector problems.1 In the case of systemic runs in dollarized economies, these safeguards may not be enough: the presence of a significant share of foreign institutions does not necessarily limit the extent of bank runs (Uruguay 2002, Argentina 2002). While the perceived support of the parent typically isolates foreign subsidiaries from a systemic run at the outset of a crisis, such support may not always materialize—particularly when the crisis proves deep and the authorities’ response slow, uncertain, or confiscatory. Under these circumstances, depositor runs are likely to extend to foreign banks. Prearranged credit lines may become unavailable in the time of most urgent need due to strong built-in safeguards. Sizeable earmarked reserves can stabilize expectations, but unless they extend to cover a significant portion of banking sector liabilities, there will be residual uncertainty and the perception of risk.
Modified Policies to Stop Bank Runs in Dollarized Economies
The additional constraints and limited role of safeguards call for more attention to financial sector soundness in dollarized economies. Where, in spite of such added vigilance, bank runs do occur, theory and country experiences suggest that a range of additional issues needs to be taken into account when formulating an appropriate strategy.
Provision of Emergency Liquidity
Liquidity support was identified as a key element in addressing systemic runs. In a dollarized economy, dollar liquidity support will be limited by the availability of resources, and it might be necessary to have a more clearly defined strategy on liquidity provision. Dollar assistance can be provided as long as sufficient dollar resources are available. Sources of liquidity can include high international reserves prior to the run (e.g., Argentina 1995); contingent facilities (swaps, repos); international financial support or bail out (e.g., Mexico 1995); or liquidity from shareholders (e.g., Uruguay 2002). Drawing down such resources can, however, have costs in terms of credibility and limitations to macroeconomic policymaking.
With limited dollar assistance possible, a decision needs to be taken whether to shift assistance in part or fully to local currency (e.g., Bulgaria 1996, Russia 1998, and Ecuador 1999). This option avoids the need for other stop-gap measures (see also below) and circumvents constraints on the lender of last resort as the central bank can print local currency. The expansion in the money supply and the currency mismatch, however, are likely to result in a combination of loss of international reserves, exchange rate depreciation, and inflation.
Depositor Protection
Depositor protection, in particular through a blanket guarantee, has proved to be of key importance in addressing systemic bank runs. In a dollarized economy, funding constraints emerge similar to those discussed above in the case of liquidity support.
A guarantee of deposits in foreign currency can be effective as long as sufficient dollar backing is available. In some countries, most notably the transition economies, banking systems tended to be small enough that a full blanket guarantee of foreign currency deposits might have been an option. In most others though, reserve and fiscal constraints would undermine the credibility of a blanket guarantee issued in foreign currency.
A blanket guarantee in local currency is unlikely to instill depositor confidence because depositors will demand their original dollars rather than local currency counterpart. Moreover, the announcement of a local currency guarantee may aggravate the crisis if depositors fear the banking system does not have sufficient dollar resources and therefore seek to buy dollars in the exchange market. Guaranteeing deposits in local currency at a fixed exchange rate (Russia 1998) implies a “haircut” to depositors when the exchange rate depreciates, and depositors are therefore likely to withdraw their funds and convert them to foreign exchange as soon as possible.
One option for depositor protection would be to transfer deposits in failed banks to a bank (public or private) that is perceived to be sound and with significant foreign exchange holdings backed up by a corresponding amount of central bank securities to ensure liquidity. In conditions of a systemic run, such banks will only exist in rare cases.
Administrative Measures
Given the added constraints of dollarization for liquidity support and depositor protection, administrative measures are more likely to be needed. Measures mentioned in the text—securitization of deposits, the extension of deposit maturities, or the imposition of bank holidays or other restrictions on deposit withdrawals—are also available in dollarized settings. Similarly, the general principles stressed for general bank runs (e.g., uniformity of policies applied to banks) to minimize the fallout from such policies apply. These measures, while allowing some breathing space, may have a potential, however, to delay the success of longer-term reintermediation efforts.
1 The issue is also relevant in economies with currency boards, which face similar challenges.Blanket guarantees distort market discipline and should be replaced with limited deposit insurance as conditions permit. The blanket guarantee should be removed once the banking system is sound. The market should be given advance notice of the removal (possibly 6–12 months). The guarantee may be phased out by lifting the guarantee of deposits of the more sophisticated creditors (e.g., interbank creditors followed by large creditors) and, once the system has been shown to be stable, the guarantee on all depositors.
Bank Interventions or Closures
As part of the policies for immediate stabilization, clearly insolvent or nonviable banks should be intervened or closed at an early stage. If the market suspects the existence of insolvent banks, failure to take action could undermine efforts to stabilize market expectations. Up-front takeovers or closures of insolvent banks are drastic actions that give other banks notice that the situation is serious and the authorities are taking appropriate measures. By removing the worst banks, market distortions and excess capacity are removed from the banking system to the benefit of remaining banks.
If a credible blanket guarantee is in place, such early closures should not affect confidence negatively. The operation might be successful even lacking such a guarantee if the prevailing view of the market is that the targeted institutions are the only insolvent ones. If this is not the case, however, upfront bank takeovers or closures with losses imposed on creditors and uninsured depositors could aggravate bank runs on the banking system as a whole.22
Reflecting fears about these pitfalls, few countries have actually imposed losses on creditors and depositors in a systemic crisis when closing banks. While suspensions or closures of banks and other financial institutions were part of the initial measures to stabilize the banking systems in Indonesia, Korea, and Thailand, only Thailand imposed losses on some creditors of the first group of 16 suspended finance companies (near-banks). In Indonesia, the creditors and depositors of the 16 banks that were closed initially were subsequently all compensated in full (Box 5).
Administrative Measures
If a package of market-oriented stabilization measures is not feasible or fails to contain the crisis, the authorities may have to adopt administrative measures to avoid losing monetary control. Such measures include securitization of deposits, forced extension of maturities, or a deposit freeze. Administrative measures are extremely disruptive and should be used with caution. They can cause major economic and political disruption, and therefore must be viewed as a final, desperate measure to stop a run on banks if all other measures fail. Banks should be treated uniformly when administrative measures are applied. Differential treatment could worsen the banking crisis and may impede restoration of financial services. The administrative measures may cover all deposits or may allow for a small amount of funds to be withdrawn.
Securitization of bank deposits would halt the deposit drain but allow some liquidity for frozen deposits. Such securities might be used for payments or as collateral for bank loans. To the extent that negotiable government bonds are used uniformly for different categories of deposits, market price quotations could be expected to emerge. Such bonds could therefore be used at their discounted market values for selected payments. Another option would be to allow individual banks to issue their own negotiable certificates of deposits against the frozen deposits. Such negotiable certificates could also become eligible payment instruments at their discounted market values. Bank certificates would probably be too heterogeneous, however, to allow clear market prices to emerge (compared to government bonds) but could give individual banks flexibility to redeem them early and thus limit the fiscal burden.
Indonesian Bank Closures—Doing It Wrong and Doing It Right
Intensified bank runs followed the closure of 16 small, deeply insolvent Indonesian banks on November 1, 1997. Partial guarantees of deposits of the closed banks, the perception that other weak banks remained in the system, a loss of confidence in the government’s overall economic management, and currency flight all fueled the runs. This experience underscores the need for closures during a systemic crisis to be part of a comprehensive restructuring strategy that is clearly explained to the public, with sound macroeconomic policies in place.
A second round of closures was undertaken on March 13, 1999. This took place concurrently with the takeover of seven banks by the Indonesian Bank Restructuring Agency (IBRA), the designation of nine other banks as eligible for public contribution to recapitalization, and the announcement of fit and proper reviews of banks viewed as viable. These actions, which resulted from the assessment of the condition of private banks that began in the fall of 1998, were taken against the background of the January 1998 three-point plan of a blanket guarantee, the creation of IBRA, and the introduction of a framework for corporate restructuring.
This second round of closures involving 38 banks was managed so that most deposits were transferred over the weekend of March 13, thus resulting in minimal disruption for depositors. The interventions and closures were well publicized through the electronic and print media, with customers getting full information about how to access their funds at banks receiving the transferred deposits. The combination of decisive action clearly communicated to the public, the existence of a credible guarantee, and evidence of a comprehensive approach to the private banks all contributed to the orderly exit of insolvent banks with minimal disruption.
Extension of deposit maturities is an alternative to securitization. While deposits with extended maturities are frozen into the banking system and not converted into negotiable instruments, measures should be adopted to improve the functioning of the payment system under such a deposit freeze. Checks could be issued on frozen deposits to give them full mobility within the banking system, and limited weekly or monthly withdrawals could also be allowed. This way individuals and corporations with frozen balances could still carry out transactions. The authorities should be aware that mobility within the system may lead to a flight to quality, as depositors move frozen deposits to the strongest banks in the system.
Undefined limitations on deposit withdrawals without preestablished rules of the game—e.g., freezes without specified minimum weekly or monthly withdrawal amounts, or outright bank holidays—are an unstable solution to deposit runs and should be avoided. If implemented, they should only be in place for limited time periods, to buy the authorities time to work out a permanent solution. Also, exit policies should be prepared in advance. The cases of Ecuador in 1998 (Box 6) and Argentina in 2002 highlight the disruptive effects of prolonged deposit freezes.
The introduction of capital or exchange controls may slow a run on the currency and the liability base of the banking system. For such controls to be effective, they must be comprehensive, fully enforced, and part of a broader policy package. Experience indicates that any beneficial effects of capital controls are bound to be temporary because they encourage circumvention and discourage legitimate transactions. They may also negatively affect market confidence. Moreover, if the private sector is shifting its portfolio from local currency to foreign-currency assets, controls would have to prohibit the holding of foreign currency, a measure that has not been successful in the past. In practice, few countries have imposed such controls during banking crises.23
Burden Sharing and Depositor Protection
One often politically contentious issue in banking crisis management and resolution relates to burden sharing. A banking crisis reflects losses (many concealed) to banks and their borrowers, which have to be shared by the different stakeholders, that is, by shareholders, holders of subordinated debt, depositors, other creditors, borrowers, and the government—ultimately the taxpayers. Shareholders and subordinated debt holders should absorb losses first. Losses, however, typically exceed their investment, often by very large amounts. Therefore, losses will have to be covered by some combination of creditors, depositors, and the government. Legislation normally stipulates preferences for the liquidation of different classes of creditors and depositors.24 What may work for individual banks in normal times, however, may not work in a systemic situation.
Deposit Freeze: The Case of Ecuador
Early in 1999, confidence in the Ecuadoran banks deteriorated rapidly. The closure of one bank in August 1998, followed by the intervention of the country’s largest bank in December 1998, and the failure of five small banks during the first two months of 1999 led to depositor unrest. Loss of confidence was exacerbated by the sluggishness of the newly created Deposit Guarantee Agency in starting payments under the blanket deposit guarantee approved in December 1998. Large capital outflows and historically low oil prices resulted in a depreciation of the local currency by 54 percent during 1998 and by a further 200 percent during 1999.
Against this background, the country’s second largest bank became illiquid. Reflecting the fear of a systemic meltdown, on March 5, 1999 the government declared a banking holiday rather than close the bank. The ensuing general strike resulted in the paralysis of the country’s economic activity. Private sector confidence worsened, leading to pressures on the currency, and the government decided to freeze most of the country’s deposits before reopening the banks on March 11, 1999.
When the banks reopened, there were withdrawal pressures, but this was limited by the deposit freeze and the general strike. The freeze accomplished its immediate objective of relieving pressure on the banks; however, their financial position remained precarious.
Efficiency of the Response to the Runs
Political pressures led to the progressive easing of the freeze: exemptions were declared for pensioners and seriously ill people; the unfreezing of dollar deposits was accelerated by six months; and banks were allowed to issue negotiable certificates of frozen deposits, which could be used as payment for durable goods, real estate, and loan cancellation. As the freeze was progressively eased, deposit runs reemerged. The pressure from the easing of the deposit freeze resulted in a second currency crisis in the last months of 1999, prompted by the default on external debt in September 1999.
Post-Run Developments
Faced with an intensifying currency crisis, the authorities opted for dollarization in January 2000. Dollarization, together with the announcement of an agreement with the IMF for a Stand-By Arrangement, permitted the unfreezing of most time deposits in March 2000. Most private banks engaged in an aggressive campaign to retain time deposits. As a result, the banking system was able to retain most of the liberated deposits, although some flight to quality occurred.
Allocation of nominal losses to depositors and other creditors is very delicate in a weak banking system, as it can exacerbate private sector fears. Creditors and depositors of banks perceived to be weak will run to stronger banks—or from the domestic banking system altogether. Furthermore, nominal loss sharing is potentially difficult because it could run counter to earlier policy assurances, implicit or explicit, or simply affect too many politically influential interests. The most common option for burden sharing is to reduce the real value of liabilities through high inflation and negative interest rates in real terms.
Where a government cannot extend a credible blanket guarantee and the run cannot be stemmed, alternative measures will be required. Such measures alter the burden sharing dramatically compared to the blanket guarantee because, typically, the depositors are hit the hardest. As discussed above, measures could include securitization of deposits, some form of withdrawal restriction, such as forced lengthening of maturities, a deposit freeze, or a uniform reduction in nominal balances. Common to these types of measures is that the exact outcome of the loss-sharing arrangement is not explicitly known upfront but will emerge over time.