50363 NOTE NUMBER 4 PUBLIC POLICY FOR THE PRIVATE SECTOR JUNE 2009 Blanket Guarantees FINANCIAL AND PRIVATE SECTOR DEVELOPMENT VICE PRESIDENCY Erik Feyen and Necessary during the Crisis, but What Next? Dimitri Vittas The exp ansion o f d e p o s it ins ur a nc e a nd int r o d uc t io n o f d e b t Erik Feyen (efeijen@world bank.org) is a financial guarantee s have p la y e d a c r uc ia l r o le in c o nt a ining t he f ina n c i a l economist, and Dimitri crisis while giving g o v e r nm e nt s t im e t o d e v e lo p s uit a b le p ol i c y Vittas (dvittas@world resp onse s. But t he s e m e a s ur e s d o no t a d d r e s s t he r o o t c a us e s o f bank.org) is a consultant, in the Financial and the crisis, and t he y le a d t o c o m p e t it iv e d is t o r t io ns , m o r a l h a z a r d , Private Sector Develop- and large f iscal c o nt ing e nt lia b ilit ie s . R o lling t he m b a c k is l i k e l y t o ment Vice Presidency of the World Bank Group. req uire an inter na t io na lly c o o r d ina t e d e f f o r t -- a nd a n a ns we r t o t h e imp ortant q uestio n, " e x it t o wha t ? " This is the fourth in a series of policy briefs on The basic objective of deposit insurance schemes But they are still rare in low-income countries, the crisis--assessing the is to avoid bank runs by small, uninformed where the institutional capacity for effectively policy responses, shedding depositors by protecting their deposits. Large operating such schemes is lacking. According to light on financial reforms currently under debate, depositors are left largely unprotected, because the International Association of Deposit Insurers, and providing insights they are deemed to be better informed and thus 119 countries had an explicit deposit insurance for emerging-market policy better able to exert discipline on banks. In nor- scheme in June 2008.1 makers. mal times deposit insurance schemes, coupled The effectiveness of deposit insurance schemes with lender-of-last-resort facilities, have a stabi- in preventing bank runs is seriously weakened when lizing effect by reducing the risk of contagion large depositors and other counterparties are to illiquid but solvent banks and facilitating the unable to assess the solvency of individual banks and THE WORLD BANK GROUP orderly resolution of troubled banks. They can lose confidence in the overall stability of the bank- also foster financial development and stability ing system.2 This becomes particularly problematic by encouraging small depositors to place their in systems where short-term wholesale funding is savings in bank accounts and strengthening the dominant. In such circumstances the offer of blan- competitive position of small, local banks. ket guarantees that extend protection to interbank Before 1980, explicit deposit insurance deposits and other bank debt may become neces- schemes were rare, found mainly in the United sary. The guarantees seek to restore investor con- States, Canada, and several Latin American fidence while avoiding the systemic implications of countries. Following financial crises in the 1970s closures and liquidations. Once issued during a cri- through the 1990s, they were gradually adopted sis, blanket guarantees give governments time and in most European and Asian countries as well. opportunity to resolve failing banks, recapitalize BLANKET GUARANTEES NECESSARY DURING THE CRISIS, BUT WHAT NEXT? and restructure viable entities, and liquidate those coverage of secured debt defeats the purpose of deemed to be nonviable. These guarantees were recourse to the underlying asset. successfully used in the 1990s by both high-income The ability of deposit insurance and blanket countries (such as Finland, Japan, and Sweden) guarantees to prevent bank runs depends largely and emerging economies (such as Indonesia, the on their credibility, based on the schemes' access Republic of Korea, and Mexico). to adequate funding. Schemes are funded by pre- miums levied on participating banks and some- Main characteristics and shortcomings times supplemented with public funds. Prefunding Essential features of explicit deposit insurance provides greater certainty but raises the issue of 2 schemes include the following:3 efficient management of accumulated funds. Compulsory participation to avoid adverse Deposit insurance schemes (whether explicit selection (in which only weak banks join the or implicit) generally weaken the incentive of scheme)4 depositors to monitor bank performance and Relatively low ceilings on insured deposits to pro- may thus induce moral hazard.7 To control this, tect the vast majority of small depositors while many schemes charge risk-based premiums. In still providing incentives to larger and better- principle, these should be high enough to dis- informed depositors to discipline banks5 courage weak banks from taking on excessive risk A coinsurance requirement, so that a specified and offering higher deposit rates. However, risk- fraction (from 10 to 25 percent) of insured based premiums are difficult to design in practice deposits is not covered, to further induce mar- and, because they typically need to be higher for ket discipline6 smaller local banks, often face strong political Timely access to insured funds to help prevent opposition. Indeed, although many countries bank runs have expanded deposit insurance coverage in The coverage of deposit insurance schemes recent years, few expansions were accompanied varies widely across countries. The schemes gen- by an increase in premiums. erally focus only on retail deposits, although they Blanket guarantees also have adverse side also cover interbank deposits in some low-income effects. Research has found that issuing such guar- countries in Africa and Asia. Local subsidiaries antees under lax policy interventions is associated of foreign banks are typically covered, but the with high fiscal outlays (Honohan and Klingebiel treatment of foreign branches varies widely. The 2003). Doubts about the credibility of blanket treatment of foreign currency deposits and of guarantees--especially when countries lack suf- nonresident deposits often depends on their rela- ficient foreign exchange reserves and have limited tive size and the impact that their abrupt with- fiscal resources or when there is uncertainty about drawal could have in causing a bank run. Where the terms of the guarantee (whether compen- there is widespread "dollarization," foreign cur- sation will be in local or foreign currency and rency deposits tend to be covered, although there whether an unfavorable exchange rate may be may be uncertainty about the currency in which used)--undermine their effectiveness in prevent- the compensation will be provided and the rate ing the outflow of foreign liabilities (Laeven and of exchange that will be used. Valencia 2008). To contain the risk of moral haz- Blanket guarantees offer unlimited protection ard and the fiscal cost, most governments limit for all nonretail deposits (wholesale and inter- the volume and duration of guaranteed debt, and bank) and generally cover new issues of short- they often impose a fee that can be linked to a and medium-term unsecured senior debt to stem bank's default risk. Guarantees are therefore tem- acute wholesale funding problems. Depending porary instruments intended to be used only as a on the severity of the crisis, existing debt may also last resort, and they are meant to be withdrawn be included. In some cases junior subordinated when normal market conditions are restored. or even secured debt, such as covered bonds, is included as well. This introduces additional Crisis responses market distortions, however, since junior debt Particularly after the collapse of Lehman Brothers is supposed to be inherently riskier, while the in September 2008, fears about systemic insta- bility spread because market participants were by the desire to avoid runs by large depositors, unable to distinguish between illiquid and insol- which represent a disproportionately large frac- vent financial institutions or to gauge the level of tion of total deposits. In the United Kingdom, for government support. Faced with eroding inves- example, 50 percent of bank deposits were held tor confidence, most high-income economies in large accounts not covered by the traditional expanded deposit insurance and introduced deposit protection scheme. In other cases cover- guarantees on bank debt. These emergency mea- age was raised because of competitive concerns. sures were replicated by some emerging econo- For example, Australia introduced a deposit mies as a prudential response to the deepening insurance scheme after other high-income coun- 3 global recession and the reversal of capital flows tries in the region had raised their own cover- that accompanied the deleveraging process. age, generating fears of destabilizing cross-border deposit flows. Responses in high-income economies Excluding blanket guarantees on deposits, Many high-income economies raised the ceil- the average deposit coverage as a multiple of ing on insured deposits in addition to providing GDP per capita increased from roughly 1.5 to ample liquidity support to the financial sector 5 on average. The United States raised the ceil- (figure 1). In most cases this action was motivated ing on insured deposits from US$100,000 to Figure Increases in retail deposit insurance coverage as of January 2009 1 Previous ceiling Luxembourg Current ceiling Coverage ceiling as a multiple of 2007 GDP per capita Unlimited coverage Coverage ceiling in 2008 US$ (thousands) Czech Republic Finland Philippines Sweden Ukraine United Kingdom Romania Russian Federation Russian Federation Netherlands Bulgaria Austria Lithuania Czech Republic Finland Spain Luxembourg Hungary Hungary United States Sweden Portugal United Kingdom Ukraine Austria Philippines Netherlands Romania Portugal Lithuania Spain Bulgaria United States Australia Australia Denmark Denmark Germany Germany Greece Greece Hong Kong (China) Hong Kong (China) Iceland Iceland Ireland Ireland Malaysia Malaysia New Zealand New Zealand Singapore Singapore 0 3 6 9 12 0 50 100 150 200 250 Sources: International Association of Deposit Insurers; International Monetary Fund. BLANKET GUARANTEES NECESSARY DURING THE CRISIS, BUT WHAT NEXT? US$250,000.8 The European Union raised the Responses in emerging economies minimum ceiling from 20,000 to 50,000 in late Most countries in Central and Eastern Europe 2008 and to 100,000 by the end of 2009, a change expanded their deposit insurance coverage that some member states adopted immediately. because of their dependence on foreign fund- In other cases the authorities provided insurance ing to finance domestic credit expansion and coverage for the first time, extended full blanket the importance of Western European banks in guarantees on deposits, or did both. their financial system. Some countries in the In addition, countries extended guarantees to Middle East (Jordan, Kuwait, Saudi Arabia, the other bank liabilities in systems with large whole- United Arab Emirates) and East Asia (Indonesia, 4 sale exposures so as to prevent the nonrenewal Malaysia, the Philippines, Thailand) took simi- of contracts--or "silent runs"--by institutional lar measures as a precautionary or competitive investors and other counterparties. The scope and response. Interestingly, these moves have not conditions of access to such guarantees differ sub- been matched by Latin American, South Asian, stantially among countries (table 1). For example, or African countries, presumably because of their Australia, Korea, the Netherlands, New Zealand, stronger starting positions or lower financial inte- and Spain offered guarantees only on new senior gration. unsecured debt issues. Eligible maturities typically Few emerging economies have taken measures range up to five years, although in some cases no beyond deposit insurance, either because of the maturity limit has been specified. Other countries, significant fiscal costs entailed or because their such as Denmark and Ireland, went even further banking systems have limited wholesale expo- and extended coverage to interbank deposits and sures. Heavily affected countries in Central and existing and new senior unsecured debt. Ireland Eastern Europe are an exception. For example, also guaranteed covered bonds and even sub- Hungary has extended guarantees to interbank ordinated debt. Guaranteed debt issuance by loans and new bank debt issues for up to five banks has grown significantly--primarily in the years. However, such guarantees are unlikely to United States, the United Kingdom, Germany, be credible in countries with sizable public sector and France--with overall gross issuance exceed- debt, a large banking system, and an open capital ing 200 billion by early March 2009. account. Selective controls on capital outflows, Table Overview of bank debt guarantee programs in selected countries 1 Feature Details United States--Temporary Liquidity Guarantee Program Coverage Program will guarantee all new senior unsecured debt issued by banks, bank holding companies, and thrifts on or before June 30, 2009, for a period of 3 years. Eligibility Participation by default, unless institutions opt out. Pricing Flat-fee pricing structure based on maturity of the instrument: 180 days or less, 50 basis points (bps); 181­364 days, 75 bps; 365 days or more, 100 bps. United Kingdom--Credit Guarantee Scheme Coverage Scheme will guarantee short- to medium-term debt issuance of up to 36 months, including certificates of deposit (CDs), commercial paper, and senior unsecured bonds and notes. Eligibility Participation is conditional on the institution having Tier 1 capital in an amount and at a date to be specified by the government. Pricing Pricing structure is based on the median 5-year credit default swap spread of the institution calculated over the period July 2007­July 2008, plus 50 bps. Germany--Stabilization Fund Coverage Fund will provide up to 400 billion in guarantees for interbank loans of up to 36 months and for newly issued bank debt. Eligibility State-owned and commercial banks, including subsidiaries of foreign banks. Pricing Flat-fee pricing structure for short-term debt (3­12 months): 10 bps to issue government-guaranteed bonds plus 50-bps fee. Risk-based pricing structure for longer-term debt (1­5 years): the median 5-year credit default swap spread of the institution calculated over the period January 2007­August 2008, plus 50 bps. Sources: Deutsche Bank; Fitch Ratings; International Monetary Fund. while not yet broadly evident in this crisis, have banks has introduced additional complexities, been introduced in a few countries. The pace of such as providing deposit insurance coverage such measures may quicken as more countries try for the foreign activities of domestic banks in to prevent "leakages" of domestic liquidity injec- home jurisdictions11 as well as for the branches tions or become unable to credibly guarantee the of foreign banks in host jurisdictions.12 Iceland, liabilities of their domestic banking industry. whose deposit insurance scheme could not sup- port the large offshore deposits of its collapsed Policy issues domestic banks, is a case in point. The crisis has raised important issues relating to 5 the design of deposit insurance schemes, while Distortionary effects of emergency measures the use of emergency measures has led to ques- The emergency policy measures have done tions about distortionary effects and how best to much to help contain the crisis. But they have roll back these measures. also led to significant distortions in the competi- tive landscape that will eventually need to be Design of deposit insurance schemes resolved. For example, cross-border financial The crisis has exposed flaws in the structure of links motivated several countries to expand deposit insurance schemes that may compromise their deposit insurance schemes primarily out both their ex ante deterrent role and their ex of competitive (rather than prudential) con- post role in orderly and efficient bank resolu- cerns, leading to a "race to the bottom" in fully tion. In particular, the crisis has led to a growing guaranteeing coverage. realization that the use of coinsurance require- Moreover, many countries have not yet charged ments created insecurity among depositors and banks a higher premium for the expanded deposit weakened the protection from bank runs. This insurance coverage. Doing so entails problems of was highlighted by the case of Northern Rock in defining an appropriate pricing mechanism and the United Kingdom. Although the proximate may create additional stress for weaker banks. But cause of the bank's failure was massive with- the absence of risk-based pricing increases moral drawal or nonrenewal of wholesale funding, hazard and may lead to excessive risk-taking by the lines of anxious depositors forming outside weaker banks. its branches prompted the U.K. authorities to Inappropriate pricing of bank debt guaran- announce the removal of coinsurance. The new tees has also given rise to distortions. For exam- EU rules prescribe abandoning coinsurance, ple, in contrast to the U.K. scheme, under the which most member states have already done, U.S. scheme banks are charged a small, flat fee or reducing it to no more than 10 percent of for their guaranteed debt issuance despite sig- insured deposits. nificant differences in their risk profiles. The The crisis has also underscored the impor- guarantees have effectively become a cheap tance of timely access to insured deposits. In the source of funding for many banks, crowding United States federal law has long required pay- out other, nonguaranteed debt issuance,13 cre- ment of insured deposits as soon as possible after ating an uneven playing field for financial insti- the failure of an insured bank. The authorities tutions without access to them,14 and, together in some other countries have announced plans with other measures, helping to keep weaker to expedite access to insured deposits in cases banks afloat. of bank failure. In the European Union payouts More broadly, the large-scale emergency mea- used to be made after three months, but recent sures adopted will inevitably shape the expecta- plans contemplate allowing access to funds after tions of banks and investors about government only three days. interventions in future financial crises. These More broadly, the crisis has shown that deposit expectations could influence their behavior, insurance schemes have not kept pace with diminishing the monitoring incentives of inves- changes in the size and structure of the banking tors while inducing moral hazard by banks. This industry9 or in the composition of bank fund- reinforces the challenge of developing a more ing.10 The increased cross-border expansion of effective regulatory framework going forward. BLANKET GUARANTEES NECESSARY DURING THE CRISIS, BUT WHAT NEXT? Exit strategies important weaknesses of deposit insurance The distortions and fiscal risks introduced by bank schemes--ranging from coverage arrangements debt guarantees and extended deposit insurance to risk-based pricing--that need to be addressed. arrangements call for their early removal. But this The central question, however, is what safety net task can only be undertaken once the conditions design will be credible after these emergency for a sustainable recovery are in place. Sunset measures are suspended. In particular, future clauses and an explicit timeline can bring clarity, reforms will need to address the dynamic trade- but their credibility would suffer if repeated exten- offs between the ex ante deterrent role of such sions become necessary. Instead, governments schemes, with its emphasis on reducing moral 6 could choose a gradual phaseout and unwind- hazard, and their ex post role of facilitating effi- ing of emergency measures while the economy cient bank resolution and minimizing depositor improves, problem institutions and assets are losses. sufficiently addressed, and adequate regulatory frameworks are put into place. However, the cross-border financial links mean that the roll- ing back of emergency measures may take longer Notes and require international coordination to ensure The authors would like to thank Diego Sourrouille a reasonably level playing field once sound condi- and Haocong Ren for research assistance, and Laura tions return in most countries. Ard, Thorsten Beck, Asli Demirgüç-Kunt, Neil Gregory, James Hanson, Alain Ize, Luc Laeven, Roberto Rocha, Conclusion Augusto de la Torre, and especially Constantinos The expansion of deposit insurance and intro- Stephanou for valuable discussions or comments. duction of debt guarantees have played a crucial 1. Where explicit deposit insurance arrangements are role in containing the crisis while giving govern- not in place, as is the case in many, particularly low- ments time to develop suitable policy responses. income, countries, this does not mean that deposits are These measures are relatively easy to implement, not implicitly protected by the government. have no significant up-front fiscal costs, and are 2. See Garcia (2000) for a general discussion on important commitment devices because they deposit insurance in financial crises, and OECD (2008 signal the willingness of the authorities to take and 2009) for recent experiences and considerations action to avert a financial market meltdown. for financial safety nets. The extraordinary circumstances in which pol- 3. See BCBS and IADI (2009) and Demirgüç-Kunt, icy makers found themselves in late 2008, when Kane, and Laeven (2008) for an overview of design faced with the potential collapse of the financial principles of such schemes. system, make objections about the desirability of 4. Most, but not all, countries that have adopted introducing them a moot point. explicit deposit insurance schemes impose compulsory Such arrangements are no panacea, however, participation. and they come at a cost. They do not address 5. In practice, the application of the ceiling on a per- the root causes of the crisis, and they lead to depositor, per-bank basis allows individual depositors competitive distortions, moral hazard, and to shield large sums of money by maintaining deposits large fiscal contingent liabilities. They need to with multiple banks. be promptly phased out once concerted action 6. Coinsurance is predicated on the argument that the has been taken to reestablish the soundness of fear of some deposit loss would force even small deposi- banks and restore confidence in the financial tors to monitor the performance of individual banks system. Because competitive pressures stemming and thus limit moral hazard. from increased cross-border financial links com- 7. Empirical research underscores the moral hazard plicates their rollback, the exit strategy will likely problem. Countries with ill-designed deposit insurance require an internationally coordinated effort systems have a higher probability of experiencing a and may take longer to complete. banking crisis, especially if they have a poor institu- Just as important as the exit strategy is the ques- tional and contracting environment (see, for example, tion, "exit to what?" The crisis has underscored Demirgüç-Kunt and Detragiache 2002). 8. The U.S. authorities also extended unlimited insur- References ance to non-interest-bearing transaction accounts, often BCBS (Basel Committee on Banking Supervision) and used by small and medium-size enterprises for payroll IADI (International Association of Deposit Insur- activities, and have insured outstanding balances by ers). 2009. Core Principles for Effective Deposit Insurance money market mutual funds, as losses suffered by some Systems. Basel. funds on supposedly prime-quality investments created Demirgüç-Kunt, A., and E. Detragiache. 2002. "Does concerns about massive withdrawals. Deposit Insurance Increase Banking System Stabil- 9. For example, the presence of a few very large banks ity? An Empirical Investigation." Journal of Monetary may limit the credibility of deposit insurance schemes Economics 49 (7): 1373­406. 7 in the absence of adequate prefunding arrangements, Demirgüç-Kunt, A., E. Kane, and L. Laeven. 2008. forcing the authorities to keep these banks alive, includ- "Deposit Insurance Design and Implementation: ing through forbearance. Policy Lessons from Research and Practice." In A. 10. The shift from retail deposits to other sources of Demirgüç-Kunt, E. Kane, and L. Laeven, eds., Deposit bank funding may limit the effectiveness of deposit Insurance around the World: Issues of Design and Imple- insurance and lead to unintended consequences. Con- mentation. Cambridge, MA: MIT Press. sider the U.S. bank IndyMac. At the time of its failure in Garcia, G. 2000. "Deposit Insurance and Crisis Manage- July 2008, it had borrowed large sums from the Federal ment." IMF Working Paper 00/57, International Home Loan Banks (FHLBs) to fund mortgage lend- Monetary Fund, Washington, DC. ing. Since this funding was overcollateralized and had Honohan, P., and D. Klingebiel. 2003. "The Fiscal Cost preferred-creditor status, the FHLBs were repaid ahead Implications of an Accommodating Approach to of the Federal Deposit Insurance Corporation (FDIC). Banking Crises." Journal of Banking and Finance 27 As a result, the IndyMac bankruptcy led to FDIC losses (8): 1539­60. estimated at US$10 billion (a third of IndyMac's size), Laeven, L., and F. Valencia. 2008. "The Use of Blanket illustrating the complex interplay between deposit Guarantees in Banking Crises." IMF Working Paper insurance and bank resolution. 08/250, International Monetary Fund, Washington, 11. To the extent that deposit insurance schemes cover DC. (but do not adequately price) the cross-border activities OECD (Organisation for Economic Co-operation and of a few banks, they can indirectly lead to the unfair Development). 2008. Financial Crisis: Further Issues subsidization of these banks' activities relative to those Regarding Deposit Insurance and Related Financial Safety of smaller domestic competitors. Net Aspects. Committee on Financial Markets back- 12. In the European Union foreign bank branches are ground document. Paris. covered by the deposit insurance scheme of their home ------. 2009. Expanded Government Guarantees for Bank country. Because the schemes are not harmonized and Liabilities: Selected Issues. Committee on Financial capital can move freely within the single market, this Markets background document. Paris. has led to competitive pressures during the crisis. By contrast, in most Latin American and Asian countries foreign bank branches are covered by the host country scheme. 13. There is anecdotal evidence of crowding out and higher pricing in the nonguaranteed bank debt markets as well as in some sovereign debt markets as a result of such guarantees. 14. This is particularly true for nonbank financial insti- tutions, which have lobbied heavily to be included in such schemes in countries (such as the United States) where they play an important role in the financial system. PRIVATIZATION TRENDS A RECORD YEAR IN 2006 crisisresponse The views published here are those of the authors and should not be attributed to the World Bank Group. Nor do any of the conclusions represent official policy of the World Bank Group or of its Executive Directors or the countries they represent. To order additional copies contact Suzanne Smith, managing editor, The World Bank, 1818 H Street, NW, Washington, DC 20433. Telephone: 001 202 458 7281 Fax: 001 202 522 3480 Email: ssmith7@worldbank.org Produced by Grammarians, Inc. Printed on recycled paper This Note is available online: http://rru.worldbank.org/PublicPolicyJournal