Financial Crises

Chapter 15. Crisis Management and Resolution: Early Lessons from the 2007–09 Financial Crisis

Stijn Claessens, Ayhan Kose, Luc Laeven, and Fabian Valencia
Published Date:
February 2014
  • ShareShare
Show Summary Details
Stijn Claessens, Ceyla Pazarbasioglu, Luc Laeven, Marc Dobler, Fabián Valencia, Oana Nedelescu and Seal Katharine The authors are grateful for guidance provided by Olivier Blanchard and José Viñals, input from Luis Cortavarria, and very useful comments from internal and external reviewers, including during an informal discussion at the IMF Executive Board.

This chapter compares the policy choices in the 2007–09 global financial crisis with those in past episodes and draws some preliminary policy lessons focusing mainly on crisis management tools and techniques.1 Country experiences in recent and past crises are examined with a particular focus on the extent to which policy choices were affected by initial conditions and the nature of the crisis. The chapter reviews the state of financial and operational restructuring during the 2007–09 crisis, as well as institutional reforms, in the light of lessons from past episodes, and provides policy implications for the near- and long-term agendas.

The 2007–09 crisis was unusual in its speed and breadth and the types of countries affected. Systemic crises—situations of significant stress in the financial sector, followed by significant policy interventions—often affect several countries at the same time. In the past, though, crises have been largely limited to specific regions or types of economies—the Nordic countries in the early 1990s, Latin America in the mid-1990s, Asia in the late 1990s, and the emerging market economies early in the first decade of the 2000s. The recent crisis was unusual in its global nature, affecting countries with a speed and virulence not seen since the Great Depression, with major advanced economies and countries recently acceding to the European Union (EU) most affected.

The chapter is organized as follows: The first section reviews the initial macroeconomic and financial conditions for two sets of countries, one drawn from past crises and the other from the 2007–09 crisis, and documents differences that help explain the choice of policies used by countries in the two samples. The second section reviews the policy choices made in the recent crisis countries in light of outcomes to date and shows how the choices involved trade-offs. The third section concludes with lessons on the best mix of policies for crisis management, the adequacy of current resolution toolkits to deal with systemic distress, and the structural reform agenda to reduce systemic risks.

What was Different in the 2007–09 Crisis?

Initial Macroeconomic and Financial Conditions

The study collected and compared two samples of crises—past (1991–2002) and recent (2007–09)—that, for each country, qualified as a systemic banking crisis (see Appendix 15A). The past crises occurred in 17 countries during the 1991–2002 period. The recent crises were the manifestations of the 2007–09 global financial crisis as it appeared in 12 countries.2

Table 15.1 provides precrisis macroeconomic, institutional, and financial data for the two samples of crisis countries (past crises and recent crises), with data as of one year before each crisis started. The past crises were most frequently among emerging and developing economies, whereas the recent crises were mostly among advanced economies. These advanced economies were, at the time, generally considered to have relatively strong institutional frameworks, a condition associated with a lower probability of a crisis. At the same time, financial systems in the sample of recent crisis countries were generally larger than those in countries in the sample of past crises, and larger systems suggest higher costs in the event of a crisis. Past crises, however, often involved twin crises (both a banking and a currency crisis), and many involved an IMF program.

Table 15.1Precrisis Indicators
Past crisesRecent crises




Medians (percent of GDP unless indicated otherwise)1
Institutional and Macroeconomic
Overall fiscal balance−
Gross public debt32.728.942.043.0
Current account−3.1−0.6−0.91.8
Private credit growth (median, t − 1 through t − 4)
Real estate prices (median growth, t- 1 through t − 4)−2.822.05.25.2
Stock prices (MSCI in US$, median percent growth, t − 1 through t − 4)18212626
Stock market capitalization293486102
Private bond market capitalization8394545
Real GDP growth (percent, t − 1)
GDP per capita (constant 2000 US$)4,52025,56127,56330,425
CPI inflation7.
Creditors’ rights (4 = best)22.533
Central bank independence (1 = best)0.630.610.830.83
Governance3 (2.5 = best)0.171.681.571.61
Deposit insurance in place (percent of cases)5075100100
Bank assets4880150156
Bank credit4372136160
Bank credit/deposits (percent)118126125124
Net interest margin (percent of revenues)5232
Size of five largest banks (global assets)23764307391
Market share of three largest banks48975862
Commercial banks/total system assets (percent)86766759
Cross-border claims91591114
Bank return on equity (percent)711111
Bank Z-score (standard deviations)
Core capital/assets (percent)9588
Other operating income/average assets (percent)2111
Loans/assets (percent)49695147
Liquid assets/customers and short-term funding (percent)23323737
Other funding/total liabilities (percent)
Sources: Thomson Reuters Datastream; Djankov, McLiesh, and Shleifer (2007); the Organization for Economic Co-operation and Development, the World Bank, and IMF World Economic Outlook for creditors’ rights; Crowe and Meade (2003) for central bank independence; World Bank for governance; Laeven and Valencia (2008) for deposit insurance; and BankScope and Organization for Economic Co-operation and Development for bank profitability. World Bank Financial Structure Database.Note: CPI = consumer price index; MSCI = Morgan Stanley Capital International stock market index

See Appendix 15A for data sample and crisis dates. All precrisis variables are measured as of one year before the start of the crisis, unless indicated otherwise. Past advanced economies are the Nordic countries and Japan; recent advanced countries are all but Latvia and Ukraine.

Median for advanced economies and Asian crisis episodes.

Aggregate of six dimensions of governance.

Japan only.

Sources: Thomson Reuters Datastream; Djankov, McLiesh, and Shleifer (2007); the Organization for Economic Co-operation and Development, the World Bank, and IMF World Economic Outlook for creditors’ rights; Crowe and Meade (2003) for central bank independence; World Bank for governance; Laeven and Valencia (2008) for deposit insurance; and BankScope and Organization for Economic Co-operation and Development for bank profitability. World Bank Financial Structure Database.Note: CPI = consumer price index; MSCI = Morgan Stanley Capital International stock market index

See Appendix 15A for data sample and crisis dates. All precrisis variables are measured as of one year before the start of the crisis, unless indicated otherwise. Past advanced economies are the Nordic countries and Japan; recent advanced countries are all but Latvia and Ukraine.

Median for advanced economies and Asian crisis episodes.

Aggregate of six dimensions of governance.

Japan only.

Credit growth and asset appreciation were stronger in the recent crises and accompanied by large external imbalances in some cases. A common precrisis condition in the past episodes was a large current account deficit (Table 15.1). Currency mismatches were often a significant source of vulnerability and losses once the exchange rate came under pressure. In the recent crises, fiscal and current accounts were, on average, close to balance, although there was, as with past crises, ample variation across countries, and global imbalances were large in absolute terms. Currency mismatches were important triggering and amplifying factors in some of the recent cases. Common to all precrisis periods was the presence of credit or asset price bubbles (Laeven and Valencia, 2008; and Reinhart and Rogoff, 2009). More pronounced in the 2007–09 crises were the sharp increases in leverage for households, also reflected in the higher increase in house prices.

Financial systems were much larger and more concentrated in the recent crises, and financial institutions much more complex. Global assets of the five largest banks were typically more than 300 percent of their home country’s GDP (Table 15.1). Many of these institutions enjoyed the benefits of being “too important to fail,” that is, borrowing at preferential rates, operating with higher levels of leverage, and engaging in riskier activities. These firms were often highly complex in their balancing of business, tax, and regulatory objectives. For efficiency purposes, management would operate through business lines from the center, with liquidity and treasury functions often centralized. Yet, for tax and regulatory reasons, trades and exposures would be booked wherever most profitable or efficient. Hence, the formal and de facto governance structures of many groups would diverge. In consequence, host and, in some cases, even home supervisors with formal oversight authority, were often at best only partially able to identify and assess a group’s financial health, including the adequacy of its capital and liquidity arrangements, the nature of risks, and the location of risks within the wider group.

High interconnectedness in the recent crises was enabled by financial innovations, especially securitizations and traded credit derivatives, and the expansion of the role of nonbank financial institutions known as the “shadow banking system.” Securitization created assets that were packaged, and then repackaged, into new layers such as collateralized debt obligations, and each new layer further spread risks while reducing the clarity of risk exposures. Credit default swaps led to a separation of credit risks on and off balance sheets, and facilitated the concentration of risks in single entities that went undetected. And although nonbanks were a cause of instability in some of the past crises, the role of the shadow banking system in the recent crises was much more important through banks’ use of conduits and collateralized transactions. Nonbank institutions like mortgage lenders, broker-dealers, and money market funds had grown rapidly, and some became systemically important. Large, internationally active financial institutions, large cross-border claims exceeding 90 percent of GDP (six time larger than in past crises), and complex linkages combined to transmit stress rapidly, most notably related to problems in the U.S. subprime mortgage market (Claessens and others, 2010).

As in past precrisis periods, traditional financial system indicators and favorable macroeconomic conditions obscured risks in the run-up to the recent crises. Conventional indicators of financial soundness based on balance sheet data showed precrisis bank profitability, liquidity, and capital to be higher than in the past, at least when comparing across only advanced economies. One exception was the increasing leverage of many large financial institutions, but that was generally thought to be consistent with improvements in risk management that allowed for the more efficient use of capital. An increased reliance on wholesale funding was also not generally considered a concern. Taken together, however, these misinterpretations meant an underestimation both of the risks accumulating outside banks’ balance sheets and of large-scale systemic liquidity withdrawals.

The Policy Responses

Policy responses in the recent crises were initially similar to those in past crises, but over time have diverged. Past crisis responses typically involved three phases: first, containment, to deal with acute liquidity stress and to stabilize financial liabilities; second, resolution and balance sheet restructuring, which involves removing insolvent financial institutions from the system and recapitalizing viable ones; and, finally, operational restructuring to restore the financial soundness and profitability of viable institutions and asset management to rehabilitate nonperforming loans. The recent crises followed this pattern through the first phase, but subsequent policy responses were less forceful, at least for the major countries.3

The sequence and range of policy responses in the past crises differed in important respects from those in the recent crises. Figure 15.1 provides the timing of interventions by depicting the evolution of liquidity support and the timing of guarantees and recapitalizations by governments around the onset of crises. Figure 15.2 compares the frequency of policies used as of end-2009. (Appendix 15B provides details on timelines and intervention measures.) As in past crises, liquidity support and guarantees were deployed in the early stages, although more extensively relative to GDP. However, this support was followed more rapidly in the recent crises compared with the past crises with recapitalization across the board in many countries, which mitigated the real effects of the crises. After these general interventions, policy approaches in the recent crises became less forceful than those typically followed in the past. In particular, progress with comprehensive operational and asset restructuring was slower.

Figure 15.1Timing of Interventions and Amount of Liquidity Support

Source: Laeven and Valencia, Chapter 2, in this volume.

Note: “Past crises” refers to pre-2007; “Recent crises” refers to crises since 2007. All dates are relative to crisis peaks, with periods referring to quarters before or after.

Figure 15.2Containment and Resolution Policies

(percent of sample countries adopting)

Source: Laeven and Valencia, Chapter 2, in this volume.

Note: “Past crises” refers to pre-2007 and “Recent crises” refers to crises since 2007.

However, in Iceland, Ireland, and Ukraine, the sequence and types of responses more closely resembled those of past crises, including due diligence on the viability of financial institutions and quality of assets, government recapitalization, removal of nonperforming assets, operational restructuring, and the adoption of IMF programs. The details of policy mixes varied, reflecting differences in the causes and severity of countries’ respective crises, including whether they also involved a currency or a sovereign debt crisis, types of defunct assets, and political economy considerations. In Iceland—and to a lesser degree in Ukraine, where foreign exchange exposures were large—wholesale funding runs and withdrawal of foreign capital led to crises and created pressures on currencies, which then reduced the repayment capacity of borrowers. In Ireland, problems were predominantly real estate–related and mostly affected commercial banks.

Liquidity and Other Central Bank Support Measures

With capital losses, severe funding pressures emerged relatively early on. Losses on securitized assets, reflecting expectations of default and deteriorating economic conditions, quickly appeared on institutions’ balance sheets (traded assets are normally marked to market or recorded at fair value). Price declines and ratings downgrades of securitized assets during 2007–08 quickly impacted firms’ capital through valuation losses. By end-2007, more than 70 percent of banks’ losses came from structured products and securitized positions. Because reputational risks compelled many banks to put previously off balance sheet obligations on their books, liquidity needs rose sharply. Given the large reliance on wholesale funding, liquidity needs spiked across many markets, and interbank market spreads widened dramatically. In tandem, asset prices and solvency positions worsened further.

Central banks responded quickly with liquidity support on a more massive and widespread scale than in the past crises (Stone, Fujita, and Ishi, 2011). They extended the duration of liquidity facilities and eased counterparty and collateral requirements. New facilities were established to alleviate liquidity shortfalls in specific markets. Liquidity support to financial institutions was accompanied or followed by large-scale asset purchases and other unconventional, quantitative interventions. The U.S. Federal Reserve and the Bank of England purchased large amounts of securities. The European Central Bank introduced a covered bond purchase program. Coordinated policy responses, unprecedented in many ways, then followed, including central bank swap facilities. Altogether, central banks’ balance sheets expanded much more than in previous crises, and support was more flexible.

Funding strains prompted the provision of guarantees, including to shadow banks, but more selectively than in past crises, when they covered a wide set of liabilities and were mostly in the form of blanket guarantees. In the recent episodes, formal guarantees were largely applied to specific banks (such as Northern Rock in the United Kingdom) or new debt issuance only. With deposit insurance schemes already in place, countries also quickly increased the coverage limits, substantially in some cases.4 In addition, guarantees were extended to some nonbank financial institutions, notably in the United States, where a run on money market funds, then a $3 trillion industry, led to the provision of guarantees traditionally used to protect bank deposits. Finally, governments made statements expressing their support for the whole sector.

Accommodative Monetary and Fiscal Policies

Expansionary monetary policies during the recent crises were critical in supporting banks and markets. Monetary policy was relaxed significantly early on by quickly adjusting short-term interest rates to historical lows (Figure 15.3), with major central banks taking coordinated actions. Taking advantage of their reserve currency status, several central banks committed themselves, at least conditionally, to maintaining low interest rates for prolonged periods. Those moves were opposite of the efforts of central banks in many past crises in which nominal rates were kept high or sometimes even raised to support currencies. In the recent crises, the low policy rates and ample liquidity often allowed banks to preserve their intermediation margins in spite of higher costs of other funding. Accommodative monetary policy also helped support overall asset values, reduced the risk of an adverse debt-deflation spiral, and limited nonperforming loans, at least initially, thus protecting some of the banks’ profit streams and balance sheets despite losses on traded securities.

Figure 15.3Monetary and Fiscal Policies during Crises

Sources: Haver Analytics; IMF, International Financial Statistics, and IMF, World Economic Outlook.

Note: M and Y on x-axes refer to month and year of the onset of the crisis, respectively.

Accommodative fiscal policies were important to maintaining aggregate demand and asset values, thus indirectly supporting financial institutions. Better initial conditions allowed for larger fiscal deficits than in past crises; most policymakers opted to allow automatic stabilizers to operate, and many undertook countercyclical fiscal measures. By supporting aggregate demand, fiscal stimulus helped reduce expected defaults on bank loans and thus reduced banks’ recapitalization needs.5 This approach also differed from that in past crises, when fiscal policy was often contractionary (Figure 15.3). Furthermore, fiscal policy responses were more coordinated across countries than in the past, further helping to support economies.

Capital Support

The private sector (including sovereign wealth funds) contributed much to recapitalizing financial institutions in the 2007–09 crises, albeit to varying degrees across regions. From September 2007 to September 2008, private capital injections for U.S., European, and Asian institutions amounted to 71, 78, and 94 percent, respectively, of announced losses (IMF, 2008) and greatly reduced balance sheet pressures during that period. Also, in some smaller countries, recapitalization came from foreign banks that dominated markets. During the entire 2007–09 crisis period, private investors contributed about 61 percent of capital, but more so in the United States than in the euro area (about 86 percent versus 47 percent).

Changes to accounting and valuation practices also alleviated capital pressures. It was perceived that fair-value accounting might contribute to a fire sale of assets and exacerbate solvency and liquidity concerns. After much political pressure, accounting standards boards allowed banks in October 2008 to reclassify certain assets, including complex structured securities, as held-to-maturity, which meant they could be reported on a historical or amortized cost basis. Also, firms could rely more on their own assumptions and models in valuing assets, including mortgage-backed securities, during illiquid or inactive market conditions. This greater flexibility in valuing assets also limited the need to raise new capital (Huizinga and Laeven, 2009).

Government recapitalizations were, proportionately, much lower than in past crises. Together with accommodating policies that supported asset values and held down losses, the rapid and large private recapitalization meant that government recapitalizations took place at a point when banking solvency was much stronger than in past crises. Government capital support totaled US$441 billion—$245 billion in the United States under the Troubled Asset Relief Program (TARP) and US$196 billion in the European Union—which, at only 5 percent of GDP on average, is about one-third of the amount provided in past crises. However, by aiming to have a rapid effect, avoid stigmatization, and support lending, the government recapitalizations were spread too broadly, forgoing the benefits of separating viable from nonviable institutions.

Asset Restructuring

In a typical crisis, nonperforming loans rise steeply, even before its onset, as banks acknowledge the expected deterioration in corporate and household repayment capacity. Although the recent crises broadly exhibited the characteristics of a typical collapse following a boom (Lindgren, Garcia, and Saal, 1996; Dooley and Frankel, 2003; and Reinhart and Rogoff, 2009), the rise in nonperforming loans was much less pronounced (Figure 15.4). Write-downs of impaired assets showed up only gradually and as of mid-2011 were lower than in past crises. This was in part due to the types of assets involved, with the drop in the value of securitized loans occurring earlier than in other crises, before the end of the cycle. Actual defaults followed only when the crisis affected the real economy and corporate sector and household conditions had worsened. Furthermore, corporate sectors were generally not overleveraged.

Figure 15.4Nonperforming Loans in Past and Recent Crises

Sources: BankScope; and IMF, Global Financial Stability Report, 2010.

Note: t = 0 denotes the year of the onset of the crisis.

Partly for these reasons, asset restructuring was far more limited in the recent crises than in the past. Restructuring refers to two processes. The first is diagnosing the value of a bank’s loans and investments, stating the value of its securities at market prices rather than acquisition cost, provisioning for and writing off part of nonpaying loans, and possibly selling off securities and loans, and deleveraging the institution. The second is to ensure that borrowers’ financial conditions are sound and their creditworthiness is restored, which typically involves both financial restructuring (extending the maturity of loans, reducing interest rates and amount owed) and operational restructuring (selling of assets, reducing labor and administrative costs, and the like).

In the recent episodes, many countries applied asset restructuring on a case-by-case basis, with government relief provided mainly through guarantees against a large deterioration in asset values; less frequent in the recent crises was the use of “bad banks” (Table 15.2). Asset guarantees require little up-front funding and do not involve immediate loss recognition or recapitalization, unlike purchasing impaired assets at a discount. Given the size and complexity of nonperforming assets—including the many securitized portfolios and mortgages to be restructured—guarantees were often the only, or at least the preferred, option. Also, high government debt levels in some countries may have prevented asset transfers. Although guarantees reduce uncertainty for financial institutions and help with their funding, the government takes on higher contingent costs.

Table 15.2Selected Asset Relief Measures during Recent and Past Crises
Type of measureUseCountriesBeneficiariesAsset typesBillion US$% of GDP
Asset guaranteesRecent

BelgiumDexia (FSAM)


Structured assets10.52.2
GermanyWest LBStructured assets7.020.2
Bayern LB6.70.2
NetherlandsINGRMBS, mortgage loans35.144.4
UnitedRBSPools of assets524.024.0
United StatesCitigroupReal estate-related301.02.1
Bank of America118.050.8




West LB
Structured assets

Toxic and nonstrategic

LatviaParexImpaired and nonstrategic assets1.25.1
UnitedNorthern RockMortgage loans and121.065.6
KingdomBradford & Bingleyother loans79.363.6
United StatesBear StearnsRMBS, CDOs30.00.2
Past crisesSwedenNordbanken

Real estate–related and

corporate loans
AMCs / asset purchaseRecent chasesIrelandBanksDistressed real estaterelated (purchased by NAMA)97.844.0
United StatesGovernment-sponsored entitiesNew MBS (purchased by asset managers for the Treasury)197.61.4
Banks“Legacy” MBS/loans (PPIFs)14.20.1
Past crisesThailandBanksBad loans (purchased by TAMC)17.013.7
Korea, Rep. ofFinancial institutionsBad loans (purchased by KAMCO)90.019.5
Sources: Borio, Vale, and von Peter (2010); Boudghene, Maes, and Schmeicher (2010); European Commission State Aid Register; Fung and others (2004); country authorities; and IMF staff.Note: AMC = asset management company; CDO = collateralized debt obligation; KAMCO = Korea Asset Management Corporation; MBS = mortgage-backed security; PPIF = Public-Private Investment Funds; RMBS = residential mortgage-backed security; TAMC = Thai Asset Management Company.

For asset guarantees, amount refers to guaranteed (book) value of portfolio or otherwise as footnoted. For asset purchases, amounts generally refer to the book value of assets transferred, that is, before any writedowns, except in cases of marketable securities bought, in which case they refer to the actual market values.

Swap facility.

Includes asset-backed securities portfolio and loan to Sealink portfolio.

Backup facility.

The guarantee was provided in January 2009; however, the arrangement was never implemented and Bank of America paid an exit fee in September 2009.

Refers to the total asset size of the institutions as of January 1, 2010.

Represents lending to AIG special purpose vehicles by the Federal Reserve Bank of New York (FRBNY), which is less than the total assets of the “bad bank” for AIG. The loans from FRBNY were repaid from the liquidation of AIG assets.

Sources: Borio, Vale, and von Peter (2010); Boudghene, Maes, and Schmeicher (2010); European Commission State Aid Register; Fung and others (2004); country authorities; and IMF staff.Note: AMC = asset management company; CDO = collateralized debt obligation; KAMCO = Korea Asset Management Corporation; MBS = mortgage-backed security; PPIF = Public-Private Investment Funds; RMBS = residential mortgage-backed security; TAMC = Thai Asset Management Company.

For asset guarantees, amount refers to guaranteed (book) value of portfolio or otherwise as footnoted. For asset purchases, amounts generally refer to the book value of assets transferred, that is, before any writedowns, except in cases of marketable securities bought, in which case they refer to the actual market values.

Swap facility.

Includes asset-backed securities portfolio and loan to Sealink portfolio.

Backup facility.

The guarantee was provided in January 2009; however, the arrangement was never implemented and Bank of America paid an exit fee in September 2009.

Refers to the total asset size of the institutions as of January 1, 2010.

Represents lending to AIG special purpose vehicles by the Federal Reserve Bank of New York (FRBNY), which is less than the total assets of the “bad bank” for AIG. The loans from FRBNY were repaid from the liquidation of AIG assets.

During the recent crises in comparison with most past episodes, such as the Nordic and Asian crises, fewer assets were removed from the balance sheets of financial institutions through government asset management companies (AMCs) or other programs (Table 15.2). One important exception was Ireland, where distressed loans with a book value equivalent to 44 percent of GDP were transferred to a government AMC. Other asset-targeted programs, like the U.S. Public Private Investment Program, amounted to only 0.1 percent of GDP. Although the Federal Reserve and other central banks also purchased large amounts of private securities to support targeted markets, those efforts were not directly aimed at asset restructuring.

Policy Choices and Preliminary Lessons

This section reviews the outcomes of policies chosen and compares these policies with both good practice and lessons from past crises. It first assesses the costs—using the metrics of output losses, fiscal costs, and increases in government debt—in comparison with past crises. Considering various trade-offs and differences in country circumstances, it then compares recent policy responses with those adopted in the past and with good practice guidelines for resolving crises (see Appendix 15C). It provides some preliminary lessons for effectiveness.

The Costs of Crises

The costs of crises can be assessed in different ways, including their direct fiscal costs, encompassing direct outlays to support the financial system and for resolving nonperforming assets; their broader fiscal costs, measured as the increase in government debt over some chosen horizon (plus the direct fiscal costs); and the real output losses.

The direct fiscal support has been lower than in past crises. The fiscal costs attributable to direct support in the recent crises reached 5 percent of GDP on average as of end-2009, against 15 percent for past crises (Figure 15.5). With the important caveat that the crisis was still unfolding in many countries, especially in EU countries, this lower cost reflected the more accommodative monetary and fiscal policies and the lesser need for and use of government recapitalization. The management of distressed assets was more decentralized than in past crises, a reflection of the more limited nature of government interventions and the greater use of guarantees. Guarantees (including liability guarantees, liquidity, and other contingent support) reduce the need for up-front fiscal outlays, but impose higher contingent fiscal costs. Broader fiscal costs, however, were larger than in past crises. Projected increases in debt for the four-year period after the onset of the crisis are higher for the recent crises (about 25 percent of GDP). These increases come on top of the already-large government debt burdens in many advanced economies.

Figure 15.5Median Costs of Recent and Past Crises

Source: Laeven and Valencia, Chapter 2, in this volume.

Note: Costs are measured as of the end of 2009.

Output losses were lower than in the past but still significant. At 25 percent of GDP, cumulative over four years, the median output losses were lower than the 35 percent for past crises, reflecting in great part the beneficial effects of the extraordinary policy measures implemented during the recent crises (and also reflecting the fact that in the recent crises, the affected countries had lower growth trends preceding the crisis).6 Measured for the whole world, however, output losses from the recent crises were 3.3 percent versus 0.2 percent for the past. That global output losses were larger this time is no surprise given the size of the affected economies and the ensuing spillovers.

Costs relative to output in the recent crises varied greatly across countries but were more comparable relative to banking system assets. For major advanced economies, the direct costs were relatively small, 3 percent to 5 percent of GDP. For some of the smaller countries, however, the direct costs were much larger, up to 17 percent of GDP for Iceland (excluding Icesave) and 25 percent for Ireland (as of end-2010). These countries engaged in larger-scale government recapitalizations and removals of bad assets from banks’ balance sheets. Furthermore, their banking systems were relatively larger (as measured by assets as a percentage of GDP). Indeed, as a share of banking assets, the direct costs were more comparable across countries. At the same time, whereas government debt in these smaller countries rose because of recapitalization programs, major countries had larger overall debt increases.

The Policy Choices and Preliminary Lessons

The following discussion offers preliminary lessons and suggestions for further reforms to the policy sequence and mix in crisis management and resolution; the diagnosis of financial institutions; the operational restructuring of weak institutions; the restructuring of assets, in particular of household debt; and measures to restore proper incentives and market discipline.

An Overall Accommodative Policy Mix Should not Preclude Deep Restructuring

Responses in the recent crises primarily relied on accommodative monetary and fiscal policy to contain the potential spillovers to the real sector. Affected countries included mostly advanced economies that had the ability to conduct countercyclical monetary and fiscal policy without undue concern, at least initially, about the impact on interest rates, exchange rates, or government debt. The accommodative policies contained the crisis by forestalling sharp increases in interest rates and large currency depreciations (currencies even appreciated in some countries)—which can degrade borrowers’ solvency and increase bank losses—and thus by directly and indirectly propping up bank asset quality and values.

This mix of policies may have transferred the costs to the future, however, in the form of higher government debt and possibly slower economic recoveries. Although the complexity of the crises may have justified more emphasis on the restoration of confidence and less extensive restructuring early on, it precluded thorough due diligence of individual banks and might have reduced incentives to restructure assets. Instead of a policy of targeted, diagnosis-based resolution and early asset restructuring, the stance was a muddling-through approach of accounting and regulatory forbearance, guarantees, and implicit government support that delayed addressing nonviable banks and nonperforming assets. The presumption should therefore remain in favor of deep restructuring early on, even when pursuing accommodative general policies.

Diagnosis of Financial Institutions Should Precede Recapitalization

Thorough, independent examinations were typically conducted in past crises to assess asset values and the viability of financial institutions in an effort to judge the appropriateness of recapitalization. Nonviable institutions would be closed or viable parts sold off and the rest of the institution liquidated. Once the size of the losses was determined, recapitalization plans for viable institutions would be announced and implemented. This process could take considerable time: in Indonesia, the government announced a blanket guarantee in January 1998, began examinations in August 1998, and implemented recapitalizations in March 1999.

In the recent crises, policymakers in the major advanced economies focused on reducing systemic consequences and therefore often opted for providing quick support to all institutions, including weak and potentially nonviable ones. Governments faced unprecedented complexity, and were hampered by limited information and limited tools for addressing systemic and cross-border entities. Therefore, they rapidly engaged in measures, first ad hoc and then more systematic, to stem contagion and restore market stability. Ad hoc assistance was provided to institutions embodying important counterparty risks (notably, in the United States, to AIG, to specialized municipal bond insurers, and to the two giant government-sponsored housing-related enterprises, Fannie Mae and Freddie Mac; and in Europe to a number of banks).7

The more rapid interventions typically did not allow for a separation of viable institutions from less-viable ones. Systematic assessments of institutions were conducted through stress tests and publicly disseminated in the United States and the EU (in May 2009 and July 2010, respectively), but only after initial government recapitalizations.8 These stress tests restored short-term investor confidence, but their long-term impact was uneven, in part because market participants had mixed views on the credibility of the assumptions used and the remedial actions announced subsequent to the tests. EU authorities were compelled to engage in a new round of stress tests. Furthermore, government support of institutions required little in the way of their operational restructuring, in contrast to earlier crises. In the recent crises in Iceland, Ireland, Latvia, and Ukraine, however, the sequence was more typical: diagnosis, recapitalization, and the removal of nonperforming assets or the creation of bad and good banks.

The lesson is that diagnosis needs to be conducted, including through strict and transparent stress tests, even while accommodative policies are being put in place. Given the circumstances, many governments had no alternative but to apply quick support measures. However, those quick measures should have been immediately followed by forward-looking measures, including asset and operational restructuring. Delaying the restructuring hampered the economic recovery because institutions were weighed down by troubled assets. As a result of residual uncertainty, confidence in many systems was still very dependent on implicit and explicit government and central bank support. Stress tests should have been conducted in many countries, accompanied as needed by credible recapitalization plans, or restructuring of institutions’ liabilities, without adding to sovereign debt burdens.

Dealing with Distressed Institutions Requires Operational Restructuring

Although most countries imposed limits on compensation to management and shareholders, more-intrusive measures—including cost cutting, downsizing, changes in management, and forced write-downs of shareholder value—were used less than in the past, except where governments took majority ownership of or fully nationalized institutions. Rather, obligations were placed on banks to continue to provide support to the real sector. The less-intrusive measures reflected institutions’ stronger solvency and continued majority private ownership, conditions different from those in past crises. But it also reflected the fact that the rapid and broad-based government support was mostly oriented toward financial stability and to limiting adverse short-term impacts.

In EU countries, additional conditions were imposed on state support measures, whereas in the United States both initial and ongoing conditions attached to state support (under the TARP program) were more limited. For those EU institutions that participated, government guarantees on liabilities carried restrictions on balance sheet growth, dividends, and employee compensation. For institutions benefiting from recapitalization and asset relief, significant balance sheet and operational conditions (e.g., restrictions on acquisitions, refocus on core activities, divestments of businesses and assets) were included for competition policy reasons. In the United States, capital assistance under TARP required that institutions be adequately capitalized. As a result, the only restriction imposed on these institutions was on executives’ compensation and advance U.S. Treasury permission for any increase in dividend payments.

Bank solvency has improved in many of the countries affected by the 2007–09 crisis, but mainly on the basis of balance sheet restructuring, including recapitalization; the support of abnormally low interest rates, which improve profits from lending and investing; and enormous fiscal stimulus, which supports loan performance. Although interest rates have remained low for some time now in many advanced economies, those conditions cannot be expected to continue. With many financial systems still overextended and subject in the coming years to new regulatory requirements,9 profitability will be under pressure. For institutions faced with limited prospects, the current incentive structures and competition for profits may again foster risky behavior. Uncertainty about possible further losses and shortages of capital may dissuade others from lending to the real economy and they may instead continue to deleverage. Although it is difficult to distinguish between demand and supply effects in the provision of credit, evidence suggests that recapitalization aids the speed and sustainability of recovery (Laeven and Valencia, 2011).

With fewer government levers to do the job, operational restructuring in the major countries depended largely on market pressures. For banks with large government ownership interests, restructuring efforts depended directly on government actions. And over time, the government must sell off its stake. Market pressures will force many other institutions, including those that benefited from government support, to rebuild balance sheets and restructure operations. Regulatory reforms addressing gaps and shortcomings could have helped speed this process along. But the problems were large and complex, market conditions were depressed, and the economic recovery in the advanced economies was still weak. Those challenges underscore the importance of creating appropriate incentives for bank managers and owners and for supervisors and markets to monitor banks and ensure prudent governance.

Dealing with Distressed Borrowers Often Lagged

Although the across-the-board policies improved conditions at many financial institutions and supported economic activity, they may have reduced the incentives for restructuring the asset side of bank balance sheets. In many markets, asset quality remained uncertain. The prices of various financial assets improved following government support measures, but they remained low in many markets. Incomplete or dubious disclosures of asset quality, attributable in part to accounting changes, hindered market transparency and liquidity. The complexity of the task notwithstanding, asset restructuring was not as far along four years after the start of the crises than it was at similar stages in past crises. In the Asian and Nordic crises, government asset management companies (AMCs) and bad banks were accustomed to removing nonperforming loans—especially real estate loans—from the balance sheets of banks taken over by the authorities, thereby incurring upfront fiscal outlays. In the 2007–09 crisis, reflecting the limited government intervention in institutions, asset restructuring was largely left to the financial institutions themselves in most large advanced economies.

Although loss recognition was not always swift in past crises either, banks nonetheless underwent more rapid dispositions of their problem assets than they did in the 2007–09 crises. The limited use of AMCs in the 2007–09 crises reflects the complex nature of the assets involved, which do not permit easy centralized restructuring. But the chosen alternative route—injection of capital into the banks while leaving nonperforming loans on their balance sheets—poses the risk of continued losses, further weakening banks’ profitability and absorbing management capacity. It could also foster forbearance from formal regulations because it extends the government safety net and thereby impedes a full recovery of confidence. These alternative paths—disposal of bad assets versus capital injections with government funds and delay of clean up—are illustrated by the experiences of Sweden and Japan in their past crises (Box 15.1). Sweden took a comprehensive approach to dealing with distressed assets (primarily commercial real estate) and implemented it quickly. In contrast, by taking much longer, Japan showed that such delays can impose enormous costs.

Box 15.1Bank Restructuring and Asset Management: Sweden and Japan

In Sweden, the authorities responded to the initial signs of financial strain in the fall of 1990 with a series of ad hoc interventions. When these measures failed to restore stability, a new bank resolution agency was established to deal comprehensively with the crisis. The agency evaluated the financial condition of troubled banks on a forward-looking basis, categorized banks as solvent or insolvent, forced shareholders to recapitalize the former, and took control of the latter. Asset management companies (AMCs) were set up for two nationalized banks, and problem assets, conservatively valued—particularly real estate loans—were transferred to the AMCs. The process helped put a floor under real estate prices and facilitated the return of investors. As early as 1993, confidence in the financial system began to recover.

The Swedish experience provides useful lessons for the 2007–09 crises but needs to be placed in context (Klingebiel, 2000). The ultimate fiscal costs were relatively small, totaling only 4 percent of GDP, partly because of a global economic recovery and the competitive benefits of a 30 percent depreciation in the currency. Problem assets were mainly relatively simple commercial and residential real estate loans, and the banks operated domestically rather than cross border. These features, which enabled quick valuation and a centralized approach to restructuring and asset management funded by the state, were not characteristics of the 2007–09 crises.

In Japan. a financial crisis was triggered in 1995 when several regional deposit-takers failed. The authorities responded with emergency liquidity and a government guarantee of bank liabilities. Recapitalization schemes in 1998 and 1999 failed to restore confidence, and although a zero interest rate policy (adopted in February 1999) and quantitative easing (introduced in early 2001 and increased substantially in 2002) ultimately eased liquidity problems, those measures did not address the root causes of the crisis—uncertainty about the solvency of banks. In fall 2002, authorities finally set quantitative targets for the disposal of nonperforming loans and conducted rigorous examinations with more stringent provisioning standards. Along with two AMCs established to underpin asset prices, these measures ultimately helped restore stability.

Their special circumstances notwithstanding, the contrasting experiences of Sweden and Japan offer two key warnings about the management of financial crises: First, delays in recognizing problem loans may exacerbate a financial crisis and postpone recovery. Weak accounting practices and regulatory forbearance may blunt incentives for remedial action, sustain uncertainty about asset values and solvency, and hinder price formation. Second, liquidity provision may mitigate the immediate effects of a systemic crisis but mask fundamental problems in the banking system. Without comprehensive measures to recognize losses and address resulting capital shortfalls, the extended provision of exceptional liquidity may delay necessary restructuring.

Sources: Drees and Pazarbasioglu (1998); Ingves and Lind (2008); Ishi (2009); Syed, Kang, and Tukuoka (2009).

Unless interventions in the banks include restructuring, especially of household debt, the recovery is likely to lengthen. Before the crisis, households in many countries had accumulated large debts, especially for home purchases; all told, debt amounted to more than 130 percent of disposable income in the United States and more than 160 percent in some European countries. For many households these debts became too onerous to service given the unfavorable economic conditions. Although low interest rates eased the debt-servicing burden and reduced the pressure on lenders to adjust borrowers’ debt levels, risks will increase when interest rates return to normal levels. Efforts by banks to reduce the burden of household loans were not enough; likewise, government programs to restructure home mortgage loans were small in scope and largely ineffective as indicated by the recurrent defaults on restructured loans. Restructuring needs to be accelerated through a more effective mix of private and government actions (see Laeven and Laryea, 2009, and Chapter 17 of this volume for some best practices).

Reducing Systemic Risks and Preventing Moral Hazard Require More Reforms

Some of the structural characteristics that contributed to the buildup of systemic risks in financial sectors are still in place today, and moral hazard increased during the crisis. In most countries, the structure of the financial system changed little. In fact, because large banks acquired failing institutions, concentration increased on average—for the twelve 2007–09 crisis countries, the assets of the five largest banks combined rose from 307 percent of GDP before the crisis to 335 percent in 2009—complicating resolution efforts. The large-scale government support provided to institutions and markets—a contingent liability equivalent to one-quarter of GDP at the peak of the crisis—exacerbated perceptions of “too important to fail” (Goldstein and Veron, 2011). Failing firms may be resolved in a number of ways (see Appendix 15D), but in the 2007–09 crises, few creditors were forced to write down claims because of the risk of contagion. The shielding of creditors restored confidence more quickly, but at the cost of more moral hazard and the perpetuation of too-important-to-fail problem (and stretched sovereign balance sheets).

Countries need to implement measures that reduce moral hazard and that lower both the odds of a new systemic crisis and the effects it would have. Governments had to wean banks off their implicit government support, scale down deposit insurance schemes, and restore creditor discipline. For the longer term, they must also begin, through regulations and supervisory actions, to reduce incentives for complexity so as to facilitate restructuring in a crisis and diminish the expectations of bailouts and their adverse effects. Measures need to be well targeted and globally coordinated, yet flexible enough to reflect local factors. Examples include living wills and recovery and resolution plans; restrictions on the types of activity undertaken (such as the Volcker rule in the United States, which limits proprietary trading); a capital charge or levy on institutions commensurate with the systemic risk they create; eliminating government support provided to systemically important banks, including through carefully designed and monitored contingent capital or bail-in instruments, with clear triggers, so that losses are shared fairly; and reining in the proliferation of complex financial instruments. Policymakers will need to be cognizant, however, of possible unintended consequences given that many of these measures remain untested.

The enhanced frameworks and tools adopted by several countries to resolve complex bank, and sometimes nonbank, institutions are only a start; more progress is needed, especially on cross-border resolution. Since the 2007–09 crisis, several countries have adopted more effective resolution regimes for large financial institutions that allow losses to be borne by uninsured creditors, but more countries need to do so.10 Much is yet to be done toward enhancing supervision of cross-border exposures and related risks. And in all cases, the ability of the new regimes to deal with actual failures of large, cross-border institutions remains an unknown.


The financial upheavals of 2007–09 exposed serious weaknesses in crisis management and resolution. In many ways, the crisis is ongoing and further analysis is needed, but this chapter provides some preliminary lessons on the basis of experience in 12 countries in the recent crisis and 17 more countries in past crises going back to 1991. The major lessons and the policies requiring urgent attention include those in the following areas:

  • The overall policy mix and sequencing. Each of the major advanced economies dealt with the 2007–09 crisis differently, and except for the initial period, less decisively from the ways countries dealt with past crises. In the 2007–09 crisis, they quickly enacted accommodative monetary and fiscal policies and sustained them for extended periods. These measures helped reestablish confidence and stabilize economies. But unlike the responses in past crises, the governments made little effort to conduct in-depth diagnoses of banks’ balance sheets and follow-on restructuring (removal of bad loans and other assets devalued by the crisis). The resulting persistent weaknesses at banks likely retarded economic recovery. The in-depth restructuring of weak financial institutions and nonperforming assets remain on the agenda in many countries for dealing with the 2007–09 crises. In designing responses to future crises, striking the right balance between containment and restructuring policies is a major policy challenge.

  • Institutional tools for resolution. In the 2007–09 crisis, countries had little ability to wind down large cross-border banks and systemic nonbank financial institutions in an orderly fashion. The ongoing challenge is to design the framework—the institutional infrastructure—for such resolutions, including principles for burden sharing, to reduce moral hazard and enhance financial stability. Measures need to limit government bailouts proactively by providing greater capital and liquidity buffers and better cost-sharing arrangements with creditors in cases of distress. Establishing the framework is even more urgent today because concentration in the financial sector has increased.

  • The approaches to reducing systemic risks. The 2007–09 crisis showed that systemic risk had built to cataclysmic levels during the preceding boom. Because the lenses through which markets and supervisors looked for such risk kept it mostly hidden, national and international bodies will need to provide for greater public transparency on exposures and other aspects of systemic risk to facilitate supervisors’ work and enhance market discipline. Greater supervisory cooperation, including through supervisory colleges, will be needed. Developing methods for containing the buildup of systemic vulnerabilities will make a systemic crisis less likely and make it easier to deal with should it occur. Measures being considered include those that encourage institutions to reduce complexity or prohibit them from engaging in risky activities. The effectiveness of these measures and trade-offs in the efficient allocation of resources, however, require further analysis.

Appendix 15A. Definitions, Data Sample, and Fiscal Costs of Crises

Definition of Systemic Banking Crisis

A systemic crisis is defined as an episode of stress in the banking sector followed by significant policy intervention (Laeven and Valencia, 2010). Because stress is difficult to measure, a crisis is defined to be systemic when any three out of five commonly used crisis resolution policies are applied extensively: liquidity support, restructuring, asset purchases, significant guarantees, and nationalizations.

Sample of Recent and Past Crises

Episodes in 12 countries during the crisis of 2007–09 were identified as meeting the definition of a systemic crisis. Appendix Table 15A.1 shows the policy measures used in each country (all as of end-2009). Some crisis countries adopted additional measures after end-2009, some of which are noted in the text. Also, euro area periphery countries faced difficulties since, but interventions were limited as of 2009 and hence were not included in these comparisons.

Appendix Table 15A.1Recent Crises(all measures are as of end-2009)
CountryExtensive liquidity supportSignificant restructuring costsSignificant asset purchasesSignificant guarantees on liabilitiesSignificant nationalizationsIncome level1
United KingdomA
United StatesA
Source: Laeven and Valencia, Chapter 2, in this volume.Note: A = advanced economy; E = emerging market economy.
Source: Laeven and Valencia, Chapter 2, in this volume.Note: A = advanced economy; E = emerging market economy.

Appendix 15BTimeline of Events and Policy Responses

Source: IMF staff.

Note: AMC = asset management company; CPP = capital purchase program; DPB = Det Private Beredskab (Danish Bankers Association); GSE = government-sponsored entity (Fannie Mae and Freddie Mac); PDCF = Primary Dealer Credit Facility; PPIP = public-private investment program; SoFFiN = Special Financial Market Stabilization Funds; TALF = Term Asset-Backed Securities Loan Facility; TARP = Troubled Asset Relief Program; TSLF = Term Securities Lending Facility.

Appendix 15CThe ABCs of Crisis Resolution and Experiences in the 2007–09 Crisis
ABCs of Crisis ResolutionExperiences in the 2007–09 Crisis
Leadership and transparency. Appoint a single, accountable body with a clear mandate for financial stability and formal arrangements for cooperation with other agencies. Announce measures and procedures on a timely basis.Countries had a variety of different arrangements for coordination among authorities but improvements were needed. Inadequacies were exposed in all regimes with respect to identifying systemic risks, managing information flows, assigning decision-making responsibility, and communicating transparently and promptly. In most countries, clear responsibilities for financial stability were neither collectively nor individually assigned to agencies. Legislation was drafted or enacted in several countries to set up specific bodies to identify and respond to financial stability risks (such as the Financial Stability Oversight Council in the United States).
International coordination. For crises with cross-border incidence, policy responses should be coordinated as much as possible across the spectrum of measures from intervention on individual institutions through to sector-wide programs and macroeconomic policy responses.Crisis containment measures were initially uncoordinated and, although coordination improved at the macro policy level, cross-border resolution remained an issue. Deposit guarantees were raised to different levels on an uncoordinated basis across countries in the fall of 2008. However, guarantees were soon followed by joint interest rate cuts by several major central banks and a G-20 announcement on coordinated fiscal stimulus. Currency swap lines were provided by the U.S. Federal Reserve to the central banks of 14 countries. The resolution of cross-border firms proved to be very difficult, in most cases leading to uncoordinated decisions and suboptimal outcomes.
Diagnosis and analysis. Comprehensive and intrusive diagnostics of the depth and breadth of problems in the financial sector to identify insolvent banks should be undertaken. Financial institutions should continue to be monitored throughout the crisis, including, where necessary, strengthening regulatory reporting.Publishing the results of system-wide stress tests was a new feature of the 2007–09 crisis. But many firms in the United States and Europe received capital injections from the government sector before the stress tests, suggesting the tests may have been used more as a crisis containment tool to address information asymmetries and uncertainty than as a diagnostic exercise to inform decision makers.
Protection of depositors. In addition to the protection of insured depositors, targeted and credible guarantees of creditors may be necessary to prevent contagion and facilitate the closure of weak banks. Wide creditor guarantees incur risk for the state.Targeted guarantees were widely deployed. All of the 12 sample countries announced measures to enhance protection for retail depositors. Countries either fully guaranteed the majority of retail deposits or increased the coverage of their deposit insurance schemes. Eight out of the 12 countries additionally guaranteed other liabilities such as wholesale deposits. Blanket guarantees for all creditors were adopted only in Ireland whereas they were deployed in half of all previous crises in advanced economies.
Equitable and time-consistent burden sharing. Subject to preserving financial stability, the authorities should ensure consistent treatment of creditors independent of the size, complexity, and ownership of failed firms. This requires effective resolution regimes, intrusive supervision, and effective contingency planning.As with past crises, the treatment of creditors was inconsistent with previously stated policy, with ongoing moral hazard consequences. Despite a small number of notable exceptions, many large financial institutions had to be nationalized with creditors (other than shareholders) made whole. Studies suggest that the too-big-to-fail subsidy increased as a result of these bailouts.
Conditionality. Government support should be conditioned on the implementation of measures to address operational failure and ensure proper incentives, including improving banks’ risk management; replacing management and owners; rigorous audit followed by prudent writedown of assets; measures to cut costs and eliminate excess capacity; and where necessary, closing or transferring part of an insolvent firm to another firm.Conditions attached to recapitalization programs were initially less extensive than those applied during past crises, except in cases in which state-aid rules applied. Limits were placed on compensation to management and shareholders but “traditional” restructuring measures such as cost-cutting measures, downsizing, and forced writedown of shareholders reportedly were less prevalent, at least initially. In EU countries, such conditions were subsequently imposed as part of state aid approval. What was new during the 2007–09 crisis were the conditions placed on banks to extend new lending to support the economy, including targets on net lending to business customers.
Impaired-asset management. Early action on impaired assets is essential to preventing creditor discipline from further eroding. A variety of institutional arrangements and techniques can be chosen to balance rapid resolution and recovery of the value of impaired assets. Removing nonperforming loans from banks’ balance sheets may be necessary to address banks’ stock problem.Traditional asset management companies were less frequently deployed in the 2007–09 crisis in comparison with past crises. However, the asset management phase of the 2007–09 crisis is arguably still incomplete and AMC’s may be less relevant for dealing with complex structured products. Instead asset guarantees were used extensively in the 2007–09 crisis. Asset guarantees were provided on both “good” and “bad” assets and were seemingly deployed primarily as crisis-containment tools to reassure creditors that banks were sufficiently capitalized rather than to restructure banks’ balance sheets.
Resolution regimes. Resolution regimes should be strengthened to ensure that failing financial institutions (including systemically important nonbank financial firms), can be resolved promptly and in ways that minimize risks to financial stability and government sector costs.Resolution tools for dealing with failing financial firms were inadequate. Most countries did not have special administration regimes to allow early intervention before insolvency. Resolution options were therefore limited to liquidation, bailout, or nationalization if private sector solutions failed. Special resolution regimes that did exist generally did not extend to nonbank entities. Consequently, the U.S. authorities were unable to prevent Lehman Brothers from going into liquidation after efforts to sell the firm failed, and were forced to nationalize AIG days later.
Corporate insolvency regimes. Orderly and effective insolvency regimes are needed to ensure predictable and equitable outcomes. Reforms may be needed in crises to establish fast-track procedures to deal with many failures.Corporate insolvency regimes largely proved adequate. The majority of the crisis countries were advanced economies with well-established and funded judicial systems. However, economic recovery is incomplete, with some countries falling back into recessions, and further corporate insolvencies may follow.
Source: IMF staff.
Source: IMF staff.
Appendix 15DResolution Approaches, Restructuring, and Moral Hazard
Type of resolutionBearer of costsImplications for moral hazardComments
Liquidation of the whole bank, and insured depositors paid off.Shareholders, uninsured creditors, and the Deposit Guarantee Fund (DGF).No moral hazard.This tool was rarely used in the 2007–09 crises (and in past crises) and only for small banks, because of concerns about contagion.
Good/bad bank split of the firm into liabilities that are protected to prevent contagion and to preserve financial stability plus good assets; and other liabilities and “bad assets.” The former are sold to another firm and the latter are liquidated.Shareholders, creditors, the DGF, and the authorities if significant wider liabilities are rescued.To the extent that all (or most) uninsured creditors are left behind in the liquidation, the effect on moral hazard will be zero (or low). Shareholders and creditors left behind will receive what they would have received in whole-company liquidation (typically zero for shareholders). But if significant noninsured creditors are rescued, for example, wholesale deposits, then moral hazard will rise.This transaction is called a purchase and assumption in the United States. The rescued liabilities and the good assets are sold to a third party, perhaps via the intermediate step of a bridge bank. This tool was mainly used for small banks in the 2007–09 crisis, with some exceptions.
Recapitalization by the government or nationalization of a failing bank.Shareholders and the authorities. It is very unlikely that the creditors incur losses, unless the recapitalization fails and the firm is subsequently put into insolvency.Creditors are bailed out, creating significant moral hazard. If shareholders are only diluted or receive compensation, moral hazard will be further exacerbated.This tool was used extensively in the 2007–09 crisis (as in past crises).
Open bank assistance, which allows a failed firm to survive under original ownership. Assistance can take the form of
  • subsidized funding

  • subsidized asset guarantees or asset purchases

  • liability guarantees.

The authorities and shareholders, depending upon the terms of the assistance. For example, if assets are purchased at less than book value, or asset guarantees include a first-loss piece for the firm, shareholders incur losses.Significant moral hazard. Creditors will only incur losses if not guaranteed and the firm is subsequently placed into liquidation. Worse still, shareholders remain owners of the firm and may not face significant losses, depending on the degree of subsidy in the government assistance.These measures were extensively deployed in the 2007–09 crisis. These measures were typically deployed in conjunction with recapitalization by the authorities.
Source: IMF staff.
Source: IMF staff.

The following systemic crises in emerging and advanced economies that took place between 1991 and the 2007–09 crises are used for illustrative comparisons: Nordic crises: Finland, Norway, and Sweden (all 1991); Latin American crises: Brazil (1994), Mexico (1994), and Jamaica (1996); Asian crises: Indonesia, Japan, the Republic of Korea, Malaysia, and Thailand (all 1997); and emerging markets crises: Colombia (1998), Ecuador (1998), the Russian Federation (1998), Turkey (2000), Argentina (2001), and Uruguay (2002). Because these crises are a subset of historical cases, summary statistics reported here can differ from those in cited references.

Fiscal Costs of Crises

For the crises that began in 2007, direct fiscal costs consist of fiscal outlays committed to the financial sector through end-2009. For each past crisis, direct fiscal costs are the total fiscal outlays during the episode. See Laeven and Valencia in this volume for country-specific figures.


    AghionP. D. Hemous and E. Kharroubi2009Credit Constraints, Cyclical Fiscal Policy and Industry GrowthNBER Working Paper No. 15119 (Cambridge, Massachusetts: National Bureau of Economic Research).

    BorioC. B. Vale and G. von Peter2010Resolving the Financial Crisis: Are We Heeding the Lessons from the Nordics?BIS Working Paper No. 311 (Basel: Bank for International Settlements).

    BoudgheneY. S. Maes and M. Schmeicher2010Asset Relief Measures in the EU—Overview and Issues” (unpublished; Brussels: European Commission Directorate-General for Competition).

    CalomirisC. D. Klingebiel and L. Laeven2003Financial Crisis Policies and Resolution Mechanisms: A Taxonomy from Cross-Country Experience” in Systemic Financial Distress: Containment and Resolution ed. by P. Honohan and L. Laeven (Cambridge, United Kingdom: Cambridge University Press).

    ClaessensS. G. Dell’AricciaD. Igan and L. Laeven2010Cross-Country Experience and Policy Implications from the Global Financial CrisisEconomic Policy Vol. 62 pp. 26993.

    ClaessensS. D. Klingebiel and L. Laeven2003Financial Restructuring in Banking and Corporate Sector Crises: What Policies to Pursue?” in Managing Currency Crises in Emerging Markets ed. by M. Dooley and J. Frankel (Chicago: University of Chicago Press) pp. 14780.

    CroweC. and E. Meade2007Evolution of Central Bank Governance around the WorldJournal of Economic Perspectives Vol. 21 No. 4 pp. 6990.

    DjankovS. C. McLiesh and A. Shleifer2007Private Credit in 129 CountriesJournal of Financial Economics Vol. 84 pp. 299329.

    DooleyM. and J. Frankel eds. 2003Managing Currency Crises in Emerging Markets Proceedings of a National Bureau of Economic Research conference (Chicago: University of Chicago Press).

    DreesB. and C. Pazarbasioglu1998The Nordic Banking Crises: Pitfalls in Financial Liberalization?Occasional Paper No. 161 (Washington: International Monetary Fund).

    DziobekC. and C. Pazarbasioglu1998Lessons from Systemic Bank Restructuring” Economic Issues Paper 14 (Washington: International Monetary Fund).

    European Union (EU)2008Declaration on a Concerted European Action Plan of the Euro Area CountriesSummit of the Euro Area CountriesOctober12.

    FungB. J. GeorgeS. Hohl and G. Ma2004Public Asset Management Companies in East Asia: A Comparative StudyOccasional Paper No. 3 (Basel: Bank for International Settlements).

    GoldsteinM. and N. Veron2011Too Big to Fail: The Transatlantic DebateWorking Paper 495 (Brussels: Bruegel).

    HoelscherD. and M. Quintyn2003Managing Systemic Banking CrisesIMF Occasional Paper No. 224 (Washington: International Monetary Fund).

    HonohanP. and L. Laeven eds. 2005Systemic Financial Crises: Containment and Resolution” (Cambridge, United Kingdom: Cambridge University Press).

    HuizingaH. and L. Laeven2009Accounting Discretion of Banks during a Financial CrisisEuropean Bank Center Discussion Paper No. 2009-17 (Tilburg, Netherlands: Tilburg University).

    IshiK.2009Have the Right Lessons Been Learned from the Banking Crisis of the 1990s?Attachment II of Sweden: 2009 Article IV Consultation IMF Country Report 09/247 (Washington: International Monetary Fund).

    IngvesS. and G. Lind2008Stockholm SolutionsFinance and Development Vol. 45 No. 4.

    IngvesS. and G. LindM. ShirakawaJ. Caruana and G.O. Martinez2009Lessons Learned from Previous Banking Crises: Sweden, Japan, Spain and MexicoGroup of Thirty Occasional Paper 79 (Washington: Group of Thirty).

    International Monetary Fund (IMF)2008Global Financial Stability ReportOctober (Washington: International Monetary Fund).

    International Monetary Fund (IMF)2010Global Financial Stability ReportOctober (Washington: International Monetary Fund).

    KlingebielD.2000The Use of Asset Management Companies in the Resolution of Banking CrisesWorld Bank Research Working Paper 2294 (Washington: World Bank).

    LaevenL. and T. Laryea2009Principles of Household Debt RestructuringStaff Position Note 9/15 (Washington: International Monetary Fund).

    LaevenL. and F. Valencia2008Systemic Banking Crises: A New DatabaseIMF Working Paper 08/224 (Washington: International Monetary Fund).

    LaevenL. and F. Valencia2010Resolution of Banking Crises: The Good, the Bad, and the UglyIMF Working Paper 10/146 (Washington: International Monetary Fund).

    LaevenL. and F. Valencia2011The Real Effects of Financial Sector Interventions during CrisesWorking Paper 11/45 (Washington: International Monetary Fund).

    LindgrenC.-J. G. Garcia and M.I. Saal1996Bank Soundness and Macroeconomic Policy” (Washington: International Monetary Fund).

    PanettaF. T. FaehG. GrandeC. HoM. KingA. LevyF.M. SignorettiM. Taboga and A. Zaghini2009An Assessment of Financial Sector Rescue ProgrammesBIS Paper No. 48 (Basel: Bank for International Settlements).

    ReinhartC. and K. Rogoff2009This Time Is Different: Eight Centuries of Financial Folly (Princeton, New Jersey: Princeton University Press).

    StoneM. K. Fujita and K. Ishi2011Should Unconventional Balance Sheet Policies Be Added to the Central Bank Toolkit?Working Paper 11/145 (Washington: International Monetary Fund).

    SyedM. K. Kang and K. Tukuoka2009Lost Decade in Translation: What Japan’s Crisis Could Portend about Recovery from the Great RecessionIMF Working Paper No. 09/282 (Washington: International Monetary Fund).

Dziobek and Pazarbasioglu (1998); Calomiris, Klingebiel, and Laeven (2003); Hoelscher and Quintyn (2003); Honohan and Laeven (2005); Laeven and Valencia (2008); and Reinhart and Rogoff (2009) review the patterns in macroeconomic and financial variables surrounding crises and the policies used in containment and resolution.

Past: Nordic crises: Finland, Norway, Sweden (all 1991); Latin American crises: Brazil (1994), Mexico (1994), and Jamaica (1996); Asian crises: Indonesia, Japan, the Republic of Korea, Malaysia, and Thailand (all 1997); and the emerging markets crises: Colombia (1998), Ecuador (1998), the Russian Federation (1998), Turkey (2000), Argentina (2001), and Uruguay (2002). Recent (as of end-2009): Austria, Belgium, Denmark, Germany, Iceland, Ireland, Latvia, Luxembourg, the Netherlands, Ukraine, the United Kingdom, and the United States (see Appendix 15A).

Other papers reviewing policy responses in past and recent crises include Claessens, Klingebiel, and Laeven (2003); Ingves and Lind (2008); Ingves and others (2009); and Panetta and others (2009).

Belgium, Luxembourg, and the Netherlands increased limits from €20,000 to €100,000 as part of an EU-wide decision; the United States more than doubled them; and some countries guaranteed all retail deposits.

Fiscal policy has a greater effect on firms that are relatively dependent on external finance (Aghion, Hemous, and Kharroubi, 2009; and Laeven and Valencia, 2011).

Output losses are defined as the deviation of real GDP from its trend (computed as the Hodrick-Prescott-filtered series for the 20-year period ending a year before the crisis) for the four-year period beginning with the first crisis year.

In the European Union, key policies were subsequently coordinated (e.g., EU, 2008), followed by agreements to coordinate recapitalization, guarantees, asset insurance, and transfer schemes.

A first round of EU-wide stress tests was conducted in September 2009, but results were not made public.

The Basel III agreement on international banking standards reached by the Basel Committee on Banking Supervision calls for a substantial increase in the quantity and quality of capital and liquidity buffers, new regulations and tougher standards for nonbank firms, and other macroprudential rules.

In the United Kingdom, the Banking Act 2009 allows for early intervention using a wide array of tools to deal with failing banks. In Germany, a temporary regime was replaced by a permanent resolution framework for systemically important banks. The United States (through the 2010 Dodd-Frank Act) and several European countries (including Belgium) extended resolution regimes to cover systemically important nonbank financial institutions. Although progress is being made in the EU (e.g., in January 2011 the European Commission made proposals for crisis preparedness and cross-border resolution in the European Union, and in September 2012 it was agreed to pursue a banking union), important issues still remain to be agreed upon, including burden sharing, cross-border deposit insurance claims, secrecy laws, and other legal impediments.

    Other Resources Citing This Publication