Financial Crises

Chapter 7. The Global Financial Crisis: How Similar? How Different? How Costly?

Stijn Claessens, Ayhan Kose, Luc Laeven, and Fabian Valencia
Published Date:
February 2014
  • ShareShare
Show Summary Details
Stijn Claessens, M. Ayhan Kose and Marco E. Terrones This chapter was originally published in the Journal of Asian Economics, Vol. 21, No. 3, pp. 247–64. The authors would like to thank David Fritz and Ezgi Ozturk for providing outstanding research assistance.

The global economy is recovering from the deepest recession in the post–World War II era. The recession was triggered, albeit slowly, by the severe 2007–09 financial crisis in key advanced economies that coincided with the freezing of global financial markets and a collapse in global trade flows. The crisis quickly resulted in deep recessions in a number of advanced economies; the emerging market and developing economies were also seriously affected, but the impact varied across regions and countries.

Although the economic recovery is under way, the nature and implications of the crisis are still at the center of academic and policy discussions.1 For example, there has been an intensive discussion about the similarities and differences between the latest crisis and past episodes. Some commentators, especially in the media, argue that the latest crisis was different. Its root causes are thought to lie in excessive global savings (a “saving glut”), flowing through a poorly regulated shadow banking system in the United States to its housing market (Krugman, 2009). Others claim that the idea of this crisis being different is misleading because an analysis of earlier crises presents remarkable similarities to the latest episode. In particular, the excessive accumulation of debt that took place in various forms ahead of the 2007–09 crisis was also a feature of previous crises (Reinhart and Rogoff, 2008).

Another dimension of the ongoing discussions about the crisis focuses on its global spread and cost. The crisis originated in the United States, but it took place in a highly integrated global economy in which the widespread use of sophisticated financial instruments, along with massive international financial flows, facilitated its rapid spread across markets and borders. Although it was not surprising that a global crisis led to a significant decline in global activity, explaining the extent and duration of this decline has been a major area of research. Recessions associated with the global financial crisis, while displaying patterns similar to those of previous recession episodes, reflect an unlikely confluence of factors. Specifically, these recessions are associated with serious financial disruptions, including credit crunches, house price busts, equity price busts, and outright banking crises in some countries.

This chapter provides a brief analysis of the nature and cost of the crisis to shed light on these issues. In particular, it addresses three major questions. First, how similar is the latest crisis to previous episodes? Second, how different is the crisis from earlier ones? Third, how costly are recessions coinciding with serious disruptions in financial markets? The first two questions are studied in the first two sections. The basic conclusion is that the massive financial crisis that gripped the global economy beginning in 2008 was the result of a multitude of factors. Some of these factors are similar to those observed during the buildup to past financial crises, but some others are distinctly new. Although ranking the relative contributions of these various factors is difficult, together they help explain the latest episode’s considerable scale and scope. Irrespective of any similarities or differences though, the crisis has been a very costly one for both the real and the financial sectors, as discussed in the third and fourth sections.

How similar was the 2007–09 crisis to previous episodes? The first section examines four primary similarities in the buildup to the crisis and other similarities in the postcrisis busts. How different was the latest crisis from previous crisis episodes? As presented in the second section, at least four new dimensions played important roles in the severity and global scale of the crisis that included surprising disruptions and breakdowns of several markets in the fall of 2008. How costly was the 2007–09 episode? The global financial crisis resulted in recessions in almost all advanced economies. Most of these recessions were accompanied by credit crunches, house price busts, and outright financial crises. The third section provides an analysis of how recessions associated with credit crunches, house price busts, and financial crises differ from other recessions. The results suggest that recessions with credit crunches or house price busts result in more costly macroeconomic outcomes than do those without such disruptions. When recessions are accompanied by financial crises, the costs are larger and much more pronounced for consumption and investment.

The fourth section presents the dynamics of the lingering recession in the United States, the epicenter of the current crisis, and compares them with those of past recessions in advanced economies. It also provides a short discussion of the speed and extent of deterioration of activity in the United States. The findings suggest that the U.S. recession in the wake of the global crisis was clearly an outlier in many respects. The financial crisis has taken a heavy toll on the real economy as evidenced by deep and long recessions in a number of advanced economies in addition to the United States. The cost of a recession is, of course, affected by a number of factors. The final section presents a brief discussion of these factors, discusses policy implications, and concludes.

The Crisis: How Similar?2

The buildup to the ongoing financial crisis has four features similar to earlier episodes: First, asset prices rapidly increased in a number of countries before the crisis. Second, a number of key economies experienced episodes of credit booms ahead of the crisis. Third, a variety of marginal loans, particularly in the mortgage markets of several advanced economies expanded dramatically, which led to a sharp increase in systemic risk. Fourth, the regulation and supervision of financial institutions failed to keep up with developments.

Asset Price Booms

The exuberant pattern of asset prices in the United States and other advanced economies before the 2007–09 crisis is reminiscent of those observed in earlier major financial crisis episodes in the post-World War II period. The overall size of the U.S. housing boom and its dynamics—including house prices rising in excess of 30 percent in the five years preceding the crisis and peaking six quarters before the beginning of the crisis—was remarkably similar to house price developments in the so-called Big Five banking crisis episodes (Finland, 1990–93; Japan, 1993; Norway, 1988; Spain, 1978–79; and Sweden, 1990–93).3

The house price boom in the United States ahead of the 2007–09 crisis was, however, unusual both in its strength and duration. Sharp increases in house prices were also a common feature in other countries hit hard by the crisis, including the United Kingdom, Iceland, and many Eastern European countries. This synchronicity of house price increases across countries before the crisis may be surprising considering that housing is the quintessential nontradable asset. However, other analysis shows that such highly synchronized episodes were not uncommon in the past (Claessens, Kose, and Terrones, 2009). During the latest period, however, house price booms were partly fueled by low (short-and long-term) interest rates resulting from abundant global liquidity and high demand for safe assets (Caballero, 2010).

Credit Booms

The prolonged credit expansion in the run-up to the crisis was also similar to other episodes. Mendoza and Terrones (2008) document the main features of episodes of unusually sharp expansions in real credit that often ended in crisis. Credit booms generally coincide with large cyclical fluctuations in economic activity—with real output, consumption, and investment rising above trend during the buildup phase of credit booms and falling below trend in the unwinding phase (see Figure 7.1). In the upswing, the current account tends to deteriorate, often accompanied by a surge in private capital inflows. Increases in house prices and the real exchange rate often accompany such credit booms. At least for advanced economies, however, credit booms are not always associated with surges in inflation. Credit booms in these countries are also more likely when preceded by a period of gains in total factor productivity or financial sector reforms.4

Figure 7.1Credit Booms and Macroeconomic Variables

(cross-country means and medians of cyclical components)

Source: Authors’ calculations.

Note: The solid (dashed) line represents the median (mean) of each variable around the time of a credit boom. Peak in the cyclical component of per capita real credit at year 0.

Most of these movements were also features of the credit booms that took place in Iceland, Spain, the United Kingdom, the United States, and some other advanced economies ahead of the 2007–09 crisis. However, unlike this crisis, credit booms in the advanced economies were only occasionally associated with currency and banking crises. In the past, advanced economies experiencing credit booms were more likely to have currency crises than banking crises. The credit boom that preceded the 2007–09 crisis was also not limited to advanced economies, but extended to varying degrees to several emerging market countries caught in the storm. In the run-up to the crisis, credit aggregates grew very fast in several Eastern European countries and often fueled real estate booms.

As in past episodes, international financial integration helped facilitate some of these trends (Cardarelli, Kose, and Elekdag, 2010). Specifically, large capital inflows were associated with accelerating GDP growth and, for many countries, with credit expansion. In addition, output growth was accompanied by large swings in aggregate demand and in the current account balance, with strong deterioration of the current account during the inflow period (see Figure 7.2).

Figure 7.2Selected Macroeconomic Variables in Periods Surrounding Large Capital Inflows

Source: Authors’ calculations.

Note: Median across all completed episodes. “Before” denotes averages of the macro variables in the two years before the episodes. “After” denotes averages of the macro variables in the two years after the episodes. The arrows indicate that the difference between medians is significant at a 10 percent level or better. For example, in panel a, the average real GDP growth in the two years after the episode is statistically significantly different from the average real GDP growth during the episodes. Real effective exchange rate appreciation is the cumulative change within periods.

Marginal Loans and Systemic Risk

The rapid growth of credit was often directed toward households and resulted in sharply increased household leverage. The boom in household credit was associated with the creation of marginal assets whose viability relied on continued favorable conditions. In the United States (and to some extent the United Kingdom), a large portion of the mortgage expansion consisted of loans extended to subprime borrowers—households with limited credit and short employment histories. Therefore, debt servicing and repayment were vulnerable to economic downturns and changes in credit and monetary conditions. This situation maximized default correlations across loans, generating portfolios highly exposed to declines in house prices, confirmed ex post through the large nonperforming loans when house prices declined.5

A similar pattern led to large portions of domestic credit denominated in foreign currency, particularly in emerging Europe. Large foreign currency exposures in the corporate and financial sectors had been a common feature in the Asian crisis of the late 1990s. In the 2007–09 crisis, in several Eastern European economies large portions of domestic credit (including to households) were denominated in foreign currency (euros, Swiss francs, and yen—Árvai, Driessen, and Ötker-Robe, 2009). Although lower interest rates relative to local currency increased affordability, borrowers’ ability to service loans and their creditworthiness depended on continued exchange rate stability. As with U.S. subprime loans, this meant high default risk correlations across loans and systemic exposure to macroeconomic shocks.

As a result of buoyant housing and corporate financing markets, derivative markets in many forms expanded greatly. In particular, favorable conditions spurred the emergence of large-scale derivative markets, such as mortgage-backed securities and collateralized debt obligations with payoffs that depended in complex ways on underlying asset prices. The pricing of these instruments was often based on a continuation of increasing house prices that facilitated the refinancing of underlying mortgages. The corporate credit default swap market also expanded dramatically based on favorable spreads and low volatility.

Regulation and Supervision

Episodes of large credit expansion have reflected not only macroeconomic conditions, but also various structural deficiencies, such as explicit or implicit government guarantees, herding behavior by investors, reduced lending standards, excessive competition, and information asymmetries. These episodes have also been associated with rapid financial liberalization and poorly supervised and unregulated financial innovation.

Evidence shows that past crises often followed credit expansions triggered by financial liberalization not accompanied by necessary regulatory and prudential reforms (Demirgüç-Kunt and Detragiache, 1998). Moreover, imbalances often resulted from poorly sequenced regulatory reforms. Underdeveloped domestic financial systems were often unable to intermediate large capital inflows in the wake of capital account liberalization. Poorly designed financial reforms and deficient supervision often led to currency and maturity mismatches and to large and concentrated credit risks.

In the run-up to the 2007–09 crisis, regulatory approaches to and supervisory oversight of financial innovation were insufficient, although perhaps in more subtle forms. As in previous crises, but this time in advanced economies, financial companies, merchant banks, investment banks, and off-balance-sheet vehicles of commercial banks operated—to varying degrees—outside banking regulations. However, while this shadow banking system provided increasingly important avenues for intermediation, it grew without adequate oversight and led to systemic risks. Unhealthy turf competition between various supervisory agencies in some countries and conflict of interest issues at rating agencies exacerbated problems. Regulators also underestimated the conflicts of interest and information problems associated with the originate-to-distribute model.6 Not only did this harm consumers of financial services, but it also created the potential for a chain reaction leading to systemic risk.

Dynamics of the Bust

As in earlier crises, the increase in asset prices, the rapid growth of credit combined with poor lending practices, the increase in systemic risk, and failures in regulation and supervision created many vulnerabilities. Although only a small number of credit booms end in banking crises—about one-quarter of all asset price booms end in busts (Helbling and Terrones, 2003)—the probability of a crisis increases with a boom (Dell’Ariccia, Barajas, and Levchenko, 2008). More generally, research documenting the main features of these types of credit booms highlights the strong association with subsequent busts (Mendoza and Terrones, 2008). Furthermore, the larger the size and duration of a boom, the greater the likelihood it will result in a crisis. The mechanisms linking credit booms to crises include increases in leverage of borrowers (and lenders) and a decline in lending standards. In the U.S. episode, both channels were at work (Dell’Ariccia, Igan, and Laeven, 2008).

When asset booms turn into busts, significant output losses often result. The outcome depends on the nature of the asset booms, with important differences between house price busts and equity busts (Claessens, Kose, and Terrones, 2008). First, the magnitude of the asset price fall during a bust depends on the size of the run-up in prices before the bust. But price corrections during house price busts are smaller than those during equity price busts. This reflects, in part, the lower volatility and liquidity of housing markets. Second, the association between booms and busts is stronger for housing than for equity prices. The implied probability of a house price boom being followed by a bust is about 40 percent. Third, house price busts last longer than equity price busts do. Moreover, the output loss associated with a typical house price bust is twice as large as that associated with an equity price bust.

Fourth, bank-based financial systems suffer larger output losses than market-based financial systems during house price busts, while market-based systems tend to suffer larger output losses than bank-based systems during equity price busts (Helbling and Terrones, 2003). This outcome is consistent with the high exposure of banks to real estate lending, and the larger importance of equities in households’ assets in market-based systems. Last, both equity and house price busts are often synchronized across countries, but the degree of synchronization in equity price busts is particularly high. This time, however, the downturn in house prices was highly synchronized across countries, with implications for global economic activity. Indeed, as of 2010, most advanced economies had experienced a recession for at least a year (Kose, Loungani, and Terrones, 2010).

The degree of international financial integration before the crisis also affects the bust. Cardarelli, Kose, and Elekdag (2010) examine developments subsequent to surges in private capital inflows for a group of emerging market and open advanced economies since 2000. After a period of capital inflows, growth can drop significantly. In fact, average GDP growth in the two years after episodes that end abruptly tends to be about 3 percentage points lower than during the episode, and about 1 percentage point lower than during the two years before the episode (see Figure 7.2). Past episodes characterized by sharper post-inflow declines in GDP growth tend to be those with faster acceleration in domestic demand, sharper increases in inflation, and larger real appreciation during the inflow period (see Figure 7.3).

Figure 7.3Large Capital Inflows: Post-Inflow GDP Growth and Selected Macroeconomic Variables

Source: Authors’ estimates.

Note: Values reported are medians for the two groups of episodes. Episodes with the weakest (strongest) post-inflow GDP growth are those with above (below) median differences between average GDP growth in the two years after the episode and the average during the episodes.

* indicates that the difference between medians is significant at a 10 percent level or better.

1 Average real GDP growth in the two years after the episodes less average during episodes.

2 Average during episodes minus average in the two years before the episode.

3 Average during episodes. CPI = consumer price index.

4 Cumulative change during episodes.

The surge in capital inflows also appears to be associated with a real effective exchange rate appreciation. Hence, the sharper post-inflow decline in GDP growth seems to be associated with persistent, expansionary capital inflows, which compound external imbalances and sow the seeds of the eventual sharp reversal. Then, a sharp reversal also often occurs in the current account. The end of the inflow episodes typically entailed a sharp reversal of non-foreign direct investment flows, while foreign direct investment proved much more resilient. This was the pattern in several Eastern European countries during the 2007–09 crisis.

As often before, poor crisis management played an important role in aggravating the financial crisis. For instance, similar to past episodes, it was difficult to get ahead of a quickly evolving situation to contain the financial turmoil and reduce its impact on the real economy (Cecchetti, 2009). The chronology of the crisis (Calomiris, 2009; and Gorton, 2009) shows how events and market developments triggered and conditioned specific subsequent developments and policy responses, that, in retrospect at least, probably made the crisis more severe (Taylor, 2009). The focus of authorities typically remained primarily on the liquidity and insolvency of individual institutions, rather than on the resilience of the whole financial system. Incomplete information and partial assessments of the serious financial problems led to ad hoc and piecemeal interventions, which at times created further disruptions and loss of confidence among creditors and investors. These shortcomings meant an underestimation of the probability and costs of systemic risk in many countries.

At the international level, insufficient coordination among regulators and supervisors and the absence of clear procedures for the resolution of global financial institutions has been a long-standing problem. In the 2007–09 crisis especially, these issues hindered efforts to prevent and contain the impact and transmission of the crisis (Claessens, 2009). As clearly demonstrated by the failures of Lehman Brothers and some Icelandic banks (among many other financial institutions), countries could not deal with large, complex, globally active financial institutions on their own because these institutions affect many markets and countries. Various government interventions, although necessary and often unavoidable, led to unintended effects on other countries, creating large distortions in international capital flows and financial intermediation. Overall, the lack of global agreements on tools for intervention made the crisis worse.

The Crisis: How Different?

New dimensions played important roles in the severity and global scale of the 2007–09 crisis—particularly with respect to its transmission and amplification—including surprising disruptions and breakdowns of several markets in the fall of 2008. The crisis was different from previous ones in at least four respects. First, there was widespread use of complex and opaque financial instruments. Second, the interconnectedness among financial markets, nationally and internationally, with the United States at the core, had increased in a short period. Third, the degree of leverage of financial institutions had accelerated sharply. Fourth, the household sector played a central role. These new elements combined to create unprecedented sell-offs in the fall of 2008 that resulted in the global financial crisis.

Increased Opaqueness

Securitization spurred by the use of innovative (but complex) financial instruments was a critical element of the credit expansion, particularly mortgage credit, in the United States. Securitization—a long-standing practice for prime loans conforming to the underwriting standards of government sponsored entities—changed in scope beginning in the mid-1990s, with more than 70 percent of nonconforming mortgages in the United States being securitized by 2007, up from less than 35 percent in 2000 (Ashcraft and Schuermann, 2007). Other assets were increasingly packaged as well, and cash-flow streams from securities were further separated and tranched into other securities such as collateralized debt obligations (Blanchard, 2009).

The increased recourse to securitization and the expansion of the originate-and-distribute model exacerbated agency problems (Furceri and Morurougane, 2009). The progressive expansion of more-opaque and complex securities and the increasing delinking of borrowers from lenders further worsened agency problems. Risk assignments became increasingly unclear and incentives for due diligence decreased, leading to insufficient monitoring of loan originators and an emphasis on boosting volumes to generate fees. The distribution model led to widespread reliance on ratings for the pricing of credit risks, with investors often unable or unwilling to fully assess underlying values and risks themselves.

As discussed in Mishkin (2009), the quality of balance sheets of households and firms is a key element of the financial accelerator mechanism, because some of the assets of each borrower may serve as collateral for its liabilities, which helps mitigate the problem of asymmetric information. In case of default, the lender can take title to the borrower’s collateral and recover some or all of the value of the loan. In a macroeconomic downturn, however, the value of many forms of collateral decreases. This, in turn, exacerbates the impact of frictions in credit markets and reinforces the propagation of adverse feedback loops.

Financial Integration and Interconnectedness

Financial integration has increased dramatically since 1990. Capital account openness and financial market reform have led to massive increases in cross-border gross positions, especially among member countries of the Organization for Economic Cooperation and Development (OECD). The presence of foreign intermediaries has also increased in several banking systems, including in many emerging markets (Goldberg, 2009). As a result, international risk sharing and competition and efficiency have increased, but so has the risk of rapid spread of financial shocks across borders. Several emerging markets have experienced sudden stops in this period.

Financial integration can result in indirect and catalytic growth benefits (Kose and others, 2009). Far more important than the direct growth effects of access to more capital is the potential for capital flows to generate collateral benefits (so called because they may not be countries’ primary motivations for undertaking financial integration). In particular, as reviewed by Kose and others (2009), growing evidence shows that financial openness can promote development of the domestic financial sector, impose discipline on macroeconomic policies, generate efficiency gains among domestic firms by exposing them to competition from foreign entrants, and unleash forces that result in better government and corporate governance. These collateral benefits could enhance efficiency and, by extension, total factor productivity growth (see Kose and others, 2010).

However, the 2007–09 financial crisis serves as a reminder of the risks of financial integration for both advanced and emerging market economies (Obstfeld, 2009). Specifically, increasing interconnectedness of financial institutions and markets, and more highly correlated financial risks, intensified cross-border spillovers early on through many channels—including liquidity pressures, a global sell-off in equities (particularly financial stocks), and depletion of bank capital. The sheer size of the U.S. financial market and its central role as an investment destination contributed to the spread of the crisis. Any shock to the U.S. financial markets is bound to have global effects. Before the crisis, U.S. financial assets represented about 31 percent of global financial assets, and the U.S. dollar share in reserve currency assets was about 62 percent. In the years before the crisis especially, U.S. financial assets were perceived to offer the combination of safety and liquidity attractive to private and public investors alike.

The crisis also triggered an unwinding of imbalances in other countries. In part because of closer international financial integration, benign financial and macroeconomic conditions—notably, low interest rates and narrower risk spreads—were in place on a global basis and asset price booms developed in many economies. However, for similar reasons, the busts came in a highly synchronized manner as well, in more intense and different ways compared with previous crises.

The Role of Leverage

The buildup of an unusually high degree of leverage of financial institutions and borrowers contributed to the propagation of shocks (Brunnermeier, 2009). Leverage increased sharply in the financial sector, directly at commercial banks in Europe, and through the shadow banking system and the rising share of investment banks and non-deposit-taking institutions in the United States. The leverage buildup among households especially differed from previous crises. In the run-up to Japan’s real estate crisis in the 1990s, for example, although the household debt-to-income ratio increased sharply, measures of household leverage (the household debt-to-assets ratio) declined, suggesting that Japanese homeowners built equity in their properties as real estate prices soared.

The high leverage preceding the 2007–09 crisis limited the system’s ability to absorb even small losses and contributed to the rapid decline in confidence and increase in counterparty risk early on. Loan-to-income values larger than in the past left households highly exposed to shocks, while at the same time high loan-to-value mortgages caused even moderate declines in house prices to push many households into negative equity. In the financial sector, high leverage meant that initial liquidity concerns quickly gave way to solvency worries. The buildup in leverage (including rising household indebtedness) was not restricted to advanced economies.

The Role of Households

Problems in the household sector played a more prominent role in the 2007–09 crisis than in previous crises. Most previous episodes of financial distress stemmed from problems in the official sector (e.g., Latin America’s debt crises of the 1980s) or the corporate sector (e.g., the Asian crises of the late 1990s). The 2007–09 crisis, however, largely originated from overextended households, in particular with respect to subprime mortgage loans. Although aggregate credit growth in the United States was less pronounced than in previous episodes, reflecting slower corporate credit expansion and the securitization of mortgages, the growth of household debt was excessive. Credit to households rose rapidly after 2000, driven largely by mortgages outstanding, with interest rates below historical averages and financial innovation contributing to an increase in outstanding household debt. Despite low interest rates, debt service relative to disposable income reached a historical high. The increased leverage left households vulnerable to a decline in house prices, a tightening in credit conditions, and a slowdown in economic activity. Similar patterns existed in several crisis countries.

Household balance sheet vulnerabilities also built up in other advanced economies and several emerging markets. Household debt-to-income ratios also rose sharply in several Western European countries (most notable in Ireland, Spain, and the United Kingdom). In several emerging markets, household credit expanded rapidly as well, leading to sharp increases in leverage and vulnerabilities. The decline in real estate prices adversely affected the quality of loan portfolios and put financial intermediaries at risk, especially in markets in which values had grown rapidly. This rapid growth of household debt had major implications for the transmission of the crisis from the financial to the real sector and complicated the resolution mechanisms and policy responses.

Old and New Elements Combined in Causing the Crisis

The various new elements combined with those factors observed in more “traditional” boom and bust cycles resulted in an unprecedented financial crisis. In the United States, a vicious cycle of rising foreclosures, falling home values, and disappearing securitization markets quickly developed. Vulnerable cohorts of borrowers became increasingly susceptible to rising interest rates and falling home values, and could no longer refinance their mortgages, leading to higher monthly payments and rising delinquency and default rates.

A wave of failures in financial companies—suddenly no longer able to securitize subprime mortgages—led to a virtual breakdown in mortgage origination and more abrupt adjustments in prices. Adverse feedback loops—of rising foreclosures placing additional downward pressures on house prices—started. With U.S. house prices declining on a national basis for the first time since the Great Depression, many heavily indebted borrowers confronted with substantial negative home equity faced incentives to “walk away” from their mortgages.

Tighter standards for new mortgages and consumer credit led to a sharp compression in consumer spending that compounded already difficult situations in the real sector. With households’ savings and net assets already at historical lows, financial constraints imposed by financial institutions under stress directly translated into reduced consumer spending, leading to initially localized but gradually spreading cycles of declines in corporate sector profitability, increases in layoffs and unemployment, and slowing economies—resulting in more foreclosures (Furceri and Morurougane, 2009).

Although initial recapitalizations were relatively large and rapid (including through the participation of sovereign wealth funds), they were limited to only a few banks and increasingly fell short of losses. As financial institutions incurred large losses and wrote down illiquid securities, solvency concerns across markets fueled a process of rapid deleveraging and forced asset sales. Mark-to-market rules forced further deleveraging and fire sales. Hedge funds—facing financing constraints and redemption pressures—further fueled this rapid unwinding process. This flurry of deleveraging led to further asset price declines, prompting distressed asset sales and rising recapitalization needs, resulting in further loss of confidence, causing a near meltdown in October 2008.

During fall 2008, increased balance sheet opaqueness and reliance on wholesale funding increased systemic fragility (Gorton and Metrick, 2009). Once U.S. house prices began to decline and defaults began to rise (affecting the expected value of the assets underlying mortgage-backed securities and collateralized debt obligations), the complexity of instruments undermined price discovery and led to market illiquidity and a freeze in securitization activity. The increased opaqueness of balance sheets (including that caused by the widespread recourse to off-balance-sheet instruments) made it difficult to distinguish healthy from unhealthy institutions. The resulting adverse selection problems contributed to the freezing of interbank markets and forced further sales of securities to raise funds. The increased centrality and systemic importance in many countries of highly leveraged, underregulated intermediaries relying on wholesale and short-term funding exacerbated problems.

Housing market vulnerabilities also came home to roost in several countries, notably in Europe. In the United Kingdom, mortgage lenders came under intense pressure, beginning in the fall of 2007 with a bank run on Northern Rock, which had been heavily reliant on interbank markets rather than deposits for funding. Large pressures also hit Hungary, Iceland, and the Baltic countries, where imbalances were more pronounced. The increased connections and simultaneous buildup of systemic risks across multiple countries made the management of shocks more complex, especially in light of institutional deficiencies in many countries—including the inability to quickly resolve large, cross-border financial institutions—and led to the rapid global spreading of turmoil.

Mortgage-backed securities and other U.S.-originated instruments were widely held by institutions in other advanced economies and the official sector in several emerging markets. Through these direct exposures and associated funding problems, spillovers quickly surfaced among European banks, including in Germany (IKB, July 2007) and France (BNP Paribas’ money market fund, August 2007). Because troubled intermediaries hit by losses and scrambling for liquidity were forced to sell other assets and cut lending, the crisis gradually spread to other markets and institutions through common lender effects.

Emerging markets—especially those that had relied heavily on external financing, and paradoxically those with more liquid markets—were affected through capital account and bank funding pressures. Amid global deleveraging, heightened investor risk aversion, and repatriation of funds, many emerging market economies suddenly found foreign funding sources increasingly scarce and were confronted with sudden stops or reversals of capital flows. In addition, emerging market corporations faced much higher borrowing costs, limited opportunity to issue equity, and few alternative sources of financing. Although official financing filled some of the gaps, a number of emerging markets had to make rapid adjustments, leading to real economic dislocations.

With the crisis still ongoing in many parts of the world, it is premature to opine on its implications for the broader debate on the costs and benefits of international financial integration. Nevertheless, there are two preliminary observations that are pertinent. First, the differential effects of the crisis across countries confirm that it is not just financial openness, but a country’s structural features and its precrisis policy choices that determined the overall impact of the crisis on a country. Second, outflows of capital triggered by the crisis did not lead to a resurgence of capital controls in emerging market economies.

Recessions and Financial Market Turmoil: How Costly?

The global financial crisis resulted in recessions in almost all advanced economies. As discussed in the previous sections, most of these recessions were accompanied by credit crunches or house price busts. This raises two specific questions about recessions associated with disruptions in credit and housing markets: How do recessions associated with credit crunches or house price busts differ from other recessions? And are recessions that coincide with financial crises more costly and longer than other recessions?

Building on earlier research (Claessens, Kose, and Terrones, 2009; 2010), this section analyzes the features of recession episodes that coincide with disruptions in credit or housing markets. To complement and expand on other studies focusing on the parallels between the 2007–09 financial crisis and past crises (Reinhart and Rogoff, 2008; 2009), it also examines the implications of recessions associated with financial crisis episodes.7

This section first briefly describes the database and methodology. Next, it discusses the characteristics of recessions associated with credit crunches or house price busts compared with other types of recessions. Finally, an analysis is made of recession episodes coinciding with financial crises, and the implications of such episodes are compared with those from recessions without a crisis.

Database and Methodology

A comprehensive database of key macroeconomic and financial variables for 21 OECD countries over the period 1960–2007 is used. The data are quarterly series mostly from the OECD Analytical Database and the IMF International Financial Statistics Database. The advantages of using OECD countries are the frequency and good quality of the data. Analyzing a large sample of emerging markets and developing countries would mean using annual data, a frequency at which detecting business cycles is much more challenging.8 The quarterly time series of macroeconomic variables are seasonally adjusted whenever necessary and are in constant prices. The financial variables considered are credit, house prices, and equity prices. All financial variables are converted into real terms by deflating them by the respective country’s consumer price index.

Before analyzing recessions and their interactions with financial crises, it is necessary to determine the dates of these events. The methodology focuses on changes in the levels of variables to identify cycles. This is consistent with the guiding principles of the National Bureau of Economic Research (NBER), which is the unofficial arbiter of U.S. business cycles. This methodology assumes that a recession begins just after the economy reaches a peak and ends as the economy reaches a trough. The methodology determines the peaks and troughs of any given series by first searching for maximums and minimums during a given period. It then selects pairs of adjacent, locally absolute maximums and minimums that meet censoring rules requiring a certain minimum duration of cycles and phases.

In particular, the analysis uses the algorithm introduced by Harding and Pagan (2002), which extends the so-called BB algorithm developed by Bry and Boschan (1971), to identify the cyclical turning points in the log-level of a series. A complete cycle goes from one peak to the next peak with its two phases, the contraction (recession) phase (from peak to trough) and the expansion phase (from trough to peak). The algorithm requires the minimum duration of the complete cycle to be at least five quarters and each phase to be at least two quarters. This methodology closely replicates the dates of U.S. business cycles as determined by the NBER.

With this methodology, cycles in output (GDP) are identified to provide a broad measure of economic activity for the 21 OECD countries. The exercise identifies 122 recessions, implying that a typical OECD country experienced about six recessions during 1960–2007. A recession on average lasts about four quarters (one year) with substantial variation across episodes—the shortest recession is 2 quarters and the longest 13 quarters. The typical decline in output from peak to trough, the recession’s amplitude, tends to be about 2 percent. A measure of cumulative loss is also computed, combining information about both duration and amplitude to proxy the overall cost of a recession. The cumulative loss of a recession is typically about 3 percent of GDP, but this number varies quite a bit across episodes.

Using the same methodology, the periods of decline in real credit and house prices are determined. The main focus is on those disruptions in credit or housing markets characterized by a peak-to-trough decline that falls into the top quartile of all credit or house price declines. These episodes are called credit crunches and house price busts, respectively. This method identifies 113 contractions and 28 crunches in credit, and 114 declines and 28 busts in house prices.

The episodes of credit crunches and house price busts tend to be long and deep. While a credit contraction episode typically lasts about six quarters, a credit crunch lasts a year longer. Credit contractions typically mean a 4 percent decrease in credit from peak to trough, but for crunches, the decrease is more than three times larger. House price busts tend to last even longer than credit crunches do. The typical episode of a decline in house prices is about nine quarters, whereas a house price bust usually persists twice as long. A typical house price decline is only 6 percent, but house prices tend to fall by five times as much during a house price bust.

Next, a simple episode-dating rule is used to determine whether a specific recession is associated with a credit crunch or house price bust period. If a recession episode starts at the same time as or after the beginning of an ongoing credit crunch or house price bust, the recession is considered to be associated with the credit crunch or house price bust. This rule, by definition, basically describes a timing association (a coincidence) between the two events but does not imply a causal link. With this rule, 48 recession episodes are identified to be associated with at least a credit crunch or house price bust. Out of these 48 episodes, 33 episodes are associated with house price busts and 21 with credit crunches.

Because the features of recessions associated with financial crises are also of interest, the relevant crisis episodes in the sample of advanced economies need to be identified. Following the same logic used above, Terrones, Scott, and Kannan (2009) identify whether a specific recession is associated with a financial crisis. They define financial crises as episodes during which there is widespread disruption to financial institutions and the functioning of financial markets. If a recession episode starts at the same time or after the beginning of an ongoing financial crisis, they classify that recession as being associated with the respective crisis. They report that using this rule, 15 recession episodes are associated with financial crises for the sample of countries used in this analysis.9

Recessions Associated with Disruptions in Credit or Housing Markets

Recessions associated with disruptions in credit or housing markets are simply different from other recessions without such disruptions. The analysis of these differences first focuses on the main characteristics of recessions: their duration and amplitude (Harding and Pagan, 2002). The duration of a recession, D, is the number of quarters, k, between a peak and the next trough. The amplitude of a recession, A, measures the change in output from a peak (y0) to the next trough (yk)—that is, A = yk – y0.

Another widely used measure, the cumulative loss, is also used to analyze the adverse impact of recessions on output. This measure combines information about the duration and amplitude of a phase to proxy the overall cost of a recession. To provide a sense of the variation in the recessions, the features of recessions coinciding with severe credit crunches or house price busts are also examined. These severe crunch or bust episodes consist of the top 12.5 percent of all credit contractions or house price declines (or the top half of all credit crunches or house price busts). Six recessions are accompanied by both a house price bust and a credit crunch. Some 26 recessions coincide with either a severe credit crunch or a severe house price bust.10

There are a number of statistically significant differences between recessions coinciding with credit crunches or house price busts and those not coinciding (Table 7.1). In particular, recessions associated with such episodes are on average more than a quarter longer than those without busts (4.3 versus 3.2 quarters). Moreover, output declines (and corresponding cumulative losses) are typically much larger in recessions with crunches or busts: 2.5 (4.8) percent versus 1.6 (2.3) percent in those without crunches or busts.

Table 7.1Recessions Associated with House Price Busts or Credit Crunches(percent change unless indicated otherwise)
Median valuesMean values
Without busts and crunchesWith busts or crunchesWith severe busts or crunchesWithout busts and crunchesWith busts or crunchesWith severe busts or crunches
Cumulative loss–2.30–4.8**–5.23**–3.67–10.60***–14.17***
Components of output
Total investment–3.45–6.07**–6.07–4.33–8.40**–8.46*
Residential investment–1.96–6.85**–7.52*–4.13–10.63***–12.50***
Nonresidential investment–2.85–4.31–4.44–3.95–6.87–6.51
Net exports (% of GDP)20.391.20***1.29**0.241.57***1.58**
Current account (% of GDP)20.170.92**0.63*0.171.15**1.20*
Other macroeconomic variables
Industrial production–3.97–4.79–5.31–3.80–4.29–4.78
Unemployment rate20.471.18**1.160.801.74***1.77***
Inflation rate2–0.10–0.63–0.33–0.16–0.44–0.12
Financial variables
House prices–0.24–5.96***–6.30***–0.02–8.44***–10.13***
Equity prices–8.85–0.58*–2.63–6.79–0.48*–0.40
Source: Authors’ estimates.Note: A severe house price bust or credit crunch is a bust or crunch in the top half of all busts or crunches. In each cell, the mean (median) change in the respective variable from peak to trough of recessions associated with house price busts or credit crunches is reported, unless indicated otherwise. The symbols *, **, and *** indicate that the difference between means (medians) of recessions with house price busts or credit crunches and recessions without house price busts or credit crunches is significant at the 10 percent, 5 percent, and 1 percent levels, respectively.

Number of quarters.

Change in the levels.

Source: Authors’ estimates.Note: A severe house price bust or credit crunch is a bust or crunch in the top half of all busts or crunches. In each cell, the mean (median) change in the respective variable from peak to trough of recessions associated with house price busts or credit crunches is reported, unless indicated otherwise. The symbols *, **, and *** indicate that the difference between means (medians) of recessions with house price busts or credit crunches and recessions without house price busts or credit crunches is significant at the 10 percent, 5 percent, and 1 percent levels, respectively.

Number of quarters.

Change in the levels.

These sizable differences also extend to the other macroeconomic variables, including consumption, investment, and the unemployment rate. For example, although consumption typically does not contract much in recessions, there is a statistically significant decline in consumption in recessions associated with credit crunches or house price busts, and for severe crunches and busts a 1 percentage point greater decline. The large decline likely reflects the substantial adverse effects of the lack of credit and erosion of housing wealth on consumption during these episodes. These findings indicate that recessions with credit crunches or house price busts result in more costly macroeconomic outcomes than do those without such disruptions. This is consistent with a large body of literature suggesting that credit and housing market developments play an important role in driving business cycles (Leamer, 2007; Mendoza and Terrones, 2008).

Trade variables also exhibit substantial differences between the recessions coinciding with crunches or busts and other types of recessions. In part reflecting the substantial decline in domestic demand, imports fall more in recessions with credit crunches or with house price busts. Along with an increase in exports, both net exports and the current account balance improve significantly more in recessions with such financial shocks.

With respect to financial outcomes, credit, house, and equity prices fall much more in recessions with credit crunches or house price busts. In particular, although credit continues to grow, albeit at a slower rate, during recessions without severe credit market problems, it contracts by about 1.6 percent during recessions coinciding with crunches or busts. House prices tend to register a decline of roughly 6 percent during these episodes. Equity prices also drop during these types of recessions.

The lag between the start of a credit crunch and the beginning of the corresponding recession is also examined. If a recession is associated with a credit crunch, it typically starts three quarters after the onset of the crunch. Because credit crunches last longer than recessions do, the latter tend to end two quarters before their corresponding credit crunch episodes. These findings suggest that the phenomenon of creditless recoveries is not specific to sudden stop episodes observed in emerging markets (Calvo, Izquierdo, and Talvi, 2006) but is also a feature of business cycles in advanced economies.

Similar to those recessions associated with credit crunches, recessions associated with house price busts tend to begin three quarters after the start of their respective house price busts. However, they end nine quarters ahead of the corresponding house price busts because house price busts typically last three times longer than recessions do. Moreover, when a recession is associated with a house price bust, residential investment stays depressed for a prolonged period and typically recovers only three to five quarters after the end of that recession.

These observations imply that recessions can end, and recoveries start, without a revival in credit growth and improvement in asset prices. This raises a natural question: What drives recoveries after recessions associated with credit crunches and house price busts? There could be several explanations. First, not all forms of demand depend on the availability of credit. In particular, consumption is typically the most important contributor to output growth during recoveries. Investment (especially nonresidential) recovers only with a lag, with the contribution of fixed investment growth to recovery often relatively small. Because consumption can be less credit intensive, a recovery could start without the stress in financial markets having been overcome. Second, firms and households may be able to get external financing from sources other than commercial banks. These sources are not captured in the aggregate credit series focused on by this analysis. Third, there can be a switch from more to less credit-intensive sectors in such a way that overall credit does not expand, yet, because of productivity gains, output increases. The aggregate data used in this analysis hide such reallocations of credit across sectors, including between corporations and households that vary in their credit intensity.

Recessions Associated with Financial Crises

The analysis now turns to the characteristics of recessions associated with financial crises. Some of these episodes also coincide with house price busts or credit crunches. Table 7.2 presents the findings, and compares the changes in the main macroeconomic and financial variables during recessions associated with crisis episodes and other recessions. Following Reinhart and Rogoff (2008), the implications of the Big Five financial crises are studied separately. The statistics associated with those recessions are reported under the column “with severe crises.”

Table 7.2Recessions Associated with Financial Crises(percent change unless indicated otherwise)
Median valuesMean values
Without crisesWith crisesWith severe crisesWithout crisesWith crisesWith severe crises
Cumulative loss–2.64–4.9***–4.90–5.24–14.68–27.20
Components of output
Total investment–3.82–10.44*–11.09–5.12–11.56–18.65
Residential investment–3.67–10.98***–12.27–5.69–13.24*–17.27
Nonresidential investment–3.52–9.78–17.44–4.34–10.27–19.78
Net exports (% of GDP)20.561.140.180.701.171.32
Current account (% of GDP)20.460.790.410.530.720.32
Other macroeconomic variables
Industrial production–3.92–5.66–2.79–3.92–4.47–3.07
Unemployment rate20.561.38**4.66**0.942.545.83
Inflation rate2–0.20–1.06***–4.13**–0.04–1.97**–3.15*
Financial variables
House prices–1.84–4.97**–6.21**–2.68–8.68–16.60
Equity prices–5.28–9.78–17.16–3.81–8.74–8.26
Source: Authors’ calculations.Note: A severe crisis refers to one of the Big Five crises. In each cell, the mean (median) change in the respective variable from peak to trough of recessions associated with financial crises is reported, unless indicated otherwise. The symbols *, **, and *** indicate that the difference between means (medians) of recessions with financial crises and reessions without financial crises is significant at the 10 percent, 5 percent, and 1 percent levels, respectively.

Number of quarters.

Change in level.

Source: Authors’ calculations.Note: A severe crisis refers to one of the Big Five crises. In each cell, the mean (median) change in the respective variable from peak to trough of recessions associated with financial crises is reported, unless indicated otherwise. The symbols *, **, and *** indicate that the difference between means (medians) of recessions with financial crises and reessions without financial crises is significant at the 10 percent, 5 percent, and 1 percent levels, respectively.

Number of quarters.

Change in level.

The average duration of a recession associated with a (severe) financial crisis exceeds that without a crisis by two (three) quarters. There is typically a larger output decline in recessions associated with crises compared with other recessions, −2.5 percent versus −1.8 percent, or a 0.7 percentage point difference (although not statistically significant). For recessions with severe crises, the difference in output decline is even larger, 0.9 percentage point, but is also statistically insignificant.

The cumulative output loss of recessions associated with a (severe) crisis is typically significantly larger than those without. In particular, the median cumulative loss of a recession associated with a crisis is roughly two times that of a recession without a crisis. Recessions with a crisis are generally associated with greater contractions in consumption, investment, industrial production, employment, exports, and imports, compared with those recessions without a crisis. These differences are significant for most variables. Recessions associated with financial crises often coincide with a rapid acceleration of the rate of unemployment. In particular, the increase in unemployment during recessions associated with severe financial crises is almost eight times larger than those recessions without crises. This suggests that the welfare costs of recessions with financial crises are much larger.

Credit, almost by construction, registers much larger (and statistically significant) declines in recessions with severe financial crises than those without. House prices also fall statistically significantly more in recessions with crises than those without. This might stem from the high sensitivity of housing activity to credit conditions. Equity prices also decrease much more in recessions with crises.

Dynamics of Recessions Associated with Financial Crises

Next, this chapter examines how the various macroeconomic and financial variables behave around recessions associated with crises compared with recessions that do not coincide with crises episodes (Figure 7.4). It focuses on patterns in the year-over-year growth in each variable over a six-year window—12 quarters before and 12 quarters after a peak. All panels include the median growth rate, that is, the typical behavior of events under consideration. As noted earlier, the sample includes 15 recessions associated with financial crises and 107 other recession episodes.

Figure 7.4Recessions and Financial Crises

(percent change from a year earlier)
(percent change from a year earlier)

Source: Authors’ calculations.

Note: The solid line denotes recessions not associated with financial crises, while the dotted line represents those recessions associated with financial crises. Zero on the x-axes is the quarter after which a recession begins (peak in the level of output).

The pattern of output growth around these recessions is as expected. Output registers a larger decline and takes a longer time to recover during recessions associated with financial crises than for other recessions. Consumption, investment, and industrial production also follow similar patterns. It is interesting that in recessions without crises, the growth rate of consumption slows down but does not fall below zero. In contrast, consumption contracts during recessions associated with financial crises. In recessions without a crisis, investment tends to take five to six quarters to expand again on an annual basis, but it often takes up to 10 quarters to do so during a recession accompanied by a crisis. Industrial production also exhibits a protracted period of contraction during recessions with crises. The rate of unemployment continues to rise up to three years after the recession starts when it is combined with a financial crisis.

With regard to financial variables, the growth rate of credit slows down in a typical recession whereas credit contracts somewhat during recessions associated with crises. House prices also decline more sharply during recessions with crises. Ahead of recessions associated with crises, both credit and house prices tend to grow at much higher rates than they do before other recessions, confirming the boom-bust cycles in these variables discussed in the previous section. Equity prices take a longer time to recover during recessions accompanied by crises.

An Anatomy of Recessions After The Financial Crisis

As noted previously, the majority of advanced economies in the sample entered into recession in late 2007 or early 2008. How similar to or different from earlier recessions are these recessions associated with the 2007–09 crisis? This section addresses this question using the data up to mid-2009, paying particular attention to the recession dynamics in the United States, the epicenter of the crisis.

As in the earlier section, this section studies the behavior of key macroeconomic and financial variables. It focuses on patterns in the year-over-year growth in each variable over a six-year window—12 quarters before and 12 quarters after a peak (Figure 7.5).11 All panels include the median year-over-year growth rate of these variables for all 122 recessions in 21 OECD countries in the sample, along with their upper and lower quartile bands. These bands allow the likelihood of various outcomes to be gauged, with the lower band representing worse than typical outcomes. Overlaid on each chart is information for the United States as of end 2009.

Figure 7.5Recessions in Organization for Economic Cooperation and Development Countries and Recent United States Recession

(percent change from a year earlier)
(percent change from a year earlier)
(percent change from a year earlier)

Source: Authors’ calculations.

Note: The solid line denotes the recent U.S. recession with 2007:Q4 as t = 0. The thick dotted line denotes the median of all observations while the thin dotted lines correspond to upper and lower quartiles. Zero on the x-axis is the quarter after which a recession begins (peak in the level of output). Unemployment rate is the level in percent. The date of the latest observation for the United States is 2009:Q2. (For total investment and unemployment it is 2009:Q3.)

The U.S. recession was clearly an outlier in many respects. First, confirming its severity, output registered a rate of growth below the median of the lower quartile of previous recession episodes five quarters after the beginning of the recession. Second, private sector demand also exhibited lower than typical growth. In particular, private consumption growth in the United States fell below the lower quartile band as households tried to cope with the sharp losses in their wealth and rebuilt their balance sheets. Investment growth declined more sharply than typical, reflecting the collapse in residential investment. The collapse in U.S. residential investment growth was exceptional, reflecting the bust in house prices and disruptions in credit markets.

Third, industrial production registered a much sharper decline than that of the lower quartile of all recessions. This suggests that the manufacturing cycle was more severe than in the past owing in part to the sharp decline in durables consumption. Moreover, unemployment climbed above the upper quartile of earlier episodes. The steep increase in unemployment reflects the sharp downsizing in many sectors of the U.S. economy, particularly in the financial sector.

With respect to asset prices, the U.S. recession was also quite different from previous ones. Although the decline in U.S. house prices was as steep as those observed during the Big Five episodes discussed previously, there was a much sharper decline in the growth rate of house prices than is typical in the OECD recessions. This is related, of course, to the sharp drop in residential investment. Although equity prices had increased until a few quarters before the recession began, a pattern not usually seen in the run-up to a recession, this quickly reversed and equity prices registered sharper than typical declines. Although house and equity prices rebounded in late 2009, they were still well below their precrisis highs.

Credit growth also started to slow down before the onset of the recession as the signs of financial stress began to emerge. This is more evidence of the negative feedback between asset prices, credit, and domestic demand, which, as discussed in the previous section, is common in severe recessions associated with financial crises. The growth rates of exports and imports both collapsed as the forces of recession became more intense over 2009. This observation is related to the highly synchronized nature of national recessions.

Another important feature of the 2007–09 recession is its global reach. Kose, Loungani, and Terrones (2010) analyze the implications of three previous global recessions (1975, 1982, and 1991) and compare these with the one in 2007–09. They define a global recession as a contraction in world real per capita GDP accompanied by a broad decline in various other measures of global economic activity. They reported at that time that the 2007–09 global recession easily qualifies as the most severe of the four global recessions: output—depending on the measure—was projected to fall between four and six times as much as it did on average in the three other global recessions, and unemployment was likely to increase twice as much. The collapse in world trade in 2009 dwarfs that in past global recessions. Moreover, no previous global recession has had so many countries in a state of recession simultaneously, both in advanced and developing economies.


This chapter provides a brief analysis of the three central questions about the global financial crisis. First, how similar is the most recent crisis to previous episodes? The latest crisis featured some close similarities to earlier ones, including the presence of credit and asset price booms fueled by rapid debt accumulation in a number of advanced economies. Second, how different is the most recent crisis from earlier episodes? As much similarity as the latest crisis had with the earlier episodes, it also featured some significant differences, such as in the explosion of opaque and complex financial instruments and in highly integrated global financial markets. Third, how costly were the recessions that followed the crisis? To answer this question, the chapter first examined whether recessions associated with financial market disruptions or outright financial crises are more damaging than other “normal” recessions. The findings indicate that the recessions in the former group result in much larger declines in economic activity and tend to last much longer. The analysis also considered the depth of the recession in the United States following the 2007–09 financial crisis and examined its severity in light of earlier recession episodes in a large sample of OECD countries. The latest recession is indeed an outlier in a number of respects.

In addition to the issues discussed in this chapter, the global financial crisis and associated recessions have led to an extensive discussion about the ability of macroeconomic and financial sector policies to mitigate the costs stemming from such episodes. The cost of a recession is, of course, affected by a number of factors. First, changes in credit and asset prices can have important implications for the severity of the recession. Second, prevailing economic conditions at the onset of a recession, such as global economic conditions and oil prices, can also be associated with different recession outcomes. Third, countercyclical macroeconomic and financial sector policies might mitigate the cost of a recession.

Although some observers argue that these macroeconomic and financial sector policies can help moderate recessions, others claim that they can worsen recession outcomes. Recent work, however, suggests that discretionary monetary and fiscal policies could help reduce the duration of recessions in the advanced economies (Terrones, Scott, and Kannan, 2009). In particular, evidence indicates that discretionary monetary policy is associated with shorter recessions, although fiscal policy does not have a significant impact on duration. By contrast, expansionary discretionary fiscal policies tend to shorten the duration of recessions associated with financial crises. This finding is consistent with evidence that fiscal policy is particularly effective when agents face tighter liquidity constraints.

The evidence on the effects of policies on the amplitude of a recession is, however, less robust. Claessens, Kose, and Terrones (2009) report that fiscal and monetary policies do not seem to have a significant impact on the depth of recessions. This finding could reflect several potential factors, including the coarse nature of the fiscal and monetary policy proxies they employ; lags on the policy effects, particularly with regard to fiscal policy; and several instances in which procyclical policies were in place to fight inflation. In summary, the evidence on the effectiveness of countercyclical policies during recessions is at best mixed, indicating fertile ground for future research.

The crisis has also provided important lessons about financial sector policies. In particular, it has exposed flaws in the precrisis regulatory framework and has shown the limits of policy measures in dealing with financial meltdowns. Although many elements of existing regulatory frameworks remain valid, the crisis has forced the relevant actors to think about the future architecture of regulatory policies. Although improvements in microprudential regulations are needed to reduce financial market procyclicality, rules calling for well-capitalized and transparent banks adhering to sound accounting standards are still critical. The crisis has made clear, however, that greater coordination between macroeconomic and financial policy is needed. Prudential regulation has to acquire a more macro, systemwide, dimension. The global nature of the financial crisis has also shown that although financially integrated markets have benefits, they also have risks, with large real economic consequences. It has shown that the international financial architecture is still far from institutionally matching the closely integrated financial systems.

The crisis has also had major financial and economic repercussions for emerging markets and developing economies, even though many of them were innocent bystanders. Some of these countries benefited from their improved fundamentals and were better able to tackle the adverse effects of the crisis on their economies. Short-term policy responses, involving more accommodative fiscal and monetary policies and better, restructured frameworks, were more effective than they were in earlier periods. However, the crisis also highlighted some specific financial sector reform challenges for emerging markets and developing economies.

Although there are a number of lessons for macroeconomic policy and regulation of the financial sector, many areas remain for which further policy research would be useful. These include competition policy for a more stable financial system, integration of macroeconomic and financial policy choices, approaches to consumer protection in financial services, and resolution of the political economy pressures regarding financial deregulation, financial openness, and financial crises.


    ÁrvaiZsófiaKarlDriessen and InciÖtker-Robe2009Regional Financial Interlinkages and Financial Contagion Within EuropeIMF Working Paper 09/6 (Washington: International Monetary Fund).

    AshcraftAdam and TilSchuermann2007Understanding the Securitization of Subprime Mortgage CreditStaff Report No. 318 (New York: Federal Reserve Bank of New York).

    BlanchardOlivier2009The Crisis: Basic Mechanisms, and Appropriate PoliciesIMF Working Paper 09/80 (Washington: International Monetary Fund).

    BorioClaudio2008The Financial Turmoil of 2007–?: A Preliminary Assessment and Some Policy ConsiderationsBIS Working Paper No. 251 (Basel: Bank for International Settlements).

    BrunnermeierMarkus K.2009Deciphering the 2007–08 Liquidity and Credit CrunchJournal of Economic Perspectives Vol. 23 No. 1 pp. 77100.

    BryGerhard and CharlotteBoschan1971Cyclical Analysis of Economic Time Series: Selected Procedures and Computer ProgramsNBER Technical Working Paper No. 20 (Cambridge, Massachusetts: National Bureau of Economic Research).

    CaballeroRicardo2010The ‘Other’ Imbalance and the Financial CrisisNBER Working Paper No. 15636 (Cambridge, Massachusetts: National Bureau of Economic Research).

    CalomirisCharles W.2009The Subprime Turmoil: What’s Old, What’s New, and What’s NextJournal of Structured Finance Vol. 15 No. 1 pp. 652.

    CalvoGuillermo A. AlejandroIzquierdo and ErnestoTalvi2006Phoenix Miracles in Emerging Markets: Recovering without Credit from Systemic Financial CrisesNBER Working Paper No. 12101 (Cambridge, Massachusetts: National Bureau of Economic Research).

    CardarelliRobertoM. Ayhan Kose and SelimElekdag2010Capital Inflows: Macroeconomic Implications and Policy ResponsesEconomic Systems Vol. 34 No. 4 pp. 33356.

    CecchettiStephen G.2009Crisis and Responses: The Federal Reserve in the Early Stages of the Financial CrisisJournal of Economic Perspectives Vol. 23 No. 1 pp. 5175.

    ClaessensStijn2009Lessons from the Recent Financial Crisis for Reforming National and International Financial SystemsPaper prepared for the Annual Bank Conference on Development EconomicsThe Road Ahead to a Sustainable Global Economic SystemSeoulJune 22–24.

    ClaessensStijnGiovanniDell’AricciaDenizIgan and LucLaeven2010Cross-Country Experience and Policy Implications from the Global Financial CrisisEconomic Policy Vol. 25 No. 2 pp. 26793.

    ClaessensStijnM. Ayhan Kose and Marco E. Terrones2009What Happens during Recessions, Crunches and Busts?Economic Policy Vol. 60 pp. 653700.

    ClaessensStijnM. Ayhan Kose and Marco E. Terrones2010Recessions and Financial Disruptions in Emerging MarketsEconomía Chilena Vol. 13 No. 2 pp. 5584.

    Dell’AricciaGiovanniAdolfoBarajas and AndreiLevchenko2008Credit Booms: The Good, the Bad, and the Ugly” (unpublished; Washington: International Monetary Fund).

    Dell’AricciaGiovanniDenizIgan and LucLaeven2008Credit Booms and Lending Standards: Evidence from the U.S. Subprime Mortgage MarketIMF Working Paper 08/106 (Washington: International Monetary Fund).

    Demirgüç-KuntAsli and EnricaDetragiache1998Financial Liberalization and Financial FragilityIMF Working Paper 98/93 (Washington: International Monetary Fund).

    FurceriDavide and AnnabelleMourougane2009Financial Crises: Past Lessons and Policy Implications” (Paris: Organization for Economic Cooperation and Development).

    GoldbergLinda S.2009Understanding Banking Sector GlobalizationIMF Staff Papers Vol. 56 No. 1 pp. 17197.

    GortonGary2009Slapped in the Face by the Invisible Hand: Banking and the Panic of 2007Paper prepared for the Federal Reserve Bank of Atlanta’s 2009 Financial Markets ConferenceFinancial Innovation and CrisisMay 11–13.

    GortonGary and AndrewMetrick2009HaircutsNBER Working Paper No. 15273 (Cambridge, Massachusetts: National Bureau of Economic Research).

    HardingDon and AdrianPagan2002Dissecting the Cycle: A Methodological InvestigationJournal of Monetary Economics Vol. 49 pp. 36581.

    HelblingThomas and MarcoTerrones2003Real and Financial Effects of Bursting Asset Price BubblesWorld Economic OutlookApril (Washington: International Monetary Fund).

    HongKiseokJohn-WhaLee and HsiaChink Tang2009Crises in Asia: Historical Perspectives and ImplicationsJournal of Asian Economics Vol. 21 No. 3 pp. 26579.

    KoseM. AyhanPrakashLoungani and MarcoTerrones2010Global Recessions and Recoveries” (unpublished; Washington: International Monetary Fund).

    KoseM. AyhanEswarPrasadKennethRogoff and Shang-JinWei2009Financial Globalization: A ReappraisalIMF Staff Papers Vol. 56 No. 1 pp. 862.

    KoseM. AyhanEswarPrasadKennethRogoff and Shang-JinWei2010Financial Globalization and Economic Policies” in Handbook of Development Economics Vol. 5 ed. by DaniRodrik and MarkRosenzweig (Amsterdam: Elsevier) pp. 4283362.

    KrugmanPaul2009Revenge of the GlutNew York TimesMarch1.

    LeamerEdward E.2007Housing IS the Business CycleNBER Working Paper No. 13428 (Cambridge, Massachusetts: National Bureau of Economic Research).

    MayerChristopherKarenPence and Shane M. Sherlund2009The Rise in Mortgage DefaultsJournal of Economic Perspectives Vol. 23 No. 1 pp. 2750.

    MendozaE. and M.E. Terrones2008An Anatomy of Credit Booms: Evidence from Macro Aggregates and Micro DataNBER Working Paper No. 14049 (Cambridge, Massachusetts: National Bureau of Economic Research).

    MishkinFrederic S.2009Is Monetary Policy Effective during Financial Crises?NBER Working Paper No. 14678 (Cambridge, Massachusetts: National Bureau of Economic Research).

    ObstfeldMaurice2009International Finance and Growth in Developing Countries: What Have We Learned?IMF Staff Papers Vol. 56 pp. 63111.

    ReinhartCarmen M. and Kenneth S. Rogoff2008Is the 2007 U.S. Subprime Crisis So Different? An International Historical ComparisonAmerican Economic Review Vol. 98 No. 2 pp. 33944.

    ReinhartCarmen M. and Kenneth S. Rogoff2009This Time Is Different: Eight Centuries of Financial Folly (Princeton, New Jersey: Princeton University Press).

    ShinHyun Song2009Reflections on Northern Rock: The Bank Run that Heralded the Global Financial CrisisJournal of Economic Perspectives Vol. 23 No. 1 pp. 10119.

    TaylorJohn B.2009The Financial Crisis and the Policy Responses: An Empirical Analysis of What Went WrongNBER Working Paper No. 14631 (Cambridge, Massachusetts: National Bureau of Economic Research).

    TerronesMarco E. AlasdairScott and PrakashKannan2009From Recession to Recovery: How Soon and How Strong?World Economic OutlookApril (Washington: International Monetary Fund).

A number of papers provide detailed discussions about the evolution of the crisis (Borio, 2008; Brunnermeier, 2009; Calomiris, 2009; Gorton, 2009; and Shin, 2009).

A number of papers examine the differences and similarities between the latest episode and past crises (Furceri and Mourougane, 2009). Some parts of the first and second sections extend Claessens (2009) and Claessens and others (2010).

Reinhart and Rogoff (2008) examine the run-up in house prices in the United States before 2007 and the Big Five crises in advanced economies and confirm significant increases in house prices before financial crises, and marked declines in the crisis year and in subsequent years. However, they note that the run-up in house prices before the U.S. crisis exceeded the run-ups in house prices before the Big Five.

Some 40 percent of the credit booms in these countries followed large gains in total factor productivity, 33 percent followed significant financial sector reform, and 27 percent followed large capital inflows.

Mayer, Pence, and Sherlund (2009) document that mortgage defaults and delinquencies were particularly concentrated among borrowers whose mortgages were classified as subprime or near prime. They report that many such borrowers put down small or no down payments when they purchased their homes, and were likely to have negative equity in their homes when house prices fell. This implies that they often were unable to sell before the bank could foreclose.

Gorton (2009) describes the trend toward the originate-to-distribute model and explains how it led to a decline in lending standards. He claims that banks increasingly financed their asset holdings with shorter maturity instruments, which left them particularly exposed to the drying up of funding liquidity.

Reinhart and Rogoff (2008) focus on the so-called Big Five financial crisis episodes, which include Finland (1990–93), Japan (1993), Norway (1988), Spain (1978–79), and Sweden (1990–93). These crises took a long time to resolve and all led to substantial fiscal costs.

Hong, Lee, and Tang (2009) examine the impact of shocks in 21 industrial, mostly OECD, countries on 21 developing Asian economies using annual data. They show that developing Asian (OECD) countries on average are in recession about 13.0 (8.5) percent of the time, and each recession lasted about 1.6 (1.3) years, with a cumulative loss of about 12.0 (2.6) percent.

The recession episodes associated with financial crises are the following: Australia, 1990:Q2–1991:Q2; Denmark, 1987:Q1–1988:Q2; Finland, 1990:Q2–1993:Q2*; France, 1992:Q2–1993:Q3; Germany, 1980:Q2–1980:Q4; Greece, 1992:Q2–1993:Q1; Italy, 1992:Q2–1993:Q3; Japan, 1993:Q2–1993:Q4*; Japan, 1997:Q2–1999:Q1; New Zealand, 1986:Q4–1987:Q4; Norway, 1988:Q2–1988:Q4*; Spain, 1978:Q3–1979:Q1*; Sweden, 1990:Q2–1993:Q1*; United Kingdom, 1973:Q3–1974:Q1; United Kingdom, 1990:Q3–1991:Q3. * Denotes the Big Five financial crises in Reinhart and Rogoff (2008, 2009), who provide a detailed history of these and other crisis episodes.

The sample includes 20 recessions associated with severe house price busts and 11 recessions associated with severe credit crunches.

At the quarterly frequency, year-over-year changes in the growth rates are used because quarter-over-quarter changes can be quite volatile and provide a noisy presentation of recession dynamics. For the unemployment rate the level rate is used.

    Other Resources Citing This Publication