## INTRODUCTION (wpiea2019083)

## Source details

**Canonical URL:** [INTRODUCTION (wpiea2019083)](https://www.imf.org/-/media/files/publications/wp/2019/wpiea2019083.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2019/wpiea2019083.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2019/wpiea2019083.pdf.json)

---

### Overview
- The 2008 financial crisis was the most severe shock to hit the global economy in more than 70 years; the most acute phase followed the September 15, 2008 collapse of the investment bank Lehman Brothers.
- Ten years after the crisis:
  - Median general government debt-GDP ratio stands at 51 percent, up from 36 percent before the crisis.
  - Central bank balance sheets, particularly in advanced economies, are several multiples of the size they were before the crisis.
  - Emerging market and developing economies account for 60 percent of global GDP in purchasing-power-parity terms (compared with 44 percent in the decade before the crisis).
- The paper addresses:
  - How output evolved across countries compared with precrisis trends.
  - How capital, labor inputs, and total factor productivity advanced after the crisis and their contributions to persistent output shortfalls.
  - Which policies and structural attributes account for cross-country variation in postcrisis performance and which actions limited output losses over the medium term (2015–17).
- Sample: 180 countries covering advanced, emerging market, and low-income developing economies.
- Methodological note: Precrisis trends are adjusted for influences such as credit growth that may affect the path of output beyond typical demand fluctuations. Even with adjustments, measured output deviations may capture slow-moving structural changes in trend growth rates unrelated to the crisis.

### Context and motivation
- Builds on previous WEO analysis (October 2009) but focuses on aftermath of the 2008 crisis.
- Cautionary examples:
  - United States: a slowdown in total productivity growth that predates the 2008 crisis has contributed to lower potential growth over time.
  - China: major structural shifts spanning the 2008 crisis and transition to slower growth.

*Source: IMF staff calculations (from the content unit "INTRODUCTION" of wpiea2019083).*

### Key empirical observations (overview)
- After the global financial meltdown in late 2008:
  - 91 economies representing two-thirds of global GDP in purchasing-power-parity terms experienced a decline in output in 2009.
  - By comparison, during the 1982 global recession 48 economies accounting for 46 percent of world GDP registered output declines compared with the previous year.
- Persistence:
  - The post-2008 output deviations exhibit strong persistence over time.
  - The correlation coefficient between GDP deviations for 2011–13 and 2015–17 is about 0.90.
- Adjustments to trend estimation are important because accommodative financial conditions likely contributed to unsustainable growth pre-2008; measured deviations may still reflect country-specific changes in trend growth rates unrelated to the crisis.

---

### QUANTIFYING LOSSES

### Measurement approach
- Postcrisis deviations of output are measured from the level that would have prevailed had output followed its pre-2009 trend growth rate (Ball 2014), with adjustments for factors such as credit growth where relevant.
- Precrisis components removed by low pass filters following Gourinchas and Obstfeld (2012) using two-sided Hodrick-Prescott (HP) filter with smoothing parameter set at 100.
- Trend of filtered series calculated over 2000–08.
- Deviations of post-crisis GDP from its pre-crisis trend are calculated as the average differences for 2011–13 and 2015–17.

### Empirical findings on output deviations
- Strong persistence of output deviations: correlation between 2011–13 and 2015–17 deviations ≈ 0.90.
- Output deviations a decade after 2008 are more skewed toward losses than after the 1982 global recession.
- Economies with larger output and employment losses in the initial aftermath registered greater increases in income inequality relative to precrisis average.
- Employment losses measured as the gap between employed workers and number consistent with employment growing at the same rate during the postcrisis period as the economically active cohort aged 15–65.

---

### 2. Employment Deviations — Patterns, Channels, and Correlates

### Persistent losses: aggregate patterns and magnitudes
- Banking crises in 2007–08 (Laeven-Valencia definition): 24 countries, 18 in advanced economies.
- Among the 24 banking crisis countries, about 85 percent still show negative deviations from the pre-2009 trend a decade after the 2008 meltdown.
- Output remains below precrisis trends in about 60 percent of economies (including many without banking crises in 2007–08).
- By 2017, on average, investment was about 25 percent below precrisis trend.
- Close to 80 percent of economies that suffered a banking crisis in 2007–08 experienced shortfalls in capital relative to precrisis trends.
- Among economies without a banking crisis in 2007–08, capital stocks of about 65 percent appear lower than they would be if capital accumulation had followed the extrapolated precrisis trend path.
- Median share of output per worker deviation accounted for by total factor productivity, 2015–17:
  - Countries without banking crisis in 2007–08: 70.4
  - 2007–08 banking crisis countries: 80.5
- Median share of GDP per worker deviation accounted for by total factor productivity, 2015–17:
  - Countries without banking crisis in 2007–08: 79.3
  - 2007–08 banking crisis countries: 78.2

### Channels — proximate causes: investment, capital, and TFP shortfalls
- Output per worker deviations mirror aggregate output deviations, implying changes in labor input cannot account for the bulk of observed output losses.
- Sluggish investment is a primary proximate cause:
  - Investment shortfalls may reflect lack of credit access and weak expectations of future growth/profitability.
  - Consequences:
    - Large capital shortfalls relative to precrisis trends.
    - Slower technology adoption where technology is embodied in equipment, reflected in residual TFP deviations.
- Growth-accounting decomposition attributes a large role to residual TFP in explaining output per worker deviations once capital per worker deviations are accounted for.
- Caveat: TFP residual also reflects measurement error in factors of production and changes in capacity utilization; cross-country data do not allow full separation of TFP deviations into those due to sluggish investment versus efficiency or other factors.

### Variation in postcrisis performance — main correlates (empirical approach)
- Cross-sectional OLS regressions estimate ∆y_i (output deviations during 2011–13 and in some specifications 2015–17) as a function of precrisis (2005–08 averages) macrofinancial vulnerabilities, policy space, structural rigidities, and a 2007–08 banking crisis dummy.
- Nature of the shock:
  - Countries with banking crises in 2007–08 suffered, on average, a 4 percentage point higher output loss during 2011–13 relative to the precrisis trend than those that did not.
- Macroeconomic imbalances and financial factors:
  - Weaker current account balances entering the crisis associated with bigger output losses relative to precrisis trends.
  - Faster precrisis credit growth associated with larger postcrisis losses in tighter financial conditions.
  - Greater integration into global financial markets (larger external assets and liabilities relative to GDP) associated with larger deviations from precrisis trend.
- Labor market structure:
  - Economies with greater difficulty of dismissal (Cambridge University CBR index) suffered larger postcrisis output losses relative to precrisis trends.
- Spillovers:
  - Economies more exposed to demand from advanced economies suffered larger output losses after the crisis.
  - Demand exposure to China was an offsetting influence for many exporters: exporters more exposed to China benefited disproportionately from China’s 4 trillion yuan stimulus during 2008–11.
- Precrisis policies and policy frameworks:
  - Stronger restrictions on certain banking activities (Barth, Caprio, and Levine 2013) in 2006 associated with lower probability of a banking crisis during 2007–08.
  - Countries with smaller increases in general government debt over 2005–08 experienced smaller losses relative to trends.
  - Exchange rate flexibility is associated with less damage; the euro area periphery experienced notably larger median output losses in 2011–13 and wider divergence through 2015–17.

### Policy-relevant implications and considerations
- Financial-sector resilience before a crisis matters: restrictions on certain bank activities and stronger regulatory/supervisory frameworks associated with lower probability of banking crises.
- Preserving policy space precrisis (lower precrisis public borrowing requirements) associated with smaller postcrisis output losses.
- Exchange rate flexibility can buffer asymmetric shocks; absence of such flexibility magnified losses in the euro area periphery.
- Addressing sluggish investment and capital shortfalls is central to closing persistent output gaps; policies to restore credit intermediation, rebuild balance sheets, and support productive investment can matter for recovery and TFP outcomes.
- Trade and financial spillovers: exposure to advanced-economy demand amplified losses, while exposure to China’s stimulus attenuated them for some exporters.

*Source: IMF staff calculations and analysis (section “2. Employment Deviations” from the provided chapter).*

---

### 1. Median — Policy Actions and Outcomes

### Postcrisis monetary policy actions
- Advanced economy monetary policy actions included: quantitative easing (purchases mainly of government bonds, mortgage-backed securities, and corporate bonds), state-dependent forward guidance, negative interest rates, and yield-curve control.
- Quantitative easing: estimates of impact on interest rates and financial conditions vary; generally believed to have had positive effects on domestic output in advanced economies and on imports from trading partners (IMF 2014).
- Controversy: policymakers in emerging market and developing economies raised concerns about adverse spillovers from advanced economy central banks’ unconventional monetary policy approaches.

### Fiscal and quasi-fiscal support for the financial sector
- Focus on fiscal and quasi-fiscal measures to support the financial sector after the crisis (Table 6).
- G20 economies, on average, injected discretionary fiscal stimulus of just over 2 percent of GDP in 2009 and 2010.
- Regression evidence (Table 7 and Figure 11) — conditional on size of initial losses during 2011–13 — shows headline support for the financial sector has a statistically significant positive correlation with subsequent output deviations from trend; capital injections and guarantees appear to have helped limit subsequent output losses.
- Selected reported regression coefficients and significance levels:
  - Total Headline Support for Financial and Other Sectors 0.20**
  - Capital Injections 1.90*
  - Purchase of Assets and Lending by Treasury 0.21
  - Central Bank Support with Treasury Backing –14.35
  - Central Bank Liquidity Support –0.25
  - Guarantees (excluding Deposit Insurance) 0.24*
  - Upfront Government Financing 0.31
  - Crisis-Related Discretionary Fiscal Stimulus –0.78
  - Banking Crisis in 2007–08 –0.17 –1.74 2.88 3.54* 3.06 –1.35 1.71 2.25
  - GDP Deviation 2011–13 1.12*** 1.05*** 1.10*** 1.08*** 1.10*** 1.06*** 1.09*** 1.33***
  - Constant –5.95*** –5.08*** –4.79** –4.04** –2.04 –5.12** –4.72** –1.33
  - Observations 29 29 29 29 29 28 29 19
  - R2 0.62 0.60 0.53 0.54 0.54 0.60 0.53 0.50

### Reported cross-country aggregates (Table 6 / panel labels preserved)
- Capital Injection (A), Purchase of Assets, Lending by Treasury (B), Central Bank Support with Treasury Backing (C), Central Bank Liquidity Support (D), Guarantees (E), Total (A+B+C+D+E)
  - G20 Average (PPP GDP weights) 2.03 3.3 1.0 9.2 14.3 29.8
  - Advanced Economies 2.95 5.0 1.2 12.9 21.3 43.3
  - Advanced Europe 2.43 3.6 2.1 1.0 19.5 28.6
  - Emerging Markets 0.30 0.1 0.3 1.8 0.2 2.7

### Multilateral and IMF actions
- IMF provided unconditional financial resources through a general allocation of SDR 204 billion ($316 billion) during August–September 2009.
- IMF approved SDR 420 billion in support to its members during 2008–13, of which SDR 119 billion was drawn during that interval.

### Summary of empirical findings and policy implications (median section)
- Persistent output losses followed the 2008 financial meltdown across a broad set of countries, not only those with banking crises.
- Protracted weak investment was a major contributing factor to persistent shortfalls in capital and total factor productivity and slower technology adoption among countries hit harder by the crisis.
- Stronger banking regulation (proxied by restrictions on certain aspects of bank activity) appears to have lowered the probability of a banking crisis in 2007–08.
- Countries with stronger fiscal positions entering the crisis suffered smaller losses, indicating greater room for policy maneuver helped defend against harm.
- Extraordinary fiscal and quasi-fiscal actions to support the financial sector after the crisis appear to have helped lessen output losses over the medium term; economies that moved quickly to assess banking-system health and recapitalize banks appeared to suffer smaller subsequent output losses.
- China’s large fiscal stimulus during 2008–11 appears to have had favorable spillovers to trading partners.
- Unconventional monetary policy actions by advanced economy central banks helped limit output declines and employment losses at home while supporting imports from abroad.
- Side effects of policy efforts over the past decade include buildup of financial vulnerabilities from an extended period of ultralow interest rates in advanced economies and large accumulation of public debt that has eroded fiscal buffers.

---

### Tests of Distributions and Banking Crisis Probabilities

### Tests of equality of distributions (2015–17 deviations) — two-sample Wilcoxon rank-sum (Mann-Whitney)
- GDP: Average Percentile 39.45; Expected Percentile 0.3; P-Value 0.052
- Capital Stock: Average Percentile 47.75; Expected Percentile 0.3; P-Value 0.630
- Total Factor Productivity: Average Percentile 41.55; Expected Percentile 0.5; P-Value 0.079
- Interpretation: The null hypothesis of equality of distributions is rejected for output and total factor productivity deviations, but not rejected for capital stock deviations.
- Note: Analysis omits countries with large output deviations caused by war or political strife.

### Regression model for probability of a banking crisis (2007–08)
- Model specification: Pr(banking crisis) = f(regulation, Ɵ) (equation (3) in source).
- Regulation index: Barth, Caprio, and Levine (2013).
- Main finding: Strength of restriction on banking activities in 2006 associated with a lower probability of a banking crisis in 2007–08; coefficient statistically significant.
- Robustness: Robustness tests on the probit regression presented in Annex Table 4.

### Annex Table 3 — Probability of Banking Crisis and Strength of Restrictions on Banking Activities (reported lines)
- Strength of Restrictions on Banking Activities –0.72*** –1.27*** –0.18***
- Constant –1.04*** –1.79*** 0.19***
- Observations 116 116 116
- R2 0.17
- Note: LPM = linear probability model. Significance: *** p < 0.01, ** p < 0.05, * p < 0.1.

### Expanded coefficient lines (as presented)
- Strength of Restrictions on Banking Activities –0.72*** –0.71*** –0.61*** –0.60*** –0.60** –0.65** –0.43 –0.46*
- Fraction of Bank Application Denied –1.50*** –1.55*** –1.60* –1.07*** –1.13*** –1.32** –0.90** –1.25*
- Bank Concentration 0.05
- Supervisory Power –0.10
- Capital Regulation –0.16
- Share of Interest Borrowing from G5 –0.01 0.27 0.35
- Financial Openness 1.21** 2.48** 1.94*
- Demand Exposure to Advanced Economies 3.13** 4.89** 3.36
- Constant lines and Observations preserved as reported in the source layout.
- Note: G5 = Group of Five (France, Germany, Netherlands, United Kingdom, and United States). Significance: *** p < 0.01, ** p < 0.05, * p < 0.1.

---

### Annex: Definitions, Data Construction, and Robustness
- Banking crises defined per Laeven and Valencia (2013) — significant financial distress and significant government intervention; sample includes all banking crises that started between 2007–08.
- Deviations from pre-crisis trends:
  - Two-sided HP lowpass filter with smoothing parameter set at 100 used to remove transitory pre-crisis components.
  - Trend calculated over 2000–08; deviations averaged for 2011–13 and 2015–17.
- Employment deviations: approach by Schanzenbach and others (2017), extended to 102 countries; employment gap formula reported.
- TFP deviations: calculated using standard Cobb-Douglas production function for output per worker and comparing observed post-crisis values in labor productivity and output per worker with pre-crisis trends.
- Treatment of explanatory variables in regressions:
  - Explanatory variables averaged over 2005–08.
  - Regressors (except banking crisis dummy) standardized to zero mean and standard deviation unity.
  - Regressors winsorized to alleviate influence of outliers.
- Robustness checks: estimates robust to allowing for a structural break (modelled as 5-σ shock to potential GDP in 2009); multivariate filter yields estimates in agreement with HP filter (lambda=100).

*Source: IMF staff calculations; figures, tables, and quoted numeric values are from the source content provided.*

### INTRODUCTION ________________________________________________________5

### INTRODUCTION

### Overview
- The 2008 financial crisis was the most severe shock to hit the global economy in more than 70 years. The most acute phase followed the September 15, 2008 collapse of the investment bank Lehman Brothers.
- Ten years after the crisis:
  - Median general government debt-GDP ratio stands at 51 percent, up from 36 percent before the crisis.
  - Central bank balance sheets, particularly in advanced economies, are several multiples of the size they were before the crisis.
  - Emerging market and developing economies account for 60 percent of global GDP in purchasing-power-parity terms (compared with 44 percent in the decade before the crisis).
- The paper addresses:
  - How output evolved across countries compared with precrisis trends.
  - How capital, labor inputs, and total factor productivity advanced after the crisis and their contributions to persistent output shortfalls.
  - Which policies and structural attributes account for cross-country variation in postcrisis performance and which actions limited output losses over the medium term (2015–17).
- Sample: 180 countries covering advanced, emerging market, and low-income developing economies.
- Methodological note: Precrisis trends are adjusted for influences such as credit growth that may affect the path of output beyond typical demand fluctuations. Even with adjustments, measured output deviations may capture slow-moving structural changes in trend growth rates unrelated to the crisis.

### Context and motivation
- The paper builds on previous WEO analysis (October 2009) that examined output performance after financial crises during 1970–2002, but zeroes in on the aftermath of the 2008 crisis.
- Key caution: Some countries experienced temporarily elevated potential growth rates before the crisis that subsequently reverted to long-run average; examples include:
  - United States: a slowdown in total productivity growth that predates the 2008 crisis has contributed to lower potential growth over time.
  - China: major structural shifts spanning the 2008 crisis and transition to slower growth.

---

### QUANTIFYING LOSSES

### Quantifying Post-Crisis Deviations in Output from Pre-Crisis Trends
- After the global financial meltdown in late 2008:
  - 91 economies representing two-thirds of global GDP in purchasing-power-parity terms experienced a decline in output in 2009.
  - By comparison, during the 1982 global recession 48 economies accounting for 46 percent of world GDP registered output declines compared with the previous year.
- Measurement approach: Postcrisis deviations of output are measured from the level that would have prevailed had output followed its pre-2009 trend growth rate (Ball 2014), with adjustments for factors such as credit growth where relevant.
- Persistence:
  - The post-2008 output deviations exhibit strong persistence over time.
  - The correlation coefficient between GDP deviations for 2011–13 and 2015–17 is about 0.90.
  - As shown in Annex Figure 3, output deviations close to a decade after the 2008 crisis are more skewed toward losses than those registered at a similar interval after the 1982 global recession.
- Inequality:
  - Economies with larger output and employment losses in the initial aftermath of the crisis registered greater increases in income inequality compared with their precrisis average.
  - Employment losses are measured as the gap between the number of employed workers and the number consistent with employment growing at the same rate during the postcrisis period as the economically active cohort between the ages of 15 and 65.

### Key empirical observations (as presented)
- Figure 1 (Correlation of GDP Deviations between Periods) documents the strong persistence of deviations (2011–13 versus 2015–17) and shows average percent deviations from precrisis trend.
- Adjustments to trend estimation are important because generally accommodative financial conditions likely contributed to unsustainable growth in many countries prior to 2008; nevertheless, measured deviations may still reflect country-specific changes in trend growth rates unrelated to the crisis.

---

*Source: IMF staff calculations (from the content unit "INTRODUCTION" of wpiea2019083).*

### 2. Employment Deviations

### 2. Employment Deviations

### Persistent losses: aggregate patterns and magnitudes
- According to the Laeven-Valencia definition, there were banking crises in 24 countries during 2007–08, 18 of those in advanced economies.
- Among the 24 economies in the banking crisis group, about 85 percent still show negative deviations from the pre-2009 trend a decade after the 2008 meltdown.
- Output remains below precrisis trends in about 60 percent of economies (including many that did not suffer a banking crisis in 2007–08).
- By 2017, on average, investment was about 25 percent below precrisis trend.
- Close to 80 percent of economies that suffered a banking crisis in 2007–08 experienced shortfalls in capital relative to precrisis trends.
- Among economies without a banking crisis in 2007–08, capital stocks of about 65 percent appear to be lower than they would be if capital accumulation had followed the extrapolated precrisis trend path.
- The median share of output per worker deviation accounted for by total factor productivity, 2015–17:
  - Countries without banking crisis in 2007–08: 70.4
  - 2007–08 banking crisis countries: 80.5
  - Median share of GDP per worker deviation accounted for by total factor productivity, 2015–17:
    - Countries without banking crisis in 2007–08: 79.3
    - 2007–08 banking crisis countries: 78.2

### Channels — proximate causes: investment, capital, and TFP shortfalls
- Output per worker deviations mirror aggregate output deviations, implying changes in labor input cannot account for the bulk of observed output losses.
- Sluggish investment is a primary proximate cause:
  - Investment shortfalls may reflect lack of credit access and weak expectations of future growth/profitability (secular stagnation narrative).
  - Two consequences of sluggish investment:
    - Large capital shortfalls relative to precrisis trends (see capital stock shortfall statistics above).
    - Slower technology adoption where technology is embodied in equipment, reflected in residual TFP deviations.
- Growth-accounting decomposition attributes a large role to residual TFP in explaining output per worker deviations once capital per worker deviations are accounted for.
- Caveat: TFP residual also reflects measurement error in factors of production and changes in capacity utilization; cross-country data do not allow full separation of TFP deviations into those due to sluggish investment versus efficiency or other factors.

### Variation in postcrisis performance — empirical approach and main correlates
- Empirical approach:
  - Cross-sectional OLS regressions estimate ∆y_i (output deviations during 2011–13 and in some specifications 2015–17) as a function of precrisis (2005–08 averages) macrofinancial vulnerabilities, policy space, structural rigidities, and a 2007–08 banking crisis dummy.
- The nature of the shock matters:
  - Countries that experienced banking crises in 2007–08 suffered, on average, a 4 percentage point higher output loss during 2011–13 relative to the precrisis trend than those that did not experience banking crises in 2007–08.
- Macroeconomic imbalances and financial factors:
  - Countries with current account balances weaker than levels consistent with fundamentals entering the crisis suffered bigger output losses relative to precrisis trends.
  - Faster precrisis credit growth is associated with larger postcrisis losses in tighter financial conditions.
  - Greater integration into global financial markets (larger external assets and liabilities relative to GDP) is associated with larger deviations from precrisis trend.
- Labor market structure:
  - Economies with greater difficulty of dismissal (Cambridge University CBR index) suffered larger postcrisis output losses relative to precrisis trends, suggesting labor rigidities impeded postcrisis recovery.
- Spillovers:
  - Economies more exposed to demand from advanced economies suffered larger output losses after the crisis.
  - Demand exposure to China was an offsetting influence for many exporters: economies whose export baskets were more exposed to China before the crisis benefited disproportionately from China’s 4 trillion yuan stimulus during 2008–11.
- Precrisis policies and policy frameworks:
  - Stronger restrictions on certain banking activities (per Barth, Caprio, and Levine 2013) in 2006 are associated with a lower probability of a banking crisis during 2007–08.
  - Countries with smaller increases in general government debt over 2005–08 experienced smaller losses relative to trends, consistent with greater fiscal space to support demand in the aftermath.
  - Exchange rate flexibility is associated with less damage; the euro area periphery experienced notably larger median output losses in 2011–13 and wider divergence through 2015–17, reflecting limited nominal exchange rate adjustment and other architecture gaps.

### Policy-relevant implications and considerations
- Financial-sector resilience before a crisis matters: restrictions on certain bank activities and stronger regulatory/supervisory frameworks are associated with lower probability of banking crises.
- Preserving policy space precrisis (lower precrisis public borrowing requirements) is associated with smaller postcrisis output losses.
- Exchange rate flexibility can serve as a buffer to absorb asymmetric shocks; absence of such flexibility magnified losses in the euro area periphery.
- Addressing sluggish investment and capital shortfalls is central to closing persistent output gaps; policies to restore credit intermediation, rebuild balance sheets, and support productive investment can matter for recovery and TFP outcomes.
- Trade and financial spillovers highlight the importance of external linkages: exposure to advanced-economy demand amplified losses, while exposure to China’s stimulus attenuated them for some exporters.

*Source: IMF staff calculations and analysis (section “2. Employment Deviations” from the provided chapter).*

### 1. Median

### 1. Median

### Postcrisis monetary policy actions
- Advanced economy monetary policy actions included: quantitative easing (massive balance sheet expansion with purchases mainly of government bonds, mortgage-backed securities, and corporate bonds), state-dependent forward guidance, negative interest rates, and yield-curve control.
- Quantitative easing: estimates of impact on interest rates and financial conditions vary; generally believed to have had positive effects on domestic output in advanced economies and on imports from trading partners (IMF 2014).  
- Controversy: policymakers in emerging market and developing economies raised concerns about adverse spillovers from advanced economy central banks’ unconventional monetary policy approaches.

### Fiscal and quasi-fiscal support for the financial sector
- The analysis focuses on fiscal and quasi-fiscal measures to support the financial sector after the crisis (Table 6).
- Group of Twenty (G20) economies, on average, injected discretionary fiscal stimulus of just over 2 percent of GDP in 2009 and 2010.
- Regression evidence (Table 7 and Figure 11) — conditional on size of initial losses during 2011–13 — shows headline support for the financial sector has a statistically significant positive correlation with subsequent output deviations from trend; capital injections and guarantees appear to have helped limit subsequent output losses.
- Selected reported regression coefficients and significance levels (as shown):
  - Total Headline Support for Financial and Other Sectors 0.20**
  - Capital Injections 1.90*
  - Purchase of Assets and Lending by Treasury 0.21
  - Central Bank Support with Treasury Backing –14.35
  - Central Bank Liquidity Support –0.25
  - Guarantees (excluding Deposit Insurance) 0.24*
  - Upfront Government Financing 0.31
  - Crisis-Related Discretionary Fiscal Stimulus –0.78
  - Banking Crisis in 2007–08 –0.17 –1.74 2.88 3.54* 3.06 –1.35 1.71 2.25
  - GDP Deviation 2011–13 1.12*** 1.05*** 1.10*** 1.08*** 1.10*** 1.06*** 1.09*** 1.33***
  - Constant –5.95*** –5.08*** –4.79** –4.04** –2.04 –5.12** –4.72** –1.33
  - Observations 29 29 29 29 29 28 29 19
  - R2 0.62 0.60 0.53 0.54 0.54 0.60 0.53 0.50
- Reported cross-country aggregates (Table 6 / panel labels preserved):
  - Capital Injection Purchase of Assets, Lending by Treasury Central Bank Support with Treasury Backing Central Bank Liquidity Support Guarantees Total (A)(B)(C)(D)(E)(A+B+C+D+E)
  - G20 Average (PPP GDP weights)2.03.31.09.214.329.8
  - Advanced Economies2.95.01.212.921.343.3
  - Advanced Europe2.43.62.11.019.528.6
  - Emerging Markets0.30.10.31.80.22.7

### Multilateral and IMF actions
- IMF provided unconditional financial resources through a general allocation of SDR 204 billion ($316 billion) during August–September 2009.
- IMF approved SDR 420 billion in support to its members during 2008–13, of which SDR 119 billion was drawn during that interval.

### Summary of empirical findings and policy implications
- Persistent output losses followed the 2008 financial meltdown across a broad set of countries, not only those with banking crises.
- Protracted weak investment was a major contributing factor to persistent shortfalls in capital and total factor productivity and slower technology adoption among countries hit harder by the crisis.
- Stronger banking regulation (proxied by restrictions on certain aspects of bank activity) appears to have lowered the probability of a banking crisis in 2007–08, suggesting a preventive role for regulation.
- Countries with stronger fiscal positions entering the crisis suffered smaller losses, indicating greater room for policy maneuver helped defend against harm.
- Extraordinary fiscal and quasi-fiscal actions to support the financial sector after the crisis appear to have helped lessen output losses over the medium term; economies that moved quickly to assess banking-system health and recapitalize banks appeared to suffer smaller subsequent output losses.
- China’s large fiscal stimulus during 2008–11 appears to have had favorable spillovers to trading partners.
- Unconventional monetary policy actions by advanced economy central banks helped limit output declines and employment losses at home while supporting imports from abroad.
- Side effects of policy efforts over the past decade include buildup of financial vulnerabilities from an extended period of ultralow interest rates in advanced economies and large accumulation of public debt that has eroded fiscal buffers.

### Annex: definitions and data construction (key points)
- Definition of banking crises: based on Laeven and Valencia (2013) — significant financial distress and significant government intervention; sample includes all banking crises that started between 2007–08. Annex Table 1 lists systemic and borderline cases (countries and start year as shown).
- Deviations from pre-crisis trends:
  - First, transitory pre-crisis components removed by low pass filters; approach follows Gourinchas and Obstfeld (2012) using two-sided Hodrick-Prescott (HP) lowpass filter with smoothing parameter set at 100 (higher than standard business cycle detrending).
  - Second, trend of filtered series calculated over 2000–08.
  - Finally, deviations of post-crisis GDP from its pre-crisis trend are calculated as the average differences for 2011–13 and 2015–17.
- Employment deviations: calculated using approach by Schanzenbach and others (2017), extended to 102 countries; employment gap formula reported as (equation 1).
- TFP deviations: calculated using standard Cobb-Douglas production function for output per worker and comparing observed post-crisis values in labor productivity and output per worker with pre-crisis trends (equation 2).
- Treatment of explanatory variables in regressions:
  - Explanatory variables averaged over 2005-08.
  - Regressors (except banking crisis dummy) standardized to zero mean and standard deviation unity.
  - Regressors winsorized to alleviate influence of outliers.
- Robustness checks: estimates robust to allowing for a structural break (modelled as 5-σ shock to potential GDP in 2009); multivariate filter yields estimates in agreement with HP filter (lambda=100).  

*Source: IMF staff calculations; figures, tables, and quoted numeric values are from the source content provided.*

### 2. The table shows the rejection of the null hypothesis of equality of distributions in the cases

### wpiea2019083 - 2. The table shows the rejection of the null hypothesis of equality of distributions in the cases

### Tests of equality of distributions (2015–17 deviations)
- The two-sample Wilcoxon rank-sum (Mann-Whitney) test results:
  - GDP: Average Percentile 39.45; Expected Percentile 0.3; P-Value 0.052
  - Capital Stock: Average Percentile 47.75; Expected Percentile 0.3; P-Value 0.630
  - Total Factor Productivity: Average Percentile 41.55; Expected Percentile 0.5; P-Value 0.079
- Interpretation: The null hypothesis of equality of distributions is rejected for output and total factor productivity deviations, but not rejected for capital stock deviations.
- Note: The analysis omits countries with large output deviations that were caused by war or political strife.

### Regression model for probability of a banking crisis (2007–08)
- Model specification (as presented):
  - Pr(banking crisis) = f(regulation, Ɵ)  (equation (3) in source)
  - regulation is a measure of various aspects of banking regulation; Ɵ is the set of parameters to be estimated.
- Banking regulation index source: Barth, Caprio, and Levine (2013).
- Main finding: Strength of restriction on banking activities in 2006 is associated with a lower probability of a banking crisis in 2007–08 and the coefficient is statistically significant.
- Robustness: Robustness tests on the probit regression are presented in Annex Table 4.

### Annex Table 3 — Probability of Banking Crisis and the Strength of Restrictions on Banking Activities
- Reported coefficient lines (as presented in source):
  - Strength of Restrictions on Banking Activities –0.72***–1.27***–0.18***
  - Constant –1.04***–1.79***0.19***
  - Observations 116116116
  - R2 0.17
  - Source: Barth, Caprio and Levine (2013); IMF staff calculations.
  - Note: LPM = linear probability model.
  - Significance: *** p < 0.01, ** p < 0.05, * p < 0.1.

- Expanded coefficient table lines (as presented in source):
  - Probit Logit LPM
  - Strength of Restrictions on Banking Activities–0.72***–0.71***–0.61***–0.60***–0.60**–0.65**–0.43–0.46*
  - Fraction of Bank Application Denied –1.50*** –1.55***–1.60* –1.07***–1.13***–1.32**–0.90**–1.25*
  - Bank Concentration 0.05
  - Supervisory Power –0.10
  - Capital Regulation –0.16
  - Share of Interest Borrowing from G5 –0.01 0.27 0.35
  - Financial Openness 1.21** 2.48** 1.94*
  - Demand Exposure to Advanced Economies 3.13** 4.89** 3.36
  - Constant –1.04***–1.14***–0.74***–0.83***–0.88***–1.01***–0.98***–0.90***–1.15***–1.17**–0.81***–1.11***–1.20***–0.87***
  - Observations 50 51 51 116 115 53 54 54 51 11 65 4 52 98 98 111 (15) [numbers preserved as in source layout]
  - Source: Barth, Caprio, Levine (2013); and IMF staff calculations.
  - Note: Group of Five (France, Germany, Netherlands, United Kingdom, and United States). G5 = Group of Five.
  - Significance: *** p < 0.01, ** p < 0.05, * p < 0.1.

### Annex Table 4 — Banking Crisis and Regulations: Probit Regression
- The source lists multiple probit specifications (columns (1)–(14)) with observations counts and coefficients reported in Annex Table 4; robustness tests are presented there (table details preserved in source).

### Key methodological and data notes
- Regulation index: drawn from Barth, Caprio, and Levine (2013).
- Regression framework: probit (with logit and LPM reported for comparison).
- Statistical significance notation preserved: *** p < 0.01, ** p < 0.05, * p < 0.1.

*Source: IMF staff calculations; Barth, Caprio and Levine (2013).*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2019/wpiea2019083.pdf_