## wpiea2020276-print-pdf

## Source details

**Canonical URL:** [wpiea2020276-print-pdf](https://www.imf.org/-/media/files/publications/wp/2020/english/wpiea2020276-print-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2020/english/wpiea2020276-print-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2020/english/wpiea2020276-print-pdf.pdf.json)

---

### I. Introduction and dataset
- COVID-19 characterization and macro impact:
  - Described as an unprecedented shock with far-reaching economic and financial repercussions worldwide.
  - As of December 10, 2020: over 69.3 million confirmed cases of COVID-19 in 190 countries, with more than 1.5 million deaths.
  - IMF projection: global real GDP growth is projected to contract by 4.4 percent in 2020, resulting in a cumulative loss of more than US$10 trillion over 2020-21.
- Key mechanisms of corporate stress:
  - Massive disruptions in supply networks and a collapse in private-sector demand jeopardize corporate profitability and deplete firms’ cash buffers.
  - Policy responses (extensive fiscal stimulus, lowering interest rates, relaxing macroprudential regulations) implemented, but uncertainty has depressed risk appetite and pushed borrowing costs higher.
  - Systemic risk amplified by record-breaking debt levels among nonfinancial firms.
- Dataset and coverage:
  - More than 537,000 nonfinancial firms from 14 emerging market economies over 1998–2018.
  - Total of 2,152,671 firm-year observations.
  - Orbis firm-level financial data; values in nominal US dollars.
  - Country selection threshold: 10,000 annual observations per country.
  - Data cleaning: drop observations with missing or negative total assets, tangible fixed assets, employment, operating revenue, sales, short-term loans, long-term debt, or where total assets ≠ total liabilities and equity; winsorized at the 1st and 99th percentiles.
  - Geographic concentration: Asia accounts for 95 percent of nonfinancial firms in the sample.
- Dependent variables and controls:
  - Sales growth: rate of change in sales.
  - Profitability: ratio of earnings before interest and tax to total assets.
  - Net fixed investment: difference between tangible assets in current and previous period scaled by total assets at end of previous year.
  - Firm controls: firm age (log of years since establishment), firm size (log of total assets), leverage (short- and long-term debt over total assets), cash flow (ratio of cash flow to total assets), asset tangibility (tangible fixed assets to total assets).
  - Pandemic measure: number of confirmed infectious-disease cases (including Ebola, malaria, SARS, yellow fever) from EM-DAT and WHO, scaled by population.
  - Country controls: real GDP per capita, real GDP growth, trade openness (exports + imports as share of GDP), financial development (domestic credit to private sector as share of GDP), health spending, institutional quality from IMF WEO and World Bank WDI.
  - Stationarity: Im-Pesaran-Shin (2003) panel unit root tests indicate variables are stationary after logarithmic transformation or first differencing.

### II. Empirical methodology
- Core specification (notation preserved):
  - y_{i s c t} = α1 PANDEMIC_{c t} + α2 FIRM_{i s c t-1} + α3 MACRO_{c t-1} + η_i + η_{s t} + η_{s c} + ε_{i s c t}
  - PANDEMIC is number of confirmed infectious disease cases scaled by population in country c at time t.
  - Firm vector includes size, leverage, cash flow, asset tangibility, age; Macro includes real GDP per capita, real GDP growth, trade openness, financial development.
  - Fixed effects: η_i (firm), η_{s t} (sector-year), η_{s c} (sector-country).
  - Robust standard errors clustered at firm level.
- Estimation approaches:
  - OLS baseline.
  - System GMM (Arellano and Bover, 1995; Blundell and Bond, 1998) for dynamic models; instrument proliferation mitigated following Roodman (2009).
  - Validation via AR(1), AR(2) tests and Hansen J-test: high first-order autocorrelation expected; no significant second-order autocorrelation; Hansen J-test supports validity of internal instruments.
- Firm survival analysis:
  - Failure defined as receivership, liquidation, or dissolved.
  - Survival analysis accounts for left truncation and right censoring; first appearance in dataset used as time at risk; firms that drop out without failure excluded.
  - Proportional hazard specification: θ_{i t} = θ(t)·X'_{i t} β.
  - Discrete-time complementary log-log model: h_i(X_{i t}) = 1 − exp{−exp(X'_{i t} β + θ(t))}
  - Extension with unobserved heterogeneity: h_i(X_{i t}) = 1 − exp{−exp(X'_{i t} β + θ(t)) + μ_i}, μ_i is firm-specific random effect.
  - Controls in survival models: age, size, profitability, leverage, cash flow, asset tangibility; sector and country fixed effects; robust standard errors clustered at firm level.

### III. Main estimation results
- Baseline OLS pandemic effects (Table 2; robust standard errors clustered at firm level in brackets):
  - Pandemics:
    - Sales growth: -0.598*** [0.043]
    - Profitability: -0.083*** [0.013]
    - Fixed investment: -0.244*** [0.025]
  - Sample sizes and fit:
    - Number of observations: Sales growth 1,551,353; Profitability 1,543,485; Fixed investment 712,724
    - Number of firms: Sales growth 531,252; Profitability 529,054; Fixed investment 353,724
    - Adj. R-squared: Sales growth 0.053; Profitability 0.008; Fixed investment 0.088
  - Interpretation: past pandemics (confirmed infectious-disease cases scaled by population) have economically and statistically significant negative effects on sales growth, profitability, and fixed investment, with the largest adverse impact on sales growth.
- Firm-level controls (selected coefficients, lagged unless noted; Table 2):
  - Leverage: Sales growth 0.210*** [0.007]; Profitability 0.005 [0.004]; Fixed investment -0.179*** [0.008]
  - Cash flow: Sales growth -0.023** [0.009]; Profitability -0.013* [0.007]; Fixed investment -0.020*** [0.004]
  - Asset tangibility: Sales growth 0.063*** [0.008]; Profitability -0.035*** [0.005]; Fixed investment -0.758*** [0.011]
  - Firm size: Sales growth -0.103*** [0.000]; Profitability -0.011*** [0.000]; Fixed investment -0.013*** [0.001]
  - Firm age: Sales growth -0.097*** [0.007]; Profitability -0.036*** [0.002]; Fixed investment -0.044*** [0.006]
- Macroeconomic controls (lagged; Table 2):
  - GDP per capita: Sales growth -0.162*** [0.033]; Profitability -0.052*** [0.011]; Fixed investment 0.106*** [0.022]
  - GDP growth: Sales growth -1.383*** [0.047]; Profitability 0.018 [0.021]; Fixed investment 0.241*** [0.036]
  - Trade openness: Sales growth -0.606*** [0.007]; Profitability -0.057*** [0.003]; Fixed investment -0.093*** [0.005]
  - Financial development: Sales growth -0.376*** [0.010]; Profitability 0.011*** [0.004]; Fixed investment -0.112*** [0.008]

### IV. Heterogeneity, dynamics, and survival
- Size and age heterogeneity (Appendix Tables A4–A6; coefficients and SEs preserved):
  - Sales growth pandemic coefficients:
    - Small firms: -1.22 [0.174]
    - Large firms: -0.35 [0.063]
    - Young firms: -0.84 [0.222]
    - Old firms: -0.39 [0.064]
  - Profitability pandemic coefficients:
    - Small firms: -0.163** [0.137]
    - Large firms: -0.041** [0.019]
    - Young firms: -0.101** [0.079]
    - Old firms: -0.018** [0.020]
  - Conclusion: infectious-disease outbreaks have a significantly greater negative effect on smaller and younger firms across outcomes.
- Dynamics and persistence:
  - System GMM (dynamic model; Appendix Table A7) confirms negative impact of pandemics; for sales growth the magnitude of pandemics is at least twice as large as in the static model.
- Firm survival (proportional Cox hazard model; Table 3):
  - Pandemics (hazard coefficients; robust standard errors clustered at firm level in brackets):
    - All firms: 1.450*** [0.391]
    - Post GFC: 3.092*** [0.396]
    - Young firms: 3.161*** [0.699]
    - Old firms: 0.054 [0.770]
    - Small firms: 4.108*** [0.972]
    - Large firms: 0.257 [0.540]
  - Interpretation: a percent increase in the number of infectious-disease cases per population is associated with an increase of 1.5 percent in the likelihood of failure among nonfinancial firms in developing countries.
  - Sample and failure counts:
    - Number of observations: All 1,547,234; Post GFC 1,223,839; Young 366,737; Old 453,358; Small 342,673; Large 418,269
    - Number of firms: All 529,912; Post GFC 484,796; Young 204,326; Old 142,247; Small 176,973; Large 109,642
    - Number of failures: All 55,279; Post GFC 29,673; Young 16,196; Old 13,494; Small 16,565; Large 9,009

### V. COVID-19 scaling and quantified projections
- Scaling headline estimates to COVID-19 prevalence in sample countries (textual scaling):
  - Negative impact increases by 17.5 times to:
    - Over 10 percent on sales growth
    - Profitability down by 1.5 percent
    - Fixed investment down by 4.3 percent
  - Estimated increase in probability of corporate failure during the first year of COVID-19:
    - More than 25 percentage points for nonfinancial firms overall
    - 53 percentage points for young firms
    - 70 percentage points for small companies
- Caveats to scaling:
  - Estimates may be an upper bound because projected economic growth in developing countries is expected to decline by 7 percentage points in 2020 relative to 2019, while past pandemics lowered growth by 0.6 percentage points on average.
  - Estimates may understate some effects because the COVID-19 contraction is much deeper than past pandemics.
  - Exceptional policy measures implemented in many countries could mitigate some impacts.

### VI. Robustness checks (representative results)
- Robustness exercises include adding liquidity and capital intensity at firm level; adding rule of law at country level; restricting sample to post-global financial crisis (2010–2018); subsamples by firm size and age; dynamic System GMM.
- Representative robustness (Appendix Table A4; sales growth and Appendix Table A6 for fixed investment):
  - Sales growth pandemic coefficients across specifications:
    - Additional controls: -0.372*** [0.050]
    - Post-GFC period: -0.630*** [0.057]
    - Small firms: -1.221*** [0.174]
    - Large firms: -0.350*** [0.063]
    - Young firms: -0.841*** [0.222]
    - Old firms: -0.387*** [0.064]
  - Fixed investment pandemic coefficients (Appendix Table A6):
    - (1) Additional controls: -0.083*** [0.028]
    - (2) Post-GFC period: -0.255*** [0.032]
    - (3) Small firms: -0.218** [0.099]
    - (4) Large firms: -0.015* [0.034]
    - (5) Young firms: -0.166* [0.101]
    - (6) Old firms: -0.064* [0.034]
  - Appendix Table A6 sample and fit details:
    - Number of observations: (1) 503,723; (2) 374,951; (3) 54,030; (4) 180,479; (5) 93,488; (6) 164,753
    - Number of firms: (1) 149,813; (2) 124,177; (3) 21,234; (4) 45,657; (5) 40,584; (6) 49,387
    - Adj R-squared: (1) 0.088; (2) 0.132; (3) 0.067; (4) 0.091; (5) 0.206; (6) 0.053
  - Robust standard errors clustered at firm level; sector-year and sector-country fixed effects included in all regressions.

### VII. Policy implications and conclusion
- Policy-relevant findings:
  - Pandemics materially and negatively affect sales growth, profitability, fixed investment, and firm survival.
  - Smaller and younger firms and firms in contact-intensive sectors (accommodation, food services, retail trade) and complex production networks (automotive) are disproportionately harmed.
  - Exceptional policy measures during COVID-19 (temporary changes in bankruptcy laws; easing financial conditions; prudential measures to strengthen banks’ lending capacity; reducing wage expenditures; grants and revenue support; debt moratoria and tax deferrals; deferring legal action against insolvent debtors) have shielded many nonfinancial firms but may keep nonviable “zombie” firms alive and impair resource allocation and potential growth.
- Final assessment:
  - The resilience of nonfinancial firms during the coronavirus pandemic depends critically on the magnitude and duration of the economic shock and how losses are shared across stakeholders; prolonged economic uncertainty raises the likelihood of firm failure and long-lasting damage to economic potential.

*Source: wpiea2020276-print-pdf*

### References .............................................................................................................

### References ......................................................................................................................................................................... 22

### I. INTRODUCTION
- COVID-19 described as an unprecedented shock with far-reaching economic and financial repercussions worldwide.
- As of December 10, 2020: over 69.3 million confirmed cases of COVID-19 in 190 countries, with more than 1.5 million deaths.
- IMF projection: global real GDP growth is projected to contract by 4.4 percent in 2020, resulting in a cumulative loss of more than US$10 trillion over 2020-21.
- Key mechanisms of corporate stress highlighted:
  - Massive disruptions in supply networks and a collapse in private-sector demand jeopardize corporate profitability and deplete firms’ cash buffers.
  - Policy responses (extensive fiscal stimulus, lowering interest rates, relaxing macroprudential regulations) have been implemented, but uncertainty has depressed risk appetite and pushed borrowing costs higher.
  - Systemic risk amplified by record-breaking debt levels among nonfinancial firms.
- Literature context:
  - Firm- and sector-specific determinants of corporate performance reviewed (Myers, 1984; Titman and Wessels, 1988; Harris and Raviv, 1991; Booth and others, 2001; Baker and Wurgler, 2002; Lemmon, Roberts, and Zender, 2008; Frank and Goyal, 2009; Gungoraydinoglu and Öztekin, 2011; Graham, Leary, and Roberts, 2015; De Angelo and Roll, 2015; Öztekin, 2015).
  - Macro- and institutional determinants cited (Borio, 1990; Rajan and Zingales, 1995; Kayo and Kimura, 2011; Cevik and Miryugin, 2018).
  - Sparse cross-country firm-level research on infectious diseases (Fornaro and Wolf, 2020; Hassan and others, 2020; Ma, Rogers, and Zhou, 2020).
- Contribution and headline empirical findings:
  - Dataset: more than 537,000 nonfinancial firms from 14 emerging market economies over 1998–2018.
  - Pandemics measured by number of confirmed infectious disease cases scaled by population have an economically and statistically significant negative effect on nonfinancial firm performance across sales growth, corporate profitability, and fixed investment spending.
  - Smaller and younger firms suffer significantly more than larger and more established corporations.
  - Robustness checks including System GMM validate baseline results.
  - Infectious-disease outbreaks increase probability of firm failure in developing countries.
- Scaling to COVID-19 prevalence in sample countries:
  - Negative impact increases by 17.5 times to over 10 percent on sales growth.
  - Profitability declines by 1.5 percent.
  - Fixed investment spending declines by 4.3 percent.
  - Probability of corporate failure increases by more than 25 percentage points among nonfinancial firms during the first year.
- Caveat: these estimates are an upper bound given projected economic growth declines and exceptional policy measures that have mitigated some effects.
- Sectoral vulnerability noted:
  - More contact-intensive sectors (accommodation, food services, retail trade) and firms in complex production networks (automotive) suffer more.
  - Less contact-intensive sectors (information, communication, technology) suffer less.

### II. DATA OVERVIEW
- Data source: Orbis firm-level financial data (balance sheets and income statements).
- Sample size and period:
  - 537,392 nonfinancial firms from 14 developing countries during 1998–2018.
  - Total of 2,152,671 firm-year observations.
- Orbis coverage and preprocessing:
  - Orbis complete sample contains more than 365 million firms from over 100 countries.
  - All values in Orbis reported in nominal US dollars.
  - Country selection threshold: 10,000 annual observations per country.
  - Data cleaning: drop observations with missing or negative total assets, tangible fixed assets, employment, operating revenue, sales, short-term loans, long-term debt, or where total assets ≠ total liabilities and equity.
  - Winsorization at the 1st and 99th percentiles.
- Sectoral distribution (Table 1):
  - Agriculture: 6,760 firms, 1.26 percent
  - Mining: 2,828 firms, 0.53 percent
  - Manufacturing: 158,597 firms, 29.51 percent
  - Utilities: 7,089 firms, 1.32 percent
  - Construction: 67,968 firms, 12.65 percent
  - IT: 21,165 firms, 3.94 percent
  - Other service activities, households, extraterritorial bodies: 13,585 firms, 2.53 percent
  - Wholesale and retail trade, accommodation: 165,844 firms, 30.86 percent
  - Transport and storage: 17,819 firms, 3.32 percent
  - Real estate: 21,232 firms, 3.95 percent
  - Professional and administrative activities: 54,505 firms, 10.14 percent
  - Total: 537,392 firms, 100.0 percent
- Geographic concentration: Asia accounts for 95 percent of nonfinancial firms in the sample.
- Dependent variables:
  - Sales growth: rate of change in sales.
  - Profitability: ratio of earnings before interest and tax to total assets.
  - Net fixed investment: difference between tangible assets in current and previous period scaled by total assets at end of previous year.
- Firm controls:
  - Firm age (log of years since establishment), firm size (log of total assets), leverage (short- and long-term debt over total assets), cash flow (ratio of cash flow to total assets), asset tangibility (tangible fixed assets to total assets).
- Pandemic measure: number of confirmed infectious-disease cases (including Ebola, malaria, SARS, yellow fever) from EM-DAT and WHO, scaled by population.
- Country controls: real GDP per capita, real GDP growth, trade openness (exports + imports as share of GDP), financial development (domestic credit to private sector as share of GDP), health spending, measures of institutional quality from IMF WEO and World Bank WDI.
- Stationarity: Im-Pesaran-Shin (2003) panel unit root tests indicate variables are stationary after logarithmic transformation or first differencing (results available upon request).

### III. EMPIRICAL METHODOLOGY
- Core empirical specification (notation preserved):
  - y_{i s c t} = α1 PANDEMIC_{c t} + α2 FIRM_{i s c t-1} + α3 MACRO_{c t-1} + η_i + η_{s t} + η_{s c} + ε_{i s c t}
  - Subscripts: i = firm, s = sector, c = country, t = time.
  - PANDEMIC is number of confirmed infectious disease cases scaled by population in country c at time t.
  - Firm vector includes size, leverage, cash flow, asset tangibility, age.
  - Macro includes real GDP per capita, real GDP growth, trade openness, financial development.
  - η_i: firm fixed effects; η_{s t}: sector-year fixed effects; η_{s c}: sector-country fixed effects.
  - Robust standard errors clustered at firm level.
- Estimation approaches:
  - Ordinary least squares (OLS) baseline.
  - System GMM (Arellano and Bover, 1995; Blundell and Bond, 1998) for dynamic models to address lagged dependent variables and potential endogeneity.
  - Instrument proliferation issues mitigated following Roodman (2009); validation via AR(1), AR(2) tests (p-values for first- and second-order autocorrelation) and Hansen J-test for overidentifying restrictions.
  - Findings: high first-order autocorrelation expected; no significant second-order autocorrelation; Hansen J-test supports validity of internal instruments.
- Firm survival analysis:
  - Failure defined as receivership, liquidation, or dissolved.
  - Survival analysis accounts for left truncation and right censoring; first appearance in dataset used as time at risk; firms that drop out without failure excluded.
  - Proportional hazard specification: θ_{i t} = θ(t)·X'_{i t} β.
  - Discrete-time complementary log-log model for hazard:
    - h_i(X_{i t}) = 1 − exp{−exp(X'_{i t} β + θ(t))}
  - Extension with unobserved heterogeneity:
    - h_i(X_{i t}) = 1 − exp{−exp(X'_{i t} β + θ(t)) + μ_i}, where μ_i is firm-specific random effect.
  - Controls: age, size, profitability, leverage, cash flow, asset tangibility; sector and country fixed effects; robust standard errors clustered at firm level.

### IV. ESTIMATION RESULTS
- Dataset strengths reiterated:
  - More than 537,000 nonfinancial firms from 14 emerging market economies; 2,152,671 firm-year observations over 1998–2018.
- Main empirical results summary:
  - Consistent negative effect of pandemics (infectious-disease cases scaled by population) on firm performance across sales growth, profitability, and fixed investment.
  - Smaller and younger firms more adversely affected than larger and more established firms.
  - Robustness: results hold across alternative specifications and System GMM dynamic estimations.
  - Pandemic exposure increases firm failure probability in developing countries.
- COVID-19 scaling implications (reiterated):
  - Estimated negative impact scaled to COVID-19 prevalence increases by 17.5 times to over 10 percent on sales growth.
  - Profitability down by 1.5 percent.
  - Fixed investment down by 4.3 percent.
  - Probability of corporate failure increases by more than 25 percentage points during the first year.
- Policy context noted:
  - Exceptional policy measures (temporary changes in bankruptcy laws, easing financial conditions, maintaining access to credit, wage and cost reductions, grants, debt moratoria, tax deferrals, deferring legal action) have mitigated some liquidity pressures and insolvency risks.
  - Sectoral heterogeneity in impact emphasized (contact-intensive and complex production-network sectors worse affected).

*Source: wpiea2020276-print-pdf - References .............................................................................................................*

### introduction of country-specific macroeconomic control variables does not materially alter the

### introduction of country-specific macroeconomic control variables does not materially alter the magnitude and statistical significance of estimated coefficients on firm-level variables

### Baseline estimations: pandemics and firm performance
- Data and scope:
  - Sample: more than 537,000 nonfinancial firms from 14 emerging market economies over the period 1998–2018.
- Main baseline coefficients (Table 2; robust standard errors clustered at firm level in brackets):
  - Pandemics:
    - Sales growth: -0.598*** [0.043]
    - Profitability: -0.083*** [0.013]
    - Fixed investment: -0.244*** [0.025]
  - Firm-level controls (lagged unless noted):
    - Leverage: Sales growth 0.210*** [0.007]; Profitability 0.005 [0.004]; Fixed investment -0.179*** [0.008]
    - Cash flow: Sales growth -0.023** [0.009]; Profitability -0.013* [0.007]; Fixed investment -0.020*** [0.004]
    - Asset tangibility: Sales growth 0.063*** [0.008]; Profitability -0.035*** [0.005]; Fixed investment -0.758*** [0.011]
    - Firm size: Sales growth -0.103*** [0.000]; Profitability -0.011*** [0.000]; Fixed investment -0.013*** [0.001]
    - Firm age: Sales growth -0.097*** [0.007]; Profitability -0.036*** [0.002]; Fixed investment -0.044*** [0.006]
  - Macroeconomic controls (lagged):
    - GDP per capita: Sales growth -0.162*** [0.033]; Profitability -0.052*** [0.011]; Fixed investment 0.106*** [0.022]
    - GDP growth: Sales growth -1.383*** [0.047]; Profitability 0.018 [0.021]; Fixed investment 0.241*** [0.036]
    - Trade openness: Sales growth -0.606*** [0.007]; Profitability -0.057*** [0.003]; Fixed investment -0.093*** [0.005]
    - Financial development: Sales growth -0.376*** [0.010]; Profitability 0.011*** [0.004]; Fixed investment -0.112*** [0.008]
  - Sample sizes and fit:
    - Number of observations: Sales growth 1,551,353; Profitability 1,543,485; Fixed investment 712,724
    - Number of firms: Sales growth 531,252; Profitability 529,054; Fixed investment 353,724
    - Adj. R-squared: Sales growth 0.053; Profitability 0.008; Fixed investment 0.088
- Interpretation:
  - Past pandemics (measured as number of confirmed infectious-disease cases scaled by population) have economically and statistically significant negative effects on sales growth, profitability, and fixed investment, with the largest adverse impact on sales growth.

### Firm-level heterogeneity and detailed findings
- Size and age heterogeneity (robustness and subsample analysis; Appendix Tables A4–A6):
  - Sales growth pandemic coefficients:
    - Small firms: -1.22 (Appendix Table A4, column for Small firms) [0.174]
    - Large firms: -0.35 [0.063]
    - Young firms: -0.84 [0.222]
    - Old firms: -0.39 [0.064]
  - Profitability pandemic coefficients (Appendix Table A5, robustness subsamples):
    - Small firms: -0.163** [0.137]
    - Large firms: -0.041** [0.019]
    - Young firms: -0.101** [0.079]
    - Old firms: -0.018** [0.020]
  - Conclusion: infectious-disease outbreaks have a significantly greater negative effect on smaller and younger firms across outcomes.
- Dynamics and persistence:
  - System GMM (dynamic model with lagged dependent variable; Appendix Table A7): confirms negative impact of pandemics; for sales growth the magnitude of pandemics is at least twice as large as in the static model.

### Firm survival implications (proportional Cox hazard model)
- Main survival coefficients (Table 3; stratified results; robust standard errors clustered at firm level in brackets):
  - Pandemics (hazard coefficients):
    - All firms: 1.450*** [0.391]
    - Post GFC: 3.092*** [0.396]
    - Young firms: 3.161*** [0.699]
    - Old firms: 0.054 [0.770]
    - Small firms: 4.108*** [0.972]
    - Large firms: 0.257 [0.540]
  - Interpretation in percentage terms (textual result):
    - A percent increase in the number of infectious-disease cases per population is associated with an increase of 1.5 percent in the likelihood of failure among nonfinancial firms in developing countries.
  - Sample sizes and events:
    - Number of observations: All 1,547,234; Post GFC 1,223,839; Young 366,737; Old 453,358; Small 342,673; Large 418,269
    - Number of firms: All 529,912; Post GFC 484,796; Young 204,326; Old 142,247; Small 176,973; Large 109,642
    - Number of failures: All 55,279; Post GFC 29,673; Young 16,196; Old 13,494; Small 16,565; Large 9,009
- Scaled COVID-19 projections (textual scaling using sample countries’ confirmed cases per population):
  - Scaling estimated coefficients to COVID-19 prevalence increases the negative impact by 17.5 times to:
    - Over 10 percent on sales growth
    - 1.5 percent on profitability
    - 4.3 percent on fixed investment spending
  - Estimated increase in probability of corporate failure during the first year of COVID-19:
    - More than 25 percentage points for nonfinancial firms overall
    - 53 percentage points for young firms
    - 70 percentage points for small companies
  - Caveats noted in text:
    - These estimates may be an upper bound because projected economic growth in developing countries is expected to decline by 7 percentage points in 2020 relative to 2019, while past pandemics lowered growth by 0.6 percentage points on average.
    - Estimates may understimate some effects because the COVID-19 contraction is much deeper than past pandemics; exceptional policy measures implemented in many countries could mitigate some impacts.

### Robustness checks and additional controls
- Robustness exercises:
  - Added firm-level variables: liquidity and capital intensity.
  - Added country-level institutional control: rule of law.
  - Restricted sample to post-global financial crisis period (2010–2018).
  - Subsample estimations by firm size (below 25th percentile vs above 75th percentile) and firm age (below 25th percentile vs above 75th percentile).
  - Dynamic specification via System GMM to capture persistence.
- Representative robustness results (Appendix Table A4, Sales growth with additional controls and subsamples):
  - Pandemics coefficients across specifications:
    - Additional controls: -0.372*** [0.050]
    - Post-GFC period: -0.630*** [0.057]
    - Small firms: -1.221*** [0.174]
    - Large firms: -0.350*** [0.063]
    - Young firms: -0.841*** [0.222]
    - Old firms: -0.387*** [0.064]
  - Robustness results confirm the significant negative impact of pandemics on nonfinancial corporate performance.

### Policy implications and overarching conclusion
- Key policy-relevant findings:
  - Pandemics materially and negatively affect sales growth, profitability, fixed investment, and firm survival, disproportionately harming smaller and younger firms and firms in contact-intensive sectors (accommodation, food services, retail trade) and complex production networks (automotive).
  - Exceptional policy measures enacted during COVID-19 (temporary changes in bankruptcy laws; easing financial conditions; prudential measures to strengthen banks’ lending capacity; reducing wage expenditures; grants and revenue support; debt moratoria and tax deferrals; deferring legal action against insolvent debtors) have shielded many nonfinancial firms but may keep nonviable “zombie” firms alive and thus impair resource allocation and potential growth.
- Final assessment:
  - The resilience of nonfinancial firms during the coronavirus pandemic depends critically on the magnitude and duration of the economic shock and how losses are shared across stakeholders; prolonged economic uncertainty raises the likelihood of firm failure and long-lasting damage to economic potential.

*Source: wpiea2020276-print-pdf - introduction of country-specific macroeconomic control variables does not materially alter the magnitude and statistical significance of estimated coefficients on firm-level variables*

### Appendix Table A6. Robustness Checks: Infectious Diseases and Fixed Investment

### Appendix Table A6. Robustness Checks: Infectious Diseases and Fixed Investment

### Key findings
- The estimated effect of Pandemics on fixed investment is negative and statistically significant across multiple robustness checks and subsamples:
  - Column (1) Additional controls: Pandemics = -0.083*** [0.028]
  - Column (2) Post-GFC period: Pandemics = -0.255*** [0.032]
  - Column (3) Small firms: Pandemics = -0.218** [0.099]
  - Column (4) Large firms: Pandemics = -0.015* [0.034]
  - Column (5) Young firms: Pandemics = -0.166* [0.101]
  - Column (6) Old firms: Pandemics = -0.064* [0.034]

### Firm-level controls (coefficients and reported standard errors)
- Leverage (lag)
  - (1) -0.175*** [0.008]
  - (2) -0.162*** [0.010]
  - (3) -0.051*** [0.019]
  - (4) -0.121*** [0.012]
  - (5) -0.290*** [0.026]
  - (6) -0.115*** [0.013]
- Cash flow (lag)
  - (1) -0.020*** [0.004]
  - (2) -0.000 [0.004]
  - (3) 0.005 [0.004]
  - (4) -0.028*** [0.008]
  - (5) -0.030*** [0.008]
  - (6) -0.037** [0.015]
- Tangibility (lag)
  - (1) -0.753*** [0.011]
  - (2) -0.863*** [0.014]
  - (3) -0.613*** [0.043]
  - (4) -0.608*** [0.017]
  - (5) -1.569*** [0.039]
  - (6) -0.570*** [0.020]
- Size (lag)
  - (1) -0.018*** [0.001]
  - (2) -0.126*** [0.004]
  - (3) -0.146*** [0.010]
  - (4) -0.022*** [0.001]
  - (5) -0.091*** [0.004]
  - (6) -0.015*** [0.001]
- Age
  - (1) -0.040*** [0.006]
  - (2) 0.048*** [0.009]
  - (3) 0.126*** [0.024]
  - (4) -0.126*** [0.008]
  - (5) -0.073 [0.072]
  - (6) -0.065** [0.028]
- Liquidity (lag)
  - (1) 0.000 [0.000]
- Capital intensity (lag)
  - (1) -0.003*** [0.000]

### Macroeconomic controls (coefficients and reported standard errors)
- GDP per capita (lag)
  - (1) 0.028 [0.023]
  - (2) 1.343*** [0.071]
  - (3) 0.038 [0.124]
  - (4) 0.106*** [0.021]
  - (5) 0.060 [0.204]
  - (6) 0.151*** [0.044]
- GDP growth (lag)
  - (1) -0.079 [0.078]
  - (2) -0.568*** [0.085]
  - (3) 0.114 [0.146]
  - (4) 0.032 [0.084]
  - (5) -0.020 [0.241]
  - (6) 0.057 [0.093]
- Trade openness (lag)
  - (1) -0.072*** [0.011]
  - (2) -0.088*** [0.023]
  - (3) 0.063* [0.037]
  - (4) -0.114*** [0.014]
  - (5) -0.354*** [0.051]
  - (6) -0.061*** [0.015]
- Financial development (lag)
  - (1) 0.007 [0.012]
  - (2) -0.186*** [0.018]
  - (3) -0.072 [0.046]
  - (4) 0.060*** [0.014]
  - (5) 0.013 [0.064]
  - (6) -0.032** [0.016]
- Rule of law (lag)
  - (1) -0.017*** [0.006]

### Sample, fit, and estimation details
- Number of observations:
  - (1) 503,723
  - (2) 374,951
  - (3) 54,030
  - (4) 180,479
  - (5) 93,488
  - (6) 164,753
- Number of firms:
  - (1) 149,813
  - (2) 124,177
  - (3) 21,234
  - (4) 45,657
  - (5) 40,584
  - (6) 49,387
- Adj R-squared:
  - (1) 0.088
  - (2) 0.132
  - (3) 0.067
  - (4) 0.091
  - (5) 0.206
  - (6) 0.053
- Estimation notes:
  - Robust standard errors clustered at the firm level are reported in brackets.
  - Fixed effects included in all regressions are sector-year and sector-country.

*Appendix Table A6. Robustness Checks: Infectious Diseases and Fixed Investment — content as provided*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2020/english/wpiea2020276-print-pdf.pdf_