## wpiea2020287-print-pdf

## Source details

**Canonical URL:** [wpiea2020287-print-pdf](https://www.imf.org/-/media/files/publications/wp/2020/english/wpiea2020287-print-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2020/english/wpiea2020287-print-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2020/english/wpiea2020287-print-pdf.pdf.json)

---

### I. Introduction — key context and motivations
- The COVID-19 pandemic is described as "an unprecedented global shock" with far-reaching repercussions.
- IMF projection: global real GDP growth is projected to contract by 4.4 percent in 2020.
- Projected cumulative loss: "more than US$10 trillion over 2020-21."
- Pandemic consequences: disruptions in supply networks, collapse in private-sector demand, jeopardized corporate profitability, depleted firms’ cash buffers, higher borrowing costs amid elevated uncertainty.
- Corporate leverage context:
  - "nonfinancial corporate leverage has remained high at an average of about 90 percent of GDP in advanced economies."
  - "it surged in developing countries from 56 percent of GDP in 2008 to 96 percent in 2018."
- Empirical contribution:
  - Two-pronged firm-level analysis using "more than 1.8 million nonfinancial firms from 52 countries over the period 1997–2018."
  - First: determinants of corporate leverage (total debt-to-assets ratio), accounting for country-specific characteristics.
  - Second: effect of corporate debt overhang on fixed investment.
- Main empirical findings (summary):
  - "Nonfinancial corporate leverage is highly vulnerable to disruptions in profitability and cash flow at the firm level and economic growth at the aggregate level."
  - "Highly leveraged companies tend to have lower levels of fixed investment."
  - Example: "a manufacturing firm with a leverage ratio of 100 percent invests about 4 percentage points less than a company with a debt-to-asset ratio of 50 percent in the same sector."
- Policy implication: "corporate debt overhang could become a strenuous burden on nonfinancial firms during severe economic downturns," amplified if the COVID-19 pandemic lingers.

### II. Data overview — sample construction and variables
- Core dataset:
  - "harmonized firm-level financial data, including balance sheets and income statements, on 1,867,227 nonfinancial firms in 52 countries during the period 1997–2018."
  - Unbalanced panel: 1,867,227 unique firms from 52 countries (30 advanced and 22 developing) with a total of 10,426,274 firm-year observations during 1997–2018.
- Data source and units:
  - Firm data drawn from Orbis; "All values reported in the Orbis database are in nominal US dollars."
- Cleaning and trimming procedures:
  - Country selection threshold: "10,000 annual observations per country."
  - Drop observations where key items are missing/negative or where total assets ≠ total liabilities and equity.
  - Winsorization at the "1th and 99th percentile of the distribution."
  - Robustness: results remain robust if winsorizing 5 percent on both tails.
- Sectoral distribution (Num. of obs. and Percent):
  - Agriculture: 377,802 2.60%
  - Mining: 56,536 0.39%
  - Manufacturing: 3,102,690 21.35%
  - Utilities: 234,161 1.61%
  - Construction: 2,028,255 13.96%
  - IT: 439,603 3.02%
  - Other service activities, households, extraterritorial bodies: 766,552 5.27%
  - Wholesale and retail trade, accommodation: 4,330,233 29.79%
  - Transport and storage: 649,077 4.47%
  - Real estate: 1,112,347 7.65%
  - Professional and administrative activities: 1,436,389 9.88%
  - Total: 14,533,645 100%
- Key firm-level measures:
  - Corporate leverage: "short-term and long-term debt over total assets."
  - Fixed investment: "difference between tangible assets in the current period and those in the previous period scaled by total assets at the end of the previous year."
  - Firm controls: firm age (log), firm size (log total assets), profitability (profit before tax / total assets), cash flow (cash flow / total assets), sales growth, asset tangibility, effective interest rate.
- Macroeconomic controls:
  - Real GDP per capita, real GDP growth, trade openness, financial development (domestic credit to private sector / GDP), and institutional quality measures (used in robustness checks).

### III. Empirical methodology — model specifications and identification
- First-stage model (determinants of corporate leverage):
  - Dependent variable: Lev = total debt / total assets.
  - Firm vector (Firm): total assets, profitability, cash flow, sales growth, asset tangibility, effective interest rate, firm age.
  - Macro vector (Macro): real GDP per capita, real GDP growth, trade openness, financial development.
  - Fixed effects: firm fixed effects (η_i), sector-year fixed effects (η_si_t), country-sector fixed effects (η_csi), and time effects.
  - Robust standard errors clustered at the firm level.
- Second-stage model (leverage → fixed investment):
  - Dependent variable: Inv = net fixed investment / total assets at beginning of year.
  - Main explanatory variable: Lev (total debt / total assets) with same Firm and Macro controls.
  - Interpretation: sector-year fixed effects allow leverage coefficient to be read as effect "relative to a firm’s sector peers at time t."
- Estimation notes:
  - All specifications include firm, sector, country and time fixed effects.
  - Panel unit root testing: Im-Pesaran-Shin (2003) — variables stationary after log transformation or first differencing.
  - Robust standard errors clustered at the firm level.

### IV. Estimation results — determinants of corporate leverage (selected coefficients and sample details)
- Firm-level coefficients (leverage = total debt / total assets):
  - Profitability (lag): -0.140***, -0.141***, -0.142***, -0.142***, -0.121***, -0.122***.
  - Cash flow (lag): -0.003***, -0.006***, -0.003***, -0.005***, -0.012***, -0.012***.
  - Sales growth (lag): -0.000***, -0.001***, -0.001***, -0.001***, 0.002***, 0.002***.
  - Tangibility (lag): 0.107***, 0.105***, 0.108***, 0.107***, 0.075***, 0.074***.
  - Total assets (lag): 0.044***, 0.043***, 0.045***, 0.044***, 0.028***, 0.029***.
  - Interest rate (lag): -0.003***, -0.004***, -0.003***, -0.004***, -0.000, -0.001**.
  - Age: -0.071***, -0.065***, -0.072***, -0.065***, -0.046***, -0.050***.
- Macroeconomic coefficients:
  - Real GDP per capita: 0.013*** and 0.015*** (pooled/all); -0.066*** (emerging market subsample).
  - Real GDP growth: -0.340***, -0.372***, -0.170***.
  - Trade openness: -0.035***, -0.042***, -0.054***.
  - Financial development: -0.040***, -0.046***, 0.047***.
- Sample and model details (Table 2 snapshots):
  - Number of observations: 10,425,382; 10,426,274; 9,769,226; 9,769,639; 656,156; 656,635.
  - Number of firms: 1,867,105; 1,867,227; 1,728,796; 1,728,859; 138,309; 138,368.
  - Fixed effects: S*T*C, S*T+S*C, S*T*C, S*T+S*C, S*T*C, S*T+S*C.
  - Adj R-squared: 0.028, 0.029, 0.029, 0.030, 0.015, 0.022.
- Key interpretations:
  - Profitability and cash flow reduce corporate leverage across country groups; profitability impact marginally greater in developing countries, cash flow impact twice as large in advanced economies.
  - Sales growth generally negatively related to leverage, but positively associated with leverage in emerging market economies.
  - Firm size and asset tangibility positively related to leverage, larger effects in advanced economies.
  - Effective interest rate negatively associated with leverage, more pronounced in advanced economies.
  - Firm age negatively associated with leverage; effect less pronounced in developing countries.
  - Real income per capita positive for leverage overall and in advanced economies, negative among emerging market economies.
  - Real GDP growth negatively associated with firm-level debt accumulation across countries; more pronounced in advanced economies.
  - Trade openness and financial development generally dampen corporate leverage, with heterogeneity across country groups.

### V. Estimation results — determinants of fixed investment (leverage → investment link and firm/macroeconomic controls)
- Leverage (lag) coefficients on fixed investment (Table 3 and robustness tables):
  - -0.045***, -0.053***, -0.045***, -0.053***, -0.042***, -0.040***.
- Interpretation:
  - Firms with higher leverage undertake significantly less fixed investment across country groups and specifications; effect marginally higher in advanced economies.
  - Example from introduction: manufacturing firm with leverage 100 percent invests about 4 percentage points less than a company with debt-to-asset ratio 50 percent in same sector.
- Firm-level determinants of fixed investment (selected coefficients):
  - Profitability (lag): 0.038***, 0.022***, 0.038***, 0.028***, 0.041***, 0.005.
  - Cash flow (lag): 0.000***, -0.000, 0.000***, 0.000, -0.000, -0.000*.
  - Sales growth (lag): -0.000***, -0.000***, -0.000***, -0.000***, 0.000**, 0.000**.
  - Tangibility (lag): -0.362***, -0.378***, -0.360***, -0.378***, -0.405***, -0.396***.
  - Total assets (lag): -0.078***, -0.044***, -0.078***, -0.048***, -0.103***, -0.049***.
  - Interest rate (lag): -0.000***, -0.000***, -0.000***, -0.000***, -0.000*, -0.000.
  - Age: -0.016***, -0.054***, -0.016***, -0.047***, -0.008, -0.025*.
- Macroeconomic determinants of fixed investment:
  - Real GDP per capita: -0.445***, -0.752***, 0.096***.
  - Real GDP growth: -0.820***, -0.661***, -0.873***.
  - Trade openness: -0.321***, -0.460***, 0.170***.
  - Financial development: -0.065***, -0.091***, -0.137***.
- Sample and model details (Table 3 snapshots):
  - Number of observations: 4,278,402; 4,007,046; 3,996,966; 3,759,078; 281,436; 247,966.
  - Number of firms: 1,190,211; 1,145,398; 1,100,452; 1,067,118; 89,759; 78,280.
  - Fixed effects: S*T*C, S*T+S*C, S*T*C, S*T+S*C, S*T*C, S*T+S*C.
  - Adj R-squared: 0.045, 0.033, 0.046, 0.034, 0.044, 0.032.
- Additional interpretations:
  - Profitability positively affects fixed investment across country groups.
  - Cash flow matters for fixed investment in full sample and advanced economies, but not in developing countries.
  - Sales growth negative effect on investment notably greater in emerging market economies.
  - Firm size and asset tangibility dampen fixed investment.
  - Effective interest rate associated with lower capital spending; not statistically significant in emerging market economies.
  - Older firms invest more than younger corporations, especially in developing countries.

### VI. Sectoral results (Appendix Table A4 — fixed investment by sector)
- Leverage (lag) coefficients by sector (selected):
  - AGR: -0.068*** [0.012]
  - MIN: -0.062** [0.027]
  - MFG: -0.083*** [0.005]
  - UTI: -0.094*** [0.021]
  - CON: 0.006 [0.007]
  - IT: -0.036** [0.015]
  - OTH: -0.095*** [0.013]
  - TRD: -0.040*** [0.005]
  - TRA: -0.088*** [0.011]
  - EST: -0.015* [0.008]
  - ADM: -0.073*** [0.009]
- Tangibility (lag): strongly negative across sectors (examples: AGR -0.411*** [0.019]; MFG -0.468*** [0.006]; IT -0.424*** [0.025]).
- Total assets (lag): negative across sectors (examples: AGR -0.031*** [0.002]; CON -0.054*** [0.001]; IT -0.077*** [0.003]).
- Interest rate (lag): generally negative and small across sectors.
- Sample sizes by sector (observations / firms examples):
  - MFG: 975,832 observations; 254,733 firms.
  - TRD: 1,147,000 observations; 329,775 firms.
  - AGR: 110,734 observations; 31,557 firms.
- Adj R-squared range across sectors: 0.015 (MIN) to 0.050 (IT).
- Notes: Robust standard errors clustered at firm level; firm fixed effects included; significance: *** p<0.01, ** p<0.05, * p<0.1.

### VII. Robustness checks and sub-sample analyses (Appendix Tables A5–A6)
- Additional controls (Appendix Table A5):
  - Leverage (lag) remains negative and significant across specifications:
    - All (col 1): -0.043*** [0.003]
    - All (col 2): -0.054*** [0.003]
    - Advanced (col 3): -0.043*** [0.003]
    - Advanced (col 4): -0.053*** [0.003]
    - Developing (col 5): -0.043*** [0.013]
    - Developing (col 6): -0.068*** [0.017]
  - GDP per capita (All): -0.586*** [0.010]; GDP growth (All): -0.882*** [0.045]; Trade openness (All): -0.337*** [0.008]; Financial development (All): -0.073*** [0.003]; Rule of law (All): 0.015*** [0.002].
  - Number of observations (examples): Col (1): 3,721,468; Col (2): 3,440,685.
- Sub-sample results (Appendix Table A6):
  - Leverage (lag) by sub-sample:
    - Pre-GFC: -0.083*** [0.004]
    - Post-GFC: -0.017*** [0.006]
    - Small: -0.125*** [0.005]
    - Large: -0.012*** [0.004]
    - Low leverage: 0.050*** [0.010]
    - High leverage: -0.250*** [0.006]
  - Tangibility (lag) larger negative effects post-GFC and for high-leverage firms (examples: Post-GFC -0.581*** [0.009]; High leverage -0.756*** [0.009]).
  - Macroeconomic coefficients vary substantially across sub-samples (examples: Real GDP per capita — Large: 0.921*** [0.016]; High leverage: -1.032*** [0.023]).
  - Sample sizes (examples): Pre-GFC observations 883,991; Post-GFC 1,257,531; Small 1,375,768; Large 2,313,731.
- Robustness summary:
  - Introducing liquidity, capital intensity, and institutional controls (corruption, rule of law) confirms baseline findings.
  - Estimates for subperiods, firm-size splits, and leverage-percentile splits remain broadly consistent with baseline.

### VIII. Conclusion — key messages and policy-relevant implications
- Context:
  - The global economy faces an unprecedented slump from the coronavirus pandemic amid record corporate debt levels.
  - Protracted downturn risks a vicious cycle exacerbating corporate vulnerabilities, macro-financial instability, and long-lasting damage to economic potential.
- Empirical summary:
  - Firm-level balance sheet analysis covering more than 1.8 million nonfinancial firms from 52 countries over 1997–2018.
  - Corporate leverage is highly vulnerable to disruptions in profitability and cash flow at the firm level and to economic growth at the aggregate level.
  - Corporate debt overhang can become a strenuous burden on nonfinancial firms during severe downturns.
  - Highly leveraged firms tend to have lower levels of fixed investment spending.
- Policy implication:
  - "In times of greater macro-financial uncertainty, corporate debt overhang is likely to depress nonfinancial capital investment for years, especially if the COVID-19 pandemic lingers and the global downturn becomes protracted."

### IX. Key statistics and dataset snapshots
- Total unique firms: 1,867,227.
- Firm-year observations (selected totals):
  - Dataset total referenced: 10,426,274 firm-year observations during 1997–2018.
  - Appendix yearly total: 14,533,645 observations (yearly breakdown and country-group aggregates provided).
- Summary statistics (selected):
  - Leverage: Min 0.00; p25 0.05; p50 0.20; p75 0.42; Max 2.34; Average 0.27; Std. dev. 0.27; Num. of obs. 14,533,645.
  - Fixed investment: Min 0.00; p25 0.01; p50 0.03; p75 0.09; Max 5.13; Average 0.10; Std. dev. 0.24; Num. of obs. 6,062,631.
  - Total assets (log): Min 0.00; p25 12.77; p50 13.86; p75 15.08; Max 32.51; Average 14.00; Std. dev. 1.84; Num. of obs. 14,533,645.
  - Profitability: Min -3.21; p25 -0.01; p50 0.02; p75 0.07; Max 1.22; Average 0.02; Std. dev. 0.19; Num. of obs. 14,453,593.
  - GDP per capita (log): Min 6.80; p25 8.98; p50 9.97; p75 10.70; Max 11.63; Average 9.75; Std. dev. 1.08; Num. of obs. 951.

*Source: wpiea2020287-print-pdf*

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

### References

### I. Introduction — key context and motivations
- The COVID-19 pandemic is described as "an unprecedented global shock" with far-reaching repercussions.
- IMF projection: global real GDP growth is projected to contract by 4.4 percent in 2020.
- Projected cumulative loss: "more than US$10 trillion over 2020-21."
- Pandemic consequences noted: disruptions in supply networks, collapse in private-sector demand, jeopardized corporate profitability, depleted firms’ cash buffers, higher borrowing costs amid elevated uncertainty.
- Corporate leverage context:
  - "nonfinancial corporate leverage has remained high at an average of about 90 percent of GDP in advanced economies."
  - "it surged in developing countries from 56 percent of GDP in 2008 to 96 percent in 2018."
- Empirical contribution:
  - Two-pronged firm-level analysis using a large dataset of "more than 1.8 million nonfinancial firms from 52 countries over the period 1997–2018."
  - First: determinants of corporate leverage (total debt-to-assets ratio), accounting for country-specific characteristics.
  - Second: effect of corporate debt overhang on fixed investment.
- Main empirical findings (summary):
  - "Nonfinancial corporate leverage is highly vulnerable to disruptions in profitability and cash flow at the firm level and economic growth at the aggregate level."
  - "Highly leveraged companies tend to have lower levels of fixed investment."
  - Example quantitative finding: "a manufacturing firm with a leverage ratio of 100 percent invests about 4 percentage points less than a company with a debt-to-asset ratio of 50 percent in the same sector."
- Policy implication highlighted: "corporate debt overhang could become a strenuous burden on nonfinancial firms during severe economic downturns," with amplified effects if the COVID-19 pandemic lingers.

### II. Data overview — sample construction and variables
- Core dataset:
  - "harmonized firm-level financial data, including balance sheets and income statements, on 1,867,227 nonfinancial firms in 52 countries during the period 1997–2018."
  - After cleaning and processing: "an unbalanced panel of 1,867,227 unique firms from 52 countries (30 advanced and 22 developing) with a total of 10,426,274 firm-year observations during the period 1997–2018."
- Data source and units:
  - Firm data drawn from Orbis; "All values reported in the Orbis database are in nominal US dollars."
- Cleaning and trimming procedures:
  - Country selection threshold: "10,000 annual observations per country."
  - Drop observations where "total assets, tangible fixed assets, employment, operating revenue, sales and short-term loans and long-term debt in any given year are missing or negative, and where total assets do not equal to total liabilities and equity."
  - Winsorization: firm-level variables winsorized at the "1th and 99th percentile of the distribution" (note: the text uses "1th and 99th").
  - Robustness note: "The estimation results remain robust if we winsorize 5 percent of observations on both tails of the distribution. These results are available upon request."
- Sectoral distribution (Table 1 excerpts — Num. of obs. and Percent):
  - Agriculture: 377,802 2.60%
  - Mining: 56,536 0.39%
  - Manufacturing: 3,102,690 21.35%
  - Utilities: 234,161 1.61%
  - Construction: 2,028,255 13.96%
  - IT: 439,603 3.02%
  - Other service activities, households, extraterritorial bodies: 766,552 5.27%
  - Wholesale and retail trade, accommodation: 4,330,233 29.79%
  - Transport and storage: 649,077 4.47%
  - Real estate: 1,112,347 7.65%
  - Professional and administrative activities: 1,436,389 9.88%
  - Total: 14,533,645 100%
- Dependent variables and key firm-level measures:
  - Corporate leverage: "short-term and long-term debt over total assets."
  - Fixed investment: "difference between tangible assets in the current period and those in the previous period scaled by total assets at the end of the previous year."
  - Firm controls included: firm age (log of years since establishment), firm size (log of total assets), profitability (profit before tax to total assets), cash flow (cash flow to total assets), sales growth (rate of change in sales), asset tangibility (tangible fixed assets to total assets), effective interest rate (interest payment in the current period to total debt at the end of the previous year).
- Macroeconomic controls:
  - Real GDP per capita, real GDP growth, trade openness (sum of exports and imports in GDP), financial development (domestic credit to the private sector as a share of GDP), and measures of institutional quality (used in robustness checks).

### III. Empirical methodology — model specifications and identification
- First-stage model (determinants of corporate leverage):
  - Dependent variable: Lev = ratio of total debt to total assets.
  - Firm vector (Firm): total assets, profitability, cash flow, sales growth, asset tangibility, effective interest rate, firm age.
  - Macro vector (Macro): real GDP per capita, real GDP growth, trade openness, financial development.
  - Fixed effects included: firm fixed effects (η_i), sector-year fixed effects (η_si_t), country-sector fixed effects (η_csi), and time effects; robust standard errors clustered at the firm level.
- Second-stage model (link between leverage and fixed investment):
  - Dependent variable: Inv = ratio of net fixed investment in a given year to total assets at the beginning of the year.
  - Main explanatory variable: Lev (total debt to total assets), including same Firm and Macro controls as above.
  - Interpretation: inclusion of sector-year fixed effects allows leverage coefficient to be interpreted as effect of indebtedness "relative to a firm’s sector peers at time t."
- Estimation notes:
  - All specifications include firm, sector, country and time fixed effects.
  - Panel unit root testing: Im-Pesaran-Shin (2003) procedure applied; "results... indicate that the variables used in the analysis are stationary after logarithmic transformation or upon first differencing."
  - Robust standard errors clustered at the firm level.

### IV. Estimation results — main empirical findings (high-level)
- Dataset scale reiterated: "more than 1.8 million nonfinancial firms from 52 countries... with a total of 10.4 million firm-year observations."
- Consistency: "Estimation results present a consistent picture across different specifications and econometric methodologies."
- Determinants of leverage:
  - Corporate leverage "is highly vulnerable to disruptions in profitability and cash flow at the firm level and economic growth at the aggregate level."
  - Country-specific macroeconomic and institutional factors are important determinants of leverage.
- Leverage and investment link:
  - "Highly leveraged companies tend to have lower levels of fixed investment."
  - Quantitative example: "a manufacturing firm with a leverage ratio of 100 percent invests about 4 percentage points less than a company with a debt-to-asset ratio of 50 percent in the same sector."
- Policy-relevant scenario:
  - "In times of greater macro-financial uncertainty, corporate debt overhang is likely to depress nonfinancial capital investment for years to come, especially if the COVID-19 pandemic lingers and global downturn becomes protracted."

*Source: wpiea2020287-print-pdf*

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

### wpiea2020287-print-pdf - introduction of country-specific macroeconomic control variables does not materially alter the

### A. Determinants of Corporate Leverage
- Firm-level relationships (leverage = total debt / total assets):
  - Profitability (lag): coefficients -0.140***, -0.141***, -0.142***, -0.142***, -0.121***, -0.122*** (robust SEs in brackets as reported).
  - Cash flow (lag): coefficients -0.003***, -0.006***, -0.003***, -0.005***, -0.012***, -0.012***.
  - Sales growth (lag): coefficients -0.000***, -0.001***, -0.001***, -0.001***, 0.002***, 0.002***.
  - Tangibility (lag): coefficients 0.107***, 0.105***, 0.108***, 0.107***, 0.075***, 0.074***.
  - Total assets (lag): coefficients 0.044***, 0.043***, 0.045***, 0.044***, 0.028***, 0.029***.
  - Interest rate (lag): coefficients -0.003***, -0.004***, -0.003***, -0.004***, -0.000, -0.001**.
  - Age: coefficients -0.071***, -0.065***, -0.072***, -0.065***, -0.046***, -0.050***.
- Macroeconomic relationships:
  - Real GDP per capita: coefficients 0.013*** and 0.015*** for pooled/all specifications; coefficient -0.066*** reported for emerging market subsample.
  - Real GDP growth: coefficients -0.340***, -0.372***, -0.170***.
  - Trade openness: coefficients -0.035***, -0.042***, -0.054***.
  - Financial development: coefficients -0.040***, -0.046***, 0.047***.
- Sample and model details (from Table 2):
  - Number of observations: 10,425,382; 10,426,274; 9,769,226; 9,769,639; 656,156; 656,635.
  - Number of firms: 1,867,105; 1,867,227; 1,728,796; 1,728,859; 138,309; 138,368.
  - Fixed effects: S*T*C, S*T+S*C, S*T*C, S*T+S*C, S*T*C, S*T+S*C.
  - Adj R-squared: 0.028, 0.029, 0.029, 0.030, 0.015, 0.022.
- Key interpretations:
  - Profitability and cash flow dampen corporate leverage across country groups; profitability impact marginally greater in developing countries, cash flow impact twice as large in advanced economies.
  - Sales growth generally negatively related to leverage, but positively associated with leverage in emerging market economies.
  - Firm size and asset tangibility positively related to leverage, with larger effects in advanced economies.
  - Effective interest rate has a negative coefficient moderating leverage, more pronounced in advanced economies.
  - Firm age negatively associated with leverage; effect less pronounced among nonfinancial firms in developing countries.
  - Real income per capita has a positive effect on leverage overall and in advanced economies, but negative among emerging market economies.
  - Real GDP growth negatively associated with firm-level debt accumulation across countries; more pronounced in advanced economies.
  - Trade openness and financial development generally dampen corporate leverage, but magnitudes and signs vary across country groups (trade openness stronger negative in emerging markets; financial development shows opposing influences across advanced and developing countries).

### B. Determinants of Fixed Investment
- Main findings on leverage and investment (Table 3):
  - Leverage (lag) coefficients on fixed investment: -0.045***, -0.053***, -0.045***, -0.053***, -0.042***, -0.040***.
  - Interpretation: Firms with higher leverage undertake significantly less fixed investment across all country groups and specifications; effect marginally higher in advanced economies.
- Firm-level determinants of fixed investment (coefficients from Table 3):
  - Profitability (lag): 0.038***, 0.022***, 0.038***, 0.028***, 0.041***, 0.005.
  - Cash flow (lag): 0.000***, -0.000, 0.000***, 0.000, -0.000, -0.000*.
  - Sales growth (lag): -0.000***, -0.000***, -0.000***, -0.000***, 0.000**, 0.000**.
  - Tangibility (lag): -0.362***, -0.378***, -0.360***, -0.378***, -0.405***, -0.396***.
  - Total assets (lag): -0.078***, -0.044***, -0.078***, -0.048***, -0.103***, -0.049***.
  - Interest rate (lag): -0.000***, -0.000***, -0.000***, -0.000***, -0.000*, -0.000.
  - Age: -0.016***, -0.054***, -0.016***, -0.047***, -0.008, -0.025*.
- Macroeconomic determinants of fixed investment (coefficients from Table 3):
  - Real GDP per capita: -0.445***, -0.752***, 0.096***.
  - Real GDP growth: -0.820***, -0.661***, -0.873***.
  - Trade openness: -0.321***, -0.460***, 0.170***.
  - Financial development: -0.065***, -0.091***, -0.137***.
- Sample and model details (from Table 3):
  - Number of observations: 4,278,402; 4,007,046; 3,996,966; 3,759,078; 281,436; 247,966.
  - Number of firms: 1,190,211; 1,145,398; 1,100,452; 1,067,118; 89,759; 78,280.
  - Fixed effects: S*T*C, S*T+S*C, S*T*C, S*T+S*C, S*T*C, S*T+S*C.
  - Adj R-squared: 0.045, 0.033, 0.046, 0.034, 0.044, 0.032.
- Additional empirical findings:
  - Profitability positively affects fixed investment across country groups.
  - Cash flow matters for fixed investment in the full sample and advanced economies, but not in developing countries.
  - Sales growth negative effect on investment notably greater in emerging market economies.
  - Firm size and asset tangibility dampen fixed investment; magnitudes comparable across country groups.
  - Effective interest rate associated with lower capital spending; not statistically significant in emerging market economies.
  - Older firms invest more than younger corporations, especially in developing countries.
- Robustness checks:
  - Introducing additional firm-level variables (liquidity and capital intensity) and country-level institutional controls (corruption and the rule of law) confirms baseline findings (Appendix Table A5).
  - Estimates for subperiods 1997-2007 and 2008-2018, separated samples by firm size and by leverage percentiles (small/large firms: below 25th vs above 75th percentile; low-indebted vs highly indebted firms: leverage below 25th vs above 75th percentile) remain broadly consistent with baseline findings (Appendix Table A6).

### V. Conclusion — key messages and policy-relevant implications
- Context:
  - The global economy faces an unprecedented slump from the coronavirus pandemic amid record-breaking corporate debt levels.
  - Protracted downturn risks a vicious cycle exacerbating corporate vulnerabilities, macro-financial instability, and long-lasting damage to economic potential.
- Empirical summary:
  - Analysis uses firm-level balance sheet data for more than 1.8 million nonfinancial firms from 52 countries over 1997–2018.
  - Corporate leverage is highly vulnerable to disruptions in profitability and cash flow at the firm level and to economic growth at the aggregate level.
  - Corporate debt overhang can become a strenuous burden on nonfinancial firms during severe downturns.
  - Highly leveraged firms tend to have lower levels of fixed investment spending.
- Implication for policy and macro-financial stability:
  - In times of greater macro-financial uncertainty, corporate debt overhang is likely to depress nonfinancial capital investment for years, especially if the COVID-19 pandemic lingers and the global downturn becomes protracted.

*Content derived from the supplied PDF excerpt.*

### Appendix Table A1. Breakdown by Year

### Appendix Table A1. Breakdown by Year

### Yearly observations
- 1997: 2,317
- 1998: 2,965
- 1999: 10,197
- 2000: 14,024
- 2001: 251,143
- 2002: 291,733
- 2003: 389,718
- 2004: 477,346
- 2005: 525,481
- 2006: 614,713
- 2007: 855,231
- 2008: 953,859
- 2009: 1,018,369
- 2010: 1,087,112
- 2011: 1,097,868
- 2012: 1,131,531
- 2013: 1,165,906
- 2014: 1,159,494
- 2015: 1,106,179
- 2016: 1,144,517
- 2017: 1,035,979
- 2018: 197,963
- Total: 14,533,645

### Country groups (aggregate counts)
- Advanced economies: 13,511,498 observations
- Developing economies: 1,022,147 observations
- Total: 14,533,645

### Appendix Table A3. Summary Statistics

### Firm characteristics (selected variables; Unit, Min, p25, p50, p75, Max, Average, Std. dev., Num. of obs.)
- Leverage: Ratio; Min 0.00; p25 0.05; p50 0.20; p75 0.42; Max 2.34; Average 0.27; Std. dev. 0.27; Num. of obs. 14,533,645
- Fixed investment: Ratio; Min 0.00; p25 0.01; p50 0.03; p75 0.09; Max 5.13; Average 0.10; Std. dev. 0.24; Num. of obs. 6,062,631
- Total assets: Log; Min 0.00; p25 12.77; p50 13.86; p75 15.08; Max 32.51; Average 14.00; Std. dev. 1.84; Num. of obs. 14,533,645
- Profitability: Ratio; Min -3.21; p25 -0.01; p50 0.02; p75 0.07; Max 1.22; Average 0.02; Std. dev. 0.19; Num. of obs. 14,453,593
- Asset tangibility: Ratio; Min 0.00; p25 0.05; p50 0.20; p75 0.47; Max 0.99; Average 0.29; Std. dev. 0.28; Num. of obs. 14,477,526
- Sales growth: Ratio; Min -1.00; p25 -0.16; p50 0.00; p75 0.19; Max 15.03; Average 0.09; Std. dev. 0.76; Num. of obs. 14,257,481
- Cash flow: Ratio; Min -0.64; p25 -0.02; p50 0.00; p75 0.03; Max 2.27; Average 0.01; Std. dev. 0.13; Num. of obs. 14,289,598
- Interest rate: Ratio; Min 0.00; p25 0.03; p50 0.06; p75 0.11; Max 16.36; Average 0.20; Std. dev. 0.84; Num. of obs. 12,926,990
- Age: Log; Min 0.00; p25 1.95; p50 2.49; p75 3.00; Max 4.61; Average 2.48; Std. dev. 0.73; Num. of obs. 14,533,645
- Liquidity: Ratio; Min 0.00; p25 0.89; p50 1.29; p75 2.13; Max 216.3; Average 2.84; Std. dev. 8.91; Num. of obs. 14,412,022
- Capital intensity: Ratio; Min -5.06; p25 -0.43; p50 -0.20; p75 -0.09; Max 0.00; Average -0.35; Std. dev. 0.45; Num. of obs. 12,454,432

### Macroeconomic controls (selected)
- GDP per capita: Log; Min 6.80; p25 8.98; p50 9.97; p75 10.70; Max 11.63; Average 9.75; Std. dev. 1.08; Num. of obs. 951
- GDP growth: Ratio; Min -0.15; p25 0.01; p50 0.03; p75 0.05; Max 0.25; Average 0.03; Std. dev. 0.03; Num. of obs. 951
- Trade openness: Ratio; Min 0.16; p25 0.58; p50 0.81; p75 1.22; Max 4.37; Average 1.00; Std. dev. 0.71; Num. of obs. 951
- Private credit: Ratio; Min 0.00; p25 0.47; p50 0.84; p75 1.19; Max 3.09; Average 0.87; Std. dev. 0.50; Num. of obs. 951
- Corruption Index: Min 1.00; p25 2.50; p50 3.00; p75 4.50; Max 6.00; Average 3.47; Std. dev. 1.31; Num. of obs. 858
- Rule of law Index: Min 1.00; p25 4.00; p50 5.00; p75 6.00; Max 6.00; Average 4.61; Std. dev. 1.23; Num. of obs. 858

### Appendix Table A4. Fixed Investment Estimations by Sectors

### Dependent variable: Fixed investment — Firm-level coefficients (selected variables across sectors AGR, MIN, MFG, UTI, CON, IT, OTH, TRD, TRA, EST, ADM)
- Leverage (lag): coefficients by sector
  - AGR: -0.068*** [0.012]
  - MIN: -0.062** [0.027]
  - MFG: -0.083*** [0.005]
  - UTI: -0.094*** [0.021]
  - CON: 0.006 [0.007]
  - IT: -0.036** [0.015]
  - OTH: -0.095*** [0.013]
  - TRD: -0.040*** [0.005]
  - TRA: -0.088*** [0.011]
  - EST: -0.015* [0.008]
  - ADM: -0.073*** [0.009]
- Tangibility (lag): strongly negative across all sectors, examples:
  - AGR: -0.411*** [0.019]
  - MFG: -0.468*** [0.006]
  - IT: -0.424*** [0.025]
  - TRD: -0.409*** [0.007]
  - ADM: -0.354*** [0.012]
- Total assets (lag): negative across sectors, examples:
  - AGR: -0.031*** [0.002]
  - CON: -0.054*** [0.001]
  - IT: -0.077*** [0.003]
  - TRD: -0.058*** [0.001]
  - ADM: -0.069*** [0.001]
- Interest rate (lag): generally negative and small, examples:
  - AGR: -0.000** [0.000]
  - MFG: -0.000*** [0.000]
  - IT: -0.001*** [0.000]
  - TRD: -0.001*** [0.000]
  - ADM: -0.001*** [0.000]
- Age: negative in most sectors, examples:
  - AGR: -0.050*** [0.011]
  - MFG: -0.077*** [0.004]
  - UTI: -0.097*** [0.013]
  - OTH: -0.088*** [0.011]
  - ADM: -0.033*** [0.010]

### Macroeconomic controls (examples of coefficients)
- Real GDP per capita: negative and significant in many sectors, examples:
  - MFG: -0.264*** [0.010]
  - CON: -0.607*** [0.029]
  - IT: -0.764*** [0.073]
  - TRD: -0.328*** [0.017]
  - ADM: -0.555*** [0.043]
- Real GDP growth: mixed signs; examples:
  - AGR: -0.381*** [0.141]
  - UTI: 0.457*** [0.168]
  - IT: -1.673*** [0.274]
  - TRD: -1.661*** [0.075]
  - EST: 0.893*** [0.146]
- Trade openness: coefficients vary, many negative and significant, examples:
  - AGR: 0.186*** [0.023]
  - MFG: -0.161*** [0.008]
  - CON: -0.582*** [0.021]
  - IT: -0.714*** [0.049]
  - EST: -0.501*** [0.038]
- Financial development: mixed; examples:
  - MFG: -0.081*** [0.004]
  - UTI: 0.047*** [0.018]
  - IT: -0.204*** [0.025]
  - TRD: -0.099*** [0.006]
  - EST: 0.159*** [0.013]

### Sample sizes and model fit
- Number of observations by sector:
  - AGR: 110,734
  - MIN: 18,165
  - MFG: 975,832
  - UTI: 60,514
  - CON: 539,967
  - IT: 106,286
  - OTH: 176,411
  - TRD: 1,147,000
  - TRA: 193,075
  - EST: 320,315
  - ADM: 353,277
- Number of firms by sector:
  - AGR: 31,557
  - MIN: 4,591
  - MFG: 254,733
  - UTI: 15,818
  - CON: 161,324
  - IT: 33,251
  - OTH: 54,353
  - TRD: 329,775
  - TRA: 52,459
  - EST: 96,667
  - ADM: 109,748
- Fixed effects: T+C in all regressions
- Adj R-squared ranges (examples): 0.015 (MIN) to 0.050 (IT)

### Notes
- Robust standard errors clustered at the firm level reported in brackets.
- Firm fixed effects included in all regressions.
- Sector abbreviations: AGR – Agribusiness, MIN – Mining, MFG – Manufacturing, UTI – Utilities, CON – Construction, IT – Information technology, OTH – Other service activities, households, extra territorial bodies, TRD – Wholesale and retail trade, accommodation, TRA – Transport and storage, EST – Real estate, ADM – Professional and administrative activities.
- Significance: *** p<0.01, ** p<0.05, * p<0.1

### Appendix Table A5. Fixed Investment Estimations—Additional Controls

### Dependent variable: Fixed investment — Full-sample and split by Advanced/Developing
- Leverage (lag):
  - All (col 1): -0.043*** [0.003]
  - All (col 2): -0.054*** [0.003]
  - Advanced (col 3): -0.043*** [0.003]
  - Advanced (col 4): -0.053*** [0.003]
  - Developing (col 5): -0.043*** [0.013]
  - Developing (col 6): -0.068*** [0.017]
- Profitability (lag):
  - All (1): 0.054*** [0.004]
  - All (2): 0.041*** [0.004]
  - Advanced (3): 0.052*** [0.004]
  - Advanced (4): 0.046*** [0.004]
  - Developing (5): 0.081*** [0.015]
  - Developing (6): 0.039** [0.017]
- Tangibility (lag):
  - All (1): -0.374*** [0.003]
  - All (2): -0.389*** [0.004]
  - Advanced (3): -0.372*** [0.003]
  - Advanced (4): -0.391*** [0.004]
  - Developing (5): -0.412*** [0.015]
  - Developing (6): -0.361*** [0.017]
- Total assets (lag):
  - All (1): -0.076*** [0.001]
  - All (2): -0.043*** [0.000]
  - Advanced (3): -0.076*** [0.001]
  - Advanced (4): -0.047*** [0.000]
  - Developing (5): -0.098*** [0.004]
  - Developing (6): -0.056*** [0.003]
- Other notable coefficients (macroeconomic):
  - GDP per capita (All): -0.586*** [0.010]
  - GDP per capita (Advanced): -0.853*** [0.012]
  - GDP per capita (Developing): 0.169*** [0.040]
  - GDP growth (All): -0.882*** [0.045]
  - Trade openness (All): -0.337*** [0.008]
  - Financial development (All): -0.073*** [0.003]
  - Rule of law (All): 0.015*** [0.002]
- Number of observations:
  - Col (1): 3,721,468
  - Col (2): 3,440,685
  - Col (3): 3,510,217
  - Col (4): 3,293,636
  - Col (5): 211,251
  - Col (6): 147,043
- Number of firms:
  - Col (1): 1,029,214
  - Col (2): 977,855
  - Col (3): 961,454
  - Col (4): 931,525
  - Col (5): 67,760
  - Col (6): 46,329
- Fixed effects: S*T+S*C in all specifications
- Adj R-squared:
  - Col (1): 0.047
  - Col (2): 0.035
  - Col (3): 0.047
  - Col (4): 0.036
  - Col (5): 0.050
  - Col (6): 0.046

### Notes
- Robust standard errors clustered at the firm level reported in brackets.
- Firm fixed effects included in all regressions.
- Singleton observations excluded in specifications with all three fixed effects interacted.
- Significance: *** p<0.01, ** p<0.05, * p<0.1

### Appendix Table A6. Fixed Investment Estimations—Sub-samples

### Dependent variable: Fixed investment — Sub-samples (Pre-GFC, Post-GFC, Small, Large, Low leverage, High leverage)
- Leverage (lag):
  - Pre-GFC: -0.083*** [0.004]
  - Post-GFC: -0.017*** [0.006]
  - Small: -0.125*** [0.005]
  - Large: -0.012*** [0.004]
  - Low leverage: 0.050*** [0.010]
  - High leverage: -0.250*** [0.006]
- Profitability (lag):
  - Pre-GFC: 0.009* [0.005]
  - Post-GFC: 0.049*** [0.009]
  - Small: -0.021*** [0.007]
  - Large: 0.152*** [0.004]
  - Low leverage: 0.010 [0.009]
  - High leverage: 0.019** [0.009]
- Tangibility (lag):
  - Pre-GFC: -0.474*** [0.008]
  - Post-GFC: -0.581*** [0.009]
  - Small: -0.519*** [0.007]
  - Large: -0.472*** [0.006]
  - Low leverage: -0.404*** [0.017]
  - High leverage: -0.756*** [0.009]
- Total assets (lag):
  - Pre-GFC: -0.082*** [0.002]
  - Post-GFC: -0.198*** [0.002]
  - Small: -0.029*** [0.001]
  - Large: -0.405*** [0.003]
  - Low leverage: -0.072*** [0.001]
  - High leverage: -0.057*** [0.001]

### Macroeconomic controls (selected coefficients, examples)
- Real GDP per capita:
  - Pre-GFC: -0.043* [0.022]
  - Post-GFC: -0.347*** [0.020]
  - Small: -0.262*** [0.021]
  - Large: 0.921*** [0.016]
  - Low leverage: -0.449*** [0.034]
  - High leverage: -1.032*** [0.023]
- Real GDP growth:
  - Pre-GFC: -0.325*** [0.085]
  - Post-GFC: 0.630*** [0.066]
  - Small: -0.256*** [0.075]
  - Large: -0.324*** [0.054]
  - Low leverage: -1.170*** [0.137]
  - High leverage: 0.167** [0.074]
- Trade openness:
  - Pre-GFC: -0.144*** [0.034]
  - Post-GFC: -0.876*** [0.013]
  - Small: -0.131*** [0.017]
  - Large: -0.259*** [0.010]
  - Low leverage: -0.064* [0.036]
  - High leverage: -0.571*** [0.016]
- Financial development:
  - Pre-GFC: -0.006 [0.006]
  - Post-GFC: -0.170*** [0.008]
  - Small: 0.065*** [0.005]
  - Large: 0.037*** [0.006]
  - Low leverage: 0.071*** [0.012]
  - High leverage: -0.110*** [0.008]

### Sample sizes and fit
- Number of observations:
  - Pre-GFC: 883,991
  - Post-GFC: 1,257,531
  - Small: 1,375,768
  - Large: 2,313,731
  - Low leverage: 529,514
  - High leverage: 893,888
- Number of firms:
  - Pre-GFC: 287,080
  - Post-GFC: 379,643
  - Small: 445,722
  - Large: 800,934
  - Low leverage: 210,286
  - High leverage: 299,117
- Fixed effects: S*T+S*C in all sub-sample regressions
- Adj R-squared:
  - Pre-GFC: 0.056
  - Post-GFC: 0.200
  - Small: 0.025
  - Large: 0.332
  - Low leverage: 0.041
  - High leverage: 0.078

### Notes
- Robust standard errors clustered at the firm level reported in brackets.
- Firm fixed effects included in all regressions.
- Singleton observations excluded in specifications with all three fixed effects interacted.
- Significance: *** p<0.01, ** p<0.05, * p<0.1

*Source: wpiea2020287-print-pdf - Appendix Tables A1–A6*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2020/english/wpiea2020287-print-pdf.pdf_