## 1. Credit to the Public Sector

## Source details

**Canonical URL:** [1. Credit to the Public Sector](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2013/_wp13206.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2013/_wp13206.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2013/_wp13206.pdf.json)

---

### I. Research question and dataset
- Purpose: Test whether government ownership of commercial banks is associated with a relaxed financing constraint for public sector entities, weaker fiscal discipline, and crowding out of private-sector credit.
- Panel dataset: 123 countries over 2008–10.
- Key regression specification:
  - Dependent variable: credit to the public sector as share of total banking system assets.
  - Main explanatory variable: government ownership of the banking system (assets of state-owned banks as share of total banking system assets).
  - Controls: lagged real GDP growth, inflation, overall fiscal balance, gross national savings, real interest rate, broad money as percent of GDP, multidimensional governance indicator, dummies for currency crisis, stock market crashes, banking crises.
  - Estimation: country fixed effects and time effects employed; fixed effects preferred after Hausman test in almost all specifications.

### II. Taxonomy and roles of state-owned banks (paper focus)
- State-owned retail commercial banks:
  - Perform deposit-taking and lending like private banks but are typically assigned policy objectives:
    - providing credit to specific sectors;
    - promoting access to banking services for underserved population groups or regions;
    - mitigating market failures from information asymmetries;
    - financing socially valuable (possibly financially unprofitable) projects;
    - competing with private institutions to lower financial intermediation costs.
- Other categories (context): development banks, quasi-narrow banks, development agencies.
- Performance concerns: objectives often reduce profitability and increase riskiness and misallocation of capital; weak governance exacerbates lower profitability and higher non-performing loans.

### III. Conceptual channel: soft financing constraint
- Distinction:
  - “soft financing constraint”: public entity obtains financing it would not access absent government ownership of the lender.
  - “soft budget constraint”: direct transfers/bailouts.
- Mechanism: government ownership of banks can remove private-lender monitoring, allowing public entities to access credit that would otherwise be unavailable or only at higher cost.
- Implication: state-owned banks can be captive sources of financing for the public sector, potentially weakening fiscal discipline and temporarily hiding fiscal imbalances within state-owned banks.

### IV. Stylized facts on government ownership and credit allocation
- In 96 countries where the government owns some commercial banks:
  - state-owned banks hold on average 21 percent of banking system assets; median is 15 percent.
- Descriptive sample (Table 1, averages 2008–10):
  - Sample of 75 countries: state-owned banks' assets are on average about 20 percent of total banking system assets; 13 percent of banking system assets are claims on the public sector.
  - In 24 countries with no government participation in the banking system, the banking system holds on average 9 percent of its assets in the public sector.
- Scatter-plot regression (relationship between state-owned banks' share and credit to public sector):
  - y = 0.274x + 7.809, prob(b) = 0.000

### V. Empirical findings — Credit to the public sector
- Main result: higher government participation in the banking system is associated with more financing to the public sector.
- Magnitude:
  - A one percentage point increase in the share of assets of the banking system owned by the government is associated with an increase in the share of credit to the public sector as a percent of total assets ranging between 0.4 and 0.5 percentage points (result robust across specifications and sample changes).
- Governance effect: the multidimensional governance indicator is significant and negative (better governance → less credit to public sector, conditional on controls).
- Estimation details: fixed effects and time effects included to absorb crisis-wide changes.

### VI. Empirical findings — Fiscal discipline (descriptive and regression evidence)
- Descriptive statistics (Table 3, averages):
  - In 93 countries where the government owns banks, public debt is about 45 percent of GDP.
  - This is 7 percentage points lower than in the 25 countries where the government does not own banks; the difference is not statistically significant.
- Key regression associations (selected reported magnitudes and statistics):
  - Each additional percentage point in the share of banking system’s assets owned by the government is associated with an increase in public debt of 0.2–0.3 percent of GDP.
  - A one percentage point increase in the share of assets of banks owned by the government is associated with an increase in the overall deficit of about 0.15 percent of GDP.
  - Example coefficients reported (Overall balance regressions): lagged real GDP growth: -0.723***, -0.745***, -0.673*** (standard errors: (0.171), (0.170), (0.153)); inflation examples: 0.169**, 0.154**, 1.225*** (standard errors: (0.069), (0.069), (0.315)).
  - Government ownership coefficients in overall-balance regressions: 0.190, 0.158, 0.192*, 0.191* (standard errors: (0.138), (0.137), (0.111), (0.114)).
  - Example overall-balance coefficient on public debt in debt regressions: -0.548** (standard error (0.229)).
  - Observations and R-squared examples reported across specifications: Observations 349, 349, 250, 250, 119, 119, 119; R-squared 0.158, 0.179, 0.308, 0.308, 0.424, 0.441, 0.443.
  - Countries in samples examples: 119, 119, 87, 87, 41, 41, 41.
- Appendix cross-section 2008 examples (selected estimates):
  - Model (1), Credit to the public sector: Government ownership coefficient 0.239** (0.099); Observations: 983; R-squared: 0.188.
  - Model (5), Debt: Government ownership coefficient 0.184* (0.106); Observations: 4114; R-squared: 0.115.
- FGLS examples (selected estimates):
  - Model (1), Credit to the public sector: Government ownership 0.370*** (0.021); Observations: 287; Countries: 98.
  - Model (5), Debt: Government ownership 0.131** (0.060); Observations: 3491; Countries: 119.
  - Notes: Includes country and time effects. Standard errors in parentheses, *** p<0.01, ** p<0.05, * p<0.1.

### VII. Crowding out of private-sector credit
- Sample averages:
  - In the 92 countries with government participation in the banking system, credit to the private sector averages 81 percent of total assets.
  - This is 2 percentage points lower than in the 26 countries without government participation.
- Regression evidence:
  - Government participation is significantly associated with crowding out of private-sector credit in most specifications.
  - Examples of government ownership coefficients in private-sector credit regressions: -0.058, -0.059, -0.126*, -0.123*, -0.147*, -0.154**, -0.153** (standard errors reported per specification).
  - When normalized by GDP, the estimated effect is that a one percentage point increase in government-owned bank asset share reduces the private-vs-public credit share by slightly more than 0.5 of a percentage point.
- Appendix FGLS private-public differential examples:
  - Model (7): Government ownership -0.119* (0.062); Observations: 193; Countries: 1596.
  - Model (8): Government ownership -0.569*** (0.074); Observations: 1596; Countries: 834.

### VIII. Robustness checks and diagnostics
- Exercises performed:
  - Cross-section regressions for 2008 (pre-crisis, pre-bailouts) — broadly consistent; effect on overall balance turns insignificant in some specs but coefficients close to 10 percent significance threshold.
  - Inclusion of output gap in panel specifications — broadly consistent, reduced observations.
  - Controls for country-group effects via regional dummies and advanced vs. emerging/developing economy dummy — no noteworthy variations.
  - Tests for serial correlation and heteroskedasticity; models re-estimated using Feasible Generalized Least Squares (FGLS) — similar coefficient magnitudes and significance.
  - Substitution of multidimensional governance indicator with subcomponents — only control of corruption shows a negative and significant coefficient when dependent variable is public debt.
- Selected robustness regression examples (coefficients reported exactly as in source): government ownership coefficients -0.040, 0.005, -0.734***, -0.664***, -0.660***, -0.595***, -0.566***, -0.545**; inflation coefficients -0.103, -0.097, -1.085***, -0.594***, -0.597***, -0.619***, -0.622***, -0.617***.
- Observations and R-squared examples in robustness sets: Observations 324, 306, 220, 204, 204, 99, 99, 99; R-squared 0.029, 0.047, 0.181, 0.245, 0.246, 0.556, 0.564, 0.564.
- Countries in robustness samples examples: 117, 111, 81, 76, 73, 73, 73, 73.

### IX. Conclusions and policy implications
- Main conclusions:
  - Government participation in the banking industry is associated with more financing from banks to the public sector, higher fiscal deficits, higher public debt, and a relatively lower share of total credit going to the private sector.
  - Results are robust to multiple control variables and alternative specifications.
- Policy implications and recommendations emphasized by authors:
  - It may be insufficient to insist on fiscal discipline without establishing conditions to achieve it.
  - Close monitoring of public sector entities’ access to financing from banks owned by the government should be considered.
  - Privatization of state-owned banks is proposed as a potential, more extreme solution to address underlying governance problems, but must be carefully evaluated because state-owned banks can fulfill functions not performed by private banks and finance projects with positive externalities.
  - Assessing financing practices of state-owned banks concerning the public sector should be considered in any strategy to pursue fiscal discipline.

*Source: Authors' calculations.*

### 1. Credit to the Public Sector .........................................................................................

### 1. Credit to the Public Sector

### I. Research question and dataset
- Purpose: Test whether government ownership of commercial banks is associated with a relaxed financing constraint for public sector entities, weaker fiscal discipline, and crowding out of private-sector credit.
- Panel dataset: 123 countries over 2008–10.
- Key regression specification: dependent variable is credit to the public sector as share of total banking system assets; main explanatory variable is government ownership of the banking system (assets of state-owned banks as share of total banking system assets). Model includes lagged real GDP growth, inflation, overall fiscal balance, gross national savings, real interest rate, broad money as percent of GDP, a multidimensional governance indicator, and dummies for currency crisis, stock market crashes, and banking crises. Country fixed effects and time effects employed.

### II. Taxonomy and roles of state-owned banks (focus of the paper)
- State-owned retail commercial banks (paper’s focus) perform deposit-taking and lending like private banks but are typically assigned policy objectives:
  - providing credit to specific sectors;
  - promoting access to banking services for underserved population groups or regions;
  - mitigating market failures from information asymmetries;
  - financing socially valuable (possibly financially unprofitable) projects;
  - competing with private institutions to lower financial intermediation costs.
- Other categories (context): development banks, quasi-narrow banks, development agencies.
- Performance concerns: objectives often reduce profitability and increase riskiness and misallocation of capital; weak governance exacerbates lower profitability and higher non-performing loans.

### III. Conceptual channel: soft financing constraint
- Distinction: “soft financing constraint” (public entity obtains financing it would not access absent government ownership of the lender) vs. “soft budget constraint” (direct transfers/bailouts).
- Mechanism: government ownership of banks can remove private-lender monitoring, allowing public entities to access credit that would otherwise be unavailable or only at higher cost.
- Implication: state-owned banks can be captive sources of financing for the public sector, potentially weakening fiscal discipline and temporarily hiding fiscal imbalances within state-owned banks.

### IV. Stylized facts on government ownership and credit allocation
- In 96 countries where the government owns some commercial banks, state-owned banks hold on average 21 percent of banking system assets; median is 15 percent.
- Descriptive sample (Table 1, averages 2008–10):
  - Sample of 75 countries: state-owned banks' assets are on average about 20 percent of total banking system assets; 13 percent of banking system assets are claims on the public sector.
  - In 24 countries with no government participation in the banking system, the banking system holds on average 9 percent of its assets in the public sector.
- Scatter-plot regression shown: y = 0.274x + 7.809, prob(b) = 0.000 (relationship between state-owned banks' share of banking system assets and banking system credit to public sector as share of total assets).

### V. Empirical findings — Credit to the public sector
- Main result: higher government participation in the banking system is associated with more financing to the public sector.
- Magnitude: a one percentage point increase in the share of assets of the banking system owned by the government is associated with an increase in the share of credit to the public sector as a percent of total assets ranging between 0.4 and 0.5 percentage points (result robust across specifications and sample changes).
- Among controls, the multidimensional governance indicator is significant and negative (better governance → less credit to public sector, conditional on controls).
- Estimation approach: fixed effects preferred after Hausman test in almost all specifications; time effects included to absorb crisis-wide changes.

### VI. Empirical findings — Fiscal discipline (descriptive evidence)
- Descriptive statistics (Table 3, averages):
  - In 93 countries where the government owns banks, public debt is about 45 percent of GDP.
  - This is 7 percentage points lower than in the 25 countries where the government does not own banks; the difference is not statistically significant.
- Overall-balance relationship: descriptive analysis does not show a strong correlation between government ownership of the banking system and the overall balance (text indicates no strong correlation; detailed regression results follow in subsequent sections).

### VII. Implications highlighted by the paper
- Government participation in the banking system can soften public-sector financing constraints without direct transfers, by providing access to credit that private lenders might withhold or provide at higher cost.
- This softening can:
  - lead to larger credit allocations to the public sector;
  - be associated with larger fiscal deficits and public debt levels (empirical evidence supports the credit association and indicates links to weaker fiscal indicators in regressions reported elsewhere in the chapter);
  - crowd out credit to the private sector (evidence and regressions on private-sector credit and private-public credit differential are discussed in later sections of the paper).

*Source: IMF Working Paper chapter "1. Credit to the Public Sector" (content unit provided).*

### 3.4 percent of GDP, only 0.4 percent of GDP higher that in the 27 countries in which the

### _wp13206 - 3.4 percent of GDP, only 0.4 percent of GDP higher that in the 27 countries in which the government has no participation in the banking system.

### Key empirical findings
- Government participation in the banking system is associated with more financing from the banking system to the public sector.
- Each additional percentage point in the share of banking system’s assets owned by the government is associated with an increase in public debt of 0.2–0.3 percent of GDP.
- A one percentage point increase in the share of assets of banks owned by the government is associated with an increase in the overall deficit of about 0.15 percent of GDP.
- In the 92 countries with government participation in the banking system, credit to the private sector averages 81 percent of total assets, which is 2 percentage points lower than in the 26 countries without government participation.
- An increase in the share of assets of banks owned by the government of one percentage point is associated with a decrease in the share of credit to the private sector (relative to the share devoted to the public sector) of slightly more than 0.5 of a percentage point.

### Regression evidence — public debt and overall balance (selected coefficients and statistics as reported)
- Table 4 (Public debt regressions) summary statements:
  - Government ownership of banking system coefficient: associated with higher public debt; magnitude 0.2–0.3 percent of GDP per percentage point increase in government-owned bank assets.
  - GDP growth contributes to lower public debt.
  - Higher inflation associated with higher indebtedness.
  - Larger deficits increase indebtedness.
  - Better governance associated with lower public debt.
- Table 5 (Overall balance regressions) selected reported coefficients and statistics:
  - Lagged real GDP growth coefficients (examples): -0.723***, -0.745***, -0.673*** (standard errors: (0.171), (0.170), (0.153)).
  - Inflation coefficients (examples): 0.169**, 0.154**, 1.225*** (standard errors: (0.069), (0.069), (0.315)).
  - Government ownership of banking system coefficients (examples): 0.190, 0.158, 0.192*, 0.191* (standard errors: (0.138), (0.137), (0.111), (0.114)).
  - Overall balance coefficient on public debt regressions example: -0.548** (standard error (0.229)).
  - Constants reported: 43.170***, 43.280***, 27.994***, 28.041***, 40.929***, 41.913***, 42.210*** (standard errors listed per specification).
  - Observations and R-squared examples: Observations 349, 349, 250, 250, 119, 119, 119; R-squared 0.158, 0.179, 0.308, 0.308, 0.424, 0.441, 0.443.
  - Countries in samples: 119, 119, 87, 87, 41, 41, 41.
- Notes provided: Includes time and fixed effects. Standard errors in parentheses, *** p<0.01, ** p<0.05, * p<0.1. Source: Authors' calculations.

### Crowding out of private-sector credit
- Evidence:
  - Regression results indicate government participation in the banking system is significantly associated with crowding out of credit to the private sector in most specifications.
  - Government ownership of banking system coefficients in private-sector credit regressions: examples include -0.058, -0.059, -0.126*, -0.123*, -0.147*, -0.154**, -0.153** (standard errors reported per specification).
  - When normalized by GDP, the estimated effect is that a one percentage point increase in government-owned bank asset share reduces the private-vs-public credit share by slightly more than 0.5 percentage point.
- Sample averages:
  - Credit to the private sector averages for countries with state-owned banks: 81 percent of total assets (Mean/Median entries and sample counts provided in Table 6).
  - Difference vs. countries without state-owned banks: private credit share is 2 percentage points lower in countries with state-owned banks (92 countries) versus 26 countries without.

### Robustness checks and diagnostics
- Robustness exercises performed:
  - Cross-section regressions for the year 2008 (pre-crisis, pre-bailouts) — results broadly consistent; effect of government ownership on overall balance turns insignificant but coefficients close to 10 percent significance threshold.
  - Inclusion of output gap in panel specifications — results broadly consistent, but reduced observations.
  - Controls for country-group effects via regional dummies and advanced vs. emerging/developing economy dummy — results do not show noteworthy variations.
  - Tests for serial correlation and heteroskedasticity produced mixed results; models re-estimated using Feasible Generalized Least Squares (FGLS) — Appendix II, Table 2 shows similar coefficient magnitudes and significance.
  - Substitution of multidimensional governance indicator with subcomponents (government effectiveness, regulatory quality, rule of law, control of corruption) — only control of corruption shows a negative and significant coefficient when the dependent variable is public debt; otherwise no single governance subcomponent drives results.
- Selected robustness regression excerpts (examples of reported coefficients and significance):
  - Government ownership of banking system coefficients in robustness regressions: -0.040, 0.005, -0.734***, -0.664***, -0.660***, -0.595***, -0.566***, -0.545** (standard errors listed).
  - Inflation coefficients in robustness regressions: -0.103, -0.097, -1.085***, -0.594***, -0.597***, -0.619***, -0.622***, -0.617*** (standard errors listed).
  - Observations and R-squared examples: Observations 324, 306, 220, 204, 204, 99, 99, 99; R-squared 0.029, 0.047, 0.181, 0.245, 0.246, 0.556, 0.564, 0.564.
  - Countries in robustness samples: 117, 111, 81, 76, 73, 73, 73, 73.
- Notes: Includes time and fixed effects. Standard errors in parentheses, *** p<0.01, ** p<0.05, * p<0.1. Source: Authors' calculations.

### Conclusions and policy implications
- Main conclusions:
  - Government participation in the banking industry is associated with more financing from banks to the public sector, higher fiscal deficits, higher public debt, and a relatively lower share of total credit going to the private sector.
  - Results are robust to multiple control variables and alternative specifications.
- Policy implications and recommendations emphasized by authors:
  - It may be insufficient to insist on fiscal discipline without establishing conditions to achieve it.
  - Close monitoring of public sector entities’ access to financing from banks owned by the government should be considered.
  - Privatization of state-owned banks is proposed as a potential, more extreme solution to address underlying governance problems, but must be carefully evaluated because state-owned banks can fulfill functions not performed by private banks and finance projects with positive externalities.
  - Assessing financing practices of state-owned banks concerning the public sector should be considered in any strategy to pursue fiscal discipline.

*Source: Authors' calculations.*

### Appendix I. Data

### Appendix I. Data

### Variable descriptions and sources
- Credit to public sector — Ratio of banking system's credit to public sector as a share of total banking system's assets, percent. Source: World Bank, Global Financial Development, 2013 (Database and Survey)
- Overall balance — Ratio of net lending/borrowing of the general government to GDP, percent. Source: WEO
- Debt — Ratio of gross debt of the general government to GDP, percent. Source: WEO
- Private-public credit differential — Difference between banking system's credit to the private sector and banking system's credit to the public sector as a share of GDP, percent. Source: World Bank, Global Financial Development, 2013 (Database and Survey)
- Real GDP growth — Growth rate of real GDP, percent. Source: WEO
- Inflation — Growth rate of the Consumer Price Index, average, percent. Source: WEO
- Government ownership of banking system — Ratio of state-owned banks' assets to banking system's assets, percent. Source: World Bank, Global Financial Development, 2013 (Database and Survey)
- Gross national savings — Ratio of gross national savings to GDP, percent. Source: WEO
- Real interest rate — Nominal interest rate minus the inflation rate, percent. Source: WEO
- Broad money — Ratio of broad money to GDP, percent. Source: WEO
- Governance — Additive index of government effectiveness, regulatory quality, rule of law, and control of corruption. Source: World Governance Indicators
- Currency crisis — Dummy taking the value 1 if there is an annual depreciation versus the US dollar (or the relevant anchor currency) of 15 percent or more, and/or a reduction in the metallic content of coins in circulation of 5 percent or more, and/or a currency reform where a new currency replaces a much-depreciated earlier currency in circulation, zero otherwise. Source: Reinhart and Rogoff, 2010
- Stock market crash — Dummy taking the value 1 if there are cumulated multi-year real returns of –25 percent or less, zero otherwise. Source: Reinhart and Rogoff, 2010
- Banking crisis — Dummy taking the value 1 if there is a bank run that leads to the closure, merging, or takeover by the public sector of one or more financial insitutions and/or there is no run, the closure, merging, takeover, or large-scale government assistance of an important financial institution (or groups of institutions) that marks the start of a string of similar outcomes for other financial institutions. Source: Reinhart and Rogoff, 2010
- Trade openness — Ratio of the sum of imports and exports of goods and services to GDP, percent. Source: WEO
- Oil exports — Ratio of oil exports to GDP, percent. Source: WEO

### Appendix II — Robustness tests: Cross-section regressions for 2008 (Table A2.1)
Notes: Robust standard errors in parentheses, *** p<0.01, ** p<0.05, * p<0.1. Source: Authors' calculations.

Selected coefficient estimates (coefficient (standard error)):
- Model (1) — Dependent: Credit to the public sector
  - Lagged real GDP growth: 0.560 (0.373)
  - Inflation: 0.044 (0.073)
  - Government ownership of banking system: 0.239** (0.099)
  - Constant: -0.471 (1.224)
  - Observations: 983
  - R-squared: 0.188
- Model (2) — Dependent: Credit to the public sector
  - Lagged real GDP growth: 0.478 (0.486)
  - Inflation: 0.635 (0.396)
  - Government ownership of banking system: 0.333* (0.174)
  - Constant: -11.351 (9.552)
  - Observations: 1119
  - R-squared: 0.615
- Model (3) — Dependent: Overall balance
  - Lagged real GDP growth: -0.198* (0.115)
  - Inflation: -0.017 (0.014)
  - Government ownership of banking system: 0.009 (0.024)
  - Constant: 1.295 (1.366)
  - Observations: 4419
  - R-squared: 0.045
- Model (4) — Dependent: Overall balance
  - Lagged real GDP growth: -0.724* (0.363)
  - Inflation: 0.008 (0.153)
  - Government ownership of banking system: 0.038 (0.047)
  - Constant: 3.654 (2.709)
  - Observations: 1194
  - R-squared: 0.328
- Model (5) — Dependent: Debt
  - Lagged real GDP growth: -2.676*** (0.876)
  - Inflation: 0.193* (0.115)
  - Government ownership of banking system: 0.184* (0.106)
  - Constant: 54.304*** (6.602)
  - Observations: 4114
  - R-squared: 0.115
- Model (6) — Dependent: Debt
  - Lagged real GDP growth: -1.955 (1.324)
  - Inflation: -0.695 (0.774)
  - Government ownership of banking system: 0.333* (0.182)
  - Constant: 54.516*** (14.203)
  - Observations: 11434
  - R-squared: 0.398
- Model (7) — Dependent: Private-public credit differential
  - Lagged real GDP growth: 1.626* (0.855)
  - Inflation: 3.234*** (0.836)
  - Government ownership of banking system: -0.726* (0.434)
  - Constant: -83.109*** (11.438)
  - Observations: 11434
  - R-squared: 0.219
- Model (8) — Dependent: Private-public credit differential
  - Lagged real GDP growth: 0.410 (1.034)
  - Inflation: -0.380 (1.412)
  - Government ownership of banking system: 0.705** (0.298)
  - Constant: -17.380 (24.771)
  - Observations: 11434
  - R-squared: 0.788

### Appendix II — Robustness tests: FGLS regressions (Table A2.2)
Notes: Includes country and time effects. Standard errors in parentheses, *** p<0.01, ** p<0.05, * p<0.1. Source: Authors' calculations.

Selected coefficient estimates (coefficient (standard error)):
- Model (1) — Dependent: Credit to the public sector
  - Lagged real GDP growth: 0.017 (0.012)
  - Inflation: 0.030*** (0.004)
  - Government ownership of banking system: 0.370*** (0.021)
  - Constant: -0.471 (1.224)
  - Observations: 287
  - Countries: 98
- Model (2) — Dependent: Credit to the public sector
  - Lagged real GDP growth: -0.105* (0.057)
  - Inflation: -0.157* (0.092)
  - Government ownership of banking system: 0.593*** (0.091)
  - Constant: -7.715* (4.590)
  - Observations: 883
  - Countries: 98
- Model (3) — Dependent: Overall balance
  - Lagged real GDP growth: -0.012 (0.016)
  - Inflation: -0.026*** (0.008)
  - Government ownership of banking system: -0.089*** (0.009)
  - Constant: 1.295 (1.366)
  - Observations: 349
  - Countries: 111
- Model (4) — Dependent: Overall balance
  - Lagged real GDP growth: -0.211*** (0.058)
  - Inflation: 0.304*** (0.058)
  - Government ownership of banking system: -0.130*** (0.041)
  - Constant: -5.760 (6.000)
  - Observations: 1263
  - Countries: 194
- Model (5) — Dependent: Debt
  - Lagged real GDP growth: -0.762*** (0.041)
  - Inflation: 0.054* (0.032)
  - Government ownership of banking system: 0.131** (0.060)
  - Gross national savings: -0.207*** (0.050)
  - Lagged public debt: -0.013*** (0.002)
  - Constant: 41.063*** (2.365)
  - Observations: 3491
  - Countries: 119
- Model (6) — Dependent: Debt
  - Lagged real GDP growth: -0.045 (0.092)
  - Inflation: 0.127 (0.090)
  - Government ownership of banking system: 0.196** (0.099)
  - Gross national savings: -0.160*** (0.048)
  - Real interest rate: 0.209*** (0.071)
  - Constant: 5.402 (4.589)
  - Observations: 1263
  - Countries: 159
- Model (7) — Dependent: Private-public credit differential
  - Lagged real GDP growth: -0.145*** (0.033)
  - Inflation: -0.116*** (0.028)
  - Government ownership of banking system: -0.119* (0.062)
  - Broad money: -0.053* (0.030)
  - Governance: -4.620*** (0.730)
  - Stock market crash: -2.386*** (0.710)
  - Lagged public debt: -0.096*** (0.020)
  - Trade openness: -0.086*** (0.026)
  - Oil exports: 0.351*** (0.081)
  - Constant: 4.671** (2.011)
  - Observations: 193
  - Countries: 1596
- Model (8) — Dependent: Private-public credit differential
  - Lagged real GDP growth: -0.015 (0.078)
  - Inflation: -0.736*** (0.071)
  - Government ownership of banking system: -0.569*** (0.074)
  - Real interest rate: -0.347*** (0.071)
  - Broad money: -0.501*** (0.039)
  - Governance: -1.924*** (0.578)
  - Currency crisis: -1.732*** (0.483)
  - Stock market crash: -1.135* (0.615)
  - Constant: 27.604*** (4.040)
  - Observations: 1596
  - Countries: 834

Italic source: Authors' calculations.

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2013/_wp13206.pdf_