## wp17143

## Source details

**Canonical URL:** [wp17143](https://www.imf.org/-/media/files/publications/wp/2017/wp17143.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2017/wp17143.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2017/wp17143.pdf.json)

---

### Introduction and motivation
- Study premise: financial development is not a magic bullet; relationship between finance, productivity, and investment is complex and not unambiguously positive across countries and time.
- Four dimensions of financial development analyzed:
  - (i) financial depth;
  - (ii) financial efficiency;
  - (iii) financial stability and openness;
  - (iv) access to financial services.
- Summary of high-level findings:
  - Financial stability and stock market efficiency are the only dimensions that appear related to productivity and capital accumulation across time and countries.
  - Effects of other dimensions (depth, efficiency, openness, access) depend on country characteristics such as income level and region.
  - Financial openness can boost productivity in low-income countries, but the positive effect vanishes in advanced countries.
  - Financial depth contributed positively to growth in the pre-2005 period; effects can vary over time.
  - Mild evidence of differential regional effects for CCA and MENA countries; CCA and MENA oil-importing countries may gain extra benefits from increasing banking sector efficiency (interest rate spreads and overhead costs).

### Contribution to literature, scope, and data
- Three complementarities with existing literature:
  - Sample extension: unbalanced panel data for 145 countries from 1960 to 2011, including the recent global financial crisis.
  - Breadth of indicators: 16 different indicators capturing depth, efficiency, stability/openness, and access.
  - Focus on sources of growth: decomposes growth into total factor productivity (TFP) growth and capital stock accumulation.
- Data and measurement:
  - Sample period: 1960 to 2011; unbalanced panel for 145 countries.
  - Sources of growth derived from Penn World Tables; TFP growth and capital stock accumulation constructed with methods accounting for heterogeneity in labor income and decomposed assets.
  - 16 financial development indicators (primary source: Global Financial Development Database); variables transformed into natural logarithms except Chinn Ito index (original form).
  - Controls: logarithm of initial real GDP per capita; average years of schooling; trade-to-GDP ratio; ratio of government consumption to GDP; oil prices.
  - Data treatment: multi-year non-overlapping averages when possible; three-year non-overlapping averages or annual observations for shorter series.

### Empirical methodology
- Baseline dynamic panel specification:
  - g_it = α y_{i,t−1} + β′ F_{i,t} + γ′ X_{i,t} + η_i + λ_t + ε_{i,t}
  - g_it: growth in either TFP or capital stock; F_{i,t}: financial dimension; X_{i,t}: macro controls; η_i: country fixed effects; λ_t: time effects.
- Estimation approach:
  - System GMM (Arellano and Bond 1997) used to mitigate weak instruments from first-differencing (Arellano and Bond 1991) bias.
  - Hansen test and AR(2) test used to assess instrument validity and serial correlation.
  - Capital accumulation regressions: two lags of dependent variable; TFP regressions: one lag.
  - Instruments restricted to fewer than the number of countries to avoid instrument proliferation.
- Heterogeneity and non-linearity:
  - Five interaction specifications: financial variables interacted with income level, inflation regime, quality of institutions, level of financial development, and regional dummies (CCA; MENAPOE; MENAPOI).
  - Dummy variables for (i–iv) created by splitting samples equally into three ranked subgroups.

### Main empirical results and robust relationships
- Overall conclusion: no unambiguously positive relationship between aggregate “finance” and the sources of growth; relationship is nuanced and can be detrimental when financial development is excessive.
- Most consistent links:
  - Financial stability and efficiency indicators are most consistently linked to growth outcomes.
  - Nonperforming loans (NPLs) and stock market turnover are the two indicators exhibiting a general, robust, and economically significant relationship with both productivity growth and capital accumulation.
- Quantitative illustrative result on NPLs:
  - A one standard-deviation decline in NPLs, corresponding to a decline in the ratio of NPLs to gross loans from 13 to 6 percent, is associated with a 1.2 and 1.5 percentage point improvement in annual TFP and investment growth, respectively.
- Stock market liquidity:
  - Stock market turnover ratio has a significantly positive effect on TFP and capital accumulation.
- Financial depth and openness:
  - Financial depth and financial openness show significant associations with either TFP or investment in benchmark specifications, but significance often driven by specific country-year observations; generalizability across full sample is limited.
- Unconditional correlations and plots:
  - Simple scatter plots and unconditional correlation coefficients show little evidence of a strong overarching relationship; when significant, effects are small and omit other macroeconomic, institutional, regional, or developmental contributions.

### Heterogeneity by income, inflation, institutions, and region
- Income thresholds (time-varying categories via bottom/middle/top thirds):
  - Low-income countries:
    - More borrowing from foreign banks → significantly faster TFP growth.
    - Deeper credit markets support TFP; coefficient significant at the 10 percent level.
    - Poor asset quality of banks → more pronounced detrimental effects on investment.
    - Stock market development does not emerge as an important determinant of growth.
  - Middle-income countries:
    - Middle-income interaction terms are never significant.
  - Advanced economies:
    - Deeper credit markets and more foreign borrowing depress TFP in advanced countries (estimates in columns 3 and 9); negative sign on advanced country interaction disappears and becomes insignificant when excluding the most recent observation.
  - Overall: positive effects of financial development accrue primarily in low-income countries; advanced countries can suffer from “excessive” financial development.
- Inflation and institution thresholds:
  - Interaction terms generally not significant, with two exceptions:
    1. Premium on safeguarding financial stability larger in low-inflation and high-institutional-quality environments.
    2. Macroeconomic stability and resilient institutions can reveal beneficial effects from financial development (e.g., financial integration under strong institutions and larger stock markets in low-inflation environments lead to faster investment growth).
- Regional heterogeneity (CCA; MENA oil exporters; MENA oil importers):
  - CCA:
    - Large benefits from increasing banking-sector efficiency; regulatory barriers have protected incumbent banks and reduced efficiency incentives.
  - MENA oil exporters:
    - Efficiency measured by bank overhead costs to total assets is the only significant financial dimension; greater financial efficiency associated with detrimental effects on productivity in some specifications.
  - MENA oil importers:
    - Weaker productivity in response to tighter net interest margins and greater financial stability (bank Z score and loan-to-deposit ratios) in some specifications.
  - Interpretation: financial development supports growth only if institutional prerequisites (creditor rights, supervision) are in place; increasing financial activity without adequate infrastructure can undermine growth prospects.

### Policy implications and conclusions
- Primary policy priority:
  - Safeguarding financial stability must be the first priority for countries seeking to accelerate financial development.
- Rationale and cautionary mechanisms:
  - A country’s risk of suffering a financial crisis increases with the size of its financial sector (Schularick and Taylor 2012); rising incidence of financial crises since 1990 helps explain vanishing growth effects (Rousseau and Wachtel 2011).
  - Rapid financial development can engender costs via externalities:
    - Skilled workers drawn into finance rather than R&D may reduce productivity-enhancing activity (Cecchetti and Kharroubi 2012).
    - Financial sectors focusing on asset transfers (real estate mortgages) rather than purchases of goods and services may destabilize asset prices and produce limited positive impact on real activity (Bezemer and others 2014).
- Prescriptive summary:
  - Financial development needs to proceed according to country-specific circumstances.
  - Financial stability is a pre-requisite to avoid financial development undermining growth.
  - CCA and some MENA countries stand to gain from increasing efficiency and competition in banking sectors, provided institutional reforms (creditor rights, supervision) accompany financial deepening.

### Key statistics and selected table highlights
- Summary statistics (counts, means, standard deviations, minima, maxima):
  - Private Credit: Count 5803; Mean 38.7; Standard Deviation 37.7; Minimum 0.0; Maximum 284.6
  - Liquid Liabilities: Count 5809; Mean 47.0; Standard Deviation 38.3; Minimum 0.3; Maximum 399.1
  - Stock Market Capitalization: Count 1972; Mean 48.1; Standard Deviation 57.2; Minimum 0.0; Maximum 569.5
  - Interest Rate Spread: Count 3756; Mean 8.0; Standard Deviation 7.6; Minimum 0.0; Maximum 91.8
  - Stock Market Turnover: Count 1935; Mean 46.0; Standard Deviation 59.4; Minimum 0.0; Maximum 538.2
  - Net Interest Margin: Count 2156; Mean 5.0; Standard Deviation 3.3; Minimum 0.0; Maximum 39.2
  - Bank Overhead Costs: Count 2167; Mean 4.1; Standard Deviation 3.7; Minimum 0.0; Maximum 90.3
  - Bank Non-Performing Loans: Count 1267; Mean 7.4; Standard Deviation 7.9; Minimum 0.1; Maximum 74.1
  - Stock Price Volatility: Count 1408; Mean 23.1; Standard Deviation 14.2; Minimum 2.4; Maximum 141.6
  - Bank Z-Score: Count 2143; Mean 14.9; Standard Deviation 10.0; Minimum -21.2; Maximum 65.3
  - Loan-to-Deposit Ratio: Count 7250; Mean 98.6; Standard Deviation 60.0; Minimum 1.1; Maximum 898.1
  - Capital Adequacy Ratio: Count 1276; Mean 15.8; Standard Deviation 5.2; Minimum 2.5; Maximum 48.6
  - Foreign Claims to GDP: Count 4425; Mean 36.2; Standard Deviation 76.0; Minimum 0.0; Maximum 957.1
  - Chinn Ito Index: Count 6137; Mean 0.4; Standard Deviation 0.4; Minimum 0.0; Maximum 1.0
  - Assets and Liabilities to GDP: Count 6337; Mean 238.5; Standard Deviation 1079.0; Minimum 5.3; Maximum 24074.9
  - Bank Branches: Count 1268; Mean 18.7; Standard Deviation 19.6; Minimum 0.3; Maximum 126.1
- Selected table insights (significance notation preserved as in source):
  - Table 2 (Five-year, Three-year, Annual averages):
    - Private Credit (Five Year): -0.071* (Productivity Growth)
    - Assets and Liabilities to GDP (Five Year): -0.095*** (Capital Stock Growth)
    - Stock Market Turnover (Three Year): 0.120*** (Capital Stock Growth)
    - Interest Rate Spread (Three Year): -0.119*** (Capital Stock Growth)
    - Bank Overhead Costs (Annual): 0.124*** (Productivity Growth); -0.180*** (Capital Stock Growth)
    - Capital Adequacy Ratio (Annual): 0.222*** (Productivity Growth); -0.142*** (Capital Stock Growth)
    - Bank Branches (Annual): -0.118*** (Productivity Growth); -0.148*** (Capital Stock Growth)
  - Diagnostic and coefficient examples from Tables 3–19 (selected values preserved):
    - Trade to GDP coefficients often positive and significant in many regressions (examples: 2.575***, 2.322**, 2.327***, 2.177**, 2.843***, 2.881***).
    - Government Consumption to GDP frequently negative and significant in many regressions (examples: -2.337***, -2.298***, -2.710***, -2.524***, -2.336***, -2.590***).
    - Years of Schooling often positive and significant in some specifications (examples: 2.148*, 3.312**, 2.639**).
    - Lag structure:
      - Lag 1 of TFP Growth: coefficients reported (examples include 0.110*, 0.098, 0.078).
      - Lag 1 of Capital Stock Growth: strongly positive and highly significant across many tables (e.g., 0.663***, 0.617***, 0.647***, 0.746***, 0.637***, 0.582***).
      - Lag 2 of Capital Stock Growth: often negative small coefficients (examples: -0.097*, -0.074, -0.054).
    - Selected financial-variable coefficients (examples preserved as reported):
      - Stock Market Capitalization: coefficients vary; examples include -3.725** (-2.45), -4.974*** (-2.92).
      - Interest Rate Spread: interaction coefficients often positive and significant (examples include 0.927** (2.05), 0.975** (2.11)).
      - Stock Market Turnover: mixed signs; examples include -3.865** (-2.10) in one specification and -2.191* (-1.92) in another.
      - Bank Overhead Costs: strong negative interactions in regional dummies (examples include -1.619* (-1.96) for CCA; -1.398*** (-3.14) for MENAPOI; -2.753** (-2.43) for MENAPOX).
      - Non-Performing Loans: interaction coefficients mixed; examples include -2.366** (-2.24) for Medium Financial Depth interaction in one specification.
      - Net Interest Margin: some large negative interactions in specific subsamples (example: -11.358*** (-3.39) for Net Interest Margin*Medium Financial Depth in one column).
      - Bank Z-Score and Stock Price Volatility: coefficients on capital often negative and significant (examples: Stock Price Volatility capital coefficient -3.862*** (-2.94); Bank Z-Score capital coefficient -1.820*** (-3.37)).
- Diagnostics and estimation notes preserved:
  - System GMM diagnostic statistics reported across tables (Observations, Number of clusters, Number of instruments, Hansen test p-value, A-B AR(1) p-value, A-B AR(2) p-value, lowest/highest/average number of observations included).
  - Significance legends retained as reported in source: * p<0.1  ** p<0.05  *** p<0.01.

*Source: IMF Working Paper WP/17/143, "Financial Development and Source of Growth: New Evidence" (wp17143).*

### Section 1

### Financial Development and Source of Growth: New Evidence

### I. Introduction and Motivation
- Study premise: financial development is not a magic bullet for economic development; the relationship between finance, productivity, and investment is complex and not unambiguously positive across countries and time.
- Four dimensions of financial development analyzed: (i) financial depth; (ii) financial efficiency; (iii) financial stability and openness; (iv) access to financial services.
- Findings summary:
  - Financial stability and stock market efficiency are the only dimensions that appear related to productivity and capital accumulation across time and countries.
  - Effects of other dimensions (depth, efficiency, openness, access) depend on country characteristics such as income level and region.
  - Financial openness can boost productivity in low-income countries, but the positive effect vanishes in advanced countries.
  - Financial depth contributed positively to growth in the pre-2005 period; effects can vary over time.
- Regional focus: CCA and MENA countries examined for differential effects.
  - Mild evidence of differential regional effects.
  - CCA and MENA oil-importing countries may gain extra benefits from increasing banking sector efficiency, measured by interest rate spreads and overhead costs — likely reflecting weak competition and resulting high prices for borrowers and high operational costs.

### II. Contribution to Literature and Scope
- Three complementarities with existing literature:
  - Sample extension: unbalanced panel data for 145 countries from 1960 to 2011, including the recent global financial crisis.
  - Breadth of indicators: 16 different indicators capturing depth, efficiency, stability/openness, and access.
  - Focus on sources of growth: decomposes growth into total factor productivity (TFP) growth and capital stock accumulation rather than growth per se.
- Literature context:
  - Earlier consensus generally supported a positive finance-growth relationship; more recent work shows complexity, thresholds, and possible vanishing effects.
  - Noted literature observations:
    - Positive finance-growth relationship may depend on income/financial development thresholds (Deidda and Fattouh 2002; Rioja and Valev 2004a,b).
    - Cechetti and Karroubi (2012) and Arcand et al. (2012) find financial development good for growth only up to a point between 90 percent and 100 percent of GDP, beyond which effects may turn negative.
    - Post-crisis studies (Rousseau and Watchel 2011; Beck et al. 2012b; Bezemer et al. 2014) find a lower or negative effect of finance on growth, potentially linked to crises and shifts in credit composition (household vs enterprise credit).

### III. Data
- Sample and period:
  - Unbalanced panel data for 145 countries from 1960 to 2011.
  - Number of observations varies by model and variable availability.
- Sources of growth:
  - Growth decomposed into total factor productivity (TFP) growth and capital stock accumulation using Penn World Tables; methodology accounts for heterogeneity in labor income and constructs capital stock based on decomposed assets.
- Financial development indicators:
  - 16 indicators across four dimensions: depth, efficiency, stability/openness, access.
  - Primary data source: Global Financial Development Database.
  - Variables transformed into natural logarithms (except Chinn Ito index used in original form).
- Control variables:
  - Logarithm of initial real GDP per capita for convergence.
  - Average years of schooling.
  - Trade-to-GDP ratio.
  - Ratio of government consumption to GDP.
  - Oil prices.
- Data treatment:
  - Use of multi-year, non-overlapping averages when possible to remove business cycle effects; for shorter series, three-year non-overlapping averages or annual observations used.

### IV. Empirical Methodology
- Baseline regression specification (general form):
  - g_it = α y_{i,t−1} + β′ F_{i,t} + γ′ X_{i,t} + η_i + λ_t + ε_{i,t}
  - g_it represents growth in either TFP or capital stock; F_{i,t} is the financial dimension; X_{i,t} contains macro controls; η_i country fixed effects; λ_t time effects.
- First-difference and GMM approach:
  - First-difference leads to Arellano and Bond (1991) bias; system GMM (Arellano and Bond 1997) employed to mitigate weak instruments by stacking regressions in differences and levels.
  - Hansen test and AR(2) test used to assess consistency of system GMM.
  - Capital accumulation regressions use two lags of the dependent variable; TFP regressions use one lag.
  - Instruments restricted to fewer than the number of countries to avoid instrument proliferation.
- Heterogeneity and non-linearity:
  - Five interaction specifications tested by interacting financial variables with:
    1. Income level
    2. Inflation regime
    3. Quality of institutions
    4. Level of financial development
    5. Regional dummies (CCA; MENAPOE; MENAPOI)
  - Dummy variables for (i–iv) created by splitting samples equally into three ranked subgroups.

### V. Empirical Results
- Overall conclusion: no unambiguously positive relationship between aggregate “finance” and the sources of growth; relationship is nuanced with possible detrimental effects from excessive financial development.
- Key robust findings:
  - Financial stability and efficiency indicators are most consistently linked to growth.
  - Nonperforming loans (NPLs) and stock market turnover are the two indicators that exhibit a general, robust, and economically significant relationship with both productivity growth and capital accumulation.
- Quantitative illustrative result on NPLs:
  - A one standard-deviation decline in NPLs, corresponding to a decline in the ratio of NPLs to gross loans from 13 to 6 percent, is associated with a 1.2 and 1.5 percentage point improvement in annual TFP and investment growth, respectively.
- Stock market liquidity:
  - Stock market turnover ratio has a significantly positive effect on TFP and capital accumulation, consistent with theories that emphasize liquidity’s role in improving capital allocation efficiency.
- Financial depth and openness:
  - Financial depth and financial openness show significant associations with either TFP or investment in benchmark specifications, but significance appears driven by specific country-year observations, limiting generalizability across the full sample.
- Unconditional correlations and data plots:
  - Simple scatter plots and unconditional correlation coefficients show little evidence of a strong overarching relationship between the financial indicators and sources of growth; when significant, effects are small and omit other macroeconomic, institutional, regional, or developmental contributions.

*Source: IMF Working Paper WP/17/143, "Financial Development and Source of Growth: New Evidence" (Section 1).*

### Section 2

### wp17143 - Section 2

### Main findings on financial development and growth
- Previous studies found that deeper financial markets—banking and non-banking—do not lead to better growth outcomes. The paper generalizes this finding: financial development is no magic bullet for generating growth.
- The relationship between financial development and the sources of growth is marked by thresholds and varies across country characteristics:
  - (1) income level;
  - (2) policy regime;
  - (3) institutional quality; and
  - (4) region.
- The effectiveness of financial development to accelerate productivity and investment depends on the specific economic circumstances of individual countries.
- The one indicator that stands out in importance for economic growth is financial stability, particularly in the banking sector.

### Income thresholds
- Method:
  - Time-varying income categories created each period: bottom third = low-income, middle third = middle-income, top third = advanced.
  - Dummy variables designate each country observation to one of the three income categories.
  - Interaction terms between income dummies and financial indicators measure differential impacts.
- Low-income countries:
  - More borrowing from foreign banks → significantly faster TFP growth.
  - Deeper credit markets support TFP; coefficient significant at the 10 percent level.
  - Poor asset quality of banks → more pronounced detrimental effects on investment.
  - Stock market development does not emerge as an important determinant of growth in low-income countries.
  - Interpretation: integrating domestic banking into international markets can deliver significant benefits at early stages, if countries safeguard financial stability; possible channels include access to foreign technology and markets (Giannetti and others 2002).
- Middle-income countries:
  - Middle-income interaction terms are never significant.
  - Contrasts with the inverted U-shape hypothesis (Rioja and Valev 2004b) that posits greatest importance at intermediate stages.
- Advanced economies:
  - Deeper credit markets and more foreign borrowing from international banks depress TFP in advanced countries (estimates in columns 3 and 9).
  - Cross-border banking flows can be destabilizing via current account reversals and exchange rate volatility; noted during the global financial crisis in 2008/09.
  - Unreported results excluding the most recent observation make the negative sign on the advanced country interaction disappear and the coefficient become insignificant.
- Overall: positive effects of financial development accrue primarily in low-income countries; advanced countries can suffer from “excessive” financial development.

### Inflation and institution thresholds
- Method:
  - For each period, construct three categories (low, medium, high) for inflation and institutional quality based on bottom/middle/top thirds of the distributions.
  - Interact resulting dummy variables with relevant financial indicators.
- Results:
  - Interaction terms of financial development with inflation and institution thresholds are generally not significant, with two exceptions:
    1. The premium on safeguarding financial stability is larger in environments characterized by low inflation and high institutional quality.
       - Rationale: in high-inflation/bad-institution environments, marginal impact from financial instability is smaller because the environment already saps growth.
    2. Macroeconomic stability and resilient institutions can bring out beneficial effects from financial development.
       - Financial integration under strong institutions and larger stock markets in low-inflation environments lead to faster investment growth.
       - Important for advanced countries to guard against unwanted consequences of “excessive” financial development.

### Regional heterogeneity
- Focus groups: Caucasus and Central Asia (CCA), MENA oil exporters, and MENA oil importers.
- CCA:
  - Economically large benefits from increasing efficiency in banking sectors.
  - Regulatory barriers have protected incumbent banks, reducing pressure to increase efficiency in banking operations and credit allocation.
- MENA oil exporters:
  - Only significant financial dimension is efficiency measured by bank overhead costs to total assets.
  - Dummy carries the wrong sign: greater financial efficiency associated with detrimental effects on productivity.
- MENA oil importers:
  - Appear to suffer weaker productivity in response to tighter net interest margins and greater financial stability (measured by bank Z score and loan to deposit ratios).
- Interpretation:
  - Financial development supports growth only if institutional prerequisites are in place.
  - MENA countries often have financial systems dominated by banks, weak creditor rights, and weak supervisory regimes (Rocha et al, 2011).
  - Increasing financial activity without adequate infrastructure can undermine growth prospects.
  - Evidence that CCA and some MENA countries stand to gain from increasing efficiency and competition in banking sectors.

### Conclusion and policy implications
- The study fails to establish positive relationships between most financial development indicators, investment, and TFP growth across the full sample.
- Most dimensions of financial development are effective only conditional on country characteristics.
- Key policy implications:
  - Safeguarding financial stability must be the first priority for countries seeking to accelerate financial development.
  - Reducing risks to financial stability is crucial because a country’s risk of suffering a financial crisis increases with the size of its financial sector (Schularick and Taylor 2012).
  - The rising incidence of financial crises since 1990 helps explain the vanishing growth effects of financial development (Rousseau and Wachtel 2011).
  - Rapid financial development can engender costs via externalities:
    - Skilled workers drawn into finance rather than R&D may reduce productivity-enhancing activity (Cecchetti and Kharroubi 2012).
    - Financial sectors focusing on asset transfers (real estate mortgages) rather than purchases of goods and services may destabilize asset prices and produce limited positive impact on real activity (Bezemer and others 2014).
  - Financial development needs to proceed according to country-specific circumstances; financial stability is a pre-requisite to avoid financial development undermining growth.
- Summary of nuanced outcomes:
  - Low- and middle-income countries: deeper financial markets tend to reduce productivity growth and investment, potentially linked to uncompetitive markets.
  - Low-income countries: productivity growth seems to benefit from more open financial markets.
  - Middle-income countries: better financial access associated with higher growth in investment.
  - CCA and MENA regions: indicators of financial stability and efficiency sometimes carry unexpected signs, highlighting the need for effective creditor rights and banking supervision.

### Key statistics (as reported in summary statistics table)
- Private Credit: Count 5803; Mean 38.7; Standard Deviation 37.7; Minimum 0.0; Maximum 284.6
- Liquid Liabilities: Count 5809; Mean 47.0; Standard Deviation 38.3; Minimum 0.3; Maximum 399.1
- Stock Market Capitalization: Count 1972; Mean 48.1; Standard Deviation 57.2; Minimum 0.0; Maximum 569.5
- Interest Rate Spread: Count 3756; Mean 8.0; Standard Deviation 7.6; Minimum 0.0; Maximum 91.8
- Stock Market Turnover: Count 1935; Mean 46.0; Standard Deviation 59.4; Minimum 0.0; Maximum 538.2
- Net Interest Margin: Count 2156; Mean 5.0; Standard Deviation 3.3; Minimum 0.0; Maximum 39.2
- Bank Overhead Costs: Count 2167; Mean 4.1; Standard Deviation 3.7; Minimum 0.0; Maximum 90.3
- Bank Non-Performing Loans: Count 1267; Mean 7.4; Standard Deviation 7.9; Minimum 0.1; Maximum 74.1
- Stock Price Volatility: Count 1408; Mean 23.1; Standard Deviation 14.2; Minimum 2.4; Maximum 141.6
- Bank Z-Score: Count 2143; Mean 14.9; Standard Deviation 10.0; Minimum -21.2; Maximum 65.3
- Loan-to-Deposit Ratio: Count 7250; Mean 98.6; Standard Deviation 60.0; Minimum 1.1; Maximum 898.1
- Capital Adequacy Ratio: Count 1276; Mean 15.8; Standard Deviation 5.2; Minimum 2.5; Maximum 48.6
- Foreign Claims to GDP: Count 4425; Mean 36.2; Standard Deviation 76.0; Minimum 0.0; Maximum 957.1
- Chinn Ito Index: Count 6137; Mean 0.4; Standard Deviation 0.4; Minimum 0.0; Maximum 1.0
- Assets and Liabilities to GDP: Count 6337; Mean 238.5; Standard Deviation 1079.0; Minimum 5.3; Maximum 24074.9
- Bank Branches: Count 1268; Mean 18.7; Standard Deviation 19.6; Minimum 0.3; Maximum 126.1

*Source: wp17143 - Section 2*

### Section 3

### wp17143 - Section 3

### Table 2: Summary Statistics of Raw Financial Variables (Five-year, Three-year, Annual)
- Five Year Averages (variable — Productivity Growth ; Capital Stock Growth):
  - Private Credit: -0.071* 0.037
  - Liquid Liabilities: -0.022 0.040
  - Loan-to-Deposit Ratio: -0.011 0.036
  - Foreign Claims to GDP: -0.083* -0.052
  - Chinn Ito Index: 0.011 0.017
  - Assets and Liabilities to GDP: -0.003 -0.095***
- Three Year Averages:
  - Stock Market Capitalization: -0.076* 0.041
  - Interest Rate Spread: 0.023 -0.119***
  - Stock Market Turnover: 0.034 0.120***
  - Net Interest Margin: 0.171*** -0.053
  - Non-Performing Loans: 0.021 -0.103*
  - Stock Price Volatility: 0.061 0.016
  - Bank Z-Score: -0.063 0.097**
- Annual Observations:
  - Bank Overhead Costs: 0.124*** -0.180***
  - Capital Adequacy Ratio: 0.222*** -0.142***
  - Bank Branches: -0.118*** -0.148***
- Note: Financial variables are included in log forms.
- Significance notation reported in source:
  - * denotes p<0.05, ** denotes p<0.01, and *** denotes p<0.001

### Table 3: Unconditional Correlations and Regression Diagnostic Statistics
- Sample statistics and selected coefficients (columns and t-statistics in parentheses):
  - Coefficients row (first block): -0.999**  -0.995**  -1.136**  -0.998  -1.205*  -0.659  -0.756  -0.146  -0.769  -1.356*  -1.207  -0.871
    - t-statistics: (-1.98) (-2.15) (-2.06) (-1.56) (-1.76) (-1.01) (-1.19) (-0.18) (-1.19) (-1.94) (-1.46) (-1.14)
  - Coefficients row (second block): -0.502 0.342 -0.648 -1.467 -0.459 -0.930 0.509 -0.783 0.352 0.329 0.373 0.574
    - t-statistics: (-0.82) (0.39) (-0.85) (-1.63) (-0.76) (-1.10) (0.65) (-0.75) (0.43) (0.29) (0.47) (0.70)
  - Trade to GDP coefficients: 0.457 1.329 1.007 -0.275 0.616 -0.464 2.575*** 2.322** 2.327*** 2.177** 2.843*** 2.881***
    - t-statistics: (0.63) (1.43) (1.43) (-0.21) (0.94) (-0.41) (2.99) (2.53) (2.73) (2.01) (3.51) (2.62)
  - Government Consumption to GDP coefficients: 0.336 -0.276 -0.066 1.065 0.444 -0.431 -2.337*** -2.298*** -2.710*** -2.524*** -2.336*** -2.590***
    - t-statistics: (0.60) (-0.43) (-0.10) (0.98) (0.82) (-0.65) (-4.10) (-3.94) (-4.21) (-2.98) (-3.99) (-3.71)
  - Years of Schooling coefficients: 2.148* 3.312** 2.639** 3.597* 1.620 2.841* -0.583 -0.212 0.153 1.454 -0.830 -0.606
    - t-statistics: (1.73) (2.04) (1.99) (1.68) (1.53) (1.82) (-0.40) (-0.15) (0.11) (0.76) (-0.59) (-0.36)
  - Lag 1 of TFP Growth coefficients: 0.110* 0.098 0.078 0.079 0.102* 0.032
    - t-statistics: (1.95) (1.56) (1.40) (1.30) (1.70) (0.47)
  - Lag 1 of Capital Stock Growth coefficients (where reported): 0.663*** 0.617*** 0.647*** 0.746*** 0.637*** 0.582***
    - t-statistics: (7.20) (6.44) (7.38) (8.63) (8.14) (5.22)
  - Lag 2 of Capital Stock Growth coefficients (where reported): -0.097* -0.074 -0.054 -0.089 -0.082 -0.089
    - t-statistics: (-1.71) (-1.21) (-1.00) (-1.54) (-1.38) (-1.58)
- Panel and diagnostic counts (reported across columns):
  - Observations: 724 724 705 547 724 724 824 824 803 625 824 824
  - Number of clusters: 109.000 109.000 109.000 97.000 109.000 109.000 129.000 129.000 129.000 113.000 129.000 129.000
  - Number of instruments: 73.000 93.000 93.000 82.000 93.000 62.000 73.000 93.000 93.000 83.000 93.000 63.000
  - Hansen test p-value: 0.181 0.487 0.203 0.302 0.484 0.127 0.042 0.053 0.114 0.060 0.224 0.025
  - A-B AR(1) test p-value: 0.000 (reported for all columns)
  - A-B AR(2) test p-value: 0.379 0.391 0.372 0.867 0.458 0.730 0.921 0.757 0.478 0.428 0.999 0.759
  - Lowest number of observations included: 1.000 (reported for all columns)
  - Highest number of observations included: 10.000 10.000 10.000 7.000 10.000 10.000 9.000 9.000 9.000 7.000 9.000 9.000
  - Average number of observations included: 6.642 6.642 6.468 5.639 6.642 6.642 6.388 6.388 6.225 5.531 6.388 6.388
- Model notes:
  - Dummy variables for time periods were also included in each model. Results for these are not reported here.
  - t-statistics in parentheses based on cluster-robust standard errors
  - Significance legend: * p<0.1  ** p<0.05  *** p<0.01

### Table 4: Private Credit — Key Coefficients, Diagnostics, and Sample Sizes
- Interaction and main-variable coefficient examples (columns and t-statistics):
  - First coefficient block: -0.194 -0.470 -0.513 -0.061 0.065 0.039 -0.903 -0.374 -1.117 -1.542 -1.025 -1.117
    - t-statistics: (-0.30) (-0.73) (-0.75) (-0.09) (0.10) (0.05) (-0.87) (-0.37) (-1.25) (-1.36) (-0.97) (-1.20)
  - Second coefficient block: -1.142* -0.367 -0.940 -1.947* -1.146** -1.648* 0.081 -0.588 -0.162 -0.455 -0.000 0.218
    - t-statistics: (-1.98) (-0.44) (-1.59) (-1.92) (-2.21) (-1.79) (0.10) (-0.70) (-0.20) (-0.40) (-0.00) (0.28)
  - Trade to GDP: 0.396 1.270 0.852 0.339 0.614 -0.267 2.466*** 2.264*** 2.050** 2.187* 2.872*** 2.533**
    - t-statistics: (0.56) (1.43) (1.20) (0.25) (0.81) (-0.20) (2.91) (2.78) (2.57) (1.93) (3.74) (2.37)
  - Government Consumption to GDP: 0.657 -0.022 0.124 0.452 0.693 -0.018 -2.115*** -1.964*** -2.505*** -2.463*** -2.211*** -2.287***
    - t-statistics: (1.17) (-0.04) (0.20) (0.46) (1.36) (-0.03) (-3.30) (-3.16) (-3.39) (-2.94) (-3.81) (-3.51)
  - Years of Schooling: 2.400* 3.448** 2.344* 3.189 1.813* 3.091* 0.533 0.481 1.537 3.210 0.340 0.347
    - t-statistics: (1.89) (2.31) (1.70) (1.40) (1.69) (1.67) (0.35) (0.32) (0.97) (1.55) (0.27) (0.19)
  - Lag 1 of TFP Growth: 0.092 0.075 0.078 0.042 0.104* -0.019
    - t-statistics: (1.62) (1.34) (1.33) (0.67) (1.84) (-0.30)
  - Lag 1 of Capital Stock Growth (where reported): 0.650*** 0.620*** 0.609*** 0.715*** 0.621*** 0.609***
    - t-statistics: (7.27) (7.31) (7.40) (7.54) (8.03) (5.74)
  - Lag 2 of Capital Stock Growth: -0.085 -0.081 -0.042 -0.088 -0.085 -0.092
    - t-statistics: (-1.45) (-1.41) (-0.74) (-1.48) (-1.49) (-1.49)
- Panel and diagnostic counts:
  - Observations: 729 729 710 548 717 729 827 827 806 627 817 827
  - Number of clusters: 108.000 108.000 108.000 96.000 108.000 108.000 128.000 128.000 128.000 112.000 128.000 128.000
  - Number of instruments: 73.000 93.000 93.000 82.000 93.000 62.000 73.000 93.000 93.000 83.000 93.000 64.000
  - Hansen test p-value: 0.111 0.346 0.158 0.224 0.410 0.122 0.018 0.056 0.080 0.041 0.223 0.011
  - A-B AR(1) test p-value: 0.000 (reported for all columns)
  - A-B AR(2) test p-value: 0.433 0.412 0.377 0.954 0.376 0.848 0.749 0.639 0.508 0.374 0.916 0.787
  - Lowest number of observations included: 1.000 (reported for all columns)
  - Highest number of observations included: 10.000 10.000 10.000 7.000 10.000 10.000 9.000 9.000 9.000 7.000 9.000 9.000
  - Average number of observations included: 6.750 6.750 6.574 5.708 6.639 6.750 6.461 6.461 6.297 5.598 6.383 6.461
- Model notes:
  - Dummy variables for time periods were also included in each model. Results for these are not reported here.
  - t-statistics in parentheses based on cluster-robust standard errors
  - Significance legend: * p<0.1  ** p<0.05  *** p<0.01

### Table 5: Liquid Liabilities — Key Coefficients, Diagnostics, and Sample Sizes
- Selected coefficient blocks and diagnostics:
  - First coefficient block: 0.454 1.341** 0.823 0.090 0.228 0.165 0.761 1.352** 0.528 1.009 0.860 0.710
    - t-statistics: (0.53) (1.99) (1.17) (0.16) (0.35) (0.24) (1.53) (2.31) (0.78) (1.55) (1.59) (0.97)
  - Second coefficient block: -2.549 -4.143** -3.516** -2.052 -1.087 -2.471* -0.682 -1.928 -1.307 -2.829 -2.241* -0.364
    - t-statistics: (-1.53) (-2.21) (-2.08) (-1.45) (-0.96) (-1.75) (-0.68) (-1.23) (-0.92) (-1.52) (-1.66) (-0.33)
  - Trade to GDP: 4.180 4.995*** 2.865 3.324* 5.427** 4.009* 3.835* 3.861** 0.646 1.512 4.533*** 4.606**
    - t-statistics: (1.63) (2.83) (1.29) (1.93) (2.62) (1.81) (1.97) (2.55) (0.27) (0.85) (3.11) (2.04)
  - Government Consumption to GDP: 2.086 3.824 -0.516 -0.136 0.520 0.240 -6.044*** -5.365*** -7.562*** -8.008*** -6.495*** -6.681**
    - t-statistics: (0.49) (1.00) (-0.13) (-0.05) (0.15) (0.09) (-3.04) (-3.18) (-2.93) (-2.80) (-4.13) (-2.33)
  - Years of Schooling: 2.408 2.888 6.752* 4.819 0.195 3.087 1.860 1.454 7.715** 8.950** 1.603 1.342
    - t-statistics: (0.69) (1.13) (1.96) (1.26) (0.07) (0.88) (0.56) (0.50) (2.15) (2.13) (0.51) (0.35)
  - Lag 1 of TFP Growth: -0.046 -0.032 -0.035 -0.049 -0.014 -0.048
    - t-statistics: (-0.72) (-0.46) (-0.64) (-0.77) (-0.24) (-0.83)
  - Lag 1 of Capital Stock Growth (where reported): 0.387*** 0.385*** 0.381*** 0.391*** 0.375*** 0.413***
    - t-statistics: (7.51) (8.19) (6.31) (5.72) (6.53) (10.01)
  - Lag 2 of Capital Stock Growth: 0.003 0.009 0.040 -0.008 -0.007 0.023
    - t-statistics: (0.05) (0.15) (0.78) (-0.12) (-0.13) (0.38)
- Panel and diagnostic counts:
  - Observations: 589 589 574 558 585 589 617 617 603 580 614 617
  - Number of clusters: 92.000 92.000 92.000 87.000 92.000 92.000 100.000 100.000 100.000 94.000 100.000 100.000
  - Number of instruments: 44.000 56.000 56.000 56.000 56.000 60.000 47.000 59.000 59.000 59.000 59.000 63.000
  - Hansen test p-value: 0.154 0.291 0.225 0.269 0.346 0.569 0.660 0.651 0.711 0.580 0.854 0.419
  - A-B AR(1) test p-value: 0.003 0.003 0.002 0.003 0.002 0.002 0.002 0.001 0.002 0.002 0.002 0.001
  - A-B AR(2) test p-value: 0.252 0.410 0.038 0.217 0.207 0.213 0.265 0.271 0.359 0.174 0.262 0.349
  - Lowest number of observations included: 1.000 (reported for all columns)
  - Highest number of observations included: 8.000 (reported for all columns)
  - Average number of observations included: 6.402 6.402 6.239 6.414 6.359 6.402 6.170 6.170 6.030 6.170 6.140 6.170
- Model notes:
  - Dummy variables for time periods were also included in each model. Results for these are not reported here.
  - t-statistics in parentheses based on cluster-robust standard errors
  - Significance legend: * p<0.1  ** p<0.05  *** p<0.01

*Source: wp17143 - Section 3*

### Section 4

### wp17143 - Section 4

### Table 6: Stock Market Capitalization
- Coefficients for "Stock Market Capitalization" across models (columns (1)–(6) first panel; (1)–(6) second panel):
  - -3.725** (-2.45); -1.794 (-1.36); -2.230 (-0.98); -1.962 (-1.07); -2.730* (-1.92); -4.974*** (-2.92)
  - -2.847 (-1.43); -2.069 (-1.18); -2.196 (-0.95); -4.616** (-2.00); -2.337 (-1.42); -0.643 (-0.33)
- Coefficients for other key regressors (selected):
  - Trade to GDP: 1.073 (0.49); 1.012 (0.54); -0.129 (-0.03); 3.680 (1.50); 0.306 (0.10); 0.227 (0.11)
  - Government Consumption to GDP: -2.614 (-0.97); 0.510 (0.20); -1.190 (-0.45); -2.664 (-0.60); -2.403 (-1.13); -2.254 (-0.64)
  - Years of Schooling: 10.214*** (2.69); 7.325*** (2.71); 10.179*** (2.64); 11.676** (2.45); 6.876*** (2.87); 13.957*** (3.31)
  - Lag 1 of TFP Growth: -0.021 (-0.41); -0.049 (-0.93); -0.021 (-0.33); -0.045 (-0.69); -0.004 (-0.07); -0.024 (-0.55)
  - Lag 1 of Capital Stock Growth (selected columns): 0.348*** (3.17); 0.387*** (4.01); 0.321*** (2.98); 0.245** (2.30); 0.341*** (3.96); 0.375*** (3.91)
  - Lag 2 of Capital Stock Growth (selected): 0.029 (0.66); 0.038 (0.97); 0.065 (1.25); -0.029 (-0.54); 0.030 (0.73); 0.037 (0.89)
- Sample, instruments, and diagnostics:
  - Observations: 773; 773; 747; 611; 715; 773; 911; 911; 880; 689; 831; 911
  - Number of clusters: 102.000; 102.000; 102.000; 90.000; 102.000; 102.000; 120.000; 120.000; 120.000; 105.000; 120.000; 120.000
  - Number of instruments: 53.000; 67.000; 67.000; 65.000; 67.000; 72.000; 56.000; 70.000; 70.000; 68.000; 70.000; 75.000
  - Hansen test p-value: 0.694; 0.432; 0.437; 0.150; 0.436; 0.852; 0.040; 0.080; 0.168; 0.153; 0.408; 0.089
  - A-B AR(1) test p-value: 0.018; 0.028; 0.027; 0.050; 0.039; 0.013; 0.003; 0.001; 0.006; 0.014; 0.001; 0.002
  - A-B AR(2) test p-value: 0.308; 0.489; 0.340; 0.445; 0.564; 0.250; 0.279; 0.336; 0.401; 0.954; 0.310; 0.283
  - Lowest number of observations included: 2.000; 2.000; 1.000; 2.000; 1.000; 2.000; 2.000; 2.000; 2.000; 2.000; 1.000; 2.000
  - Highest number of observations included: 11.000; 11.000; 11.000; 9.000; 11.000; 11.000; 11.000; 11.000; 11.000; 9.000; 11.000; 11.000
  - Average number of observations included: 7.578; 7.578; 7.324; 6.789; 7.010; 7.578; 7.592; 7.592; 7.333; 6.562; 6.925; 7.592
- Notes:
  - Dummy variables for time periods were also included in each model. Results for these are not reported here.
  - t-statistics in parentheses based on cluster-robust standard errors
  - Significance legend: * p<0.1  ** p<0.05  *** p<0.01

### Table 7: Interest Rate Spread
- Coefficients for "Interest Rate Spread" interaction terms (columns (1)–(6) first panel; (1)–(6) second panel):
  - 0.927** (2.05); 0.975** (2.11); 0.850** (2.40); 0.849* (1.74); 0.779 (1.59); 0.676 (1.60)
  - 0.707** (2.16); 1.101** (2.32); 0.824*** (2.75); 0.900* (1.84); 1.165** (2.54); 0.856** (2.24)
- Coefficients for other key regressors (selected):
  - Trade to GDP: 3.424** (2.02); 3.977** (2.03); 2.392 (1.58); 2.046 (1.27); 3.456** (2.33); 2.770* (1.73)
  - Government Consumption to GDP: 1.827 (0.55); 2.455 (0.79); -1.054 (-0.32); -2.193 (-0.68); -0.609 (-0.21); -0.473 (-0.17)
  - Years of Schooling: 3.888 (0.84); 3.265 (0.94); 5.527 (1.36); 8.443* (1.95); 4.617 (1.24); 6.786 (1.43)
  - Lag 1 of TFP Growth: -0.035 (-0.53); -0.031 (-0.42); -0.024 (-0.39); -0.061 (-0.94); -0.062 (-0.83); -0.038 (-0.73)
  - Lag 1 of Capital Stock Growth (selected): 0.396*** (8.05); 0.375*** (7.57); 0.363*** (6.89); 0.414*** (7.89); 0.365*** (6.60); 0.399*** (8.24)
  - Lag 2 of Capital Stock Growth (selected): -0.020 (-0.34); -0.004 (-0.10); 0.018 (0.35); -0.011 (-0.15); -0.019 (-0.30); 0.002 (0.03)
- Sample, instruments, and diagnostics:
  - Observations: 581; 581; 566; 552; 577; 581; 608; 608; 594; 573; 605; 608
  - Number of clusters: 91.000; 91.000; 91.000; 86.000; 91.000; 91.000; 99.000; 99.000; 99.000; 93.000; 99.000; 99.000
  - Number of instruments: 44.000; 56.000; 56.000; 56.000; 56.000; 60.000; 47.000; 59.000; 59.000; 59.000; 59.000; 63.000
  - Hansen test p-value: 0.172; 0.101; 0.419; 0.383; 0.159; 0.614; 0.465; 0.081; 0.591; 0.421; 0.208; 0.392
  - A-B AR(1) test p-value: 0.003; 0.005; 0.004; 0.002; 0.005; 0.002; 0.002; 0.001; 0.001; 0.002; 0.002; 0.002
  - A-B AR(2) test p-value: 0.353; 0.418; 0.052; 0.347; 0.223; 0.296; 0.134; 0.198; 0.377; 0.201; 0.219; 0.270
  - Lowest number of observations included: 1.000 (for all columns)
  - Highest number of observations included: 8.000 (for all columns)
  - Average number of observations included: 6.385; 6.385; 6.220; 6.419; 6.341; 6.385; 6.141; 6.141; 6.000; 6.161; 6.111; 6.141
- Notes:
  - Dummy variables for time periods were also included in each model. Results for these are not reported here.
  - t-statistics in parentheses based on cluster-robust standard errors
  - Significance legend: * p<0.1  ** p<0.05  *** p<0.01

### Table 8: Stock Market Turnover
- Coefficients for "Stock Market Turnover" interaction terms (columns (1)–(6) first panel; (1)–(6) second panel):
  - -0.420 (-0.24); -1.012 (-0.60); -1.437 (-0.83); 0.767 (0.53); -3.865** (-2.10); -0.753 (-0.24)
  - -0.857 (-0.74); -0.052 (-0.04); -1.721 (-1.42); 0.384 (0.38); -2.191* (-1.92); -1.690 (-0.84)
- Coefficients for other key regressors (selected):
  - Trade to GDP: 5.766** (2.15); 4.965* (1.96); 5.580** (2.46); 7.448*** (3.03); 2.155 (1.32); 6.976* (1.69)
  - Government Consumption to GDP: -1.277 (-0.47); -3.077 (-0.93); 0.577 (0.25); 0.595 (0.17); -0.048 (-0.02); -5.690 (-0.97)
  - Years of Schooling: 4.328 (1.55); 2.320 (0.86); 3.902 (1.38); 2.762 (0.59); 3.113* (1.67); 1.445 (0.27)
  - Lag 1 of TFP Growth: 0.077 (1.25); 0.145** (2.21); 0.095 (1.51); 0.108 (1.54); 0.147** (2.10); 0.136 (1.63)
  - Lag 1 of Capital Stock Growth (selected): 0.745*** (6.92); 0.739*** (7.03); 0.715*** (5.83); 0.730*** (4.94); 0.735*** (7.35); 0.705*** (5.36)
  - Lag 2 of Capital Stock Growth (selected): -0.076 (-0.51); -0.072 (-0.53); -0.102 (-0.60); -0.106 (-0.57); -0.078 (-0.53); -0.110 (-0.89)
- Sample, instruments, and diagnostics:
  - Observations: 423; 423; 414; 378; 403; 423; 496; 496; 484; 438; 475; 496
  - Number of clusters: 108.000; 108.000; 108.000; 96.000; 107.000; 108.000; 128.000; 128.000; 128.000; 112.000; 127.000; 128.000
  - Number of instruments: 90.000; 96.000; 96.000; 96.000; 96.000; 46.000; 91.000; 97.000; 97.000; 97.000; 97.000; 49.000
  - Hansen test p-value: 0.292; 0.357; 0.372; 0.393; 0.248; 0.249; 0.239; 0.332; 0.361; 0.158; 0.739; 0.156
  - A-B AR(1) test p-value: 0.020; 0.015; 0.007; 0.017; 0.003; 0.033; 0.059; 0.052; 0.069; 0.142; 0.045; 0.053
  - A-B AR(2) test p-value: 0.059; 0.138; 0.076; 0.114; 0.096; 0.105; 0.776; 0.741; 0.801; 0.447; 0.646; 0.842
  - Lowest number of observations included: 1.000 (for all columns)
  - Highest number of observations included: 4.000 (for all columns)
  - Average number of observations included: 3.917; 3.917; 3.833; 3.938; 3.766; 3.917; 3.875; 3.875; 3.781; 3.911; 3.740; 3.875
- Notes:
  - Dummy variables for time periods were also included in each model. Results for these are not reported here.
  - t-statistics in parentheses based on cluster-robust standard errors
  - Significance legend: * p<0.1  ** p<0.05  *** p<0.01

*Source: wp17143 - Section 4 (pdf).*

### Section 5

### Section 5

### Net Interest Margin (Table 9)
- Coefficients and t-statistics (selected, by column (1)–(6) and (1)–(6) for two panels as presented):
  - Net Interest Margin*MENAPOX Dummy: -0.793 (-0.32), -1.323 (-0.62), 3.215 (1.48), 1.382 (0.42), 1.529 (0.88), -3.813 (-0.95)
  - Net Interest Margin*MENAPOI Dummy: 0.245 (0.36), -0.099 (-0.14), 0.149 (0.16), -0.541 (-0.66), 0.128 (0.13), 0.520 (0.53)
  - Net Interest Margin*Medium Financial Depth: -1.471 (-0.78), -3.669 (-1.19), 0.052 (0.02), -3.799 (-1.40), -0.430 (-0.27), -11.358*** (-3.39)
  - Net Interest Margin*Low Financial Depth: -0.765 (-1.12), -0.985 (-1.43), -0.651 (-1.25), -0.918 (-1.08), -1.032 (-1.33), -1.412* (-1.96)
  - Trade to GDP: 9.218** (2.14), 9.027* (1.87), 5.725** (2.10), 7.606 (1.54), 3.194* (1.82), 4.610 (0.80); 2.357*** (3.36), 2.580*** (3.39), 2.735*** (2.79), 1.824** (2.61), 1.600** (2.26), 2.712*** (2.96)
  - Government Consumption to GDP: 15.144** (2.07), 9.876 (1.32), 2.169 (0.57), 21.667** (2.36), -1.498 (-0.67), 15.233*** (2.63); 0.034 (0.03), -0.415 (-0.42), 0.115 (0.10), -0.527 (-0.58), -0.709 (-0.52), -0.272 (-0.27)
  - Years of Schooling: -0.122 (-0.02), 2.244 (0.41), 2.761 (0.75), 3.656 (0.36), 4.072 (1.47), 15.119 (1.65); 0.662 (0.31), 0.987 (0.48), -0.661 (-0.43), 3.206 (1.25), 1.487 (0.78), 2.687 (0.95)
  - Lag 1 of TFP Growth: 0.053 (0.38), -0.007 (-0.04), 0.216** (2.13), 0.012 (0.07), 0.209** (2.53), 0.039 (0.25)
  - Lag 1 of Capital Stock Growth: 0.765*** (5.40), 0.775*** (6.66), 0.755*** (7.26), 0.782*** (4.95), 0.719*** (5.60), 0.761*** (7.71)
  - Lag 2 of Capital Stock Growth: -0.036 (-0.77), -0.009 (-0.17), -0.034 (-0.67), 0.019 (0.34), 0.026 (0.54), -0.011 (-0.23)
- Model diagnostics and sample:
  - Observations: 1252, 1252, 1234, 1123, 1162, 1252; 1466, 1466, 1441, 1302, 1367, 1466
  - Number of clusters: 108.000, 108.000, 108.000, 96.000, 107.000, 108.000; 128.000, 128.000, 128.000, 112.000, 127.000, 128.000
  - Number of instruments: 54.000, 68.000, 68.000, 68.000, 68.000, 75.000; 57.000, 71.000, 71.000, 71.000, 71.000, 78.000
  - Hansen test p-value: 0.138, 0.045, 0.117, 0.097, 0.044, 0.334; 0.315, 0.253, 0.383, 0.073, 0.041, 0.194
  - A-B AR(1) test p-value: 0.003, 0.055, 0.000, 0.012, 0.000, 0.009; 0.011, 0.013, 0.011, 0.029, 0.032, 0.010
  - A-B AR(2) test p-value: 0.824, 0.598, 0.585, 0.904, 0.302, 0.909; 0.982, 0.905, 0.663, 0.599, 0.886, 0.950
  - Lowest number of observations included: 1.000 (all columns)
  - Highest number of observations included: 12.000 (all columns)
  - Average number of observations included: 11.593, 11.593, 11.426, 11.698, 10.860, 11.593; 11.453, 11.453, 11.258, 11.625, 10.764, 11.453
- Notes:
  - Dummy variables for time periods were also included in each model. Results for these are not reported here.
  - t-statistics in parentheses based on cluster-robust standard errors. Significance: * p<0.1, ** p<0.05, *** p<0.01.

### Bank Overhead Costs (Table 10)
- Key coefficients and t-statistics (selected):
  - Bank Overhead Costs*CCA Dummy: -1.619* (-1.96), -1.376** (-2.14), -2.411** (-2.24), -1.759* (-1.93), -2.911*** (-3.60), -1.770** (-2.13)
  - Bank Overhead Costs*MENAPOI Dummy: -1.398*** (-3.14), -0.752 (-1.60), -2.159** (-2.47), -1.077** (-2.18), -1.256*** (-3.12), -1.191** (-2.25)
  - Bank Overhead Costs*MENAPOX Dummy: -2.753** (-2.43), -2.719** (-2.58), -3.227*** (-3.80), -4.425*** (-2.81), -1.626** (-2.39), -3.254** (-2.51)
  - Trade to GDP: 1.837 (0.57), 2.621 (0.89), 0.414 (0.18), 2.967 (0.91), 1.493 (1.33), 3.837 (0.85); 0.010 (0.00), -1.017 (-0.71), -0.142 (-0.06), 1.646 (1.41), 0.641 (0.42), 0.228 (0.11)
  - Government Consumption to GDP: 0.946 (0.24), 0.880 (0.32), 1.899 (0.70), -0.546 (-0.13), -0.062 (-0.03), 1.296 (0.31); -4.169** (-2.00), -2.642 (-1.12), -1.155 (-0.55), -1.494 (-0.85), -4.057* (-1.81), -5.132*** (-2.75)
  - Years of Schooling: 3.405 (0.85), 1.844 (0.50), 3.173 (1.02), 3.583 (0.69), 2.210 (1.13), 3.209 (0.52); 5.235* (1.68), 3.995 (1.50), 3.610 (1.37), 2.152 (1.01), 5.583* (1.82), 6.374* (1.75)
  - Lag 1 of TFP Growth: -0.056 (-0.75), -0.069 (-0.94), -0.033 (-0.45), -0.123 (-1.59), -0.042 (-0.53), -0.105 (-1.35)
  - Lag 1 of Capital Stock Growth: 0.613*** (3.22), 0.610*** (3.72), 0.558** (2.56), 0.694*** (6.48), 0.602*** (3.84), 0.662*** (4.25)
  - Lag 2 of Capital Stock Growth: -0.027 (-0.39), -0.031 (-0.49), 0.015 (0.18), -0.056 (-1.18), -0.030 (-0.54), -0.049 (-0.75)
- Model diagnostics and sample:
  - Observations: 397, 397, 389, 378, 382, 397; 410, 410, 403, 391, 395, 410
  - Number of clusters: 90.000, 90.000, 90.000, 84.000, 90.000, 90.000; 95.000, 95.000, 95.000, 89.000, 95.000, 95.000
  - Number of instruments: 44.000, 52.000, 52.000, 52.000, 52.000, 56.000; 47.000, 55.000, 55.000, 55.000, 55.000, 59.000
  - Hansen test p-value: 0.121, 0.252, 0.082, 0.203, 0.255, 0.386; 0.355, 0.593, 0.401, 0.196, 0.583, 0.622
  - A-B AR(1) test p-value: 0.116, 0.147, 0.062, 0.153, 0.064, 0.232; 0.069, 0.040, 0.070, 0.004, 0.051, 0.032
  - A-B AR(2) test p-value: 0.133, 0.174, 0.105, 0.100, 0.062, 0.073; 0.675, 0.553, 0.229, 0.313, 0.536, 0.698
  - Lowest number of observations included: 1.000 (all columns)
  - Highest number of observations included: 5.000 (all columns)
  - Average number of observations included: 4.411, 4.411, 4.322, 4.500, 4.244, 4.411; 4.316, 4.316, 4.242, 4.393, 4.158, 4.316
- Notes:
  - Dummy variables for time periods were also included in each model. Results for these are not reported here.
  - t-statistics in parentheses based on cluster-robust standard errors. Significance: * p<0.1, ** p<0.05, *** p<0.01.

### Non-Performing Loans (Table 11)
- Key coefficients and t-statistics (selected):
  - Non-Performing Loans*MENAPOX Dummy: 1.735 (0.60), -0.442 (-0.17), 2.391 (0.65), 0.088 (0.04), -1.524 (-0.47), 2.088 (1.04)
  - Non-Performing Loans*MENAPOI Dummy: 0.309 (0.08), -0.435 (-0.24), -0.952 (-0.31), -0.288 (-0.17), -0.651 (-0.27), -1.764 (-1.15)
  - Non-Performing Loans*Medium Financial Depth: -1.656 (-1.41), -2.366** (-2.24), -2.833* (-1.78), 0.128 (0.07), -2.040 (-1.06), -2.361 (-1.37)
  - Non-Performing Loans*Low Financial Depth: 0.296 (0.11), -0.494 (-0.25), 0.148 (0.09), 0.244 (0.17), 0.950 (0.63), -0.797 (-0.43)
  - Trade to GDP: 1.339 (0.73), 0.040 (0.04), 1.478 (0.74), 2.987* (1.67), 1.603 (0.79), 2.036 (0.85); 3.743 (1.25), 3.821** (2.21), 2.131 (0.79), 2.600 (1.34), 5.540** (2.23), 2.431 (1.38)
  - Government Consumption to GDP: 2.944 (0.84), 2.170 (0.48), 1.876 (0.59), 1.880 (0.46), 0.198 (0.06), 2.378 (0.70); -7.031 (-1.26), -6.675* (-1.73), -6.741 (-1.35), -3.673 (-1.14), -8.038* (-1.77), -9.491*** (-3.42)
  - Years of Schooling: 5.131* (1.68), 5.775 (1.65), 7.359* (1.90), -0.762 (-0.21), 7.071 (1.52), 6.793* (1.75); 2.954 (0.42), 5.027 (1.03), -0.278 (-0.07), 2.309 (0.80), 0.010 (0.00), 5.184 (1.44)
  - Lag 1 of TFP Growth: 0.039 (0.29), 0.034 (0.34), 0.069 (0.64), 0.069 (0.99), 0.028 (0.18), 0.018 (0.15)
  - Lag 1 of Capital Stock Growth: 0.370*** (4.39), 0.366*** (5.77), 0.338*** (3.93), 0.382*** (5.13), 0.366*** (4.60), 0.386*** (4.32)
  - Lag 2 of Capital Stock Growth: 0.003 (0.04), -0.001 (-0.02), 0.006 (0.11), -0.022 (-0.36), -0.017 (-0.22), 0.028 (0.45)
- Model diagnostics and sample:
  - Observations: 435, 435, 422, 410, 421, 435; 436, 436, 423, 413, 422, 436
  - Number of clusters: 73.000, 73.000, 73.000, 72.000, 73.000, 73.000; 75.000, 75.000, 75.000, 74.000, 75.000, 75.000
  - Number of instruments: 40.000, 48.000, 48.000, 41.000, 48.000, 51.000; 42.000, 50.000, 50.000, 44.000, 50.000, 53.000
  - Hansen test p-value: 0.198, 0.169, 0.456, 0.009, 0.193, 0.573; 0.207, 0.271, 0.639, 0.387, 0.359, 0.703
  - A-B AR(1) test p-value: 0.075, 0.011, 0.082, 0.019, 0.054, 0.079; 0.038, 0.035, 0.041, 0.046, 0.032, 0.030
  - A-B AR(2) test p-value: 0.770, 0.993, 0.717, 0.524, 0.922, 0.624; 0.396, 0.212, 0.292, 0.089, 0.147, 0.746
  - Lowest number of observations included: 1.000 (all columns)
  - Highest number of observations included: 16.000, 16.000, 16.000, 9.000, 16.000, 16.000; 15.000, 15.000, 15.000, 9.000, 15.000, 15.000
  - Average number of observations included: 5.959, 5.959, 5.781, 5.694, 5.767, 5.959; 5.813, 5.813, 5.640, 5.581, 5.627, 5.813
- Notes:
  - Dummy variables for time periods were also included in each model. Results for these are not reported here.
  - t-statistics in parentheses based on cluster-robust standard errors. Significance: * p<0.1, ** p<0.05, *** p<0.01.

*Source: wp17143 - Section 5*

### Section 6

### wp17143 - Section 6

### Stock Price Volatility (Table 12)
- Reported coefficient estimates (selected, by column grouping):
  - Constant / baseline terms:
    - 1.176 (1.08)
    - 0.489 (0.43)
    - 0.418 (0.43)
    - 1.634 (1.15)
    - -0.468 (-0.44)
    - 0.683 (0.63)
    - 0.510 (0.72)
    - 0.740 (1.00)
    - -0.209 (-0.25)
    - 1.037 (1.20)
    - -0.549 (-0.74)
    - 0.361 (0.28)
  - Capital (coefficients with significance):
    - -3.862*** (-2.94)
    - -2.757** (-2.43)
    - -3.027*** (-3.40)
    - -5.365*** (-3.18)
    - -1.610* (-1.80)
    - -4.668* (-1.94)
    - -0.898 (-0.70)
    - -1.053 (-1.26)
    - -0.005 (-0.01)
    - -0.726 (-0.59)
    - -0.709 (-0.55)
    - -0.019 (-0.02)
  - Trade to GDP:
    - 7.460*** (3.01)
    - 7.728*** (2.87)
    - 6.225** (2.57)
    - 7.876** (2.54)
    - 4.321** (2.10)
    - 11.216*** (3.87)
    - 1.943 (1.58)
    - 2.013* (1.79)
    - 1.271 (0.70)
    - 1.628 (1.29)
    - 2.342* (1.98)
    - 3.089** (2.16)
  - Government Consumption to GDP:
    - -1.213 (-0.46)
    - -1.957 (-0.70)
    - -1.560 (-0.62)
    - 0.479 (0.14)
    - 0.298 (0.14)
    - -1.336 (-0.32)
    - -3.169* (-1.81)
    - -3.158** (-2.21)
    - -4.027* (-1.77)
    - -1.247 (-0.78)
    - -3.240*** (-2.85)
    - -0.931 (-0.59)
  - Years of Schooling:
    - 2.235 (0.62)
    - 1.195 (0.36)
    - 1.678 (0.53)
    - 1.720 (0.31)
    - 1.141 (0.54)
    - 2.954 (0.44)
    - 2.019 (0.54)
    - 1.000 (0.42)
    - 1.590 (0.52)
    - 1.157 (0.30)
    - 1.902 (0.54)
    - -1.753 (-0.57)
  - Lag 1 of TFP Growth:
    - 0.117** (2.01)
    - 0.164** (2.47)
    - 0.135** (2.12)
    - 0.124* (1.79)
    - 0.153** (2.23)
    - 0.189* (1.87)
  - Lag 1 of Capital Stock Growth (selected):
    - 0.731*** (6.26)
    - 0.744*** (7.17)
    - 0.677*** (4.89)
    - 0.757*** (5.76)
    - 0.673*** (5.22)
    - 0.782*** (7.83)
  - Lag 2 of Capital Stock Growth (selected):
    - -0.065 (-0.43)
    - -0.042 (-0.27)
    - -0.083 (-0.44)
    - -0.088 (-0.49)
    - -0.061 (-0.39)
    - -0.047 (-0.33)
- Model diagnostics and sample sizes:
  - Observations: 419, 419, 410, 374, 399, 419, 489, 489, 477, 433, 468, 489
  - Number of clusters: 107.000, 107.000, 107.000, 96.000, 106.000, 107.000, 125.000, 125.000, 125.000, 111.000, 124.000, 125.000
  - Number of instruments: 90.000, 96.000, 96.000, 96.000, 96.000, 46.000, 91.000, 97.000, 97.000, 97.000, 97.000, 49.000
  - Hansen test p-value: 0.365, 0.254, 0.425, 0.376, 0.301, 0.834, 0.167, 0.331, 0.229, 0.250, 0.582, 0.172
  - A-B AR(1) test p-value: 0.004, 0.006, 0.002, 0.006, 0.006, 0.033, 0.077, 0.075, 0.088, 0.144, 0.082, 0.073
  - A-B AR(2) test p-value: 0.079, 0.111, 0.093, 0.168, 0.077, 0.106, 0.484, 0.524, 0.614, 0.301, 1.000, 0.600
  - Lowest number of observations included: 1.000 (all models)
  - Highest number of observations included: 4.000 (all models)
  - Average number of observations included: 3.916, 3.916, 3.832, 3.896, 3.764, 3.916, 3.912, 3.912, 3.816, 3.901, 3.774, 3.912

### Bank Z-Score (Table 13)
- Reported coefficient estimates (selected, by column grouping):
  - Constant / baseline terms:
    - -0.544 (-0.76)
    - -1.167* (-1.73)
    - -0.362 (-0.48)
    - -1.259 (-1.25)
    - -0.816 (-0.93)
    - -0.395 (-0.47)
    - -0.125 (-0.24)
    - 0.246 (0.39)
    - 0.290 (0.51)
    - -0.318 (-0.48)
    - -0.995 (-1.65)
    - 0.207 (0.30)
  - Capital (coefficients with significance):
    - -1.820*** (-3.37)
    - -0.392 (-0.52)
    - -1.714*** (-2.88)
    - -2.851*** (-2.91)
    - -0.970* (-1.94)
    - -2.071*** (-3.32)
    - -0.669 (-1.03)
    - -1.121 (-1.58)
    - -1.070 (-1.59)
    - -0.565 (-0.65)
    - -0.289 (-0.65)
    - -0.732 (-1.03)
  - Trade to GDP:
    - 0.103 (0.12)
    - 1.162 (1.44)
    - 0.351 (0.43)
    - -0.033 (-0.03)
    - 0.724 (0.93)
    - -0.838 (-0.84)
    - 2.474*** (3.10)
    - 2.463*** (3.13)
    - 2.484*** (3.16)
    - 2.705** (2.56)
    - 2.512*** (3.09)
    - 2.720*** (3.26)
  - Government Consumption to GDP:
    - 0.398 (0.82)
    - -0.073 (-0.15)
    - 0.265 (0.53)
    - 0.100 (0.08)
    - 0.395 (0.79)
    - -0.425 (-0.74)
    - -2.161*** (-3.27)
    - -1.709** (-2.49)
    - -2.338*** (-3.45)
    - -3.027*** (-4.06)
    - -2.081*** (-3.65)
    - -2.213*** (-2.98)
  - Years of Schooling:
    - 3.825** (2.42)
    - 3.244** (2.20)
    - 3.227** (2.19)
    - 5.213** (2.59)
    - 2.092* (1.84)
    - 4.680*** (2.77)
    - 1.492 (0.87)
    - 0.712 (0.50)
    - 2.193 (1.41)
    - 1.805 (0.87)
    - -0.562 (-0.45)
    - 0.575 (0.36)
  - Lag 1 of TFP Growth:
    - 0.030 (0.56)
    - 0.049 (0.75)
    - 0.058 (1.10)
    - -0.021 (-0.32)
    - 0.081 (1.39)
    - -0.026 (-0.45)
  - Lag 1 of Capital Stock Growth (selected):
    - 0.583*** (7.28)
    - 0.589*** (7.24)
    - 0.623*** (8.77)
    - 0.655*** (7.83)
    - 0.616*** (9.20)
    - 0.562*** (6.44)
  - Lag 2 of Capital Stock Growth (selected):
    - -0.067 (-1.15)
    - -0.048 (-0.85)
    - -0.071 (-1.41)
    - -0.111 (-1.56)
    - -0.103* (-1.74)
    - -0.080 (-1.27)
- Model diagnostics and sample sizes:
  - Observations: 755, 755, 735, 564, 716, 755, 879, 879, 856, 649, 816, 879
  - Number of clusters: 108.000, 108.000, 108.000, 96.000, 108.000, 108.000, 128.000, 128.000, 128.000, 112.000, 128.000, 128.000
  - Number of instruments: 73.000, 93.000, 93.000, 82.000, 93.000, 62.000, 73.000, 93.000, 93.000, 83.000, 93.000, 63.000
  - Hansen test p-value: 0.062, 0.204, 0.224, 0.124, 0.217, 0.054, 0.064, 0.095, 0.179, 0.061, 0.350, 0.009
  - A-B AR(1) test p-value: 0.000, 0.000, 0.000, 0.004, 0.000, 0.001, 0.001, 0.001, 0.000, 0.000, 0.000, 0.001
  - A-B AR(2) test p-value: 0.820, 0.808, 0.561, 0.686, 0.407, 0.772, 0.661, 0.796, 0.573, 0.549, 0.911, 0.591
  - Lowest number of observations included: 1.000 (all models)
  - Highest number of observations included: 10.000, 10.000, 10.000, 7.000, 10.000, 10.000, 9.000, 9.000, 9.000, 7.000, 9.000, 9.000
  - Average number of observations included: 6.991, 6.991, 6.806, 5.875, 6.630, 6.991, 6.867, 6.867, 6.688, 5.795, 6.375, 6.867

### Loan-to-Deposit Ratio (Table 14)
- Reported coefficient estimates (selected, by column grouping):
  - Constant / baseline terms:
    - 6.712** (2.20)
    - 2.496 (0.55)
    - 7.199** (2.53)
    - 5.295* (1.78)
    - 2.074 (0.43)
    - 3.879 (1.42)
    - 0.138 (0.09)
    - 0.800 (0.83)
    - -0.248 (-0.15)
    - 0.466 (0.49)
    - -0.363 (-0.24)
    - -0.066 (-0.06)
  - Capital (coefficients with significance and signs):
    - -0.103 (-0.07)
    - -1.026 (-0.77)
    - -0.324 (-0.19)
    - -1.107 (-0.77)
    - -0.680 (-0.44)
    - -0.475 (-0.28)
    - -0.410 (-0.88)
    - -0.340 (-0.70)
    - -0.525 (-0.81)
    - -0.209 (-0.45)
    - -0.323 (-0.58)
    - -0.364 (-0.77)
  - Trade to GDP:
    - 5.281 (1.31)
    - 2.268 (0.67)
    - 4.360 (1.19)
    - 2.730 (0.80)
    - 1.894 (0.71)
    - 5.487* (1.90)
    - -0.253 (-0.28)
    - 0.186 (0.34)
    - 0.141 (0.17)
    - 0.629 (1.25)
    - -0.251 (-0.28)
    - 0.179 (0.30)
  - Government Consumption to GDP:
    - 3.977 (0.88)
    - 4.669 (1.02)
    - 0.783 (0.19)
    - 3.188 (0.78)
    - 1.349 (0.43)
    - -2.852 (-0.78)
    - -2.414 (-1.54)
    - -1.762* (-1.69)
    - -1.578 (-1.48)
    - -1.268* (-1.87)
    - -3.323** (-2.06)
    - -2.785** (-2.54)
  - Years of Schooling:
    - 2.924 (0.84)
    - 6.405 (1.57)
    - 7.394 (1.61)
    - 10.467** (2.43)
    - 5.713* (1.87)
    - 2.126 (0.44)
    - 1.057 (0.73)
    - 0.351 (0.26)
    - 0.301 (0.23)
    - 1.184 (1.12)
    - 1.349 (0.98)
    - 0.634 (0.47)
  - Lag 1 of TFP Growth:
    - 0.310*** (4.36)
    - 0.261*** (4.27)
    - 0.249*** (2.85)
    - 0.295*** (3.97)
    - 0.206*** (3.24)
    - 0.255*** (3.15)
  - Lag 1 of Capital Stock Growth (selected):
    - 0.981*** (7.85)
    - 1.056*** (11.58)
    - 0.940*** (8.43)
    - 1.093*** (12.87)
    - 1.023*** (13.29)
    - 0.942*** (11.34)
  - Lag 2 of Capital Stock Growth (selected):
    - -0.207*** (-3.21)
    - -0.215*** (-3.22)
    - -0.213*** (-3.39)
    - -0.199** (-2.31)
    - -0.201*** (-3.16)
    - -0.208*** (-3.71)
  - Additional notable coefficients:
    - -0.583* (-1.91)
    - 0.248* (1.70)
    - 1.036* (1.81)
- Model diagnostics and sample sizes:
  - Observations: 1009, 1009, 1000, 964, 941, 1009, 1067, 1067, 1058, 1022, 999, 1067
  - Number of clusters: 90.000, 90.000, 90.000, 84.000, 90.000, 90.000, 95.000, 95.000, 95.000, 89.000, 95.000, 95.000
  - Number of instruments: 49.000, 61.000, 61.000, 61.000, 61.000, 67.000, 52.000, 64.000, 64.000, 64.000, 64.000, 70.000
  - Hansen test p-value: 0.625, 0.281, 0.510, 0.717, 0.225, 0.443, 0.387, 0.496, 0.456, 0.254, 0.312, 0.686
  - A-B AR(1) test p-value: 0.000, 0.000, 0.000, 0.000, 0.000, 0.000, 0.004, 0.002, 0.001, 0.002, 0.003, 0.002
  - A-B AR(2) test p-value: 0.677, 0.637, 0.650, 0.831, 0.878, 0.640, 0.098, 0.089, 0.087, 0.124, 0.109, 0.102
  - Lowest number of observations included: 1.000 (all models)
  - Highest number of observations included: 13.000 (all models)
  - Average number of observations included: 11.211, 11.211, 11.111, 11.476, 10.456, 11.211, 11.232, 11.232, 11.137, 11.483, 10.516, 11.232

*Dummy variables for time periods were also included in each model. Results for these are not reported here. t-statistics in parentheses based on cluster-robust standard errors. * p<0.1  ** p<0.05  *** p<0.01*

*Source: wp17143 - Section 6 (wp17143 - Section 6)*

### Section 7

### wp17143 - Section 7

### Table 15: Capital Adequacy Ratio — reported coefficients and diagnostics
- Reported coefficient rows (selected regressors as printed):
  - First coefficient row (columns (1)–(6) and (1)–(6)): -0.537**, -0.669*, -0.610, -0.763*, -0.629***, -0.421, -0.621*, -0.733**, -0.570*, -1.125***, -0.536, -0.420
    - t-statistics (corresponding): (-2.18), (-1.83), (-1.55), (-1.67), (-2.84), (-1.32), (-1.94), (-2.49), (-1.81), (-3.15), (-1.62), (-1.23)
  - Second coefficient row: -1.067*, -0.146, -1.177, -1.359, -0.488, -1.402**, -0.202, -0.193, -0.135, -0.979, -0.748, -0.607
    - t-statistics: (-1.78), (-0.20), (-1.42), (-1.30), (-0.86), (-1.99), (-0.28), (-0.28), (-0.19), (-1.37), (-1.07), (-0.70)
  - Trade to GDP (columns shown): 1.330, 1.717, 0.810, 1.587, 1.173, 1.303, 3.013***, 3.061***, 2.857***, 2.201**, 3.053***, 2.089**
    - t-statistics: (1.20), (1.38), (0.54), (1.16), (1.29), (1.03), (3.32), (3.46), (3.37), (2.19), (3.62), (2.59)
  - Government Consumption to GDP: 0.816, 0.203, 0.370, 1.531, 1.345, -0.404, -2.726***, -2.520***, -3.072***, -2.700***, -2.832***, -2.662***
    - t-statistics: (0.82), (0.19), (0.26), (1.21), (1.33), (-0.41), (-4.48), (-4.53), (-4.34), (-4.59), (-4.71), (-4.71)
  - Years of Schooling: 3.397**, 3.079**, 3.933*, 3.045, 2.384, 4.113**, -0.121, 0.492, -0.244, 2.279, -0.040, 0.943
    - t-statistics: (2.48), (2.10), (1.78), (1.34), (1.65), (2.36), (-0.06), (0.28), (-0.14), (1.26), (-0.02), (0.46)
  - Lag 1 of TFP Growth: 0.032, 0.077, -0.000, 0.049, 0.089, 0.010
    - t-statistics: (0.58), (1.20), (-0.01), (0.88), (1.41), (0.16)
  - Lag 1 of Capital Stock Growth (where reported): 0.708***, 0.704***, 0.650***, 0.675***, 0.626***, 0.663***
    - t-statistics: (10.34), (9.99), (9.42), (10.77), (8.26), (9.78)
  - Lag 2 of Capital Stock Growth: -0.077, -0.061, -0.070, -0.045, -0.092, -0.090
    - t-statistics: (-1.13), (-0.78), (-1.34), (-0.66), (-1.59), (-1.39)
- Diagnostics and sample information:
  - Observations: 589, 589, 570, 530, 545, 589, 714, 714, 690, 615, 653, 714
  - Number of clusters: 107.000, 107.000, 107.000, 95.000, 107.000, 107.000, 127.000, 127.000, 127.000, 111.000, 127.000, 127.000
  - Number of instruments: 66.000, 78.000, 78.000, 78.000, 78.000, 59.000, 67.000, 79.000, 79.000, 79.000, 79.000, 62.000
  - Hansen test p-value: 0.135, 0.239, 0.036, 0.284, 0.265, 0.093, 0.084, 0.274, 0.101, 0.210, 0.420, 0.110
  - A-B AR(1) test p-value: 0.000, 0.000, 0.002, 0.001, 0.000, 0.001, 0.000, 0.000, 0.000, 0.000, 0.000, 0.000
  - A-B AR(2) test p-value: 0.879, 0.555, 0.556, 0.842, 0.364, 0.992, 0.685, 0.621, 0.074, 0.204, 0.387, 0.825
  - Lowest number of observations included: 2.000, 2.000, 2.000, 2.000, 1.000, 2.000, 2.000, 2.000, 2.000, 2.000, 1.000, 2.000
  - Highest number of observations included: 7.000, 7.000, 7.000, 7.000, 7.000, 7.000, 7.000, 7.000, 7.000, 7.000, 7.000, 7.000
  - Average number of observations included: 5.505, 5.505, 5.327, 5.579, 5.093, 5.505, 5.622, 5.622, 5.433, 5.541, 5.142, 5.622
- Notes:
  - Dummy variables for time periods were also included in each model. Results for these are not reported here.
  - t-statistics in parentheses based on cluster-robust standard errors
  - Significance notation: * p<0.1  ** p<0.05  *** p<0.01

### Table 16: Foreign Claims to GDP — reported coefficients and diagnostics
- Reported coefficient rows (selected regressors as printed):
  - First coefficient row (columns (1)–(6) and (1)–(6)): -0.441, -1.796, 0.725, -0.511, -0.898, -0.441, 0.602, 1.042, 0.792, -0.045, 0.373, 0.944
    - t-statistics: (-0.47), (-1.62), (0.55), (-0.51), (-1.13), (-0.48), (0.77), (1.03), (0.74), (-0.04), (0.38), (0.97)
  - Second coefficient row: -1.704***, -0.718, -2.377***, -2.496***, -1.085**, -1.822***, -1.182*, -1.231*, -1.365, -1.277*, -0.210, -0.982
    - t-statistics: (-2.89), (-0.99), (-3.01), (-3.39), (-2.11), (-2.93), (-1.92), (-1.71), (-1.54), (-1.68), (-0.35), (-1.32)
  - Trade to GDP: 1.272, 1.812*, 0.728, 0.176, 1.326*, 1.083, 2.725***, 2.868***, 1.920**, 2.333**, 2.892***, 2.228***
    - t-statistics: (1.31), (1.94), (0.77), (0.11), (1.85), (1.22), (3.73), (3.93), (2.16), (2.22), (3.42), (2.87)
  - Government Consumption to GDP: 0.459, 0.713, 0.071, 1.613, 0.729, 0.151, -1.932***, -1.940***, -2.147***, -2.304***, -2.297***, -2.252***
    - t-statistics: (0.62), (1.08), (0.10), (1.40), (0.86), (0.19), (-3.13), (-3.23), (-3.49), (-3.47), (-3.77), (-3.20)
  - Years of Schooling: 3.287*, 2.614, 4.797***, 4.529*, 2.256*, 3.044*, 1.667, 1.499, 2.962, 3.494*, 0.195, 0.611
    - t-statistics: (1.91), (1.53), (2.64), (1.84), (1.69), (1.73), (1.07), (0.97), (1.46), (1.78), (0.15), (0.31)
  - Lag 1 of TFP Growth: 0.012, 0.036, 0.026, 0.009, 0.068, -0.037
    - t-statistics: (0.20), (0.62), (0.51), (0.13), (1.05), (-0.70)
  - Lag 1 of Capital Stock Growth (where reported): 0.632***, 0.599***, 0.627***, 0.656***, 0.615***, 0.541***
    - t-statistics: (7.10), (7.20), (8.42), (8.27), (7.12), (5.98)
  - Lag 2 of Capital Stock Growth: -0.050, -0.040, -0.040, -0.058, -0.062, -0.093
    - t-statistics: (-0.86), (-0.72), (-0.68), (-0.94), (-1.02), (-1.46)
- Diagnostics and sample information:
  - Observations: 743, 743, 723, 570, 682, 743, 887, 887, 864, 651, 805, 887
  - Number of clusters: 107.000, 107.000, 106.000, 95.000, 106.000, 107.000, 126.000, 126.000, 125.000, 110.000, 125.000, 126.000
  - Number of instruments: 70.000, 86.000, 86.000, 80.000, 86.000, 60.000, 71.000, 87.000, 87.000, 81.000, 87.000, 63.000
  - Hansen test p-value: 0.033, 0.091, 0.146, 0.298, 0.230, 0.030, 0.008, 0.043, 0.046, 0.102, 0.128, 0.005
  - A-B AR(1) test p-value: 0.000, 0.000, 0.000, 0.001, 0.000, 0.000, 0.001, 0.001, 0.000, 0.000, 0.000, 0.001
  - A-B AR(2) test p-value: 0.992, 0.627, 0.777, 0.696, 0.491, 0.558, 0.935, 0.988, 0.294, 0.166, 0.695, 0.683
  - Lowest number of observations included: 1.000, 1.000, 2.000, 1.000, 1.000, 1.000, 1.000, 1.000, 2.000, 1.000, 1.000, 1.000
  - Highest number of observations included: 9.000, 9.000, 9.000, 7.000, 9.000, 9.000, 9.000, 9.000, 9.000, 7.000, 9.000, 9.000
  - Average number of observations included: 6.944, 6.944, 6.821, 6.000, 6.434, 6.944, 7.040, 7.040, 6.912, 5.918, 6.440, 7.040
- Notes:
  - Dummy variables for time periods were also included in each model. Results for these are not reported here.
  - t-statistics in parentheses based on cluster-robust standard errors
  - Significance notation: * p<0.1  ** p<0.05  *** p<0.01

### Table 17: Chinn Ito Index — reported coefficients and diagnostics
- Reported coefficient rows (selected regressors as printed):
  - First coefficient row (columns (1)–(6) and (1)–(6)): -0.179, -0.404, 0.164, 0.469, -0.215, 0.249, -0.911**, -0.826, -0.856, -1.414***, -1.009**, -0.879**
    - t-statistics: (-0.39), (-0.96), (0.32), (0.76), (-0.47), (0.34), (-2.02), (-1.60), (-1.59), (-3.59), (-2.22), (-2.02)
  - Second coefficient row: -1.619**, -0.463, -2.006***, -2.919***, -1.116*, -2.262**, -0.198, -0.391, -0.354, 0.339, 0.169, 0.011
    - t-statistics: (-2.53), (-0.47), (-2.65), (-3.05), (-1.67), (-2.21), (-0.27), (-0.40), (-0.41), (0.37), (0.26), (0.02)
  - Trade to GDP: 0.797, 1.486*, 0.113, -0.353, 0.943, 0.415, 3.020***, 3.281***, 3.024***, 3.144***, 3.111***, 2.827***
    - t-statistics: (0.94), (1.72), (0.12), (-0.23), (1.26), (0.38), (3.46), (3.60), (3.33), (3.00), (3.34), (4.14)
  - Government Consumption to GDP: 0.727, 0.214, 0.270, 1.132, 0.805, -0.295, -2.156***, -1.966***, -1.951***, -2.408***, -1.845***, -2.077***
    - t-statistics: (0.91), (0.28), (0.30), (0.85), (0.78), (-0.30), (-2.97), (-2.87), (-2.84), (-3.44), (-3.56), (-3.53)
  - Years of Schooling: 3.597**, 3.006**, 4.840**, 5.859***, 2.571**, 5.019**, -0.097, -0.598, 0.052, -0.703, -0.960, -0.847
    - t-statistics: (2.24), (2.25), (2.58), (3.04), (2.30), (2.41), (-0.05), (-0.40), (0.03), (-0.37), (-0.72), (-0.49)
  - Lag 1 of TFP Growth: 0.022, 0.073, 0.018, 0.021, 0.057, -0.002
    - t-statistics: (0.44), (1.19), (0.35), (0.32), (0.81), (-0.04)
  - Lag 1 of Capital Stock Growth (where reported): 0.604***, 0.597***, 0.610***, 0.661***, 0.598***, 0.534***
    - t-statistics: (8.13), (8.59), (9.02), (9.75), (8.33), (7.50)
  - Lag 2 of Capital Stock Growth: -0.063, -0.051, -0.058, -0.105*, -0.087, -0.079
    - t-statistics: (-1.05), (-0.82), (-1.09), (-1.83), (-1.45), (-1.28)
- Diagnostics and sample information:
  - Observations: 751, 751, 732, 584, 692, 751, 900, 900, 878, 668, 816, 900
  - Number of clusters: 108.000, 108.000, 108.000, 97.000, 108.000, 108.000, 128.000, 128.000, 128.000, 113.000, 128.000, 128.000
  - Number of instruments: 70.000, 86.000, 86.000, 80.000, 86.000, 61.000, 71.000, 87.000, 87.000, 81.000, 87.000, 64.000
  - Hansen test p-value: 0.075, 0.235, 0.211, 0.205, 0.198, 0.046, 0.055, 0.075, 0.174, 0.157, 0.139, 0.137
  - A-B AR(1) test p-value: 0.000, 0.000, 0.000, 0.001, 0.000, 0.000, 0.001, 0.001, 0.000, 0.000, 0.000, 0.001
  - A-B AR(2) test p-value: 0.904, 0.643, 0.744, 0.917, 0.537, 0.927, 0.897, 0.918, 0.340, 0.253, 0.743, 0.816
  - Lowest number of observations included: 2.000, 2.000, 2.000, 2.000, 1.000, 2.000, 2.000, 2.000, 2.000, 2.000, 1.000, 2.000
  - Highest number of observations included: 9.000, 9.000, 9.000, 7.000, 9.000, 9.000, 9.000, 9.000, 9.000, 7.000, 9.000, 9.000
  - Average number of observations included: 6.954, 6.954, 6.778, 6.021, 6.407, 6.954, 7.031, 7.031, 6.859, 5.912, 6.375, 7.031
- Notes:
  - Dummy variables for time periods were also included in each model. Results for these are not reported here.
  - t-statistics in parentheses based on cluster-robust standard errors
  - Significance notation: * p<0.1  ** p<0.05  *** p<0.01

*Source: wp17143 - Section 7 (tables and notes as provided).*

### Section 8

### Section 8

### Regression results: Assets and Liabilities to GDP
- Coefficients for "Assets and Liabilities to GDP" interactions and related terms (columns labeled (1) through (6) twice):
  - Assets and Liabilities to GDP*MENAPOX Dummy / (1)(2)(3)(4)(5)(6)(1)(2)(3)(4)(5)(6)
    - -0.781, 0.091, -1.615, 0.237, 1.582, 1.083, 0.486, -0.040, -0.660, 0.061, 0.860, 1.530
    - t-statistics: (-0.31), (0.03), (-0.70), (0.07), (0.68), (0.37), (0.56), (-0.05), (-0.90), (0.08), (1.43), (1.37)
  - (Second row of related coefficients)
    - -0.440, -1.079, -0.740, -2.561, 0.169, -2.276, 0.168, 0.253, -0.112, 0.519, 0.207, -0.823
    - t-statistics: (-0.22), (-0.42), (-0.44), (-0.67), (0.12), (-1.28), (0.31), (0.45), (-0.13), (0.42), (0.45), (-0.76)

- Trade to GDP
  - Coefficients: 5.767, 2.353, 7.263**, 4.285, 0.983, 6.465, 0.196, 0.223, 0.836, 0.328, -0.591, 1.017
  - t-statistics: (1.27), (0.54), (2.21), (1.33), (0.43), (1.43), (0.15), (0.23), (0.63), (0.37), (-0.41), (1.00)

- Government Consumption to GDP
  - Coefficients: 0.711, -1.749, -1.082, -2.950, -3.096, -3.298, -1.925**, -1.991, -0.997, -2.121, -2.865**, -1.164
  - t-statistics: (0.15), (-0.50), (-0.24), (-0.91), (-0.87), (-0.80), (-2.03), (-1.37), (-0.63), (-1.48), (-2.18), (-0.74)

- Years of Schooling
  - Coefficients: -4.204, 2.144, -6.743, 2.116, 1.454, -2.622, -1.941, -1.747, -1.915, -1.220, -0.027, -1.181
  - t-statistics: (-0.81), (0.36), (-1.37), (0.35), (0.49), (-0.52), (-1.00), (-1.04), (-0.69), (-0.49), (-0.01), (-0.35)

- Lag 1 of TFP Growth
  - Coefficients reported: 0.245***, 0.213***, 0.210**, 0.243**, 0.186***, 0.193***
  - t-statistics: (3.40), (2.81), (2.27), (2.28), (2.72), (2.81)

- Lag 1 of Capital Stock Growth (selected columns)
  - Coefficients: 0.965***, 0.964***, 0.938***, 0.913***, 0.993***, 0.900***
  - t-statistics: (6.78), (7.08), (8.82), (4.98), (7.86), (6.88)

- Lag 2 of Capital Stock Growth (selected columns)
  - Coefficients: -0.072, -0.077, -0.061, -0.037, -0.078, -0.100
  - t-statistics: (-0.89), (-0.94), (-0.84), (-0.38), (-0.72), (-1.18)

- Additional reported coefficients and t-statistics (displayed in the table in multiple cells):
  - 1.191 (0.36)
  - 0.166 (0.18)
  - 0.011 (0.01)
  - -0.109 (-0.44)
  - -0.130 (-0.21)
  - -0.639 (-1.18)
  - -0.423 (-1.27)
  - -0.096 (-0.41)
  - -1.123 (-0.55)
  - 0.397 (0.52)
  - -1.310 (-1.14)
  - 0.159 (0.27)
  - 1.660 (1.34)
  - -0.001 (-0.00)
  - 0.821** (2.19)
  - 0.115 (0.86)
  - -1.284 (-0.56)
  - -0.509 (-0.40)
  - -2.173 (-1.03)
  - -0.056 (-0.19)
  - 1.271 (0.38)
  - 1.297 (1.15)

- Sample, diagnostics, and model statistics
  - Observations: 689, 689, 685, 613, 637, 689, 796, 796, 792, 694, 743, 796
  - Number of clusters: 100.000, 100.000, 100.000, 89.000, 99.000, 100.000, 117.000, 117.000, 117.000, 102.000, 116.000, 117.000
  - Number of instruments: 52.000, 66.000, 65.000, 66.000, 65.000, 71.000, 55.000, 69.000, 68.000, 69.000, 68.000, 74.000
  - Hansen test p-value: 0.390, 0.290, 0.554, 0.454, 0.729, 0.722, 0.429, 0.290, 0.067, 0.627, 0.714, 0.256
  - A-B AR(1) test p-value: 0.000, 0.000, 0.000, 0.000, 0.000, 0.000, 0.048, 0.044, 0.001, 0.131, 0.009, 0.047
  - A-B AR(2) test p-value: 0.506, 0.600, 0.698, 0.451, 0.469, 0.720, 0.945, 0.922, 0.148, 0.884, 0.067, 0.828
  - Lowest number of observations included: 2.000, 2.000, 2.000, 2.000, 1.000, 2.000, 1.000, 1.000, 1.000, 1.000, 1.000, 1.000
  - Highest number of observations included: 10.000 (for all columns)
  - Average number of observations included: 6.890, 6.890, 6.850, 6.888, 6.434, 6.890, 6.803, 6.803, 6.769, 6.804, 6.405, 6.803

- Notes accompanying the table
  - Dummy variables for time periods were also included in each model. Results for these are not reported here.
  - t-statistics in parentheses based on cluster-robust standard errors.
  - Significance notation: * p<0.1  ** p<0.05  *** p<0.01

### Table 19: Bank Branches
- Table title present: "Table 19:  Bank Branches"
- Listed interaction terms and variables (headers in the table):
  - Bank Branches*CCA Dummy
  - Bank Branches*MENAPOI Dummy
  - Bank Branches*MENAPOX Dummy
  - Bank Branches*Medium Institutional Quality Dummy
  - Bank Branches*Low Financial Depth
  - Bank Branches*Medium Financial Depth
  - Bank Branches*Low Income Dummy
  - Bank Branches*Middle Income Dummy
  - Bank Branches*High Inflation Dummy
  - Bank Branches*Middle Inflation Dummy
  - Bank Branches*Low Institutional Quality Dummy
  - Capital
  - Bank Branches
  - Initial GDP per Capita
  - TFP

*Source: wp17143 - Section 8*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2017/wp17143.pdf_