## 8. The Impact of Policy Variables on the Different Dimensions of Financial Development

## Source details

**Canonical URL:** [8. The Impact of Policy Variables on the Different Dimensions of Financial Development](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2015/_wp15173.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2015/_wp15173.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2015/_wp15173.pdf.json)

---

### Scope, motivation, and conceptual framework
- Reviews debate on finance and growth citing Lucas (1988), Schumpeter (1911), Goldsmith (1969), King & Levine (1993), Rajan & Zingales (1998), Levine et al (2000), Beck et al (2000), Levine & Zervos (1998).
- Notes consensus that financial institutions influence development by channeling savings, monitoring borrowers, pooling/diversifying risk, and facilitating trade, but highlights nonlinearities and adverse effects from excessive financial development (Rioja & Valev (2004); Aghion et al (2005); Demetriades & Law (2006); Arcand et al (2012); Dabla-Norris & Srivisal (2013); Rajan (2005); Beck et al (2012)).
- Introduces “optimal” level of financial development and the financial possibility frontier as benchmarking tools (Beck & de la Torre (2007); Beck & Feyen (2013); Barajas et al (2013a)).
- Financial possibility frontier and policy typology:
  - Market developing policies: legal/structural reforms, long-run fiscal changes — push the frontier outwards.
  - Market enabling policies: short- to medium-term measures (macroeconomic policy, regulatory reforms, financial and trade openness) — push countries toward the frontier. Study primarily focuses on market enabling policies.
  - Market harnessing policies: risk oversight, macro-prudential management, financial literacy, consumer protection — prevent persistent overshooting of the frontier.
- Benchmarking approach:
  - Structural benchmarks use a large cross-country panel to produce time-varying benchmarks for financial development indicators based on structural variables assumed invariant in the short run.
  - Benchmarks isolate the structural element; the remainder is modeled as a function of policy and institutional factors.
  - Caveats: benchmarks assume all key structural variables are accounted for and represent a “median country” path; may not capture country-specific initial conditions, crises, or leapfrogging.

### Research objective, data, indicators, and limitations
- Objectives and extensions:
  - Extends Barajas et al (2013a) by analyzing three dimensions of financial development—depth, efficiency, and stability—using annual data and a dynamic specification with lagged regressand and dynamic GMM estimators.
  - Focuses on banking-sector indicators due to benchmarking data constraints; results aimed at countries where banking dominates formal finance (notably many developing countries).
  - Employs panel techniques to exploit cross-sectional and time-series variation and to identify complementarities and trade-offs across dimensions.
- Primary indicators:
  - Depth: private credit as a percentage of GDP.
  - Efficiency: net interest margin.
  - Stability: non-performing loans as a percentage of total gross loans.
- Robustness alternatives included: domestic deposits to GDP, banks’ assets to GDP, banks’ cost to income ratio, lending to deposit spread, banks’ Z-score, private credit to deposits ratio.
- Data coverage:
  - Compiled dataset spanning 180 countries and 28 years (1984-2011); final panel unbalanced.
  - Banking sector depth: 2,396 observations across 115 countries (1984-2011).
  - Efficiency: 1,179 observations across 103 countries (1998-2011).
  - Stability: 930 observations across 85 countries (1998-2011).
- Data sources: FinStats 2014 (World Bank), WEO (IMF), WDI, GFDD (World Bank); trade openness measured as ratio of real imports to real GDP; financial openness from Chinn & Ito (2008); polity score (Polity IV); composite risk rating (ICRG/PRS).
- Limitations:
  - Analysis limited to banking institutions and available banking indicators.
  - Benchmarks assume structural variables fully capture fundamentals; complementarity/trade-off analysis is basic and does not model political economy.

### Methodology and estimation strategy
- Dynamic specifications due to high persistence in financial development indicators; include lagged regressand.
- Estimators:
  - Difference GMM (Arellano & Bond (1991)) with forward orthogonal deviations (Arellano & Bover (1995)).
  - System GMM (Arellano & Bover (1995); Blundell & Bond (1998)) favored for highly persistent regressands.
- Instrument strategy:
  - Higher-order lags of endogenous variables used as internal instruments.
  - Instrument count limited via restricting lags and collapsing (Holtz-Eakin et al (1988); Roodman (2009)).
- Diagnostics:
  - Arellano-Bond test for serial correlation in differenced residuals: first-order serial correlation expected; inability to reject no second-order serial correlation is crucial.
  - Sargan/Hansen tests for joint validity of instruments; watch for excessively large Hansen statistic when instrument count is too high.
- Model specifications:
  - Gap specification (Equation No. 1): regress relative deviation from structural benchmark (gap) on policy variables; for comparability, sign reversed for efficiency and stability gaps.
  - Level specification (Equation No. 2): regress levels on structural and policy variables to check for omitted-variable bias.
  - Policy variables treated in three ways: exogenous, predetermined, endogenous; lags used where appropriate.
  - Complementary pooled OLS and fixed effects estimations reported as robustness checks.

### Key empirical findings (dynamic panel estimation)
- Policy variables found to significantly affect different dimensions of financial development:
  - inflation
  - trade openness
  - institutional quality (composite risk rating)
  - occurrence of banking crises
- General pattern: most policy variables exhibit complementary effects across dimensions; some variables produce contradictory effects implying policy trade-offs.
- Diagnostics and model comparisons:
  - Arellano-Bond AR(2) p-values and Hansen test p-values support model validity across reported tables.
  - Lags of the dependent variable are significant in each regression.
  - Gap (Equation No.1) and level (Equation No.2) specifications yield largely similar determinants; a few variables lose significance switching models but no sign reversals reported.
  - For financial depth, second order lags were added to obtain satisfactory results.

### Determinants by dimension — summarized empirical results
- Financial depth (private credit/GDP):
  - Persistence: depth and depth gap exhibit very high persistence; emphasis on system GMM.
  - Statistically significant determinants:
    - Growth: positive effect on depth; significance depends on treatment of growth (significant when lagged exogenous or predetermined; loses significance when endogenised).
    - Banking crises: occurrence significantly reduces depth.
    - Trade openness: increased openness positively influences depth.
    - Institutional quality (composite risk rating): better institutions positively influence depth.
- Financial efficiency (net interest margin; lower margin = greater efficiency):
  - Efficiency gap sign reversed for comparability.
  - Key determinants:
    - Inflation: higher inflation detrimental to efficiency.
    - Banking crises: Table 3b suggests crises may improve efficiency (possible purging effect), but effect loses significance in level specification (Table 6b).
    - Institutions: stronger institutions tend to promote efficiency (lower significance).
    - Lerner index (competitiveness): insignificant in gap analysis (Table 3b) but significant in level specification (Table 6b); expected relationship is greater competitiveness increases efficiency.
- Financial stability (non-performing loans % of gross loans; lower = greater stability):
  - Stability gap sign reversed for comparability.
  - Principal determinants:
    - Growth: positive effect on stability; effect extremely significant and persistent across specifications.
    - Banking crises: reduce stability (significant in some specifications; loses significance in second specification).
    - Inflation: Table 4b suggests inflation might aid stability but effect largely disappears in Table 7b; possible mechanism is inflation reducing real interest rates and easing debt burdens.
    - Institutions (risk rating): improved institutions increase stability.

### Robustness checks and selected empirical statistics
- Robustness scenarios:
  - Trim sample to low and middle income countries (exclude high income countries).
  - Limit sample to pre-global financial crisis period (up to 2007).
  - Trim out top and bottom 2.5% of regressand values (middle 95%).
- Financial depth robustness highlights:
  - Growth consistently plays an important positive role.
  - Institutional quality becomes even more important in developing-country samples.
  - Polity score (democracy) becomes significant positive in some developing-country specifications.
  - Remittances sometimes show negative effects on depth in robustness checks, but effect disappears in the second model specification.
  - Selected point estimates from Table 2c (Depth Gap (t - 1)): 1.0790***, 1.0719***, 1.0111***, 1.0358***, 1.0674***, 1.0831***, 1.0993***, 1.1188***, 1.0602***, 1.0791***, 1.0922***, 1.1319***, 1.0609***, 1.1081***, 1.0191***, 1.0956***, 1.0291***, 1.1011***.
  - Growth coefficients in Table 2c robustness: 0.0093***, 0.0090***, 0.0095***, 0.0083***, -0.0009, -0.0008, 0.0101***, 0.0099***, 0.0092***, 0.0088***, -0.0007, -0.0007, 0.0054***, 0.0062***, 0.0056***, 0.0056***, -0.0009, -0.0008.
  - Banking Crisis Dummy in Table 2c robustness: -0.1180***, -0.1046***, -0.0904***, -0.0893***, -0.0389, -0.0206, -0.1131***, -0.1084***, -0.0977***, -0.1007***, 0.0115, -0.0013, -0.0562***, -0.0607***, -0.0481**, -0.0446***, -0.0085, -0.0027.
- Financial efficiency robustness highlights:
  - Negative inflation effect on efficiency holds across most robustness checks.
  - Institutional quality still plays a role, though weaker in some developing-country samples.
  - Lerner Index remains significant in most specifications and is particularly significant in developing countries.
  - Selected estimates from Table 3c:
    - Efficiency Gap (t - 1): 0.3763***, 0.3854**, 0.3349*, 0.4306***, 0.3137**, 0.4127***, 0.6031***, 0.1241***, 0.5046***, 0.1867***, 0.4423***, 0.1861***, 0.3477***, 0.3249***, 0.3562***, 0.4820***, 0.3306***, 0.4752***.
    - Inflation coefficients in Table 3c robustness: -0.0060***, -0.0108***, -0.0054***, -0.0063**, -0.0078, -0.0038, -0.0043*, -0.0137***, -0.0031, -0.0048*, -0.0038, -0.0035, -0.0037*, -0.0088***, -0.0002, -0.0023, -0.0074, -0.0113**.
    - Lerner Index coefficients in Table 3c robustness: -0.6063***, -0.2636*, -0.5443*, -0.3351, -0.7994**, -0.3467, -0.4007**, -0.3818***, -0.2277, -0.4199, -0.0576, -0.0811, -0.2942***, -0.2082**, -0.0129, 0.0115, -0.3134, -0.2752.
  - Banking crisis dummy often loses significance in robustness checks.
- Financial stability robustness highlights:
  - Smaller sample sizes limit stability results; when limited to developing countries, explanatory power is reduced.
  - Trimmed sample: growth, inflation, banking crises, and institutional quality are significant determinants.
  - Selected estimates from Table 4c:
    - Stability Gap (t - 1): 0.9231***, 1.0271***, 0.9156***, 0.9826***, 0.8771***, 0.9306***, 0.6036***, 0.5121, 0.4972***, 0.6782***, 0.5166***, 0.5801***, 0.8038***, 0.9364***, 0.7615***, 0.8901***, 0.7769***, 0.8645***.
    - Growth coefficients in Table 4c robustness: 0.0122, 0.0062, 0.0092, 0.0051, 0.0745**, 0.0616***, 0.0349, 0.0461, 0.0176, 0.0286, 0.0509**, 0.0228, 0.0183**, 0.0140*, 0.0232***, 0.0256***, 0.0504***, 0.0606***.
    - Banking Crisis Dummy in Table 4c robustness shows mixed significance: -0.0390, -0.0403, -0.3533, -0.2951, -0.0536, -0.4908*, 0.1471, -0.0232, -0.3638, -0.1109, -0.7711*, -0.5131, -0.1429*, -0.0619, -0.2341, -0.1422*, -0.1910, -0.1982**.

### Conceptual and policy implications
- Reinforces use of the financial possibility frontier and benchmarking to assess financial system health and guide policy sequencing: market-enabling, market-developing, and market-harnessing instruments depending on position relative to frontier.
- Emphasizes optimizing financial development (not maximizing) because sustained overshooting of benchmarks associates with higher probability of bad credit booms.
- Multi-dimensionality of financial development requires policies that consider depth, access, efficiency, and stability together given complementarities and trade-offs.
- Practical implication: identifying effects of individual policy variables across dimensions can inform the design of an appropriate policy mix, but results are intended to guide intuition for country-specific analyses rather than prescribe global solutions.

*Source: _wp15173 - Sections 5–6; 6.1 Estimation results; 6.2 Robustness checks; 8. The Impact of Policy Variables on the Different Dimensions of Financial Development*

### 8. The Impact of Policy Variables on the Different Dimensions of Financial Development ..38

### 8. The Impact of Policy Variables on the Different Dimensions of Financial Development

### Scope and motivation
- Reviews longstanding debate on finance and economic growth, citing diverse views from Lucas (1988) to Schumpeter (1911) and empirical studies (Goldsmith (1969); King & Levine (1993); Rajan & Zingales (1998); Levine et al (2000); Beck et al (2000); Levine & Zervos (1998)).
- Notes consensus that financial institutions positively influence economic development by channeling savings, monitoring borrowers, pooling and diversifying risk, and facilitating trade.
- Highlights non-linearities and potential adverse effects of excessive financial development (Rioja & Valev (2004); Aghion et al (2005); Demetriades & Law (2006); Arcand et al (2012); Dabla-Norris & Srivisal (2013); Rajan (2005); Beck et al (2012)).
- Introduces the concept of an “optimal” level of financial development and the financial possibility frontier as a benchmarking tool (Beck & de la Torre (2007); Beck & Feyen (2013); Barajas et al (2013a)).

### Research objective and contributions
- Extends Barajas et al (2013a) in two ways:
  - Extends analysis to three dimensions of financial development—depth, efficiency, and stability—based on data and benchmarks.
  - Uses annual data and a dynamic specification with lagged regressand and dynamic GMM estimators to address persistence and endogeneity.
- Focuses on banking-sector indicators due to current benchmarking data constraints; results aimed to be relevant to a large set of countries, particularly developing countries where banking dominates the formal financial sector.
- Provides a more nuanced assessment of policy effects across multiple financial development dimensions, identifying complementarities and trade-offs.
- Employs panel data techniques to exploit cross-sectional and time-series variation.

### Indicators and robustness
- Primary indicator choices:
  - Depth: private credit as a percentage of GDP.
  - Efficiency: net interest margin.
  - Stability: non-performing loans as a percentage of total gross loans.
- Robustness checks (re-runs of regressions reported in Tables 5-7b) used alternative indicators, including:
  - domestic deposits to GDP, banks’ assets to GDP, banks’ cost to income ratio, lending to deposit spread, banks’ Z-score, and private credit to deposits ratio.
- Notes that size and significance of estimated coefficients vary across alternate indicators, but robustness tests confirm broad findings.

### Empirical strategy and limitations
- Uses annual data and a dynamic specification that includes a lagged regressand to account for persistence and adjustment toward steady state.
- Employs dynamic generalized method of moments (GMM) estimators to allow for endogeneity.
- Current availability of benchmarking data limits analysis to banking institutions and indicators of banking-sector development.

### Key empirical findings (summary from dynamic panel estimation)
- The following policy variables significantly affect different dimensions of financial development:
  - inflation
  - trade openness
  - institutional quality
  - occurrence of banking crises
- Most policy variables exhibit complementary effects across different dimensions of financial development; some variables produce contradictory effects, implying policy trade-offs for decision makers.

### Conceptual and policy implications
- Reinforces the financial possibility frontier and benchmarking as conceptual tools to assess financial system health and guide policy.
- Emphasizes that optimizing financial development (rather than maximizing it) is important, given evidence that sustained overshooting of benchmarks is associated with higher probability of bad credit booms.
- Highlights the multidimensionality of financial development and the need to consider depth, access, efficiency, and stability when designing policy, given potential complementarities and trade-offs across these dimensions.

*Source: _wp15173 - 8. The Impact of Policy Variables on the Different Dimensions of Financial Development ..38*

### Section 5 describes the data used in the analysis. Section 6 reports the estimation results, and

### _wp15173 - Sections 5–6 (Data, Measurement, Sources; Empirical Analysis introduction)

### Financial Possibility Frontier — conceptual framework
- Financial intermediaries arise because of market frictions: uncertainty, informational asymmetries, limited enforceability, transaction costs, and network externalities.
- These frictions generate both demand for intermediation (risk pooling, diversification, insurance) and supply constraints (fixed costs, risk of default, network effects).
- Network effects increase incumbent market power and can act as barriers to new entrants.
- The financial possibility frontier denotes the maximum sustainable level of an economy’s financial system development at a given time, influenced by state variables and market frictions.
- Policy typology based on position relative to the frontier (Beck & Feyen (2013); Barajas et al (2013a)):
  - Market developing policies: legal/structural reforms, long-run fiscal changes — push the frontier outwards.
  - Market enabling policies: short- to medium-term measures (macroeconomic policy, regulatory reforms, financial and trade openness) — push countries toward the frontier. This study primarily focuses on market enabling policies.
  - Market harnessing policies: risk oversight, macro-prudential management, financial literacy, consumer protection — prevent persistent overshooting of the frontier.
- Benchmarking approach:
  - Structural benchmarks use a large cross-country panel to produce time-varying benchmarks for financial development indicators based on structural variables assumed invariant in the short run.
  - The benchmark isolates the structural element of financial development; the remainder is modeled as a function of policy and institutional factors.
  - Caveats: benchmarks assume all key structural variables are accounted for; they represent a “median country” path and may not capture country-specific initial conditions, crises, or leapfrogging.

### Structural variables used in benchmarks (as employed in Feyen et al (2013))
- Economic Development:
  - GDP per capita
  - Square of GDP per capita
- Demographics:
  - Population size
  - Population density
  - Age dependency ratio, old
  - Age dependency ratio, young
- Special circumstances:
  - Offshore financial centre dummy
  - Oil exporter dummy
  - Transition economy dummy
- Time dummies are included to capture global cycle effects in pooled regressions.

### Determinants of financial development — policy and institutional factors surveyed
- Institutions and legal/regulatory environment: property rights protection, creditor rights, checks and balances affect financial development and crisis mitigation.
- Political economy: interest groups, trade and financial openness, democracy/regime stability affect financial development.
- Macroeconomic stability: inflation impacts financial development; studies indicate finance–growth nexus breaks down when inflation consistently exceeds 13%.
- Remittances: evidence of a significant and positive influence on financial development in many studies.
- Competitiveness / market structure: mixed evidence; cross-country studies often find positive effects.
- Policy variables used in analyses:
  - Macroeconomic variables:
    - (Real) growth rate
    - Inflation rate
    - Remittance inflows
    - Banking crisis dummy
  - Openness variables:
    - (De jure) financial openness
    - Trade openness
  - Institutional variables:
    - Polity score
    - Composite risk rating
  - Market power / competitiveness:
    - Lerner index
- Structural variables used in estimations (when included): real GDP per capita, population size, population density, age dependency ratio.

### Methodology — model specification and estimation strategy
- Objective: identify key policy determinants of three dimensions of financial development using an annual unbalanced panel with dynamic specifications due to persistence in financial development indicators.
- Two model specifications:
  - Gap specification: regress the relative deviation (from structural benchmark) of a financial development indicator on policy variables. The gap is defined as the difference between the benchmark and the actual indicator (sign reversed for efficiency and stability to facilitate comparison across dimensions).
  - Level specification: regress the level of a financial development indicator on structural variables and policy variables to check for omitted-variables bias from the benchmarking approach.
- Error term decomposition in both specifications includes:
  - Fixed effects (country)
  - Time-varying global components (captured via time dummies)
  - Idiosyncratic error term (independent, mean zero)
- Dynamic panel bias: the lagged dependent variable on the RHS is endogenous to fixed effects (Nickell (1981)), so OLS and fixed effects are biased.

### Dynamic panel GMM estimation details and diagnostics
- Estimators used:
  - Difference GMM (Arellano & Bond (1991)) with forward orthogonal deviations (Arellano & Bover (1995)) to reduce data loss in unbalanced panels.
  - System GMM (Arellano & Bover (1995); Blundell & Bond (1998)) favored when the regressand is highly persistent.
- Instrument strategy and precautions:
  - Use higher-order lags of endogenous variables as internal instruments (differenced or differenced instruments as per system GMM).
  - Limit the number of instruments to less than the number of countries by restricting the number of lags and using the collapsing method of Holtz-Eakin, Newey, and Rosen (1988), following Roodman (2009) guidance.
- Diagnostic tests:
  - Arellano-Bond test for serial correlation in differenced residuals: first-order serial correlation expected; crucial test is inability to reject no second-order serial correlation.
  - Sargan/Hansen test for joint validity of instruments; watch for excessively large Hansen statistic when instrument count is too high.
- Treatment of policy variables:
  - Three formulations for policy variables: treated as exogenous, predetermined, and endogenous respectively.
  - In exogenous and predetermined formulations, policy variables are lagged by one period to mitigate contemporaneous bias.
  - In level specification, structural variables are always treated as exogenous and lagged by one period.
- Complementary checks:
  - Pooled OLS and fixed effects estimations are included as robustness checks (provide useful checks on the coefficient of the lagged dependent variable despite being technically incorrect for dynamic panels).

### Data, measurement, and sources (sample coverage and indicators)
- Compiled data set spanning 180 countries and 28 years (1984-2011). The final panel is unbalanced and indicator coverage varies by financial development dimension.
- Coverage by dimension:
  - Banking sector depth:
    - 2,396 observations across 115 countries between 1984 and 2011.
    - Indicator: Private credit as a percentage of GDP (FinStats 2014 database, World Bank).
  - Efficiency:
    - 1,179 observations across 103 countries (1998-2011).
    - Indicator: Net interest margin.
  - Stability:
    - 930 observations across 85 countries (1998-2011).
    - Indicator: Non-performing loans as a percentage of total gross loans.
- Data sources:
  - Financial development indicators and benchmarks: FinStats 2014 database (World Bank).
  - Macroeconomic data: World Economic Outlook (WEO) data set (IMF), World Development Indicators (WDI), Global Financial Development (GFDD) data set (World Bank).
  - Trade openness measure: constructed as ratio of real imports to real GDP (chosen over net exports/GDP for the stated reasons).
  - Financial openness: de jure measure constructed by Chinn & Ito (2008).
  - Political/institutional quality measures:
    - Polity score (Polity IV project).
    - Average annual composite risk rating (International Country Risk Guide (ICRG), Political Risk Services (PRS)).
- Variable definitions, and summary statistics are reported in Tables 1 and Tables 2-7a (not reproduced here).

### Analysis approach and assumptions highlighted
- The analysis assumes different dimensions of financial development are orthogonal; complementarities and trade-offs are assessed based on the signs of estimated coefficients across different dimensions.
- Acknowledged limitations:
  - Complementarities and trade-offs analysis is basic; more detailed work (interaction terms, extensive correlation analysis among policy variables) is beyond scope.
  - Benchmarking assumes all key structural variables are captured; may not reflect country-specific initial conditions or heterogeneity.
  - Benchmarks reflect the median-country development path; policy importance may vary significantly across countries and time.

*Source: _wp15173 (Sections 5–6 excerpt provided).*

### 6.1 Estimation results

### 6.1 Estimation results

### Overview and model diagnostics
- Main results presented in Tables 2-7b. Tables 2b, 3b, and 4b report estimates of Equation No. 1 using development gaps for depth, efficiency, and stability, respectively. Tables 5b, 6b, and 7b report estimates of Equation No. 2 using indicators for banking sector depth, efficiency, and stability.
- Arellano-Bond test for second order serial correlation (AR(2)) p-values are sufficiently high in each table, supporting the absence of problematic AR(2) correlation.
- Hansen test p-values confirm that the instruments are appropriate in each case.
- Lags of the y-variable are significant in each regression, validating the choice of a dynamic specification.
- Comparing results from the gap analysis (Equation No.1) and the second specification (Equation No.2) shows largely similar policy determinants across dimensions; a few determinants lose significance when switching models, but no variable reverses sign.
- Note: In the case of financial depth, second order lags were added to obtain satisfactory results. (See footnote 17.)
- Note: See Table 1 for detailed definitions of all variables used. (See footnote 18.)

### Financial depth (Tables 2b and 5b)
- Indicator: private credit as a fraction of GDP (higher indicates greater depth).
- By construction, a positive financial depth gap implies the financial sector is over-performing on depth.
- Persistence:
  - Financial depth and the depth gap exhibit a very high level of persistence; focus placed on system GMM estimations.
- Statistically significant determinants reported:
  - Growth: positive effect on financial depth; significance depends on treatment of growth (significant when treated as lagged exogenous and lagged predetermined; loses significance when endogenised).
  - Banking crises: occurrence significantly reduces financial depth.
  - Trade openness: increased openness positively influences financial depth.
  - Quality of institutions (composite risk rating): better institutions positively influence financial depth.

### Financial efficiency (Tables 3b and 6b)
- Indicator: net interest margin in the banking sector (lower margin implies greater efficiency).
- Efficiency gap constructed by reversing sign to facilitate comparison across dimensions.
  - Thus, a variable that positively influences efficiency has a positive coefficient in Table 3b and a negative coefficient in Table 6b.
- Key findings:
  - Inflation: higher inflation appears detrimental to financial efficiency.
  - Banking crises: Table 3b suggests crises have a positive effect on efficiency (possible "purging" effect where weak, inefficient banks are wiped out), but this effect loses significance in Table 6b.
  - Institutions: stronger institutions tend to promote efficiency, though significance is lower.
  - Lerner index (competitiveness): insignificant in gap analysis (Table 3b) but significant in the second specification (Table 6b); relationship as expected—greater competitiveness spurs increased efficiency.

### Financial stability (Tables 4b and 7b)
- Indicator: fraction of total loans classified as non-performing (lower value implies greater stability).
- Stability gap defined by reversing the sign; a positive policy effect on stability corresponds to a positive coefficient in Table 4b and a negative coefficient in Table 7b.
- Principal determinants:
  - Growth: positive effect on stability; effect is extremely significant and persists across all specifications.
    - Economic rationale: higher growth increases real returns and capacity to repay loans, reducing non-performing loans.
  - Banking crises: shown to reduce financial stability (significant in some specifications, loses significance in the second specification).
  - Inflation: Table 4b suggests an increase in inflation might aid financial stability, but this effect largely disappears in Table 7b; possible mechanism is inflation reducing real interest rates and easing debt burdens.
  - Institutions (risk rating): improved quality of institutions increases financial stability.

*Source: _wp15173 - 6.1 Estimation results*

### 6.2 Robustness checks

### _wp15173 - 6.2 Robustness checks

### Robustness check design
- Three limiting situations re-run:
  - Trim sample to low and middle income countries by excluding all data from high income countries—both within and outside the OECD.
  - Limit data sample to the period before the global financial crisis (i.e. up to 2007).
  - Test impact of outliers in the regressand by repeating analysis on the middle 95% of the regressand values; i.e. trim out the top and bottom 2.5% values of the y-variable for each regression (middle 95%).
- Results of robustness checks presented in Tables 2-7c.

### Findings — Financial depth
- Most results are resilient to robustness checks despite reduced sample breadth when trimming to developing countries.
- Persistent effects across specifications:
  - Growth “plays an important, positive role in increasing financial depth.”
  - Institutional quality “becomes even more important” in trimmed/developing-country samples.
- Additional specific effects:
  - The level of democracy (polity score) “becomes a significant positive factor in some of the specifications for the developing country sample.”
  - Remittances “seem to have a negative effect on financial depth in our robustness checks,” but “this effect disappears in the second model specification.”
- Pre-crisis sample and trimmed sample show the same general results with growth, banking crises, and institutional quality being the most significant explanatory variables.
- Empirical details (from tables):
  - Robustness checks tables for depth: Table 2c and Table 5c report estimations across Trimmed Sample, Developing Countries, Pre-GFC (upto 2007), and other instrument/lag specifications.
  - Example point estimates (selected, preserved as in source tables):
    - Depth Gap (t - 1) coefficients in robustness table (Table 2c): 1.0790***, 1.0719***, 1.0111***, 1.0358***, 1.0674***, 1.0831***, 1.0993***, 1.1188***, 1.0602***, 1.0791***, 1.0922***, 1.1319***, 1.0609***, 1.1081***, 1.0191***, 1.0956***, 1.0291***, 1.1011***.
    - Growth coefficients in Table 2c robustness: 0.0093***, 0.0090***, 0.0095***, 0.0083***, -0.0009, -0.0008, 0.0101***, 0.0099***, 0.0092***, 0.0088***, -0.0007, -0.0007, 0.0054***, 0.0062***, 0.0056***, 0.0056***, -0.0009, -0.0008.
    - Banking Crisis Dummy in Table 2c robustness: -0.1180***, -0.1046***, -0.0904***, -0.0893***, -0.0389, -0.0206, -0.1131***, -0.1084***, -0.0977***, -0.1007***, 0.0115, -0.0013, -0.0562***, -0.0607***, -0.0481**, -0.0446***, -0.0085, -0.0027.
  - Standard significance notation preserved: *** p<0.01, ** p<0.05, * p<0.1.

### Findings — Financial efficiency
- Robustness checks validate previously reported results.
- Persistent effects across specifications:
  - Negative impact of high inflation on financial efficiency “holds through most of the robustness checks.”
  - Institutional quality “still seem[s] to play a role,” although not for the developing countries’ sample in some specifications.
  - Measure of competitiveness (Lerner Index) remains significant through most specifications and “seems particularly significant in developing countries.”
- Banking crisis dummy:
  - “The coefficient of the banking crisis dummy variable loses significance” in some robustness specifications; authors note this is not altogether surprising.
- Empirical details (from tables):
  - Efficiency Gap (t - 1) coefficients in robustness Table 3c: 0.3763***, 0.3854**, 0.3349*, 0.4306***, 0.3137**, 0.4127***, 0.6031***, 0.1241***, 0.5046***, 0.1867***, 0.4423***, 0.1861***, 0.3477***, 0.3249***, 0.3562***, 0.4820***, 0.3306***, 0.4752***.
  - Inflation coefficients in Table 3c robustness: -0.0060***, -0.0108***, -0.0054***, -0.0063**, -0.0078, -0.0038, -0.0043*, -0.0137***, -0.0031, -0.0048*, -0.0038, -0.0035, -0.0037*, -0.0088***, -0.0002, -0.0023, -0.0074, -0.0113**.
  - Lerner Index coefficients in Table 3c robustness (market power): -0.6063***, -0.2636*, -0.5443*, -0.3351, -0.7994**, -0.3467, -0.4007**, -0.3818***, -0.2277, -0.4199, -0.0576, -0.0811, -0.2942***, -0.2082**, -0.0129, 0.0115, -0.3134, -0.2752.

### Findings — Financial stability
- Reduced sample sizes impact stability results most severely.
- When limited to developing countries (sample nearly halved):
  - “Only growth and banking crises retaining some explanatory power in developing countries.”
- Trimmed sample results:
  - “Growth, inflation, banking crises, and institutional quality are significant determinants of financial stability.”
- Empirical details (from tables):
  - Stability Gap (t - 1) coefficients in Table 4c robustness: 0.9231***, 1.0271***, 0.9156***, 0.9826***, 0.8771***, 0.9306***, 0.6036***, 0.5121, 0.4972***, 0.6782***, 0.5166***, 0.5801***, 0.8038***, 0.9364***, 0.7615***, 0.8901***, 0.7769***, 0.8645***.
  - Growth coefficients in Table 4c robustness: 0.0122, 0.0062, 0.0092, 0.0051, 0.0745**, 0.0616***, 0.0349, 0.0461, 0.0176, 0.0286, 0.0509**, 0.0228, 0.0183**, 0.0140*, 0.0232***, 0.0256***, 0.0504***, 0.0606***.
  - Banking Crisis Dummy in Table 4c robustness shows mixed significance across specifications; select values include -0.0390, -0.0403, -0.3533, -0.2951, -0.0536, -0.4908*, 0.1471, -0.0232, -0.3638, -0.1109, -0.7711*, -0.5131, -0.1429*, -0.0619, -0.2341, -0.1422*, -0.1910, -0.1982**.

### Overall conclusions from the paper (as presented in this section)
- Study aim: identify key policy determinants of financial development across three dimensions: depth, efficiency, and stability, using the financial possibility frontier framework which separates structural fundamentals and policy/institutional factors.
- Method: benchmark economies on structural fundamentals and analyze gaps between actual indicators and structural benchmarks to identify policy determinants; alternate model specifications and robustness checks applied.
- Practical implications:
  - The financial possibility frontier guides which policy tools are appropriate (market-enabling, market-developing, or market-harnessing) depending on where a country stands relative to its structural benchmark.
  - Identifying effects of individual policy variables across dimensions allows design of an appropriate policy mix.
- Caveats and scope:
  - The study treats policy variables as exogenous for analysis and does not explicitly model the political economy of financial development or financial policy.
  - Results intended to “guide intuition in country-specific policy analyses, rather than to prescribe global solutions for local maladies.”
- Summary impact matrix (Table 8) preserved:
  - Growth: Depth ↑-↑ ; Efficiency (blank) ; Stability ↑-↑
  - Inflation: Depth -↓ ; Efficiency (blank) ; Stability ↑-↓-
  - Remittance Inflows: ------
  - Banking Crisis Dummy: Depth ↓-↓ ; Efficiency (blank) ; Stability ↓--
  - Financial Openness: ------
  - Trade Openness: Depth ↑-↓ ; Efficiency ↑-↓ ; Stability (blank)
  - Polity Score: ------
  - Composite Risk Rating: Depth ↑↑↑ ; Efficiency ↑↑↑ ; Stability (figure repeats) ↑↑↑
  - Lerner Index (Market power): -↓

### REFERENCES

### _wp15173 - REFERENCES

### References list

- Acemoglu, Daron, Simon Johnson, and James A. Robinson. "Institutions as a fundamental 
cause of long-run growth." Handbook of economic growth 1 (2005): 385-472. 
- Aggarwal, Reena, Aslı Demirgüç-Kunt, and Maria Soledad Martínez Pería. "Do remittances 
promote financial development?" Journal of development economics 96.2 (2011): 255-264. 
- Aghion, Philippe, Peter Howitt, and David Mayer-Foulkes. "The Effect of Financial 
Development on Convergence: Theory and Evidence." The quarterly journal of 
economics 120.1 (2005): 173-222. 
- Arcand, Jean-Louis, Enrico Berkes, and Ugo Panizza. "Too much finance?" International 
Monetary Fund working paper 12/161 (2012). 
- Arellano, Manuel, and Olympia Bover. "Another look at the instrumental variable estimation 
of error-components models." Journal of econometrics 68.1 (1995): 29-51. 
- Arellano, Manuel, and Stephen Bond. "Some tests of specification for panel data: Monte 
Carlo evidence and an application to employment equations." The review of economic 
studies 58.2 (1991): 277-297. 
- Bagehot, Walter. Lombard Street: A Description of the Money Market. Scribner, Armstrong 
& Co., 1873. 
- Baltagi, Badi H., Panicos O. Demetriades, and Siong Hook Law. "Financial development 
and openness: Evidence from panel data." Journal of development economics 89.2 (2009): 
285-296. 
- Baltagi, Badi. Econometric analysis of panel data. Vol. 1. John Wiley & Sons, 2008. 
- Barajas, Adolfo, Thorsten Beck, Era Dabla-Norris, and Seyed Reza Yousefi. "Too Cold, Too 
Hot, Or Just Right? Assessing Financial Sector Development Across the Globe." 
International Monetary Fund working paper 13/81 (2013). [Barajas et al 2013a] 
- Barajas, Adolfo, Ralph Chami, and Seyed Reza Yousefi. "The Finance and Growth Nexus 
Re-Examined: Do All Countries Benefit Equally?" International Monetary Fund working 
paper 13/130 (2013). [Barajas et al 2013b] 
- Beck, Thorsten, and Augusto de la Torre. "The basic analytics of access to financial 
services." Financial markets, institutions & instruments 16.2 (2007): 79-117. 
- Beck, Thorsten, and Erik Feyen. "Benchmarking Financial Systems: Introducing the 
Financial Possibility Frontier." World Bank Policy Research working paper 6615 (2013). 
- Beck, Thorsten, Aslı Demirgüç-Kunt, and Ross Levine. "Finance, inequality and the 

 40 
poor." Journal of economic growth 12.1 (2007): 27-49. 
- Beck, Thorsten, Aslı Demirgüç-Kunt, and Ross Levine. "Financial institutions and markets 
across countries and over time: The updated financial development and structure 
database." The World Bank economic review 24.1 (2010): 77-92. 
- Beck, Thorsten, Erik Feyen, Alain Ize, and Florencia Moizeszowicz. "Benchmarking 
financial development." World Bank Policy Research working paper 4638 (2008). 
- Beck, Thorsten, Ross Levine, and Norman Loayza. "Finance and the Sources of 
Growth." Journal of financial economics 58.1 (2000): 261-300. 
- Beck, Thorsten, Tao Chen, Chen Lin, and Frank Song. "Financial innovation: The bright and 
the dark sides." Hong Kong Institute for Monetary Research working paper 5 (2012) 
- Beck, Thorsten. "Bank competition and financial stability: friends or foes?" World Bank 
Policy Research working paper 4656 (2008). 
- Beck, Thorsten. "The Role of Finance in Economic Development–Benefits, Risks, and 
Politics." Oxford handbook of capitalism (2012). 
- Bencivenga, Valerie R., and Bruce D. Smith. "Financial intermediation and endogenous 
growth." The review of economic studies 58.2 (1991): 195-209. 
- Benmelech, Efraim, and Tobias J. Moskowitz. "The political economy of financial 
regulation: evidence from US state usury laws in the 19th century." The journal of 
finance 65.3 (2010): 1029-1073. 
- Blundell, Richard, and Stephen Bond. "Initial conditions and moment restrictions in dynamic 
panel data models." Journal of econometrics 87.1 (1998): 115-143. 
- Bond, Stephen R. "Dynamic panel data models: a guide to micro data methods and 
practice." Portuguese economic journal 1.2 (2002): 141-162. 
- Boyd, John H., Ross Levine, and Bruce D. Smith. "The impact of inflation on financial 
sector performance." Journal of monetary economics 47.2 (2001): 221-248. 
- Chinn, Menzie D., and Hiro Ito. "A new measure of financial openness." Journal of 
comparative policy analysis 10.3 (2008): 309-322. 
- Chinn, Menzie D., and Hiro Ito. "What matters for financial development? Capital controls, 
institutions, and interactions." Journal of development economics 81.1 (2006): 163-192. 
- Čihák, Martin, Aslı Demirgüç-Kunt, Erik Feyen, and Ross Levine. "Benchmarking financial 
systems around the world." World Bank Policy Research working paper 6175 (2012). 

 41 
- Claessens, Stijn, Gergely Dobos, Daniela Klingebiel, and Luc Laeven. "The Growing 
Importance of Networks in Finance and Its Effects on Competition." Innovations in 
Financial and Economic Networks (2003): 110-135. 
- Dabla-Norris, Era, and Narapong Srivisal. "Revisiting the Link Between Finance and 
Macroeconomic Volatility." International Monetary Fund working paper 13/29 (2013). 
- Demetriades, Panicos O., and Khaled A. Hussein. "Does financial development cause 
economic growth? Time-series evidence from 16 countries." Journal of development 
Economics 51.2 (1996): 387-411. 
- Demetriades, Panicos O., and Siong Hook Law. "Finance, institutions and economic 
development." International journal of finance & economics 11.3 (2006): 245-260. 
- Demetriades, Panicos O., and Svetlana Andrianova. "Sources and effectiveness of financial 
development: what we know and what we need to know." United Nations University 
research paper 2005/76 (2005). 
- Demirgüç-Kunt, Aslı, and Enrica Detragiache. "Cross-country empirical studies of systemic 
bank distress: a survey." National Institute economic review192.1 (2005): 68-83. 
- Demirgüç-Kunt, Aslı, and Ross Levine. "Finance, financial sector policies, and long-run 
growth." World Bank Policy Research working paper series, vol (2008). 
- Dubey, Pradeep, John Geanakoplos, and Martin Shubik. "Default and punishment in general 
equilibrium." Econometrica 73.1 (2005): 1-37. 
- Feyen, Erik, Katie Kibuuka, and Diego Sourrouille. "Fin Stats 2014: A ready-to-use tool to 
benchmark financial sectors across countries and over time." World Bank Financial Systems 
Global Practice (2013). 
- Girma, Sourafel, and Anja Shortland. "The political economy of financial development."  
Oxford economic papers 60.4 (2008): 567-596. 
- Goldsmith, Raymond W. Financial structure and economic development. Yale University 
Press, 1969. 
- Haber, Stephen, and Enrico Perotti. "The political economy of financial systems." Tinbergen 
Institute Discussion Paper 08-045/2 (2008). 
- Hansen, Lars Peter. "Large Sample Properties of Generalized Method of Moments 
Estimators." Econometrica 50.4 (1982): 1029-54. 
- Honohan, Patrick. "Financial development, growth, and poverty: how close are the links?" 
World Bank Policy Research working paper 3203 (2004). 

 42 
- Jalilian, Hossein, and Colin Kirkpatrick. "Does financial development contribute to poverty 
reduction?" Journal of development studies 41.4 (2005): 636-656. 
- Jalilian, Hossein, and Colin Kirkpatrick. "Financial development and poverty reduction in 
developing countries." International journal of finance & economics 7.2 (2002): 97-108. 
- Jeanneney, Sylviane Guillaumont, and Kangni Kpodar. "Financial development and poverty 
reduction: Can there be a benefit without a cost?" The Journal of development studies 47.1 
(2011): 143-163. 
- King, Robert G., and Ross Levine. "Finance and Growth: Schumpeter Might be Right." The 
quarterly journal of economics 108.3 (1993): 717-737. 
- La Porta, Rafael, Florencio Lopez-De-Silanes, Andrei Shleifer, and Robert W. Vishny. 
"Legal Determinants of External Finance." Journal of finance (1997): 1131-1150. 
- Levine, Ross, Norman Loayza, and Thorsten Beck. "Financial intermediation and growth: 
Causality and causes." Journal of monetary economics 46.1 (2000): 31-77. 
- Levine, Ross, and Sara Zervos. "Stock markets, banks, and economic growth." American 
economic review 88.3 (1998): 537-558. 
- Levine, Ross. "Finance and growth: theory and evidence." Handbook of economic growth 1 
(2005): 865-934. 
- Levine, Ross. "Financial development and economic growth: views and agenda." Journal of 
economic literature (1997): 688-726. 
- Levine, Ross. "The legal environment, banks, and long-run economic growth." Journal of 
money, credit and banking (1998): 596-613. 
- Lucas Jr, Robert E. "On the mechanics of economic development." Journal of the monetary 
economics 22.1 (1988): 3-42. 
- McKinnon, Ronald I. Money and capital in economic development. Brookings Institution 
Press, 1973. 
- Miller, Merton H. "Financial markets and economic growth." Journal of applied corporate 
finance 11.3 (1998): 8-15. 
- Nickell, Stephen J. "Biases in Dynamic Models with Fixed Effects." Econometrica 49.6 
(1981): 1417-26. 
- Peria, Maria Soledad Martinez, and Ashoka Mody. "How foreign participation and market 
concentration impact bank spreads: evidence from Latin America." Journal of money, credit 
and banking (2004): 511-537. 

 43 
- Rajan, Raghuram G. "Has finance made the world riskier?" European financial 
management 12.4 (2006): 499-533. 
- Rajan, Raghuram G., and Luigi Zingales. "Financial Dependence and Growth." American 
economic review 88.3 (1998): 559-586. 
- Rajan, Raghuram G., and Luigi Zingales. "The great reversals: the politics of financial 
development in the twentieth century." Journal of financial economics 69.1 (2003): 5-50. 
- Rioja, Felix, and Neven Valev. "Does one size fit all?: a reexamination of the finance and 
growth relationship." Journal of development economics 74.2 (2004): 429-447. 
- Roodman, David. "How to do xtabond2: An introduction to difference and system GMM in 
Stata." Stata journal 9.1 (2009): 86. 
- Rousseau, Peter L., and Paul Wachtel. "Inflation thresholds and the finance–growth 
nexus." Journal of international money and finance 21.6 (2002): 777-793. 
- Sargan, J. D. "The Estimation of Economic Relationships using Instrumental 
Variables." Econometrica 26.3 (1958): 393-415. 
- Schumpeter, Joseph Alois. The theory of economic development. Harvard University 
Press, 1911. 
- Shaw, Edward S. Financial deepening in economic development. Oxford University 
Press, 1973. 
- Shubik, Martin, and Charles Wilson. "The optimal bankruptcy rule in a trading economy 
using fiat money." Journal of economics 37.3 (1977): 337-354. 

*References list from _wp15173 - REFERENCES*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2015/_wp15173.pdf_