## _wp1681

## Source details

**Canonical URL:** [_wp1681](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2016/_wp1681.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2016/_wp1681.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2016/_wp1681.pdf.json)

---

### I. Introduction and objectives
- Growth has slowed in Latin America and the Caribbean (LAC), and the region’s medium-term growth prospects have been marked down.
- Deepening financial systems through better market access, liquidity, and diversity of instruments can:
  - unleash new growth sources,
  - better reap the benefits from globalization,
  - transition to higher income levels.
- Paper objectives and contributions:
  - Construct a measure of financial development for a large sample of advanced, emerging, and developing countries.
  - Estimate financial development gaps relative to current fundamentals.
  - Evaluate the relationship of financial development with growth and stability.
  - Contributions relative to prior literature include:
    - a consistent empirical framework to estimate financial development gaps and long-term relationships between growth/stability and financial development;
    - refinement of the index of financial development constructed in IMF (Sahay and others 2015a);
    - improved specification of the growth/stability regressions;
    - use of a more comprehensive stability measure;
    - exploration of the state of financial development in LAC.

### II. Measuring financial development — index construction and data
- Index structure:
  - Two major components: financial institutions (FI) and financial markets (FM).
  - Each component broken into sub-components: access, depth, and efficiency.
  - Sub-components built from underlying variables tracking development in each area.
- Normalization formulas (individual variables normalized to range 0–1):
  - For variables where increase indicates development:
    - I_xit = (x_it - min(x_it)) / (max(x_it) - min(x_it))
  - For variables where decrease indicates development (Interest Rate Spread, Bank Asset Concentration, Overhead Costs, Net Interest Margin, Non-Interest Income):
    - I_xit = 1 - (x_it - min(x_it)) / (max(x_it) - min(x_it))
- Aggregation and weighting:
  - Weights estimated using five methods: principal component (levels and differences), factor analysis (levels and differences), and equal weights within subcomponent.
  - Aggregation results robust across methods; index with equal weights is used for simplicity.
- Database and coverage:
  - Database includes 122 countries for 1995–2013.
  - Annual data availability varies by series; some series (e.g., ATMs per thousands of adults) available only from 2004 in many cases.
- Key index variables (selected examples):
  - Institutions — Access: Bank branches per 100,000 adults; ATMs per 100,000 adults; Domestic Bank Deposits / GDP (%)
  - Institutions — Depth: Insurance Company Assets / GDP (%); Mutual Fund Assets / GDP (%); Domestic Credit to Private Sector / GDP (%)
  - Institutions — Efficiency: Bank concentration (%); Bank lending-deposit spread; Overhead cost/total assets; Bank net interest Margin; Non-interest income/total income
  - Markets — Access/Depth/Efficiency: Total number of issuers of debt; Market capitalization excluding top 10 companies to total market capitalization; Stock market capitalization to GDP; Stock market total value traded to GDP (%); Stock of government debt securities in % of GDP; Debt securities of financial sector by local firms in % of GDP; Debt securities of non-financial sector by local firms in % of GDP; Stock market turnover ratio (value traded/stock market capitalization)
- Data processing and cleaning:
  - Missing observations supplemented with national sources and IMF FSAPs; linear interpolation used to fill gaps when necessary.
  - Exponential smoothing applied to extrapolate backwards to 1995 for every variable per country.
  - Forward extrapolation used last observation available for missing forward values.
  - Outlier detection: two-stage procedure using HP filter (lambda = 6.25) for trend/cycle and a 3 standard deviation rule on gaps; secondary check on growth rates when outliers cluster; outliers dropped and replaced with linear interpolation.
- Coverage summary (Table A2 reported observations):
  - East Asia Pacific 13247
  - Europe and Central Asia 43817
  - Latin America 22418
  - Middle East and North Africa 12228
  - North America 238
  - South Asia 6114
  - Sub-Sahara Africa 24456
  - Total 1222,318

### III. Where LAC stands on financial development
- Overall assessment:
  - LAC compares unfavorably with other emerging markets (EMs); only low-income countries (LICs) lag behind LAC.
  - LAC scores higher on financial institutions than on financial markets, a pattern shared with LICs.
  - LAC lags other EM regions on depth and efficiency of financial institutions and on all sub-components of financial market development.
  - LAC excels relative to other EMs in access to financial institutions (bank and ATM networks), but usage by households remains low.
- Cross-country variation and notable country facts:
  - Chile:
    - Domestic bond market represents almost 40 percent of GDP.
    - Market value of listed companies in the equity market is about 90 percent of GDP.
    - Reforms since mid-1970s and a fully-funded pension system supported institutional investor demand and market depth.
  - Brazil:
    - Insurance company assets to GDP more than doubled in the past decade.
    - Mutual fund assets grew from 30 percent of GDP to 50 percent of GDP.
    - Brazil became sixth in the world (excluding financial centers) by mutual fund assets.
    - Government debt management reforms lengthened maturities and built yield curve benchmarks.
  - Mexico:
    - Post-1994 reforms strengthened regulations, reformed deposit insurance, improved collateral execution and information sharing among credit bureaus.
    - Reforms promoted financial education and competition; contributed to acceleration in credit growth given low credit to GDP ratio.
  - Other LAC progress:
    - Colombia and Peru significantly developed financial institutions; number of commercial bank branches more than quadrupled.
    - Ecuador saw dramatic growth in bank branches driven by expansion of two large banks and conversions of cooperatives into commercial banks.
    - Market-side progress limited in most LAC countries, except for the Bahamas and El Salvador.

### IV. Financial development gaps and their drivers in LAC
- Definition:
  - Gaps = actual composite index minus predicted norms based on fundamentals (income per capita, government size, macro stability, etc.).
- Regression performance fragment:
  - "The regressions explain a large portion of the variation in financial development, with R-squares of 0.74 and" (text fragment preserved).
  - Additional reported fit: 0.61 for institutions and market regressions, respectively.
- Examples of negative gaps and drivers:
  - Dominican Republic (financial crisis in 2003): erosion of trust in financial institutions and depressed demand for credit.
  - Uruguay (banking crisis in 2002): low access to financial institutions and markets.
  - Peru: negative efficiency gap linked to weak frameworks for obtaining or seizing collateral.
  - Jamaica: negative efficiency gap reflecting high bank concentration and historical investment dependence on low risk government debt, hindering banks’ capacity for risk assessment and driving up spreads.
- Examples of positive gaps and potential risks:
  - Bolivia: regulated interest rates and credit quotas may pose risks to banks’ profitability and generate inefficient allocation of credit.
  - Honduras: rapid credit growth beyond fundamentals has largely fueled consumption due to scant investment opportunities.
  - Several Central American cases: small, illiquid stock markets with few listed firms, weak legal/contractual infrastructure, and low affordability for domestic companies.
- Recommendation summary:
  - Countries should explore causes behind financial development gaps and tailor policies to address country-specific distortions—macroeconomic fundamentals are often difficult to change in the short term.

### V. Empirical framework: financial development, growth, and stability
- Empirical approach:
  - Dynamic panel regression using 5-year averages and Arellano-Bond/system GMM to address dynamics and endogeneity.
  - Dependent variables:
    - Financial instability (itFS): first principal component of inverse distance to distress (z-score), real credit growth volatility, real and nominal interest rate volatility.
    - Growth volatility (itGV): standard deviation of GDP growth.
    - Real GDP growth (itYD).
  - Controls included benchmarking regressions plus ratio of FDI to GDP, capital account openness, initial income per capita, government consumption to GDP, trade openness, changes in terms of trade, growth in per capita income, capital flows to GDP, exchange rate regime, political stability, and offshore financial center indicator.
  - Two forms of financial development variables:
    - Aggregated index specification: fFinDevit = FD1it b1 + FD2it b2
    - Subcomponents: FI and FM with interaction FM*FI
- Non-linear hypotheses:
  - Financial development may have diminishing returns and thresholds: initial deepening lowers instability and raises growth, but beyond a turning point further development can increase instability and lower marginal growth benefits.
  - Mechanisms include diversion to speculative activities, excessive leverage, and risk-taking when regulation/supervision are inadequate.

### VI. Regression evidence — key coefficients and findings (selected)
- Significance notation: *** p<0.01, ** p<0.05, * p<0.1.
- Selected estimated coefficients (standard errors in parentheses) from Table 1:
  - Financial Instability dependent variable:
    - FD: -6.457* (3.814)
    - FD2: 6.263 (5.735)
    - ∆ FD: 5.283** (2.160)
    - ∆ FI: 4.753** (2.114)
    - ∆ FM: 3.190* (1.672)
  - Growth Volatility dependent variable:
    - FD: -21.42*** (7.270)
    - FD2: 23.74** (10.82)
    - ∆ FD: 8.423** (4.008)
    - FI: -13.75** (5.419)
    - FI2: 18.64** (8.123)
    - FM: -0.772 (3.119)
    - FM2: 3.360 (4.886)
    - FM*FI: -5.140 (9.730)
    - ∆ FI: 14.08*** (3.708)
    - ∆ FM: -2.335 (2.846)
  - Growth dependent variable:
    - FD: 11.47* (6.279)
    - FD2: -12.38* (6.556)
    - ∆ FD: 5.698* (3.075)
    - FI: -27.89*** (9.533)
    - FI2: 36.38** (14.45)
    - FM: -6.779 (5.345)
    - FM2: 18.02** (8.324)
    - FM*FI: 27.27** (13.16)
    - ∆ FI: 7.088** (2.958)
    - ∆ FM: 0.508 (2.222)
- Observations reported in Table 1: 143, 143, 158, 158, 301, 301.
- Main empirical findings:
  - Non-linear (inverted-U or threshold) relationships between financial development and both growth and stability.
  - Financial development initially lowers macroeconomic instability, but beyond a turning point further development can increase instability.
  - Non-linearity particularly pronounced for institutional depth and growth.
  - Financial services efficiency shows a more linear positive relationship with growth, though stability costs may arise if efficiency reduces bank profitability and incentivizes risk-taking.
  - Financial institution development effects are robust; financial market development effects are weaker at low levels but become important at higher development levels and are complementary to institutions for raising growth.
  - Excessive market development at early stages of institutional development may harm stability.

### VII. Interpretations, country positioning, and caveats
- Most LAC countries have not yet reached the turning point where marginal growth dividends from additional financial development become negative.
- Brazil and Chile are nearest this “optimum” level of financial development; Dominican Republic, Paraguay, and Honduras are on the opposite side of the spectrum.
- Empirical relationships assume other growth determinants held constant; causality is difficult to disentangle despite use of instrumental variables (rule of law and legal origin dummies) and system GMM to mitigate endogeneity.

### VIII. Policy implications and recommendations
- General building blocks for well-functioning financial systems:
  - (i) strong property rights;
  - (ii) efficient legal system;
  - (iii) low incidence of corruption;
  - (iv) sufficient financial information;
  - (v) good corporate governance;
  - (vi) sound prudential regulation and supervision of the banking system.
- Policy guidance by country circumstance:
  - Crisis-recovering countries:
    - improve credibility of financial systems;
    - strengthen capital and liquidity buffers;
    - ensure credible deposit insurance;
    - address balance-sheet mismatches (example: reforms in Mexico after 1994 crisis).
  - Countries with negative gaps in institutional depth and efficiency (e.g., Dominican Republic, Jamaica, Peru):
    - strengthen property rights and collateral frameworks;
    - improve efficiency of courts and credit reporting systems.
  - Countries with underdeveloped bond markets (e.g., Costa Rica, Uruguay):
    - adopt market-friendly debt management and issuance strategies to foster secondary markets (use of standardized simple instruments with conventional maturities);
    - strengthen legal and regulatory frameworks.
  - Countries with underdeveloped or inefficient stock markets (majority of LAC):
    - strengthen macroeconomic environment, institutional and legal frameworks;
    - promote investor rights and information disclosure;
    - consider policies to increase market size (pension reforms, carefully sequenced financial liberalization, corporate governance and tax reforms);
    - note that for smaller economies domestic equity markets may not be justified.
  - Countries with positive development gaps:
    - enhance supervisory vigilance to improve credit quality and underwriting;
    - strengthen macroprudential policy frameworks.
- Sequencing and prudential caveats:
  - Gradual sequencing recommended: secure institutional development before promoting rapid market development.
  - Regulation and supervision should be developed consistent with existing financial development levels and flexible to future deepening.
  - Careful management of financial integration: coordinate across supervisory and regulatory agents to capture benefits (economies of scale, risk diversification) while containing risks.

### IX. Costa Rica — conclusions and specific recommendations
- Key conclusions for Costa Rica:
  - Costa Rica’s financial system deepened notably in the past decade, but continues to lag behind other emerging markets and the level implied by its macroeconomic fundamentals.
  - Given sticky fundamentals in the short term, Costa Rica should remove distortions that prevent the country from reaching its financial development potential.
  - In the longer term, as fundamentals evolve toward higher income per capita, further financial development would be advantageous provided adequate regulatory oversight.
- Institutional and collateral reforms (short term):
  - Follow through on modernization of the collateral framework while balancing with proper regulation and supervision.
  - 2015 secured transactions law features (as reported):
    - establishes a functional secured transactions system and a modern, centralized, notice-based collateral registry;
    - broadened the range of assets that can be used as collateral, including intangibles such as intellectual property rights;
    - allowed a general description of assets granted as collateral;
    - permitted out-of-court enforcement of collateral.
  - Careful monitoring warranted to hinder abuse as the new system is tested.
- Financial intermediation and bank structure:
  - Improve efficiency by ensuring a level playing field for private banks versus public banks.
  - Remove the explicit guarantee currently given by the state to all colon-denominated deposits in state banks.
- Debt management and government securities markets:
  - Use standardized simple instruments with conventional maturities and strengthen legal/regulatory frameworks to foster secondary markets.
- Stock market and institutional investor development:
  - Promote a more robust macroeconomic environment and stronger institutional/legal frameworks to promote investor rights and information disclosure.
  - Encourage development of an institutional investor base; allow private participation in sectors previously state monopolies (example: recent insurance sector reform).
  - Strengthen protection of minority investors (Costa Rica does not score well in Doing Business indicators).
  - Review tax treatment of securities issuance and investment to make the tax system more attractive to issuers, consistent with fiscal sustainability.

*Source: IMF staff calculations and analysis as presented in the provided content.*

### REFERENCES ___________________________________________________________30

### _wp1681 - REFERENCES ___________________________________________________________30

### I. INTRODUCTION
- Growth has slowed in Latin America and the Caribbean (LAC), and the region’s medium-term growth prospects have been marked down.
- Even during the most recent commodity super-cycle, LAC’s growth performance improved relative to the previous decade but still underperformed other emerging market (EM) regions.
- Deepening financial systems through better market access, liquidity, and diversity of instruments can help:
  - unleash new growth sources,
  - better reap the benefits from globalization,
  - transition to higher income levels.
- Paper objectives and contributions:
  - Construct a measure of financial development for a large sample of advanced, emerging, and developing countries.
  - Estimate financial development gaps relative to current fundamentals.
  - Evaluate the relationship of financial development with growth and stability.
  - Contributions compared to existing literature:
    - (i) consistent empirical framework to estimate financial development gaps and long-term relationships between growth/stability and financial development;
    - (ii) refinement of the index of financial development constructed in IMF (Sahay and others 2015a);
    - (iii) improved specification of the growth/stability regressions;
    - (iv) use of a more comprehensive stability measure than in prior literature;
    - (v) exploration of the state of financial development in LAC.

### II. MEASURING FINANCIAL DEVELOPMENT
- Traditional proxies are narrow: private credit to GDP, liquid liabilities to GDP, stock market capitalization, turnover ratio.
- Non-bank financial institutions (pension funds, insurance companies, mutual funds) have grown significantly, expanding opportunities for consumption smoothing, investment funding, and risk diversification (see Figure 1).
- The IMF index (Sahay and others 2015a) is employed with modifications:
  - Two major components: financial institutions (FI) and financial markets (FM).
  - Each component broken into sub-components: access, depth, and efficiency.
  - Sub-components built from underlying variables tracking development in each area (see Figure 2).
- Data coverage and processing:
  - Database includes 122 countries for 1995–2013.
  - Appendix 1 describes data processing and transformations.
- Normalization of individual variables into an index ranging between zero and 1 using:
  - For variables where increase indicates development:
    - I_xit = (x_it - min(x_it)) / (max(x_it) - min(x_it))
  - For variables where decrease indicates development (Interest Rate Spread, Bank Asset Concentration, Overhead Costs, Net Interest Margin, Non-Interest Income):
    - I_xit = 1 - (x_it - min(x_it)) / (max(x_it) - min(x_it))
- Aggregation:
  - Weights estimated using five methods: principal component (levels and differences), factor analysis (levels and differences), and equal weights within subcomponent.
  - Aggregation results robust across methods.
  - For simplicity, index with equal weights is used.
- Index components and example variables (as presented):
  - Institutions — Access: Bank branches per 100,000 adults; ATMs per 100,000 adults; Domestic Bank Deposits / GDP (%)
  - Institutions — Depth: Insurance Company Assets / GDP (%); Mutual Fund Assets / GDP (%); Domestic Credit to Private Sector / GDP (%)
  - Institutions — Efficiency: Bank concentration (%); Bank lending-deposit spread; Overhead cost/total assets; Bank net interest Margin; Non-interest income/total income
  - Markets — Access/Depth/Efficiency: Total number of issuers of debt; Market capitalization excluding top 10 companies to total market capitalization; Stock market capitalization to GDP; Stock market total value traded to GDP (%); Stock of government debt securities in % of GDP; Debt securities of financial sector by local firms in % of GDP; Debt securities of non-financial sector by local firms in % of GDP; Stock market turnover ratio (value traded/stock market capitalization)
- Robustness note:
  - There are striking differences between the composite index and traditional measures (Figure 3), illustrated by country examples (Honduras, Trinidad and Tobago).

### III. FINANCIAL DEVELOPMENT: WHERE DOES LAC STAND?
- Overall assessment:
  - LAC compares unfavorably with other emerging markets (EMs) on financial development; only low-income countries (LICs) lag behind LAC (Figure 4).
- Component-wise findings:
  - LAC scores higher on financial institutions than on financial markets, a pattern shared with LICs.
  - LAC lags other EM regions on depth and efficiency of financial institutions and on all sub-components of financial market development.
  - LAC excels relative to other EMs in access to financial institutions, reflecting emphasis on financial inclusion via improved bank and ATM networks.
  - Despite access improvements, LAC still lags other EM regions on the level of usage of financial services by households.
- Cross-country variation within LAC (Figure 5):
  - Chile and Brazil rank highest in development of financial markets and financial institutions, respectively.
  - Peru, Colombia, and Mexico follow; Mexico has made major strides since its 1994 crisis.
- Country highlights and exact figures:
  - Chile:
    - Domestic bond market represents almost 40 percent of GDP.
    - Market value of listed companies in the equity market is about 90 percent of GDP.
    - Reforms since mid-1970s and a fully-funded pension system supported institutional investor demand and market depth.
  - Brazil:
    - Insurance company assets to GDP more than doubled in the past decade.
    - Mutual fund assets grew from 30 percent of GDP to 50 percent of GDP.
    - Brazil became sixth in the world (excluding financial centers) by mutual fund assets.
    - Government debt management reforms lengthened maturities and built yield curve benchmarks, contributing to market development.
  - Mexico:
    - Post-1994 reforms strengthened regulations, reformed deposit insurance, improved collateral execution and information sharing among credit bureaus.
    - Reforms promoted financial education and competition; contributed to acceleration in credit growth given low credit to GDP ratio.
  - Other LAC progress:
    - Colombia and Peru significantly developed financial institutions; the number of commercial bank branches more than quadrupled.
    - Ecuador saw dramatic growth in bank branches driven by expansion of two large banks and conversions of cooperatives into commercial banks.
    - Market-side progress in most LAC countries was limited, except for the Bahamas and El Salvador.

### IV. FINANCIAL DEVELOPMENT AND MACROECONOMIC FUNDAMENTALS
- Alignment assessment:
  - For most LAC countries, current financial development is not fully aligned with macroeconomic fundamentals.
- Financial development gaps:
  - Defined as deviation of the composite index from a prediction based on fundamentals (income per capita, government size, macro stability, etc.).
  - Gaps can identify potential distortions or sources of under-/over-development.
  - Analysis is suggestive; normative assessment of links to growth and stability presented in next sections (not included here).
- Regression specification summary:
  - Fundamentals (X_it) included:
    - initial income per capita,
    - government consumption to GDP,
    - inflation,
    - trade openness,
    - educational attainment proxied by average years of secondary schooling for people 25+,
    - population growth,
    - capital account openness,
    - size of the shadow economy,
    - rule of law.
  - Instruments (Z_it) used, e.g., rule of law and legal origin dummies.
  - Predicted norms computed using equation:
    - FI_it = X_it' δ + Z_it' δ' + ε_it  (as presented in the source text)
  - Gaps (Figure 5) are actual index minus calculated norms.
- Regression performance note (text fragment preserved):
  - "The regressions explain a large portion of the variation in financial development, with R-squares of 0.74 and"

### FIGURES AND TABLES (as listed in the source)
- Figures (titles and placement):
  - Figure 1: Non Bank Assets (Regional averages in percent of GDP)
  - Figure 2: How to Measure Financial Development
  - Figure 3: Composite Financial Development Index vs. Traditional Measures, 2013
  - Figure 4: Inter-Regional Variation in Financial Development
  - Figure 5: Financial Development Progress and Remaining Gaps
  - Figure 6: Financial Institutions and Markets Development, and Economic Growth
  - Figure 7: Financial Development, Growth, and Stability
- Tables and Appendices:
  - Table 1. Estimated Equations
  - Appendix 1: Data Description and Processing
  - Appendix 2: An Example of Framework Application to Costa Rica

*Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2016/_wp1681.pdf*

### 0.61 for institutions and market regressions, respectively. Nonetheless, the lack of a solid theory on the factors

### _wp1681 - 0.61 for institutions and market regressions, respectively. Nonetheless, the lack of a solid theory on the factors

### Financial development gaps in Latin America and the Caribbean (LAC)
- Shortfalls are common in institutional efficiency and depth as well as market access and efficiency.
- Examples of negative gaps and their drivers:
  - Dominican Republic (financial crisis in 2003): erosion of trust in financial institutions and depressed demand for credit.
  - Uruguay (banking crisis in 2002): low access to financial institutions and markets.
  - Peru: negative efficiency gap linked to weak frameworks for obtaining or seizing collateral.
  - Jamaica: negative efficiency gap reflecting high bank concentration and historical investment dependence on low risk government debt, hindering banks’ capacity for risk assessment and driving up spreads.
- Positive gaps can indicate potential excess or inefficiency:
  - Bolivia: regulated interest rates and credit quotas may pose risks to banks’ profitability and generate inefficient allocation of credit.
  - Honduras: rapid credit growth beyond fundamentals has largely fueled consumption due to scant investment opportunities.
  - Several Central American cases: stock markets have few listed firms, little trading activity, weak legal/contractual infrastructure, and are not viewed as affordable financing sources by most domestic companies.
- Recommendation: countries should explore causes behind financial development gaps and tailor policies to address country-specific distortions—macroeconomic fundamentals are often difficult to change in the short term.

### Nexus between financial development, stability, and growth — empirical framework
- Financial development generally positively related to economic growth; however, diminishing returns and thresholds exist where further deepening generates decreasing returns to growth and stability.
- Potential mechanisms for non-linear effects:
  - Large financial systems can divert resources to speculative and risky financial investments (Minsky, 1975).
  - Excessive leverage and risk-taking can increase economic and financial volatility, especially with inadequate regulation/supervision.
- Empirical approach:
  - Dynamic panel regression framework using 5-year averages and Arellano-Bond/system GMM techniques to address dynamics and endogeneity.
  - Dependent variables:
    - Financial instability (itFS): first principal component of inverse distance to distress (z-score), real credit growth volatility, real and nominal interest rate volatility.
    - Growth volatility (itGV): standard deviation of GDP growth.
    - Real GDP growth (itYD).
  - Controls included same as benchmarking regressions plus ratio of FDI to GDP and capital account openness; other controls for instability and volatility included initial income per capita, government consumption to GDP, trade openness, changes in terms of trade, growth in per capita income, capital flows to GDP, exchange rate regime, political stability, and offshore financial center indicator.
  - Two forms of financial development variables:
    - Aggregated index: fFinDevit = FD1it b1 + FD2it b2
    - Subcomponents: FI and FM with interaction FM*FI

### Regression evidence and key coefficients (Table 1)
- Estimated equations (selected coefficients and standard errors in parentheses):
  - For Financial Instability dependent variable:
    - FD: -6.457* (3.814)
    - FD2: 6.263 (5.735)
    - ∆ FD: 5.283** (2.160)
    - ∆ FI: 4.753** (2.114)
    - ∆ FM: 3.190* (1.672)
  - For Growth Volatility dependent variable:
    - FD: -21.42*** (7.270)
    - FD2: 23.74** (10.82)
    - ∆ FD: 8.423** (4.008)
    - FI: -13.75** (5.419)
    - FI2: 18.64** (8.123)
    - FM: -0.772 (3.119)
    - FM2: 3.360 (4.886)
    - FM*FI: -5.140 (9.730)
    - ∆ FI: 14.08*** (3.708)
    - ∆ FM: -2.335 (2.846)
  - For Growth dependent variable:
    - FD: 11.47* (6.279)
    - FD2: -12.38* (6.556)
    - ∆ FD: 5.698* (3.075)
    - FI: -27.89*** (9.533)
    - FI2: 36.38** (14.45)
    - FM: -6.779 (5.345)
    - FM2: 18.02** (8.324)
    - FM*FI: 27.27** (13.16)
    - ∆ FI: 7.088** (2.958)
    - ∆ FM: 0.508 (2.222)
- Observations: 143, 143, 158, 158, 301, 301 (as reported in Table 1).
- Significance notation: *** p<0.01, ** p<0.05, * p<0.1.
- Main empirical findings:
  - Non-linear (inverted-U or threshold) relationships between financial development and both growth and stability.
  - Financial development initially lowers macroeconomic instability, but beyond a turning point further development can increase instability.
  - Non-linearity particularly pronounced for institutional depth and growth.
  - Financial services efficiency has a more linear positive relationship with growth, though stability costs may arise if efficiency reduces bank profitability and incentivizes risk-taking.
  - Financial institution development effects are robust across regressions; financial market development effects are weaker at low levels but become important at higher development levels and are complementary to institutions for raising growth.
  - Too much market development at early stages of institutional development may harm stability (market volatility dominates when institutions are weak).

### Interpretations and country positioning
- Most LAC countries have not yet reached the turning point where marginal growth dividends from additional financial development become negative.
- Brazil and Chile are nearest this “optimum” level of financial development; Dominican Republic, Paraguay, and Honduras are on the opposite side of the spectrum.
- Empirical relationships assume other growth determinants are held constant; causality is difficult to disentangle despite instrumental variables (rule of law and legal origin dummies) and system GMM used to mitigate endogeneity.

### Policy implications and recommendations
- General building blocks for well-functioning financial systems: (i) strong property rights; (ii) efficient legal system; (iii) low incidence of corruption; (iv) sufficient financial information; (iv) good corporate governance; (v) sound prudential regulation and supervision of the banking system.
- Policy guidance by country circumstance:
  - Crisis-recovering countries: improve credibility of financial systems, strengthen capital and liquidity buffers, ensure credible deposit insurance, address balance-sheet mismatches (example: reforms in Mexico after 1994 crisis).
  - Countries with negative gaps in institutional depth and efficiency (e.g., Dominican Republic, Jamaica, Peru): strengthen property rights and collateral frameworks; improve efficiency of courts and credit reporting systems.
  - Countries with underdeveloped bond markets (e.g., Costa Rica, Uruguay): adopt market-friendly debt management and issuance strategies to foster secondary markets (use of standardized simple instruments with conventional maturities); strengthen legal and regulatory frameworks.
  - Countries with underdeveloped or inefficient stock markets (majority of LAC): strengthen macroeconomic environment, institutional and legal frameworks; promote investor rights and information disclosure; consider policies increasing market size (pension reforms, carefully sequenced financial liberalization, corporate governance and tax reforms); note that for smaller economies domestic equity markets may not be justified.
  - Countries with positive development gaps: enhance supervisory vigilance to improve credit quality and underwriting, and strengthen macroprudential policy frameworks.
- Sequencing and prudential caveats:
  - Gradual sequencing recommended: secure institutional development before promoting rapid market development.
  - Regulation and supervision should be developed consistent with existing financial development levels and flexible to future deepening.
  - Careful management of financial integration: benefits include regional economies of scale and risk diversification; risks require coordination across supervisory and regulatory agents.

*Source: IMF staff calculations and analysis as presented in the provided content.*

### APPENDIX 1: DATA DESCRIPTION AND PROCESSING

### APPENDIX 1: DATA DESCRIPTION AND PROCESSING

### Sources and data gathering
- Data sources are listed in Table A1 and include: World Bank (WDI, FinStats, Global Financial Development), IMF (Financial Access Survey, IFS, FSAPs), Dealogic, BIS (Debt securities statistics), World Federation of Exchanges, Bankscope, Federal Reserve Bank of St. Louis, and national sources (central banks and national statistical agencies).
- Annual data were available starting in the 1960s for some series, but "for nearly all of the countries the data were only available from mid 1990s."
- The period of analysis is set from 1995 to 2013.
- Some variables have more limited coverage; for example, ATMs per thousands of adults data are available only from 2004 in many cases.
- For countries with missing data, supplementation used national sources and IMF FSAPs; the Federal Reserve Bank of St. Louis and the World Federation of Exchanges were used alternatively.

### Index components and variables (as reported in Table A1)
- Automated teller machines (ATMs) (per 100,000 adults) — World Bank, WDI
- Number of Branches Per 100,000 Adults, Commercial Banks — World Bank, FinStats; IMF, Financial Access Survey
- Domestic credit to private sector / GDP (%) — World Bank, WDI
- Mutual Fund Assets / GDP (%) — World Bank, FinStats and NBFI database
- Insurance Company Assets / GDP (%) — World Bank, FinStats and NBFI database
- Domestic Bank Deposits / GDP (%) — World Bank, FinStats and WDI; IMF, IFS
- Interest rate spread (lending rate minus deposit rate, %) — World Bank, WDI
- Bank net interest margin (%) — World Bank, Global Financial Development (GFD)
- Non-Interest Income / Total income (%) — World Bank, FinStats; Bankscope
- Overhead Costs / Total Assets (%) — World Bank, FinStats; Bankscope
- 3 Bank Asset Concentration (%) — World Bank, FinStats; Bankscope
- Total number of issuers of debt (domestic and external, NFCs and Financial) — Dealogic, supplement from SPR
- Market capitalization excluding top 10 companies to total market capitalization (%) — World Bank, GFD; World Federation of Exchanges
- Market capitalization of listed companies (% of GDP) — World Bank, WDI
- Stocks traded, total value (% of GDP) — World Bank, WDI
- Outstanding International Public Debt Securities / GDP (%) — BIS, Debt securities statistics
- Debt securities of financial sector by sub nationality in % of GDP — Dealogic
- Debt securities of non-financial sector by sub nationality in % of GDP — Dealogic
- Efficiency: Stock market turnover ratio (value traded/stock market capitalization) — World Bank, WDI

### Building a consistent dataset: supplementation, interpolation, and extrapolation
- To obtain a balanced panel and longest possible sample, the authors:
  - Supplemented missing observations with national sources and IMF FSAPs.
  - Used linear interpolation to fill gaps when no supplementary data were available (described as "the same as to fill the gaps with a linear projection between two available data points").
- Outliers were identified and treated (see next subsection).
- After cleaning and interpolation, the dataset was further extrapolated backwards to 1995 using exponential smoothing.
  - Exponential smoothing was applied "to every variable per country back to 1995."
  - The authors note they "also tried other methods instead, but this last seemed to be the most conservative of all."
- Forward extrapolation: missing values forward were replaced with the last observation available in the sample; typically only one year forward required extrapolation.

### Outlier detection and treatment
- Two-stage outlier identification that accounts for trending series:
  1. For each country-variable series, estimate the trend component and calculate the gap between actual and trend. The trend and cycle estimation uses a Hodrick-Prescott (HP) filter with a lambda equal to 6.25.
     - An outlier is defined as a point in time that is 3 standard deviations above or below the calculated gap.
  2. For rare cases where 2 or 3 outliers cluster and pull the trend, a secondary procedure is applied: calculate the growth rate of the variable and define an outlier as a point where the growth rate exceeds 3 standard deviations of the sample period.
- All detected outliers were dropped and replaced using linear interpolation.

### Aggregation and index construction
- The cleaned, gap-filled, and extrapolated dataset was aggregated using principal component and factor analysis as well as equal weights.
- Factor weights are identified as theoretically most appropriate for capturing the common factor underlying variable movements, but the results across weighting methods were "very similar."
- For simplicity, results are presented for the index constructed with equal weights.

### Differences with Sahay and others 2015a
- The main differences highlighted:
  - This index is supplemented with additional data from IMF Financial Sector Assessment Programs and country reports and is extended back to 1995, while Sahay and others 2015a extend missing series back to 1980.
  - Exponential extrapolation is used here in contrast to average growth rates of all non-missing observations used in Sahay and others 2015a.
  - Outlier treatment differs: this paper eliminates observations that exceed three standard deviations above or below the trend and fills resulting gaps with simple linear interpolation; Sahay and others 2015a eliminate observations below the 5th and above the 95th percentiles and use winsorization to replace outliers.
  - This paper uses equal weights for the final index while Sahay and others 2015a employ principal component analysis for index construction.

### Coverage summary (Table A2: Regions and Coverage)
- Region-level reported observations (as presented in Table A2):
  - East Asia Pacific 13247
  - Europe and Central Asia 43817
  - Latin America 22418
  - Middle East and North Africa 12228
  - North America 238
  - South Asia 6114
  - Sub-Sahara Africa 24456
  - Total 1222,318

*Source: APPENDIX 1: DATA DESCRIPTION AND PROCESSING (from the provided IMF content).*

### Conclusions and Policy Recommendations

### Conclusions and Policy Recommendations

### Key conclusions
- Costa Rica’s financial system deepened notably in the past decade, but continues to lag behind those of other emerging markets as well as the level of development implied by its macroeconomic fundamentals.
- Given that the fundamentals are sticky in the short term, Costa Rica should aim at removing distortions that prevent the country from reaching its full financial development potential given the current state of macroeconomic fundamentals.
- In the longer term, as fundamentals continue to evolve, including toward a higher income per capita, further financial development would be advantageous for Costa Rica in terms of growth and stability, provided there is adequate regulatory oversight to prevent excesses.

### Institutional and collateral reforms (short term)
- Follow through on the modernization of the collateral framework while balancing it with proper regulation and supervision.
  - In 2015, the country adopted a new secured transactions law that:
    - establishes a functional secured transactions system and a modern, centralized, notice-based collateral registry;
    - broadened the range of assets that can be used as collateral, including intangibles such as intellectual property rights;
    - allowed a general description of assets granted as collateral; and
    - permitted out-of-court enforcement of collateral.
  - Careful monitoring is warranted at this stage to hinder abuse as the new system is being tested.

### Financial intermediation and bank structure
- Improve efficiency of financial intermediation by ensuring a level playing field for private banks compared to public banks.
  - An important first step would be to remove the explicit guarantee currently given by the state to all colon-denominated deposits in state banks.

### Debt management and government securities markets
- Follow market-friendly debt management and issuance strategies to help foster secondary markets for government securities, such as:
  - use of standardized simple instruments with conventional maturities;
  - strengthening legal and regulatory frameworks.

### Stock market and institutional investor development
- Promote development of the stock market through:
  - a more robust macroeconomic environment;
  - stronger institutional and legal frameworks which promote investor rights and information disclosure;
  - policies that increase market size, in particular those supporting the development of an institutional investor base.
- Recent reform: allowing private participation into the insurance sector where it used to be a state monopoly; more could be done to encourage further entry.
- Strengthen protection of minority investors — an area where Costa Rica does not score well in Doing Business indicators.
- Review tax treatment of securities issuance and investment to make the tax system more attractive to issuers, provided it does not jeopardize fiscal sustainability objectives.

### Figures and indices referenced
- Figure A2-2: Costa Rica’s Financial Development Gaps (presented in source).
- Figure A2-3: Financial Development Indices and Sub-indices by Country (presented in source).
- Tables referenced in source: Table A2-1 Index Components; Table A2-2: Weight Comparison; Table A2-3: Regression Variables, Source and Transformation.

*Source: Conclusions and Policy Recommendations (content unit: _wp1681 - Conclusions and Policy Recommendations).*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2016/_wp1681.pdf_