## _wp11240

## Source details

**Canonical URL:** [_wp11240](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2011/_wp11240.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2011/_wp11240.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2011/_wp11240.pdf.json)

---

### I. Introduction: objective and scope
- Objective: assess the effectiveness of the interest-rate transmission mechanism in Central America (CADR) and provide recommendations to enhance monetary policy effectiveness in CADR.
- Sample/period and comparators:
  - Benchmark group: LA6 = Brazil, Chile, Colombia, Mexico, Peru, Uruguay.
  - Panel sample for cross-country estimation: 40 countries that publish a policy rate.
  - Annual data used for panel estimation: 2004–10.
- Key finding summarized:
  - The interest-rate transmission is weaker and slower in CADR than in the LA6.
  - Transmission strength varies across CADR according to financial dollarization, exchange rate flexibility, bank concentration, financial sector development, institutional quality, and fiscal dominance.

### II. Determinants of interest-rate transmission (literature synthesis)
- Factors affecting first-step pass-through (policy rate → bank lending/deposit rates):
  - Financial dollarization:
    - High financial dollarization may reduce pass-through to local-currency bank rates; borrowers can switch to foreign-currency instruments.
    - De-dollarization in Peru and Uruguay has strengthened transmission (Acosta-Ormaechea and Coble, 2011).
  - Exchange rate flexibility:
    - Greater flexibility allows the policy rate to be the primary monetary instrument, improves signaling, encourages hedging, and supports de-dollarization (Freedman and Otker-Robe, 2010).
  - Bank concentration:
    - Higher concentration (market power) lowers banks' responsiveness to policy rate changes; changes may affect spreads rather than market rates (Cotarelli and Korelis, 1994; De Bondt, 2002).
  - Financial sector development:
    - More developed money and interbank markets and domestic capital markets improve pass-through (Yang et al., 2011; Leiderman et al., 2006).
  - Financial system health:
    - Weaker banks may hoard liquidity or widen margins after central bank liquidity injections or rate cuts, limiting credit response (IMF, 2010; Archer and Turner, 2006).
  - Central bank independence:
    - Greater de jure and de facto independence enhances signaling and transmission (Cukierman, Webb, and Neyapti, 1992).
  - Fiscal dominance:
    - Central bank claims on government (percent of GDP) and direct central bank lending to government undermine signaling and pass-through (Laurens, 2005; IMF, 2010).
  - Institutional and regulatory quality:
    - Weak institutions raise asymmetric information and enforcement costs, reduce elasticity of loan demand, and weaken pass-through; proxies include World Bank governance indicators (Kaufmann, Kraay, and Mastruzzi, 2009).

### III. Stylized empirical evidence on transmission in CADR
- Correlation analysis (monthly data; methodology follows Mishra et al., 2010):
  - General result: pass-through is generally weaker in CADR than in LA6; strength varies by country.
  - Within CADR:
    - Costa Rica has the highest correlation between policy and market rates.
    - Nicaragua, with its exchange rate anchor, has the lowest correlation.
  - Within LA6:
    - Except for Peru (and Uruguay for deposit rates), LA6 countries have higher long-term correlation coefficients than CADR.
  - Lending rates tend to have a higher correlation with the policy rate than deposit rates in most Latin American countries.
- Speed of transmission (cumulative response estimates):
  - LA6:
    - Peru: full transmission at about four months (despite lowest long-term correlation coefficient).
    - Most other LA6 countries: full transmission within six and eight months.
  - CADR:
    - Costa Rica: fastest in CADR — about six months to full transmission to lending rates.
    - Dominican Republic, Guatemala, and Honduras: slowest in CADR — transmission within eight months to a year.
  - Nicaragua: cumulative pass-through estimated but excluded from the main figure; details in Appendix III.
- Cross-country correlations (summary):
  - Positive correlation with: exchange rate flexibility; financial intermediation (bank credit to private sector, percent of GDP); institutional environment (World Bank regulatory quality).
  - Negative correlation with: financial dollarization; bank concentration (Hirschman-Herfindahl index); fiscal dominance (central bank claims on government, percent of GDP).

### IV. Panel estimation setup and diagnostics
- Objective: measure pass-through from policy rates to banks' lending rates and test interactions with structural determinants.
- Model specification (as presented in source):
  - y_it = α + β1 y_it−1 + β2 x_it + μ z_it + θ x_it z_it + δ_it (with time dummies δ denoting time effects).
  - Dependent variable: change in bank lending rate (y_it).
  - Policy rate: x_it.
  - Structural conditioning variables (z_it) interacted with policy rate:
    - Deposit dollarization: ratio of foreign currency deposits to total deposits.
    - Exchange rate flexibility: index 0–10 based on standard deviation of daily exchange rate within one month (higher = more flexible).
    - Size of banking sector: bank deposits to GDP.
    - Banking concentration: dummy = 1 if concentration index > median, 0 otherwise.
  - Expected interaction signs:
    - Policy rate × exchange rate flexibility → positive.
    - Policy rate × deposits-to-GDP → positive.
    - Policy rate × deposit dollarization → negative.
    - Policy rate × banking concentration → negative.
- Econometric approach:
  - Estimators: pooled OLS, least-squares dummy variables with fixed effects (LSDV), and System-GMM (Arellano and Bover, 1995; Blundell and Bond, 1998).
  - Monthly data collapsed into yearly averages for System-GMM (large n, small t).
  - Stationarity tests confirmed no unit root; variables treated in levels.
- Estimation diagnostics and sample:
  - No. of observations = 235.
  - Number of countries = 40.
  - No. of instruments = 24 (in one specification) and 40 (in system GMM two-step).
  - System GMM (two-step) results reported in column 6; instrument set of 40 instruments. Hansen test p-value = 0.149.
  - A-B AR(1) test statistics: -1.839 and -2.797; A-B AR(1) test p-values: 0.0659 and 0.00516.
  - A-B AR(2) test statistics: -1.470 and -1.709; A-B AR(2) test p-values: 0.1410 and 0.199.
  - Robust standard errors reported; significance: *** p<0.01, ** p<0.05, * p<0.1.

### V. Main quantitative panel findings
- Baseline pass-through estimates:
  - Pass-through in the first year (system GMM with interactions, column 6): 0.55 (a one percentage change in the policy rate is associated with a 0.55 percentage point increase in the lending rate).
- Policy rate coefficients across specifications: 0.191***, 0.499***, 0.657***, 0.277***, 0.526***, 0.551***.
- Lagged lending rate coefficients (percent): 0.869*** (OLS), 0.378*** (LSDV), 0.428*** (Windmeijer SE correction in one specification), 0.848***, 0.343***, 0.618*** (various specifications).
- Interaction coefficients (selected, column 6 where applicable):
  - FX deposits (percent of total deposits): coefficient -0.00711 (standard error 0.0361).
  - FX flexibility (index): coefficient -0.118 (standard error 0.118).
  - Deposits (percent of GDP): coefficient -0.000935 (standard error 0.00489).
  - Bank concentration (dummy): coefficient -0.209 (standard error 0.511).
  - Policy rate × FX deposits: -0.00722* (standard error 0.00359).
  - Policy rate × FX flexibility: 0.0657* (standard error 0.0355).
  - Policy rate × Deposits (percent of GDP): 0.00197* (standard error 0.00101).
  - Policy rate × Bank concentration: -0.107 (standard error 0.100).
- Model fit and joint significance:
  - Selected R-squared values: 0.952, 0.542, 0.955, 0.611.
  - Selected Adjusted R-squared values: 0.949, 0.519, 0.952, 0.585.
  - F-test p value (joint significance of financial variables) = 0.0336** (jointly significant at the five percent level).

### VI. Interpretation of interactions and economic magnitudes
- Dollarization (foreign currency deposits):
  - Interaction with the policy rate is negative and statistically significant in GMM (column 6).
  - An increase in the ratio of deposits in foreign currency by one standard deviation, from 29 to 53 percent, would reduce the pass-through by about 0.1.
- Exchange rate flexibility:
  - Positive and statistically significant impact on pass-through in columns 4 and 5; significance decreases in column 6.
  - An increase in the index of exchange flexibility by one standard deviation (from 5.4 to 8.1) would increase the pass-through from about 0.50 to about 0.65.
- Financial sector size (deposits to GDP):
  - Positive correlation with pass-through.
  - Were the ratio of deposits to GDP to increase by one standard deviation (from 35.1 to 61.2 percent), the pass-through would increase from 0.5 to 0.6.
  - Interaction coefficient of policy rate and bank deposits to GDP is positive and statistically significant in equations 4 to 6.
- Bank concentration:
  - Interaction with the policy rate shows the expected negative sign, indicating concentration tends to lower pass-through.
  - Statistical significance only confirmed in equation 4; limitations noted in concentration index construction.

### VII. Cross-country and subgroup patterns
- Using system GMM (column 6) to compute pass-through at mean values:
  - On average, LA6 have a higher interest-rate pass-through than CADR countries.
  - Within LA6: Brazil and Chile have the highest pass-through; Peru and Uruguay (dollarized economies) have the lowest.
  - CADR country pass-throughs are relatively similar in the Dominican Republic, Costa Rica, and Guatemala; pass-through range for CADR ≈ 0.5–0.7.
- Exclusions:
  - Panel analysis excludes Nicaragua (it does not have a policy rate).
- Country context:
  - Higher pass-through coefficients in Dominican Republic, Costa Rica and Guatemala are consistent with relatively more advanced monetary policy frameworks; Guatemala has adopted IT, Dominican Republic and Costa Rica are moving toward it but face constraints (limited exchange rate flexibility, shallow capital markets).

### VIII. Conclusions and policy recommendations
- Overall conclusion:
  - Pass-through of the policy rate to market rates is generally weaker and slower in CADR than in LA6 countries.
- Key constraints identified:
  - Limited exchange rate flexibility, financial dollarization, bank concentration, shallow financial markets, fiscal dominance, and weak regulatory/institutional environment.
- Panel results summary:
  - Estimated pass-through to lending ≈ 0.55 in the sample.
  - Dollarization has a negative and statistically significant relationship with pass-through.
  - Exchange rate flexibility and financial system development have positive and statistically significant relationships.
  - Bank concentration shows a negative relationship but statistical significance is not robust.
- Recommended near-term priorities for CADR:
  - Increase exchange rate flexibility.
  - Preserve the absence of fiscal dominance.
  - Ensure central bank instrument independence (Medina Cas, Carrión-Menéndez, Frantischek, 2011).
- Additional reforms to follow or phase in:
  - Strengthen political independence of central banks and tackle central bank balance-sheet weaknesses.
  - Develop central bank capacity for forecasting inflation.
  - Improve transparency and accountability of central banks.
  - Enhance financial supervision and regulation.
  - Implement structural reforms to develop and diversify financial markets.

### IX. Data, frequency, and sources
- Frequency: Monthly for most series; some yearly and daily series noted; quarterly data for some indicators.
- Data and sources include: Haver Analytics; Standardized Report Form (IMF, data provided by authorities); Bloomberg LLP; dXdata; Secretaría Ejecutiva del Consejo Monetario Centroamericano (SECMCA); World Bank Governance Indicators; Beck and Al-Husaini (2009); country authorities.
- FX flexibility measures:
  - FX flexibility 1/ calculated as a log of the monthly average weekly standard deviation of the exchange rate.
  - FX flexibility 3/ calculated as the monthly average daily standard deviation of exchange rate, standardized as a 0-10 index.
- Appendix IV: panel regression data sources and country-level data coverage for policy rates, lending rates, deposit currency, CB claims on the central government, private credit, CB independence & regulatory quality, bank concentration, and FX flexibility.

*Source: IMF staff working paper (sections I–III and Appendix IV excerpted from the referenced PDF).*

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

### _wp11240 - References.

### I. Introduction: objective and scope
- Objective: assess the effectiveness of the interest-rate transmission mechanism in Central America (CADR) and provide recommendations to enhance monetary policy effectiveness in CADR.
- Sample/period and comparators:
  - Benchmark group: LA6 = Brazil, Chile, Colombia, Mexico, Peru, Uruguay.
  - Panel sample for cross-country estimation: 40 countries that publish a policy rate.
  - Annual data used for panel estimation: 2004–10.
- Key finding summarized in introduction:
  - The interest-rate transmission is weaker and slower in CADR than in the LA6.
  - Transmission strength varies across CADR according to financial dollarization, exchange rate flexibility, bank concentration, financial sector development, institutional quality, and fiscal dominance.

### II. Determinants of interest-rate transmission (literature review)
- Factors reviewed as determinants of the first-step pass-through (policy rate → bank lending/deposit rates):
  - Financial dollarization:
    - High financial dollarization may reduce pass-through of policy rate to local-currency bank rates; borrowers can switch to foreign-currency instruments.
    - De-dollarization in Peru and Uruguay has strengthened transmission (Acosta-Ormaechea and Coble, 2011).
  - Exchange rate flexibility:
    - Greater flexibility allows the policy rate to be the primary monetary instrument, improves signaling, encourages hedging, and supports de-dollarization (Freedman and Otker-Robe, 2010).
  - Bank concentration:
    - Higher concentration (market power) lowers banks' responsiveness to policy rate changes; policy-rate changes may affect banks' spreads rather than market rates (Cotarelli and Korelis, 1994; De Bondt, 2002).
  - Financial sector development:
    - More developed money and interbank markets and domestic capital markets improve pass-through by increasing alternatives to bank financing and strengthening the interbank channel (Yang, Davies, Wang, Dunn, and Wu, 2011; Leiderman et al., 2006).
  - Financial system health:
    - Weaker banks may hoard liquidity or widen margins after central bank liquidity injections or rate cuts, limiting credit response (IMF, 2010; Archer and Turner, 2006).
  - Central bank independence:
    - Greater de jure and de facto independence enhances signaling and transmission (Cukierman, Webb, and Neyapti, 1992).
  - Fiscal dominance:
    - Central bank claims on government (percent of GDP) and direct central bank lending to government undermine signaling and pass-through (Laurens, 2005; IMF, 2010).
  - Institutional and regulatory quality:
    - Weak institutions raise asymmetric information and enforcement costs, reduce elasticity of loan demand, and weaken pass-through; proxies include World Bank governance indicators (Kaufmann, Kraay, and Mastruzzi, 2009).

### III. Interest-rate transmission in CADR — stylized empirical evidence
- Correlation analysis (monthly data; methodology follows Mishra, et al., 2010; Appendix I for details):
  - General result: pass-through is generally weaker in CADR than in LA6; strength varies by country.
  - Within CADR:
    - Costa Rica has the highest correlation between policy and market rates.
    - Nicaragua, with its exchange rate anchor, has the lowest correlation.
  - Within LA6:
    - With the exception of Peru (and Uruguay for deposit rates), LA6 countries have higher long-term correlation coefficients than CADR.
  - Lending rates tend to have a higher correlation with the policy rate than deposit rates in most Latin American countries.
- Speed of transmission (cumulative response estimates; Appendix III explains calculation):
  - LA6:
    - Peru: full transmission at about four months (despite having the lowest long-term correlation coefficient).
    - Most other LA6 countries: full transmission within six and eight months.
  - CADR:
    - Costa Rica: fastest in CADR — about six months to full transmission of policy rate to lending rates.
    - Dominican Republic, Guatemala, and Honduras: slowest in CADR — transmission within eight months to a year.
  - Nicaragua: cumulative pass-through estimated but excluded from the main figure; details in Appendix III.
- Visual and tabular materials referenced (see source):
  - Table 1: Monetary and exchange rate frameworks in CADR and LA6 (as of end-April, 2010).
  - Table 2: Short and long-term correlations between policy rate and bank lending/deposit rates (average 2004–2010 unless noted otherwise).
  - Figure 1: Pass-through of the policy rate to market rates in Latin America.
  - Figure 2: CADR + LA6: Policy-to-Lending Pass-through Cumulative Response Over Time.
  - Appendix II: plots of policy and lending rates in LA6 and CADR.
- Cross-country correlations of pass-through with candidate determinants (summary of empirical associations shown in figures and text):
  - Positive correlation with: exchange rate flexibility; financial intermediation (bank credit to private sector, percent of GDP); institutional environment (World Bank regulatory quality).
  - Negative correlation with: financial dollarization; bank concentration (Hirschman-Herfindahl index); fiscal dominance (central bank claims on government, percent of GDP).

### IV. Panel estimation setup (empirical strategy)
- Goal: measure pass-through from policy rates to banks' lending rates using panel estimation and interactions with structural determinants.
- Key model features:
  - Dependent variable: change in bank lending rate (y_it).
  - Policy rate: x_it.
  - Structural conditioning variables (z_it) interacted with policy rate:
    - Deposit dollarization: ratio of foreign currency deposits to total deposits in the banking system.
    - Exchange rate flexibility: index from 0 to 10 based on the standard deviation of daily exchange rate within one month (higher = more flexible).
    - Size of the banking sector: ratio of bank deposits to GDP.
    - Banking concentration: dummy = 1 if concentration index > median, 0 otherwise.
  - Expected interaction signs:
    - Policy rate × exchange rate flexibility → positive.
    - Policy rate × deposits-to-GDP (banking sector size) → positive.
    - Policy rate × deposit dollarization → negative.
    - Policy rate × banking concentration → negative.
- Econometric approach:
  - Concern: independent variables not strictly exogenous (correlation with error term).
  - Estimators used: pooled OLS, least-squares dummy variables with fixed effects (LSDV), and System-GMM (Arellano and Bover, 1995; Blundell and Bond, 1998).
  - Data preprocessing: monthly data collapsed into yearly averages for System-GMM implementation (large n, small t).
  - Stationarity: tests performed and confirmed no unit root; variables treated in levels.
- Model equation structure (as presented in source):
  - y_it = α + β1 y_it−1 + β2 x_it + μ z_it + θ x_it z_it + δ_it (with time dummies δ denoting time effects).
  - Pass-through represented by β2 and modified by θ x_it z_it.

### V. Interim empirical conclusions highlighted in the text
- Panel and correlation evidence point to a set of policy-relevant determinants that can strengthen the interest-rate transmission:
  - Reducing dollarization.
  - Increasing exchange rate flexibility.
  - Further developing the financial sector (interbank markets, domestic capital markets).
  - Improving institutional and regulatory quality and reducing fiscal dominance.
- Policy implication emphasized:
  - Strengthening the pass-through of the policy rate to market interest rates is important to improve the effectiveness of monetary policy in CADR and help central banks ensure price stability.

*Source: IMF staff working paper (sections I–III excerpted from the referenced PDF).*

### Appendix IV has details on the sample of countries, data, and sources.

### _wp11240 - Appendix IV has details on the sample of countries, data, and sources.

### Estimation approach and diagnostics
- System GMM (two-step) results reported in column 6; instrument set of 40 instruments. Hansen test p-value = 0.149.
- Columns 1–3 present models with no interactions; columns 4–6 include interactions.
- All estimations include time dummies.
- No. of observations = 235.
- Number of countries = 40.
- No. of instruments = 24 (in one specification) and 40 (in system GMM two-step).
- A-B AR(1) test statistics: -1.839 and -2.797; A-B AR(1) test p-values: 0.0659 and 0.00516.
- A-B AR(2) test statistics: -1.470 and -1.709; A-B AR(2) test p-values: 0.1410 and 0.199.
- Robust standard errors in parentheses. Significance: *** p<0.01, ** p<0.05, * p<0.1.

### Main quantitative findings (panel results)
- Pass-through in the first year (system GMM with interactions, column 6): 0.55 (a one percentage change in the policy rate is associated with a 0.55 percentage point increase in the lending rate).
- Lagged lending rate (percent) coefficients reported: 0.869*** (OLS), 0.378*** (LSDV), 0.428*** (Windmeijer SE correction in one specification), 0.848***, 0.343***, 0.618*** (various specifications).
- Policy rate (percent) coefficients reported across specs: 0.191***, 0.499***, 0.657***, 0.277***, 0.526***, 0.551***.
- Individual financial-variable interaction coefficients (selected, column 6 where applicable):
  - FX deposits (percent of total deposits): coefficient -0.00711 (standard error 0.0361).
  - FX flexibility (index): coefficient -0.118 (standard error 0.118).
  - Deposits (percent of GDP): coefficient -0.000935 (standard error 0.00489).
  - Bank concentration (dummy): coefficient -0.209 (standard error 0.511).
  - Policy rate × FX deposits: -0.00722* (standard error 0.00359).
  - Policy rate × FX flexibility: 0.0657* (standard error 0.0355).
  - Policy rate × Deposits (percent of GDP): 0.00197* (standard error 0.00101).
  - Policy rate × Bank concentration: -0.107 (standard error 0.100).
- R-squared and adjusted R-squared reported (selected): R-squared 0.952, 0.542, 0.955, 0.611; Adjusted R-squared 0.949, 0.519, 0.952, 0.585.
- F-test p value (joint significance of financial variables) = 0.0336** (jointly significant at the five percent level).

### Interpretation of interactions and economic magnitudes
- Dollarization (foreign currency deposits):
  - Interaction with the policy rate is negative and statistically significant in GMM (column 6).
  - An increase in the ratio of deposits in foreign currency by one standard deviation, from 29 to 53 percent, would reduce the pass-through by about 0.1.
- Exchange rate flexibility:
  - Positive and statistically significant impact on interest-rate pass-through in columns 4 and 5; significance decreases in column 6.
  - An increase in the index of exchange flexibility by one standard deviation (from 5.4 to 8.1) would increase the pass-through from about 0.50 to about 0.65.
- Financial sector size (deposits to GDP):
  - Positive correlation with pass-through.
  - Were the ratio of deposits to GDP to increase by one standard deviation (from 35.1 to 61.2 percent), the pass-through would increase from 0.5 to 0.6.
  - Interaction coefficient of policy rate and bank deposits to GDP is positive and statistically significant in equations 4 to 6.
- Bank concentration:
  - Interaction with the policy rate shows expected negative sign, indicating concentration tends to lower pass-through.
  - Statistical significance only confirmed in equation 4; limitations may stem from the concentration index construction (accounts only for the assets of the three largest banks as a share of all commercial banks) or competitive dynamics.

### Cross-country and subgroup results
- Panel estimates (system GMM, column 6) used to compute pass-through at mean values for LA6 and CADR.
- On average, LA6 have a higher interest-rate pass-through than CADR countries.
- Within LA6: Brazil and Chile have the highest pass-through; Peru and Uruguay (dollarized economies) have the lowest.
- CADR country pass-throughs are relatively similar in the Dominican Republic, Costa Rica, and Guatemala; pass-through range for CADR ~ 0.5–0.7.
- Panel analysis excludes Nicaragua (it does not have a policy rate).
- Higher pass-through coefficients in Dominican Republic, Costa Rica and Guatemala are consistent with relatively more advanced monetary policy frameworks; Guatemala has adopted IT, Dominican Republic and Costa Rica are moving toward it but face constraints (e.g., limited exchange rate flexibility, shallow capital markets).

### Conclusion and policy recommendations
- Overall conclusion: Pass-through of the policy rate to market rates is generally weaker and slower in CADR than in LA6 countries.
- Key constraints identified: limited exchange rate flexibility, financial dollarization, bank concentration, shallow financial markets, fiscal dominance, and weak regulatory/institutional environment.
- Panel results summary: Estimated pass-through to lending ≈ 0.55 in the sample; dollarization has a negative and statistically significant relationship with pass-through; exchange rate flexibility and financial system development have positive and statistically significant relationships; bank concentration shows a negative relationship but statistical significance is not robust.
- Recommended near-term priorities for CADR:
  - Increase exchange rate flexibility.
  - Preserve the absence of fiscal dominance.
  - Ensure central bank instrument independence (Medina Cas, Carrión-Menéndez, Frantischek, 2011).
- Additional reforms to follow or phase in after the above:
  - Strengthen political independence of central banks and tackle central bank balance-sheet weaknesses.
  - Develop central bank capacity for forecasting inflation.
  - Improve transparency and accountability of central banks.
  - Enhance financial supervision and regulation.
  - Implement structural reforms to develop and diversify financial markets.

### Data, frequency, and sources (Appendices)
- Frequency: Monthly for most series; some yearly and daily series noted; quarterly data for some indicators.
- Sources include: Haver Analytics; Standardized Report Form (IMF, data provided by authorities); Bloomberg LLP; dXdata; Secretaría Ejecutiva del Consejo Monetario Centroamericano (SECMCA); World Bank Governance Indicators; Beck and Al-Husaini (2009); country authorities.
- FX flexibility 1/ calculated as a log of the monthly average weekly standard deviation of the exchange rate (Appendix I note).
- FX flexibility 3/ calculated as the monthly average daily standard deviation of exchange rate, standardized as a 0-10 index (Appendix IV note).
- Appendix IV provides panel regression data sources and country-level data coverage for policy rates, lending rates, deposit currency, CB claims on the central government, private credit, CB independence & regulatory quality, bank concentration, and FX flexibility.

*Source: _wp11240 - Appendix IV has details on the sample of countries, data, and sources.*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2011/_wp11240.pdf_