## 9. BCP compliance and bank stability: Checking for endogeneity

## Source details

**Canonical URL:** [9. BCP compliance and bank stability: Checking for endogeneity](https://www.imf.org/-/media/files/publications/wp/2017/wp17161.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2017/wp17161.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2017/wp17161.pdf.json)

---

### Introduction and objectives
- Examine whether compliance with Basel Core Principles (BCPs) for effective banking supervision affects bank stability and risk taking by comparing conventional and Islamic banks.
- Focus: banks operating mainly in developing and emerging countries over the period 1999–2013.
- Use BCP index (IMF/World Bank FSAP database) as main regulatory compliance measure; CPIFR data unavailable for the period, but CPIFRs are argued to be closely benchmarked to BCPs.

### Data and main variables
- Initial sample: 761 conventional and Islamic banks in 19 countries covering 1999–2013.
- Final bank-level sample: 651 conventional banks and 110 Islamic banks.
- BCP chapter waves matched to periods: 1999 wave for 1999–2004; 2005 wave for 2005–11; 2012 wave for 2012–13 (example: Saudi Arabia reported in 2004 and 2011; 2009 wave used for 2004–10 and 2011 wave for 2011–13).
- Dependent variable: Z-score (natural logarithm used, LnZ-score). Z-score defined as ([return on average assets + equity/assets]/[standard deviation of the return on average assets] over (t, t–3)).
- Main independent variable: country BCP compliance index (aggregate of seven chapters: Ch. 1–7).
- Estimation features: variables winsorized at the 1 percent and 99 percent levels; clustering at the bank level; random-effect GLS regressions; interaction term between Islamic dummy and BCP to test differential effects.

### Key descriptive statistics
- BCP index: Mean = 84.95 percent; Median = 83.33; SD = 12.14; N = 285.
- Country examples (upper 10 percent distribution): Saudi Arabia 97.66 percent; UK 94.22 percent; Malaysia 91.73 percent; United Arab Emirates 90.71 percent.
- Z-score (Table 2, Panel A and C):
  - Conventional banks (CBs) — Z-score: N = 5031, Mean = 3.61, Median = 3.6, SD = 5.63.
  - Islamic banks (IBs) — Z-score: N = 637, Mean = 3.19, Median = 3.22, SD = 1.13.
- Individual BCP chapter means (N = 285):
  - Chapter 1: Mean = 84.15, Median = 87.5, SD = 14.55.
  - Chapter 2: Mean = 74.27, Median = 77.5, SD = 18.42.
  - Chapter 3: Mean = 80.09, Median = 85, SD = 15.71.
  - Chapter 4: Mean = 87.89, Median = 100, SD = 16.30.
  - Chapter 5: Mean = 75.64, Median = 75, SD = 19.61.
  - Chapter 6: Mean = 81.22, Median = 83, SD = 16.66.
  - Chapter 7: Mean = 81.94, Median = 83, SD = 16.70.
- Governance and macro controls (N = 285):
  - wgi: Mean = -0.42, Median = -0.63, SD = 0.65.
  - gdpg: Mean = 4.03, Median = 4.3, SD = 2.96.
  - inf: Mean = 6.33, Median = 4.5, SD = 7.79.
  - oil: Mean = 5.11, Median = 1.06, SD = 9.74.
  - gaz: Mean = 2.2, Median = 0.78, SD = 2.91.
  - mineral: Mean = 0.35, Median = 0, SD = 0.81.

### Main empirical findings
- Overall association:
  - BCP compliance index is positively and significantly associated with bank stability (LnZ-score).
  - Conventional banks: positive and significant at the 1 percent level.
  - Islamic banks: positive and significant at the 5 percent level.
  - Full sample: positive and significant at the 1 percent level.
- Magnitude:
  - Estimated coefficients on BCP compliance in Models 1, 5, and 9 vary between 0.015 and 0.017, indicating that a one-unit increase in the BCP compliance index is associated with an increase of nearly two percentage points in the Z-score.
  - Baseline regression highlights (Table 3, Panel A): Conventional banks BCP (α_BCP) coefficient on Z-score = 0.015*** (standard error 0.002). Full sample BCP coefficient on Z-score = 0.017*** (standard error 0.002).
- Mechanisms and decomposition:
  - Effects mainly driven by bank capital ratios (equity to assets, TETA).
  - Conventional banks: BCP compliance associated with lower ROAA (negative at 1 percent), lower volatility of returns (SDROAA) (negative at 10 percent), and higher capitalization (TETA) (positive at 1 percent).
  - Islamic banks: capital association positive in some specifications (Model 8 positive at the 10 percent level; Model 12 positive at the 5 percent level).
- Bank-level controls:
  - lnta (bank size) negatively correlated with Z-score (effect driven by negative effect of size on capital).
  - gtap (growth of total assets) negatively associated with Z-score.
  - cirp (cost-to-income ratio) negatively associated with Z-score.
  - Liquidity (ladstfp) mixed: negative on profits and positive on capital, yielding weaker net effects for Islamic banks.
- Country-level controls:
  - Banks more stable in countries with better GDP growth (gdpg), higher mineral rents (mineral), lower gas rents (gaz), and lower inflation (inf).
  - Positive effect of GDP and mineral rents driven by ROAA; negative effects of gas rents and inflation driven by SDROAA.

### Disaggregated BCP chapters and heterogeneity
- Individual chapter impacts on Z-score (selected coefficients, Table 4 excerpt):
  - chapter 1: 0.004 (0.003)
  - chapter 2: 0.015*** (0.002)
  - chapter 3: 0.01*** (0.002)
  - chapter 4: 0.012*** (0.002)
  - chapter 5: 0.009*** (0.002)
  - chapter 6: 0.003** (0.001)
- Conventional banks: five out of seven chapters significantly positive at the 10 percent level or better; Chapters 2 (Licensing and Structure) and 7 (Cross-Border Banking) most pronounced; Chapter 1 least.
- Islamic banks: three out of seven chapters significantly positive at the 5 percent level or more; Chapter 2 most pronounced.
- Including all chapters simultaneously weakens results due to multicollinearity.

### Alternative risk measures
- BCP index negatively associated with credit-risk proxies and SDNIM for conventional banks:
  - LLRGLP: -0.049*** (0.019) for Conventional banks; -0.05*** (0.019) for Full sample.
  - LLPTLP: -0.012** (0.005) for Conventional banks; -0.014*** (0.005) for Full sample.
  - NPLGLP: -0.059** (0.029) for Conventional banks; -0.058* (0.029) for Full sample.
  - SDNIM: -0.009*** (0.002) for Conventional banks; -0.012*** (0.003) for Full sample.
- For Islamic banks significance with alternative risk measures is weaker and mainly for SDNIM.

### Robustness checks and alternative estimations
- Subsamples (regions, UK exclusion, listing status, crisis periods, institutional environments, efficiency):
  - Excluding GCC / SEA / MENA: conventional bank coefficients remain positive (examples: 0.014*** (0.003); 0.008* (0.004); 0.014*** (0.003)).
  - For Islamic banks, positive association mainly driven by SEA and GCC; often insignificant when excluding these regions.
  - Excluding UK: results similar for conventional banks; become significant for Islamic banks.
  - Listed banks: stronger BCP effect on Z-score, especially for Islamic banks.
  - Period subsamples: effect stronger in pre-crisis period (1999–2006); less effective in periods of economic distress.
  - Institutional environment: BCP compliance negative in countries with weak protection of depositors (example: CBs -0.21* (0.119); Full -0.296*** (0.098)); positive and significant for highly efficient banks (CBs 0.019*** (0.004); Full 0.025*** (0.003)).
- Quantile regressions (Table 6):
  - Conventional banks: Q25 0.013*** (0.003); Q50 0.017*** (0.003); Q75 0.018*** (0.004).
  - Full sample: Q25 0.017*** (0.003); Q50 0.019*** (0.003); Q75 0.019*** (0.004).
  - For Islamic banks quantile evidence weaker.
- Alternative estimation techniques (Table 8):
  - Truncated regressions, bootstrapped standard errors (100 resamples), and White heteroscedasticity correction affirm positive BCP effect on Z-score for conventional banks (examples: Truncated 0.019*** (0.003); Bootstrap 0.015*** (0.002); White 0.016*** (0.002)).
  - Islamic banks: one truncated regression reported insignificant (0.005 (0.006)).
- Propensity Score Matching (PSM) (Table 8, Panel B):
  - BCP compliance dummy = 1 if BCP ≥ median.
  - Matching methods: K-nearest neighbors (n=10, n=15, n=20), Gaussian Kernel, radius matching.
  - Matched results: conventional banks in higher-BCP countries have higher Z-scores relative to matched lower-BCP peers; similar but smaller effects for Islamic banks.
  - Reported PSM differences (examples): Z-score differences vary between 0.123 and 0.288 percent for conventional banks and between 0.123 and 0.276 percent for Islamic banks (source reports ranges as provided).

### Endogeneity and selection-bias checks
- Instrumental Variables (IV) approach — instruments:
  - rule of law (Heritage Foundation’s Economic Freedom index).
  - business regulation (Fraser Institute’s Economic Freedom of the World (EFW) index).
- First-stage diagnostics:
  - F-test of excluded exogenous variables rejects instrument irrelevance at the 1 percent level in all models.
  - First-stage shows higher Z-score in countries with better institutional environment (rule of law, business regulation).
- Second-stage estimation: 2SLS and GMM used.
- IV results (Table 9 summary):
  - Clear evidence of positive and significant association at the 1 percent level between BCP compliance index and Z-score for conventional banks (examples: Conventional banks IV 0.013*** (0.003); Conventional banks 2SLS / GMM 0.019*** (0.004)).
  - Islamic banks: IV results need caution because Sargan and Hansen J tests are significant.
  - Full sample: 훼_BCP positive and significant; (훼_BCP + 훼_inter) positive and significant in full-sample specification.
- Selection correction (Heckman 1979):
  - First-step probit: dependent dummy = 1 if country’s BCP compliance index ≥ median; instruments and controls included.
  - Second-stage regression includes inverse Mills ratio.
  - Heckman results (Table 9, Panels A and B, Models 5, 10, and 15) continue to suggest both conventional and Islamic banks are more stable in countries with higher BCP compliance index.

### Contributions, policy implications, and next steps
- Contributions:
  - Empirical evidence that BCP compliance positively impacts stability of conventional banks (strong effect) and Islamic banks (positive but less pronounced).
  - Regulatory compliance enhances stability primarily via higher capital ratios and by discouraging excessive risk taking (lower ROAA and lower volatility of returns).
  - Comparative evidence shows relative similarity in regulatory determinants of stability across bank types, with heterogeneity by region and institutional environment.
- Policy implications:
  - Findings support the role of BCP standards in improving bank stability and suggest CPIFRs—benchmarking closely to BCPs—should also positively affect Islamic bank financial soundness.
  - Raises question whether BCP standards should be amended to account for Islamic-bank specificities (e.g., IAH treatment, Rate of Return risk, Sharia’a governance) given differences highlighted by CPIFRs.
  - Emphasize importance of institutional environment and bank efficiency for BCP effectiveness: BCP compliance stronger for efficient banks in countries with better institutional and political systems.
- Limitations and avenues for future research:
  - Overall significance and interpretation depend on sample size, country choice, and validity of accounting measures for bank stability.
  - Future work: explore effect of CPIFRs on Islamic-bank stability and compare to BCP effects; identify which BCP and CPIFR chapters drive significant effects; investigate whether BCP and CPIFR guidelines affect Islamic bank efficiency using nonparametric scores or market-based data.
  - Data note: IFSB asked banks to start reporting CPIFR data as of January 2016; data likely to be available in 2017 (beyond the current study period).

*Source: wp17161 — "9. BCP compliance and bank stability: Checking for endogeneity" (excerpted content provided).*

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

### References

### Tables

- 1. Overview of Basel Core Principles studies in conventional banking ..................................................... 24
- 2. Descriptive statistics ............................................................................................................................... 25
- 3. BCP compliance and bank stability: Islamic vs. conventional banks ..................................................... 26
- 4. BCP compliance and bank stability: Individual factors .......................................................................... 27
- 5. BCP compliance and bank stability: alternative samples ....................................................................... 29
- 6. BCP compliance and bank stability: A quantile regression approach .................................................... 32
- 7. BCP compliance and alternative measures of risk .................................................................................. 33
- 8. Robustness checks: Alternative estimation techniques ........................................................................... 34

*Source: wp17161 - References (wp17161 - References; https://www.imf.org/-/media/files/publications/wp/2017/wp17161.pdf)*

### 9. BCP compliance and bank stability: Checking for endogeneity .........................................................

### 9. BCP compliance and bank stability: Checking for endogeneity

### Introduction and objectives
- Examine whether compliance with Basel Core Principles (BCPs) for effective banking supervision affects bank stability and risk taking by comparing conventional and Islamic banks.
- Extend prior work to include Islamic banks and focus on banks operating mainly in developing and emerging countries over the period 1999–2013.
- Note: CPIFR guidelines published in 2015; data on Islamic banks’ compliance with CPIFRs are not available for the study period, so the analysis uses the BCP index and argues that CPIFRs should have similar effects for reasons given in the source.

### Data and sample
- Initial sample: 761 conventional and Islamic banks in 19 countries covering 1999–2013.
- Final bank-level sample: 651 conventional banks and 110 Islamic banks (total 761 initial; sample reduced to banks operating in 19 countries).
- BCP chapters collected in waves and matched to periods: 1999 wave for 1999–2004; 2005 wave for 2005–11; 2012 wave for 2012–13. Example: Saudi Arabia reported in 2004 and 2011; 2009 wave used for 2004–10 and 2011 wave for 2011–13.
- Main dependent variable: Z-score (natural logarithm used, LnZ-score). Z-score defined as ([return on average assets + equity/assets]/[standard deviation of the return on average assets] over (t, t–3)).
- Main independent variable: country BCP compliance index from IMF/World Bank FSAP database, based on 25 principles aggregated into seven chapters: (Ch. 1) Preconditions for Effective Banking Supervision; (Ch. 2) Licensing and Structure; (Ch. 3) Prudential Regulation and Requirements; (Ch. 4) Methods of Ongoing Supervision; (Ch. 5) Information Requirements; (Ch. 6) Formal Powers of Supervisors; (Ch. 7) Cross-Border Banking.
- Variable handling: all variables winsorized at the 1 percent and 99 percent levels; clustering at the bank level; use of random-effect GLS regressions; interaction term between Islamic dummy and BCP to test differential effects.

### Key descriptive statistics
- Mean of BCP compliance index (BCP index): 84.95 percent.
- BCP index, upper 10 percent distribution examples: Saudi Arabia 97.66 percent; UK 94.22 percent; Malaysia 91.73 percent; United Arab Emirates 90.71 percent.
- Observations coverage: N obs. percent higher for conventional banks (58.4 percent) than for Islamic banks (52.1 percent).

### Main empirical findings
- Overall association:
  - BCP compliance index is positively and significantly associated with bank stability (LnZ-score).
  - For conventional banks: positive and significant at the 1 percent level.
  - For Islamic banks: positive and significant at the 5 percent level.
  - For the full sample: positive and significant at the 1 percent level.
- Magnitude reported: estimated coefficients on BCP compliance in Models 1, 5, and 9 vary between 0.015 and 0.017, indicating that a one-unit increase in the BCP compliance index is associated with an increase of nearly two percentage points in the Z-score.
- Mechanisms and components (results on Z-score components):
  - Findings mainly driven by bank capital ratios (equity to assets, TETA).
  - Conventional banks: BCP compliance associated with lower return on average assets (ROAA) (negative at 1 percent), lower volatility of returns (SDROAA) (negative at 10 percent), and higher capitalization (TETA) (positive at 1 percent).
  - Islamic banks: less pronounced effects; association with capital positive in some specifications (Model 8 positive at the 10 percent level; Model 12 positive at the 5 percent level for capital ratios).
- Bank-level controls:
  - Bank size (lnta) negatively correlated with Z-score (effect driven by negative effect of size on capital).
  - Growth of total assets (gtap) negatively associated with Z-score.
  - Cost-to-income ratio (cirp) negatively associated with Z-score.
  - Liquidity ratios have mixed effects: negative on profits and positive on capital, producing weaker net effects for Islamic banks.
- Country-level controls:
  - Banks more stable in countries with better GDP growth (gdpg), higher mineral rents (mineral), lower gas rents (gas), and lower inflation (inf).
  - Positive effect of GDP and mineral rents driven by ROAA; negative effects of gas rents and inflation driven by SDROAA.

### Robustness checks and extensions
- BCP chapters (disaggregate analysis):
  - For conventional banks, five out of seven individual chapters are significantly positive at the 10 percent level or better; Chapters 2 (Licensing and Structure) and 7 (Cross-Border Banking) have the most pronounced effects; Chapter 1 (Preconditions) the least.
  - For Islamic banks, three out of seven chapters are significantly positive (Models cited in source) at the 5 percent level or more, with Chapter 2 having the most pronounced effect.
  - Including all chapters simultaneously weakens results due to multicollinearity (consistent with prior literature).
- Subsamples (regions, UK exclusion, listing status, crisis periods, institutional environments, efficiency):
  - Association between BCP compliance and conventional banks’ Z-score remains positive when excluding GCC, SEA, and MENA regions; coefficients vary between 0.008 and 0.015 in specific exclusions (interpreted economically as Z-score increases between 0.75 percentage point and 1.5 percentage points depending on exclusion).
  - For Islamic banks, positive association is mainly driven by banks in the SEA and GCC regions; results often become insignificant when excluding these regions.
  - Excluding UK (which represents 26 percent, 167 banks of the sample) leaves results similar for conventional banks and becomes significant for Islamic banks.
  - Listed banks show stronger effects: the effect of BCP compliance on Z-score is stronger when banks are publicly listed, especially for Islamic banks.
  - Period subsamples: effect stronger in the pre-crisis period (1999–2006); BCP compliance appears less effective in periods of economic distress.
  - Institutional environment: BCP compliance has negative impact on stability in countries with weak protection of depositors and insignificant effect in countries with less stable political institutions; effect positive and significant for highly efficient banks.
- Quantile regressions:
  - For conventional banks and the full sample, coefficients on BCP compliance are positive at the 25th (Q25), 50th (Q50), and 75th (Q75) quantiles; for Islamic banks, evidence weaker.
  - Coefficients increase across quantiles for conventional banks, but Wald tests do not show significant differences between quantiles.
- Alternative risk measures:
  - Using credit-risk proxies LLRGLP (loan loss reserves to gross loans), LLPTLP (loan loss provision to total loans), NPLGLP (nonperforming loans to gross loans), and SDNIM (standard deviation of net interest margins), the BCP compliance index is negatively and significantly associated with credit risk and SDNIM for conventional banks; for Islamic banks significance is mainly for SDNIM.
  - Reported coefficient range: estimated coefficients on the BCP index in Models 1–4 vary between 0.012 and 0.059, indicating that a one-unit increase in the BCP compliance index is associated with a decrease in credit risk between a one-unit decrease (when using LLPTLP) and a nearly 6 percent decrease (when using NPLGLP).
- Alternative estimation techniques:
  - Truncated regressions, bootstrapped standard errors (100 resamples), and White heteroscedasticity correction all affirm a significantly positive BCP effect on Z-score in most models; exception noted for one truncated regression on Islamic banks (Model 4).
  - BCP compliance has a significantly positive effect on capital ratios of Islamic banks at the 1 percent level in selected estimations (Models 8 and 9 referenced).
- Propensity Score Matching (PSM):
  - Created BCP compliance dummy = 1 if BCP >= median, 0 otherwise; logit model estimated on controls and year fixed effects; matching methods: K-nearest neighbors (n=10, n=15, n=20), Gaussian Kernel, radius matching.
  - Matched results: conventional banks in countries with higher BCP compliance have higher Z-scores compared to matched conventional banks in countries with lower BCP compliance; similar but less pronounced results for Islamic banks.
  - Reported PSM differences (T statistics reported in source): Z-score differences vary between 0.123 and 0.288 percent for conventional banks and between 0.123 and 0.276 percent for Islamic banks (source text ends mid-range reporting for full sample; quoted ranges preserved as provided).

### Contributions and policy implications
- Contributions:
  - Evidence that BCP compliance positively impacts stability of conventional banks (strong effect) and Islamic banks (positive but less pronounced).
  - Shows regulatory compliance enhances stability primarily via higher capital ratios and by discouraging excessive risk taking (lower ROAA and lower volatility of returns).
  - Comparative evidence suggesting relative similarity between the two bank types in regulatory determinants of stability.
- Policy implications:
  - Findings provide empirical support for the role of BCP standards in improving bank stability and suggest that CPIFRs—being closely benchmarked to BCPs—should also positively affect Islamic bank financial soundness.
  - Open question: whether BCP standards should be amended to account for Islamic bank specificities; argument in favor exists given more stable financial systems in countries where both bank types operate and given differences highlighted by CPIFRs (e.g., IAH treatment, Rate of Return risk, Sharia’a governance).
  - Emphasize importance of institutional environment and bank efficiency for BCP effectiveness: BCP compliance is stronger for efficient banks in countries with better institutional and political systems.

*Source: IMF working paper section "9. BCP compliance and bank stability: Checking for endogeneity" (excerpt provided).*

### 0.465 percent for the full sample. These differences are statistically significant at the 1 percent

### wp17161 - 0.465 percent for the full sample. These differences are statistically significant at the 1 percent

### H. Addressing Endogeneity and Selection Bias — Instrumental Variable (IV) approach
- Instruments used:
  - rule of law (Heritage Foundation’s Economic Freedom index): defined as the capacity of a country’s government and legal system to recognize and ensure the protection of property and fight corruption.
  - business regulation (Fraser Institute’s Economic Freedom of the World (EFW) index): defined as the extent to which regulations and bureaucracy procedures restrain entry into business and increase the cost of production.
- First-stage diagnostics:
  - F-test of excluded exogenous variables (following Barth et al. (2009)) — null hypothesis that instrument does not explain cross-sectional differences in capital regulatory guidelines and measures is rejected at the 1 percent level in all models.
  - First-stage results (Table 9, Models 1, 6, and 11) mainly show that a bank’s Z-score is higher in countries with a better institutional environment in terms of rule of law and business regulation.
- Second-stage estimation techniques:
  - Two-stage least squares regression (2SLS) (Ashraf et al., 2016).
  - Generalized method of moments (GMM) (Bitar et al. 2017b).
- IV results:
  - Clear evidence of a positive and significant association at the 1 percent level between the BCP compliance index and the Z-score for the sample of conventional banks.
  - For the Islamic banks sample, results need to be treated with caution because both the Sargan and the Hansen J tests are significant.
  - For the full sample, Panels A and B in Table 9 indicate:
    - 훼_BCP is positive and significant (effect on conventional banks’ stability).
    - (훼_BCP + 훼_inter) is positive and significant (combined effect on Islamic banks’ stability in the full-sample specification).

### H. Addressing Endogeneity and Selection Bias — Selection correction (Heckman 1979)
- First-step probit:
  - Dependent dummy = 1 if country’s BCP compliance index ≥ median, 0 otherwise.
  - Explanatory variables: the two instruments (rule of law and business regulation), bank-level controls, country-level controls, and year-fixed effect from baseline model.
- Second-stage regression:
  - Dependent variable: Z-score.
  - Independent variable: BCP compliance index.
  - Controls: same bank and country-level controls as baseline.
  - Self-selection parameter: inverse Mills ratio estimated from first stage.
- Heckman results (Table 9, Panels A and B, Models 5, 10, and 15):
  - Findings continue to suggest that both conventional and Islamic banks are more stable in countries with a higher BCP compliance index.

### VI. Concluding remarks — main findings and implications
- Main empirical findings:
  - This study suggests a positive effect of BCP compliance on the stability of banks in 19 developing countries.
  - Results are robust to inclusion of individual BCP chapters, though effects are more pronounced for conventional banks than for Islamic banks.
  - Decomposition of Z-score indicates results are mainly driven by capital ratios for both bank types.
- Policy implications:
  - Since BCPs are effective in improving the stability of Islamic banks as well, CPIFRs (Core Principles for Islamic Finance Regulation), which are benchmarked closely to BCPs, can also positively affect Islamic bank stability.
- Limitations and avenues for future research:
  - Overall significance and interpretation depend on sample size, choice of countries, and validity of accounting measures used to proxy bank stability.
  - Increasing sample size depends on future IMF and World Bank surveys.
  - Measurement-error concerns addressed via a variety of proxies and econometric techniques.
  - Next steps:
    - Explore the effect of CPIFRs on Islamic-bank stability and compare to BCP effects.
    - Identify which BCP and CPIFR chapters (especially those considering Islamic-bank specificities) drive significant effects.
    - Investigate whether BCP and CPIFR guidelines have the same effect on Islamic bank efficiency using nonparametric scores or market-based data (stock returns, Tobin’s Q).
  - Data note: IFSB asked banks to start reporting CPIFR data as of January 2016; data are likely to be available in 2017, beyond the current study period.

### Key quantitative statistics and robustness checks reported in the source
- Sample and central descriptive statistics (Table 2, Panel A and C):
  - Conventional banks (CBs) — Z-score: N = 5031, Mean = 3.61, Median = 3.6, SD = 5.63.
  - Islamic banks (IBs) — Z-score: N = 637, Mean = 3.19, Median = 3.22, SD = 1.13.
  - Country-level BCP index: N = 285, Mean = 84.95, Median = 83.33, SD = 12.14.
  - Individual BCP chapter means (N = 285):
    - Chapter 1: Mean = 84.15, Median = 87.5, SD = 14.55.
    - Chapter 2: Mean = 74.27, Median = 77.5, SD = 18.42.
    - Chapter 3: Mean = 80.09, Median = 85, SD = 15.71.
    - Chapter 4: Mean = 87.89, Median = 100, SD = 16.30.
    - Chapter 5: Mean = 75.64, Median = 75, SD = 19.61.
    - Chapter 6: Mean = 81.22, Median = 83, SD = 16.66.
    - Chapter 7: Mean = 81.94, Median = 83, SD = 16.70.
  - Governance and macro controls (N = 285):
    - wgi: Mean = -0.42, Median = -0.63, SD = 0.65.
    - gdpg: Mean = 4.03, Median = 4.3, SD = 2.96.
    - inf: Mean = 6.33, Median = 4.5, SD = 7.79.
    - oil: Mean = 5.11, Median = 1.06, SD = 9.74.
    - gaz: Mean = 2.2, Median = 0.78, SD = 2.91.
    - mineral: Mean = 0.35, Median = 0, SD = 0.81.
- Baseline regression highlights (Table 3, Panel A):
  - Conventional banks: BCP (α_BCP) coefficient on Z-score = 0.015*** (standard error 0.002).
  - Full sample: BCP coefficient on Z-score = 0.017*** (standard error 0.002).
  - Statistical significance notation: * 10% level; ** 5% level; *** 1% level.
- Robustness:
  - Results stand up to robustness checks allowing for omitted variables, endogeneity concerns, selection bias, and alternative estimation techniques.
  - IV analyses show positive and significant associations at the 1 percent level for conventional banks; caution flagged for Islamic-bank IV tests due to Sargan and Hansen J test significance.

*Italic: Source — wp17161 (excerpted content provided).*

### chapter 1 0.003

### Table 4 — BCP compliance and bank stability: Individual chapters (excerpt)

### Impact of individual BCP chapters on Z-score (selected coefficient estimates)
- chapter 1: 0.004 (0.003)
- chapter 2: 0.015*** (0.002)
- chapter 3: 0.01*** (0.002)
- chapter 4: 0.012*** (0.002)
- chapter 5: 0.009*** (0.002)
- chapter 6: 0.003** (0.001)

### Islamic bank indicator and interactions (Panel A, models reported)
- Islamic: -0.721 (0.649); -0.621 (0.519); -0.104 (0.413); -0.258 (0.491); -0.344 (0.458); -0.095 (0.373); 0.258 (0.497)
- Islamic × chapter 1: 0.006 (0.008)
- Islamic × chapter 2: 0.005 (0.007)
- Islamic × chapter 3: -0.003 (0.005)
- Islamic × chapter 4: 0.000 (0.006)
- Islamic × chapter 5: 0.001 (0.005)
- Islamic × chapter 6: -0.002 (0.005)

### Bank- and country-level control variables (coefficients, standard errors)
- lnta:
  - -0.011 (0.019)
  - -0.023 (0.019)
  - -0.029 (0.019)
  - -0.024 (0.019)
  - -0.02 (0.019)
  - -0.013 (0.019)
  - -0.031 (0.019)
  - -0.041** (0.019)
  - 0.163** (0.081)
  - 0.149* (0.080)
  - 0.171** (0.085)
  - 0.15 (0.092)
  - 0.162** (0.081)
  - 0.159* (0.084)
  - 0.151* (0.085)
  - 0.125 (0.096)
- gtap:
  - -0.003*** (0.001)
  - -0.003*** (0.001)
  - -0.003*** (0.001)
  - -0.002*** (0.001)
  - -0.003*** (0.001)
  - -0.003*** (0.001)
  - -0.003*** (0.001)
  - -0.002*** (0.001)
  - -0.000 (0.001)
  - -0.000 (0.001)
  - -0.000 (0.001)
  - 0.001 (0.001)
  - -0.001 (0.001)
  - -0.001 (0.001)
  - -0.000 (0.001)
  - 0.001 (0.002)
- cirp:
  - -0.007*** (0.001)
  - -0.007*** (0.001)
  - -0.007*** (0.001)
  - -0.007*** (0.001)
  - -0.007*** (0.001)
  - -0.007*** (0.001)
  - -0.007*** (0.001)
  - -0.007*** (0.001)
  - -0.003*** (0.001)
  - -0.003*** (0.001)
  - -0.003*** (0.001)
  - -0.003** (0.001)
  - -0.003*** (0.001)
  - -0.003*** (0.001)
  - -0.003*** (0.001)
  - -0.004*** (0.001)
- niitip:
  - -0.029 (0.101)
  - -0.05 (0.103)
  - -0.03 (0.102)
  - -0.129 (0.107)
  - -0.046 (0.102)
  - -0.035 (0.100)
  - -0.028 (0.102)
  - -0.095 (0.104)
  - -0.220 (0.233)
  - -0.307 (0.232)
  - -0.172 (0.215)
  - -0.334 (0.245)
  - -0.287 (0.224)
  - -0.326 (0.247)
  - -0.243 (0.239)
  - -0.181 (0.233)
- ladstfp:
  - 0.001 (0.001)
  - 0.001 (0.001)
  - 0.001 (0.001)
  - 0.001* (0.001)
  - 0.001 (0.001)
  - 0.001** (0.001)
  - 0.001 (0.001)
  - 0.001 (0.001)
  - -0.000 (0.001)
  - -0.000 (0.001)
  - -0.000 (0.001)
  - -0.001 (0.001)
  - -0.000 (0.001)
  - -0.000 (0.001)
  - -0.000 (0.001)
  - -0.000 (0.001)
- wgi:
  - 0.255*** (0.050)
  - 0.269*** (0.051)
  - 0.256*** (0.052)
  - 0.193*** (0.0514)
  - 0.231*** (0.050)
  - 0.255*** (0.050)
  - 0.262*** (0.049)
  - 0.291*** (0.064)
  - 0.154 (0.126)
  - 0.175 (0.126)
  - 0.140 (0.134)
  - -0.0331 (0.160)
  - 0.0472 (0.147)
  - 0.199 (0.124)
  - 0.150 (0.132)
  - 0.331 (0.314)
- gdpg:
  - 0.037*** (0.008)
  - 0.04*** (0.008)
  - 0.043*** (0.008)
  - 0.038*** (0.008)
  - 0.038*** (0.008)
  - 0.036*** (0.009)
  - 0.039*** (0.008)
  - 0.036*** (0.009)
  - 0.034 (0.024)
  - 0.034 (0.023)
  - 0.03 (0.023)
  - 0.023 (0.023)
  - 0.033 (0.023)
  - 0.032 (0.023)
  - 0.028 (0.024)
  - 0.036 (0.025)
- inf:
  - -0.026*** (0.004)
  - -0.022*** (0.004)
  - -0.028*** (0.004)
  - -0.022*** (0.004)
  - -0.024*** (0.004)
  - -0.025*** (0.005)
  - -0.022*** (0.004)
  - -0.026*** (0.004)
  - 0.021** (0.010)
  - 0.019* (0.010)
  - 0.025** (0.011)
  - 0.017* (0.010)
  - 0.021** (0.010)
  - 0.021** (0.010)
  - 0.023** (0.010)
  - 0.012 (0.012)
- oil:
  - -0.011*** (0.003)
  - -0.009*** (0.003)
  - -0.006 (0.005)
  - -0.012*** (0.003)
  - -0.01*** (0.003)
  - -0.012*** (0.003)
  - -0.011*** (0.003)
  - 0.001 (0.006)
  - -0.019*** (0.005)
  - -0.014*** (0.005)
  - -0.02*** (0.005)
  - -0.02*** (0.005)
  - -0.017*** (0.005)
  - -0.02*** (0.004)
  - -0.017*** (0.005)
  - -0.017* (0.009)
- gaz:
  - -0.028* (0.016)
  - -0.007 (0.016)
  - -0.018 (0.016)
  - -0.024 (0.016)
  - -0.028* (0.016)
  - -0.028* (0.016)
  - -0.019 (0.016)
  - -0.016 (0.019)
  - 0.022 (0.025)
  - 0.055** (0.028)
  - 0.016 (0.025)
  - 0.019 (0.027)
  - 0.009 (0.026)
  - 0.02 (0.025)
  - 0.028 (0.025)
  - 0.097** (0.046)
- mineral:
  - 0.083*** (0.023)
  - 0.089*** (0.022)
  - 0.114*** (0.022)
  - 0.093*** (0.022)
  - 0.074*** (0.022)
  - 0.075*** (0.022)
  - 0.095*** (0.022)
  - 0.109*** (0.024)
  - -0.012 (0.049)
  - 0.011 (0.046)
  - 0.002 (0.053)
  - 0.046 (0.047)
  - -0.015 (0.043)
  - -0.004 (0.047)
  - 0.007 (0.051)
  - 0.039 (0.049)

### Constants, sample sizes, goodness-of-fit, and diagnostics
- Constant (various models):
  - 4.260*** (0.350)
  - 3.496*** (0.362)
  - 4.010*** (0.332)
  - 3.846*** (0.353)
  - 3.915*** (0.320)
  - 4.313*** (0.315)
  - 3.936*** (0.333)
  - 3.997*** (0.429)
  - 0.585 (1.538)
  - -0.293 (1.369)
  - 0.633 (1.484)
  - 0.211 (1.517)
  - 0.442 (1.419)
  - 1.520 (1.384)
  - 1.121 (1.400)
  - -0.430 (1.583)
- Obs.: 2,896; 2,975; 2,848; 2,733; 2,975; 2,975; 2,873; 2,606; 342; 350; 329; 301; 350; 350; 342; 280
- YFE: Yes (reported for all models)
- R2: 0.1163; 0.1397; 0.1414; 0.1285; 0.1371; 0.1202; 0.1412; 0.1460; 0.3374; 0.3441; 0.3384; 0.3488; 0.3265; 0.2976; 0.3271; 0.3937
- Chi2: 0.00*** for all reported models

### Notes (as reported)
- Standard errors are clustered at the bank level and are reported in parentheses below their coefficient estimates.
- * Statistical significance at the 10% level.
- ** Statistical significance at the 5% level.
- *** Statistical significance at the 1% level.

*Source: IMF Working Paper (excerpted table content provided in the source PDF).*

### chapter 7       0.012***

### chapter 7

### Main regression findings (BCP compliance and bank stability)
- BCP (α_BCP) coefficients on Z-score (bank stability) are consistently positive and often statistically significant:
  - 0.012*** (0.002)
  - 0.014*** (0.003)
  - 0.015*** (0.003)
  - 0.017*** (0.003)
  - 0.018*** (0.003)
  - 0.019*** (0.003)
- Interaction effects (BCP × Islamic, α_BCP_inter) on Islamic banks’ stability relative to conventional banks are mostly small and generally not statistically significant in many specifications:
  - 0.01 (0.007)
  - 0.021*** (0.006)
  - 0.007 (0.005)
  - 0.012** (0.006)
  - 0.01** (0.005)
  - 0.001 (0.004)
  - 0.005 (0.006)

- Model summary statistics (selected):
  - Obs.: ranges reported include 3238, 3325, 3177, 3034, 3325, 3225, 3215.
  - R2: 0.1227, 0.149, 0.1439, 0.1331, 0.1435, 0.1216, 0.1472.
  - Chi2: 0.00*** across many specifications.

### Subsample and regional breakdowns (alternative samples)
- Panel A.1 (Excluding GCC / SEA / MENA) — BCP (α_BCP) effects:
  - Excluding GCC: CBs 0.014*** (0.003); IBs 0.009 (0.010); Full 0.014*** (0.003).
  - Excluding SEA: CBs 0.008* (0.004); IBs 0.001 (0.010); Full 0.009** (0.004).
  - Excluding MENA: CBs 0.014*** (0.003); IBs 0.016* (0.009); Full 0.016*** (0.003).
- Interaction (BCP × Islamic) in these subsamples:
  - -0.008 (0.008), -0.001 (0.012), -0.007 (0.009) — generally small and not significant.
- Obs. and R2 (examples):
  - Obs.: 2369, 172, 2541; 1756, 178, 1934; 2190, 246, 2436.
  - R2: 0.1575, 0.2635, 0.1618; 0.1064, 0.4582, 0.1138; 0.1633, 0.4152, 0.1628.

### Other subsamples and robustness of interaction effect
- Excluding UK / unlisted / listed banks (Panel B.2) — BCP (α_BCP):
  - Excluding UK: CBs 0.015*** (0.003); IBs 0.019** (0.009); Full 0.017*** (0.003).
  - Excluding unlisted banks: CBs 0.02*** (0.004); IBs 0.03** (0.015); Full 0.021*** (0.004).
  - Excluding listed banks: CBs 0.017*** (0.005); IBs 0.01 (0.014); Full 0.018*** (0.005).
- Panel B.2 reported interaction impacts after exclusions:
  - 0.014* (0.008), 0.024** (0.009), 0.01 (0.011).

- Crisis-period subsamples (Panel C.1) — BCP (α_BCP):
  - Excluding period before 2007/2009 crisis: CBs 0.006* (0.003); IBs 0.016* (0.008); Full 0.008** (0.003).
  - Excluding 2007/2009 crisis period: CBs 0.021*** (0.003); IBs 0.005 (0.010); Full 0.021*** (0.003).
  - Excluding period after 2007/2009 crisis: CBs 0.012*** (0.003); IBs 0.004 (0.011); Full 0.014*** (0.003).
- Interaction impacts across time subsamples:
  - 0.01 (0.008), 0.011 (0.009), 0.001 (0.010).

### Alternative subsampling considerations (Panel D.1)
- Selected BCP (α_BCP) estimates:
  - Unstable political environment: CBs 0.007 (0.006); IBs 0.031 (0.026); Full 0.009 (0.006).
  - Weak protection of depositors: CBs -0.21* (0.119); IBs -0.328** (0.159); Full -0.296*** (0.098).
  - Highly efficient banks: CBs 0.019*** (0.004); IBs 0.015* (0.009); Full 0.025*** (0.003).
- Interaction impacts on selected other subsamples:
  - 0.006 (0.009), -0.272*** (0.097), 0.013 (0.008) — illustrating heterogeneity across subsamples.

### Quantile regression results (Table 6)
- Panel A: BCP (α_BCP) across quantiles for Conventional banks:
  - Q25 0.013*** (0.003)
  - Q50 0.017*** (0.003)
  - Q75 0.018*** (0.004)
- For Full sample quantiles:
  - Q25 0.017*** (0.003)
  - Q50 0.019*** (0.003)
  - Q75 0.019*** (0.004)
- Panel B: Impact of BCP compliance on different quantiles of Islamic banks’ stability (interaction):
  - 0.017* (0.008)
  - 0.009 (0.006)
  - 0.009 (0.011)

### Alternative measures of risk (Table 7)
- Panel A: Impact of BCP (α_BCP) on alternative risk measures:
  - LLRGLP: -0.049*** (0.019) for Conventional banks; -0.05*** (0.019) for Full sample.
  - LLPTLP: -0.012** (0.005) for Conventional banks; -0.014*** (0.005) for Full sample.
  - NPLGLP: -0.059** (0.029) for Conventional banks; -0.058* (0.029) for Full sample.
  - SDNIM: -0.009*** (0.002) for Conventional banks; -0.012*** (0.003) for Full sample.
- Panel B: Interaction (BCP + α_inter) impact on Islamic banks’ risk measures:
  - -0.053 (0.070), 0.009 (0.032), -0.047 (0.088), -0.003 (0.007) — not uniformly significant.

### Robustness checks: alternative estimation techniques and standards errors (Table 8)
- Panel A: BCP (α_BCP) remains positive and significant across estimation techniques and standard error treatments:
  - Conventional banks (Truncated): 0.019*** (0.003)
  - Conventional banks (Bootstrap): 0.015*** (0.002)
  - Conventional banks (White): 0.016*** (0.002)
  - Islamic banks (Truncated): 0.005 (0.006)
  - Full sample (Truncated): 0.018*** (0.003), (Bootstrap) 0.017*** (0.002), (White) 0.018*** (0.002)
- Panel B (Propensity score matching) — Average treated-minus-control differences (examples):
  - K-Nearest neighbors n = 10: Conventional 3.737 0.169 1.68*; Full 3.727 0.465 5.36***
  - K-Nearest neighbors n = 15: Conventional 3.737 0.288 3.00***; Full 3.727 0.425 4.94***
  - Kernel: Conventional 3.737 0.123 1.2; Full 3.727 0.443 5.28***
  - Radius: Conventional 3.737 0.261 6.65***; Islamic 3.595 0.384 3.11***; Full 3.727 0.273 7.28***

### Endogeneity checks (Table 9)
- IV / Heckman / 2SLS / GMM specifications find BCP (α_BCP) positive and often significant:
  - Conventional banks IV: 0.013*** (0.003)
  - Conventional banks 2SLS / GMM: 0.019*** (0.004)
  - Islamic banks IV/2SLS (selected): 0.014*** (0.002), 0.02*** (0.004)
  - Full sample IV: 0.014*** (0.002), Full sample 2SLS: 0.02*** (0.004)
- Panel B: Impact of BCP compliance on Islamic banks’ stability (interaction) under endogeneity checks:
  - 0.016*** (0.005), 0.015*** (0.005), 0.018** (0.008)

### Key control variable signs and notable coefficients (selected)
- wgi (world governance index) generally positive and statistically significant in many models: e.g., 0.242*** (0.047), 0.261*** (0.047).
- gdpg (GDP growth) positive and significant: e.g., 0.039*** (0.008), 0.042*** (0.008), 0.044*** (0.008).
- inf (inflation) negative and significant: e.g., -0.017*** (0.004), -0.015*** (0.004).
- cirp (cost-to-income ratio) negative and significant: e.g., -0.006*** (0.001).
- gtap (growth of total assets) negative and often significant: e.g., -0.002*** (0.001).
- mineral (share of mineral rents) positive and sometimes significant: e.g., 0.055** (0.022), 0.07*** (0.021).

### Appendix: comparison and variable definitions (selected)
- Table A.1: Mapping between Basel Core Principles (BCPs) and IFSB Core Principles for Islamic Finance Regulation (CPIFR) — several principles retained unchanged; others adapted (e.g., Principle 2 Permissible activities; Principle 6 Capital adequacy; Principle 16 Rate of return risk).
- Table A.2: Variable definitions (selected):
  - Z-score: natural logarithm of ((ROAAP+TETAP)/SDROAA).
  - AROAA: return on average assets divided by standard deviation of ROAA.
  - LLRGLP: Bank reserves for loan losses divided by gross loans times 100.
  - LLPTLP: Bank provisions for loan losses divided by total loans times 100.
  - NPLGLP: Bank non-performing loans divided by gross loans times 100.
  - SDNIM: standard deviation of Net interest margin for a three-year period.
  - lnta: natural logarithm of total assets (Bankscope).
  - gtap: current year growth rate of bank total assets (Bankscope).
  - cirp: share of bank costs to bank income before provisions times 100 (Bankscope).
  - ladstfp: ratio of liquid assets to deposits and short term funding (Bankscope).
  - BCP index: overall index computed as average of seven chapters, values between 0 and 100 (IMF/World Bank FSAP database).

*Italic: Source — wp17161 - chapter 7 (from the supplied chapter content).*

### Chapter 1 This index is a normalized sum of the rates of compliance with sub-principles of principle 1 and

### Chapter 1 This index is a normalized sum of the rates of compliance with sub-principles of principle 1 and 

### Overview
- Purpose: Measures the extent to which the preconditions for effective banking supervision have been met.
- Scale: This index takes values between 0 and 100, with values closer to 100 indicate better adherence to these preconditions.
- Data source: IMF/World Bank Basel Core Financial Sector Assessment Program (FSAP) database.

### Components (sub-principles aggregated)
- 1(1): There should be clear responsibilities and objectives set by legislation for each supervisory agency.
- 1(2): Each supervisory agency should possess adequate resources to meet the objective set, provided on terms that do not undermine the autonomy, integrity and independence of supervisory agency.
- 1(3): A suitable framework of banking laws, setting bank minimum standard, including provisions related to authorization of banking establishments and their supervision.
- 1(4): The legal framework should provide power to address compliance with laws as well as safety and soundness concerns.
- 1(5): The legal framework should provide protection of supervisors for actions taken in good faith in the course of performing supervisory duties.
- 1(6): There should be arrangements of interagency cooperation, including with foreign supervisors, for sharing information and protecting the confidentiality of such information.

### Key interpretation
- Higher index values (closer to 100) indicate better adherence to the listed preconditions for effective banking supervision.
- The index is a normalized sum of compliance rates across the six listed sub-principles.

---

### Chapter 2 — Licensing and structural influence (principles 2–5)
- Index definition: Normalized sum of the compliance rates of principles 2-5.
- Components:
  - 2: Definition of permissible activities.
  - 3: Right to set licensing criteria and reject applications for establishments that do not meet the standard sets.
  - 4: Authority to review and reject proposals for significant ownership changes.
  - 5: Authority to establish criteria for reviewing major acquisitions or investments.
- Scale: Values between 0 and 100, with values closer to 100 indicate greater power of supervisors to licence and influence structure.
- Data source: IMF/World Bank Basel Core Financial Sector Assessment Program (FSAP) database.

### Chapter 3 — Capital adequacy and prudential requirements (principles 6–15)
- Index definition: Normalized sum of the rates of compliance with principles 6–15.
- Components (6–15): 
  - 6: Prudent and appropriate risk-adjusted capital adequacy ratios must be set.
  - 7: Supervisors should evaluate banks’ credit policies.
  - 8: Banks should adhere to adequate loan evaluation and loan-loss provisioning policies.
  - 9: Supervisors should set limits to restrict large exposures, and concentration in bank portfolios should be identifiable.
  - 10: Supervisors must have in place requirements to mitigate the risks associated with related lending.
  - 11: Policies must be in place to identify, monitor, and control country risks, and to maintain reserves against such risks.
  - 12: Systems must be in place to accurately measure, monitor, and adequately control markets risks, and supervisors should have powers to impose limits or capital charge on such exposures.
  - 13: Banks must have in place a comprehensive risk management process to identify, measure, monitor, and control all other material risks and, if needed, hold capital against such risks.
  - 14: Banks should have internal control and audit systems in place.
  - 15: Adequate policies, practices, and procedures should be in place to promote high ethical and professional standards and prevent the bank being used by criminal elements.
- Scale: Values between 0 and 100, with values closer to 100 indicating a greater compliance cost for banks of adherence to the minimum capital requirements.
- Data source: IMF/World Bank Basel Core Financial Sector Assessment Program (FSAP) database.

### Chapter 4 — Ongoing supervision (principles 16–20)
- Index definition: Normalized sum of the rates of compliance rates with principles 16–20.
- Components (16–20):
  - 16: An effective supervisory system should consist of on-site and off-site supervision.
  - 17: Supervisors should have regular contact with bank management.
  - 18: Supervisors must have a means of collecting, reviewing, and analyzing prudential reports and statistics returns from banks on a solo and consolidated basis.
  - 19: Supervisors must have a means of independent validation of supervisory information, either through on-site examinations or use of external auditors.
  - 20: Supervisors must have the ability to supervise banking groups on a consolidated basis.
- Scale: Values between 0 and 100, with values closer to 100 suggesting higher levels of on-going supervision.
- Data source: IMF/World Bank Basel Core Financial Sector Assessment Program (FSAP) database.

### Chapter 5 — Internal records and disclosure (principle 21)
- Variable definition: Normalized compliance rate for principle 21.
- Principle 21: Each bank must maintain adequate records that enable the supervisor to obtain a true and fair view of the financial condition of the bank, and must publish on a regular basis financial statements that fairly reflect its condition.
- Scale: Values between 0 and 100, with values closer to 100 suggesting more requirements for information disclosure on banks by supervisors.
- Data source: IMF/World Bank Basel Core Financial Sector Assessment Program (FSAP) database.

### Chapter 6 — Formal supervisory powers (principle 22)
- Index definition: Normalized compliance rate of principle 22.
- Principle 22: Adequate supervisory measures must be in place to bring about corrective action when banks fail to meet prudential requirements when there are regulatory violations, or when depositors are threatened in any other way. This should include the ability to revoke the banking license or recommend its revocation.
- Scale: Values between 0 and 100, with values closer to 100 indicating greater supervisory powers.
- Data source: IMF/World Bank Basel Core Financial Sector Assessment Program (FSAP) database.

### Chapter 7 — Global consolidated supervision (principles 23–25)
- Index definition: Normalized sum of the compliance rates of principles 23-25.
- Components (23–25):
  - 23: Supervisors must practice global consolidated supervision over internationally active banks, adequately monitor, and apply prudential norms to all aspects of the business conducted by these banks.
  - 24: Consolidated supervision should include establishing contact and information exchange with the various supervisors involved, primarily host country supervisory authorities.
  - 25: Supervisors must require the local operations of foreign banks to be conducted at the same standards as required of domestic institutions, and must have powers to share information needed by the home country supervisors of those banks.
- Scale: Values between 0 and 100, with values closer to 100 suggesting a movement towards global consolidated supervision.
- Data source: IMF/World Bank Basel Core Financial Sector Assessment Program (FSAP) database.

---

### Additional variables and data sources (definitions)
- wgi: The world governance index is the average of six governance dimensions including: (1) voice and accountability, (2) political stability and absence of violence, (3) government effectiveness, (4) regulatory quality, (5) rule of law, and (6) control of corruption. — World governance indicators database (The World Bank and Kaufmann et al. (2013))
- gdpg: Growth rate of GDP — World Development Indicators (WDI)
- inf: Inflation rate, based on changes in the consumer price index — World Development Indicators (WDI)
- oil: Oil rents are the difference between the value of crude oil production at world prices and total costs of production. — World Development Indicators (WDI)
- gaz: Natural gas rents are the difference between the value of natural gas production at world prices and total costs of production. — World Development Indicators (WDI)
- mineral: Mineral rents are the difference between the value of production for a stock of minerals (tin, gold, lead, zinc, copper, nickel, silver, bauxite, and phosphate) at world prices and total costs of production. — World Development Indicators (WDI)

*IMF/World Bank Basel Core Financial Sector Assessment Program (FSAP) database; World governance indicators database (The World Bank and Kaufmann et al. (2013)); World Development Indicators (WDI).*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2017/wp17161.pdf_