## _wp0782

## Source details

**Canonical URL:** [_wp0782](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2007/_wp0782.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2007/_wp0782.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2007/_wp0782.pdf.json)

---

### Introduction and objectives
- Purpose: investigate the relationship between components of regulatory governance (independence, accountability/integrity, transparency) and the quality of banking supervision.
- Databases and scope:
  - BCP assessments: 116 countries.
  - MFP Transparency Code assessments: 53 countries.
  - Countries common to both databases: 51.
  - Assessments considered as of end-2004.
- Method: qualitative scores mapped on a 4-point scale and converted into quantitative indices for empirical analysis.
- Key high-level findings:
  - Banking supervisors’ compliance with the operational independence principle was among the least fully compliant principles assessed.
  - Substantial weaknesses exist in preventive and remedial powers of banking supervisors.
  - Transparency requirements are easier and cheaper to fulfill than BCP requirements.
  - Positive correlation between supervisor transparency and effectiveness of banking supervision.
  - Better accountability and integrity practices are associated with higher BCP compliance, especially where supervisory independence is higher.
  - Europe and Central Asia score highest in BCP and transparency assessments; South East Asia and the Pacific shows above-average compliance in remedial measures and making financial policies public.

### The three main components of regulatory governance
- Independence
  - Degree to which RSAs are insulated from political pressure and industry influence.
  - Distinction between goal independence and instrument independence.
  - Instrument independence dimensions: regulatory, supervisory, institutional, budgetary.
- Accountability (together with integrity)
  - Degree to which RSAs are responsible for actions, required to explain and justify them, and to make amends and avoid recurrence.
  - Mechanisms: periodic reporting, appearances before parliament, external representation, internal inspectorate, financial audit.
  - RSA accountability is more complex than central bank accountability due to multiple objectives, confidentiality needs, broader set of stakeholders, and coercive sanctioning powers.
- Transparency
  - Degree to which information about RSA activities is verifiable and communicated timely.
  - Three disclosure components:
    - (i) clarity of roles, responsibilities and objectives and periodic performance explanation;
    - (ii) well-founded legal, institutional and economic framework;
    - (iii) provision of data and information to create an informed view on financial policies.
- Interrelations
  - Accountability and independence are complementary; accountability supports independence.
  - Transparency enables accountability (by disclosing actions) and supports independence (by highlighting interference).

### International codes on banking supervision — empirical measures and patterns (BCP database)
- BCP composition and assessment coverage:
  - BCP composed of 25 core principles with CP 1 having 6 sub-principles, totaling 30 assessable principles.
  - Assessment missions: IMF and WB, 1999 to 2004.
  - Total theoretically assessable principles: 3,480 (116 countries × 30 assessable principles).
  - Assessed principles: 3,364.
  - Not applicable/no answer assessable principles: 116.
- Aggregate assessment distribution (over the 3,364 assessed principles):
  - 38.32 percent rated as compliant (1,289 out of 3,364).
  - 32.04 percent rated as largely compliant (1,078).
  - 22.5 percent rated as materially non-compliant (757).
  - 7.14 percent rated as non-compliant (240).
  - Interpretation: over 70 percent of assessed core principles have at least a largely compliant assessment.
- Compliance by chapter (selected highlights):
  - Chapter 1 (Preconditions for effective banking supervision): 76.55 percent of assessed principles scored as compliant or largely compliant.
  - Chapter 2 (Licensing and approval powers): 83.05 percent scored as compliant or largely compliant.
  - Weakest chapters: chapter 3 (prudential regulations and requirements) and chapter 6 (remedial measures available to supervisors) — almost 40 percent of assessed principles in each were scored as materially non-compliant or non-compliant.
- Compliance by stage of financial development:
  - Sample composition: 24 advanced countries, 26 emerging economies, 66 developing countries.
  - Observed decreasing trend of compliance from advanced to emerging and developing countries across all chapters.

### Analytical approach and index construction
- Conversion of qualitative 4-point assessments into quantitative indices for cross-country analysis:
  - Scoring: 0 = non-compliant/not observed; 1 = materially non-compliant/Partially observed; 2 = largely compliant/broadly observed; 3 = compliant/fully observed.
  - BCP index: unweighted aggregation of numerical scores divided by theoretical maximum 90 (3 * 30 CP).
  - Transparency index: unweighted aggregation of numerical scores divided by theoretical maximum 108 (3 * 36 practices).
  - Weighting for not applicable/not assessed principles: chapter-level weighting coefficient = (number of possible assessments in chapter) / (number of actually assessed principles/practices).
  - If all principles/practices of a chapter are not applicable/not assessed, chapter given score 0.
- Appendix I describes construction methodology for indices of compliance/observance for BCP and MFP Transparency Code in banking supervision.

### Cross-country compliance — chapters 4 and 5 and overall sample
- Chapters 4 (methods of ongoing supervision) and 5 (information requirements) degrees of compliance (compliant plus largely compliant):
  - Chapter 4: 95.62 percent.
  - Chapter 5: 100 percent.
  - Suggestion: strong financial market development proceeds in parallel with a high level of supervision and information requirements.
- BCP sample size: 116 countries.
- Compliance categories used: Compliant, Largely compliant, Materially non-compliant, Non-compliant.
- Regional and income classifications used in analysis: ADV=Advanced, EME=Emerging, DEV=Developing; EU/CA=Europe and Central Asia, ME/NA=Middle East and North Africa, SEA/P=South-East Asia and Pacific, SSA=Sub-Saharan Africa, WH=Western Hemisphere; H=High income, UM=Upper middle income, LM=Lower Middle Income, L=Low income.

### Compliance by geographic area
- Country counts by geographic area:
  - Europe and Central Asia: 39 countries
  - Middle East and North Africa: 12 countries
  - South-East Asia and the Pacific: 19 countries
  - Sub-Saharan Africa: 16 countries
  - Western Hemisphere: 30 countries
- Key findings:
  - Europe and Central Asia exhibit the highest degree of compliance overall.
  - Middle East and North Africa rank closely behind Europe and Central Asia and outperform in chapters 1 and 5 in some cases.
  - South-East Asia and the Pacific achieves 73.68 percent of compliant and largely compliant assessments in chapter 6 (existence of remedial measures for banks not meeting requirements).
  - Other geographic areas show generally homogeneous compliance across chapters, except chapter 6.

### Compliance by income level
- Country counts by income level:
  - High-income: 37 countries
  - Upper middle-income: 29 countries
  - Lower middle-income: 34 countries
  - Low-income: 16 countries
- Key finding:
  - A gap of 30-40 percent in the degree of full compliance exists between high-income and non-high-income economies.

### Selected characteristics of compliance rates (core principles)
- Core Principles with highest numbers of compliant assessments (fully compliant and largely compliant) in 116 countries:
  - CP 2 (Definition of permissible activities to banking institutions (chapter 2)): 111 compliant
  - CP 1.1 (Clear objectives and responsibilities for supervisory agency (chapter 1)): 103 compliant
  - CP 1.3 (Existence of legal framework for banking supervision (chapter 1)): 101 compliant
  - Licensing authority right to set criteria for licensing banks (chapter 2): 94 compliant
  - Power for supervisors to reject significant transfers of ownership in banks (chapter 2): 93 compliant
- Core Principles with largest numbers of non-compliant assessments:
  - CP 12 (Banks having systems to measure, monitor and control market risk (chapter 3)): 61 non-compliant
  - CP 11 (Banks having policies for identifying and managing country and transfer risk (chapter 3)): 57 non-compliant
  - CP 13 (Banks having systems to measure, monitor and control all other material risk (chapter 3)): 55 non-compliant
  - CP 15 (Setting of rules for the prevention of fraud and money laundering (chapter 3)): 52 non-compliant
  - CP 20 (Ability of supervisors to oversee banking groups on a consolidated basis (chapter 4)): 51 non-compliant
- Specific weaknesses and implications:
  - Weak compliance with CP 11, CP 12, CP 13 is worrying in light of the new Basel Capital Accord at the end of 2006.
  - Urgent action needed on weaknesses in money laundering (CP 15).
  - Urgent action needed on consolidated supervision (CP 20) due to financial conglomerates and blurring boundaries between financial products and markets.
- Sub-components of CP 1:
  - CP 1.2 (operational supervisory independence): 31 out of 116 fully compliant (lowest number of fully compliant assessments among CP 1 components).
  - CP 1.5 (adequate legal protection for supervisors): 63 fully compliant (second highest) and 19 non-compliant (third highest number of non-compliant assessments).

### Not applicable assessments (summary)
- Overall distribution of number of not applicable assessments per country (out of 30 BCPs):
  - 0 not applicable assessments: 72 countries
  - 1 not applicable assessment: 14 countries
  - 2 not applicable assessments: 14 countries
  - 3 not applicable assessments: 9 countries
  - 4 not applicable assessments: 5 countries
  - 5 not applicable assessments: 1 country
  - >5 not applicable assessments: 1 country (outlier with 22 CP out of 30 assessed as not applicable)
  - Total: 116 assessments/countries
- Distribution highlights by country group (selected):
  - Advanced countries (total 24): 20 with 0, 2 with 1, 1 with 2, 1 with >5.
  - Emerging countries (total 26): 22 with 0, 4 with 1.
  - Developing countries (total 66): 30 with 0, 8 with 1, 13 with 2, 9 with 3, 5 with 4, 1 with 5.
- Not applicable assessments concentrate in developing countries.
- Specific CPs with notable numbers of not applicable assessments:
  - CP 20: 18 not applicable assessments
  - CP 23: 25 not applicable assessments
  - CP 24: 34 not applicable assessments
  - Interpretation: banks from developing countries are less internationally active.

### IMF MFP Transparency Code on Banking Supervision (end-2004)
- MFP Transparency Code grading: fully observed, broadly observed, partially observed, not observed (4-point scale mapped to 3, 2, 1, 0 in indices).
- As of end-2004: finalized assessments for banking supervision for 53 countries, 51 overlapping with BCP database.
- Higher incidence of not applicable/not assessed practices than in BCP:
  - 469 out of 1,908 assessable practices reported as not applicable or not assessed (36 practices × 53 countries = 1,908).
  - Averaging almost 9 not applicable/not assessed practices per country versus 1 per country for the BCP.
  - Two countries from Middle East and North Africa and 1 from South-Asia received 20, 17 and 18 not applicable/not assessed scores respectively.
- Seven transparency practices assessed as not applicable or not assessed in two thirds of countries or more (counts of not ass./not appl.):
  - 5.3 Public disclosure of the role of oversight agencies on payment systems: 46
  - 5.3.1 Disclosure of general policy principles affecting robustness of payment systems: 48
  - 5.4 Disclosure of relationship between agencies overseeing payment systems: 43
  - 5.5 Enforcement to self-regulatory organizations of the same transparency practices specified for financial agencies: 44
  - 6.1.3 Public disclosure of the regulations for the operation of organized markets: 36
  - 7.6 Public disclosure of information on deposit insurance guarantees, if applicable: 38
- Interpretation: insufficient development of financial sector appears origin of non-applicability for practices dealing with payment systems, deposit insurance, and consumer protection.

### Aggregate observance and comparison with BCP
- Percentage of fully observed practices among assessed transparency practices: 71.86 percent (1,034 out of 1,439 assessed practices).
- Comparison figure for fully compliant core principles in the BCP database: 38.32 percent.
- Interpretation: transparency requirements in banking supervision are more likely to be fulfilled (and scored as fully observed) than BCP requirements.

### Observance across MFP practices 5–8 (accountability/transparency areas)
- Degree of full observance about 10 percent higher for:
  - Practice 6 (formulation and reporting of financial policies) and Practice 7 (making information on those policies available to the public)
  - than for Practice 5 (setting clear objectives, roles and responsibilities for the RSA) and Practice 8 (accountability and assurance of integrity by financial agencies).
- Strong individual practice results:
  - Provision of public information services (practice 7.4): 46 fully observed out of 53 assessments.
  - Availability of the text of regulations, directives and guidelines (practice 7.5): 48 fully observed out of 53 assessments.
- Not observed practices:
  - Percentage of not observed practices for practice 8: 13.99 percent.
- Sample composition note: nearly 80 percent emerging and developing countries; potential gap in independence and accountability where supervisors are located in central banks.

### Observance by stage of financial development and region
- Stage of development:
  - Advanced economies present higher proportions of fully and broadly observed practices than emerging and developing countries.
  - Emerging markets perform better than developing countries.
- Geographic highlights:
  - Europe and Central Asia: large proportions of full and broad observance for all practices (linked to transition economies and EU-entry reforms).
  - Western Hemisphere: practice 8 (accountability and integrity) non-observance 31.98 percent; full observance for practice 8: 44.44 percent.
  - South-East Asia and the Pacific: practice 7 full observance 85.11 percent (consistent with post-1997-98 crisis transparency reforms).
- Income-level observation:
  - Smaller gap between high income and non-high income economies for transparency practices than for BCP compliance.
  - Low-income countries’ transparency is close to upper middle- and lower middle-income countries and higher for practice 8 in some measures.

### BCP and Transparency indices — country group means and variability
- Overall sample means and standard deviations:
  - BCP Mean: 0.6660; BCP Standard Deviation: 0.1968.
  - Transparency Mean: 0.7630; Transparency Standard Deviation: 0.1467.
- By stage of development:
  - Advanced: BCP Mean 0.8600, BCP Std Dev 0.1350; Transparency Mean 0.8608, Transparency Std Dev 0.0902.
  - Emerging: BCP Mean 0.7171, BCP Std Dev 0.1432; Transparency Mean 0.7828, Transparency Std Dev 0.1259.
  - Developing: BCP Mean 0.5753, BCP Std Dev 0.1758; Transparency Mean 0.6960, Transparency Std Dev 0.1582.
- By geographic area (selected):
  - Europe and Central Asia: BCP Mean 0.7621, BCP Std Dev 0.1656; Transparency Mean 0.8431, Transparency Std Dev 0.0859.
  - Middle East and North Africa: BCP Mean 0.6985, BCP Std Dev 0.1791; Transparency Mean 0.6020, Transparency Std Dev 0.2575.
  - South-East Asia: BCP Mean 0.6194, BCP Std Dev 0.2374; Transparency Mean 0.7506, Transparency Std Dev 0.1095.
  - Sub-Saharan Africa: BCP Mean 0.5463, BCP Std Dev 0.1222; Transparency Mean 0.7154, Transparency Std Dev 0.0993.
  - Western Hemisphere: BCP Mean 0.6214, BCP Std Dev 0.1964; Transparency Mean 0.7071, Transparency Std Dev 0.1290.
- By income level:
  - High income: BCP Mean 0.8440, BCP Std Dev 0.1230; Transparency Mean 0.8485, Transparency Std Dev 0.0852.
  - Upper middle income: BCP Mean 0.6259, BCP Std Dev 0.1692; Transparency Mean 0.7906, Transparency Std Dev 0.0989.
  - Lower middle income: BCP Mean 0.5598, BCP Std Dev 0.1729; Transparency Mean 0.6959, Transparency Std Dev 0.1935.
  - Low income: BCP Mean 0.5526, BCP Std Dev 0.1447; Transparency Mean 0.7062, Transparency Std Dev 0.1240.
- Interpretations:
  - In almost all cases the transparency index is higher than the BCP index.
  - Europe and Central Asia show the highest BCP and transparency means (0.76 and 0.84).
  - Sub-Saharan Africa has the lowest BCP mean and standard deviation.
  - Middle East and North Africa: transparency mean lower than BCP mean and highest transparency standard deviation.
  - Income grouping: gap between high income and non-high income countries is much higher for the BCP index than for the transparency index.

### Regulatory governance and the quality of banking supervision — empirical associations
- Focus: empirical relationship between BCP index and three components of regulatory governance (independence, accountability and transparency).
- CP 1’s six sub-principles (SP) relate to independence, adequacy of staff/resources, legal protection for regulatory staff, enforcement powers, clarity/transparency of objectives and responsibilities, and confidentiality of shared information.

### Independence and BCP correlations
- SP 1.2 (RSA operational independence and adequacy of resources) normalized average score: 0.61 across 116 jurisdictions (least observed among CP 1 sub-principles).
- Normalized average for overall chapter 1: 0.73.
- Correlations between chapter 1 sub-index and other chapter sub-indices (All countries, statistically significant at the 99 percent level):
  - Ch. 2: 0.5997***
  - Ch. 3: 0.6182***
  - Ch. 4 to 6: 0.6460***
  - Ch. 7: 0.6044***
  - BCP minus CP 1: 0.6944***
- Correlations by development group (selected):
  - Advanced economies: Ch. 2: 0.3756*; Ch. 3: 0.4653**; Ch. 4–6: 0.5380***; Ch. 7: 0.4723**; BCP minus CP 1: 0.5458***.
  - Emerging markets: Ch. 2: 0.5224***; Ch. 3: 0.3765*; Ch. 4–6: 0.6035***; Ch. 7: 0.4729**; BCP minus CP 1: 0.5501***.
  - Developing countries: Ch. 2: 0.6001***; Ch. 3: 0.5599***; Ch. 4–6: 0.5581***; Ch. 7: 0.5703***; BCP minus CP 1: 0.6566***.
- Interpretation:
  - Strong and significant relationship between CP 1 (regulatory governance) and other chapters representing effectiveness of supervision.
  - Correlation strongest between chapter 1 and sum of other six chapters.
  - Correlation magnitudes inversely related to level of financial development: weakest in advanced economies, strongest in developing countries.

### Association of SP 1.2 (operational independence) with BCP sub-indices — correlation ratios
- Use of correlation ratio (η2) due to SP 1.2 taking only four scores (3, 2, 1, 0).
- Correlation Ratios (η2) of BCP Sub-Indices on SP 1.2:
  - Overall sample: Ch. 2 = 0.1213; Ch. 3 = 0.2091; Ch. 4 to 6 = 0.2712; Ch. 7 = 0.1305; BCP minus = 0.2603.
  - Advanced countries: Ch. 2 = 0.0256; Ch. 3 = 0.0285; Ch. 4 to 6 = 0.0461; Ch. 7 = 0.0871; BCP minus = 0.0559.
  - Emerging countries: Ch. 2 = 0.0801; Ch. 3 = 0.1152; Ch. 4 to 6 = 0.3664; Ch. 7 = 0.0921; BCP minus = 0.2829.
  - Developing countries: Ch. 2 = 0.1260; Ch. 3 = 0.1878; Ch. 4 to 6 = 0.2491; Ch. 7 = 0.1180; BCP minus = 0.2345.
- Interpretation:
  - Higher operational independence associated with higher mean compliance with BCP, especially for emerging markets and developing countries.
  - For advanced countries conditional means of single chapters are not much affected by SP 1.2 values.

### Transparency and BCP-minus-chapter1 correlations
- Analysis on 51 countries common to both databases:
  - Overall sample correlation coefficient = 0.65; statistically significant at the 99 percent level. Trend line slope = 0.97; significant at the 99 percent level.
- By stage of financial sector development:
  - Advanced countries: correlation = 0.79; significant at the 99 percent level.
  - Emerging countries: correlation = 0.29; statistically non-significant.
  - Developing countries: correlation = 0.55; significant at the 99 percent level.
- Supervisory structure subgroup analysis:
  - When supervisory functions are separated (central bank has no supervisory responsibility): correlation ≈ 0.76.
  - When monetary and financial policies jointly entrusted to central bank: correlation ≈ 0.60.
  - Both correlations positive and statistically significant; sample for subgroup analysis small (17 observations).

### Transparency in accountability and integrity versus BCP-minus-chapter1
- Constructed MFP sub-index for accountability and integrity (practices included: 5.1.3, 6.3, 7.1, 7.3, 7.4.2, 8.1, 8.2, 8.2.1, 8.2.2, 8.3, 5.1.4, 8.4, 8.4.1).
- Correlation results (BCP-minus-chapter1 vs. Transparency in Accountability and Integrity):
  - Overall sample correlation coefficient = 0.57; statistically significant at the 99 percent level.
  - Advanced economies: correlation = 0.88; statistically significant at the 99 percent level.
  - Developing countries: correlation = 0.44; statistically significant at the 95 percent level.
  - Emerging countries: correlation = 0.41; statistically not significant.
- Interpretation:
  - Strongest association in advanced countries; evidence of complementarity between independence and accountability of regulatory and supervisory agencies.

### Key empirical findings and implications (Conclusions)
- Best BCP scores:
  - Preconditions for effective banking supervision, licensing process and approval for changes in banks’ structure.
- Weakest BCP areas:
  - Chapter 3 (prudential regulations and requirements) and Chapter 6 (formal power for supervisors to take adequate remedial measures).
  - Implication: RSA preventive and remedial powers should be substantially improved; more human and financial resources needed.
- Transparency observations:
  - Transparency requirements more often fully observed than BCP requirements.
  - Highest full observance in transparency practices: practice 6 (formulation and reporting of financial policies) and practice 7 (making information on those policies available to the public).
  - Practice 8 (accountability and assurance of integrity) has percentage of not observed practices twice the average of non-observance for other main practices; especially pronounced in the Western Hemisphere.
- Regional observations:
  - South-East Asia and the Pacific: excellent performance in remedial measures for banks (chapter 6), reflecting reforms post-1997-98 financial crisis.
  - Europe and Central Asia: highest BCP and transparency values; influenced by Eastern European transition economies and EU accession reforms.
- Governance and supervision interdependence:
  - Countries with high observance of the transparency code tend to have high compliance with the BCP.
  - Positive significant correlation between regulatory governance (CP 1) and effectiveness of banking supervision (CP 2-25); correlation inversely related to level of financial development.
  - Degree of compliance with BCP improves when operational independence increases, especially for emerging markets and developing countries.
  - Better accountability and integrity practices of RSA strongly associated with higher RSA independence, which in turn is associated with better BCP compliance.
- Causality not examined; authors note that good governance leading to high quality supervision seems more likely.

### Appendix I — Construction of indices for BCP and MFP Transparency Code
- Scoring:
  - 0 = non-compliant/not observed
  - 1 = materially non-compliant/Partially observed
  - 2 = largely compliant/broadly observed
  - 3 = compliant/fully observed
- Weighting for not applicable/not assessed principles/practices:
  - Chapter weighting coefficient = (number of possible assessments in chapter) / (number of actually assessed principles/practices).
  - Example for chapter 2 of BCP (4 possible assessments CP 2–CP 5):
    - If CP assessments: CP 2 = Compliant (3); CP 3 = Materially non-compliant (1); CP 4 = Largely compliant (2); CP 5 = Not applicable
    - Number of assessed principles = 3; weighting coefficient = 4/3 = 1.333
    - Chapter 2 total score = (3*1.333 + 1*1.333 + 2*1.333) = 8
  - If all principles/practices of a chapter are not applicable/not assessed, chapter is given score 0.
- Overall index:
  - Unweighted aggregation of assessed principle scores multiplied by chapter weighting coefficients where applicable.
  - Normalized indices obtained by dividing overall score by maximum value:
    - BCP maximum value = 90.
    - MFP Transparency Code maximum value = 108.

*Source: _wp0782 - References; authors’ elaborations and calculations.*

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

### _wp0782 - References..............................................................................................................

### Introduction and objectives
- Purpose: investigate the relationship between components of regulatory governance (independence, accountability/integrity, transparency) and the quality of banking supervision.
- Databases and scope:
  - BCP assessments: 116 countries.
  - MFP Transparency Code assessments: 53 countries.
  - Countries common to both databases: 51.
  - Assessments considered as of end-2004.
- Method: qualitative scores mapped on a 4-point scale and converted into quantitative indices for empirical analysis.
- Key high-level findings:
  - Banking supervisors’ compliance with the operational independence principle was among the least fully compliant principles assessed.
  - Substantial weaknesses exist in preventive and remedial powers of banking supervisors.
  - Transparency requirements are easier and cheaper to fulfill than BCP requirements.
  - Positive correlation between supervisor transparency and effectiveness of banking supervision.
  - Better accountability and integrity practices are associated with higher BCP compliance, especially where supervisory independence is higher.
  - Europe and Central Asia score highest in BCP and transparency assessments; South East Asia and the Pacific shows above-average compliance in remedial measures and making financial policies public.

### The three main components of regulatory governance
- Components defined and their roles:
  - Independence:
    - Degree to which RSAs are insulated from political pressure and industry influence.
    - Distinction between goal independence and instrument independence.
    - Instrument independence dimensions: regulatory, supervisory, institutional, budgetary.
  - Accountability (together with integrity):
    - Degree to which RSAs are responsible for actions, required to explain and justify them, and to make amends and avoid recurrence.
    - Mechanisms: periodic reporting, appearances before parliament, external representation, internal inspectorate, financial audit.
    - RSA accountability is more complex than central bank accountability due to multiple objectives, confidentiality needs, broader set of stakeholders, and coercive sanctioning powers.
  - Transparency:
    - Degree to which information about RSA activities is verifiable and communicated timely.
    - Three disclosure components: (i) clarity of roles, responsibilities and objectives and periodic performance explanation; (ii) well-founded legal, institutional and economic framework; (iii) provision of data and information to create an informed view on financial policies.
- Interrelations:
  - Accountability and independence are complementary; accountability supports independence.
  - Transparency enables accountability (by disclosing actions) and supports independence (by highlighting interference).

### International codes on banking supervision — empirical measures and patterns
- Basel Core Principles (BCP) database details:
  - BCP composed of 25 core principles with CP 1 having 6 sub-principles, totaling 30 assessable principles.
  - Assessment missions: IMF and WB, 1999 to 2004.
  - Total theoretically assessable principles: 3,480 (116 countries × 30 assessable principles).
  - Assessed principles: 3,364.
  - Not applicable/no answer assessable principles: 116.
- BCP aggregate assessment distribution (over the 3,364 assessed principles):
  - 38.32 percent rated as compliant (1,289 out of 3,364).
  - 32.04 percent rated as largely compliant (1,078).
  - 22.5 percent rated as materially non-compliant (757).
  - 7.14 percent rated as non-compliant (240).
  - Interpretation: over 70 percent of assessed core principles have at least a largely compliant assessment.
- Compliance by chapter (selected highlights):
  - Preconditions for effective banking supervision (chapter 1): 76.55 percent of assessed principles scored as compliant or largely compliant.
  - Licensing and approval powers (chapter 2): 83.05 percent scored as compliant or largely compliant.
  - Chapters with weakest observance: chapter 3 (prudential regulations and requirements) and chapter 6 (remedial measures available to supervisors) — almost 40 percent of assessed principles in each were scored as materially non-compliant or non-compliant, indicating substantial room for improvement in preventive and remedial powers.
- Compliance by stage of financial development:
  - Sample composition: 24 advanced countries, 26 emerging economies, 66 developing countries.
  - A decreasing trend of compliance from advanced to emerging and developing countries is observed across all chapters.

### Analytical approach and indices
- Conversion of qualitative 4-point assessments into quantitative indices for cross-country analysis and to measure associations among:
  - BCP compliance indices (normalized and by chapter/sub-principle).
  - MFP Transparency Code observance indices.
- Appendix I describes construction methodology for indices of compliance/observance for BCP and MFP Transparency Code in banking supervision.

*Source: _wp0782 - References..............................................................................................................*

### chapter 7 on cross-border banking – are chapters 4 (methods of ongoing supervision) and 5

### chapter 7 on cross-border banking – are chapters 4 (methods of ongoing supervision) and 5 (information requirements) with degrees of compliance (compliant plus largely compliant) equal to 95.62 and 100 percent respectively, suggesting that strong financial market development proceeds in parallel with a high level of supervision and information requirements.

### Cross-country compliance (summary of Figure 2 and overall sample)
- Chapters 4 and 5 show degrees of compliance (compliant plus largely compliant) equal to 95.62 and 100 percent respectively.
- Sample size for Basel Core Principles (BCP) assessments: 116 countries.
- Compliance categories used: Compliant, Largely compliant, Materially non-compliant, Non-compliant.
- Regional and income classifications used in analysis: ADV=Advanced, EME=Emerging, DEV=Developing; EU/CA=Europe and Central Asia, ME/NA=Middle East and North Africa, SEA/P=South-East Asia and Pacific, SSA=Sub-Saharan Africa, WH=Western Hemisphere; H=High income, UM=Upper middle income, LM=Lower Middle Income, L=Low income.

### Compliance by geographic area
- Country counts by geographic area:
  - Europe and Central Asia: 39 countries
  - Middle East and North Africa: 12 countries
  - South-East Asia and the Pacific: 19 countries
  - Sub-Saharan Africa: 16 countries
  - Western Hemisphere: 30 countries
- Key findings:
  - Europe and Central Asia exhibit the highest degree of compliance overall.
  - Middle East and North Africa rank closely behind Europe and Central Asia and outperform in chapters 1 and 5 in some cases.
  - South-East Asia and the Pacific achieves 73.68 percent of compliant and largely compliant assessments in chapter 6 (existence of remedial measures for banks not meeting requirements), likely reflecting stronger sanctioning powers after the 1997-98 financial crisis.
  - The three other geographic areas (outside EU/CA and ME/NA) show a generally homogeneous degree of compliance across chapters, except for chapter 6 as noted.

### Compliance by income level
- Country counts by income level:
  - High-income: 37 countries
  - Upper middle-income: 29 countries
  - Lower middle-income: 34 countries
  - Low-income: 16 countries
- Key finding:
  - A gap of 30-40 percent in the degree of full compliance exists between high-income and non-high-income economies, suggesting a positive correlation between availability of resources and quality of banking supervision.

### Selected characteristics of compliance rates (core principles)
- Core Principles (CP) with the highest numbers of compliant assessments (measured as fully compliant and largely compliant) in the sample of 116 countries:
  - CP 2 (Definition of permissible activities to banking institutions (chapter 2)): 111 compliant
  - CP 1.1 (Clear objectives and responsibilities for supervisory agency (chapter 1)): 103 compliant
  - CP 1.3 (Existence of legal framework for banking supervision (chapter 1)): 101 compliant
  - Licensing authority right to set criteria for licensing banks (chapter 2): 94 compliant
  - Power for supervisors to reject significant transfers of ownership in banks (chapter 2): 93 compliant
- Core Principles with largest numbers of non-compliant assessments:
  - CP 12 (Banks having systems to measure, monitor and control market risk (chapter 3)): 61 non-compliant
  - CP 11 (Banks having policies for identifying and managing country and transfer risk (chapter 3)): 57 non-compliant
  - CP 13 (Banks having systems to measure, monitor and control all other material risk (chapter 3)): 55 non-compliant
  - CP 15 (Setting of rules for the prevention of fraud and money laundering (chapter 3)): 52 non-compliant
  - CP 20 (Ability of supervisors to oversee banking groups on a consolidated basis (chapter 4)): 51 non-compliant
- Specific weaknesses highlighted:
  - Weak compliance with CP 11, 12, 13 (monitoring and control of financial risks) is worrying in light of the new Basel Capital Accord at the end of 2006, which entails greater responsibilities for supervisors to evaluate banks’ internal capital assessment processes.
  - Urgent action needed on weaknesses in money laundering (CP 15).
  - Urgent action needed on consolidated supervision (CP 20) to address financial conglomerates and blurring boundaries between financial products and markets.
- Sub-components of CP 1 (clear objectives and legal framework) with notable results:
  - CP 1.2 (operational supervisory independence): lowest number of fully compliant assessments among the six components of CP 1 — 31 out of 116 fully compliant.
  - CP 1.5 (adequate legal protection for supervisors): second highest number of fully compliant assessments (63) and third highest number of non-compliant assessments (19) across the sample.

### Not applicable assessments (Table 2 summary)
- Overall distribution of number of not applicable assessments per country (out of 30 BCPs):
  - 0 not applicable assessments: 72 countries
  - 1 not applicable assessment: 14 countries
  - 2 not applicable assessments: 14 countries
  - 3 not applicable assessments: 9 countries
  - 4 not applicable assessments: 5 countries
  - 5 not applicable assessments: 1 country
  - >5 not applicable assessments: 1 country (outlier in which 22 CP out of 30 were assessed as not applicable)
  - Total: 116 assessments/countries
- Distribution by country group:
  - Advanced countries: 20 with 0, 2 with 1, 1 with 2, 0 with 3, 0 with 4, 0 with 5, 1 with >5 — total 24
  - Emerging countries: 22 with 0, 4 with 1, 0 with 2, 0 with 3, 0 with 4, 0 with 5, 0 with >5 — total 26
  - Developing countries: 30 with 0, 8 with 1, 13 with 2, 9 with 3, 5 with 4, 1 with 5, 0 with >5 — total 66
  - Europe & Central Asia: 30 with 0, 3 with 1, 4 with 2, 1 with 3, 0 with 4, 0 with 5, 1 with >5 — total 39
  - Middle East & North Africa: 8 with 0, 2 with 1, 2 with 2, 0 with 3, 0 with 4, 0 with 5, 0 with >5 — total 12
  - South-East Asia & Pacific: 11 with 0, 5 with 1, 0 with 2, 1 with 3, 1 with 4, 1 with 5, 0 with >5 — total 19
  - Sub-Saharan Africa: 4 with 0, 2 with 1, 3 with 2, 3 with 3, 4 with 4, 0 with 5, 0 with >5 — total 16
  - Western Hemisphere: 19 with 0, 2 with 1, 5 with 2, 4 with 3, 0 with 4, 0 with 5, 0 with >5 — total 30
  - High income: 29 with 0, 4 with 1, 3 with 2, 0 with 3, 0 with 4, 0 with 5, 1 with >5 — total 37
  - Upper middle income: 19 with 0, 0 with 1, 3 with 2, 4 with 3, 3 with 4, 0 with 5, 0 with >5 — total 29
  - Lower middle income: 19 with 0, 7 with 1, 5 with 2, 2 with 3, 0 with 4, 1 with 5, 0 with >5 — total 34
  - Low income: 5 with 0, 3 with 1, 3 with 2, 3 with 3, 2 with 4, 0 with 5, 0 with >5 — total 16
- Not applicable assessments concentrate in developing countries.
- Specific CPs with notable numbers of not applicable assessments:
  - CP 20 (conduct of consolidated supervision): 18 not applicable assessments
  - CP 23 (global consolidated supervision over internationally active banks): 25 not applicable assessments
  - CP 24 (information exchange with other supervisors): 34 not applicable assessments
  - These reflect that banks from developing countries are less internationally active.

### IMF MFP Transparency Code on Banking Supervision (end-2004 database summary)
- MFP Transparency Code assesses transparency practices for banking supervision using a 4-point grading scale: fully observed, broadly observed, partially observed, not observed.
- As of end-2004 the MFP Transparency Code database contained finalized assessments for banking supervision for 53 countries, 51 of which overlap with the BCP database.
- Larger number of transparency practices reported as not applicable or not assessed compared to BCP:
  - 469 out of 1,908 assessable practices were reported as not applicable or not assessed (36 practices × 53 countries = 1,908), averaging almost 9 not applicable/not assessed practices per country versus 1 per country for the BCP.
  - Two countries from the Middle East and North Africa and 1 from South-Asia received respectively 20, 17 and 18 not applicable/not assessed scores.
- Seven transparency practices assessed as not applicable or not assessed in two thirds of countries or more (Table 3):
  - 5.3 Public disclosure of the role of oversight agencies on payment systems: 46 not ass./not appl.
  - 5.3.1 Disclosure of general policy principles affecting robustness of payment systems: 48 not ass./not appl.
  - 5.4 Disclosure of relationship between agencies overseeing payment systems: 43 not ass./not appl.
  - 5.5 Enforcement to self-regulatory organizations of the same transparency practices specified for financial agencies: 44 not ass./not appl.
  - 6.1.3 Public disclosure of the regulations for the operation of organized markets: 36 not ass./not appl.
  - 7.6 Public disclosure of information on deposit insurance guarantees, if applicable: 38 not ass./not appl.
- Interpretation:
  - Insufficient development of the financial sector appears to be the origin of non-applicability for practices dealing with transparency in payment systems, deposit insurance, and consumer protection.

*Source: authors’ elaborations.*

### 7.7 Transparency on consumer protection arrangements 35

### 7.7 Transparency on consumer protection arrangements 35

### Aggregate observance and comparison with BCP
- Percentage of fully observed practices among assessed transparency practices: 71.86 percent (1,034 out of 1,439 assessed practices).
- Comparison figure for fully compliant core principles in the BCP database: 38.32 percent.
- Interpretation provided: transparency requirements in banking supervision are more likely to be fulfilled (and scored as fully observed) than BCP requirements, likely because disclosure is easier and cheaper than building a regulatory framework with adequate expertise and resources.

### Observance across all countries (Practices 5–8 of MFP Transparency Code)
- Degree of full observance is about 10 percent higher for:
  - Practice 6 (formulation and reporting of financial policies) and
  - Practice 7 (making information on those policies available to the public)
  than for:
  - Practice 5 (setting clear objectives, roles and responsibilities for the RSA) and
  - Practice 8 (accountability and assurance of integrity by financial agencies).
- Strong individual practice results:
  - Provision of public information services (practice 7.4): 46 fully observed out of 53 assessments.
  - Availability of the text of regulations, directives and guidelines (practice 7.5): 48 fully observed out of 53 assessments.
- Not observed practices:
  - Percentage of not observed practices for practice 8: 13.99 percent (about twice the average proportion of non-observance for the other three practices).
- Sample composition note: nearly 80 percent emerging and developing countries; potential gap in independence and accountability where supervisors are located in central banks.

### Observance by stage of financial development
- General trend: advanced economies present higher proportions of practices assessed as fully and broadly observed than emerging economies and developing countries.
- Emerging markets perform better than developing countries in the same comparison.

### Observance by geographic region
- Europe and Central Asia: large proportions of full and broad observance for all practices; attributed to presence of transition economies undertaking EU-entry-related reforms.
- Western Hemisphere:
  - High proportion of non-observance for practice 8 (accountability and integrity): 31.98 percent non-observance.
  - Low proportion of full observance for practice 8: 44.44 percent full observance.
- South-East Asia and the Pacific:
  - High level of full observance for making financial policies available to the public (practice 7): 85.11 percent; consistent with post-1997-98 crisis transparency reforms.

### Observance by income level
- Smaller gap between high income economies and non-high income economies for transparency practices than for BCP compliance.
- Transparency in low-income countries is very close to that of upper middle- and lower middle-income countries and is even higher in the case of accountability and assurance of integrity for financial agencies (practice 8).
- Interpretation: providing a transparent supervisory setting appears less resource constrained than achieving high quality banking supervision.

### Basel Core Principles (BCP) and Transparency Indices
- Two normalized indices constructed:
  - BCP index: unweighted aggregation of numerical scores divided by theoretical maximum 90.
  - Transparency index: unweighted aggregation of numerical scores divided by theoretical maximum 108.
  - Minimum theoretical value for both indices: 0.
- Comparative result: in almost all cases the transparency index is higher than the BCP index; explanation: disclosure is cheaper/easier than building legal, political and social frameworks and training staff.

### Country group means and variability (Figure 5 highlights)
- When grouped by stage of financial development:
  - Highest average values in advanced economies: 0.86 for both BCP and Transparency Code.
  - Lowest in developing countries: 0.5753 for BCP and 0.6960 for Transparency.
  - Standard deviations increase from advanced to emerging to developing countries, suggesting more heterogeneity among developing countries.

### Cross-country means and standard deviations for normalized indices (Table 4)
- Overall sample:
  - BCP Mean: 0.6660
  - BCP Standard Deviation: 0.1968
  - Transparency Mean: 0.7630
  - Transparency Standard Deviation: 0.1467
- By stage of development:
  - Advanced: BCP Mean 0.8600, BCP Std Dev 0.1350, Transparency Mean 0.8608, Transparency Std Dev 0.0902
  - Emerging: BCP Mean 0.7171, BCP Std Dev 0.1432, Transparency Mean 0.7828, Transparency Std Dev 0.1259
  - Developing: BCP Mean 0.5753, BCP Std Dev 0.1758, Transparency Mean 0.6960, Transparency Std Dev 0.1582
- By geographic area:
  - Europe and Central Asia: BCP Mean 0.7621, BCP Std Dev 0.1656, Transparency Mean 0.8431, Transparency Std Dev 0.0859
  - Middle East and North Africa: BCP Mean 0.6985, BCP Std Dev 0.1791, Transparency Mean 0.6020, Transparency Std Dev 0.2575
  - South-East Asia: BCP Mean 0.6194, BCP Std Dev 0.2374, Transparency Mean 0.7506, Transparency Std Dev 0.1095
  - Sub-Saharan Africa: BCP Mean 0.5463, BCP Std Dev 0.1222, Transparency Mean 0.7154, Transparency Std Dev 0.0993
  - Western Hemisphere: BCP Mean 0.6214, BCP Std Dev 0.1964, Transparency Mean 0.7071, Transparency Std Dev 0.1290
- By income level:
  - High income: BCP Mean 0.8440, BCP Std Dev 0.1230, Transparency Mean 0.8485, Transparency Std Dev 0.0852
  - Upper middle income: BCP Mean 0.6259, BCP Std Dev 0.1692, Transparency Mean 0.7906, Transparency Std Dev 0.0989
  - Lower middle income: BCP Mean 0.5598, BCP Std Dev 0.1729, Transparency Mean 0.6959, Transparency Std Dev 0.1935
  - Low income: BCP Mean 0.5526, BCP Std Dev 0.1447, Transparency Mean 0.7062, Transparency Std Dev 0.1240

### Interpretations from indices and groupings
- Europe and Central Asia show the highest BCP and transparency means (0.76 and 0.84), influenced by transition economies and EU-entry reforms.
- Sub-Saharan Africa has the lowest BCP mean and standard deviation, indicating resource and trained human capital constraints for supervision quality.
- Middle East and North Africa: transparency index mean is lower than BCP mean; influenced by a very low-compliance outlier and characterized by the highest transparency standard deviation among regions.
- Income grouping interpretation:
  - Higher income levels increase the quality of banking supervision more than the ability to fulfill transparency practices.
  - Gap between high income and non-high income countries is much higher for the BCP index than for the transparency index.
  - High income countries are the only group where BCP and transparency indices are almost equal (0.84 and 0.85, respectively); for other income groups observance of transparency practices is stronger than BCP compliance.

### Regulatory governance and the quality of banking supervision
- Purpose: analyze empirical relationship between BCP index (quality of effective banking supervision) and three components of regulatory governance: independence, accountability and transparency.
- Components included in parts of the BCP (CP 1) and the MFP Transparency Code; emphasis on transparency in accountability and integrity practices.
- CP 1’s six sub-principles (SP) relate to independence, adequacy of staff/resources, legal protection for regulatory staff, enforcement powers, clarity/transparency of objectives and responsibilities, and confidentiality of shared information.

### Independence and Basel Core Principles
- Independence is the only component of regulatory governance directly mentioned in the BCP.
- Statistic: across 116 jurisdictions, SP 1.2 on RSA operational independence and adequacy of resources is the least observed among the six SP regarding preconditions for effective banking supervision with an average normalized score of 0.61.
- Normalized average for overall chapter 1: 0.73.
- Correlations between chapter 1 (governance) sub-index and other chapter sub-indices (Table 5):
  - All countries correlations (statistically significant at the 99 percent level):
    - Ch. 2: 0.5997***
    - Ch. 3: 0.6182***
    - Ch. 4 to 6: 0.6460***
    - Ch. 7: 0.6044***
    - BCP minus CP 1: 0.6944***
  - Advanced economies:
    - Ch. 2: 0.3756*
    - Ch. 3: 0.4653**
    - Ch. 4 to 6: 0.5380***
    - Ch. 7: 0.4723**
    - BCP minus CP 1: 0.5458***
  - Emerging markets:
    - Ch. 2: 0.5224***
    - Ch. 3: 0.3765*
    - Ch. 4 to 6: 0.6035***
    - Ch. 7: 0.4729**
    - BCP minus CP 1: 0.5501***
  - Developing countries:
    - Ch. 2: 0.6001***
    - Ch. 3: 0.5599***
    - Ch. 4 to 6: 0.5581***
    - Ch. 7: 0.5703***
    - BCP minus CP 1: 0.6566***
  - Note: Symbols *, **, *** indicate statistically significant at the 90, 95, 99 percent level of confidence.
- Interpretation:
  - A fairly strong and significant relationship exists between CP 1 (regulatory governance) and other chapters representing effectiveness of supervision.
  - Correlation is strongest between chapter 1 and the sum of the other six chapters.
  - Correlation magnitudes are inversely related to level of financial development: weakest in advanced economies, strongest in developing countries—suggesting regulatory governance is more strongly associated with supervisory quality in developing countries.

*Source: authors’ elaboration.*

### 0.63 and 0.58, respectively. However, this difference is not statistically significant.

### _wp0782 - 0.63 and 0.58, respectively. However, this difference is not statistically significant.

### Association between RSA operational independence (SP 1.2) and BCP sub-indices
- Problem with correlation coefficient when SP 1.2 assumes only four possible scores (3, 2, 1, 0); correlation ratio (η2) used as alternative measure.
- Table 6: Correlation Ratios (η2) of BCP Sub-Indices on SP 1.2 (Operational independence and adequate resources for RSA agencies – Basel SP 1.2)
  - Overall sample: Ch. 2 = 0.1213; Ch. 3 = 0.2091; Ch. 4 to 6 = 0.2712; Ch. 7 = 0.1305; BCP minus = 0.2603
  - Advanced countries: Ch. 2 = 0.0256; Ch. 3 = 0.0285; Ch. 4 to 6 = 0.0461; Ch. 7 = 0.0871; BCP minus = 0.0559
  - Emerging countries: Ch. 2 = 0.0801; Ch. 3 = 0.1152; Ch. 4 to 6 = 0.3664; Ch. 7 = 0.0921; BCP minus = 0.2829
  - Developing countries: Ch. 2 = 0.1260; Ch. 3 = 0.1878; Ch. 4 to 6 = 0.2491; Ch. 7 = 0.1180; BCP minus = 0.2345
- Interpretation:
  - Mean dependence of BCP-index-minus-SP 1.2 on SP 1.2 exists: higher operational independence associated with higher mean compliance with BCP.
  - Dependence evident for emerging markets and developing countries; for advanced countries the conditional means of single chapters are not much affected by SP 1.2 values.
  - Suggests greater need for operational independence where financial sector is less developed.

### Transparency and Basel Core Principles (BCP)
- Analysis focuses on BCP sub-index excluding chapter 1 (BCP-minus-chapter1) when comparing with MFP Transparency Code.
- Figure 6A (51 countries common to both databases):
  - Correlation coefficient 0.65 for the overall sample; statistically significant at the 99 percent level of confidence.
  - Trend line slope = 0.97; statistically significant at the 99 percent level.
- Figure 6B (by stage of financial sector development):
  - Advanced countries: correlation = 0.79; significant at the 99 percent level.
  - Emerging countries: correlation = 0.29; statistically non-significant.
  - Developing countries: correlation = 0.55; significant at the 99 percent level.
- Interpretation:
  - Positive correlation between BCP-minus-chapter1 and transparency indices overall and within groups, strongest in advanced countries.
  - Weaker and non-significant association for emerging markets possibly due to higher dispersion within the group.
- Supervisory structure subgroup analysis:
  - When supervisory functions are separated (central bank has no supervisory responsibility) correlation ≈ 0.76.
  - When monetary and financial policies are jointly entrusted to central bank correlation ≈ 0.60.
  - Both correlations positive and statistically significant; caution due to small sample (17 observations) and potential non-normality.

### Transparency in Accountability and Integrity versus BCP
- Constructed sub-index of MFP Transparency Code for accountability and integrity (practices listed include 5.1.3, 6.3, 7.1, 7.3, 7.4.2, 8.1, 8.2, 8.2.1, 8.2.2, 8.3, 5.1.4, 8.4, 8.4.1).
- Figure 7 (BCP-minus-chapter1 vs. Transparency in Accountability and Integrity):
  - Overall sample correlation coefficient = 0.57; statistically significant at the 99 percent level.
  - Advanced economies: correlation = 0.88; statistically significant at the 99 percent level.
  - Developing countries: correlation = 0.44; statistically significant at the 95 percent level.
  - Emerging countries: correlation = 0.41; statistically not significant.
- Interpretation:
  - Strongest association in advanced countries, supporting complementarity between independence and accountability of regulatory and supervisory agencies.

### Key empirical findings and implications (Conclusions)
- Best BCP scores: preconditions for effective banking supervision, licensing process and approval for changes in banks’ structure.
- Weakest BCP areas: chapter 3 (prudential regulations and requirements) and chapter 6 (formal power for supervisors to take adequate remedial measures).
  - Implication: RSA preventive and remedial powers should be substantially improved; more human and financial resources needed.
- Transparency requirements tend to be more often fully observed than BCP requirements.
  - Highest full observance in transparency practices: formulation and reporting of financial policies (practice 6) and making information on those policies available to the public (practice 7).
  - Practice 8 (accountability and assurance of integrity) has percentage of not observed practices twice the average of non-observance for other main practices; especially pronounced in the Western Hemisphere.
- Regional observations:
  - South-East Asia and the Pacific: excellent performance in remedial measures for banks (chapter 6), likely reflecting reforms post-1997-98 financial crisis.
  - Europe and Central Asia: highest BCP and transparency values; partly due to Eastern European transition economies undergoing EU accession–related reforms.
- Governance and supervision interdependence:
  - Countries with high observance of the transparency code tend to have high compliance with the BCP.
  - Positive significant correlation between regulatory governance (CP 1) and effectiveness of banking supervision (CP 2-25); correlation inversely related to level of financial development (advanced lowest, developing highest).
  - Analysis of correlation ratios shows degree of compliance with BCP improves when operational independence increases, especially for emerging markets and developing countries.
  - Better accountability and integrity practices of RSA strongly associated with higher RSA independence, which in turn is associated with better BCP compliance.
- Note: causality between governance and supervision quality is not examined; the authors note that good governance leading to high quality supervision seems more likely.

### Appendix I — Construction of indices for BCP and MFP Transparency Code
- Scoring for each assessed principle/practice:
  - 0 = non-compliant/not observed
  - 1 = materially non-compliant/Partially observed
  - 2 = largely compliant/broadly observed
  - 3 = compliant/fully observed
- Weighting scheme for not applicable/not assessed principles/practices:
  - Weighting applied inside the chapter: weighting coefficient = (number of possible assessments in chapter) / (number of actually assessed principles/practices).
  - Example for chapter 2 of BCP (4 possible assessments CP 2–CP 5):
    - If CP assessments are: CP 2 = Compliant (3); CP 3 = Materially non-compliant (1); CP 4 = Largely compliant (2); CP 5 = Not applicable
    - Number of assessed principles = 3; weighting coefficient = 4/3 = 1.333
    - Chapter 2 total score = (3*1.333 + 1*1.333 + 2*1.333) = 8
  - If all principles/practices of a chapter are not applicable/not assessed, chapter is given score 0.
- Overall index:
  - Obtained as unweighted aggregation of assessed principle scores multiplied by chapter weighting coefficients where applicable.
  - Normalized indices obtained by dividing overall score by maximum value:
    - BCP maximum value = 90 (all compliant = 3 * 30 CP)
    - MFP Transparency Code maximum value = 108 (all fully observed = 3 * 36 practices)

*Source: authors’ calculations, IMF working paper content.*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2007/_wp0782.pdf_