## A New Measure of Central Bank Independence (wpiea2024035-print-pdf)

## Source details

**Canonical URL:** [A New Measure of Central Bank Independence (wpiea2024035-print-pdf)](https://www.imf.org/-/media/files/publications/wp/2024/english/wpiea2024035-print-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2024/english/wpiea2024035-print-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2024/english/wpiea2024035-print-pdf.pdf.json)

---

### Purpose and contribution of the paper
- Constructs a new de jure central bank independence (CBI) index—presented as the first entirely new index in three decades not based on CWN.
- Key methodological contributions:
  - Simplifies prior frameworks (which included upwards of forty distinct variables) down to ten key metrics.
  - Defines variables conservatively to avoid overstating independence; e.g., treats the status of the chief executive as a composite metric requiring both a meaningful term in office and protection from removal by the executive branch.
  - Replaces subjective weightings with an empirically grounded framework using survey data from senior officials at 87 monetary authorities.

### The ten metrics included in the new index (each scored 0 or 1)
- 1. Independence of the Chief Executive
  - Yes if: Governor/CEO appointed for longer than the electoral cycle (i.e., often at least five years); appointment includes involvement of non-executive officials (approval, ratification, nomination, or consultation); can be dismissed by government officials only “for cause” and not for reasons relating to policy; is compensated; cannot hold other high ranking executive positions at the same time.
  - “Governor”/“CEO” includes titles such as “president” or “chairman.”
- 2. Independence of the Highest Governing Body
  - Yes if: members appointed for longer than the electoral cycle (i.e., often at least five years) with staggered terms; appointment by or with approval of non-executive officials; dismissible by government officials outside the central bank only “for cause” and not for policy reasons; compensated; barred from holding other high ranking executive positions at the same time.
  - Presence of ex officio government officers on the highest governing body (with or without voting rights) means the body is not independent.
- 3. Budgetary Independence
  - Yes if: the legislative power has no involvement with the central bank’s budgeting process on an annual basis (including to propose, approve, or suspend the budget).
- 4. Independence in Formulating Monetary Policy
  - Yes if: the central bank can set policy without approval of other government officials and there are no government officials (with or without voting rights) on the body that sets monetary policy (e.g., board or MPC).
  - If the central bank only advises other bodies which can adopt or reject its recommendations, there is no independence on this metric.
- 5. Primary Objective of the Central Bank
  - Yes if: price stability is the only objective or other objectives are clearly subordinated to price stability.
  - No if: statute specifies other objectives not clearly subordinated to price stability (e.g., financial stability, maximum employment, supporting government economic policies, climate change, setting government spending policy).
- 6. Long-Term Direct Lending to the Government
  - Yes if: central bank prohibited from lending long-term to the government directly or purchasing long-term securities in the primary market.
  - Long-term lending defined as any lending (loans, advances, primary market purchases, or guarantees) to the government for periods greater than 90 days.
- 7. Short-Term Direct Lending to the Government
  - Yes if: central bank prohibited from lending to the government for 90 days or less (including loans, advances, primary market purchases, or guarantees).
  - If permitted to lend in “emergencies” (e.g., war, natural disasters) then it is not prohibited and would score zero. Short-term lending also includes seasonal imbalances.
- 8. Financial Independence
  - Yes if: separately capitalized (own capital, separate from government/public entities) with paid up capital; can control distributions to the government; has a reserve fund; can exclude unrealized gains from net profit.
- 9. Lending Outside the Financial System
  - Yes if: central bank cannot provide loans, advances, or guarantees to entities not regulated and supervised by financial regulatory/supervisory authorities, and cannot conduct quasi-fiscal activities allocating credit to nonfinancial institutions.
- 10. Monetary Policy Audit
  - Yes if: law specifies a state audit body can only examine operational efficiency of the central bank and not policy-related decisions.
  - “State audit body” includes national audit office, national evaluation office, or comptroller-general.

### Methodological choices, scoring rules, and exclusions
- Composite necessary-features approach:
  - Variables like chief executive independence are treated as composite concepts requiring multiple conditions; no partial credit is allowed.
  - Rationale: avoid crediting laws that appear to offer decisional independence but contain loopholes that render it illusory.
- Scoring rule: each variable is set equal to either zero or one; no partial credit. A country can score a one only if a series of specific requirements are met.
- Simplified lending assessment:
  - Three lending types highlighted: temporary emergency lending to government, long-term lending to government, and lending outside the financial system.
  - Restrictions on lending to the government for long periods are considered more important than other lending restrictions.
- Dropped items:
  - The index drops information regarding central bank responsibility for supervision and foreign exchange policy due to complex, heterogeneous relationships to monetary policy autonomy and difficulty of universal comparison across countries.

### Empirical scope and data coverage
- Index development draws on:
  - IMF’s Central Bank Legislation Database (CBLD) and Monetary Operations and Instruments Database (MOID).
  - A new survey of 87 central banks.
- Preliminary scoring applied to enabling laws for 151 monetary authorities using the methodology, weightings, and 2020/2021 data.
- The forthcoming paper will release detailed country-level data, summary statistics, and empirical analysis using this 151-sample.

### Survey design, respondent composition, and weighting procedure
- Population invited: senior central bank officials at 193 monetary authorities.
- Response count: 87 responses.
- Respondent roles (counts shown in source figure): 20 Governors; 37 General Counsel / Head of Legal; 19 Senior Staff; 11 Other.
- Monetary unions included among respondents: Central Bank of West African States, European Central Bank, Eastern Caribbean Central Bank (three of the five monetary unions surveyed).
- Weighting procedure:
  - Respondents asked to indicate the most appropriate weight for each variable: 0 (should not be considered); 1 (not important); 2 (less important); 3 (important); 4 (very important); 5 (critical).
  - Respondents were instructed to weigh based on general importance for central banks globally, not based on their own central bank or monetary union.
  - Respondents randomly divided into four groups; each group received the overview of ten metrics in a different order to counter any ordering bias.
  - Definitions of each metric were provided and an open text box allowed additional comments.

### Survey coverage, regional and market-type patterns, and exact response distributions
- Regional and income-level representation:
  - Responses drawn from all five world regions (Africa; Asia and Pacific; Europe; Middle East and Central Asia; Western Hemisphere) and a range of income levels.
  - Response rates by group:
    - High income countries: responses from 70 percent of countries.
    - European countries: responses from 80 percent of countries.
    - African countries: responses from 17 percent of countries.
    - Low-income countries: responses from 23 percent of countries.
    - Advanced economies: responses from 86 percent of countries.
- Market type composition (counts and percentages as presented):
  - Advanced Economies 32 86%
  - Low Income Developing Countries 19 32%
  - Emerging Market and Developing Countries 33 34%
  - Monetary Union 3
- Exchange rate regime responses (counts):
  - Conventional peg 8
  - Crawl-like arrangement 8
  - Crawling peg 4
  - Currency board 3
  - Floating 20
  - Free floating 27
  - Other managed arrangement 2
  - Pegged exchange rate within horizontal bands 1
  - Stabilized arrangement 11
- Perceived importance of metrics (survey distribution by rating 0–5, exact percentages)
  - 0 = should not be considered: 0.0% 1.1% 2.3% 2.3% 1.1% 2.3% 3.4% 0.0% 0.0% 2.3%
  - 1 = not important: 0.0% 1.1% 2.3% 0.0% 0.0% 2.3% 1.1% 4.6% 0.0% 5.7% 0.0%
  - 2 = less important: 1.1% 8.0% 6.9% 5.7% 3.4% 3.4% 5.7% 1.1% 5.7% 2.3%
  - 3 = important: 9.2% 11.5% 5.7% 6.9% 13.8% 6.9% 6.9% 3.4% 12.6% 6.9%
  - 4 = very important: 27.6% 27.6% 17.2% 20.7% 23.0% 31.0% 21.8% 21.8% 17.2% 20.7%
  - 5 = critical: 62.1% 50.6% 65.5% 64.4% 56.3% 55.2% 57.5% 73.6% 58.6% 67.8%
- Notable survey findings:
  - Short-term lending to the government: comparatively less important, with twelve countries scoring the variable “less important,” “not important,” or “should not be considered.”
  - Restrictions on lending outside of the financial system: nine central banks responded that restrictions on lending outside of the financial system were either “not important” or “less important.”
  - Financial independence: 64 out of 87 monetary authorities thought that financial independence was critical, the most of any variable.

### Aggregate survey-derived weights for CBI variables (exact aggregate weightings reported)
- 4. Can the central bank set monetary policy independently from executive branch officials? 58%
- 5. Is price stability the primary objective of the central bank? 52%
- 3. Does the central bank have budgetary independence from the legislature? 51%
- 6. Is the central bank prohibited from long-term direct lending to the government? 50%
- 9. Is the central bank restricted from lending outside the financial system? 46%
- 10. Are state audit bodies with oversight powers explicitly restricted to examining operational efficiency aspects of the central bank? 46%
- 8. Does the central bank have financial independence? 43%
- 7. Is the central bank prohibited from short-term lending to the government? 32%
- 1. Is the central bank Governor / CEO independent of executive branch officials? 25%
- 2. Is the central bank’s highest governing body independent of executive branch officials? 17%

### Comparative critique of existing indices and motivation for change
- Shortcomings of prior indices emphasized:
  - Subjective and arbitrary variable weightings across existing indices.
  - CWN example: lending rules receive 50 percent of the overall index while the central bank’s role in formulating monetary policy receives 5 percent—an imbalance producing counterintuitive rankings.
  - Partial-credit granularity can produce inconsistent outcomes: e.g., a chief executive appointed by the central bank board for a ten-year term but dismissible at will by the executive branch could nevertheless receive partial credit under previous schemes.
- Examples of scoring anomalies and granularity in prior work:
  - A bank with full control over monetary policy formulation but very strict lending limits and no power to buy government securities in the primary market scores just 0.416 (out of 0.55) under prior granular schemes.
  - Romelli’s equal weighting of six categories treats reporting/disclosure as equivalent to control over monetary policy.
  - Romelli dismissal distinctions and appointment-source waterfall produce fine-grained credits criticized as excessively precise and potentially misleading.
- Relationship to other frameworks:
  - New index is closest to IAPOC in granularity and similar variables (board composition, financial independence, budgetary independence) but differs by adding weighting based on the representative central bank respondent sample and using IMF CBLD and MOID data for scoring.

### CBLD and MOID data sources, mapping, and scoring procedure
- CBLD coverage and update moments:
  - The CBLD contains central bank laws, excerpts of constitutions, and other relevant legislation of 175 IMF member countries and monetary unions, grouped into 273 unique categories.
  - The CBLD currently includes datasets from four specific update moments: 2010, 2015, 2020/2021, and 2023.
  - All collected data is in English only—either provided by the authorities or translated by the IMF.
- Scoring approach and quality control:
  - For the initial index, CBLD data for 2020/2021 based on preselected search categories for central bank independence were used.
  - Two research officers manually scored the CBLD data, cross-checking each other’s work; compiled overview reviewed with ambiguous cases flagged and random sample checks.
  - For 30 countries, the 2020/2021 legislation was not available in the CBLD, and the 2015 dataset was used to complete the scores.
  - MOID was used as an additional source to aid scoring, focusing on category A (Monetary Policy Institutional Arrangements, questions 2a – 5a).
  - Observed discrepancy example: several central banks self-reported price stability as the primary objective in the MOID survey, though the CBLD data suggested that was not the case.
- Mapping of the ten metrics to CBLD search categories (exact mappings preserved):
  - 1 Independence of the Chief Executive: 3.23 Governor: Requirements, Terms, Nomination, Selection, Appointment; 3.28 Governor: Resignation; 3.29 Governor: Subsequent Employment; 3.31 General Manager
  - 2 Independence of the Highest Governing Body: 3.09 Governing Bodies: Composition, Terms, Qualification Requirements, Nomination, Selection, Appointment; 3.12 Governing Bodies: Ineligibility/Inability to Serve in the Governing Bodies; 3.13 Governing Bodies: Disqualification and Removal
  - 3 Budgetary Independence: 2.17 Financial Autonomy; 4.03 Budget; 9.01 Central Bank as Advisor to the Government
  - 4 Independence in Formulating Monetary Policy: 2.32 Prohibited Activities of the Central Bank; 6.01 Monetary Policy: Institutional Arrangement; 9.01 Central Bank as Advisor to the Government
  - 5 Primary Objective of the Central Bank: 2.02 - 2.11 Legal Status and Objectives of the Central Bank
  - 6 Long-Term Direct Lending to the Government: 9.02 Fiscal Agency, Depository Functions, and Cashier of the Government; 9.03 Central Bank Credit to the Government and Purchase of Government Securities; 2.32 Prohibited Activities of the Central Bank
  - 7 Short-Term Direct Lending to the Government: 12.01 Lender of Last Resort Function / Objectives; 9.06 Lending and Guaranteeing for Development Purposes and other Quasi-fiscal Activities
  - 8 Financial Independence: 2.17 Financial Autonomy; 4.01 – 4.09 Financial Provisions
  - 9 Lending Outside the Financial System: 9.06 Lending and Guaranteeing for Development Purposes and other Quasi-fiscal Activities; 2.32 Prohibited Activities of the Central Bank
  - 10 Monetary Policy Audit: 17.04 – 17.08 Accounts, Financial Statements, Audits and their Publication; 3.32 Comptroller General/Audit Committee; 3.33 Internal Auditor; 3.34 External Auditor

### Preliminary comparisons, outcomes, and next steps
- Using the CBLD and MOID, the authors score the enabling statutes and associated legal provisions governing 151 monetary authorities.
- Sample comparison notes:
  - The sample is largely consistent with the sample scored by Romelli (2022) and Garriga (2016), which cover 154 and 152 statutes, respectively.
  - Preliminary comparisons suggest the new index produces lower independence scores on average across the board.
  - One reason for the downward shift and the greater variance is the use of composite metrics for several variables, which leads to scores of zero where the CWN approach would offer significant partial credit.
- Forthcoming work:
  - The forthcoming paper will release detailed country-level data, summary statistics, and empirical analysis using this 151-sample.
  - Whether there is a statistically significant difference between equal weighting of the variables and weighting based on the survey responses will be examined in more detail in the forthcoming paper.

*Source: IMF Working Paper "A New Measure of Central Bank Independence" (excerpt provided).*

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

### References

### Major themes: Purpose and contribution of the paper
- Constructs a new de jure central bank independence (CBI) index—presented as the first entirely new index in three decades not based on CWN.
- Key methodological contributions:
  - Simplifies prior frameworks (which included upwards of forty distinct variables) down to ten key metrics.
  - Defines variables conservatively to avoid overstating independence; e.g., treats the status of the chief executive as a composite metric requiring both a meaningful term in office and protection from removal by the executive branch.
  - Replaces subjective weightings with an empirically grounded framework using survey data from senior officials at 87 monetary authorities.

### The ten metrics included in the new index
- Independence of the chief executive from executive branch officials.
- Independence of the highest governing body from executive branch officials.
- Budgetary independence of the central bank from the legislature.
- Policy independence from the executive branch.
- The role of price stability among the central bank’s statutory objectives.
- Whether the central bank is prohibited from long-term direct lending to the government.
- Whether the central bank is prohibited from direct short-term lending to the government.
- Whether the central bank has financial autonomy.
- Whether the central bank is restricted from lending outside the financial system.
- Whether state audit bodies are barred from investigating monetary policy decisions.

### Empirical scope and data coverage
- Index development draws on:
  - IMF’s Central Bank Legislation Database (CBLD) and Monetary Operations and Instruments Database (MOID).
  - A new survey of 87 central banks (survey respondents described in Figures and Boxes).
- Preliminary scoring applied to enabling laws for 151 monetary authorities using the methodology, weightings, and 2020/2021 data.
- The forthcoming paper will release detailed country-level data, summary statistics, and empirical analysis using this 151-sample.

### Critique of existing measures and motivation for change
- Prior dominant indices and features summarized:
  - Bade and Parkin (BP): categorical approach (policy and finances) for 12 industrial countries, no composite scores or variable weightings.
  - Grilli, Masciandaro, and Tabellini (GMT): 15 metrics across political and economic dimensions; binary scoring with one exception receiving two points for certain supervisory arrangements.
  - Cukierman, Webb, and Neyapti (CWN, 1992): original coverage of 72 countries, 16 variables coded on a 0–1 scale with sub-weights (0.17, 0.33, 0.5, 0.67, 0.83). Variables granular (e.g., term of office scoring: 1 for >8 years, 0.75 for 6–8 years, 0.5 for 5 years, 0.25 for 4 years, 0 for shorter/no term). Dismissal rules and lending terms have similarly detailed sub-weights.
  - Later augmentations: Jâcome and Vázquez (2005) added/removed variables and increased emphasis on accountability; Dincer and Eichengreen (2014) added measures on reappointment and board members; Romelli expanded to a 52-variable CBIE with six equally weighted parts and a dataset back to 1923; Unsal, Papageorgiou, and Garbers (2022) proposed the IAPOC framework with three unweighted pillars.
- Specific shortcomings highlighted:
  - Subjective and arbitrary variable weightings across existing indices.
  - Example: CWN weights lending rules as 50 percent of the overall index while assigning 5 percent to the central bank’s role in formulating monetary policy—an imbalance that can produce counterintuitive rankings (e.g., a bank with no policy role but no lending authority could score relatively highly).

### Methodological choices and rationale
- Conservative definitional approach to variables to avoid overclaiming decisional autonomy when statutory features are insufficient in isolation.
- Empirically derived weightings based on the new survey of 87 monetary authorities—intended to align index weights with expert opinion among central bankers rather than researcher-imposed heuristics.
- Reduction of variable count to ten metrics intended to preserve substantive coverage (adding areas absent from CWN such as board composition, financial independence, and budgetary independence) while improving interpretability.

### Comparative and historical context (selected exact values and examples)
- CWN original variable sub-weights: 0.17, 0.33, 0.5, 0.67, 0.83.
- CWN chief executive term scoring: 1 for greater than eight years; 0.75 for six to eight years; 0.5 for five years; 0.25 for four years; 0 for shorter/no term.
- CWN dismissal scoring examples: 0.83 if dismissal must be for cause; 0.67 if dismissal is at the discretion of the bank’s board; 0.5 if dismissal is at the discretion of the legislative; 0.33 if legislative dismissal can be unconditional; 0.17 if the executive dismissal is permitted; 0 if unconditional dismissal by the executive is permitted.
- CWN variable weighting example: four chief executive variables each weighted 0.05 (totaling 0.2); policy formulation totals 0.15; statutory objectives count for 0.15; eight lending-related variables total 0.5.
- Historical dataset coverage examples: BP limited to 12 countries between 1972 and 1986; GMT covered 15 countries; CWN covered 72 countries; Romelli extends dataset back to 1923.
- Preliminary dataset for the new index: laws scored for 151 monetary authorities using 2020/2021 data.

*IMF Working Paper: A New Measure of Central Bank Independence (selected content from the References and introductory sections).*

### 0.5 points (out of 0.55 points allocated to these dimensions) while a bank with full control over monetary policy

### A New Measure of Central Bank Independence (wpiea2024035-print-pdf)

### Critiques of existing indices (CWN, Romelli, IAPOC)
- Examples of scoring anomalies and granularity:
  - A bank with full control over monetary policy formulation but very strict lending limits (advances and securitized lending up to only fifteen percent of government revenues, with terms set by the bank, maturities no more than six months, and at penalty rates) and no power to buy government securities in the primary market scores just 0.416 (out of 0.55).
  - Romelli’s expansion of the CWN index weights each of his six categories equally, resulting in reporting/disclosure being treated as equivalent to control over monetary policy.
  - The variable regarding price stability as a statutory objective (the only variable in the statutory objectives dimension) accounts for 16.67 percent of a country’s overall score, whereas the central bank reporting variable (one of two variables in the reporting and disclosure dimension) accounts for 8.3 percent overall, and whether the central bank has the exclusive right to determine and approve its annual budget (one of twelve variables under financial independence) is weighted just 1.4 percent overall.
  - Past disaggregation allows “partial credit” in ways that can be inconsistent: e.g., if a chief executive is appointed by the central bank board for a term of ten years and cannot hold other offices, the central bank receives 0.15 points (out of 0.2 points allocated to this dimension) even if the executive branch can dismiss the chief executive at any time for no reason.
- Questionable inference rules and arbitrary relative weightings:
  - Appointment-source waterfall: chief executive/board scores most independent if appointed by private shareholders (or self-perpetuating board); 25 percent less independent if appointed by a council of executive, legislature, and existing board; 50 percent less independent if appointed by the legislative branch alone; 75 percent less independent if appointed by a council of executive branch officials; and not independent at all if one or more executive branch officials makes the appointments alone.
  - Dismissal distinctions in Romelli’s formulation: dismissal for policy reasons at the legislative branch’s discretion credits 0.5 points; dismissal at will by the legislative branch credits 0.33 points; dismissal by the executive branch for policy reasons counts for 0.17; dismissal by the executive at will counts for zero; removal at will by the central bank board scores 0.67 points.
  - These fine-grained distinctions are criticized as excessively precise because removal “at will” versus removal “for policy reasons” can leave central bankers equally vulnerable.
- Ambiguous crediting of features with unclear effect on independence:
  - Central bank responsibility for bank supervision is penalized in some indices despite arguments that consolidating supervision under central bank control can be beneficial for monetary policy and financial conditions.
  - Reporting requirements and statutory requirements to publish certified financial statements may not necessarily enhance independence; e.g., extensive reporting requirements were imposed on the U.S. Federal Reserve in the late 1970s to reduce its independence.
- IAPOC observations:
  - IAPOC provides a useful barometer of monetary policy frameworks (policy strategy and communications) but devotes less attention to aspects of independence beyond operational independence.
  - Weightings in IAPOC are assigned on the basis of authors’ analysis and preferences; it evaluates independence and accountability using publicly available resources while keeping averaging unweighted.

### De jure vs de facto independence
- De facto variables have added value but are difficult to capture robustly:
  - Example: voluntary turnover of governors/board members shortly after a new government (e.g., every five years) versus official dismissal for non-policy reasons every three or four years—unclear which is better or worse.
  - The report argues further research is needed to develop proper de facto variables and proposes inclusion in future work.

### Construction of the new index — overview
- Purpose: a new index focused on independence of the central bank as a monetary authority — the first entirely new index focused on independence since CWN in 1992.
- Coverage: major areas addressed by CWN, GMT, and Romelli but in a more parsimonious fashion.
- Structure: consists of ten key metrics, weighted using data collected from 87 central banks.
- Data source for scoring: IMF’s central bank legislation database (catalogues authoritative versions of relevant legal provisions).
- Scoring rule: each variable is set equal to either zero or one; no partial credit. A country can score a one only if a series of specific requirements are met (composite necessary features).

### The ten metrics (each scored 0 or 1)
1. Independence of the Chief Executive
   - Yes if: Governor/CEO appointed for longer than the electoral cycle (i.e., often at least five years); appointment includes involvement of non-executive officials (approval, ratification, nomination, or consultation); can be dismissed by government officials only “for cause” and not for reasons relating to policy; is compensated; cannot hold other high ranking executive positions at the same time.
   - “Governor”/“CEO” includes titles such as “president” or “chairman.”
2. Independence of the Highest Governing Body
   - Yes if: members appointed for longer than the electoral cycle (i.e., often at least five years) with staggered terms; appointment by or with approval of non-executive officials; dismissible by government officials outside the central bank only “for cause” and not for policy reasons; compensated; barred from holding other high ranking executive positions at the same time.
   - Note: presence of ex officio government officers on the highest governing body (with or without voting rights) means the body is not independent.
3. Budgetary Independence
   - Yes if: the legislative power has no involvement with the central bank’s budgeting process on an annual basis (including to propose, approve, or suspend the budget).
4. Independence in Formulating Monetary Policy
   - Yes if: the central bank can set policy without approval of other government officials and there are no government officials (with or without voting rights) on the body that sets monetary policy (e.g., board or MPC).
   - Note: if the central bank only advises other bodies which can adopt or reject its recommendations, there is no independence on this metric.
5. Primary Objective of the Central Bank
   - Yes if: price stability is the only objective or other objectives are clearly subordinated to price stability.
   - No if: statute specifies other objectives not clearly subordinated to price stability (e.g., financial stability, maximum employment, supporting government economic policies, climate change, setting government spending policy).
6. Long-Term Direct Lending to the Government
   - Yes if: central bank prohibited from lending long-term to the government directly or purchasing long-term securities in the primary market.
   - Long-term lending defined as any lending (loans, advances, primary market purchases, or guarantees) to the government for periods greater than 90 days.
7. Short-Term Direct Lending to the Government
   - Yes if: central bank prohibited from lending to the government for 90 days or less (including loans, advances, primary market purchases, or guarantees).
   - Note: if permitted to lend in “emergencies” (e.g., war, natural disasters) then it is not prohibited and would score zero. Short-term lending also includes seasonal imbalances.
8. Financial Independence
   - Yes if: separately capitalized (own capital, separate from government/public entities) with paid up capital; can control distributions to the government; has a reserve fund; can exclude unrealized gains from net profit.
9. Lending Outside the Financial System
   - Yes if: central bank cannot provide loans, advances, or guarantees to entities not regulated and supervised by financial regulatory/supervisory authorities, and cannot conduct quasi-fiscal activities allocating credit to nonfinancial institutions.
10. Monetary Policy Audit
   - Yes if: law specifies a state audit body can only examine operational efficiency of the central bank and not policy-related decisions.
   - “State audit body” includes national audit office, national evaluation office, or comptroller-general.

### Index design choices and exclusions
- Composite necessary-features approach:
  - Variables like chief executive independence are treated as composite concepts requiring multiple conditions; no partial credit is allowed.
  - Rationale: avoid crediting laws that appear to offer decisional independence but contain loopholes that render it illusory.
- Simplified lending assessment:
  - Three lending types highlighted: temporary emergency lending to government, long-term lending to government, and lending outside the financial system.
  - Prior: restrictions on lending to the government for long periods are considered more important than other lending restrictions.
- Dropped items:
  - The index drops information regarding central bank responsibility for supervision and foreign exchange policy due to complex, heterogeneous relationships to monetary policy autonomy and difficulty of universal comparison across countries.

### Relationship to IAPOC and other indices
- New index is closest to IAPOC in granularity and similar variables (board composition, financial independence, budgetary independence).
- Distinction: IAPOC uses unweighted averaging of publicly available resources; the new index adds weighting based on a representative sample of central bank respondents and uses IMF CBLD and MOID data for country scoring.

### Survey design and weights
- Survey fundamentals:
  - Population invited: senior central bank officials at 193 monetary authorities.
  - Response count: 87 responses.
  - Monetary unions included among respondents: Central Bank of West African States, European Central Bank, Eastern Caribbean Central Bank (three of the five monetary unions surveyed).
  - Respondent roles (counts shown in source figure): 20 Governors; 37 General Counsel / Head of Legal; 19 Senior Staff; 11 Other.
- Weighting procedure:
  - Respondents asked to indicate the most appropriate weight for each variable: 0 (should not be considered); 1 (not important); 2 (less important); 3 (important); 4 (very important); 5 (critical).
  - Respondents were instructed to weigh based on general importance for central banks globally, not based on their own central bank or monetary union.
  - Respondents randomly divided into four groups; each group received the overview of ten metrics in a different order to counter any ordering bias.
  - Definitions of each metric were provided and an open text box allowed additional comments.
- Respondent coverage and representation:
  - Responses drawn from central banks in all five world regions (Africa; Asia and Pacific; Europe; Middle East and Central Asia; Western Hemisphere), a range of income levels (low to high), and all three market types (advanced economies; low income developing countries; emerging market and developing countries).
  - Response rates by group:
    - High income countries: responses from 70 percent of countries.
    - European countries: responses from 80 percent of countries.
    - African countries: responses from 17 percent of countries.
    - Low-income countries: responses from 23 percent of countries.
    - Advanced economies: responses from 86 percent of countries.

*Source: IMF Working Paper "A New Measure of Central Bank Independence" (excerpt provided).*

### Box 1. Survey Responses by Region and Income Level

### Box 1. Survey Responses by Region and Income Level

### Regional and income-level patterns
- "Nonetheless, differences based on region, income level, and market type were relatively muted."
- "One difference bears note: on average, European central banks were more likely to rate variables “critical” than central banks in other regions as were central banks in high income countries."

### Market type composition (as presented)
- "Our sample also draws on a range of foreign exchange regimes including countries with a conventional peg, crawl-like arrangement, crawling peg, currency board, floating rate, free floating rate, other managed arrangement, pegged exchange rate within horizontal bands, and stabilized arrangement."
- Market TypeResponsesPercent of Total
  - Advanced Economies3286%
  - Low Income Developing Countries1932%
  - Emerging Market and Developing Countries3334%
  - Monetary Union3

### Exchange rate regime responses (counts provided)
- Exchange Rate RegimeResponses
  - Conventional peg8
  - Crawl-like arrangement8
  - Crawling peg4
  - Currency board3
  - Floating20
  - Free floating27
  - Other managed arrangement2
  - Pegged exchange rate within horizontal bands1
  - Stabilized arrangement11

### Perceived importance of metrics (survey summary)
- "Respondents on average viewed all ten metrics as at least “very important.” However, there was still meaningful variation between metrics."
- Short-term lending to the government: "comparatively less important, with twelve countries scoring the variable “less important,” “ not important,” or “should not be considered.”"
- Restrictions on lending outside of the financial system: "nine central banks responded that restrictions on lending outside of the financial system were either “not important” or “less important.”"
- Financial independence: "64 out of 87 monetary authorities thought that financial independence was critical, the most of any variable."
- Note on CWN index: "financial independence was entirely absent from the CWN index (in Romelli’s extension of the CWN index, what we call “financial independence” is combined with measures of what we define as “budgetary independence” and “state audit” to comprise 16.67 percent of the overall score, collectively)."

### Survey response distribution by rating (0–5) across ten metrics (as presented)
- Metrics listed (by numbered question):
  1. Is the central bank Governor / CEO independent of executive branch officials?
  2. Is the central bank’s highest governing body independent of executive branch officials?
  3. Does the central bank have budgetary independence from the legislature?
  4. Can the central bank set monetary policy independently from executive branch officials?
  5. Is price stability the primary objective of the central bank?
  6. Is the central bank prohibited from long-term direct lending to the government?
  7. Is the central bank prohibited from short-term lending to the government?
  8. Does the central bank have financial independence?
  9. Is the central bank restricted from lending outside the financial system?
  10. Are state audit bodies with oversight powers explicitly restricted to examining operational
- Response category percentages shown (rows correspond to 0 = should not be considered through 5 = critical), by metric position (columns correspond to metrics 1–10) — values presented exactly as in the source:
  - 0 = should not be considered
    - 0.0% 1.1% 2.3% 2.3% 1.1% 2.3% 3.4% 0.0% 0.0% 2.3%
  - 1 = not important
    - 0.0% 1.1% 2.3% 0.0% 0.0% 2.3% 1.1% 4.6% 0.0% 5.7% 0.0%
  - 2 = less important
    - 1.1% 8.0% 6.9% 5.7% 3.4% 3.4% 5.7% 1.1% 5.7% 2.3%
  - 3 = important
    - 9.2% 11.5% 5.7% 6.9% 13.8% 6.9% 6.9% 3.4% 12.6% 6.9%
  - 4 = very important
    - 27.6% 27.6% 17.2% 20.7% 23.0% 31.0% 21.8% 21.8% 17.2% 20.7%
  - 5 = critical
    - 62.1% 50.6% 65.5% 64.4% 56.3% 55.2% 57.5% 73.6% 58.6% 67.8%

### Survey weights for CBI variables (aggregate weights reported)
- VariableWeighting
  - 4. Can the central bank set monetary policy independently from executive branch officials?58%
  - 5. Is price stability the primary objective of the central bank?52%
  - 3. Does the central bank have budgetary independence from the legislature?51%
  - 6. Is the central bank prohibited from long-term direct lending to the government?50%
  - 9. Is the central bank restricted from lending outside the financial system?46%
  - 10. Are state audit bodies with oversight powers explicitly restricted to examining operational efficiency aspects of the central bank?46%
  - 8. Does the central bank have financial independence?43%
  - 7. Is the central bank prohibited from short-term lending to the government?32%
  - 1. Is the central bank Governor / CEO independent of executive branch officials?25%
  - 2. Is the central bank’s highest governing body independent of executive branch officials?17%

### Comparison and interpretation notes
- "Figure 2 below allows a comparison of the aggregated scores of our CBI index (based on weights assigned by the survey respondents) with the results of country scores using Romelli’s index."
- "Whether there is a statistically significant difference between equal weighting of the variables and weighting based on the survey responses will be examined in more detail in our forthcoming paper."
- "Preliminary comparisons between indices suggest that our index produces lower independence scores on average across the board."
- "One reason for the downward shift and the greater variance is our use of composite metrics for several variables, which leads to scores of zero where the CWN approach would offer significant partial credit."
- Sample scoring coverage: "Using the CBLD and MOID, we score the enabling statutes and associated legal provisions governing 151 monetary authorities."
- Sample comparison: "Our sample is largely consistent with the sample scored by Romelli (2022) and Garriga (2016), which cover 154 and 152 statutes, respectively."

### CBLD and MOID data sources and scoring procedure (relevant to survey context)
- CBLD description and coverage:
  - "The CBLD is the most comprehensive central bank legislation database in the world, and contains central bank laws, excerpts of constitutions, and other relevant legislation (such as selected banking, resolution, payments, and AML/CFT laws) of 175 IMF member countries and monetary unions, grouped into 273 unique categories, see https://data.imf.org/cbld;"
  - "The CBLD currently includes datasets from four specific update moments: 2010, 2015, 2020/2021, and 2023."
  - "CBLD data can be accessed by searching by country, or by pre-set groups of countries (notably, by region, income level, exchange rate arrangements, and membership of a monetary union)."
  - "All collected data is in English only—either provided by the authorities or translated by the IMF (in which case a disclaimer would be added, noting that the provided text is not an official translation)."
  - "Figure 3 below shows CBLD data coverage for the 2020/2021 update, in addition to data coverage for the combined CBLD updates of 2010 and 2015."
- CBLD process and enhancements:
  - "The CBLD’s 273 categories cover all topics of relevance to central banks ... and allow for highly granular searches."
  - "As part of the CBLD update process, search categories are manually assigned to articles or even sentences within an article by a team of central bank lawyers as part of the CBLD’s update (see Khan (2017))."
  - "Additionally, as of 2023, Artificial Intelligence (AI) and Machine Learning (ML) tools are used by the IMF to enhance this process (AlAjmi et al. (2023))."
- Scoring approach and quality control:
  - "For our initial index, we use CBLD data for 2020/2021 based on preselected search categories for central bank independence."
  - Two research officers manually scored the CBLD data, cross-checking each other’s work; compiled overview reviewed with ambiguous cases flagged and random sample checks.
  - "For 30 countries, the 2020/2021 legislation was not available in the CBLD, and the 2015 dataset was used to complete the scores."
  - "Additionally, data from the IMF’s Monetary Operations and Instruments Database (MOID) were used as an additional source to aid scoring ... with a focus on category A (Monetary Policy Institutional Arrangements, questions 2a – 5a)."
  - Observed discrepancy example: "we found that several central banks self-reported price stability as the primary objective in the MOID survey, though the CBLD data suggested that was not the case."

### CBI-relevant CBLD search categories (mapping of the ten metrics to CBLD categories)
- 1 Independence of the Chief Executive
  - 3.23 Governor: Requirements, Terms, Nomination, Selection, Appointment
  - 3.28 Governor: Resignation
  - 3.29 Governor: Subsequent Employment
  - 3.31 General Manager
- 2 Independence of the Highest Governing Body
  - 3.09 Governing Bodies: Composition, Terms, Qualification Requirements, Nomination, Selection, Appointment
  - 3.12 Governing Bodies: Ineligibility/Inability to Serve in the Governing Bodies
  - 3.13 Governing Bodies: Disqualification and Removal
- 3 Budgetary Independence
  - 2.17 Financial Autonomy
  - 4.03 Budget
  - 9.01 Central Bank as Advisor to the Government
- 4 Independence in Formulating Monetary Policy
  - 2.32 Prohibited Activities of the Central Bank
  - 6.01 Monetary Policy: Institutional Arrangement
  - 9.01 Central Bank as Advisor to the Government
- 5 Primary Objective of the Central Bank
  - 2.02 - 2.11 Legal Status and Objectives of the Central Bank
- 6 Long-Term Direct Lending to the Government
  - 9.02 Fiscal Agency, Depository Functions, and Cashier of the Government
  - 9.03 Central Bank Credit to the Government and Purchase of Government Securities
  - 2.32 Prohibited Activities of the Central Bank
- 7 Short-Term Direct Lending to the Government
  - 12.01 Lender of Last Resort Function / Objectives
  - 9.06 Lending and Guaranteeing for Development Purposes and other Quasi-fiscal Activities
- 8 Financial Independence
  - 2.17 Financial Autonomy
  - 4.01 – 4.09 Financial Provisions
- 9 Lending Outside the Financial System
  - 9.06 Lending and Guaranteeing for Development Purposes and other Quasi-fiscal Activities
  - 2.32 Prohibited Activities of the Central Bank
- 10 Monetary Policy Audit
  - 17.04 – 17.08 Accounts, Financial Statements, Audits and their Publication
  - 3.32 Comptroller General/Audit Committee
  - 3.33 Internal Auditor
  - 3.34 External Auditor

*Source: Authors. (Box 1. Survey Responses by Region and Income Level, wpiea2024035-print-pdf)*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2024/english/wpiea2024035-print-pdf.pdf_