## 1. Properties of Fiscal Transparency Indices

## Source details

**Canonical URL:** [1. Properties of Fiscal Transparency Indices](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2012/_wp12156.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2012/_wp12156.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2012/_wp12156.pdf.json)

---

### I. Research question, channels, and estimation approach
- Research question: whether fiscal transparency affects the assessment of sovereign risk by market participants as reflected in sovereign credit ratings.
- Two channels examined:
  - Indirect channel: transparency → better fiscal outcomes → improved credit ratings.
  - Direct channel: transparency → reduced uncertainty/creditors’ risk premia → improved credit ratings.
- Estimation approach:
  - System of equations modeling:
    - credit ratings as a function of fiscal indicators, fiscal transparency, and controls;
    - fiscal outcomes (debt and primary balance to GDP) as a function of transparency and controls;
  - Seemingly Unrelated Regression Estimation (SURE); single-equation OLS also checked.
  - Instruments explored for fiscal transparency: political competition and electoral competition; legal origin considered but correlation low in sample.

### II. Measures of fiscal transparency and key data features
- Two primary indices:
  - IMF fiscal transparency ROSCs (index constructed by Hameed (2005), range 0–100).
    - Three components: Public Information; Budget Process; Assurance of Integrity.
    - Sub-component scoring: 0, 33, 66, or 100.
    - Empirical analysis uses the average of the first two components as main ROSC index.
  - International Budget Partnership (IBP) Open Budget Index (OBI) — main index (range 0–100).
    - Based on 92 questions covering quality and comprehensiveness of 8 key budget documents and related processes.
    - Vintages used: 2006, 2008, 2010; analysis uses 2010 OBI where referenced.
- One standard deviation in the fiscal transparency index corresponds to 15 points out of 100 in both the ROSC and the OBI indices.
- Data limitations:
  - Fiscal transparency measures do not vary sufficiently over time; panel estimation not feasible.

### III. Summary statistics and cross-index relationships (preserved exactly as reported)
- ROSC Public Information
  - Mean: Advanced Economies 73.2, Developing Economies 39.8
  - Standard Deviation: Advanced Economies 16.2, Developing Economies 15.6
  - Min: Advanced Economies 37.9, Developing Economies 9.4
  - Max: Advanced Economies 95.1, Developing Economies 80.6
  - Number of Countries: Advanced Economies 22, Developing Economies 56
- ROSC Budget Process
  - Mean: Advanced Economies 79.0, Developing Economies 46.6
  - Standard Deviation: Advanced Economies 16.2, Developing Economies 15.5
  - Min: Advanced Economies 46.5, Developing Economies 7.3
  - Max: Advanced Economies 100.0, Developing Economies 89.8
  - Number of Countries: Advanced Economies 22, Developing Economies 56
- OBI Index Main
  - Mean: Advanced Economies 74.2, Developing Economies 43.1
  - Standard Deviation: Advanced Economies 12.3, Developing Economies 20.7
  - Min: Advanced Economies 57.6, Developing Economies 0.0
  - Max: Advanced Economies 90.1, Developing Economies 92.3
  - Number of Countries: Advanced Economies 12, Developing Economies 59
- Correlations:
  - ROSC Public Information correlation with OBI Index Main: 0.51
  - ROSC Budget Process correlation with OBI Index Main: 0.71
- Correlation with other proxies (selected):
  - ROSC Public Information: Average Stock Flow Adjustment -0.40; Corruption Perceptions 0.62
  - ROSC Budget Process: Average Stock Flow Adjustment -0.49; Corruption Perceptions 0.65
  - OBI Index Main: Average Stock Flow Adjustment -0.28; Corruption Perceptions 0.64
- Observations:
  - Fiscal transparency is highly correlated with income per capita (Figure 1).
  - Standard deviations of transparency indices are comparable across advanced and developing economies.
  - Coverage: fiscal ROSC index coverage for advanced economies is significantly higher than for developing economies.

### IV. Empirical strategy and model specification (exact notation preserved)
- Regression system (SURE):
  - R_i = α0 + α1 fistrans_i + α2 debt_i + α3 fisbal_i + α4 X_i + ε_i
  - fisbal_i = χ0 + χ1 fistrans_i + χ2 debt_i + χ3 growth_i + χ4 frac_i + χ5 Y_i + ε_i
  - debt_i = η0 + η1 fistrans_i + η2 debt_i(2000) + η3 growth_i + η4 frac_i + η5 Z_i + ε_i
- Direct effect of a one unit increase in transparency defined as α1.
- Indirect effect expressed in text as 243 4  (notation preserved from source).

### V. Main empirical findings (preserved numeric magnitudes)
- Overall: fiscal transparency has a positive and significant effect on credit ratings.
- Advanced economies:
  - A one standard deviation increase in transparency is associated with:
    - an increase in average primary balance of 2.4 percentage of GDP (reported elsewhere as 2 percent of GDP in tables);
    - a lower debt to GDP ratio of about 17 percent of GDP (reported elsewhere as about 16 percentage points).
  - One standard deviation improvement in fiscal transparency → indirect effect on credit ratings: increase of 0.45 notches.
  - Direct effect of one standard deviation improvement in transparency → about 0.24 notches (not statistically significant).
  - Instrumental-variable evidence: political competition used as an instrument suggests some robustness to endogeneity for the transparency → debt relationship in advanced economies.
- Developing economies:
  - Effects of transparency on primary balance and debt are small and not statistically significant across specifications.
  - Indirect effect of transparency on ratings: about 0.06-0.10 notches (depending on index and sample); this effect is not statistically significant.
  - Direct effect of one standard deviation improvement in fiscal transparency: about 0.50-0.90 notch increase; statistically significant.
  - Interpretation: weak link between transparency and fiscal outcomes and between fiscal fundamentals and ratings in developing economies; transparency’s marginal value may be high via uncertainty reduction.
  - Econometric caveat: potential omitted variables and weak instrument problems for developing-country samples prevent full causal interpretation.
- Full sample:
  - Estimates mainly driven by developing economies.
  - Estimated effect of transparency on fiscal variables is small; direct effect on ratings is significant and large.
  - Including an advanced country dummy reduces the size of the estimated direct effect of transparency; without that dummy the direct effect increases to levels comparable to Hameed (2005).

### VI. Selected quantitative outputs (exact values preserved)
- One standard deviation increase in transparency ≈ 15 points.
- Advanced sample: primary balance improvement = 2 percent of GDP per one standard deviation increase in transparency.
- Advanced sample: debt to GDP reduction ≈ 16 percentage points per one standard deviation increase in transparency.
- Advanced sample: ratings improvement ≈ 0.45 notches per one standard deviation increase in transparency.
- Developing sample: direct effect on ratings ≈ 0.5-0.9 notches per one standard deviation increase in transparency.
- Instrument strength guideline cited: "Ideally, the F statistic should be higher than 10." (Staiger and Stock (1997)).

### VII. Selected regression coefficient excerpts (preserved exact table values and significance)
- Table 6 (OBI, Developing sample) selected values and diagnostics:
  - GDP per capita (2000): 2.53  *** ; 7.36  *** ; 2.11  *** ; 7.35  ***
  - Default history: -2.67  ** ; -2.68  ***
  - Gross public debt (2000): -0.04  *** ; 0.52  ***
  - Inflation (2000-2010): -0.14  *** ; -0.14  ***
  - Real growth (2000-2010): 0.35  ** ; -0.15 ; 0.94
  - OBI Index: 0.00 ; -0.14 ; 0.03  **
  - Fuel Exporter dummy: 7.14  *** ; -34.12  *** ; 7.13  *** ; -34.14  ***
  - R-squared: 0.64 ; 0.5 ; 0.53 ; 0.68 ; 0.5 ; 0.53
  - N: 37 37 37 37 37 37
  - Direct Effect: 0.00 ; 0.52  **
  - Indirect Effect: 0.09 ; 0.06 (Due to debt: 0.09 ; 0.05; Due to primary balance: 0.00 ; 0.01)
- Table 7 (ROSC, Full sample) selected values:
  - GDP per capita (2000): 2.37  *** ; 13.11  ** ; 2.06  *** ; 12.99  **
  - Gross public debt (2000): -0.04  *** ; 0.81  ***
  - ROSC Index (Average): 0.05  * ; -0.32 ; 0.06  *** ; 0.05  * ; -0.29
  - Advanced country dummy: 4.60  *** ; -0.94 ; 11.99 ; 3.21  *** ; -0.92 ; 11.44
  - R-squared: 0.83 ; 0.43 ; 0.69 ; 0.86 ; 0.43 ; 0.69
  - N: 56 56 56 56 56 56
  - Direct Effect: 0.00 ; 0.94  ***
  - Indirect Effect: 0.18 ; 0.13
- Table 8 (OBI, Full sample) selected values:
  - GDP per capita (2000): 2.71  *** ; 8.03  ** ; 2.41  *** ; 8.07  **
  - Gross public debt (2000): -0.04  ** ; 0.55  ***
  - OBI Index: 0.01 ; -0.12 ; 0.03  *
  - Fuel Exporter dummy: 7.73  *** ; -36.31  *** ; 7.74  *** ; -36.33  ***
  - R-squared: 0.84 ; 0.34 ; 0.61 ; 0.85 ; 0.34 ; 0.61
  - N: 49 49 49 49 49 49
  - Direct Effect: 0.00 ; 0.43  *
  - Indirect Effect: 0.06 ; 0.06 (Due to debt: 0.05 ; 0.04; Due to primary balance: 0.01 ; 0.02)

### VIII. Instrumental-variable results (selected, exact statistics)
- Table 9 (Fiscal Transparency and Fiscal Performance - IV):
  - Debt (2000): -0.020 .88  ***
  - Average real growth: -0.22 -12.33  ***
  - Average Polarization: 2.39  ** -11.55  **
  - ROSC Index (Average): 0.14 -0.94  *
  - Cragg - Donald statistic (weak instrument): 5.09 5.09
  - Shea's Partial R-squared: 0.24 0.24
  - R-squared: 0.47 0.92
  - Number of Observations: 21 21
- Table 10 (Fiscal Transparency and Ratings - IV):
  - GDP per capita: 1.87  ** ; 1.75  ***
  - Default history: -0.93 ; -0.75
  - Gross public debt (2000): -0.01 ; 0.00
  - Primary balance: 0.13 ; 0.20
  - Inflation: -0.16  ** ; -0.15  *
  - Real growth: 0.26  * ; 0.28  *
  - ROSC Index (average): 0.09 ; 0.12  *
  - Cragg - Donald statistic (weak instrument): 2.60 2.93
  - Hansen J statistic (p-value): 0.56 ...
  - Shea's Partial R-squared: 0.14 0.08
  - R-squared: 0.56 0.50
  - N: 42 42
- Notes: robust t-statistics in parentheses; significance: * 10%, ** 5%, *** 1%; two-stage least squares used.

### IX. Conclusions, caveats, and suggested directions
- Conclusions:
  - Enhancing fiscal transparency is associated with better credit ratings for both advanced and developing economies.
  - Two channels identified:
    - Indirect channel: transparency improves fiscal performance (primary balance, debt), which in turn improves ratings—more evident in advanced economies.
    - Direct credibility channel: transparency directly improves market perceptions and ratings—more evident in developing economies.
- Caveats:
  - Measurement issues for transparency.
  - Endogeneity and omitted variable concerns.
  - Weak instrument problems, especially for developing-country samples and some OBI specifications.
  - Lack of time series variation in transparency measures.
- Suggested directions for future research and policy implications:
  - Research determinants of fiscal transparency to develop alternative measures and stronger instruments.
  - Explore which transparency reforms are most effective for improving fiscal performance and market perceptions.

### X. Appendix excerpts (selected exact items)
- Numerical Conversion of Sovereign Credit Ratings (Fitch, Moody's and S&P) excerpt (selected lines preserved exactly):
  - AAA: 23 ; Aaa: 20 ; AAA: 22
  - AA+: 22 ; Aa1: 19 ; AA+: 21
  - AA: 21 ; Aa2: 18 ; AA: 20
  - AA-: 20 ; Aa3: 17 ; AA-: 19
  - A+: 19 ; A1: 16 ; A+: 18
  - A: 18 ; A2: 15 ; A: 17
  - A-: 17 ; A3: 14 ; A-: 16
  - BBB+: 16 ; Baa1: 13 ; BBB+: 15
  - BBB: 15 ; Baa2: 12 ; BBB: 14
  - BBB-: 14 ; Baa3: 11 ; BBB-: 13
  - BB+: 13 ; Ba1: 10 ; BB+: 12
  - BB: 12 ; Ba2: 9 ; BB: 11
  - BB-: 11 ; Ba3: 8 ; BB-: 10
  - B+: 10 ; B1: 7 ; B+: 9
  - B: 9 ; B2: 6 ; B: 8
  - B-: 8 ; B3: 5 ; B-: 7
  - CCC+: 7 ; Caa1: 4 ; CCC+: 6
  - CCC: 6 ; Caa2: 3 ; CCC: 5
  - CCC-: 5 ; Caa3: 2 ; CCC-: 4
  - CC: 4 ; Ca: 1 ; CC: 3
  - C: 3 ; C: 0 ; C: 2
  - DDD: 2 ; SD: 1 ; DD: 1 ; D: 0 ; RD: 0
- Data source and variable definition snippets (exact text preserved for select variables):
  - Average primary balance: "Average ratio of primary fiscal balance to GDP (in percent)" — WEO Database
  - Gross public debt (2000): "Gross debt to GDP in 2000 (in percent)" — WEO Database
  - Average real growth: "Average real growth during 2001-2010." — WEO Database
  - Fuel exporter dummy: "Dummy variable that equals one if the country is a fuel exporter according to the WEO classification." — WEO Database
  - Electoral Competition: "Average of the executive index of electoral competitiveness during 2000-2010. The variable taxes values between 1 and 7, where higher values correspond to higher electoral competitiveness. The executive index of electoral competitition 'eiec' variable, Database of Political Institutions, World Bank."
  - Political Competition: "Author's calculation based on the frequency of political turnover since 1975. Using the variable that captures the name of the largest government party 'gov1me', Database of Political Institutions, World Bank."
  - GDP per capita, 2000: "GDP per capita (PPP), in US Dollars" — WEO Database
  - Inflation: "Average level of inflation (in percent) during 2001-2010." — IFS Database

*Source: _wp12156 — 1. Properties of Fiscal Transparency Indices and related sections and tables as provided in the source content.*

### 1. Properties of Fiscal Transparency Indices ...........................................................................

### 1. Properties of Fiscal Transparency Indices

### Major sections (chapter structure)
- 1. Properties of Fiscal Transparency Indices ...........................................................................10
- 2. Correlation Between Different Transparency Indices .........................................................11
- 3. Fiscal Transparency and Other Proxies for Transparency ...................................................12
- 4. Fiscal Transparency, Primary Balance and Public Debt - Advanced Economies  
    Sample..................................................................................................................................20
- 5. Fiscal Transparency (ROSC Index), Primary Balance and Public Debt - Developing     
    Country Sample ...................................................................................................................21
- 6. Fiscal Transparency (OBI Index), Primary Balance and Public Debt - Developing Country  
    Sample..................................................................................................................................22
- 7. Fiscal Transparency (ROSC Index), Primary Balance and Public Debt - Full Sample .......23
- 8. Fiscal Transparency (OBI Index), Primary Balance and Public Debt - Full Sample ..........24
- 9. Fiscal Transparency and Fiscal Performance - Instrumental Variable Results ...................25
- 10. Fiscal Transparency and Ratings - Instrumental Variable Results ....................................26

### Figures listed
- Figure 1. ROSC Index and GDP Per Capita .......................................................................................11
- Figure 2. ROSC and OBI Indices ........................................................................................................12
- Figure 3. ROSC Index and Political Competition in Advanced Economies .......................................17
- Figure 4. ROSC Index and Political Competition in Developing Economies ....................................17

### Appendix tables listed
- A.1. Numerical Conversion of Sovereign Credit Ratings ........................................................27
- A.2. Data Source and Variable Definitions .............................................................................28

### Implied analytic focus (from section and figure headings)
- Examination of properties and measurement issues of fiscal transparency indices.
- Correlation analysis across different transparency indices (including ROSC and OBI).
- Comparison of fiscal transparency with other proxies for transparency.
- Empirical analysis of the relationship between fiscal transparency and fiscal outcomes:
  - Primary balance and public debt across samples: Advanced Economies, Developing Countries, and Full Sample.
  - Separate treatment of ROSC Index and OBI Index in developing-country and full-sample analyses.
- Use of instrumental variable methods to assess:
  - Fiscal transparency and fiscal performance (Section 9).
  - Fiscal transparency and sovereign credit ratings (Section 10).
- Presentation of supporting figures linking ROSC to GDP per capita and political competition.
- Appendix resources for operationalization:
  - Numerical conversion scheme for sovereign credit ratings (A.1).
  - Data source and variable definitions (A.2).

*Source: _wp12156 - 1. Properties of Fiscal Transparency Indices ...........................................................................*

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

### _wp12156 - References

### I. Introduction — purpose and channels
- Research question: whether fiscal transparency affects the assessment of sovereign risk by market participants as reflected in sovereign credit ratings.
- Two channels examined:
  - Indirect channel: transparency → better fiscal outcomes → improved credit ratings.
  - Direct channel: transparency → reduced uncertainty/creditors’ risk premia → improved credit ratings.
- Estimation approach: a system of equations modeling
  - credit ratings as a function of fiscal indicators, fiscal transparency, and controls;
  - fiscal outcomes (debt and primary balance to GDP) as a function of transparency and controls;
  - allows separation of direct and indirect effects of fiscal transparency on credit ratings.

### Key empirical findings
- Overall: fiscal transparency has a positive and significant effect on credit ratings.
- Advanced economies:
  - A one standard deviation increase in transparency is associated with:
    - an increase in average primary balance of 2.4 percentage of GDP;
    - a lower debt to GDP ratio of about 17 percent of GDP.
  - One standard deviation improvement in fiscal transparency → indirect effect on credit ratings: increase of 0.45 notches.
  - Direct effect of one standard deviation improvement in transparency → about 0.24 notches (not statistically significant).
  - Instrumental-variable evidence: political competition used as an instrument suggests some robustness to endogeneity for the transparency → debt relationship in advanced economies.
- Developing economies:
  - Indirect effect of transparency on ratings: about 0.06-0.10 notches (depending on index and sample); this effect is not statistically significant.
  - Direct effect of one standard deviation improvement in fiscal transparency: about 0.50-0.90 notch increase; statistically significant.
  - Interpretation: weak link between transparency and fiscal outcomes and between fiscal fundamentals and ratings in developing economies; transparency’s marginal value may be high via uncertainty reduction.
  - Econometric caveat: potential omitted variables and weak instrument problems for developing-country samples prevent full causal interpretation.

- Additional empirical evidence and related literature cited:
  - Countries publishing Article IV reports, ROSCs, and participating in SDDS experienced average credit spread declines of 11 percent (Glennerster and Shin (2008) for 23 emerging markets).
  - Funds invest less in less transparent economies and flee them more during crises (Gelos and Wei (2005)).
  - Hameed (2005) and Gracia et. al. (2011) find disclosure of quantitative fiscal risks is associated with about one notch higher credit ratings.

### II. Measures of fiscal transparency used
- Two primary indices:
  - IMF fiscal transparency ROSCs (index constructed by Hameed (2005), range 0–100).
    - Three main components: Public Information; Budget Process (preparation, execution, reporting); Assurance of Integrity.
    - Sub-component scoring: 0, 33, 66, or 100; empirical analysis uses the average of the first two components as main ROSC index.
    - Authors note the Assurance of Integrity component had smaller and less significant effects on ratings and fiscal performance.
  - International Budget Partnership (IBP) Open Budget Index (OBI) — main index (range 0–100).
    - Based on 92 questions covering quality and comprehensiveness of 8 key budget documents and related processes.
    - Vintages used: 2006, 2008, 2010; analysis uses 2010 OBI where referenced.
- One standard deviation in the fiscal transparency index corresponds to 15 points out of 100 in both the ROSC and the OBI indices.
  - Example comparisons: difference in the ROSC index between Spain and Estonia, Norway and Italy, Mozambique and Uganda or Costa Rica and Paraguay is about 15 points.

### III. Properties and cross-indicator relationships
- Summary statistics reported (Table 1 excerpts; all numbers preserved exactly as in source):
  - ROSC Public Information
    - Mean: Advanced Economies 73.2, Developing Economies 39.8
    - Standard Deviation: Advanced Economies 16.2, Developing Economies 15.6
    - Min: Advanced Economies 37.9, Developing Economies 9.4
    - Max: Advanced Economies 95.1, Developing Economies 80.6
    - Number of Countries: Advanced Economies 22, Developing Economies 56
  - ROSC Budget Process
    - Mean: Advanced Economies 79.0, Developing Economies 46.6
    - Standard Deviation: Advanced Economies 16.2, Developing Economies 15.5
    - Min: Advanced Economies 46.5, Developing Economies 7.3
    - Max: Advanced Economies 100.0, Developing Economies 89.8
    - Number of Countries: Advanced Economies 22, Developing Economies 56
  - OBI Index Main
    - Mean: Advanced Economies 74.2, Developing Economies 43.1
    - Standard Deviation: Advanced Economies 12.3, Developing Economies 20.7
    - Min: Advanced Economies 57.6, Developing Economies 0.0
    - Max: Advanced Economies 90.1, Developing Economies 92.3
    - Number of Countries: Advanced Economies 12, Developing Economies 59

- Correlations and associations (preserved exactly as presented):
  - Correlation between ROSC and OBI indices: high and broadly consistent.
  - Reported correlations (Table 2 / Figure cues):
    - ROSC Public Information correlation with OBI Index Main: 0.51
    - ROSC Budget Process correlation with OBI Index Main: 0.71
  - Fiscal transparency and other proxies (Table 3 excerpts):
    - Average Stock Flow Adjustment (proxy for lack of transparency) and Corruption Perceptions correlations:
      - ROSC Public Information: Average Stock Flow Adjustment -0.40; Corruption Perceptions 0.62
      - ROSC Budget Process: Average Stock Flow Adjustment -0.49; Corruption Perceptions 0.65
      - OBI Index Main: Average Stock Flow Adjustment -0.28; Corruption Perceptions 0.64
- Observations:
  - Fiscal transparency is highly correlated with income per capita (Figure 1).
  - Standard deviations of transparency indices are comparable across advanced and developing economies, indicating significant within-group variation.
  - Coverage: fiscal ROSC index coverage for advanced economies is significantly higher than for developing economies.

### IV. Empirical strategy and organization
- Transparency measures used to:
  - explain fiscal indicators (debt/GDP and primary balance/GDP);
  - explain credit ratings both directly and indirectly (via fiscal indicators).
- Discussion of endogeneity and instruments:
  - Political competition and legal origin (common or civil law) explored as possible instruments.
  - Instrumental-variable approach applied for advanced economies; weak instruments problematic for developing-country sample.
- Paper organization:
  - Section II describes transparency measures and coverage;
  - Section III outlines empirical strategy;
  - Section IV presents results;
  - Section V concludes.

*Source: _wp12156 - References.*

### 2. The corruption perceptions index is the 2010 index calculated by Transparency International.

### _wp12156 - 2. The corruption perceptions index is the 2010 index calculated by Transparency International.

### Data and measures
- Corruption perceptions index: 2010 index calculated by Transparency International.
- Fiscal transparency measures: ROSC Reports on Fiscal Transparency (ROSC) and Open Budget Index (OBI), 2010.
- Average stock-flow adjustment data: from Weber (2012); captures average difference between fiscal balances and the change in debt not explained by factors such as exchange rate movements and financial transactions.
- Sovereign credit ratings: converted from S&P, Fitch and Moody’s to numerical indices and averaged (main sovereign credit rating variable).
- Other control variables: GDP per capita (PPP, 2000), average real growth (2000-2010), default history, average inflation (2000-2010), average political fractionalization (2000-2010), fuel exporter dummy, average polarization.
- Political/instrumental variables: political competition (frequency of political turnover) and electoral competition from Database of Political Institutions, World Bank.
- Note on data limitations: fiscal transparency measures do not vary sufficiently over time; panel estimation not feasible.

### Empirical approach and model
- Baseline cross-country regressions relate average credit ratings in 2010 to:
  - fiscal transparency,
  - average primary balance (2000-2010),
  - debt to GDP ratio (2010),
  - other controls (X, Y, Z).
- Regression system specifications presented as SURE equations for ratings (R), fiscal balance (fisbal), and debt (debt) with notation:
  - R_i = α0 + α1 fistrans_i + α2 debt_i + α3 fisbal_i + α4 X_i + ε_i
  - fisbal_i = χ0 + χ1 fistrans_i + χ2 debt_i + χ3 growth_i + χ4 frac_i + χ5 Y_i + ε_i
  - debt_i = η0 + η1 fistrans_i + η2 debt_i(2000) + η3 growth_i + η4 frac_i + η5 Z_i + ε_i
- Direct effect of a one unit increase in transparency defined as α1.
- Indirect effect defined as 2 4 3 4 α η α χ + (expressed in text as 243 4 ).
- Estimation method: Seemingly Unrelated Regression Estimation (SURE); single-equation OLS also checked with comparable estimates.
- Instruments explored for fiscal transparency: political competition and electoral competition; legal origin considered but correlation low in sample.

### Key empirical findings (summary of quantitative effects preserved as reported)
- Advanced economies (Table 4):
  - A one standard deviation increase in transparency (about 15 points) improves primary balance to GDP ratio by 2 percent of GDP.
  - A one standard deviation increase in transparency reduces debt to GDP ratio by about 16 percentage points.
  - Improvement in credit ratings associated with a one standard deviation increase in transparency is about 0.45 notches.
  - When fiscal transparency is included directly in the ratings equation, the direct effect is not statistically significant and smaller than the indirect effect.
  - Conclusion for advanced economies: transparency affects credit ratings mainly through indirect effects on fiscal performance rather than a direct credibility effect.
- Developing economies (Tables 5-6):
  - Effects of transparency on primary balance and debt are small and not statistically significant across specifications.
  - Including fiscal transparency directly in the ratings equation shows a significant and large direct effect: a one standard deviation increase in fiscal transparency is associated with an increase in ratings of about 0.5-0.9 notches depending on the transparency index.
  - Interpretation: for developing economies, the direct credibility effect (reducing uncertainty about past and future fiscal policies) may matter more than headline fiscal indicators; possibility of omitted variables remains.
- Full sample (Tables 7-8):
  - Estimates mainly driven by developing economies.
  - Estimated effect of transparency on fiscal variables is small; direct effect on ratings is significant and large.
  - Including an advanced country dummy reduces the size of the estimated direct effect of transparency; without that dummy the direct effect increases to levels comparable to Hameed (2005).

### Robustness and instrumental variables
- Endogeneity concerns: fiscal transparency may be endogenous with respect to fiscal performance; omitted variable bias possible.
- Instrumental variables considered: political competition (frequency of turnover) and electoral competition; legal origin dummy examined but not used due to low correlation.
- Advanced economies (Table 9):
  - Instrumented estimates: effect of fiscal transparency on average primary balance similar to SURE results but no longer significant.
  - Effect on debt to GDP ratio (2010) is significant at almost 5 percent level and similar in size to previous estimates.
  - Caveat: Cragg-Donald statistic needs to be stronger to conclusively rule out weak instrument problems.
- Developing economies (Table 10):
  - Instrumented direct effect of transparency on ratings is positive.
  - Cragg-Donald F-statistic and Shea’s partial R-squared are quite low, suggesting weak correlation between instruments and transparency (weak instruments).
- Additional note: for full sample and OBI index Cragg-Donald statistic suggested a serious weak instrument problem; those results not reported.

### Conclusions and implications
- Enhancing fiscal transparency is associated with better credit ratings for both advanced and developing economies.
- Two channels identified:
  - Indirect channel: transparency improves fiscal performance (primary balance, debt), which in turn improves ratings—more evident in advanced economies.
  - Direct credibility channel: transparency directly improves market perceptions and ratings—more evident in developing economies.
- Empirical caveats:
  - Measurement issues for transparency.
  - Endogeneity and omitted variable concerns.
  - Lack of time series variation in transparency measures.
- Suggested directions for future research and policy-oriented implications:
  - Research what determines fiscal transparency to develop alternative measures and stronger instruments.
  - Explore which transparency reforms are most effective for improving fiscal performance and market perceptions.

### Selected quantitative outputs excerpted from Tables (preserved exactly as reported)
- One standard deviation increase in transparency ≈ 15 points.
- Advanced sample: primary balance improvement = 2 percent of GDP per one standard deviation increase in transparency.
- Advanced sample: debt to GDP reduction ≈ 16 percentage points per one standard deviation increase in transparency.
- Advanced sample: ratings improvement ≈ 0.45 notches per one standard deviation increase in transparency.
- Developing sample: direct effect on ratings ≈ 0.5-0.9 notches per one standard deviation increase in transparency (depending on index).
- Instrument strength guideline noted: "Ideally, the F statistic should be higher than 10." (Staiger and Stock (1997) reference retained as in source.)

*Source: _wp12156 - 2. The corruption perceptions index is the 2010 index calculated by Transparency International.*

### 1. The effect is in terms of notches in credit ratings. For ex

### _wp12156 - 1. The effect is in terms of notches in credit ratings. For ex

### Main finding on fiscal transparency and credit ratings
- The effect is in terms of notches in credit ratings. For example, a 1 notch improvement corresponds to moving from AA+ to AAA.
- Total effect of 15 point (one standard deviation) increase in fiscal transparency index on ratings is reported in multiple tables (OBI and ROSC indices, developing and full samples).

### Key regression coefficients (selected, preserving exact values and significance)
- Table 6 (Fiscal Transparency (OBI Index), Primary Balance and Public Debt - Developing Country Sample)
  - GDP per capita (2000): 2.53  *** ; 7.36  *** ; 2.11  *** ; 7.35  ***
  - Default history: -2.67  ** ; -2.68  ***
  - Gross public debt (2000): -0.04  *** ; 0.52  ***
  - Gross public debt (2010): -0.04  * ; -0.02
  - Inflation (2000-2010): -0.14  *** ; -0.14  ***
  - Real growth (2000-2010): 0.35  ** ; -0.15 ; 0.94
  - OBI Index: 0.00 ; -0.14 ; 0.03  **
  - Fuel Exporter dummy: 7.14  *** ; -34.12  *** ; 7.13  *** ; -34.14  ***
  - R-squared: 0.64 ; 0.5 ; 0.53 ; 0.68 ; 0.5 ; 0.53
  - N: 37 37 37 37 37 37
  - Direct Effect: 0.00 ; 0.52  **
  - Indirect Effect: 0.09 ; 0.06
    - Due to debt: 0.09 ; 0.05
    - Due to primary balance: 0.00 ; 0.01

- Table 7 (Fiscal Transparency (ROSC Index), Primary Balance and Public Debt - Full Sample)
  - GDP per capita (2000): 2.37  *** ; 13.11  ** ; 2.06  *** ; 12.99  **
  - Gross public debt (2000): -0.04  *** ; 0.81  ***
  - Gross public debt (2010): -0.02  * ; -0.01
  - Inflation (2000-2010): -0.15 ; -0.15  *
  - Real growth (2000-2010): 0.16 ; -0.22 ; -0.74 ; 0.22 ; -0.22 ; -0.71
  - Advanced country dummy: 4.60  *** ; -0.94 ; 11.99 ; 3.21  *** ; -0.92 ; 11.44
  - ROSC Index (Average): 0.05  * ; -0.32 ; 0.06  *** ; 0.05  * ; -0.29
  - Fuel Exporter dummy: 9.49  *** ; -39.91  *** ; 9.54  *** ; -40.37  ***
  - R-squared: 0.83 ; 0.43 ; 0.69 ; 0.86 ; 0.43 ; 0.69
  - N: 56 56 56 56 56 56
  - Direct Effect: 0.00 ; 0.94  ***
  - Indirect Effect: 0.18 ; 0.13

- Table 8 (Fiscal Transparency (OBI Index), Primary Balance and Public Debt - Full Sample)
  - GDP per capita (2000): 2.71  *** ; 8.03  ** ; 2.41  *** ; 8.07  **
  - Default history: -2.74  *** ; -2.62  ***
  - Gross public debt (2000): -0.04  ** ; 0.55  ***
  - Gross public debt (2010): -0.03  * ; -0.02
  - Avg. primary balance (2000-2010): 0.10 ; 0.16  *
  - Inflation (2000-2010): -0.12  *** ; -0.12  ***
  - Real growth (2000-2010): 0.36  ** ; -0.11 ; 0.13 ; 0.37  ** ; -0.11 ; 0.11
  - OBI Index: 0.01 ; -0.12 ; 0.03  *
  - Fuel Exporter dummy: 7.73  *** ; -36.31  *** ; 7.74  *** ; -36.33  ***
  - R-squared: 0.84 ; 0.34 ; 0.61 ; 0.85 ; 0.34 ; 0.61
  - N: 49 49 49 49 49 49
  - Direct Effect: 0.00 ; 0.43  *
  - Indirect Effect: 0.06 ; 0.06
    - Due to debt: 0.05 ; 0.04
    - Due to primary balance: 0.01 ; 0.02

### Instrumental variable results (selected)
- Table 9 (Fiscal Transparency and Fiscal Performance - Instrumental Variable Results)
  - Debt (2000): -0.020 .88  ***
  - Average real growth: -0.22 -12.33  ***
  - Average Polarization: 2.39  ** -11.55  **
  - ROSC Index (Average): 0.14 -0.94  *
  - Cragg - Donald statistic (weak instrument): 5.09 5.09
  - Shea's Partial R-squared: 0.24 0.24
  - R-squared: 0.47 0.92
  - Number of Observations: 21 21

- Table 10 (Fiscal Transparency and Ratings - Instrumental Variable Results)
  - GDP per capita: 1.87  ** ; 1.75  ***
  - Default history: -0.93 ; -0.75
  - Gross public debt (2000): -0.01 ; 0.00
  - Primary balance: 0.13 ; 0.20
  - Inflation: -0.16  ** ; -0.15  *
  - Real growth: 0.26  * ; 0.28  *
  - ROSC Index (average): 0.09 ; 0.12  *
  - Cragg - Donald statistic (weak instrument): 2.60 2.93
  - Hansen J statistic (p-value): 0.56 ...
  - Shea's Partial R-squared: 0.14 0.08
  - R-squared: 0.56 0.50
  - N: 42 42

Notes included with IV tables:
- "The robust t-statistics are reported in parantheses (* denotes significance at 10%, ** significance at 5% and *** significance at 1%). All specifications include a constant, which is not reported. All estimates are obtained by using two-stage least squares."

### Appendix: Rating numeric conversion (excerpt)
- AAA: 23 ; Aaa: 20 ; AAA: 22
- AA+: 22 ; Aa1: 19 ; AA+: 21
- AA: 21 ; Aa2: 18 ; AA: 20
- AA-: 20 ; Aa3: 17 ; AA-: 19
- A+: 19 ; A1: 16 ; A+: 18
- A: 18 ; A2: 15 ; A: 17
- A-: 17 ; A3: 14 ; A-: 16
- BBB+: 16 ; Baa1: 13 ; BBB+: 15
- BBB: 15 ; Baa2: 12 ; BBB: 14
- BBB-: 14 ; Baa3: 11 ; BBB-: 13
- BB+: 13 ; Ba1: 10 ; BB+: 12
- BB: 12 ; Ba2: 9 ; BB: 11
- BB-: 11 ; Ba3: 8 ; BB-: 10
- B+: 10 ; B1: 7 ; B+: 9
- B: 9 ; B2: 6 ; B: 8
- B-: 8 ; B3: 5 ; B-: 7
- CCC+: 7 ; Caa1: 4 ; CCC+: 6
- CCC: 6 ; Caa2: 3 ; CCC: 5
- CCC-: 5 ; Caa3: 2 ; CCC-: 4
- CC: 4 ; Ca: 1 ; CC: 3
- C: 3 ; C: 0 ; C: 2
- DDD: 2 ; SD: 1 ; DD: 1 ; D: 0 ; RD: 0

(Source table headed: "Numerical Conversion of Sovereign Credit Ratings" — Fitch, Moody's and S&P.)

### Data sources and variable definitions (selected, exact text excerpts)
- Average primary balance: "Average ratio of primary fiscal balance to GDP (in percent)" — WEO Database
- Gross public debt (2000): "Gross debt to GDP in 2000 (in percent)" — WEO Database
- Average real growth: "Average real growth during 2001-2010." — WEO Database
- Fuel exporter dummy: "Dummy variable that equals one if the country is a fuel exporter according to the WEO classification." — WEO Database
- Electoral Competition: "Average of the executive index of electoral competitiveness during 2000-2010. The variable taxes values between 1 and 7, where higher values correspond to higher electoral competitiveness. The executive index of electoral competitition 'eiec' variable, Database of Political Institutions, World Bank."
- Average Polarization: "Maximum polarization between the executive party and the four principle parties of the legislature, average during 2000-2010. The 'polariz' variable, Database of Political Institutions, World Bank."
- Political Competition: "Author's calculation based on the frequency of political turnover since 1975. Using the variable that captures the name of the largest government party 'gov1me', Database of Political Institutions, World Bank."
- GDP per capita, 2000: "GDP per capita (PPP), in US Dollars" — WEO Database
- Default history: "Author's calculation based on whether the country has been classified as in defauly by S&P over the past 10 years." — S&P
- Inflation: "Average level of inflation (in percent) during 2001-2010." — IFS Database

*Source: Authors' calculations and estimations.*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2012/_wp12156.pdf_