## _wp1223

## Source details

**Canonical URL:** [_wp1223](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2012/_wp1223.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2012/_wp1223.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2012/_wp1223.pdf.json)

---

### I. Introduction — scope and aims
- Study limited to sovereign ratings (of emerging markets and advanced economies) and covers the period January 2005-June 2010.
- Goal: assess whether Credit Rating Agencies (CRAs) are influential or merely lag market information, and test three hypotheses regarding CRA services: information, certification, and monitoring.
- Core conclusions:
  - CRAs’ opinions are influential and favor the information and certification role.
  - Ratings have information value beyond publicly available market information.
  - Ratings represent a rank ordering of credit risk but lack a disclosed one-to-one mapping to quantitative measures (e.g., default probabilities).
  - Ratings perform reasonably well at rank ordering sovereign default risk but accuracy deteriorates with longer evaluation horizons and during crises.

### II. Key empirical findings
- Event-study and causality evidence:
  - Negative credit warnings (“reviews,” “watches,” and “outlooks”) have a significant impact on CDS spreads.
  - A Granger-causality test finds that negative credit warnings Granger-cause changes in CDS spreads.
  - Upgrades and downgrades generally do not have a significant impact on CDS spreads, except for upgrades and downgrades in and out of investment grade categories, which are statistically significant.
  - The impact of downgrades preceded by an outlook review in the same direction is not statistically significant (no support found for the main testable implication of the monitoring-services theory).
- Market vs. rating discrimination:
  - The market appears to discriminate more than rating agencies among different kinds of issuers—particularly at lower rating grades and during crisis periods, indicating ratings may need more granularity.
- Rating accuracy and transitions:
  - Accuracy ratios (AR) indicate agencies are more successful at sorting potential defaulters among sovereign issuers (average ARs in the 80 to 90 percent range) than among corporate and structured finance issuers (average ARs in the 63 to 87 percent range).
  - Long-term credit transition matrices: higher ratings are generally more stable than lower ones and tend to remain unchanged, except during the global crisis period when heavier downgrade activity was seen among higher-rated sovereigns than among lower-rated ones.
  - Evidence of significant rating failure (defined here as three or more rating changes in one year) occurred during the recent global financial crisis, although less than in the Asian crisis.

### III. Theories tested and implications
- Information theory:
  - Prediction: significant market reaction to rating actions (changes or credit warnings).
  - Supported by significant CDS spread reactions to negative credit warnings and Granger-causality results.
- Certification theory:
  - Prediction: significant market reaction to upgrades/downgrades, particularly those crossing investment-grade thresholds.
  - Supported by statistically significant impacts for rating actions involving movement in and out of investment grade categories.
- Monitoring-services theory:
  - Prediction: significant market reaction to rating confirmations/changes conditional on prior credit watch.
  - Not supported: downgrades preceded by an outlook review in the same direction do not show a statistically significant additional impact.

### IV. Data, sample, and methodology (exact counts)
- Main analysis sample:
  - 72 sovereigns for which both Moody’s ratings and relevant CDS data from Markit are available from January 2005 to July 2010.
  - 101,176 daily spread observations.
  - 194 credit rating events:
    - 26 downgrades
    - 57 upgrades
    - 71 positive outlooks/reviews for upgrade
    - 40 negative outlooks/reviews for downgrade
- Methods employed:
  - Event studies of rating actions on CDS spreads (event window: 41 days: 20 days before to 20 days after).
  - Granger-causality tests (panel Granger-causality estimated using GMM with fixed effects).
  - Review of rating actions by S&P and Fitch in addition to Moody’s.
  - Validation tests focused on ratings’ ability to differentiate potential defaulters and non-defaulters (rank ordering), not on producing cardinal default-probability measures.

### V. Stylized facts: CDS spreads and Moody’s ratings
- Methodology:
  - For each year and each country, the average CDS spread and average credit rating is calculated (assuming a linear conversion, with 21 = AAA and 1 = D).
  - This provides a total of 71 observations each year.
  - An exponential trend line is imposed.
- Key patterns:
  - Close association between spreads and ratings with lower credit categories reflected in higher spreads.
  - The relationship between spreads and credit ratings is nonlinear; PDs increase more steeply at the lowest credit grades.
  - Since 2007 the dispersion of spreads at the lowest rating grades has widened, implying increased market discrimination among worst credits.
- Yearly average spreads (across all countries), in basis points:
  - 2005: 100 basis points
  - 2006: 63 basis points
  - 2007: 60 basis points
  - 2008: 212 basis points
  - 2009: 325 basis points
  - H1 2010: 200 basis points (the 2010 data does not include the downgrade of Greece)

### VI. Univariate regressions of log spreads on rating categories (Table 1)
- Reported estimates (preserving numeric values exactly):
  - 2005
    - Intercept: 8.34  Std. Error: 0.18  t-Statistic: 45.72  R-squared: 0.93
    - Coefficient on rating: -0.34  Std. Error: 0.01  t-Statistic: -29.17
  - 2006
    - Intercept: 7.83  Std. Error: 0.18  t-Statistic: 42.64  R-squared: 0.92
    - Coefficient on rating: -0.32  Std. Error: 0.01  t-Statistic: -27.56
  - 2007
    - Intercept: 7.67  Std. Error: 0.18  t-Statistic: 41.49  R-squared: 0.91
    - Coefficient on rating: -0.31  Std. Error: 0.01  t-Statistic: -26.12
  - 2008
    - Intercept: 8.06  Std. Error: 0.21  t-Statistic: 37.78  R-squared: 0.79
    - Coefficient on rating: -0.22  Std. Error: 0.01  t-Statistic: -16.28
  - 2009
    - Intercept: 7.90  Std. Error: 0.20  t-Statistic: 39.53  R-squared: 0.73
    - Coefficient on rating: -0.17  Std. Error: 0.01  t-Statistic: -13.74
  - 2010
    - Intercept: 7.10  Std. Error: 0.20  t-Statistic: 35.63  R-squared: 0.65
    - Coefficient on rating: -0.14  Std. Error: 0.01  t-Statistic: -11.43
- Interpretation:
  - The significance and explanatory power (R-squared) of ratings decline over time, indicating increasing impact of other determinants of spreads as the crisis developed.

### VII. Event study baseline results (Moody’s)
- Design details:
  - Event window: 41 days (20 days before to 20 days after).
  - Changes in CDS spreads measured relative to t = 0.
  - Sovereigns split into advanced economies and emerging markets.
  - One-sided t-test for directional hypotheses (downgrades → increase; upgrades → decrease).
- Baseline findings:
  - Most incremental information value transmitted through negative credit warnings (“outlooks,” “reviews,” and “watches”) rather than actual rating changes.
  - Only 45 percent (advanced economies) and 51 percent (emerging economies) of rating events are associated with spread changes in the expected direction.
  - For outlook changes, the proportion rises to 67 percent (advanced economies) and 63 percent (emerging economies).
  - Negative outlook announcements are followed by statistically significant spread widening:
    - 100 basis points for advanced economies
    - 160 basis points for emerging markets
  - Rating changes generally have insignificant market impact, except when crossing the investment-grade boundary:
    - Downgrades out of investment grade (Baa3 or better to Ba1 or lower) lead to statistically significant CDS spread widening of 37 basis points (based on a few events).
    - Table 2 reports downgrades through the investment grade threshold lead to statistically significant CDS spread widening of 38 basis points (Table 2 summary).
  - Very little market reaction to upgrades.
  - No evidence supporting the monitoring theory (downgrades preceded by warnings do not show the monitoring effect).

### VIII. Robustness checks and causality
- Robustness checks performed include:
  - Varying event window length (10 days before/after; 45 days before/after).
  - Controlling for global volatility (S&P 500 VIX), liquidity risk premium (U.S. LIBOR), and U.S. daily business conditions (Aruoba-Diebold-Scotti index).
  - Using natural logarithms of spreads.
  - Splitting sample into pre-crisis and during-crisis periods.
  - Disaggregating announcements for investment/noninvestment grade sovereigns.
- Granger-causality results:
  - Information about preceding negative outlooks helps predict changes in CDS spreads of emerging economies even after accounting for past CDS spreads.
  - Instruments used include lagged S&P 500 VIX, lagged U.S. LIBOR, lagged Aruoba-Diebold-Scotti index, and CDS spreads lagged two periods.

### IX. Multi-agency comparison (Moody’s, S&P, Fitch)
- Differences in preceding credit reviews/watches:
  - Moody’s: 27 percent of rating changes are preceded by corresponding credit watches.
  - S&P: 14 percent.
  - Fitch: 8 percent.
- Timing dynamics and market impact:
  - Moody’s tends to lag behind Fitch’s and S&P’s negative rating actions; Moody’s and Fitch tend to follow S&P negative rating actions more often than S&P follows the others.
  - When Moody’s leads the cycle, it is more likely to issue credit watches rather than immediate downgrades.
  - Consequently, Moody’s downgrades have no significant impact on financial markets but Fitch’s and S&P’s downgrades do.
- Frequency counts for first/last negative action in adjustment cycle (Table 3 entries):
  - All countries: First—Fitch 28  Moody's 13  S&P 59 ; Last—Fitch 36  Moody's 36  S&P 28
  - Advanced economies: First—Fitch 22  Moody's 4  S&P 74 ; Last—Fitch 30  Moody's 52  S&P 17
  - Emerging economies: First—Fitch 32  Moody's 17  S&P 51 ; Last—Fitch 39  Moody's 27  S&P 34

### X. What ratings measure; accuracy and stability of ratings
- Agencies’ intended meanings (summaries preserved):
  - S&P: credit ratings provide relative rankings among issuers and obligations of overall creditworthiness; ratings are not measures of absolute default probability.
  - Fitch: credit ratings express risk in relative rank order; ratings are ordinal measures of credit risk and are not predictive of a specific frequency of default or loss.
  - Moody’s: credit ratings are opinions of the relative credit risk and address the possibility that a financial obligation will not be honored as promised; they reflect both the likelihood of default and any financial loss suffered in the event of default.
- Common elements:
  - Ratings represent a rank ordering of credit risk.
  - Only credit risk is included (ratings do not measure liquidity, market, or interest rate risks).
  - Ratings convey information about potential changes in credit risk (downgrading risk or default volatility/stability).
  - Ratings are medium-term assessments not intended to fluctuate with cyclical macroeconomic changes.
- Relationship to quantitative measures:
  - No one-to-one mapping between quantitative measures (PD, LGD, EAD) and ratings.
  - Internal bank ratings under the Basel IRB approach require vetting of data inputs to statistical default or loss prediction models (BCBS, 2006).
- Stability and transition findings:
  - Higher rating categories are associated with lower downgrading probability; diagonals of transition matrices closer to 100 for highest categories.
  - One-year transition matrices are typically more stable than 5-year matrices since default probability increases with risk horizon.
  - Empirical validation: all 14 sovereign defaulters involving 12 countries between 1975 and 2009 had (S&P) ratings of BB- or below one year prior to default.
- CAP curves and accuracy ratios (S&P sovereigns):
  - AR = 92 percent over the one-year horizon.
  - AR = 82 percent over three years.
  - AR = 80 percent over five years.
  - AR = 84 percent over 10 years.
- Corporate rating ARs (S&P):
  - One-year horizon: 77 percent for financials and 81 percent for nonfinancials.
  - Five-year horizon: 63 percent for financials and 71 percent for nonfinancials.
- Interpretation:
  - CRAs are better at predicting defaults over short horizons than long ones.

### XI. Ratings philosophy: TTC versus PIT; through-crisis criteria
- Ratings are medium-term assessments and should not change due to cyclical components.
- Traditional approach: rating through the cycle (TTC) — ratings based on ability to survive a trough; current position in the cycle should not by itself prompt downgrades.
- Point-in-time (PIT) ratings focus on current conditions.
- Through-crisis criteria and new stability approaches:
  - CRAs have developed methodologies intended to extend TTC criteria to “through crisis criteria.”
  - Example: Standard & Poor’s credit stability criteria uses hypothetical stress scenarios as benchmarks calibrated across sectors and over time; each scenario broadly corresponds to one of the rating categories.
  - The stability criterion allows for hypothetical scenarios affecting trend components, making ratings measures of risk conditional on extreme scenarios.

### XII. Examination of rating stability
- Validation focus: discriminatory power (sorting defaulters vs non-defaulters) and stability of ratings.
- Empirical stability findings (S&P data, 1975–2009):
  - Higher ratings are more stable than lower ones; sovereign ratings more stable than corporate ones.
  - Example: 82.1 percent of A- rated sovereigns were still A- rated at the end of each year, versus 77.5 percent of corporates.
  - Example: 1.8 percent of A-rated sovereigns were downgraded three or more notches, versus 6.9 percent of corporates.
- Crisis-period instability:
  - Only 68 percent of ratings remained unchanged during the Asian crisis.
  - Only 63 percent of ratings remained unchanged during the current crisis (so far).
  - The proportion of sovereigns that started each period as investment grade but finished as high yield fell from 2.9 percent to 1.9 percent.
- Average rating change (Moody’s):
  - Average rating change fell from a downgrade by less 0.4 notch during the Asian crisis to a 0.07 notch during the current crisis (difference not statistically significant).
- Large downgrades (three or more notches) concentrated in investment-grade categories despite intended greater stability.
- Smoothing rules tradeoff:
  - Smoothing rules can reduce frequency of changes but can undermine stability by creating abrupt downgrades if a prescribed downgrade proves persistent (cliff effects).

### XIII. Policy recommendations
- Remove the excessive reliance of regulations on ratings to reduce certification-driven liquidity impacts and potential cliff effects; this will not affect the information value of ratings, which works mostly through outlook reviews.
- If ratings continue to play a significant role in regulations:
  - CRAs should be more transparent about the quantitative measures they calibrate in the rating process (PDs, LGD, and stability assumptions), how these measures are mapped into ratings, and whether final ratings can be used to infer the parameters used to obtain these measures.
  - This transparency is particularly relevant for external ratings used by banks employing the standardized approach in Basel II, since internal ratings systems are subject to rigorous back testing.
- Ratings should embed the notion that risk is a forward-looking dimension conditional on the macroeconomic scenario; ratings should be better tied to macroeconomic conditions, including resilience to stress scenarios.
- Recommendation for CRAs: calibrate ratings to target credit risk metrics and publish validation tests against such standards to make the rating message more transparent.

*Source: _wp1223 - References .............................................................................................................*

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

### _wp1223 - References .............................................................................................................

### I. Introduction — scope and aims
- Study limited to sovereign ratings (of emerging markets and advanced economies) and covers the period January 2005-June 2010.
- Goal: assess whether Credit Rating Agencies (CRAs) are influential or merely lag market information, and test three hypotheses regarding CRA services: information, certification, and monitoring.
- Core conclusion statements (from the analysis):
  - CRAs’ opinions are influential and favor the information and certification role.
  - Ratings have information value beyond publicly available market information.
  - Ratings represent a rank ordering of credit risk but lack a disclosed one-to-one mapping to quantitative measures (e.g., default probabilities).
  - Ratings perform reasonably well at rank ordering sovereign default risk but accuracy deteriorates with longer evaluation horizons and during crises.

### II. Key empirical findings
- Event-study and causality evidence:
  - Negative credit warnings (i.e., “reviews,” “watches,” and “outlooks”) have a significant impact on CDS spreads.
  - A Granger-causality test finds that negative credit warnings Granger-cause changes in CDS spreads.
  - Upgrades and downgrades generally do not have a significant impact on CDS spreads, except for upgrades and downgrades in and out of investment grade categories, which are statistically significant.
  - The impact of downgrades preceded by an outlook review in the same direction is not statistically significant (no support found for the main testable implication of the monitoring-services theory).
- Market vs. rating discrimination:
  - The market appears to discriminate more than rating agencies among different kinds of issuers—particularly at lower rating grades and during crisis periods, indicating ratings may need more granularity.
- Rating accuracy and transitions:
  - Accuracy ratios (AR) indicate agencies are more successful at sorting potential defaulters among sovereign issuers (average ARs in the 80 to 90 percent range) than among corporate and structured finance issuers (average ARs in the 63 to 87 percent range).
  - Long-term credit transition matrices: higher ratings are generally more stable than lower ones and tend to remain unchanged, except during the global crisis period when heavier downgrade activity was seen among higher-rated sovereigns than among lower-rated ones.
  - Evidence of significant rating failure (defined here as three or more rating changes in one year) occurred during the recent global financial crisis, although less than in the Asian crisis.

### III. Theories tested and testable implications
- Information theory:
  - Predicts significant market reaction to rating actions (changes or credit warnings) because CRAs provide information reducing asymmetric information.
  - Supported by findings of significant CDS spread reactions to negative credit warnings and Granger-causality results.
- Certification theory:
  - Predicts significant market reaction to upgrades/downgrades, particularly those crossing investment-grade thresholds, due to certification effects (e.g., regulatory or mandate-driven asset reallocation).
  - Supported by statistically significant impacts for rating actions involving movement in and out of investment grade categories.
- Monitoring-services theory:
  - Predicts significant market reaction to rating confirmations/changes conditional on prior credit watch (CRAs acting as monitors and coordinating beliefs).
  - Not supported: the impact of downgrades preceded by an outlook review in the same direction is not statistically significant.

### IV. Data, sample, and methodology (exact counts)
- Main analysis sample:
  - 72 sovereigns for which both Moody’s ratings and relevant CDS data from Markit are available from January 2005 to July 2010.
  - 101,176 daily spread observations.
  - 194 credit rating events:
    - 26 downgrades
    - 57 upgrades
    - 71 positive outlooks/reviews for upgrade
    - 40 negative outlooks/reviews for downgrade
- Methods employed:
  - Event studies of rating actions on CDS spreads.
  - Granger-causality tests.
  - Review of rating actions by S&P and Fitch in addition to Moody’s.
  - Validation tests focused on ratings’ ability to differentiate potential defaulters and non-defaulters (rank ordering), not on producing cardinal default-probability measures.

### V. Implications for market participants and supervisors (derived from findings)
- Ratings convey incremental information to markets beyond publicly available data—market participants should treat CRA warnings (reviews/watches/outlooks) as informative signals that can affect CDS spreads.
- Certification effects matter for actions crossing investment-grade thresholds—regulatory and mandate-driven constraints tied to ratings can amplify market impact when issuers move in or out of investment grade.
- Because CRAs do not disclose the quantitative parameters mapping to each rating grade, external validation can only assess ordinal (rank) performance, not cardinal default probabilities; supervisors and banks using internal ratings frameworks (Basel II IRB) should maintain higher validation standards than external ratings.
- The lower granularity of rating scales (relative to market discrimination, especially at lower grades and during crises) suggests potential benefits from finer gradations in sovereign rating scales if the goal is improved market signaling.

*Source: _wp1223 - References .............................................................................................................*

### 16.      Stylized facts. Figure 1 presents some stylized facts about the relationship between CDS

### _wp1223 - 16.      Stylized facts. Figure 1 presents some stylized facts about the relationship between CDS

### Stylized facts: CDS spreads and Moody’s ratings
- Methodology:
  - For each year and each country, the average CDS spread and average credit rating is calculated (assuming a linear conversion, with 21 = AAA and 1 = D).
  - This provides a total of 71 observations each year.
  - An exponential trend line is imposed.
- Key empirical patterns:
  - There is a close association between spreads and ratings with lower credit categories being reflected in higher spreads.
  - The relationship between spreads and credit ratings is clearly nonlinear; PDs increase more steeply at the lowest credit grades.
- Yearly average spreads (across all countries), in basis points:
  - 2005: 100 basis points
  - 2006: 63 basis points
  - 2007: 60 basis points
  - 2008: 212 basis points
  - 2009: 325 basis points
  - H1 2010: 200 basis points (the 2010 data does not include the downgrade of Greece)
- Additional observation:
  - Since 2007 the dispersion of spreads at the lowest rating grades has widened, implying increased market discrimination among worst credits.

### Univariate regressions of log spreads on rating categories (Table 1)
- Table layout context: regressions of log spreads on rating categories for each year; R-squared shown as proportion of average CDS spreads’ movement attributable to credit ratings.
- Reported estimates (preserving original numeric values exactly):
  - 2005
    - Intercept: 8.34  Std. Error: 0.18  t-Statistic: 45.72  R-squared: 0.93
    - Coefficient on rating: -0.34  Std. Error: 0.01  t-Statistic: -29.17
  - 2006
    - Intercept: 7.83  Std. Error: 0.18  t-Statistic: 42.64  R-squared: 0.92
    - Coefficient on rating: -0.32  Std. Error: 0.01  t-Statistic: -27.56
  - 2007
    - Intercept: 7.67  Std. Error: 0.18  t-Statistic: 41.49  R-squared: 0.91
    - Coefficient on rating: -0.31  Std. Error: 0.01  t-Statistic: -26.12
  - 2008
    - Intercept: 8.06  Std. Error: 0.21  t-Statistic: 37.78  R-squared: 0.79
    - Coefficient on rating: -0.22  Std. Error: 0.01  t-Statistic: -16.28
  - 2009
    - Intercept: 7.90  Std. Error: 0.20  t-Statistic: 39.53  R-squared: 0.73
    - Coefficient on rating: -0.17  Std. Error: 0.01  t-Statistic: -13.74
  - 2010
    - Intercept: 7.10  Std. Error: 0.20  t-Statistic: 35.63  R-squared: 0.65
    - Coefficient on rating: -0.14  Std. Error: 0.01  t-Statistic: -11.43
- Interpretation:
  - The significance and explanatory power (R-squared) of ratings decline over time, indicating increasing impact of other determinants of spreads as the crisis developed.

### Event study of rating changes and outlooks (methods and baseline results)
- Event study design:
  - Event window: 41 days (20 days before the event to 20 days after the event).
  - Changes in CDS spreads are calculated against the level at the beginning of the event window (t = 0).
  - Sovereigns divided into advanced economies and emerging markets; within each group the relationship between credit ratings and CDS spreads is assumed linear for tractability.
  - Significance tested with a standard one-sided t-test (hypothesis: downgrades → increase in spreads; upgrades → decrease).
- Baseline Moody’s results (Table 2 / Figure 2 highlights):
  - Most incremental information value is transmitted through negative credit warnings (“outlooks,” “reviews,” and “watches”) rather than actual rating changes.
  - Only 45 percent (advanced economies) and 51 percent (emerging economies) of rating events are associated with spread changes in the expected direction.
  - For outlook changes, the proportion rises to 67 percent (advanced economies) and 63 percent (emerging economies).
  - Negative outlook announcements are followed by statistically significant spread widening:
    - 100 basis points for advanced economies
    - 160 basis points for emerging markets
  - Rating changes generally have insignificant market impact, except when crossing the investment-grade boundary:
    - Downgrades out of investment grade (Baa3 or better to Ba1 or lower) lead to statistically significant CDS spread widening of 37 basis points (based on a few events).
    - Table 2 reports downgrades through the investment grade threshold lead to statistically significant CDS spread widening of 38 basis points (Table 2 summary).
  - There is very little market reaction to upgrades.
  - No evidence supporting the monitoring theory (i.e., downgrades preceded by warnings do not show the monitoring effect).

### Robustness checks and causality
- Robustness checks performed include:
  - Varying event window length (10 days before/after; 45 days before/after).
  - Controlling for global volatility (S&P 500 VIX), liquidity risk premium (U.S. LIBOR), and U.S. daily business conditions (Aruoba-Diebold-Scotti index).
  - Using natural logarithms of spreads.
  - Splitting sample into pre-crisis and during-crisis periods.
  - Disaggregating announcements for investment/noninvestment grade sovereigns.
- Granger-causality (panel Granger-causality estimated using GMM with fixed effects) results:
  - Information about preceding negative outlooks helps predict changes in CDS spreads of emerging economies even after accounting for past CDS spreads.
  - Instruments used include lagged S&P 500 VIX, lagged U.S. LIBOR, lagged Aruoba-Diebold-Scotti index, and CDS spreads lagged two periods.

### Multi-agency comparison (Moody’s, S&P, Fitch)
- Two notable differences when extending analysis to S&P and Fitch:
  - Preceding credit reviews/watches:
    - Moody’s: 27 percent of rating changes are preceded by corresponding credit watches.
    - S&P: 14 percent.
    - Fitch: 8 percent.
  - Timing dynamics:
    - Moody’s tends to lag behind Fitch’s and S&P’s negative rating actions; Moody’s and Fitch tend to follow S&P negative rating actions more often than S&P follows the others.
    - When Moody’s leads the cycle, it is more likely to issue credit watches rather than immediate downgrades.
    - Consequently, Moody’s downgrades have no significant impact on financial markets but Fitch’s and S&P’s downgrades do.
- Table 3 (frequency counts for first/last negative action in adjustment cycle) — preserved numeric entries:
  - All countries: First—Fitch 28  Moody's 13  S&P 59 ; Last—Fitch 36  Moody's 36  S&P 28
  - Advanced economies: First—Fitch 22  Moody's 4  S&P 74 ; Last—Fitch 30  Moody's 52  S&P 17
  - Emerging economies: First—Fitch 32  Moody's 17  S&P 51 ; Last—Fitch 39  Moody's 27  S&P 34

### What ratings measure; accuracy and stability of ratings
- Agencies’ intended meanings (quoted summaries):
  - S&P: credit ratings provide relative rankings among issuers and obligations of overall creditworthiness; ratings are not measures of absolute default probability.
  - Fitch: credit ratings express risk in relative rank order; ratings are ordinal measures of credit risk and are not predictive of a specific frequency of default or loss.
  - Moody’s: credit ratings are opinions of the relative credit risk and address the possibility that a financial obligation will not be honored as promised; they reflect both the likelihood of default and any financial loss suffered in the event of default.
- Common elements distilled:
  - Ratings represent a rank ordering of credit risk.
  - Only credit risk is included (ratings do not measure liquidity, market, or interest rate risks).
  - Ratings convey information about potential changes in credit risk (downgrading risk or default volatility/stability).
  - Ratings are medium-term assessments not intended to fluctuate with cyclical macroeconomic changes.
- Relationship to quantitative measures:
  - Ratings rank ordering takes into account default probabilities, expected losses, and possible rating actions and warnings, but there is no one-to-one mapping between quantitative measures (PD, LGD, EAD) and ratings.
  - Internal bank ratings under the Basel IRB approach require processes for vetting data inputs to statistical default or loss prediction models (BCBS, 2006).
- Stability interpretation:
  - Higher rating categories are associated with lower downgrading probability.
  - In transition-matrix terms, the diagonal should be closer to 100 for the highest categories (higher categories more likely to remain the same).
  - One-year transition matrices are typically more stable than 5-year (or longer) matrices since default probability increases with risk horizon.

*Source: IMF staff estimates.*

### 28.      Ratings are medium-term assessments and should not change due to the impact of

### _wp1223 - 28.      Ratings are medium-term assessments and should not change due to the impact of

### Ratings philosophy: TTC versus PIT
- Ratings are medium-term assessments and should not change due to the impact of cyclical components.
- Traditionally, CRAs have tried to minimize rating volatility by rating through the cycle (TTC)—as opposed to allocating point in time (PIT) ratings.
- Ratings are based on the ability of an issuer to survive a trough; the rating is allocated by assessing credit risk under a stress scenario characterized by a shock to the transitory component.
- Once allocated, ratings are changed only in response to changes in fundamental secular trends or unanticipated policy responses.
- Under TTC, a recession or tightening of global liquidity should not, in itself, be an occasion for a downgrade; vulnerability to cycles affects the rating decision, whereas the current position in the cycle does not.
- PIT rating systems tend to focus on the current conditions of an obligor.

### Through-crisis criteria and new stability approaches
- CRAs have developed methodologies intended to extend TTC criteria to “through crisis criteria.”
- Example: Standard & Poor’s new credit stability criteria uses hypothetical stress scenarios as benchmarks calibrated across sectors and over time; each scenario broadly corresponds to one of the rating categories.
- The scenario for a particular category reflects the level of stress that issuers or obligations rated in that category should be able to withstand without defaulting.
- Unlike TTC, the new stability criterion allows for hypothetical scenarios affecting trend components, making ratings measures of risk conditional on the realization of extreme scenarios.

### Examination of rating accuracy
- Validation should focus on:
  - Discriminatory power of the rating system (sorting ex-ante between potential defaulters and non-defaulters).
  - Stability of ratings.
- Tests based on calibration of PD (or other parameters) are not appropriate for CRA ratings; CRA tests are intended to assess if defaults occur among the lowest rating categories.
- Empirical finding: all 14 sovereign defaulters involving 12 countries between 1975 and 2009 had (S&P) ratings of BB- or below one year prior to default (i.e., all were noninvestment-grade one year prior and none were in investment-grade categories).
- CAP curves and accuracy ratios (ARs) results (S&P sovereigns):
  - AR = 92 percent over the one-year horizon.
  - AR = 82 percent over three years.
  - AR = 80 percent over five years.
  - AR = 84 percent over 10 years.
  - Interpretation: CRAs are better at predicting defaults over short horizons than long ones.
- Corporate rating ARs (S&P):
  - One-year horizon: 77 percent for financials and 81 percent for nonfinancials (versus 92 percent for sovereigns).
  - Five-year horizon: 63 percent for financials and 71 percent for nonfinancials (versus 80 percent for sovereigns).
- Recommendation: CRAs should calibrate ratings to target credit risk metrics (such as default probabilities) and publish validation tests against such standards to make the rating message more transparent and allow more relevant tests of accuracy.
- BCBS (2005) suggests cardinal accuracy tests but they presuppose specific default probabilities for individual rating grades; CRAs emphasize ordinal accuracy tests instead.

### Examination of rating stability
- There is a tradeoff between rating accuracy and stability; users express a preference for stability because of the certification role of ratings and transaction costs from frequent rating changes.
- Mechanisms used by CRAs to promote stability:
  - Focus on rank ordering credit risk rather than specific credit metrics (muting cyclical impacts).
  - Rate “through the cycle” to capture ex-ante ability to survive cyclical troughs.
  - Apply smoothing rules that change ratings only if (1) the anticipated rating change is expected to be persistent, and/or (2) the prescribed change is more than one notch.
- Empirical stability findings (S&P data, 1975–2009):
  - Higher ratings are more stable than lower ones; sovereign ratings are more stable than corporate ones.
  - Example: 82.1 percent of A- rated sovereigns were still A- rated at the end of each year, versus 77.5 percent of corporates.
  - Example: 1.8 percent of A-rated sovereigns were downgraded three or more notches, versus 6.9 percent of corporates.
- Crisis-period instability:
  - Only 68 percent of ratings remained unchanged during the Asian crisis.
  - Only 63 percent of ratings remained unchanged during the current crisis (so far).
  - The proportion of sovereigns that started each period as investment grade but finished as high yield fell from 2.9 percent to 1.9 percent.
- Average rating change:
  - For Moody’s, the average rating change fell from a downgrade by less 0.4 notch during the Asian crisis to a 0.07 notch during the current crisis (difference not statistically significant).
- Large downgrades (three or more notches) have been concentrated in the investment-grade categories despite their intended greater stability.
- Smoothing rules can undermine stability by creating abrupt downgrades if a prescribed downgrade proves persistent—potentially exacerbating procyclical cliff effects; through-the-crisis methodologies may reduce frequency but remain vulnerable to smoothing-induced cliff effects.

### Policy recommendations
- Remove the excessive reliance of regulations on ratings to reduce certification-driven liquidity impacts and potential cliff effects; this will not affect the information value of ratings, which works mostly through outlook reviews.
- If ratings continue to play a significant role in regulations:
  - CRAs should be more transparent about the quantitative measures they calibrate in the rating process (PDs, LGD, and stability assumptions), how these measures are mapped into ratings, and whether final ratings can be used to infer the parameters used to obtain these measures.
  - This transparency is particularly relevant for external ratings used by banks employing the standardized approach in Basel II, since internal ratings systems are subject to rigorous back testing.
- Ratings should embed the notion that risk is a forward-looking dimension conditional on the macroeconomic scenario; ratings should be better tied to macroeconomic conditions, including resilience to stress scenarios.

*Source: IMF working paper excerpt.*

### REFERENCES

### _wp1223 - REFERENCES

### References
- Alsakka, Rasha, and Owain ap Gwilym, 2010, “Leads and Lags in Sovereign Credit Ratings,” Journal of Banking & Finance, Vol. 34, No. 11, pp. 2614–26.
- Alsakka, Rasha and Owain ap Gwilym, forthcoming, “Rating Agencies’ Credit Signals: An Analysis of Sovereign Watch and Outlook,” International Review of Financial Analysis.
- Annaert, Jan, Marc De Ceuster, Patrick Van Roy, and Cristina Vespro, 2010, “What Determines Euro Area Bank CDS Spreads?” Working Paper 190 (Brussels: National Bank of Belgium).
- Aruoba, S. Borağan, Francis X. Diebold, and Chiara Scotti, 2009, “Real-Time Measurement of Business Conditions,” Journal of Business and Economic Statistics, Vol. 27, pp. 417–27.
- Basel Committee on Banking Supervision (BCBS), 2005, “Studies on the Validation of the Internal Rating Systems,” BCBS Working Papers No. 14 (Basel: Basel Committee on Banking Supervision).
- __________, 2006, “International Convergence of Capital Measurement and Capital Standards: A Revised Framework—Comprehensive Version” (Basel: Basel Committee on Banking Supervision).
- __________, 2009, “Stocktaking on the Use of Credit Ratings” (Basel: Basel Committee on Banking Supervision).
- Bhatia, Ashok Vir, 2002, “Sovereign Credit Ratings Methodology: An Evaluation,” IMF Working Paper 02/170 (Washington: International Monetary Fund).
- Boot, Arnoud W. A., Todd T. Milbourn, and Anjolein Schmeits, 2006, “Credit Ratings as Coordination Mechanisms,” Review of Financial Studies, Vol. 19, pp. 81–118.
- Brooks, Robert, Robert Faff, Joe Hillier, and David Hillier, 2004, “The National Market Impact of Sovereign Rating Changes,” Journal of Banking and Finance, Vol. 28, pp. 233–50.
- Cantor, Richard, and Christopher Mann, 2003, “Measuring the Performance of Corporate Bond Ratings,” Moody’s Investors Services Special Comment (April).
- Cantor, Richard, and Frank Packer, 1996, “Determinants and Impact of Sovereign Credit Ratings,” FRBNY Economic Policy Review, October, pp. 37–54 (New York: Federal Reserve Bank of New York).
- Chambers, John, 2010, “Outlooks: The Sovereign Credit Weathervane, 2009/2010 Update,” Standard & Poor’s RatingsDirect, April 20.
- Dichev, Ilia, and Joseph Piotroski, 2001, “The Long-Run Stock Returns Following Bond Ratings Changes,” Journal of Finance, Vol. 56, pp. 173–203.
- Ederington, Louis, Jess Yawitz, and Brian Roberts, 1987, “The Informational Content of Bond Ratings,” Journal of Financial Research, Vol. 10, pp. 211–26.
- Engelmann, Bernd, Evelyn Hayden, and Dirk Tasche, 2003, “Testing Rating Accuracy,” Risk, Vol. 16, pp. 82–86.
- Fender, Ingo, and John Kiff, 2005, “CDO Rating Methodology: Some Thoughts on Model Risk and its Implications,” Journal of Credit Risk, Vol. 1, pp. 37–58.
- Ferri, Giovanni, Li-gang Liu, and Joseph E. Stiglitz, 1999, “The Procyclical Role of Rating Agencies: Evidence from the East Asian Crisis,” Economic Notes, Vol. 28, pp. 335–55.
- Fitch Ratings, 2010, “Definitions of Ratings and Other Forms of Opinion,” Fitch Ratings, August.
- Granger, Clive W. J., 1969, “Investigating Causal Relations by Econometric Models and Cross-spectral Methods,” Econometrica, Vol. 37, pp. 424–38.
- Güttler, Andre, 2009, “Lead-lag Relationships and Rating Convergence among Credit Rating Agencies,” Research Paper 09-14 (London: European Business School).
- Hand, John, Robert Holthausen, and Richard Leftwich, 1992, “The Effect of Bond Rating Agency Announcements on Bond and Stock Prices,” Journal of Finance, Vol. 47, pp. 733–52.
- Hamilton, David T., and Richard Cantor, 2005, “Rating Transitions and Defaults Conditional on Rating Outlooks Revisited: 1995–2005,” Moody’s Investor Service Special Comment, December.
- Hartelius, Kristian, Kenichiro Kashiwase, and Laura Kodres, 2008, “Emerging Market Spread Compression: Is it Real or Is It Liquidity?” IMF Working Paper 08/10 (Washington: International Monetary Fund).
- Hessol, Gail I., Erkan Erturk, and Jason Ontko, 2007, “Credit Watch and Rating Outlooks Provide Powerful Warning Signals,” Standard & Poor’s RatingsDirect, August 7.
- Hill, Paula, and Robert Faff, 2008, “Do Credit Watch Procedures Affect the Information Content of Sovereign Credit rating Changes? Unpublished working paper.
- Hull, John, Mirela Predescu, and Alan White, 2004, “The Relationship between Credit Default Swap Spreads, Bond Yields, and Credit Ratings Announcements,” Journal of Banking and Finance, Vol. 28, pp. 2789–2811.
- Hurlin, Christophe, and Baptiste Venet, 2001, “Granger Causality Tests in Panel Data Models with Fixed Coefficients, Working Paper Eurisco 2001-09 (Paris: Université Paris IX Dauphine).
- International Monetary Fund (IMF), 1999, International Capital Markets (Washington, July).
- __________, 2010, Global Financial Stability Report, World Economic and Financial Surveys (Washington, October).
- Ismailescu, Iuliana, and Hossein Kazemi, 2010, “The Reaction of Emerging Market Credit Default Swap Spreads to Sovereign Credit Rating Changes,” Journal of Banking and Finance, forthcoming.
- Kaminsky, Graciela, and Sergio Schmukler, 2002, “Emerging Market Instability: Do Sovereign Ratings Affect Country Risk and Stock Returns?” The World Bank Economic Review, Vol.16, pp. 171–95.
- Kiff, John, Michael Kisser, and Liliana Schumacher, forthcoming, “An Inspection of the Through-the-Cycle Rating Methodology,” IMF Working Paper (Washington: International Monetary Fund).
- Larrain, Guillermo, Helmut Reisen, and Julia von Maltzan, 1997, “Emerging Market Risk and Sovereign Credit Ratings” OECD Development Centre, Working Paper 124, April 1997.
- Liu, Pu, and Anjan Thakor, 1984, “Interest Yields, Credit Ratings and Economic Characteristics of State Bonds,” Journal of Money, Credit and Banking, Vol. 16, pp. 344–51.
- Micu, Mirian, Eli Remolona, and Philip Wooldridge, 2006, “The Price Impact of Rating Announcements: Which Announcements Matter?” Working Paper 207 (Basel: Bank of International Settlements).
- Millon, Marcia, and Anjan Thakor, 1985, “Moral Hazard and Information Sharing: A Model of Financial Information Gathering Agencies,” Journal of Finance, Vol. 40, pp. 1403–22.
- Moody’s Investors Service, 2009, “Rating Symbols and Definitions,” Moody’s Investors Service, November.
- Moody’s Investors Service, 2010, “Rating Symbols and Definitions,” Moody’s Investors Service, July.
- Norden, Lars, 2008, “Credit Derivatives. Corporate News, and Credit Ratings,” University of Mannheim Working Paper, October 18.
- Reisen, Helmut, 2003, “Ratings since the Asian Crisis,” OECD Development Centre Working Paper 214 (Paris: OECD Development Centre).
- Reisen, Helmut, and Julia von Maltzan, 1999, "Boom and Bust and Sovereign Ratings," International Finance, Vol. 2, p. 273–93.
- Sobehart, Jorge R., Sean C. Keenan, and Roger M. Stein, 2000, “Benchmarking Quantitative Default Risk Models: A Validation Methodology,” Moody’s Investors Service Rating Methodology, March.
- Standard & Poor’s (S&P), 2009, “Understanding Standard & Poor's Rating Definitions,” Standard & Poor’s RatingsDirect, June 3.
- __________, 2010a, “Another Perspective on Rating Comparability and Performance,” Standard & Poor’s RatingsDirect, April 9.
- __________, 2010b, “Methodology: Credit Stability Criteria,” Standard & Poor’s RatingDirect, May 3.
- Sy, Amadou, 2002, “Emerging Market Bond Spreads and Sovereign Credit Ratings: Reconciling Market Views with Economic Fundamentals,” Emerging Markets Review, Vol. 3, pp. 380–408.

### Appendix 1: Analysis of One-Year Fitch and S&P Sovereign Rating Changes
- Figure A1.1. Average Proportion of Fitch Sovereign Ratings Unchanged Over One Year
  - Data source: Fitch Ratings.
  - Note: Sovereign ratings for 1995-2009 and corporate ratings for 1990-2009.
  - Rating categories shown (labels as in figure): AAAAA+ AAAA+ AAAA- A+ AA- BBB+ BBB BBB- BBB-BB+ BBBB-B+ BB-
  - Two series plotted: Sovereign; Corporate.
  - Y-axis units: In percent of unchanged ratings.
- Figure A1.2. Average Proportion of Fitch Sovereign Ratings Downgraded More Than Two Notches Over One Year
  - Data source: Fitch Ratings.
  - Note: Sovereign ratings for 1995-2009 and corporate ratings for 1990-2009.
  - Y-axis units: In percent of ratings downgraded more than two notches.
  - Rating categories shown (labels as in figure): AAAAA+ AAAA+ AAAA- A+ AA- BBB+ BBB BBB- BBB-BB+ BBBB-B+ BB-
  - Two series plotted: Sovereign; Corporate.
- Figure A1.3. Average Proportion of Moody's Sovereign Ratings Unchanged Over One Year
  - Data source: Moody's Investors Service.
  - Note: Both sovereign and corporate ratings for 1983-2009.
  - Rating categories shown (labels as in figure): Aaa Aa A Baa Ba BCaa-C (as formatted in figure).
  - Two series plotted: Sovereign; Corporate.
  - Y-axis units: In percent of unchanged ratings.
- Figure A1.4. Average Proportion of Moody's Sovereign Ratings Downgraded Through at Least One Grade Over One Year
  - Data source: Moody's Investors Service.
  - Note: Both sovereign and corporate ratings for 1983-2009.
  - Rating categories shown (labels as in figure): Aaa Aa A Baa BaaBaa BCaa-C (as formatted in figure).
  - Two series plotted: Sovereign; Corporate.
  - Y-axis units: In percent of ratings downgraded at least one grade.

### Appendix 2: Sovereign Downgrades and Upgrades During the Asian and Global Crisis
- Figures use Standard & Poor's and Fitch as sources for plotted rating transitions between specified dates.
- Figure A2.1. Asian Crisis: Sovereigns Rated by S&P between July 31, 1997 and December 31, 1998
  - Source: Standard & Poor's.
  - Date range: July 31, 1997 to December 31, 1998.
  - Total number of ratings = 69.
  - On the diagonal line = 53.
  - Quadrant breakdown numbers shown in figure: (2), (2), (4), (2), (4), (7), (4), (11), (7), (3), (9), (3) (as plotted around matrix).
  - Rating categories shown (labels as in figure): AAAAA+ AA AA-A+ A A- BBB+ BBB BBB- BB BB-B+ B B- CCC+ CCC CCC-C D etc.
- Figure A2.2. Current Crisis: Sovereigns Rated by S&P between July 31, 2007 and June 30, 2010
  - Source: Standard & Poor's.
  - Date range: July 31, 2007 to June 30, 2010.
  - Total number of ratings = 107.
  - On the diagonal line = 59.
  - Quadrant breakdown numbers shown in figure: (16), (2), (4), (4), (4), (2), (3), (2), (3), (4), (5), (2), (2), (2), (2), (2), (2), (3), (2), (3), (4), (3), 3, 21, 21, 3 (as plotted around matrix).
  - Rating categories shown (labels as in figure): AAA AA+ AA AA- A+ A A- BBB+ BBB BBB- BB+ BB BB- B+ B B- CCC+ CCC CCC-C D etc.
- Figure A2.3. Asian Crisis: Sovereigns Rated by Fitch between July 31, 1997 and December 31, 1998
  - Source: Fitch.
  - Date range: July 31, 1997 to December 31, 1998.
  - Total number of ratings = 50.
  - On the diagonal line = 35.
  - Quadrant breakdown numbers shown in figure: (2), (3), (4), (9), (3), (4), (2), (3), (3), (4), (2) (as plotted around matrix).
  - Note in figure: Non investment grade on December 31, 1998; Non investment grade on July 31, 1997.
  - Rating categories shown (labels as in figure): AAAAA+ AA AA-A+ A A- BBB+ BBB BBB- BB+ BB BB- B+ B B- CCC+ CCC CCC-C D etc.
- Figure A2.4. Current Crisis: Sovereigns Rated by Fitch between July 31, 2007 and June 30, 2010
  - Source: Fitch.
  - Date range: July 31, 2007 to June 30, 2010.
  - Total number of ratings = 99.
  - On the diagonal line = 63.
  - Quadrant breakdown numbers shown in figure: (15), (2), (4), (3), (4), (2), (4), (2), (4), (6), (2), (3), (6), (2), (2), (3), (2), (3), (2), (2), (3), 4, 11, 18, 3 (as plotted around matrix).
  - Note in figure: Non investment grade on July 31, 2007; Non investment grade on June 30, 2010.
  - Rating categories shown (labels as in figure): AAAAA+ AA AA-A+ A A- BBB+ BBB BBB- BB+ BB BB- B+ B B- CCC+ CCC CCC-C D etc.

*Canonical source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2012/_wp1223.pdf*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2012/_wp1223.pdf_