Sovereign ratings provided by Credit Rating Agencies (CRAs) are meant as indicators about issuer countries’ ability to pay back their debt in due time. As such, CRAs opinions affect institutional demand and market pricing of the debt of a country. Nevertheless, ratings published by CRAs often exhibit a systematic delay in capturing timely variations in the creditworthiness of an issuer. Despite the fact that this feature provides the markets with a certain degree of stability, avoiding abrupt market shocks caused by unexpected ratings, changes in the credit health of an issuer are often more timely reflected through Credit Default Swaps (CDS) spreads. In this sense, exploiting the high frequency credit market information through the construction of a simple market-implied rating system may prove useful as a real time indicator of a country’s credit rating. Of course, one should have in mind that the ability in anticipating rating changes depends on the speed and accuracy of financial markets in pricing information about an issuer’s creditworthiness as well as on their liquidity.
A notable example of the divergence between market expectations and credit rating agencies decisions on creditworthiness is Greece. In 2019 and up to early 2020, Moody's, S&P Global and Fitch had gradually upgraded the Greek sovereign rating to “Β1” (Stable Outlook), “ΒB-” (Positive Outlook) and “BB” (Positive Outlook), respectively. In November 2020, Moody’s upgraded the Greek sovereign rating to “Ba3” (Stable Outlook) on the basis of both the continuation of the reform orientation in 2020 and the positive economic prospects of the country in the coming years. In April 2021, S&P upgraded the Greek sovereign rating to “BB” (Positive Outlook), from “BB-”. Despite these positive developments, after more than a decade in the “junk” bond rating category, Greek debt remains below the “investment grade” rating despite recent improvements in economic activity and the support given from ECB’s PEPP bond-buying programme.
To analyse the links between investor perceptions and credit rating agencies, we investigate the relationship between CDS spreads and sovereign ratings by employing a simple regression framework. In particular, we employ a one-way fixed effects model on a panel dataset of 60 countries during the period starting from Q1:2004 until Q3:2021. We use the 3-month average of 5Y sovereign CDS in US dollars for each country and the long-term credit rating scores provided by the S&P, Moodys and Fitch rating agencies. Moreover, we control for common time effects by employing three global variables, namely, crude oil price, the VIX index and US Treasury 10-year bond rates. To account for possible heteroscedasticity and autocorrelation, we employ robust standard error clustered at the country level.
Given coefficient estimates, CDS-implied ratings are derived by mapping realized CDS spreads to an estimated rating score under the assumption that there is no mispricing errors by the CDS market. In order to translate the ordinal alphanumeric ranking of ratings into numeric values we assign a linear scaling transformation system starting from zero (indicating selective default (SD)) to 100 (indicating “AAA” rating). We opt to include the credit outlook decisions (denote by *-, *+) provided by each rating agency in order to improve the discreteness of rating agency scores and enrich the informational content of our model. The rating score used in our empirical analysis is the average numeric rating across the three rating agenciesς.