Are Sovereign Ratings by CRAs Consistent?

Abstract

This study is an attempt to compare and contrast the credit ratings granted by prominent agencies, the so-called Big Three namely S&P, Moody’s and Fitch, that dominate the market. The sovereign ratings are proven to motivate the CDS figures of countries empirically, and low ratings are known to increase the interest paid to liabilities by these countries. We employ the historical data over 1994-2014 on the sovereign ratings of 117 countries to test for whether the ratings assigned by CRAs are significantly different or not, with the help of paired-t and ANOVA tests. Hypothesis test results reveal that such differences are significant for many countries and country groups, suggesting that the ratings by CRAs are not consistent with each other. This is true for BRIC, OECD, and emerging market countries. Extra ANOVA tests that we conducted support our findings.

Key words: Credit Rating Agencies, Sovereign Ratings, Rating Discrepancy, Global Financial Crisis, Paired-t Test, ANOVA.
JEL: E44,F34, G01, G15, G24, H63.

Ova studija je pokušaj upoređivanja i utvrđivanja suprotnosti kreditnih rejtinga koje dodeljuju istaknute agencije, tzv. Velika Trojka, odnosno S&P, Moody’s i Fitch, koje dominiraju na tržištu. Pokazalo se da suvereni rejtinzi motivišu empirijske CDS vrednosti zemalja, a niski rejtinzi povećavaju kamatne obaveze ovih zemalja. Upotrebljavamo istorijske podatke tokom 1994-2014. godine o suverenom rejtingu 117 zemalja kako bi testirali da li su rejtinzi dodeljeni od strane CRA značajno različiti ili ne, uz pomoć uparenih t i ANOVA testova. Rezultati ispitivanja hipoteza pokazuju da su takve razlike značajne za mnoge zemlje i grupe zemalja, što ukazuje na to da rejtinzi dodeljeni od strane CRA nisu konzistentni jedni sa drugima. Ovo važi za zemlje BRIC, OECD i zemlje u razvoju. Dodatni ANOVA testovi koje smo proveli podržavaju naše nalaze.

Ključne reči: agencije za kreditni rejting, suvereni rejting, nesaglasnost rejtinga, globalna finansijska kriza, upareni t-test, ANOVA.