Time series properties of corporate credit rating transitions by regions and rating agencies

ABSTRACT

The study confirms the time series properties of the credit rating transitions and finds commonalities and differences across regions and rating agencies. Significant differences were found between regions but not between rating agencies.

KEYWORDS:

• credit ratings
• credit ratings transition matrix
• time series
• time-homogeneous Markov Chain
• unit root test

JEL CLASSIFICATION:

• G23
• G24
• G32
• G33

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

¹ The matrices in S&P and Fitch data are 7 × 8 with 56 elements for each year, while Moody’s are 8 × 9 with 72 elements for each year. The yearly average number of issuers in S&P is 6,233 in global, 3,042 in the US, 1,319 in Europe, and 1,101 in emerging markets. The yearly average number of issuers in Moody’s and Fitch are 6,184 and 3,042 respectively.

² S&P did not publish US default data for 2008; thus, 2007Q4 to 2008Q3 data are used instead.

³ The source of Real GDP of Global, US, Europe, Emerging Market are from IMF, Bureau of Economic Analysis, European Union, and Bloomberg, respectively.

⁴ Previous studies have tended to look at individual company rating changes over a fixed period of time, and there has been a lack of studies looking at aggregate level of credit rating changes on an annual basis.

⁵ Due to the sticky nature of rating changes, annual rating changes are a significant contributor to rating retention. Elements that did not change during the study period (e.g. from B to AAA) were classified as non-effective elements.