Assessment of Credit Ratings and Credit Risk Models on Public Bonds

Summary:

The article investigates the performance of two different credit risk models, the credit ratings of S&P and the market-based credit risk models of Bloomberg, using 12,679 new corporate bonds issued by 933 firms through public offerings in the United States over the 1999–2015 period.

Abstract:

The authors investigate the performance of two different credit risk models, the credit ratings of S&P and the market-based credit risk models of Bloomberg, using 12,679 new corporate bonds issued by 933 firms through public offerings in the United States over the 1999–2015 period. In particular, they divide our sample into bonds issued by financial firms and those issued by non-financial firms, and find that (1) even though both the S&P ratings and the Bloomberg models affect the yield spreads significantly, the former has more statistical power in determining the yield spreads; (2) bond ratings of financial firms are higher than those of non-financials, but financial firms pay a higher cost of debt than non-financial firms; (3) S&P credit ratings are superior to Bloomberg models in predicting actual default of the bonds; and (4) financial firms are less likely to default than non-financial firms.

Research questions:

(1) Which of the S&P ratings or the Bloomberg model has more statistical power in determining the yield spreads? (2) Bond ratings vs. cost of debt for financial and non-financial firms?
(3)Which is a better method in predicting actual default of the bonds
(4) Which type of firms, financial or non-financial firms, has a higher probability to default?

Coauthors: