Dynamic models of credit ratings and default probabilities

Abstract

[ACCESS RESTRICTED TO THE UNIVERSITY OF MISSOURI AT AUTHOR'S REQUEST.] The upheaval in the capital markets since 1997 has highlighted the need for orderly stress testing of banks' portfolios, including both their trading and lending books. Within the past few years important advances have been made in modeling credit risk at the portfolio level. Credit migration matrices are important inputs to many risk management applications; their accurate estimation is therefore critical. I estimate sector-level credit rating transition probabilities based on annual observations and use these estimates to estimate default probabilities conditioned on the previous rating. I also estimate an ordered probit model with firm sales, GDP growth, and other sector and macroeconomic factors for the migration probabilities.