REINFORCED URN PROCESSES FOR MODELING CREDIT DEFAULT DISTRIBUTIONS

Abstract

Based on a Reinforced Urn Process introduced by Muliere et al. [11], we propose a stochastic model for the probability of credit default for debt issuers belonging to the same Moody's rated class. The model predicts how a default probability belonging to a given term structure evolves in time as information about credit defaults of debt issuers with the same Moody's rating becomes available. Connections between implied credit default probabilities and credit spreads will be exploited.