Term structure of credit spreads with learning and anticipation effects

Abstract

In this paper we unify structural and reduced form models, and further extend Duffie and Lando (2001) to consider the relation between credit spreads and investors' information structures. In the first part of the paper, investors' information is incomplete; we show how investors should utilize spot markets information to learning asset process parameters and infer the default barrier through time. In the second part of the paper, investors' information set is enlarged to include future and derivative markets information. By way of conditioning on this extended information set, we show how to incorporate anticipative messages into our original model, which gives us a better estimation of the firm's survival probability. In the numerical simulation, our credit spreads fit very with with empirical results. This fully demonstrates that some components of credit spread can be attributed to learning and anticipation effects.