A portfolio optimization model for corporate bonds subject to credit risk

ABSTRACT This paper proposes a new approach for determining an optimal portfolio consisting of credit-sensitive assets such as corporate bonds. The model combines the Jarrow–Lando–Turnbull model and a multivariate Markov model, the former for evaluating future prices of defaultable discount bonds and the latter for generating scenarios of the dynamics of correlated credit ratings in multiple corporate bonds. Unlike the existing models, our model has several desirable features. In particular, the future prices of defaultable assets are evaluated under the no-arbitrage valuation framework using observable market data only, and the calculated present values are consistent with the current market values. Some numerical examples are given to demonstrate the usefulness of the model.