Credit risk modelling : an intensity based approach

Abstract

The main aim of this dissertation is to use the theory of marked point processes to shed some light over the way Maltese firms migrate from one credit rating to another. The time spent within a particular credit rating and the rate at which firms migrate between credit ratings can help banks and other financial institutions assess the credit worthiness of firms. During the course of the dissertation four credit ratings are going to be proposed. Furthermore, the way in which a firm is given a credit rating is based on a mix between empirical work, discriminant analysis and financial accounting theory. These credit ratings should not be taken as the correct cut off levels but as indicators only. To model the migration of firms between different credit ratings, three interrelated stochastic processes are going to be proposed. In particular we will present the 9-doubly stochastic Markov chain. As the name implies, the intensity of this process is itself a stochastic processes, which in turn is assumed to be influenced by time and by a set of accounting ratios that are directly related to the financial health of a firm. During the course of the dissertation we will prove that such a process exists and afterwards we will propose three distinct ways with which the cumulative intensity matrix and the survival time can be estimated from the available data set. In particular we will propose the parametric, semi parametric and non-parametric approaches.