Introduction

The examples mentioned in Chapter 7 as well as the empirical findings in Chapter 11 demonstrate that defaults are not stochastically independent across firms. Dependencies between different obligors in a credit portfolio can arise from two sources. First, we have cyclical default dependence resulting from dependencies of firms in a credit portfolio on some underlying macro-economic factors as for example the overall economic climate. As these risk factors change in time, a firm’s financial success, measured in terms of the default probability or its rating class, can vary. We refer to this source of risk as systematic risk. Methods to quantify the impact of systematic risk on a credit portfolio have been extensively discussed in Parts I and II. Most of these models rely on the conditional independence assumptions which says that, conditional on the underlying risk factors, the default events of the firms in a portfolio are independent. There exists a broad literature on the effect of cyclical default dependence in standard reduced-form models with conditional independence structure, see e.g. [46] or [89]. In these types of models, [81] and [115] showed that the conditional independence framework usually leads to default correlations between obligors which are too low to explain large portfolio loss that do occur from time to time. Hence as different channel of default dependence has to be considered.