An overview of factor models for pricing CDO tranches

We review the pricing of synthetic CDO tranches from the point of view of factor models. Thanks to the factor framework, we can handle a wide range of well-know pricing models. This includes pricing approaches based on copulas, but also structural, multivariate Poisson and affine intensity models. Factor models have become increasingly popular since there are associated with efficient semi-analytical methods and parsimonious parameterization. Moreover, the approach is not restrictive at all in the case of homogeneous credit portfolios. Easy to compute and to handle large portfolio approximations can be provided. In factor models, the distribution of conditional default probabilities is the key input for the pricing of CDO tranches. These conditional default probabilities are also closely related to the distribution of large portfolios. Therefore, we can compare different factor models by simply comparing the distribution functions of the corresponding conditional default probabilities.