Some methodological questions of bankruptcy prediction and probability of default estimation

Hungary's adoption of Basel II Capital Accord gave new impetus to applying and improving multivariate bankruptcy-prediction methods. This article compares the four most commonly applied bankruptcy-prediction techniques in literature and banking practice, in terms of estimation performance. The study attempts to decide, on the basis of empirical research, whether the simulation procedures with less rigorous conditions of application provide a more reliable bankruptcy prediction and probability of default estimation than traditional mathematic/statistical techniques do. The empirical research reveals that principle-component analysis does not necessarily guarantee better predictive power.