New bankruptcy prediction models for Polish companies

Abstract

New logit models for predicting bankruptcy of Polish companies are presented. Major features of these approaches are: (1) selection of appropriate companies to the sample as the key step of the research, (2) well defined samples, (3) the reasoning based on the unified financial state-ments and (4) acceptable results of prediction – within samples as well as for the hold-out sam-ples. In addition, the presented models of Stepien and Strak [2004], Ciesielski [2004] and Dom-eracki [2004] have been validated for the best companies on the Warsaw Stock Exchange. The validation principle states that the estimate of the probability of bankruptcy for such company shall be less than 0.5. New models for predicting bankruptcy of Polish companies well fit into the current research in the field of financial applications of microeconometrics.