Modele predykcji bankructwa i ich zastosowanie dla rynku NewConnect
Default Prediction Models and Their Application to Poland’s NewConnect Market

Summary/Abstract: This paper deals with modeling the default of enterprises listed on Poland’s NewConnect market. The study covers an overview of the empirical literature on default prediction in Poland and proposes logit models to predict the default of enterprises listed on the NewConnect market over a one-year horizon. The lack of robustness of the estimates suggests there is no stable or monotonic relation between the financial indicators and default probability on the NewConnect market. Moreover, the models estimated in the study as well as those proposed in the literature suffer from a lack of out-of-sample predictive capabilities. Despite this, default prediction models seem to be potentially useful in the selection of stocks and in weighing them in the investment portfolio. Portfolios constructed on the basis of default prediction models, both those estimated in this paper and those proposed in the literature, are more profitable than a market portfolio with equal weights in each stock.

• Details
• Contents