Modelling the Financial Failure of Romanian Stock Companies

The aim of the present article is to model and predict the financial failure of non-financial companies listed on the Bucharest Stock Exchange. The prediction models are based on the companies’ financial reports. The paper emphasizes the importance of processing outlier data and the significance of categorical independent variables. The paper contributes to bankruptcy and corporate financial failure research by presenting a Romanian situation. Results show that the model that uses 3-year financial data prior to the failure has a better accuracy. Several models have been compared, and it was found that using categorical independent variables as explanatory variables increased the accuracy of the models against those that used numerical regressors.