Are Risk Index Models Useful for Firm Failure Prediction?

Abstract

In business practice in order to evaluate clients’ risk of failure, banks often use risk index models. These models represent techniques which are very easy for practical implementation but without mathematical/statistical sophistication. Academic research starting from Altman’s seminal paper often employs different highly sophisticated techniques such as discriminant analysis, logistic regression, rough sets, neural networks, fuzzy logic, etc. Therefore, it is very interesting to investigate whether modern sophisticated modeling techniques significantly outperform risk index models in firm failure prediction. For the purpose of this paper, risk index was calculated by using five financial ratios (ROA, EBITDA Margin, Credit period, Current ratio and Solvency ratio) that were confirmed to be good discriminators by logistic regression model and artificial neural network. Empirical test conducted on the sample of 323 solvent and 323 insolvent manufacturing firms from Croatia has shown that risk index model performed pretty good in the segment of solvent firms. However, in the segment of insolvent firms risk index model has shown moderate result in comparison with logistic regression model and artificial neural network.