Predicting bankruptcy for norwegian firms : a study of Altman’s Z´´-model using alternative ratios

Abstract

In this thesis, we study whether modern accounting ratios based on deductive reasoning and modern financial statements are superior to older, conventional ratios. The focus of this study is to evaluate to what extent alternative ratios can improve bankruptcy prediction models. This is done using Altman’s revised Z´´-model as a base throughout the study. To the best of our knowledge, this is the first study using this approach. We have found no studies that aim to improve the Z´´-model by replacing the ratios with alternative ratios that consider a similar aspect. Additionally, we found no studies that directly criticize the ratios applied by Altman. We find a general limitation on the subject of bankruptcy prediction to be a lack of reasoning behind the applied ratios. We develop alternative models to the Z´´-model. These models are based on the outline of the Z´´-model and produced using the same statistical approach, namely multivariate discriminant analysis. Our models were developed using a sample of 158 Norwegian firms from 2009-2016. The sample consists of 79 bankrupt firms and 79 nonbankrupt firms. In general, we find that a majority of the alternative ratios applied in the analysis improved the Z´´-model on an individual basis. We also highlight three alternative models that produce results superior to those of the Z´´-model. These models all consist of two alternative ratios and two of Altman’s original ratios. Generally, we found Financial Assets/Liabilities to be a particularly good ratio. On the other hand, we found Working Capital/Total Assets, which was part of the original model, to be a poor ratio. The findings of this study support our hypothesis that some modern ratios are better suited to predicting bankruptcy than conventional ratios.