Comparing Logit, Probit and Multiple Discriminant Analysis Models in Predicting Bankruptcy of Companies

Abstract

Experimental studies across the world show that one of investors and other activists’ expectations in capital market is their ability in predicting the status of activity consistency or the bankruptcy of the companies. For this purpose, this study tries to predict the bankruptcy likelihood of manufacturing firms and comparing the predictability power of 3 methods of Logit, Probit, and Multiple-discriminant analysis. Statistical population of the study included all manufacturing firms listed in Stock Exchange of Tehran since 2000-2010.

The results showed that the predictability of Logit, Probit, and Multiple-discriminant models has been81, 80, and % 70 respectively. The results also showed the higher predictability and accuracy of Logit model and also Multiple discriminant analysis methods identify bankrupt firms with high accuracy in the level logit model.