Evaluating Predictive power of Data Envelopment Analysis Technique Compared with Logit and Probit Models in Predicting Corporate Bankruptcy

Abstract

Simultaneous with extensive environmental changes and the rapid development of technology which has increasingly accelerated economy, competitiveness economical enterprises have restricted earning profit and make probable closing of bankrupt firms. Thus it seems necessary to find a model that can predict financial crisis and bankruptcy of companies. Nowadays occurrence of significant progress in other sciences, such as computer and math attract the attention of the financial scholars toward designing and using more exact patterns like Data Envelopment Analysis (DEA). For this purpose, this study uses DEA technique to predict the bankruptcy likelihood of manufacturing firms and also compare its predictability with2 methods : Logit and Probit models. Study sample includes all manufacturing firms listed in Stock Exchange of Tehran from 2000-2010. The results showed that the accuracy of the designed model under DEA technique is %72 and the predictability of Logit and Probit models has been81, and %80 respectively. The results also showed DEA was proved to be an effective tool for predicting bankruptcy likelihood of manufacturing firms; but,it acted less efficient than Logit and Probit models.