A revision of Altman Z-Score model and a comparative analysis of Turkish companies' financial distress prediction

Abstract

<abstract> This paper aims to predict companies' financial distress situation with the use of four different models; Altman Z score, Revised Altman Z Score (Linear Discriminant Analysis) and Quadratic Discriminant Analysis, Random Forest Machine Learning Model with the use of same variables suggested by Altman. Also a review of Altman Z score model for Turkey case is assessed whether it is applicable to Turkish companies and how accurate the results are. This study differentiates itself from the previous studies by the content of the data; it includes both publicly traded companies and private companies. Additionally, there are only a few studies are applied for random forest model in bankruptcy prediction in Turkey. For this reason, it aims to fill the gap from both a rare used model and originality of the data. The data consisted of the 80 firms' financial ratio analysis between the years 2013 to 2018. There are 44 firms are listed in BIST; remaining 36 firms are private firms with the size of small and micro enterprises. Random forest model with use of Altman variables has shown 95% performance and surpassed the other three models. Moreover, the classification results for publicly traded companies was 100% for Random Forest whereas other models have shown greater performance for private firms than publicly traded firms. </abstract>