Bankruptcy Prediction for Financial Sector of Pakistan: Evaluation of Logit and Discriminant Analysis Approaches

Abstract

Bankruptcy prediction is one of the core area in finance that is quite rich in empirical and theoretical work. This study compares two models for measuring the financial position of financial firms listed in Karachi Stock Exchange. The study gives a comprehensive review of two models, namely Altman’s [1] Z-score and an O-Score derived from Ohlson [14]. The purpose of this paper is two folded. First to identify unique characteristics of business failure and to compare effective variables responsible for this response. Secondly to compare two popular accounting-based measures. summarize publiclyavailable information about bankruptcy. The sample period for this study is from 2009 to 2015. From the KSE listed financial firms, a total of 40 firms were selected and accounting ratios were extracted from balance sheet analysis reports published by State bank of Pakistan. The empirical results concluded that the logit model has a high rate of classification as compared to multiple discriminant analysis. The model has obtained overall 85.5% accuracy and identified three significant accounting ratios that are: retained earnings to total asset, earnings before income and taxes to the total asset, and current liabilities to total asset. The finding of this study would benefit stakeholders that are affected by bankruptcies. So in order to take an advantage, it is important to understand the phenomenon that causes bankruptcies.