Predictive Capability of Financial Ratios for Forecasting of Corporate Bankruptcy

Abstract

Bankruptcy of a business firm is an event which results substantial losses to creditors and stockholders. A model which is capable of predicting an upcoming business failure will serve as a very useful tool to reduce such losses by providing warning to the interested parties. This was the main motivation for Beaver (1966) and Altman (1968) to construct bankruptcy prediction models based on the financial data (Deakin 1972). This research study also initiated with a great interest on this subject to investigate the predictive capability of financial ratios for forecasting of corporate distress and bankruptcy events. This study is expounded on similar previous studies by Altman (1968), Ohlson (1980), Beaver (1966) by examining the effectiveness of financial ratios for predicting of corporate distress. The logistics regression analysis (LRA) statistical method is used to scan the risk factors from the previous financial year data and prediction models are constructed which can reasonably classify the expected bankruptcy group and can reasonably predict the solvency status of a firm. The research has been focused on the USA companies only. A set of bankrupted and non-bankrupted company financial data are used for constructing the bankruptcy prediction model and then a second set of bankrupted and non-bankrupted company financial data has been used to test the classification accuracy of the constructed models. The result of this study is consistent with the previous bankruptcy prediction researches outcomes. This study also investigates the time factor implication of bankruptcy prediction models using 5 years financial ratios. Like other research projects this project is not without certain limitations and weaknesses. The bankrupted company data collection and compilation was a great challenge due to most of the bankrupted companies cease to operate or cease to be existed. Thanks to the great treasure of Mergent online database which facilitated collection of bankrupted company data. In order to facilitate identifying and collecting bankrupted company data, it is presumed that the companies which show as inactive status in Mergent online database are distressed or bankrupted companies. Another practical obstacle was the functionality of SPSS software and the output interpretation of the SPSS software; I used Andy Field’s “Discovering Statistics using SPSS” book to decipher the statistical jargons and to formulate the bankruptcy equations. Our constructed prediction model cannot be used universally as the study depended upon exclusively on US firm’s financial data, therefore the constructed prediction model can proved to be very useful tool for the US financial analysts and turnaround specialists to identify the distressed firms. In the case we need to use this model in other geographical location, the coefficients of the predictor variables must be re-estimated using the particular country’s financial data.