Controlling Bankruptcy: An Early Warning Signal Model

Array

Abstract

The continued development to best fit the model in prediction of bankruptcy has been of substantial interest over the decades. In accounting models of bankruptcy predictions usually single model appears to be a generalized one for each industry, even when industries are structurally different in terms of their financial, operational, and investing needs. This study investigates the causes of financial failure of textile companies (Around 100 companies) delisted from PSX during 2002-2017. A firm is financially distressed due to improper working capital management which led the firm to go bankrupt in the long run. We used 2 sets of data; estimation sample 2002-2012 & cross validation sample 2014-2017 and examined 20 ratios from Liquidity, profitability, Leverage and efficiency, for a five-year period preceding liquidation .Multiple Discriminant Analysis (MDA) approach followed by Logit regression were used. Four ratios; working capital to total assets,  acid test, sales to total assets, and sales growth as proxy of working capital, were identified as significant predictors. The resultant model has early forewarning signals for potential bankruptcy that can occur which could not be generalized one for each industry.