Predicting bankruptcy via induction

Abstract

Research has consistently shown that auditors disclose going-concern problems for less than 50% of all business failures. As evidenced by widespread litigation, investors and creditors believe that this performance needs to be improved. This paper reports on research which analysed financial data for 60 public companies via an inductive inferencing algorithm. The end result was a simple and theoretically consistent model that was 97% accurate in predicting bankruptcy. Auditors, investors and creditors may find the model useful in improving their bankruptcy prediction capability.