Journal of Harbin Institute of Technology (New Series)
Volume 14, Sup. 2, January 2007, Pages 42-47.

Identifying significant indicators of business failure using rough sets

^a Department of Information Management, National Yunlin University of Science and Technology, Douliou, Yunlin, 640, Taiwan

^b Clayton School of Information Technology, Faculty of Information Technology, Monash University, Clayton, Victoria 3800, Australia

^c Department of Information Management, Far East University, Tainan, Taiwan

Abstract

This paper presents a rough set approach to identifying significant indicators of business failure. Rough set models with decision rules for business failure prediction are constructed in order to examine critical attributes required for rule generation. 14 financial ratios commonly used in existing business failure prediction models and particularly a non-financial variable, auditor switching, are used in rough set models  to enhance the predictive performance. An empirical study is conducted to illustrate the approach, using financial ratio data and auditor switching status of failed firms for the three years before failure. Six sets of rough set decision rules are generated individually with and without the auditor switching variable, using the three-year data respectively. The results of the empirical study show that auditor switching and cash flow ratio are the most significant indicators of business failure; and the critical indicators include quick ratio, earnings per share, and cash flow adequacy ratio. These findings strongly suggest that financial ratios alone may not form a complete set of significant indicators for business failure analysis, as conventionally used in existing models. Non-financial factors may play a significant role in business failure research.

Keywords: Rough sets; Business failure; Financial ration; prediction