Financial distress prediction of Chinese listed companies: a rough set methodology

Abstract

PurposeTo empirically estimate a rough set (RS) model in financial distress prediction for Chinese listed companies and assess its classification accuracy.Design/methodology/approachRS model is used to test the effect of financial ratios and some non‐financial ratios on the probability of financial distress with a sample of 212 financial distressed firms and 212 healthy firms through years 1998‐2005.FindingsGrowth ratio of per share of equity, net return on assets, earnings per share, interest coverage, ownership concentration coefficient, net profit margin, pledge, retained‐earnings ratio and total assets turnover have strong classification power in financial distress prediction of Chinese listed companies, especially the ownership concentration coefficient. Prediction model combining financial and non‐financial ratios outperforms the one just containing financial ratios.Research limitations/implicationsOne limitation of this research is that it relies on publicly available data and the RS method. Further research can be devoted to making comparisons between the RS method and other prediction methods, and constructing hybrid prediction models with the use of RS and other artificial intellectual methods as well.Practical implicationsIt is necessary to consider every aspect of the company when making financial distress prediction, not just financial ratios, to improve the explanatory power of the prediction model.Originality/valueThis study explores how financial ratios and non‐financial ratios, with the help of RS theory, under the restricted tradability of stocks in the emerging stock market, impact on corporate financial distress. The prediction model employed here considers not only accounting ratios, but also cash flow and corporate governance variables, thus improving the prediction accuracy.