An Empirical Analysis of Delisting Risk Factors with the GEE Model in Korea

Purpose - In most previous studies on predicting corporate failure, the main source of information for predictor variables is financial ratios on financial statements, and constructing a predicting model based on such accounting materials has several shortcomings. Therefore, the purpose of this study is to propose other sources of information, as well as accounting information, to examine the delisting risk factors and forecast corporate failure. Another purpose of this study is to present advanced statistical techniques of corporate failure. Design/Methodology/Approach - Most previous failure prediction studies have applied models analyzing cross-sectional data, including linear discriminant analysis, logit and probit models. This study employs a dynamic statistical model, specifically the GEE model. Findings - This stud y shows that audit opinion and credit assessment provide useful information for predicting corporate failure. It indicates that not only financial ratios but also variables with qualitative characteristics could serve for the corporate failure prediction model. It also shows that other delisting risk factors found empirically in this study are useful in examining and predicting corporate failure. Research Implications - This study contributes by analyzing the usefulness of audit opinion and credit assessment on the examination and the prediction of corporate failure. Moreover, the GEE model is an improvement from previous models by supplementing the shortcomings of the cross-sectional models in that the GEE model adequately reflect changes in time.