基于财务状况多分类的上市公司财务危机预警

Financial Distress Prediction of Listed Companies Based on Classified Financial Positions

目前的财务危机预警模型大多局限于二分类研究,而公司陷入财务危机往往会经历一个逐步衰败的过程,简单的二分类有时会掩盖某些上市公司财务状况逐渐变差的事实.为了更准确地判断公司的财务状况,将公司分为财务稳定公司、潜在危机公司、财务危机公司和破产公司四类,运用t-2年的财务数据构建了有序多分类logistic回归模型.实证结果表明,模型的预测能力较好,误判成本较低,能提供更加准确的信息.

Recently the studies on financial distress prediction are mostly limited to binary classification.But the process of companies＇ getting into trouble is slow,simple binary classification sometimes can cover up the facts that companies＇ status of getting worse.In order to judge companies＇ financial status exactly,the authors classify the listed companies into four categories：financial healthy companies,potential distress companies,financial distress companies and financial failure companies.On the basis of categories,They apply ordinal logistic regression to construct the financial distress prediction model which is based on the data of year-2.The empirical research results indicate that the financial distress early warning model based on ordinal Logistic model with good forecast ability and low misjudge costs can supply more definite information on financial status of the corporation.