基于不同指标类型的财务危机预测模型比较——来自我国制造业上市公司的实证研究

A Comparative Study on Different Financial Crisis Budget Modellings based on Different Indicators:Empirical Studies from China's Manufacturing PLCs

本文选取我国证券市场上446家制造业上市公司,以因财务状况异常而被ST作为公司进入财务危机为标志,分别建立了基于财务指标的危机预测模型、基于非财务指标的危机预测模型和基于混合指标的危机预测模型,并对模型的预测结果进行了比较分析。分析发现,财务指标模型对于训练样本预测精度好于非财务指标模型,但模型外推性(稳定性)不如非财务指标模型。财务指标模型和非财务指标模型的互补性较强。混合指标模型兼具了单指标类型模型的优点,对于训练样本和测试样本的预测精度都具较高水平。

This thesis explored predicting model of financial crisis by choosing 446 manufacture companies in stock market which being ST （Specially Treated） status and in financial crisis as the sample. The author found three types of financial crisis predicting models： financial-variable model, non- financial-variable model, mixed-variable model. Then, We finished a severe comparative analysis of predicting accuracy of these models. The results showed that predicting accuracy of financial-variable model is higher than non-financial-variable model for estimation sample, but lower for validation sample. Thereby Financial-variable model and non-financial-variable model are well mutually complementary. At the end, Mixed-variable model has the advantages of both financial-varible model and non-financial-variable model.