基于面板离散选择模型的企业财务预警及其行业适用性

Panel Logit Data based Model for Financial Distress Early-Warning and its Applications

基于制造业上市公司2001~2010年的平衡面板数据,在对财务指标和非财务指标进行筛选的基础上,构建了基于面板Logit模型的财务困境预警模型,并分别检验了非财务指标对财务预警模型的影响、行业差异对财务预警模型的影响以及预警临界点的选择对模型的影响。研究发现：资产负债率是财务困境预警中必须给予高度关注的指标;在非财务指标中,行业类型对模型预警能力的影响最大,应针对不同行业建立不同的财务预警模型,外部审计师出具的加说明段的无保留意见以及保留意见可以作为判断公司陷入财务困境概率的参考依据;在预警过程中可适当降低临界点以减小对财务困境公司的误判率。

Balanced panel data of manufacturing industry from 2001 to 2010 are used to build the panel data logit model for financial distress early-warning after choosing the financial indicators and non-financial indicators. Then the influence of non-financial indicator, industry difference and the cut points on financial distress early-warning model is tested. The results show that. （1） there are significant differences between indicators of ST stocks and non-ST stocks; （2） debt-to-assets ratio contributes much to prediction power, to which should be paid more attention; （3） industrial categories contribute the most to prediction power among non-financial indicators~ auditing opinion, especially clean report with explanatory paragraph and unqualified report, can be a reference to financial distress prediction; （4） prediction powers of each model at different cut points are different. Lower cut points lead to higher prediction power for ST stocks.