摘要
针对国内外中小企业财务困境实证研究的不足,本文提出中小企业财务危机判别模型。首先,针对不平衡抽样问题,修正Logistic回归截距,实证结果表明新方法改善了模型"取伪"错误率高居不下的情况;第二,将主成分分析与Logistic回归结合,以更少的指标更好地解决了小样本和解释变量多重共线性等问题。基于实证结果,笔者认为我国中小企业基本符合Luoma等人提出的财务困境渐进规律,并据此提出了针对中小企业特点的财务指标组合体系。
Empirical researches for small and medium-sized companies still remain in the early stage. This paper offers a financial distress prediction model for small and medium- sized companies. Two improvements have been made: to solve the problem of unbalanced sampling, we raise up a solution to correct the constant term in Logistic discriminate function, which helps to reduce the error of misclassi- lying ST companies; the combination of principle component and Logistic function can better solve the problem of small population and multicollinearity. In the end, we have managed to build up a new financial ratio system for small and medium- sized companies.
出处
《数量经济技术经济研究》
CSSCI
北大核心
2010年第8期102-115,共14页
Journal of Quantitative & Technological Economics
基金
教育部博士点基金(No.20050006025)
创新实验计划资助(091000627)
国家自然科学基金创新群体(70821061)