摘要
用回归方法作企业财务困境判别模型时,财务指标的取舍依赖t检验和F检验,保留的指标受主观假设和共线性问题影响.因子分析方法可将众多财务指标减少为综合反映企业财务状况的主成份,为Logit分类选择模型提供企业的财务状况预警指标.上市公司实证分析中,因子分析找到的7个主成份的累计方差占25个财务指标总方差的81.427%,且7个主成份相互独立.因子分析方法和Logit模型的结合可以避免财务指标选取时的随意性和共线性.主成份综合反映众多指标的差异,能够最大程度地区分企业财务状况好坏,有效进行企业财务困境预警.
In the discriminant analysis and prediction of corporate financial distress, selections of financial ratios in regression model rely on t-test and F-test, and the retained variables may contain the impact of subjective hypothesis and the problem of collinearity. Factor analysis can reduce many financial ratios to principal components that reflect the comprehensive financial status of enterprises. These components will serve as the input variables for the binary Logistic function of corporate financial distress forecasting. Empirical studies for listed companies found 7 mutually-independent principal components, whose cumulative variances covered (81.427%) of the total variances of 25 financial ratios.The combination of factor analysis with Logit model can avoid the capriciousness and collinearity in the selection of financial ratios. The principal components can maximize the differentiation between sound financial conditions and financial distresses, and make early warnings of corporate financial distresses.
出处
《湖南大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2005年第1期125-128,共4页
Journal of Hunan University:Natural Sciences
基金
国家自然科学基金资助项目(70172018)
关键词
财务困境
预警
LOGIT模型
因子分析
financial distress
early warning
Logit model
factor analysis
