中国证券市场三因素模型敏感系数稳定性和可预测性研究  

Coefficient Stability and Forecast Research of The Three-Factor Model in Chinese Stock Market

作　　者：邓长荣[1] 马永开[1] 

机构地区：[1]电子科技大学,成都610054

出　　处：《电子科技大学学报（社科版）》2006年第3期108-112,共5页Journal of University of Electronic Science and Technology of China(Social Sciences Edition)

基　　金：四川省软科学研究重点资助项目(032R025-017)

摘　　要：三因素模型回归系数是测度投资对象系统风险的重要指标。我们利用chow检验对证券收益三因素模型结构的稳定性进行了分析研究,用ADF检验对模型的三个回归系数的稳定性进行了实证分析,运用ARMA和GARCH模型对回归系数的预测能力进行了研究。结果表明三因素模型结构不稳定,但短期比长期结构稳定性要高;大部分组合回归系数时序稳定性较差,同时ARMA和GARCH模型对每个回归系数时间序列进行预测显示有较好的预测能力。The coefficient of three - factor model is an important systemic risk guideline of investment object, We research the stability of the three - factor Model by using chow test and research the coefficient stationary by using unit root test, and forecast the coefficient of the model using ARMA, GARCH model. The results show that the model is instability in the long run, most coefficient is non - stationary, and we can preferably forecast the coefficient by using the ARMA, GARCH model.

关 键 词：CHOW检验 三因素模型 单位根检验 ARMA GARCH 

分 类 号：F830.91[经济管理—金融学]