摘要
对相关系数的马尔可夫体制转换模型(RSDC)进行拓展,将多维随机变量的协方差阵分成方差和相关系数,方差利用GARCH模型来描述,相关系数阵则划分为不同的状态。RSDC模型被应用于中国股票市场研究板块指数的动态相关性,给出行业板块的beta系数。实证结果表明两状态恰好体现高、低不同的相关性,且状态持续的概率较大,基于RSDC模型的beta系数与常相关系数的多元GARCH模型存在一定的差异。
In this paper Markov regime switching model of the correlations（RSDC） is developed.In the RSDC model,the covariances is decomposed into variances and correlations.The variance is described with the GARCH process and the correlation matrix divided into the different states.Finally,the model is applied to the stock market in China to study the dynamic correlation of the industry block and compute the coefficient of beta.Empirical results show that the two states reflect high and low correlation and the persistent probability of states is comparatively large.There is the difference on the coefficient of beta between the RSDC and CCC model.
出处
《统计与信息论坛》
CSSCI
2010年第1期14-18,共5页
Journal of Statistics and Information
基金
国家自然科学基金项目<时间序列的因果关系分析与图模型方法研究>(10671044)
浙江省教育厅科研项目<时间序列的图模型理论及应用研究>(Y200804757)
关键词
体制转换
相关系数
BETA
regime switching
correlation coefficient
Beta
