摘要
基于现代Copula理论,选用上海股票市场各行业板块指数,包括:工业股指数(SHGY)、商业股指数(SHSY)、地产股指数(SHDC)、公用事业股指数(GYSY)的组合为研究对象,构建了多元Copula-GARCH模型.同时考虑到相关参数的动态变化性,选择时变相关SJC-Copula模型较全面地研究各行业板块指数之间的相关性.通过对比多元正态分布、多元t分布与加入copula的多元正态copula模型、多元t-copula模型以及时变相关SJC-Copula模型的拟合情况,结果表明:①Copula函数具有比多元分布更为灵活的形式,它能更好地适应和刻画现实金融序列的分布.②上海股票市场各行业板块指数收益率序列之间均有较强的正相关关系.③时变相关的SJC-Copula函数考虑到了相关参数的动态变化性,可以更好地描述随着外部环境的不断变迁,变量间的相关性也在随时间而不断变化.
Based on the modern Copula theory,this paper constructs the multivariable Copula GARCH models.Four parts in Shanghai stock market are involved in the research,including industrial sector index(SHGY),business share index(SHSY),real estate share index(SHDC),public utility share index(GYSY).In the meantime,considering the dynamic change of the related parameters,the time-varying SJC-Copula model is selected to conduct a more comprehensive study of the correlation between industry plate indexes.The results of comparison of the multivariate normal distribution,multivariate t distribution,multivariate normal Copula model,multivariate t Copula model,multivariate Clayton Copula and the time-varying SJC-Copula model show that:(1)The Copula-based models are more flexible than multivariate distribution models so that they can better depict the distribution of the financial series.(2)The four parts in Shanghai stock market index have a remarkable positive correlation.(3)The time-varying SJC-Copula can better adapt to describe the change of the correlation between variables with the ever-changing external environment over time.
出处
《河南科学》
2011年第11期1286-1291,共6页
Henan Science
基金
上海市科技发展基金软科学项目资助(11692105900)