摘要
在常规极大似然估计法中,Copula函数的参数估计受边缘分布函数拟和的影响较大,鉴于此,用基于秩的极大似然法估计Copula函数的参数,并结合常见的4类双参数非对称BBx-Copula函数,对民生银行和浦发银行这两只股票的尾部相关性进行实证分析,结果表明股票市场在低迷时期的尾部相关性高于活跃时期的尾部相关性。
Because of the disadvantage of traditional parametric estimation, we apply a rank - based method to estimate the Copula function and analyzes the tail dependence between the Mingsheng stock and the Pufa stock carefully with BBx - Copulas. The conclusion is that stock returns appear to be more highly correlated during market downtums than upturns.
出处
《统计与信息论坛》
CSSCI
2008年第6期66-71,共6页
Journal of Statistics and Information
基金
国家自然科学基金资助项目<多变量矩序列长期均衡关系及动态金融风险规避策略研究>(70471050)
关键词
秩
COPULA
半参估计
尾部相关
股票收益
rank
Copula
semi- parametric estimation
tail dependence
stock return
