摘要
构建基于BFGS算法优化的混合旋转Clayton Copula模型,选取上证指数收益率与恒生指数收益率分别代表中国内地与香港地区股票市场收益率的研究表明,我国内地与香港地区股票市场尾部呈正相依关系,且尾部相依结构具有不对称性,下尾分布较上尾分布更重;相比于单个Copula模型和混合Copula模型,基于BFGS算法优化的混合旋转Clayton Copula模型能更准确地估计参数,更全面地拟合我国内地和香港地区股票市场的尾部相依性。建议两地加强监管合作,共同制定和执行市场监管政策;通过大数据分析、机器学习等技术手段,提高风险识别的准确性和时效性,完善风险预警机制;引导投资者理性看待尾部风险,使其认识到市场收益率尾部相依结构的存在,并在投资决策中充分考虑尾部风险。
A mixed-rotation Clayton Copula model optimized with the BFGS algorithm is constructed,and the returns of SSE index and Hang Seng index selected to represent the returns of Chinese mainland and Hong Kong region stock markets to analyse the tail dependence of Chinese mainland and Hong Kong region stock markets.The empirical results show that Chinese mainland and Hong Kong region stock market returns have a positive tailed dependence structure,showing an asymmetric pattern with a heavier lower tailed distribution than the upper tailed distribution.It is confirmed in the research process that the hybrid rotated Clayton Copula model optimized by the BFGS algorithm can estimate the parameters in a more accurate manner and fit the tail dependence of the stock market returns of Chinese mainland and Hong Kong more comprehensively than the single Copula model or the hybrid Copula model.It is suggested that regulatory authorities in both markets strengthen their cooperation,improve the risk identification accuracy and timeliness and risk warning mechanism through technical means such as big data analysis and machine learning,and guide investors to develop a rational view at tail risk,aware of the existence of the tail-dependence structure of the market returns and taking tail risks into account in their investment decisions.
作者
颜玲
刘金娥
YAN Ling;LIU Jin’e(School of Economics and Management,Xiamen University of Technology,Xiamen 361024,China)
出处
《厦门理工学院学报》
2024年第2期56-65,共10页
Journal of Xiamen University of Technology
基金
福建省自然科学基金项目“大数据环境下基于文本挖掘分析投资者行为和心理对股票市场的影响机制”(2022J01230316)
厦门理工学院科技创新项目“基于Copula模型的媒体情绪与股票收益率相关性研究”(YKJCX2022033)。
