基于HAC-广义多维CoES模型的股票市场风险溢出研究

Research on Risk Spillovers among Stock Markets Using HAC Generalized Multi-CoES Model

为了精确测度多重金融风险溢出效应,本文在多维CoVaR方法的基础上提出能够度量整个尾部预期损失的广义多维CoES方法,并基于分层阿基米德Copula(HAC)给出了广义多维CoES的计算公式.基于HAC-广义多维CoES模型对中国、美国、中国香港股票市场间风险溢出效应的实证研究表明:美国与中国香港的股票市场对中国股票市场的风险溢出存在显著的叠加效应,失败率检验结果显示广义多维CoES相对于多维CoVaR具有更高的准确性;广义多维CoES的动态走势显示中国股票市场受到美国股市与中国香港股市的多重风险溢出效应,呈现顺周期性;稳健性检验结果显示,基于HAC模型度量广义多维CoES的计算方法具有较好的稳健性.HAC-广义多维CoES模型为监管者和投资者识别多个金融市场之间可能存在的风险溢出叠加效应提供了有效的实践应用工具.

In order to accurately measure the multi-dimensional financial risk spillover effect,a generalized multi-CoES method which can measure the mean value of the whole tail conditional loss is proposed based on multi-CoVaR,and the calculation formula of generalized multi-CoES is given based on Hierarchical Archimedean Copula(HAC).According to the empirical study on risk spillovers among China,the US and Hong Kong(China)stock markets using the new approach,there is a significant superposition effect on risk spillovers of stock markets from US and Hong Kong(China)to China.The back testing results show that generalized multi-CoES has higher accuracy than multi-CoVaR.Moreover,the dynamic trend of generalized multi-CoES shows that the multi-dimensional risk spillover effect of stock markets from US and Hong Kong(China)to China is pro-cyclical.Finally,the robustness test results show that the method based on HAC model to measure generalized multi-CoES is robust.The HAC generalized multi-CoES model provides an effective tool which can be applied in the practice for regulators and investors to identify the possible superposition effects of risk spillovers among multiple financial markets.