Abstract: Accurate measurement of risk spillover effect is an important prerequisite for financial risk management and portfolio construction, while the nonlinear and dynamic dependence structure between financial markets are always a difficult problem in the research of financial risk spillover effect. This paper combines marginal distribution model that can describe important styled facts to measure financial risk spillover effect between financial markets with GAS t-Copula model and CoVaR method, and taking the Chinese mainland and other four stock markets as the research object, we test the accuracy and reliability of risk spillover model. The empirical results show that there is a positive and dynamic correlation between four stock markets including Chinese mainland and the US stock market. With the outbreak of the financial crisis, the correlation between four stock markets and US stock market gradually increases to the maximum. Affected by the risk spillover effect of the US stock market, the received spillover effect of four stock markets presents significant asymmetry. The spillover effect of downside risk is stronger than that of rising risk, and the spillover effect intensity of Chinese mainland stock market is significantly smaller than that of Hong Kong, Japan and UK stock market.

Key words: asymmetric risk spillover effect, GAS t-copula model, CoVaR

摘要： 准确测度金融风险溢出效应对于金融风险管理和构建投资组合具有重要意义,而金融市场之间的非线性及动态相关结构一直是风险溢出效应研究中的难点问题之一。本文通过引入GAS t-copula模型与CoVaR方法,结合能够刻画重要典型事实特征的边缘分布模型,构建了金融市场间的风险溢出效应测度模型,以中国内地等五个股市为研究对象,测度美国股市对中国内地等四个重要股市的风险溢出效应,以检验模型的可靠性与准确性。实证结果表明:中国内地等四个股市与美国股市之间呈现出显著为正且时变相关结构,随着金融危机的爆发,相关系数逐渐增加达到最大值;中国内地等四个股市受到美国股市的风险溢出效应呈现出非对称特征,即下跌风险溢出效应强度显著大于上涨风险溢出效应;中国内地股市受到的金融风险溢出效应显著小于香港、日本以及英国股市。

关键词: 非对称风险溢出效应, GAS t-Copula模型, CoVaR