我国金融行业间风险相依性研究——基于隐马尔科夫混合Copula模型

Research on Risk Interdependency of China’s Financial Industries Based on Hidden Markov Mixed Copula Model

基于GAS边缘分布模型,将混合Copula嵌套于隐马尔科夫模型框架中,构建高维动态混合Copula模型,实证研究了我国金融行业中的银行、保险、证券和信托4个子行业间的动态相依性与最优动态路径,并对金融四大子行业间的尾部情况进行了分析。研究结果显示:基于隐马尔科夫混合Copula模型优于3个单一Copula模型和混合Copula模型,它能较好地描述金融行业间的动态相依性和动态转换路径;还能通过高相依状态来有效地捕捉加剧金融行业风险传染的重大事件;两状态的尾部相关系数显示,金融行业高状态时更易发生尾部风险,发生风险时,银行业和保险业对彼此冲击最敏感,且更易受到对方冲击的影响。

Based on GAS model,this paper constructs a high-dimensional dynamic mixed Copula model by incorporating mixed Copula into the framework of hidden markov model,empirically studies the dynamic interdependency and optimal dynamic transformation path between the four sub-sectors of banking,insurance,securities and trust in China’s financial industry.The constructed hmm mixed Copula model is compared with three single Copula models and the mixed Copula model.Moreover,the tail situation between the four sub-sectors of finance is analyzed.The results show that dynamic mixed Copula model based on hidden markov model is better than three single Copula models and the mixed Copula.This model not only depicts the dynamic dependency and the dynamic transition path between the four sub-sectors,but also effectively captures the major events that aggravate the risk contagion in financial sectors through the high dependence state.The tail correlation coefficients of two states show that tail risks are more likely to occur between financial industries at high levels.When risks occur,banking and insurance industries are most sensitive to external shocks,and more vulnerable to shocks each other.