基于D-vine-Copula-DCC-GARCH模型的比特币市场风险溢出效应研究  

Risk Spillover Effect of Bitcoin Market Based on D-vine-Copula-DCC-GARCH Model

作　　者：彭选华[1] PENG Xuan-hua(Economics School,Southwest University of Political Science and Law,Chongqing 401120,China)

机构地区：[1]西南政法大学经济学院,重庆401120

出　　处：《数理统计与管理》2023年第4期701-713,共13页Journal of Applied Statistics and Management

基　　金：重庆市自然科学基金面上项目(cstc2019jcyj-msxmX0192);重庆市教育委员会科技计划项目(KJQN202100308);重庆市教育委员会人文社科研究项目(18SKGH006)。

摘　　要：为探究比特币市场风险的溢出效应,本文考虑价格波动的时变性和动态相依性,融合D-vine-Copula理论与DCC-GARCH模型,构建D-vine-Copula-DCC-GARCH模型,得到了似然函数和参数后验分布的近似形式,接着应用蒙特卡洛马尔科夫(MCMC)方法估计模型参数,进而计算风险溢出量(△CoVaR),最后选取比特币、人民币、欧元和日元进行实证研究。结果表明:模型构建合理,参数估计的MCMC方法优于MLE方法;同最小二乘法(OLS)、分位数回归方法(Quant)比较,本模型能更好地测度比特币与法定货币之间风险溢出的非对称性,前者向后者的风险溢出可防可控,但比特币对法定货币的风险冲击不能忽略,从而拓展了Copula理论在加密货币风险管理中的应用。这有助于业界加强重视民间数字货币的风险防范。In order to explore the risk spillover effect of bitcoin,we considers the time-varying and dependent characteristics of volatility,and combines D-vine-Copula and DCC-GARCH model to construct a D-vine-Copula-DCC-GARCH model.The approximate form of likelihood function and parameter posteriori distribution is obtained.Monte Carlo Markov(MCMC)method is used to estimate model parameters,and then risk spillover(ACoVaR)is calculated.Finally,bitcoin,RMB,Euro and Yen are used for empirical research.The results show that the model is reasonable,and the MCMC method is better than the MLE method.compared with the least square method(OLS)and quantile regression method(Quant),our model can better measure the relationship between Bitcoin and sovereign currencies.The asymmetric characteristics of the risk spillover between the former and the latter are preventable and controllable,but it cannot be ignored that the impact of Bitcoin price fuctuations on sovereign currencies.Therefore,we expand the application of Copula theory in risk management which is conducive to risk prevention of private digital currency.

关 键 词：比特币 风险溢出 Vine-copula DCC-GARCH △CoVaR 

分 类 号：F224.0[经济管理—国民经济] O212[理学—概率论与数理统计]