基于DCAC-Copula模型的金融市场相依性及风险度量

Dependence and Risk Measurement of Financial Market Based on DCAC-Copula Model

基于相依关系的非线性动态结构视角,结合动态条件角相关(DCAC)模型能够捕捉瞬时变化的特征以及Copula函数能刻画非线性相依关系的特点,构建DCAC-Copula模型,旨在研究金融市场的非线性动态相依性以及组合风险。首先,通过蒙特卡洛模拟比较DCAC-Copula模型与对照模型在刻画相依关系上的有效性,模拟结果表明DCAC-Copula模型能够较好地捕捉金融时间序列间相依关系的厚尾和动态特征。其次,使用DCAC-Copula模型来预测原油市场的组合风险,并使用MCS检验评估基于VaR和ES的联合损失函数值差异。实证结果显示,DCAC-Copula模型对应的联合回测检验p值在不同置信水平下均为1,由此表明,DCAC-Copula模型能更有效地提高风险预测的精度。

Considering the nonlinear dynamic structure of dependence,this paper combines the characteristic of Dynamic Conditional Angular Correlation(DCAC)model that capture instantaneous changes and the strength of the Copula function in terms of describing nonlinear dependence,and then constructs the DCAC-Copula model to study the nonlinear dynamic dependence and portfolio risk of financial markets.Firstly,Monte Carlo simulation is conducted to assess the effectiveness of DCAC-Copula model and other models in describing dependencies,and the simulation results indicate that the DCAC-Copula model can better capture the thick tail and dynamic characteristics of the correlation between the financial time series.Secondly,DCAC-Copula model is applied to predict the portfolio risk of crude oil markets,and MCS test is used to evaluate the difference of joint loss function value based on VaR and ES.The empirical results show that the p-values of joint backtesting procedures corresponding to DCAC-Copula model are 1 at different confidence levels,which confirms that DCAC-Copula model improve the accuracy of risk prediction effectively.