Studies On The Dependence Of Variable Structure SV-Vine-Copula Model

With the rapid development of the world economy,the interdependence between economic and financial markets is gradually strengthened,and the dependency structure is also more complex.Exploring the interdependence between financial markets to analyze the financial risk infection mechanism is conducive to risk management to reduce extreme losses caused by risk infection.Quantifying risk infection and quantitative analysis of risk dependence through statistical model has good theoretical value and application prospects.The stochastic volatility model(SV)can accurately characterize the peak-thickness distribution and volatility characteristics of financial time series.By constructing a joint distribution of multiple different distributions,the Copula model can better describe the nonlinear,asymmetry and tail dependence related characteristics of financial markets.The Vine-Copula model formed by the combination with Vine provides the basis for exploring the nonlinear dependence of multivariate variables.Based on the risk infection of the six major crude oil markets,this paper combines the stochastic volatility model(SV)with the Vine-Copula model and the change point detection method to establish a variable structure SV-Vine-Copula model that can accurately and comprehensively explore financial risk infection.Firstly,the SV-MT model was used to establish the marginal distribution to eliminate typical fluctuations.The mixed C-vine-Copula,mixed D-vine-Copula,mixed R-vine-Copula,all t-R-Vine-Copula,and all Gauss-R-Vine-Copula were established and compared.Then the mixed R-Vine-Copula model with the best fitting effect is selected,and the likelihood ratio statistic change point detection method is used to detect the dependent variable structure points.By comparing the changes of the correlation coefficient before and after the change point,and combining with the granger causality test,the economic reasons for the change point are analyzed,and the risk infection effect between markets is discussed.The constructed variable structure SV-Vine-Copula model quantifies the risk infection and tail spillover effects among the six major crude oil markets.After simulation and empirical analysis,it is found that the SV-MT model can accurately fit the marginal distribution.The mixed R-Vine-Copula has a better fitting effect on the dependency modeling than the other Vine structures Copula.The likelihood ratio statistic change point detection method can effectively detect the structural change points of the Vine-Copula model.Therefore.,the established variable structure SV-Vine-Copula model can well quantify the risk infection and tail spillover effects among the six major crude oil markets.Through the study of the risk infection of six major crude oil markets by this statistical model,it is found that the impact of international events does cause risk infection among crude oil markets,among which WTI and Brent play a dominant role,while the other four markets are in a passive position.At the end of the paper,the paper puts forward some policy recommendations for the risk management of China's domestic crude oil market,and proposes a prospect for the future research directions of nonlinear dependence and risk infection research.