The Correlation Analysis Of The Copula Theory Among Gold, Stocks And Crude Oil

The Copula model is an effective tool to capture the thick and tail of the financial sequence,which shows significant advantages in nonlinear low-frequency data analysis.However,as financial products become more and more abundant and a lot of information generated by real-time transactions is ignored,the market's demand for empirical analysis of high-frequency transactions is increasing gradually and some practical analysis methods and trading strategy are needed.First of all,this paper summarizes the research process and development survey of association analysis method,and gives the main methods and application scenarios of association analysis.What's worth mentioning,we capture the association between low frequency sample data through empirical research on gold,crude oil and the Dow Jones industrial index with the vector autoregressive method,Granger causality test,the binary model of Clayton-Copula and the triplet model of t-copula,and then the Monte Carlo simulation method is used to estimate the value of VAR of the empirical object.The empirical results show that the existing model cannot determine whether there is an association between the low-frequency sample data,while the precision of estimated risk value of the low-frequency Copula model is relatively rough and the risk is overestimated by 2.5 times.Secondly,this paper extends the analysis from low frequency data to high frequency data,and studies the related concepts and main characteristics of high frequency data.The H-Granger test method is to be used to determine the high frequency data analysis sample interval,and some empirical research about the high frequency Copula model have showed based on the high frequency data and the Copula theory,where the latter is extended to high frequency financial data to verify the feasibility and effectiveness of the H-Copula high frequency model.By establishing H-Clayton-Copula high frequency model and H-Frank-Copula high frequency model,the correlation between crude oil and gold high-frequency data was captured,and the VAR value was estimated by the same simulation analysis as the low-frequency data analysis.The results show that the high-frequency Copula model reduces the error of risk compared with the low-frequency Copula model,and reduces the risk from overestimating 2.5 times to 1.3 times.In the end,this paper puts forward the Oil-Granger-Copula connected mixing trading strategies on the basis of analysis.What's more,a trading strategy of multiple frequency copula models is established for crude oil products and some standardized analysis methods have formed about crude oil analysis.