Copula-MIDAS Modeling And Its Application

Under the impact of economic globalization and financial integration,financial markets are more closely related and more complex.,which puts forward higher requirements for the study of financial markets in the academia and the industry.Accurately depicting the dependent structure among financial markets can improve the accuracy of decision-making and reduce the risk of decision-making,so as to achieve the purpose of asset allocation optimization and financial risk measurement.Therefore,the ability to accurately portray the dependent structure among financial markets has certain theoretical and practical significance for investors' investment decisions and regulators' monitoring risks.With the continuous deepening of financial markets,the dependent structure among different financial markets has become more and more complex,mainly manifested in two points:first,the dependent structure between different financial markets tends to be nonlinear;second,the financial market is an interrelated and interdependent organic whole,the dependent structure between different financial markets will not only be affected by the high frequency information of the micro level,but also affected by the low frequency information of macro level.Therefore,in view of the shortcomings of the existing research,the paper combines the MIDAS model of different frequency information and the Copula theory that can measure the nonlinear dependent structure,and constructs the Copula-MIDAS model that can comprehensively and accurately describe the dependent structure and risk spillover effect of the market.The article uses five market daily return data from January 4,2011 to December 29,2017 as samples to analyze the dependent structure and the risk spillover effect among markets.The main work and conclusions of the paper are as follows:First,the GARCH-MIDAS model is used to fit the five market daily returns as the marginal distribution,it is found that the GARCH-MIDAS-LI-skew t model can fit the five market fluctuation processes well.Second,five common Copula functions are used to describe the dependent structure among the markets,it was found that the t-Copula function has the best fitting effect on the dependent structure among the markets.Third,the static and dynamic dependent structure analysis are carried out,it is found that the dependent structures among the markets have time-varying and asymmetry.Fourth,an analysis of risk spillover effect among the markets was conducted,it was found thatthe use of unconditional risk value underestimates the actual risks,and using CoVaR can measure the actual risks more fully and accurately.The degree of risk spillover from one market to another is positively related to the size of its own risk.There are significant two-way risk spillovers among the five financial markets,and the risk spillovers are positive and asymmetric.the stock market and the futures market are the net risk spillovers,while the commodity market,the bond market and the foreign exchange market are the risk net recipients.