Short-term Investment Analysis Based On EMD-GARCH-Copula Model

In today’s world,social progress is accelerating and investment cycles are shortening.It is important to build a portfolio and monitor risk with the goal of short-term investing.Short-term investment stocks usually have better liquidity,that is,higher turnover.Due to many speculative factors,the price of these stocks fluctuates greatly,and the investment risk is correspondingly large,so there will be a lot of irrational behavior in the market.At present,the original data of volatility in most empirical literatures are intraday data.In fact,intraday volatility is the result of the superposition of short-term investment volatility and long-term investment volatility,which aggregates risk factors in the entire stock field.For stocks with high turnover,the most useful data for predicting their future development comes from the short-cycle component of intraday volatility.This paper combines the S＆P 500 manufacturing,financial,technology,energy and pharmaceutical indices to study the portfolio selection and risk measurement of short-cycle investments.First,the S＆P 500 volatility is decomposed into multiple time scales using the EMD decomposition method to simulate the behavior of traders within the corresponding investment horizon.The short-period component corresponds to high turnover,and the long-period component corresponds to long-term holdings.The stationarity test,white noise test and ARCH effect test are carried out on the decomposed short-period components.Using different ARMA-GARCH to construct the marginal distribution of volatility short-period components,and compare the model performance according to the goodness of fit.Secondly,the Vine Copula modeling is carried out on the residual error sequence after probability integral transformation,and the model parameters are estimated by the estimation method according to the maximum likelihood,and the C-Vine Copula,D-Vine Copula,and R-Vine Copula models are obtained.The model performance was compared from two perspectives,tree composition and goodness of fit.The comparison shows that the R-Vine Copula model has better sensitivity and better fitting efficiency.Through this model,a static analysis of the dependencies between the five S＆P short-term indicators is carried out,and the analysis results show that the coefficient of rank correlation can be effectively reduced by adding a conditional market.Again,out-of-sample VaR was estimated by a rolling-window Monte Carlo simulation method and Kupiec validity test was performed.The results show that the risk estimation based on the GARCH（1,1）-R Vine Copula model is the most accurate,and it further explains that the R-Vine Copula model is used to describe the dependence between the S＆P 500 short-term index return series.The most suitable.Finally,considering the dependencies between financial assets can produce the effect of optimizing the portfolio,the portfolio model combining R-Vine Copula and Mean-CVaR is used for short-period portfolio research.The results show that the resulting asset allocation method can reduce portfolio losses while increasing returns.Based on empirical analysis conclusions,speculative investors who invest in short-term investments should allocate asset weights according to their actual conditions in order to achieve the purpose of profit.The innovation of this paper is that:（1）The EMD method is introduced to decompose the volatility of the original data,and the decomposed short-cycle components cover more abundant short-cycle information.Analysis of it will help us better understand the dynamics of stock market behavior,and then conduct investment analysis..（2）Use GARCH,EGARCH and TGARCH models of normal distribution,t distribution,partial t distribution and GED distribution to construct the volatility model of short-cycle returns.From the perspective of tree structure and goodness of fit,the Vine Copula model with the best fitting effect was selected.（3）Using the Monte Carlo simulation method,the VaR outside the sample is estimated in the form of a rolling window and the Kupiec validity test is performed.Short-period portfolio research was conducted using a portfolio model combining R-Vine Copula with Mean-CVaR.