Study On The Correlation Between The Shanghai And Shenzhen Stock Market Skewness Kurtosis Based On M-Copula-GJRSK-M Model

Previous studies mainly study the variance of assets to measure the risks and the corresponding expected return. More and more scholars has paid attention to the asymmetry of financial asset volatility and thick tail phenomenon in recent years. And there are also a growing number of studies on skewness and kurtosis of the assets. However, most of the studies are based on the static angle of risk premium of skewness and kurtosis for assets, without considering the time-varying and high-order moment which could affect the asset pricing, the selection of the optimal portfolio and option pricing.In financial markets, there is often correlation between assets, which makes the return on assets can not present the normal distribution but strict rush thick tail phenomenon. That means the joint distribution of financial assets yields tend to have two kinds of asymmetric phenomenon: one is the asymmetric phenomenon characterized by a single stock yields have asymmetric distribution, namely the existence of skewness risk premium, the other related to the asymmetric between the return on financial assets, where the correlation between assets in a bear market is greater than in a normal or bull market, that is to say, asset return exists tail extreme. Therefore, measurement of asymmetric correlation and tail extreme of financial assets yield is extremely important for measurement and choosing of the portfolio risk.Most of the existing multivariate distribution function is a simple extension of single variance distribution function which can build flexible multivariate distribution using copulas connect function. In addition, copulas connect function can be marginal distribution of random variables and their related structures between separate studies, including the choice of the marginal distribution which dose not have limits. This dissertation uses copulas connect function modeling based on its advantages.based on the understanding above, this paper analyzes the influence of t Correlation between assets on higher moments risk premium using empirical study which includes m-copulas connect- GJRSK- M model is established. This paper selects 3691 sample index of the daily closing price of Shanghai composite index and shenzhen component index between 04/02/2000 to 10/05/2015 as the analysis object.Firstly, this paper introduces copulas connect function, constructed the M- copulas connect function, and expounds the advantages of m- copulas connect function, and describes the advantages of GJRSK- M model and its economic meaning. Then this paper builds m- copulas connect- GJRSK- M model connecting the edge of different types of distribution into a new multivariate distribution, and the distribution of the edge of the multivariate distribution model is decomposed into modeling and related structure modeling, so as to solve the problem of correlation between assets in the study of multidimensional high order moment "dimension disaster" problem. and the maximum likelihood estimation method is used to estimate model parameters. Finally, the empirical shows that the high logarithm yield of dynamic moment risk and dependent features of low tail end of the asymmetric is exist in the Shanghai composite index and shenzhen component index.