| In this paper, the single financial product choose EGARCH-M model to fitting, andwith the M-Copula function connect to multiple single financial product, under thecondition of GED distribution M-Copula-EGARCH-M model is established, and undera certain level of quantile use Monte Carlo to calculate different proportion ofinvestment risk, namely the value of the VaR and CVaR, further find out under certainquantile different expected return of the optimal investment proportion.The empiricalresults, this paper constructed model is very useful to investors in the financial markets.Due to a large number of empirical shows that financial products are not only there is acommon variance risk, also shows that the risk of time-varying skewness andtime-varying kurtosis, which we must consider the financial risk of product ofthird-order and fourth-order moment (instant kurtosis and skewness risk). GJRSK-Mmodel capture the higher order moments of the financial product risk is needless to say,so in this paper, using it to fit a single financial product, the M-Copula function canconnect multiple financial product distribution, and M-Copula-GJRSK-M model isestablished. Empirical shows that the risk of higher order moment of asymmetriclogarithm yield on the benchmark Shanghai composite index and shenzhen componentindex, according to get high order moment when yields down to increased risk.In this paper, research results are as follows:(1) Paper was based on M-Copula-EGARCH-M-GED portfolio model. Choosingcan depict EGARCH-M model of the risk premium on the yield of a single financialproduct modeling, the joint distribution of the link function of M-Copula connect,model parameter calculation using genetic algorithm, using risk CVaR measurement(based on GED distribution), under the certain condition of quantile, the proportion ofdifferent investment VaR and CVaR value is to use Monte Carlo simulation is obtained,and gives different yield profit under a certain risk of the optimal investment proportion.(2) M-Copula-GJRSK-M model is established to analyze the Shanghai andshenzhen two city dependency. Choose GJRSK-M model fitting of the distribution offinancial products, the asymmetry of the single financial product yield and time-varying,reoccupy M-Copula connect function to connect M-Copula-GJRSK-M model isestablished to study the multiple dependent structure of financial products and expandthe empirical research. (3) according to the characteristics of the financial market investment introductionto include a variety of related structure change of three kinds of Archimedes copulasconnect function, namely chose Gumble, Clayton, and Frank Copula connect functionto construct linear combination of mixed copulas connect function. So you can morefully describe a complex relationship between the relevance of financial markets.(4) the M-Copula connect function fitting is used to satisfy the distribution of theM statistic to test, and the financial products, to model the relationship between;EGARCH-M model of residual error sequence correlation and error distribution fittingis use Ljung-Box Q and K-S test, inspection, and to verify this EGARCH-M rationality;The use of "gramm Charlier and Leon did correction for its definition, using normaldistribution development of GJRSK-M model to estimate, so that they can capture moreprecise to tail distribution and kurtosis steep slow.(5) the selection in Chinese securities market is representative of the Shanghaicomposite index and shenzhen component index, the data processing software, thispaper puts forward model, method and the conclusion of empirical analysis, theempirical conclusions in accordance with the theoretical conclusion is derived. |
PDF Full Text Request