Research On VaR Of Stock Market Based On Implemented GARCH Class Model

With the continuous innovation of the products and models of financial markets and the fluctuating economy caused by the political changes in the world today,our country is faced with serious economic and financial risks.Therefore,a good measure of financial market risk is not only conducive to preventing the various risks of financial markets,but also conducive to preventing the "Black Swan" and "Grey Rhinoceros" events and the occurrence of risks,so as to help government workers stabilize the good operation of the national economy and help investors to carry out risk prevention and good investment,Therefore,it has very important theoretical practical significance.With the increasing strength of computer technology and the maturation of analytical technology,high-frequency sampling of financial data has become possible,and VaR research based on the realization of GARCH model has become the research focus of financial market investors and scholars.In this paper,the 5min transaction price of Shanghai Composite Index from January 4,2013 to December 28,2018 is selected as the research sample,and the VaR measurement results in different models and different distributions are studied.In this paper,the classical three methods of calculating VaR are analyzed empirically,and it is found that the calculation results will seriously underestimate or overestimate the financial risk,and considering that the peak and thick tail phenomenon is ubiquitous in the financial asset yield sequence,an attempt is made to adopt the Realized GARCH model,and because the normal distribution does not have thick tail,The residual distribution is also extended to the T distribution and GED distribution form of students,and compared with four aspects,such as parameter estimation,model verification,VaR prediction effect and volatility prediction effect,it is found that the GARCH model has a good estimation effect under the GED distribution,but it is difficult to capture the lever effect accurately.Therefore,this paper then considers the introduction of the implemented EGARCH model,the same empirical analysis,from the above four aspects of comparison,the study found that the EGARCH model has been implemented to describe the asymmetry of asset prices better than the implemented GARCH model,and in the implemented EGARCH model,The GED distribution still has a good estimation effect,but at extreme risk levels,it is difficult to measure VaR accurately by achieving EGARCH.The results of this paper show that,in non-extreme cases,VaR,which has achieved the EGARCH model,has a good predictive effect under the residual distribution and the GED distribution,and has a good predictive effect compared with the VaR of the GARCH model.