Factor Garch Model With Multiple Jumps On The Chinese Stock Market And Experimental Study Of The Ability To Capture

The main source of financial risk is the volatility of the price of financialassets，financial assets yield uncertainty and the volatility on financial assets has beenthe focus of scholars at home and abroad to study the issue of financial risks. In recentyears, affected by the new round of international financial crisis, the financial risksbecome more complex, which will require domestic and foreign scholars furtherpropose a better model to perfectly portray the changing trends of actual financialmarket price fluctuations. So, the article combined with the latest achievements ofdomestic and foreign financial time series, and propose a new time series model tocapture the rate of return and the volatility of the stock market variation.This article will introduce the double exponential jump factor to traditionalGARCH Models to take advantage of the double exponential jump factor fitting thecharacteristics of the Chinese stock market return series presents the asymmetry andvolatility leverage effect.The empirical results show that the introduction of a doubleexponential jump factor GARCH model compared with the general GARCH-JUMPmodel can better fit the dynamic process of China’s stock market yields andvolatility.In addition, the model of Hong&Li test results also show that the former ismore accurate.However, Hong&Li test results show that the new GARCH modelstill did not pass the test. so we introduce a new model based on variance jumpfactor.The empirical results show that the new model can better fit the nature of theprocess of change in the Chinese stock market yields and volatility, such asLeptokurtic、Volatility clustering、Asymmetry and so on.Besides,the Hong&Li testresults also show that with multiple jump factor GARCH model than with a doubleexponential jump factor GARCH model is more accurate, and better able to portraythe dynamic trend of China’s stock market yields and volatility.The article also made of VaR analysis based on two types of new models on theChinese stock market.The results show that with multiple jump factor GARCH modelpredicts VaR value is better than the double exponential jump factor GARCHmodel.The formerthe failure rate is closer to our confidence level corresponding tothe selected failure rate more accurately reflect the risk profile of the two types ofcomposite index.