Research On The Influence Of GARCH-like Model Selection On Risk Identification And Measurement Of Stock Index Futures Market

China's stock index futures market started late and is still in the initial stage of development.Futures contract prices are very vulnerable to the impact of external information and occur discontinuous jumps.However,the commonly used GARCH family models assume that the prices of financial assets are continuous and they cannot capture the jump characteristics of futures.Therefore,this paper uses the 5-minute high-frequency data of Shanghai-Shenzhen 300 index futures from April 16,2010 to October 31,2017,to estimate the volatility of futures market by constructing nine general GARCH models and four jump GARCH models.Estimation shows that GARCH family models can well describe the autocorrelation and conditional heteroscedasticity of stock index futures,but only jump GARCH can depict the dynamic jump of futures prices on this basis,which combines the GARCH effect and jump characteristics.At the same time,the results of jump model also show that the jump intensity(number)of futures also has aggregation and persistence.The jump intensity of futures is mainly affected by the jump intensity of the previous period,and also by the slight impact of the jump intensity error.The response of jump amplitude to positive jump and negative jump is asymmetrical.The positive jump has more impact on the value of jump amplitude than negative jump,while the jump amplitude is mainly affected by the positive impact of futures volatility.Seven kinds of statistics(MSE,MAD,QLIKE,HMSE)are used to test the above models.The results show that the fitting effect of jump GARCH model is generally better than that of GARCH family model.In jump GARCH model,the extended ARJI-R2-GARCH and ARJI-ht-GARCH can estimate volatility optimally.On the basis of choosing the optimal model,we estimate the 5-minute Va R of China's futures market with Va R risk measurement method,and compare the estimated results with those of other models.The results show that GARCH models tend to overestimate Va R,which leads investors to choose too conservative investment strategies.Jumping GARCH can not only better estimate the actual Va R volatility curve,but also more moderately estimate the Va R value.Finally,according to the above conclusions,this paper gives the government and relevant departments,financial institutions,risk managers and investors corresponding references.The main contribution of this paper is to use the GARCH model with time-varying jump to estimate the price of futures market in China,and use seven statistical indicators to evaluate the results of the model estimation.Finally,the optimal model is selected,and the systematic risk in the systematic risk of futures market is estimated by using the popular Va R method.This is not only conducive to our in-depth understanding of the systematic risk of China's futures market,but also can provide an important theoretical basis and reference for the government to carry out the systematic risk of the futures market.