The Comparative Research Of China Stock Index Futures Risk Measurement Based On The VaR-GARCH Model Under Extreme Market

Since China stock index futures market in April 2010 launched the first futures contract, the market has developed rapidly and run smoothly. In April 16.2015. the SSI:50 and CSI 500 index futures, launched which letting our market further optimized and rich. Nevertheless, the recent emergence of extreme market from July 2014 to June 2015 made the Shanghai and Shenzhen stock market rose sharply. In June 2015. the prices of China’s securities market has experienced a huge shock, which risk management of stock index futures has to be put forward higher requirements. In this case, whether the existing risk measure of the current stock index futures contracts has a good risk measure effect has become a concern of research.Value at risk measurement first proposed by the JP Morgan firm in 1976. to measure the maximum possible loss of the portfolio’s holding period in the future suffer. The variance-covariance. historical simulation and Monte Carlo simulatoin are three methods to calculate the VaR. This paper mainly used the intuitive variance covariance method to make the calculation brief and simple. The GARCH model proposed by Bollerslev in 1986 based on the ARCH model proposed by Engle. This model made the ARCH model extended, depending on the characteristics of the current series. Now IGARCH model, TGARCH model and EGARCH model,etc, all belong to the GARCH series.In this paper, we get the time series of stock index futures under extreme market, with the current mainstream VaR-GARCH model risk measurement, in order to make the analysis and compare results. Article selected CSI 300, SSE 50, CSI 500 stock index futures Log rate of return in extreme market yields as samples sequence. Normality test, unit root test and correlation test of residual square show that three kinds of sequences refusing the normal distribution are smooth sequences with obverse ARCH effects. By comparing with the conventional GARCH model, we found that GARCH (1,1) model is a better fit for the sequence of conditional variance. VaR values were obtained by calculating the three kinds of stock index futures. The actual contract losses in extreme market shows that VaR-GARCH model underestimated the risk situation. Because of the sample sequence refusal of normal distribution and financial data asymmetry typically,the article has compared the GARCH model fitting effect under TARCH model, EGARCH model based on different distributions. It was found TGARCH Model for the CSI 300, SSE 50 based on GED distribution yields better fitting effect to estimate the conditional variance. while the TGARCH mode for the CSI 500 based on Students’ t distribution has a better estimation.To compare the risk measurement effect of CSI 300, SSE 50, CSI 500 index futures contract based on the VaR-GARCH model under extreme market, we analyze the problem by using empirical methods. We also make some useful suggestion for the stock index futures market to strengthen risk management in the future and optimize the risk measurement tool.