The Research On Future-Spot Arbitrage Of CSI 300 Index

It has been five years since the first stock index futures of our country "CSI300 index futures"appeared on the market on April,16th,2010.With investors' constant deepening of their understanding in stock futures,there is an increasing number of investors who are engaged in the arbitrage of CSI300 index futures,and such arbitrage traders have become the power that can't be ignored in the market.Future-spot arbitrage,on one hand,plays an extremely important role in maintaining the stability of market price,eliminating incorrect pricing and providing market stability;on the other hand,it can acquire arbitrage earnings with low risks under the situation of guaranteeing good financial liquidity,which is of huge attraction to investors with lower risk preference.Therefore,to study and analyze the influencing factors of future-spot arbitrage and relevant arbitrage technology contributes to improving the arbitragers' investment level and enhancing the efficiency of market arbitrage.By summarizing the previous academic research results of stock index futures and combing with the author's internship experience in some large domestic comprehensive securities operating company,this paper analyzes the key points and all kinds of influencing factors of CSI300 index arbitrage according to the practical market situation in our country.In the part of empirical study,this paper selects two key factors-tracking error and system of mandatory liquidation to analyze specifically.In the research on the tracking error of spot,it respectively studies "the method of ETF" and"Fully Replication Method" constructed by spot portfolio,compares the tracking effect of ETF portfolio and CSI300ETF constructed with the optimizing model in the method of ETF,and concludes that the tracking error of ETF portfolio is slightly bigger than that of CSI300EFT,but by discussing the two-sidedness of tracking error's influence on arbitrage,the tracking error of ETF portfolio is found to be beneficial to future-spot arbitrage sometimes.By introducing the approach of alternative stock into fully replication method,it finds that this approach will give rise to large tracking errors under extreme market through historical back-test,and such a tracking error is caused by the failure in the relevance between alternative stock and the replaced one,from which relevant improvement project is proposed and verified to be effective with data.Aiming at the arbitrage risks brought by system of mandatory liquidation,author proposes to measure the risks of futures asset in arbitrage portfolio with value at risk(VaR),and brings in GARCH model family to model for the volatility of CSI300 index futures.In the studying process,it discovers that there is obvious conditional heteroskedasticity in the earnings rate of CSI300 index futures,but the "risk premium" and "asymmetric effect" found in traditional studies don't exist,while the traditional GARCH(1,1)model has already been able to well describe its fluctuation characteristics.In the meantime,two assumptions of normal distribution and t-distribution of error term is made to model and estimate VaR value separately.In the end,by comparing the effect of models under two distributions with Kupiec test method,it finds out that t-distribution can better reflect the earnings rate's characteristic of leptokurtosis and fat-tail,and the VaR value calculated by GARCH(1,1)model under t-distribution has better predicative effect and robustness.