The Research Of Statistical Arbitrage Based On The High-Frequent Data Of Index Future And Empirical Analysis

Statistical arbitrage has been used as a common strategy of investigation in the developed capital market as a mature strategy of quantitative trad-ing.The issue of stock index futures and securities margin trading make it possible to exert shorting mechanism and become valid instruments to carry out statistical arbitrage in domestic market.Statistical arbitrage is a way to implement arbitrage using the character of mean-reverting of spread calculated by assets with stable co-integration relationship on the basis of market-neutral hypothesis.Investors only need to select appropriate assets portfolio and buy or sell the assets based on the trend of spread instead of to forecast the trend of the price charts of the assets transacted.Therefore, the key of statistical arbitrage is to find out characters of spread series of the trading portfolio to establish trading signals and to carry out arbitrage trading according to the trading rules instituted.High-frequent data makes it possible to transact several times one day. In high-frequent trading, time of position portfolio is shorter and profits is smaller in each act, however, investors can accumulate the scanty profits to abundant profits according to the substantial trading capacity using low-delayed assets.In this paper, several statistical arbitrage models are introduced including constant variance model, GARCH model with time-varying variance, Ornstein-Uhlenbeck stochastic process with stochastic volatility and statistical arbitrage model based on realized volatility.In empirical analysis, this paper tries to an- alyze all the models mentioned in detail using calendar spread arbitrage be-tween IFLO and IFL1 of HS 300 index futures as an example. In addition,5 minute、15 minute and 30 minute frequent data are used to compare the re-sults of statistical arbitrage model on different frequency. Parameters calculat-ed in the sample are used to establish models outside the sample. Conclusions can be drawn through analyze the indicators like times of successful statistical arbitrage、yield rate、yield volatility, etc.