Empirical Research On High-frequency Statistical Arbitrage Based On Shanghai-hong Kong Stock Connect A+H Shares

As a neutral market strategy,statistical arbitrage does not depend on the judgment of market trends.It can perform arbitrage operations in both bull and bear markets,and it can obtain stable gains at lower risk.With the emergence and development of HFT,the statistical arbitrage strategy is widely favored by foreign HFTs due to these advantages,and thus develops into a typical HFT strategy.For companies listed on the A-share and H-share markets at the same time,due to the same fundamentals,there is a probability of arbitrage between A shares and H-share prices.With the opening of the "Shanghai-Hong Kong Stock Connect" in April 2014,domestic and foreign investors have more channels to invest in the A-shares and H-shares markets,which enhances the feasibility of implementing statistical arbitrage strategies between A shares and H shares of these dual listed companies.Based on the background,this paper first briefly analyzes the domestic and foreign research of statistical arbitrage,and then introduces related theories and models such as"Shanghai-Hong Kong Stock Connect",statistical arbitrage and high-frequency trading.Then,we use A-shares and H-shares of SH-HK Stock Connect A+H-listed companies as trading object and design a rolling-window dynamic statistical arbitrage strategy.In this strategy,we use the cointegration theory to determine the arbitrage ratio and use the GARCH model to determine the time-varying standard deviation of spreads.The maximum total yield is used as the target function to dynamically optimize the opening threshold,and the stop loss signal is designed based on VaR.Subsequently,this paper empirically analyzes the dynamic statistical arbitrage strategy.After preliminary screening and correlation analysis of the prices of A-shares and H-shares of all "SH-HK Stock Connect" A+H-listed companies,we select A-shares and H-shares of Conch Cement which has the highest correlation coefficient as empirical trading targets,and its 5-mimute,15-minute,and 30-minute day high-frequency closing price series are used as sample data to perform dynamic statistical arbitrage strategy backtesting.The empirical results show that there is indeed a high-frequency statistical arbitrage opportunity between the A shares and H shares of "SH-HK Stock Connect" A+H listed company.In addition,we also found that 30-minute high-frequency data is more suitable for this dynamic statistical arbitrage strategy,indicating that the frequency of data used for high-frequency statistical arbitrage strategy is not the higher the better,and different frequency need to consider different rolling window length in the rolling statistics arbitrage.Finally,based on the summary of the research results,this paper puts forward some suggestions for the supervision of "SH-HK Stock Connect" and makes improvements and prospects for the inadequacies of the article.