| As China’s financial system continues to develop and improve,margin financing and securities lending is officially launched on October 31,2010,which means that China’s financial market is a short-lived era.As a quantitative investment strategy,statistical arbitrage strategies will be implemented in China.For the possibility,this article first selected the 5min high-frequency trading data of the Shanghai Composite 50 Index and its constituent stocks.In the previous period,the data was cleaned and screened.Finally,45 constituent stocks were left,the sample data was divided into in-sample data and out-of-sample data,and the in-sample data and out-of-sample data were determined.The cointegration model and the principal component model were used to select the appropriate stock portfolios for the Shanghai 50 Index and its constituent stock data,and the China Merchants Bank and Baosteel Co.,Ltd.were selected as stock portfolios using the cointegration model.The stocks selected using the principal components were China Ping An and the Shanghai 50 Index are used as asset portfolios.The ARCH model and the OU process are used to fit the volatility of the spreads between asset portfolios.In the sample,the OU process parameters calculated based on the fixed standard deviation method and the data in the sample are used to determine the trading signals,so as to set a reasonable arbitrage interval,and outside the sample,the time-varying standard is fully taken into account.The difference is used to determine the time-varying trading signal,and the time-varying signal is determined by the autoregressive conditional heteroskedasticity model and the dynamic OU process,and finally the time-varying trading interval is determined to implement the arbitrage strategy.By comparing the arbitrative utility of the cointegration model and the principal component model,it can be known that using the principal component model to perform arbitrage within the sample based on constant trading signals and sample coverage under the time-varying trading signal is better than The cointegration model is better.The analysis can be based on the fact that the principal component model is a one-to-many long-term equilibrium relationship.The stability of the model itself is strong,while the cointegration model is a oneto-one long-term equilibrium relationship,which occurs in the external market.During the change,due to its relatively poor stability,it leads to worse performance than the arbitrage strategy based on the principal component.At the same time,it gives investment advice to investors and is more suitable for OUs under constant trading signals for risk-averse investors.The ARCH models under the process and time-varying trading signals are two strategies.For risk-neutral investors,they are more suitable for fixed-difference method and dynamic OU process under time-varying standard deviation for investment. |
