Analysis Of Quantitative Investment Statistical Arbitrage In Stock Time-sharing Trading Based On Co-integration Analysis

Margin trading business and stock index futures are officially introduced into Chinese securities market in 2010.Their launch signifies the realization of short mechanism,making statistical arbitrage get rapid development in our country.At the same time,they also provide the possibility of statistical arbitrage strategy.Therefore,it has strong practical significance to study statistical arbitrage strategy.This paper first selects 26 stocks of good fluidity and high turnover rate of brokerage sector and makes a correlation analysis of 26 stocks' prices with daily data,60 minutes,30 minutes,15 minutes,5 minutes and 1 minutes data.And it chooses a stock combination of northeast securities and guangfa securities as the research object of statistical arbitrage whose stock prices have the maximum correlation coefficient among these six frequency sample.It mainly uses 6 kinds of frequency data to conduct the empirical analysis of statistical arbitrage based on co-integration model,studies the statistical arbitrage opportunities that possibly exist in the day and intraday,and the effects of data frequency on the arbitrage strategy is compared.The results show that it is effective to use the different frequency data to carry out statistical arbitrage in China's stock market.Considering that political reform and institutional change may lead to structural transformation in time series data between economic variables,the variable structure co-integration model is a solution to the change of structure,so this model is innovatively introduced,the paper mainly studies the impact of a variable structure point and two variable structure points on arbitrage performance.Considering that the parameters in the model are generally time-varying,thus introducing the state space model.According to these three models to determine the point of sale,looking for statistical arbitrage trading opportunities,analysis of statistical arbitrage profits and comparing the arbitrage performance.Research results suggest that variable structure co-integration model and state space model used in statistical arbitrage strategy are effective,these two models are better than that of ordinary co-integration model,the prediction effect of the two structural change points out of sample is superior to a variable structure point,appearing more opportunities of arbitrage trading and annualized return also increased.Variable structure co-integration model and state space model have their own advantages.