A Research Into Fairs Trading Strategy From The CSI300Index Components Based On Cointegration

This paper use Co-integration method to research pairs trading strategy from the CSI300Index components. To avoid the risk of the industry, this paper classifies the CSI300Index components by industry. At the same time, size effect of the stock market in china is taken into consideration. The CSI300Index components are divided in48groups. Then stock portfolios whose correlation is low are excluded. ADF test shows that time series of stocks are all a unit root series. After EG two-step standard, the100pairs of stock portfolios we research are proved have a co-integration relationship. At simulated pairs trading stage, this paper chooses1.5times SD of regression coefficient. In sample period,99pairs of stock portfolios occurs pairs trading. When transaction cost is0.1%, the average abnormal return of pairs trading is29.95%. When the sample period is expanded, number of trades and average abnormal return raise. This paper finds that the transaction cost will increase significantly if the correlation of the stock portfolio is too high. This paper intends to use stock portfolio whose correlation is near0.85as pairs trading stocks.The innovation of this paper is CSI300Index components are divided in48groups by industry. This process assured stock portfolios are industry-neutral. In addition, the comprehensiveness of100pairs of stock portfolios would be of benefit to using hedging of stock index futures in china.