A Research Into Pairs Trading Strategy From The SSE 180 Index Components Based On Regression And Cointegration

The fundamental idea of pairs trading is to find two stocks that have similar characteristics, when the relative price difference of the two stocks exceeds a certain standard, sell the overvalued one, buy the undervalued one. When the relative price difference returns to normal, close the previous deal and get profit from the relative price difference.This paper studied the pairs trading profitability of SSE 180 Index components using regression analysis and co-integration analysis. This paper analyzed the 5-minute stock close price data of all the SSE 180 Index components in 2014 by dividing the data into training samples and simulation samples. The result shows that the pairs trading strategy has achieved very good results in training period; but it did not perform well in simulation period and even the yield in the third simulation phase is negative. That is caused by the sudden changes of market style. It leads the trading strategy parameter that we got in one market style is no longer applicable in another market style. We add the stop loss strategy to avoid the potential loss, then the yield improves well. Empirical results show that the pairs trading strategy based on regression analysis and cointegration analysis can obtain some profits when market style is stable; but when there is a big market style change, we need to adjust the pairs trading timely.The innovation of this paper is that we proved the pairs trading strategy can earn profit based on a large sample which comes from all the combination of all SSE 180 Index componenets. And for the cointegration analysis, we used codes to analyze instead of manually checking that used in many other papers. This can save a lot of time and serve as reference for further research in larger sample.