The issue of pairs trading is regarded as a sort of trading strategy, which groups two targets with certain limitations and compares these targets. When the prices of these two targets are deviate, they are anticipated to be convergent. Therefore, investors are able to buy targets with relatively low price and sell them with higher price in order to minimize system risks and get benefits from the arbitrage behavior.
This paper chose the sample from Taiwan Futures Exchange on Jan. 25, 2010. The categories are financial group and non-financial group. The data of pairs trading divides financial group into three sub-groups by the conditions of the 60% similarity of products, ownership structure and correlation coefficients above 0.6. Alternatively, the non-financial group is divided by the conditions of the same industry and the 60% similarity of products. With the aspects of buy-and-sell signals, the RSI is employed for the technical analysis. Hence, the technique of pairs trading is used to examine if investors can have abnormal returns by using this method. Secondly, with pairs trading, the data can be investigated whether or not they have the relationship of cointergration. Therefore, the investors can have higher returns by using pairs trading. Finally, if there are higher correlation coefficients between any two target stocks and to see if target stocks with higher correlation coefficients contribute higher returns.
It indicates that the ranges of annual returns are somewhere fluctuated, with the highest 5.71% and the lowest -0.79%. Comparing with the annual returns of benchmark, some pairs trading are actually resulted in higher returns. Regarding with the cointergration, those pairs with the relationship of cointergration have higher returns compared with others which are without the cointergration. Another finding concerning the correlation coefficients between any two target stocks is found that the returns with lower and higher correlation coefficients are significantly different.
