Pairs Trading Strategies Based On The GARCH Model

Pairs trading strategy is a statistical arbitrage strategy based on the stock price fluctuation.The strategy is to make gains by buying stocks that are relatively undervalued and shorting stocks with relatively high prices.But in the current market conditions,there are high threshold of margin account,high costs,limited investment targets,and many other restrictions.They restrict the pairs trading strategy implementation.According to the actual situation of China's market,this paper improves the design of the pairs trading strategy from two aspects.In the way of transaction,unilateral transaction is used instead of bilateral transaction.In the way of transaction rules,the GARCH model is used to describe the dynamic variance of the spread series.In this paper,we select the data from January 1,2007 to December 31,2016 to simulate transaction.First of all,the momentum trading strategy was used to select the first 300 of the stock in the two years of training period as the stock pool,and the correlation coefficient of the top 5 stocks was selected as the stock of the research.Secondly,the cointegration model and error correction model are set up to test whether there is a long-term stable equilibrium relationship between the pairs of stocks and short-term deviation from the long-term equilibrium relationship.Then the GARCH model is established to simulate the dynamic variance.In this paper,we choose 0.75 times the standard deviation as the signal of liquidation and 3 times the standard deviation as the signal of stop-loss.Finally,we simulate the transaction according to the trading rules.The simulated trading results of the eight trading period show that the the cumulative return of unilateral transaction pairs trading strategy was 290.27%.The SHARP ratio is 0.9349.Over the same period,the cumulative return of sample stocks was 251.35%.The SHARP ratio is 0.5948.The cumulative abnormal return of the pair trading strategy is 38.92% and the average annual excess return is 4.87%.On the whole,whether it is from the absolute income or the expected return per unit risk,improved pair trading strategy has a very good performance.The contribution of this paper is reflected in two aspects.In the way of transaction,unilateral transaction is used to replace the bilateral transaction and the limted application of bilateralt ransaction in China's securities market is avoided.In the formulation of trading rules,the GARCH model is used to describe the dynamic variance of the spread sequence.So the variance in the trading period is closer to the actual value of variance.The improved pair trading strategy is more suitable for the actual situation of China's securities market and has higher practical value.