The Application Of Paired Trading Strategy Based On GARCH Model In The Constituent Stocks Of Shanghai And Shenzhen 300 Index

From the birth of the first stock issuance by the joint-stock reform to the comprehensive promotion of the registration system in the A-share market,in just over40 years,the domestic capital market has experienced rapid development from scratch,from the region to the whole country.At present,China’s capital market has entered a stage of high-quality development,and its structure has gradually moved closer to the international mature market.The overall operation of the market is stable,attracting more and more small and medium-sized investors into the financial market.As a market neutral strategy based on mean regression theory,paired trading strategy can achieve statistical arbitrage on the basis of maintaining low risk,which is suitable for risk-sensitive small and medium-sized investors.The effect of paired trading is mainly affected by the screening of paired stocks,the setting of trading signals,and the specific operations taken when the trading signals are triggered.Therefore,based on the sample data of 5 minutes,10 minutes and 15 minutes,this paper will compare and analyze the arbitrage results when using different trading thresholds and trading methods,and give suggestions for strategy selection.The implementation of pairs trading in this paper is mainly divided into four stages.In the first stage,this paper quantitatively screens paired stock pairs based on correlation coefficient,cointegration test and shortest distance to obtain spread sequence and trading position.In the second stage,this paper will use two methods to determine the trading signal.One is to use the traditional method to set the opening threshold to±6)σbased on the standard deviationσof the spread sequence during the training period.The other is to use the GARCH model method.Based on the time-varying standard deviation,the opening threshold is set to±6).The influence of sample data frequency and model selection on arbitrage results is compared and analyzed when different standard deviations are selected.In the third stage,the trading method when using the GARCH model is optimized,and the strategy arbitrage results before and after optimization are compared,and the strategy selection suggestions are given.In the fourth stage,this paper tests the robustness of the model from two aspects:the parameter k of the opening threshold±6)σand the sample interval,and gives the risk prompt on the basis of the strategy selection suggestion.When analyzing the arbitrage results,this paper comprehensively evaluates the indicators such as yield,maximum withdrawal,Sharpe ratio,and winning rate,and draws the following conclusions:First,the use of paired transactions to complete arbitrage has certain effectiveness and stability;secondly,the GARCH model is introduced to obtain the time-varying standard deviation.Setting the trading threshold can optimize the arbitrage results,and the choice of the parameterin the trading threshold±6)is an important factor affecting the strategy income.Thirdly,the optimization of trading methods after considering transaction costs,margin trading restrictions and other issues is effective;fourthly,it is recommended that investors choose 10-minute frequency data to complete the pairing transaction when using the GARCH model,and choose 5-minute frequency sample data when using the optimized model.Fifth,it is recommended that investors pay attention to the risk of parameter6)in the transaction threshold and the change of the transaction interval when using the pair trading strategy.