Design Of Pair Trading Strategy Based On Hurst Index

The paired trading strategy is a market-neutral statistical arbitrage investment strategy.It builds a long-short position between two stocks to earn a return on the convergence of the spread.The Hurst index is usually used to study the long memory of the time series,but when the Hurst index is in the(0,0.5)interval,the time series has anti-persistence which is similar to the mean recovery process.Therefore,this article attempts to start from the anti-persistence of the Hurst index,combined with the actual of Chinese A-share matching trading strategy.This article uses the CSI 300 Index Constituents as the research object.First,we use correlation and cointegration tests to initially screen out stock pairs with strong correlation and stable spreads.Then use the generalized Hurst index method to calculate the Hurst index of the time series of stock pair spreads,and select Hurst.The five stock pairs with the smallest index constitute a portfolio to participate in pair trading.In the process of trading signal design,based on the characteristics of strong anti-sustainability of stocks and the relatively rapid frequency of the direction of the spread sequence,a take-profit operation was added to the trading signal,and the time-varying Hurst index was applied to the trading signal.When the time-varying Hurst index is greater than a certain threshold,the anti-continuity weakens,even if the signal to open a position is triggered,no position opening operation will be performed.Combining with the fact that Chinese market cannot be directly short-selling,this article designs two trading strategies that use securities lending business to borrow the underlying stock and sell it or make two paired stocks in advance for equivalent paired trading.Through the two-year back-testing of the two trading strategies in a total of 4 trading ranges,the following conclusions were obtained:(1)The Hurst index of the most preliminarily screened paired stock spread series is less than 0.5,indicating that the vast majority of the spread time series have contradictions Continuous characteristics.Although the constraints of the time-varying Hurst index on opening operations will reduce the number of opening positions,the reduced number of loss-making positions accounts for a greater proportion.(2)The two trading strategies based on the Hurst index can achieve cumulative returns that exceed the benchmark of the CSI 300 index.The securities lending strategy continues the market-neutral characteristics of traditional pairing transactions,with lower cumulative returns but smaller drawdowns.The strategy of doing long in advance loses market neutrality,and the cumulative return is higher but the risk of retracement is higher,which proves the feasibility of using Hurst index to select stock pairs in the strategy.(3)Compared with the classic co-integration method,the securities lending strategy can obtain higher returns than the classic method after traversing the parameters.