Applying Hurst Exponent in pair trading strategies on Nasdaq 100 index

This research aims to seek an alternative approach to stock selection for algorithmic investment strategy. We try to build an effective pair trading strategy based on 103 stocks listed in the NASDAQ 100 index. The dataset has a daily frequency and covers the period from 01/01/2000 to 31/12/2018, and to 01/07/2021 as an additional out-of-time data set. In this study, Generalized Hurst Exponent, Correlation, and Cointegration methods are employed to detect the mean-reverting pattern in the time series of a linear combination of each pair of stock. The result shows that the Hurst method cannot outperform the benchmark, which implies that the market is efficient. These results are quite sensitive to varying number of pairs traded and rebalancing period but they are less sensitive to financial leverage degree. Moreover, the Hurst method is better than the cointegration method but is not superior as compared to the correlation method.