Pairs trading using Hurst exponent
The study documented in this thesis uses the method called ‘detrended fluctuation analysis (DFA)' to examine the relationship between the companies listed on Standard and Poor (S&P)'s Australian Securities Exchange (ASX) 200 using ten years of data from 2010–2019. With this method based on the calculation of the Hurst exponent of a pair, a dynamic relationship is found in multiple pairs for the price series. The Hurst approach is used to present the result and is compared with a classical trading strategy, such as Engle-Granger's cointegration method of pairs selection. A robustness test is conducted on the returns of the strategy.
This study undertakes an empirical application of this method with a comparison to the classical method of cointegration. This strategy is applied to listed stocks on S&P's ASX 200. Our study's result shows that the Hurst exponent method is a better way to select mean reverting pairs compared to a classical method, such as cointegration. Our study's strategy is also found to perform better than other strategies when the benchmark falls. Finally, the study found that the Sharpe ratio performs better than the classical strategy.