A Stochastic Spread & Co-integration Approach to Pairs Trading in a Regime- Switching Environment

This paper applies two statistical arbitrage algorithms on the U.S. equities market, using daily historical prices from January 2005 to March 2012. The algorithms construct portfolios using two different frameworks, namely, the Vasicek model and co-integration approach, with a Markov regime-switching component. The empirical results show that the strategies deliver annualised returns of 12 and 10 per cent and Sharpe ratios of 0.61 and 1.49, which is found to be superior to a comparable benchmark, i.e., the broad U.S. stock index Russell 3000. Our findings corroborate previous literature on the topic of pairs trading, and in addition to this, extends the literature by introducing a regime-switching component in the co-integration framework for pairs trading