Evaluation Of A Cointegration-Based Pairs Trading Strategy

• Main Author: VIJVERMAN, ARNO
• Format: Theses and Dissertations NonPeerReviewed
• Published: [Yogyakarta] : Universitas Gadjah Mada 2013
• Subjects: Marketing; Trade
• Online Access: https://repository.ugm.ac.id/121454/; http://etd.ugm.ac.id/index.php?mod=penelitian_detail&sub=PenelitianDetail&act=view&typ=html&buku_id=61541
• Institution: Universitas Gadjah Mada

Pairs trading is a strategy that takes advantage of the temporary mispricing of two assets with a long-run equilibrium. When the assets diverge, the relatively undervalued asset is bought and the relatively overvalued asset is sold. Because profits don't depend on the movement of the market, the strategy is considered market neutral. We back test a pairs trading strategy based on cointegration using daily closing prices of all S&P 100 stocks over the period 1985-August 2012. We give an intuitive explanation of the theory and methodology and we discuss the empirical results. We find that a simple cointegration-based pairs trading strategy with conventional trading rules and time horizons can generate statistically significant annualized excess returns between 6% and 7% after accounting for transaction costs, with only half the standard deviation of the S&P 500. In times of high market volatility, such as the 2008-2009 global financial crisis, the strategy can generate annualized excess returns as high as 30%. Very low betas indicate the pairs portfolios are in fact market neutral. We also apply the strategy to S&P 500 stocks between 2003-July 2012. We analyze the performance of this broader stock sample over the last ten years and apply modifications to the algorithm. We conclude by suggesting how to incorporate fundamental factors in the pairs selection process to make the strategy more effective.