Algorithmic Trading: Winning Strategies and Their Rationale

CHAPTER 4

Mean Reversion of Stocks and ETFs

The stock market is, in a sense, the most fertile ground for finding mean-reverting instruments and for the application of those basic mean reversion trading techniques described in the previous two chapters. In theory, we can form pairs of stocks belonging to any sector and expect them to cointegrate due to their exposure to many common economic factors. Their number is large, so diversification is easy. In practice, though, there are some serious difficulties with applying these generic techniques to trading stocks and ETFs. This chapter will examine issues specific to stocks and ETFs. I will also demonstrate that simple mean-reverting strategies actually work better for ETF pairs and triplets.

But we need not limit ourselves to those strategies described in Chapter 3 when looking for mean reversion in stocks or ETFs. We find that in the short term, most stocks exhibit mean-reverting properties under normal circumstances. (Normal circumstance means there isn't any news on the stock, a topic that is taken up in Chapter 7.) This is despite the fact that stock prices follow geometric random walks over the long term. We will build a strategy to exploit this short-term, or “seasonal,” mean reversion.

Index arbitrage is another familiar mean reversion strategy. In this case, we are counting on the cointegration of stocks versus futures or stocks versus ETFs. Because little profit is left using the traditional implementation of index arbitrage, ...