Cash-Stock Pairs Trading with Correlation and Cointegration Filters

Summary

This project outlines a mean-reversion strategy for liquid, shortable stocks organized across five sectors. It first screens candidate pairs for correlation, then tests their spread for stationarity with the Augmented Dickey-Fuller test. When a qualifying pair’s price ratio moves far from its historical mean, the strategy buys the relatively cheap stock and shorts the expensive one, closing the position as the ratio converges.

The project describes a minimum one-year training period, out-of-sample evaluation, and independent sector trading, with sector capital allocation based on a minimum-variance portfolio approach. It reports using Yahoo Finance data from 2009 through 2014 and mentions further analysis of model output and Monte Carlo methods, but the article excerpt does not provide their results or the full implementation. Cointegration can support a stable mean-reverting relationship, but the author flags that the relationship may break or shift to a new equilibrium. The described tests and historical period therefore do not establish that a pair will remain profitable in live trading.

Key ideas

• Correlation screens candidate stock pairs, while a stationarity test checks whether their spread may mean-revert.
• The strategy buys the relatively cheap stock and shorts the expensive one when their price ratio departs from its mean.
• Pair selection favors liquid stocks with narrow bid-ask spreads and short availability.
• The project uses out-of-sample evaluation and allocates sector capital with a minimum-variance approach.
• Cointegration can fail when the long-term relationship between two stocks changes.

Tags

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.