Cointegration-Based Pairs Trading in Mexican Equities

Summary

This project tests a mean-reversion pairs strategy on Mexican stocks. It screens an initial equity universe for complete price histories and minimum average trading volume, then tests within-industry pairs for cointegration with an augmented Dickey-Fuller test and filters for correlation. For each remaining pair, it calculates a price-ratio z-score using a rolling 60-day mean and standard deviation. Positions short the relatively expensive stock and buy the cheaper one when the ratio moves above the mean, with the reverse trade below it. Entries are staged at two deviation thresholds, and trades exit when the z-score returns to zero.

The study divides its 2012–2017 test period into in-sample and out-of-sample windows. It reports pair-level performance variation and a higher in-sample portfolio Sharpe ratio when both entry stages are used, while out-of-sample results are lower. It also examines pair correlations and compares the strategy’s in-sample relationship with Mexico’s market index. These findings are historical and do not establish future profitability. The author notes that trading costs are excluded, short availability is assumed, and additional testing across periods, parameter choices, and exit rules—including stop losses—is needed.

Key ideas

• The strategy seeks mean reversion in price ratios of cointegrated Mexican equity pairs.
• Pairs are filtered using data completeness, trading volume, an augmented Dickey-Fuller test, and correlation.
• A rolling 60-day z-score guides staged entries, while a return to zero triggers exits.
• The study reports stronger in-sample portfolio results with both entry stages, alongside weaker out-of-sample results.
• Trading costs, short-sale availability, parameter sensitivity, and performance across other market conditions remain unresolved.

Tags

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