Country ETF Pairs Trading with Distance-Based Selection

Summary

The document describes a mean-reversion strategy for international equity ETFs. It builds dividend-inclusive cumulative return series normalized to a common starting value, ranks candidate pairs by the sum of squared differences between their normalized histories, and trades the closest pairs. Positions go long the relatively weaker ETF and short the stronger one when their spread moves beyond a stated fraction of its historical standard deviation. The portfolio holds up to five pairs, equally allocated, and closes positions after convergence or at the end of the trading window.

The implementation uses a 120-day formation history and a 20-day trading period, with daily monitoring and rebalancing. The code also shows data freshness checks, leverage, and a proportional transaction fee model. It gives no performance results or statistical validation. The examples and rules have potential implementation inconsistencies: the comments describe a two-standard-deviation entry while the condition uses 0.5, and the spread normalization and ordering may not match the stated formation-period distance method. Traders would need to validate timing, signal construction, costs, and out-of-sample behavior before relying on it.

Key ideas

• Pairs are selected by minimizing squared distance between normalized ETF price histories.
• The strategy trades up to five pairs, going long the relatively underperforming ETF and short the outperforming one.
• Positions close when the spread reverts or when the trading window ends.
• The document provides code but no evidence of profitability or robustness.

Tags

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