Testing Mean-Reversion Pairs Across Three ETF Sectors

Summary

This project tests a simple ETF pairs strategy in oil, technology, and financial sectors: USO with XLE, XLK with IYW, and XLF with PSCF. It estimates a hedge ratio by regression, evaluates spread stationarity with an Augmented Dickey-Fuller test, then enters when the spread exceeds 1.5 rolling standard deviations over a 120-day lookback and exits when it returns to its mean or moving average. The report presents in-sample results from 2011–2014 and an out-of-sample oil-pair test for 2015.

USO/XLE produced the strongest reported in-sample performance, but failed the stationarity test out of sample and generated only one trade in that period. XLK/IYW had poor backtest results, while XLF/PSCF passed the reported test but underperformed the benchmark. These mixed results illustrate that correlation or a favorable stationarity test alone does not establish profitability; spread behavior and trading opportunities matter. The study is limited by its small set of pairs, short test periods, and omission of transaction costs, so its reported returns are not evidence of robust live performance.

Key ideas

• The project tests mean reversion in three ETF pairs using regression hedge ratios and Augmented Dickey-Fuller tests.
• The strategy enters after a spread moves beyond 1.5 rolling standard deviations and exits near its mean.
• USO/XLE performed best in sample, but failed the reported stationarity test in the 2015 out-of-sample period.
• XLK/IYW underperformed despite the ETFs' strong linear relationship, while XLF/PSCF passed a test but still lagged the benchmark.
• Transaction costs were excluded, and the limited number of pairs and trades restricts the conclusions.

Tags

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