Key Takeaways
- • Mixed results on volatility-adjusted momentum: A VIX-dependent trend-following model's "Top 1" edge persisted through nearly nine years of out-of-sample testing (August 2017–June 2026), but the "Top 2" edge did not survive the extended period, suggesting selective rather than universal applicability of volatility-adjusted lookback windows.
- • Genuine out-of-sample validation period: The test period encompasses major market dislocations including the COVID volatility shock, the 2022 inflation and interest-rate crisis, multiple rapid equity reversals, and diverse market regimes, providing robust real-world validation beyond the original 2017 backtest.
- • Practical ETF implementation tested: The analysis employed transparent, implementable proxies using SPY, VXF, EFA, AGG, and BIL with VIX-derived volatility regime signals, moving beyond theoretical index backtests to address real institutional deployment constraints.
- • Volatility-adjusted momentum hypothesis partially confirmed: The original intuition that shorter momentum windows work better during high-volatility periods while longer windows perform better in calm markets showed selective merit, though the relationship proved less universal than initial research suggested.
Summary
In September 2017, Alpha Architect published research examining whether volatility information could enhance traditional trend-following allocation strategies. The core hypothesis was straightforward: momentum measurement windows should adapt to market conditions, with shorter lookbacks responding more effectively during elevated volatility and longer windows providing stability in calmer environments.
Nearly nine years of subsequent market history now provides an opportunity to validate this original research hypothesis using genuinely out-of-sample data. The extended test period from August 2017 through June 2026 captures an unusually comprehensive range of market environments, including the sharp COVID volatility shock of 2020, the significant inflation and interest-rate dislocations of 2022, multiple rapid equity-market reversals, and extended periods of relative calm. This diversity of conditions strengthens the evidentiary power of the analysis.
Rather than reproducing the original index backtest precisely, the research employed a practical, transparent ETF proxy implementation using SPY, VXF, EFA, AGG, and BIL as the investment vehicles, with volatility regime signals derived from the VIX index. This methodological choice reflects the institutional reality that investment managers implement strategies through available instruments rather than theoretical benchmarks.
The results reveal a nuanced picture. The strategy's "Top 1" edge—the strongest performing variant—demonstrated persistence across the nine-year out-of-sample period, suggesting genuine validity for at least the primary implementation. However, the "Top 2" configuration did not survive the extended period, indicating that volatility-adjusted momentum benefits are not universal across all model variations. This partial validation supports the original intuition about volatility-regime adjustment to momentum measurement but suggests the relationship is more selective and less robust than initial research indicated. The findings underscore both the value and limitations of volatility-informed tactical allocation adjustments for professional investors.
Nearly nine years of subsequent market history now provides an opportunity to validate this original research hypothesis using genuinely out-of-sample data. The extended test period from August 2017 through June 2026 captures an unusually comprehensive range of market environments, including the sharp COVID volatility shock of 2020, the significant inflation and interest-rate dislocations of 2022, multiple rapid equity-market reversals, and extended periods of relative calm. This diversity of conditions strengthens the evidentiary power of the analysis.
Rather than reproducing the original index backtest precisely, the research employed a practical, transparent ETF proxy implementation using SPY, VXF, EFA, AGG, and BIL as the investment vehicles, with volatility regime signals derived from the VIX index. This methodological choice reflects the institutional reality that investment managers implement strategies through available instruments rather than theoretical benchmarks.
The results reveal a nuanced picture. The strategy's "Top 1" edge—the strongest performing variant—demonstrated persistence across the nine-year out-of-sample period, suggesting genuine validity for at least the primary implementation. However, the "Top 2" configuration did not survive the extended period, indicating that volatility-adjusted momentum benefits are not universal across all model variations. This partial validation supports the original intuition about volatility-regime adjustment to momentum measurement but suggests the relationship is more selective and less robust than initial research indicated. The findings underscore both the value and limitations of volatility-informed tactical allocation adjustments for professional investors.
Go deeper than the headline
Research every firm in this story, track the ones that matter and get the daily briefing — free for 7 days.