Do U.S. equity exchange-traded funds (ETF) exhibit long-term momentum? In their October 2019 paper entitled “ETF Momentum”, Frank Li, Melvyn Teo and Chloe Yang investigate future performance of U.S. equity ETFs sorted on multi-year past returns. Each month starting August 2004, they:
- Sort selected ETFs into tenths (deciles) based on returns over the past two, three or four years, with focus on three years.
- Reform an equal-weighted (EW) or value-weighted (VW) portfolio that is long (short) the decile with the highest (lowest) past returns, with focus on value-weighted.
They then evaluate performances of deciles and long-short portfolios based on raw return, 4-factor (adjusting for market, size, book-to-market and momentum) alpha and 5-factor (replacing momentum with profitability and investment) alpha. Using monthly returns, market capitalizations and net asset values for all U.S. equity ETFs with capitalizations greater than $20 million and share price greater than one dollar during August 2000 through June 2018, they find that:
- At the start (end) of the sample period, there are 50 (396) qualifying U.S. equity ETFs.
- In general (see the chart below):
- Long-short multi-year momentum portfolios generate substantially positive returns.
- The optimal long-short VW portfolio based on past 3-year returns generates average gross monthly return 1.25%, gross monthly 4-factor alpha 1.20% and gross monthly 5-factor alpha 1.28%.
- EW long-short portfolios materially underperform VW portfolios, suggesting that the largest ETFs drive the effect.
- However, returns are volatile, discouraging to traders with short investment horizons.
- Neither cross-sectional stock momentum, macroeconomic variables nor liquidity explain multi-year ETF momentum.
- Findings are generally robust to:
- Estimated monthly portfolio rebalancing frictions.
- Extending the long-short portfolio holding period from one to three, six, nine or 12 months (with overlapping portfolios).
- Using returns calculated from ETF net asset values (NAV) rather than ETF prices.
The following chart, constructed from data in the paper, summarizes performance of long-short multi-year ETF momentum portfolios for both weighting schemes and all three lookback intervals. Results show that the optimal combination is VW with 3-year lookback interval.
In summary, evidence suggests that U.S. equity ETFs exhibit multi-year momentum.
Cautions regarding findings include:
- The sample period is extremely short for testing the specified strategy. For example, the full sample period has fewer than five independent 4-year lookback intervals.
- Average monthly return is not sufficient to assess strategy attractiveness. Many investors are concerned about return variability (Sharpe ratio) and maximum drawdown.
- As tested, portfolios hold too many assets in later years for most individual investors.
- Testing multiple lookback intervals and weighting schemes on the same data introduces data snooping bias, such that the best results (VW with 3-year lookback) overstates expectations.
- The specified setup may not work for ETFs tracking asset classes other than U.S. equities.
Frequently Asked Questions
What is multi-year ETF momentum?
Multi-year ETF momentum is the tendency of U.S. equity exchange-traded funds with the highest returns over the past two to four years to keep outperforming those with the lowest past returns. In the study, ETFs are sorted into deciles each month on past-return lookbacks of two, three or four years, with focus on three years. A long-short portfolio holds the top decile and shorts the bottom decile.
Does past 3-year performance predict future returns for U.S. equity ETFs?
Yes, past 3-year returns predict future U.S. equity ETF returns in the study, which covers August 2004 to June 2018. The optimal value-weighted long-short portfolio based on 3-year past returns generated an average gross monthly return of 1.25%, a 4-factor alpha of 1.20% and a 5-factor alpha of 1.28%. The effect held up after estimated rebalancing frictions, longer holding periods of up to 12 months, and use of net asset value returns, but returns were volatile.
How is an ETF momentum strategy built and tested?
Each month starting August 2004, U.S. equity ETFs are sorted into deciles by past two-, three- or four-year returns, and an equal-weighted or value-weighted portfolio goes long the top decile and short the bottom decile. Performance is evaluated using raw return, 4-factor alpha (market, size, book-to-market, momentum) and 5-factor alpha (market, size, book-to-market, profitability, investment). The sample includes U.S. equity ETFs with capitalization above $20 million and share price above $1, growing from 50 to 396 ETFs.
Is ETF momentum the same as stock momentum, and why do value-weighted and equal-weighted ETF momentum portfolios differ?
ETF momentum is not explained by cross-sectional stock momentum, as neither stock momentum, macroeconomic variables nor liquidity explain multi-year ETF momentum. Equal-weighted long-short portfolios materially underperform value-weighted portfolios, suggesting the largest ETFs drive the effect. The best combination in the study is value-weighted with a 3-year lookback.
What are the limitations and risks of an ETF momentum strategy?
The ETF momentum strategy has several limitations: the sample period is extremely short, with fewer than five independent 4-year lookback intervals, and returns are volatile. Average monthly return alone does not capture Sharpe ratio or maximum drawdown, and the portfolios hold too many assets in later years for most individual investors. Testing multiple lookbacks and weighting schemes on the same data introduces data snooping bias, so the best result (value-weighted, 3-year) likely overstates expectations, and results may not apply to ETFs tracking asset classes other than U.S. equities.