Can investors amplify stock return momentum by screening past winners and losers based on return skewness? In their April 2015 paper entitled “Expected Skewness and Momentum”, Heiko Jacobs, Tobias Regele and Martin Webee explore the interaction of expected stock return skewness and momentum. They measure expected skewness as maximum daily return over the preceding month, which predicts future skewness more accurately than does past skewness. Their benchmark is a conventional momentum portfolio that is each month long (short) the fifth, or quintile, of stocks with the highest (lowest) returns from 12 months ago to one month ago. To test the interaction of expected skewness and momentum, they first sort stocks into quintiles based on expected skewness and then sort each expected skewness quintile into quintiles based on momentum (25 total portfolios). Their skewness-weakened momentum portfolio is long (short) winners (losers) with relatively high/positive (low/negative) expected skewness. Their skewness-enhanced momentum portfolio is long (short) winners (losers) with relatively low/negative (high/positive) expected skewness. Using daily and monthly returns for a broad sample of U.S. common stocks (excluding very small and illiquid stocks) during January 1926 through December 2011, they find that:
- Over the entire sample period, value-weighted (equal-weighted) average gross monthly returns are:
- 0.81% (0.93%) for the benchmark portfolio, with value-weighted annualized gross Sharpe and Sortino ratios 0.45 and 0.53, respectively.
- 0.47% (0.21%) for the skewness-weakened momentum portfolio, with value-weighted annualized gross Sharpe and Sortino ratios 0.22 and 0.30, respectively.
- 1.65% (1.90%) for the skewness-enhanced momentum portfolio, with value-weighted annualized gross Sharpe and Sortino ratios 0.70 and 0.73, respectively.
- Gross three-factor (market, size, book-to-market) monthly alphas for the value-weighted portfolios are:
- 0.96% for the benchmark.
- 0.21% for the skewness-weakened momentum.
- 2.14% for the skewness-enhanced momentum.
- 1.87% for the skewness-enhanced portfolio restricted to stocks with market capitalizations above the NYSE median.
- Differences in profitability among the three momentum strategies come largely from the short sides of the respective portfolios. Value-weighted average gross monthly returns of loser sides are 0.61%, 0.82% and -0.37% for the benchmark, skewness-weakened momentum and skewness-enhanced momentum portfolios, respectively.
- Over the 36 months after portfolio formation, effects of expected skewness on momentum strategy profitability do not revert.
- Across available samples for 16 non-U.S. developed stock markets, a one standard deviation increase in the expected skewness of winners and decrease in expected skewness of losers reduces average gross monthly momentum return by average 0.36%.
- Findings hold in the recent past and after controlling for stock/firm characteristics such as past return, volatility and implementation cost variables.
The following chart, taken from the paper, compares gross cumulative performance of six portfolios:
- Market – the broad, value-weighted U.S. stock market.
- Momentum – the value-weighted benchmark momentum portfolio described above.
- Weakened Momentum – the value-weighted skewness-weakened momentum portfolio described above.
- Enhanced Momentum – the value-weighted skewness-enhanced momentum portfolio described above.
- Enhanced Momentum* – Enhanced Momentum scaled according to forecasted stock market variance (since momentum returns tend to be higher in calm market conditions).
- Enhanced Momentum** – Enhanced Momentum* with asset weights dynamically set based on expected return and volatility.
At inception in January 1927, each momentum portfolio consists of a $1 position in the risk-free rate and a corresponding long-short portfolio as specified. Results indicate that the skewness-enhanced momentum portfolios consistently outperform. Gross annualized Sharpe ratios for the six strategies in the order listed in the chart legend are 1.06, 0.70, 1.07, 0.45, 0.39 and 0.22, respectively.
In summary, evidence suggests that investors may be able to boost stock momentum strategy performance by selecting winners (losers) with relatively low/negative (high/positive) expected return skewness.
Cautions regarding findings include:
- As noted, performance statistics are gross, not net. Accounting for portfolio reformation frictions and shorting costs would reduce these statistics. Also, shorting may not be feasible for all specified loser stocks. This caution applies especially since the short sides of portfolios drive performance differences.
- Trying many strategies on the same data introduces data snooping bias, such that the best-performing strategy likely overstates expected performance.
- There is apparently no lag between skewness calculations and portfolio formation, a potentially material implementation issue.
- Data/processing requirements are fairly complex. Investors delegating this work to a fund manager would bear management/administrative fees.
