What about all those criticisms of momentum investing (such as high turnover/trading frictions and crash-proneness)? In the May 2014 draft of their paper entitled “Fact, Fiction and Momentum Investing”, Clifford Asness, Andrea Frazzini, Ronen Israel and Tobias Moskowitz summarize research on the momentum anomaly and rebut ten criticisms (myths) of momentum investing. Specifically, they address claims that momentum profitability is too small, too volatile/crash-prone, works mostly on the problematic short side, works well only among small stocks and does not survive trading frictions. They focus on a “standard” definition of momentum as the past 12-month return, skipping the most recent month‘s return (to avoid microstructure and liquidity biases). Using results from widely circulated and debated academic papers and data from Kenneth French‘s website, they conclude that:
- Raw and risk-adjusted momentum factor premiums are generally larger than size and value premiums. During 1991-2013, the similarly estimated annualized gross size, value and momentum premiums (Sharpe ratios) are 3.3%, 3.6% and 6.3% (0.29, 0.32 and 0.36), respectively.
- The momentum long side is at least as profitable as the short side whether using 86 years of U.S. stock data, 40 years of international equity data or 40 years of data from five other asset classes. In other words, long-only momentum returns are attractive.
- The momentum premium is not as concentrated in small stocks as the value premium. During 1991-2013, the similarly estimated annualized gross momentum (value) premium is 4.5% (0.7%) among large stocks and 8.1% (6.5%) among small stocks.
- Based on a study of actual institutional trading frictions (trading patiently and allowing some tracking error), momentum profitability easily survives trading costs despite a turnover five to six times that of value.
- Momentum has a tax burden no greater than that of value because it tends to hold winners and sell losers (long-term capital gains and short-term capital losses) and because it has lower dividend exposure. Moreover, smart trading can suppress the momentum tax burden.
- Using momentum as an ancillary screen subordinate to some other primary decision factor (like value), rather than as a primary decision factor itself, is substantially suboptimal.
- There is no evidence that momentum has weakened since publication or with decline in trading frictions. Moreover, momentum diversification benefits would persist even if its premium disappeared.
- While momentum may experience deeper crashes (during dramatic market advances) than other factors, it still generates a higher long-term Sharpe ratio. Combining value and momentum (such as 60/40 value/momentum) mutually mitigates momentum and value crashes.
- As for other factors, there are several arguably valid momentum metrics, and these metrics may perform differently on a given set of data. Choosing the simplest metric or taking an average across all reasonable metrics are ways to guard against data snooping bias.
- Both risk-based and behavioral (overreaction/underreaction) explanations of momentum offer an economic reason for its premium to exist and persist.
In summary, the body of research on price momentum refutes many widely claimed criticisms of momentum investing.
Cautions regarding conclusions include:
- Individual investors likely cannot achieve the low institutional trading frictions cited above, and would pay management/administrative fees for indirect access to low frictions.
- The authors do not refute the risk of very deep crashes with a stock momentum strategy, suggesting rather that investors diversify strategies.
