Just because ‘momentum’ fades for a while doesn’t mean you should abandon it.
This is the second article in a two-part series about the momentum factor and whether or not its premium has permanently disappeared. The first article in this series appeared here.
Earlier this week, we took a look at some of the historical evidence on the persistence of the Fama-French momentum factor. Today we’ll examine the momentum premium’s out-of-sample record, as well as its uses in portfolio diversification.
The authors of the 2013 study, “212 Years of Price Momentum,” concluded that the most recent decade-long underperformance of momentum is not unusual, and that the momentum premium shifts with “regime” changes. The study examined the long-term evidence on momentum, merging three known 19th- and early 20th-century data sources into one testable data set from 1800 to 1927.
After extending the data back to 1801, the study’s authors found seven other decade-long periods of negative returns, outside of the period from 2004-2013. The authors also found that the out-of-sample record of momentum has been exceptional.
After its published discovery, the positive premium has continued in the U.S. market, in the international markets, in currencies, and in commodities. They also cited the study “Value and Momentum Everywhere,” which traces the power of the momentum anomaly across the globe. And finally, they noted that the momentum factor has also been confirmed in 19th-century British stock prices.
One of the more common mistakes investors make is to view the risk and return of assets in isolation. The right way to view assets is to determine how their addition impacts the risk and return of the entire portfolio.
For investors who tilt their portfolios to value strategies—which creates a larger-than-market exposure to the value factor—even if the momentum premium is low or negative, the research demonstrates that exposure to the momentum factor is still valuable for diversification purposes.
The reason is that the value and momentum factors exhibit strong negative correlation. From 1927-2013, the monthly and quarterly correlations of the U.S. value and momentum factors were about -0.4, and the semiannual and annual correlations were about -0.27.
Perhaps the best illustration of the diversification benefit provided by momentum is from the paper “Momentum in Japan: The Exception that Proves the Rule.” The authors showed that while momentum on a stand-alone basis has not “worked” in Japan, a portfolio with weightings of about 30 percent momentum and 70 percent value would have produced a higher Sharpe ratio than a 100 percent value portfolio.
Keep Some Momentum In The Portfolio
The study “Fact, Fiction and Momentum Investing,” provides further support. In their Myth No. 7, the authors asked the question, If the momentum factor went to zero, would you still want to have exposure to momentum in your portfolio? They showed that because of its diversification benefits, even if the premium went to zero, you would still want at least some exposure to momentum.
As we noted earlier, there is a nonzero chance that—no matter how long your horizon may be—there will end up being no momentum premium going forward. However, this is true for every other factor as well.
Concerns existed about the value factor having disappeared in the latter half of the 1990s. Some believed the refrain that the factor’s disappearance was somehow “different this time” and that it wouldn’t return. Others credited its disappearance to research published on the value premium.
From 1995 through 1999, the average annual value premium was -5.0 percent. However, from 2000 through 2014, the average annual value premium was 6.1 percent—1.4 percentage points higher than it had been from 1927 through 1999. So much for the value premium’s supposed “disappearance.”
Believe In Momentum
In summary, the future existence of the momentum premium is so often questioned because there is a much stronger behavioral explanation for it than a risk-based one. On the other hand, no other factor—save the market itself—has nearly as long a track record, as much out-of-sample evidence or as strong and persistent a return premium as does momentum.
The bottom line is that there’s a large body of evidence that the momentum effect has been persistent and pervasive across time, markets, economic cycles and geographic regions. Whether the spreading of this knowledge has caused the momentum premium to diminish or disappear, only time will tell.
But one 10-year period, dominated by the effect of a single year when a strong reversal occurred, doesn’t appear to be a sufficient reason to abandon the strategy.
Larry Swedroe is the director of research for the BAM Alliance, a community of more than 130 independent registered investment advisors throughout the country.