We’ve worried over the last several years whether momentum and other “alpha” factors have become exploited to the point of diminishing returns. It’s an arms race out there...
Our work on the annual “momentum effect” dates back 15 years, and was originally based on equity sectors rather than asset classes. The hypothetical approach is to entirely dispense macroeconomic trends, sector fundamentals, and valuations, and base the allocation decision exclusively on momentum.
With last year’s Bridesmaid (REITs) having laid an egg, the long-term “alpha” of the Bridesmaid portfolio narrowed to +3.7% from a bit over +5% (annualized) when we first published this study more than a decade ago.
In the extreme case where one possesses no other information beyond last year’s total returns, the best single-asset strategy has been to buy the second-best performer (the “Bridesmaid”) and hold it for the next twelve months in hopes that the prior year’s momentum will carry it through. That approach has beaten the S&P 500 by 3.7% annualized over the past 48 years.
2019 was the fourth consecutive year of underperformance by the annual Bridesmaid sector pick. Those poor results have trimmed the annualized “alpha” of the strategy to just +2.2% since 1991.
Our work on the Bridesmaid momentum effect dates back to 2006, and was originally based on equity sectors rather than asset classes. Again, the hypothetical approach is to ignore macroeconomic trends, sector fundamentals, valuations, and the like, and to base sector selection solely on the prior year’s sector total return rankings.
Here are the historical annual performance results for the hypothetical Bridesmaid strategy.
While the consideration of risk seems almost a quaint notion as the bull market nears its eighth birthday, it’s nonetheless worth noting the Bridesmaid allocation strategy has generated a favorable return/volatility trade-off in relation to: (1) the seven candidate asset classes; and, (2) the strategy of owning an asset class with a prior-year total return rank other than #2.