The rise of factor-based investing over the last two decades has greatly impacted the investing landscape. From “explaining” active strategies to identifying sources of risk and return in financial markets with greater clarity, viewing asset pricing through a factor lens has been nothing short of a revelation for money management.
But amid the rush to tap into what is arguably a better way to analyze, design and implement portfolios, how much do we really know about the historical record for the major factor premia? Perhaps less than we think.
Slicing and dicing factor returns is a relatively recent innovation and so it’s no surprise to find that the track record is relatively light compared with conventional market history. Key aspects of modern factor investing was forged in the 1990s and it took a decade or so for related fund products and strategies to evolve and find traction on Wall Street and beyond.
By contrast, standard investment indexes, along with traditional active management strategies, have much longer histories. The Dow Jones Industrial Average – the original market index — dates to 1885 and active mutual funds in the U.S. have been trading since the 1920s. Factor premia history tends to be shorter, in some cases much shorter. That’s a challenge for analysis and developing perspective on expected risk and return in the niche — which is why a recent research project that fills in some of the missing data history deserves close attention.
“Do Factor Premia Vary Over Time? A Century of Evidence” is a recent research paper authored by AQR Capital Management’s Antti Ilmanen and several co-authors that examines the value, momentum, carry, and defensive equity risk factors using a century of data across six asset classes. “We attempt to shed new light on variation in factor premia using unique data from almost a century across six different asset classes,” the authors write.
The deep dive on fleshing out the track record on US and foreign equities, commodities, bonds and currencies is impressive, not to mention immensely practical in the cause of understanding the history of value, momentum, carry and defensive risk factors.
Quite simply, this a must-read study for investors intent on deploying capital via a factor-based world view.
The perspective on equities alone is worth a look. As an example, consider how the value and momentum factors for U.S. stocks compare, using a start date of 1925. The annualized mean return for the value premium is 3.2%, or less than half of momentum’s 6.8%. Notably, momentum’s Sharpe ratio is considerably higher, too: 0.44 vs. 0.24. Is there a price tag for momentum’s higher performance in terms of maximum drawdown? Just the opposite, we’re told: momentum’s deepest peak-to-trough decline is -49% — a hefty tumble, to be sure, but moderately softer relative to value’s -56%.
Overall, the paper offers a wealth of information for contextualizing the factor premia under scrutiny. Summarizing this broad-minded research is difficult in a single table or chart, but the paper’s Figure 1 does yeoman’s work. The graph below presents the distribution of factors through time for the various asset classes at the 10th and 90th percentiles, in essence offering a quick study on managing expectations for the premia. For handy reference, the authors also include a relevant market benchmark for each asset class.
“While not definitive, Figure 1 suggests that the scope for timing factors is no less promising than timing asset classes,” the authors observe. Indeed, the variation of factor returns for individual markets more or less aligns with the dispersion for broadly defined market portfolios.
The average single factor has an inter-decile range of -7.2% to 17.2% return per annum. However, the multifactor portfolios mute much of this variation over time, with the multifactor, multi-asset portfolio ranging from 0.3% to 6.3% from the 10th to the 90th percentile.
Another intriguing finding relates to the potential for diversification benefits, which the paper deems valuable and (perhaps more importantly) persistent, as shown in Figure 2.
Figure 2 plots the time-series of pairwise correlations between the factors (across all asset classes) using rolling monthly return data over the prior 10 years from 1933 to 2018. There is some variation in correlation structure across the factors, but no periods when correlations all become significantly large to diminish diversification benefits and no upward trend in correlations.
Overall, the authors conclude that a century of historical data reveal that the four factor premia are “robust and significant in every asset class over the last century” and “vary significantly over time.” It’s also fascinating to discover that there’s minimal evidence for sourcing the premia to macroeconomic, business cycle, tail risk and other drivers that tend to come up on financial analysts’ short list for modeling. As such, the persistence of the premia present challenges to asset pricing theories for explanatory purposes.
Academics may be scratching their heads after reading the results, but investors looking for empirically based opportunities for enhancing portfolio diversification will find the paper a compelling read.
By James Picerno, Director of Analytics
IMPORTANT DISCLOSURES: PLEASE REMEMBER THAT PAST PERFORMANCE MAY NOT BE INDICATIVE OF FUTURE RESULTS. DIFFERENT TYPES OF INVESTMENTS INVOLVE VARYING DEGREES OF RISK, AND THERE CAN BE NO ASSURANCE THAT THE FUTURE PERFORMANCE OF ANY SPECIFIC INVESTMENT, INVESTMENT STRATEGY, OR PRODUCT MADE REFERENCE TO DIRECTLY OR INDIRECTLY FROM THE MILWAUKEE COMPANY™, WILL BE PROFITABLE, EQUAL ANY CORRESPONDING INDICATED HISTORICAL PERFORMANCE LEVEL(S), OR BE SUITABLE FOR YOUR PORTFOLIO. DUE TO VARIOUS FACTORS, INCLUDING CHANGING MARKET CONDITIONS, THE CONTENT MAY NO LONGER BE REFLECTIVE OF CURRENT OPINIONS OR POSITIONS. MOREOVER, YOU SHOULD NOT ASSUME THAT ANY DISCUSSION OR INFORMATION CONTAINED IN THE MILWAUKEE COMPANY™ SERVES AS THE RECEIPT OF, OR AS A SUBSTITUTE FOR, PERSONALIZED INVESTMENT ADVICE FROM THE MILWAUKEE COMPANY™