On June 21, at approximately 4:45pm ET, this article was revised. We realized that USMV was not an appropriate example of a low-volatility ETF. (It is actually a minimum-variance ETF, which is formed by optimization with the objective of minimizing portfolio volatility with an emphasis on diversification.) We replaced USMV with SPLV, which is a low-volatility fund.
The superior performance of low-volatility (and the related factor of low-beta) stocks was first documented in the 1970s – by Fischer Black (in 1972), among others – even before the size and value premiums were “discovered.” The low-volatility anomaly (i.e., lower risk stocks outperformed higher risk stocks) has been shown to exist in equity markets around the world. Interestingly, this finding is true not only for stocks but for bonds as well.
The bear market of 2008 heightened investor interest in low-volatility investing. As one example, since inception in June 2011, the Invesco S&P 500® Low Volatility ETF (SPLV) has gathered $11.8 billion in assets as of June 2019. Using data from Morningstar, the following table compares the performance of SPLV to that of Vanguard S&P 500 Index Fund ETF (VOO). As you can see, in terms of risk-adjusted performance (the Sharpe ratio), SPLV performed quite well.
June 2011-May 2019
| Annualized Return (%) | Standard Deviation | Sharpe Ratio | |
| SPLV | 12.4 | 9.4 | 1.25 |
| VOO | 11.7 | 11.9 | 0.95 |
Before you jump on the low-volatility bandwagon, it’s important you understand that the research demonstrates that the performance of the low-volatility factor is actually well explained by exposure to other factors, and it is also highly regime dependent.
Exposure to other common factors explains returns to low volatility/low beta
Both Robert Novy-Marx’s 2016 study, “Understanding Defensive Equity,” and Eugene Fama and Kenneth French’s 2015 study, “Dissecting Anomalies with a Five-Factor Model,” argued that the low-volatility and low-beta anomalies are well explained by asset pricing models that include the newer factors of profitability and investment (in addition to market beta, size and value).
Stefano Ciliberti, Yves Lemperiere, Alexios Beveratos, Guillaume Simon, Laurent Laloux, Marc Potters and Jean-Philippe Bouchaud, authors of the 2017 paper “Deconstructing the Low-Vol Anomaly,” studied the factor on a global basis and found that once the common factors of value and profitability are controlled for, the performance of low volatility/low beta becomes insignificant. They concluded: “While the low vol(/low-β) effect is indeed compelling in equity markets, it is not a real diversifier in a factor driven portfolio that already has exposure to Value type strategies, in particular Earning-to-Price and Dividend-to-Price.”
In his 2012 paper, “Enhancing a Low-Volatility Strategy is Particularly Helpful When Generic Low Volatility is Expensive,” Pim van Vliet found that while, on average, low-volatility strategies tend to have exposure to the value factor, that exposure is time varying. The low-volatility factor spends about 62% of the time in a value regime and 38% of the time in a growth regime. The regime-shifting behavior affects the performance of low-volatility strategies. When low-volatility stocks have value exposure, they, on average, outperformed the market by 2.0%. However, when low-volatility stocks have growth exposure, they have underperformed by 1.4%, on average.
Luis Garcia-Feijóo, Lawrence Kochard, Rodney Sullivan and Peng Wang, authors of the 2015 study “Low-Volatility Cycles: The Influence of Valuation and Momentum on Low-Volatility Portfolios,” found that there was no alpha in a four-factor model except in extremely cheap low-volatility environments. This finding is important because, as you will see in the following table, the “curse of popularity” has changed its exposure to the value factor.
We’ll examine the current relative valuations of SPLV, comparing them to those of the Vanguard Russell 1000 Value ETF (VONV) Data is from Morningstar as of June 21, 2019.
|
Price-to-Earnings (P/E) |
Price-to-Book (P/B) |
|
|
SPLV |
20.7 |
2.6 |
|
VONV |
13.8 |
1.8 |
While the research has found that low volatility has outperformed, that was only true while it was in the value regime, which it had been about two-thirds of the time. Clearly, low volatility is not in a value regime today, SPLV’s P/E being 50% higher than that of VONV, and its P/B being 44% higher than that of VONV.
We can see the shifting nature of the exposure of SPLV to the value factor by using the regression tool available at Portfolio Visualizer. From inception in June 2011 through December 2016, SPLV’s exposure to the value factor was a statistically significant +0.29. However, from January 2017 through April 2019, its exposure had changed signs at -0.12. In other words, SPLV had shifted from a value fund to a growth fund at a time when the spread between valuations of growth stocks and value stocks is at extreme levels.
Valuation spreads
AQR Capital Management has data going back to 1984. Using a composite value measure that includes P/B, P/E, price-to-forecasted earnings and sales-to-enterprise value, they found that, neutralizing for industry exposure as of March 31, 2019, the valuation spread between U.S. value stocks and growth stocks was trading at the 85th percentile (only 15% of the time was value trading cheaper relative to growth). Note that industry neutrality presents a more “conservative” picture, as an unconstrained measure would put the U.S. valuation spread at the 97th percentile.
The research has shown that the performance of low volatility is also explained by exposure to the term premium. The findings from the following papers are all consistent in showing low volatility’s exposure to the term factor: the 2011 study “Understanding Low Volatility Strategies: Minimum Variance” by Ronnie Shah; the 2014 study “A Study of Low-Volatility Portfolio Construction Methods” by Tzee-man Chow, Jason Hsu, Li-lan Kuo and Feifei Li; and the 2014 study “Interest Rate Risk in Low-Volatility Strategies” by David Blitz, Bart van der Grient and Pim van Vliet. With this in mind, the term premium has benefited from a sharp drop in interest rates, with the yield on the 10-year Treasury falling to just 2.14% at the end of May 2019.
Summary
The low-beta anomaly was documented almost 50 years ago. It has been persistent and pervasive around the globe and across asset classes. However, research demonstrates not only that returns to the anomaly are well explained by exposure to what are now considered other common factors (mainly value, quality and term) but also that the premium is dependent on whether low volatility is in the value or growth regime and the term premium.
The returns to the premium have justified investing only when low-beta stocks are in the value regime. They are no longer in that regime. And interest rates are at historically low levels.
Today’s investors should be concerned about the curse of popularity and the resulting rise in valuations, which historically have predicted negative returns to the low-volatility anomaly. They should also be concerned about the potential impact of rising interest rates on returns, as low-volatility stocks are more bond-like.
Finally, long-only funds that don’t focus on this anomaly can benefit from screening out lottery stocks that drive the poor performance of securities in the highest quintile of beta. For example, firms such as Bridgeway Capital Management and Dimensional Fund Advisors have long screened out such stocks. (Full disclosure: My firm, Buckingham Strategic Wealth, recommends Bridgeway and Dimensional funds in constructing client portfolios.)
Larry Swedroe is the director of research for The BAM Alliance, a community of more than 130 independent registered investment advisors throughout the country.
