Swedroe: A Factor With Caveats

September 14, 2018

Performance Analysis

Before summarizing, using the regression tool at Portfolio Visualizer, I’ll analyze the performance of three low-volatility ETFs through a factor model lens to see if the results are consistent with the academic research I have reviewed. The loadings on each factor are in parentheses.

I’ll begin with the iShares Edge MSCI Min Vol U.S.A. ETF (USMV). Data is available for the period November 2011 through June 2018. The table presents the annualized alphas.



First, the loadings should not be surprising. For example, low-volatility stocks tend to be large (hence the negative loading on size) and high-quality, defensive stocks. The negative loading on credit also reflects the high quality of low-volatility stocks.

Second, in terms of alpha, the results are entirely consistent with the findings we have been discussing. While low-volatility strategies have high alphas in a single-factor CAPM and three-factor world, the alphas turn negative once the newer factors of quality and low beta, as well as the term factor, are considered. (Note that this finding is a bit surprising, as I would expect that the inclusion of the low-beta factor would drive alpha toward zero.) In other words, investors are better served by directly targeting exposure to the factors.

We see similar results looking at the performance of the Invesco S&P MidCap Low Volatility ETF (XMLV). The period covers the time for which data is available, from March 2013 through June 2018.



The only major difference from the results we saw with USMV is that, because XMLV is a midcap fund, it has positive exposure to the size factor. Again, once we account for all common exposures, the alphas turn negative.

Again, we see similar results when looking at the performance of the Invesco S&P SmallCap Low Volatility ETF (XSLV). The period also is from March 2013 through June 2018. The one main difference is that, because most of low volatility’s benefit comes from excluding high-volatility, small-cap stocks from the portfolio, the alpha is greater in small stocks for models that don’t account for fixed-income factors and the low-beta factor itself.



Among the three low-volatility funds we examined, there is almost $19 billion in assets generating large negative risk-adjusted alphas. It’s hard to imagine that investors are aware of this.

Find your next ETF

Reset All