Should you pursue a low-volatility strategy?

(Money Watch) The relative superior performance of low volatility stocks is one of the many well-documented anomalies in finance -- low-volatility stocks have produced superior risk-adjusted returns relative to high-volatility stocks. With the publication of academic papers on the subject, many asset management companies have built strategies to exploit the anomaly. I've written on this strategy before, as you're likely aware. My posts of August 20, 2013,  July 2, 2013, May 16, 2013, andAugust 2, 2012 all looked at this issue. Let's add one more.

• How being scared of leverage affects stock returns
• Trust models over instincts
• The problem with high-dividend strategies

The anomaly appears to exist because of both behavior biases of investors and structural constraints. Among the issues are investor's preferences for stocks with lottery-like characteristics, limits to arbitrage because of the high costs of shorting high-volatility stocks and aversion or restrictions on the use of margin/leverage, and the risk of tracking error (active managers failing to beat their benchmark indices). Erik Knutzen, author of the paper, "Pursuing the Low Volatility Anomaly," published in the Fall 2013 issue of the Journal of Investing, examined the issues to determine if investors should pursue the anomaly. The following is a summary of his findings:

• While the anomaly does exist, like all anomalies, it can experience multi-year periods of underperformance, testing investor discipline. Thus, you shouldn't seek to exploit the anomaly unless you are prepared to stay the course over the long term. For example, Knutzen noted that from 1996 through 1999 a low-volatility strategy underperformed by a wide margin. In 1999 alone it underperformed by 29 percent -- an amount sufficient to test the discipline of even the most sophisticated investor.
• Because human nature is what it is, there is no reason to expect the behavioral biases to change, nor the limits to arbitrage to disappear. Thus, it's quite possible that even though there is no risk story behind the low-volatility strategy, it can persist -- it's possible it can provide higher risk-adjusted returns in the future. However, there's a problem.
• As noted in my post of August 20, 2013, Knutzen shows that the demand for low-volatility strategies has changed their very nature. He shows that while in the past low-volatility strategies tended to be value strategies (buying stocks with relatively low price-to-book ratios) -- and that exposure might help explain the anomaly -- this is no longer the case. Valuations have been driven much higher by demand from the new funds. He also notes, as I did, that the superior performance of low-volatility strategies could be largely attributed to industry bets. Utility and non-durable goods had higher weights in a low-volatility portfolio than a market portfolio. Knutzen notes that the S&P Low Volatility Index has a 38 percent exposure to utilities. Note that utilities have exposure to term risk, and obviously benefited from the 40-year bull market in bonds -- which cannot be expected to be repeated.
• Knutzen presents a chart of the relative valuations of high beta and low beta stocks (beta, while measuring something different than volatility, is a good proxy for volatility). He shows that low-volatility stocks are now relatively overvalued by his measure by over 20 percent, and high beta stocks are undervalued by 25 percent -- with prices driven by investor flows and recent performance chasing. In other words, the expensive valuations of low beta (and low-volatility) stocks is driven by their recent strong performance. And if there is anything certain in the world of finance it's that high valuations are strong predictors of future poor returns.