Among the hot “smart beta” strategies into which investors are pouring assets is quality. For example, the iShares Edge MSCI USA Quality Factor ETF (QUAL | A-84), which is only about three years old, already has $2.7 billion in assets. Before you consider investing in these increasingly popular strategies, however, it’s worth understanding the sources of their returns. That, in turn, raises an important question: Is the quality factor risk-based, or is it a behavioral anomaly?
Robert Novy-Marx’s 2012 paper, “The Other Side of Value: The Gross Profitability Premium,” provided investors with new insights into the cross section of stock returns. His study built upon the 2007study “Profitability, Investment and Average Returns” by Eugene Fama and Kenneth French, who had shown that firms with high profitability measured by earnings have high subsequent returns after controlling for book-to-market ratio and investment.
Novy-Marx investigated gross profits (defined as sales minus cost of goods sold) over the period 1962 through 2010. Among his important findings were:
- Profitability, as measured by gross profits-to-assets, has roughly the same power as book-to-market ratio (a value measure) in predicting the cross section of average returns.
- Surprisingly, profitable firms generate significantly higher returns than unprofitable firms, despite having significantly higher valuation ratios (for instance, higher price-to-book ratio).
- Profitable firms tend to be growth firms, and they expand comparatively quickly. Gross profitability is a powerful predictor of future growth as well as of earnings, free cash flow and payouts.
- The most profitable firms earn average returns of 0.31% per month higher than the least profitable firms. The data is statistically significant, with a t-statistic of 2.49.
- The returns data is economically significant even among the largest, most liquid stocks.
- Gross profitability has far more power in predicting the cross section of returns than earnings-based measures of profitability.
- High asset turnover (defined as sales divided by assets, an accounting measure of operating efficiency) primarily drives the high average returns of profitable firms, while high gross margins are the distinguishing characteristic of “good growth” stocks.
- Controlling for profitability dramatically raises the performance of value strategies, especially among the largest, most liquid stocks. Controlling for book-to-market ratio improves the performance of profitability strategies.
- Strategies built on profitability are growth strategies, so they provide an excellent hedge for value strategies. Adding profitability on top of a value strategy reduces its overall volatility.
Since the publication of Novy-Marx’s work, the profitability factor has been extended to a broader quality factor (the returns to high-quality companies minus the returns to low-quality companies), which captures a larger set of quality characteristics. High-quality companies have the following traits: low earnings volatility, high margins, high asset turnover (indicating the efficient use of assets), low financial leverage, low operating leverage (indicating a strong balance sheet and low macroeconomic risk) and low specific stock risk (volatility unexplained by macroeconomic activity). Companies with these characteristics historically have provided higher returns, especially in down markets. In particular, high-quality stocks that are profitable, stable, growing and have a high payout ratio outperform low-quality stocks with the opposite characteristics.
The quality factor is referred to as quality minus junk. For the period 1927 through 2015, the quality premium had an annual average return of 3.8%. In addition, it was slightly more persistent than the value premium, and only slightly less persistent than the market beta premium.
A Risk-Based Or Behavioral-Based Explanation?
The academic research provides at least some support for both risk-based and behavioral-based explanations for the profitability premium. A problem for risk-based explanations is that, intuitively, more-profitable firms are less prone to distress and have lower operating leverage than unprofitable firms. These characteristics suggest they are less, not more, risky.
On the other hand, more profitable firms tend to be growth firms, which have more of their cash flow in the distant future. More distant cash flows are more uncertain, and should require a risk premium. Another risk-based explanation is that higher profitability should attract more competition, thus threatening profit margins. That, too, creates more risk and should require a risk premium.
Jean-Philippe Bouchaud, Stefano Ciliberti, Augustin Landier, Guillaume Simon and David Thesmar contribute to the literature on the quality factor with their paper, “The Excess Returns of ‘Quality’ Stocks: A Behavioral Anomaly,” which was published in the June 2016 issue of the Journal of Investment Strategies.
The authors first demonstrated the performance of the strategy for U.S. stocks. They write: “Quite surprisingly, this performance is quite high: even absent cross-country diversification, one already obtains a Sharpe Ratio of 1.2 over the period 1990-2012, corresponding to a highly significant t-stat ≈ [about] 6. The same strategy is statistically significant in all geographical zones, and the corresponding signal moves sufficiently slowly so that large amounts of capital can be invested without suffering from prohibitive transaction costs.” They then try to determine whether the premium is risk-based or behavioral-based.
The authors note: “One possible story could be the following: firms can choose between safe and moderately profitable projects, and risky and more profitable ones. High average cash-flow to assets might indicate that the corresponding firms are themselves, on average, operating on very profitable but riskier segments of the economy and therefore riskier to own.”
They then observe: “However, while well-known risk premia strategies are indeed rewarding investors for carrying a significant negative skewness risk, quality strategies are in fact found to have a positive skewness and a very small propensity to crash.”
Biases In The Brain
The authors subsequently posit psychological biases exist that are inconsistent with the efficient markets hypothesis, including “that investors systematically underweight the information contained in quality-like signals (cash flows, ROA, etc.).” They offer a complimentary story that through “conservatism bias” investors’ “beliefs are ‘sticky’ and this leads to under-reaction to good or bad news about the firms.”
They then perform statistical tests and found strong support for the behavioral view. For example, they found statistically significant (at the 5% confidence level) evidence that “analysts, at best, neglect the information contained in cash-flow statements—or even weight it with the wrong sign.”
They also note that they studied expected earnings by analysts (rather than expected prices) and relate the “stickiness” of analysts’ beliefs to asset mispricing. They found that these two “widespread behavioral biases” (misplaced focus and “stickiness”) are “at the heart of the quality anomaly.”
The authors also found that analysts tend to be too optimistic about growth stocks (equities with low book-to-market ratios) and high-volatility stocks. They note this suggests that the low-volatility and value anomalies might also have behavioral origins rooted in expectation mistakes, like quality.
Perhaps their most interesting finding, however, is that “analysts are clearly under-weighting operating cash-flows; in fact they even seem to be putting a slightly negative weight on that variable, even though it is a strong positive predictor of future returns. These last results strongly suggest that the quality anomaly is likely due to a significant underweighting of quality in price forecasts. For this to transpire in market prices, one needs to further assume that analysts have some influence on other market participants, or simply if their opinions are representative of those of investors.” The authors wrote that they consider this hypothesis to be “quite plausible.”
The authors concluded: “Since the returns of a strategy long the most profitable companies and short the less profitable ones has a positive skewness, it is hard to account for this anomaly using a risk premium argument, which are characterized by a negative skewness (i.e. a propensity to crash).”
They add: “By contrast, we report strong evidence that financial analysts pay insufficient attention—or even negative attention—to firm-level profitability indicators such as operating cash-flows, leading to forecast errors that are negatively correlated with quality indicators. This suggests that the quality anomaly arises from non-optimal weighting of profitability information by analysts and investors in their expectations. This behavioral bias might be due to relative excess attention to more salient accounting information, or to under-reaction to positive (or negative) news about the firm.”
Their findings are consistent with much of the recent research on the profitability and quality factors. For example, in his 2015 study, “Profitability Premium: Risk or Mispricing?”, which covered the period July 1963 through 2013, Ryan Liu noted that while profitable firms have higher unconditional returns, investors might still avoid them if their returns are the lowest during bad times because investors care the most about returns during bad times when their marginal utility of wealth is high.
However, Liu found that the premium is actually higher during economic downturns. Profitable firms do even better than unprofitable firms during bad times when marginal utility (the benefit of incremental income or wealth) is the highest. Thus, they are less susceptible to negative macroeconomic conditions. The profitability premium increases both in recessions and when the stock market is doing poorly.
Liu then investigated the mispricing hypothesis. Specifically, he examined the difference between earnings forecasted by sell-side analysts and actual earnings realized across profitability-sorted portfolios. If the low return of the unprofitable firms relative to the profitable firms was due to investors being too optimistic about their future performance, the difference between forecasted and actual earnings (the expectation error) should be larger for unprofitable firms. He found a monotonically decreasing relationship across the 10 deciles of profitability from low to high.
The expectation error was not only larger for the unprofitable firms, it was persistent for up to five years. His investigation of the data led him to conclude that investors expect the performance of profitable firms to mean-revert faster than they actually do, and they are willing to bet on the revival of the unprofitable firms despite low net income and poor current performance.
This explanation is somewhat different from the more typical glamour story in which naive investors become overly optimistic about stocks in favor due to good news or good past performance, resulting in overvaluation of the glamour stocks.
In this case, the excessive optimism is about the potential for mean-reversion of unprofitable companies, which tend to be newer, smaller firms in distress. Thus, these stocks tend to be overvalued. Due to limits to arbitrage and the costs and risks of shorting, overvaluation is harder to correct than undervaluation.
Liu concluded that the evidence makes it hard to reconcile the profitability premium with a risk-based explanation, but it is entirely consistent with the mispricing hypothesis.
Will The Anomaly Persist?
Given the evidence strongly favoring the behavioral explanation, perhaps the next logical question is: How long will investors (in aggregate) remain inattentive and sticky in an era more and more dominated by computerized data analysis?
Investors obviously believe it will persist, as cash has been flowing into quality-based ETFs. While it’s impossible to know the answer, we do know there’s a tendency for behavioral-based premiums to at least shrink post-publication. And the literature provides us with at least some clues.
F.Y. Eric C. Lam, Shujing Wang and K.C. John Wei, authors of the 2015 study “The Profitability Premium: Macroeconomic Risks or Expectation Errors?”, also explored the two alternative explanations for the profitability premium: the rational explanation based on macroeconomic risks and the mispricing explanation attributed to expectation errors.
They found that both explanations play a role, with macroeconomic risks explaining about one-third of the profitability premium and the remainder explained by a misvaluation factor based on investor sentiment.
Huijun Wang and Jianfeng Yu, authors of the December 2013 study “Dissecting the Profitability Premium,” also sought to determine whether the explanation for the profitability premium was risk-based or behavioral-based. In examining behavioral explanations, Wang and Yu hypothesized that, to the extent the profitability premium reflects mispricing, it should be larger among firms that are more difficult to arbitrage and have greater information uncertainty. The greater the level of uncertainty, the greater we should expect the impact of investor overconfidence to be on prices.
Where there are higher limits to arbitrage, the mispricing is more likely to be sustained. In addition, with greater information uncertainty, psychological biases are increased and information is more asymmetric among investors, leaving more room for mispricing. Using a large set of standard proxies in the literature for limits to arbitrage and information uncertainty, they found the profitability premium is substantially stronger among firms that are more difficult to arbitrage or have greater information uncertainty. Specifically, they found:
- The profitability premium is insignificant or only marginally significant among firms that have low information uncertainty and are easy to arbitrage.
- The profitability premium is about 1% higher per month among firms with smaller capitalization, younger age, higher return volatility, higher cash flow volatility, less analyst coverage, larger analyst forecast dispersion, fewer institutional holdings, higher idiosyncratic return volatility, lower dollar trading volume, higher bid/ask price spread, lower credit rating and higher illiquidity.
- The majority of the ROE (a measure of profitability) premium is derived from the subsequent low returns of low-ROE firms. This is consistent with the notion that overpricing is harder than underpricing for arbitrageurs to correct due to greater shorting impediments.
- The profitability premium is not driven by ex-post overreaction (there is no evidence of long- term reversion) but by ex-ante underreaction. Investors underreact to the current profitability news, and hence high-profitability (low-profitability) firms are relatively underpriced (overpriced).
Wang and Yu concluded that the profitability premium persists because of limits to arbitrage, which prevent mispricings from being corrected. However, the fact that pricing errors—instead of rational, risk-based explanations—may be responsible for the majority of the premium does not mean it is doomed to disappear.
Anomalies can indeed persist because of limits to arbitrage that prevent mispricings from being corrected. A good example is the momentum premium, which has persisted for more than 20 years since the publication of the first paper on it.
Larry Swedroe is the director of research for The BAM Alliance, a community of more than 140 independent registered investment advisors throughout the country.