The value factor—the difference in returns between glamour growth stocks and distressed value stocks—is probably the most researched of all the factors in the “factor zoo.” Unlike some other factors, its existence isn’t debated, at least not among academia.
The reason is that, as my co-author Andrew Berkin and I demonstrate in “Your Complete Guide to Factor-Based Investing: The Way Smart Money Invests Today,” the value premium has been persistent across long periods of time and different economic regions, pervasive across the globe and even asset classes (cheap assets have outperformed expensive assets) and, while the most common metric used to define value has been the book-to-market (BtM) ratio (as in Fama and French), it has been robust to many other definitions. In addition, it’s implementable (survives transaction costs).
That said, there’s great debate about the value premium regarding its source. Is it risk-based, behavioral-based or perhaps a combination of the two?
Over the past several years, several papers have contributed to the literature and the debate, the common theme being that enterprise value was used as the value metric. The enterprise multiple (EM) is calculated as enterprise value (EV = equity + debt + preferred stock - cash) divided by earnings before interest, taxes, depreciation and amortization (EBITDA = operating income + depreciation + amortization).
Today we’ll look at three papers that have investigated the ability of enterprise value to explain the cross section of returns and provide insights into the source of the value premium.
The Significance Of Enterprise Value
The first paper we’ll examine is “New Evidence on the Relation between the Enterprise Multiple and Average Stock Returns,” which was published in the December 2011 issue of the Journal of Financial and Quantitative Analysis. The authors, Tim Loughran and Jay Wellman, were motivated to investigate this issue by the increased use of enterprise value as a measure of a firm’s value.
As the authors explain, the use of enterprise value is intuitive: “Firms with a high enterprise multiple are by definition selling for a high valuation for a given dollar of operating income compared to companies with a low EM ratio. These high enterprise multiple firms have stronger growth opportunities, lower costs of capital, and thus lower expected stock returns than low enterprise multiple firms.” Their data sample covers the period July 1963 through 2009. Following is a summary of their findings:
- The EM appears similar to book-to market in terms of monthly cross-sectional regression coefficients and t-statistics.
- Though BtM and EM are correlated, the expected return-enterprise multiple relation remains significant after controlling for BtM.
- The EM factor generates a premium of 0.44% per month, or 5.28% per year, significant at the 1% level.
- The enterprise multiple factor remains significant after controlling for the Carhart four-factor model (beta, size, value and momentum) as well as the q-theory factor model (beta, size, investment and profitability). The alpha is 0.16% per month (t-statistic of 2.39) when using the Carhart four-factor model and 0.35% per month (t-statistic of 3.00) using the q-theory model.
- The EM factor loads on the investment and ROA (return on assets) factors and is thus related to the q-theory factor model, but it’s not dominated by it.
The bottom line is that these results provide strong evidence that the enterprise multiple is related to subsequent stock returns.
It would be helpful if we had evidence that the EM factor was pervasive, reducing the risk of data mining. That is the issue addressed by Christian Walkshausl and Sebastian Lobe in their study “The Enterprise Multiple Investment Strategy: International Evidence,” which appears in the August 2015 issue of the Journal of Financial and Quantitative Analysis.
As an out-of-sample test, they examined the relation between EM and average returns in a large international sample consisting of 22 non-U.S. developed markets and 18 emerging markets that together account for two-thirds of the world market capitalization. Following is a summary of their findings:
- The return predictability of EM is similarly present in developed markets and emerging markets, and robust across small ﬁrms and large ﬁrms, after controlling for common benchmark variables like size, BtM value and momentum, and is highly persistent for up to five years after portfolio formation.
- An international portfolio of low EM firms outperforms a portfolio of high EM firms by 0.95 percentage points per month, with a t-statistic of 6.97. What’s more, international EM premiums are two times higher than the corresponding U.S. EM premiums studied by Loughran and Wellman.
- The strong performance of the EM strategy comes from both ends—low-EM ﬁrms earn average excess returns of more than 1% per month, while high-EM ﬁrms yield average excess returns that are statistically indistinguishable from zero.
- The EM premiums are significantly positive in all 40 non-U.S. countries. Eighteen of the 22 developed markets have significant t-statistics (> 2), and eight of the 18 emerging markets have significant t-statistics (>2).
These strong results make it unlikely that the U.S. experience is simply due to chance.
Predictive Power Of The Enterprise Multiple
The third and most recent paper we’ll examine is the October 2016 study “Why do Enterprise Multiples Predict Expected Stock Returns?” The authors, Steven Crawford, Wesley Gray, Jack Vogel and Yang Xu, build on the work of the two aforementioned studies on the EM.
They began by examining the return to the U.S. EM factor for the period 1972 through 2015. They found that, consistent with the existing literature, there has been a strong EM premium in the U.S.—low-EM firms significantly outperform high-EM firms.
They then focused on the great debate—investigating whether high average expected returns associated with value stocks and low average expected returns earned by glamour stocks are compensation for risk, or a result of systematic mispricing.
To understand what drives the EM effect, they used the work of Joseph Piotroski and Eric So, authors of the 2011 study “Identifying Expectation Errors in Value/Glamour Strategies: A Fundamental Analysis Approach” to help differentiate between the risk-based and the mispricing-based hypotheses.
The authors classified and allocated observations into value and glamour (growth) portfolios on the basis of each firm’s BtM ratio. A firm’s BtM ratio reflects the market’s expectations about future performance. Firms with higher expectations will have higher prices and a lower BtM ratio, while firms with weak expectations will have lower prices and a higher BtM ratio. Thus, the book-to-market ratio serves as a proxy for the relative strength of the market’s expectations about a firm’s future performance.
Piotroski and So classified a firm’s financial strength by using the aggregate statistic F-Score, which is based on nine financial signals designed to measure three different dimensions of a firm’s financial condition: profitability, change in financial leverage/liquidity, and change in operational efficiency.
The F-Score is an early example of a composite quality factor (such as the QMJ, or quality minus junk, developed by AQR Capital). Firms with the poorest signals have the greatest deterioration in fundamentals and are classified as low F-Score firms. Firms receiving the highest score have the greatest improvement in fundamentals and are classified as high F-Score firms.
Prior research demonstrated that the F-Score is positively correlated with future earnings growth and future profitability levels: Low F-Score firms experience continued deterioration in future profitability, and high F-Score firms experience overall improvement in profitability. Following is a summary of Piotroski and So’s findings:
- Among firms where the expectations implied by their current value/glamour classification were consistent with the strength of their fundamentals, the value/glamour effect in realized returns is statistically and economically indistinguishable from zero, arguing against a risk-based explanation.
- The returns to traditional value/glamour strategies are concentrated among those firms where the expectations implied by their current value/glamour classification are incongruent ex ante with the strength of their fundamentals.
- Returns to this “incongruent value/glamour strategy” are robust and significantly larger than the average return generated by a traditional value/glamour strategy.
Glamour Investors Overlook Value
In the academic literature, the explanation for the mispricing of value stocks relative to growth stocks is that, behavioral errors—such as optimism, anchoring and confirmation biases—cause investors to underweight or ignore contrarian information.
As Piotroski and So noted, “Investors in glamour stocks are likely to under-react to information that contradict their beliefs about firms’ growth prospects or reflect the effects of mean reversion in performance. Similarly, value stocks, being inherently more distressed than glamour stocks, tend to be neglected by investors; as a result, performance expectations for value firms may be too pessimistic and reflect improvements in fundamentals too slowly.”
Piotroski and So’s findings were consistent with the mispricing explanations for the value premium. They found that the value/glamour effect was concentrated among the subset of firms where expectations implied by BtM ratios were not aligned with the strength of the firms’ fundamentals (F-Score). More importantly, the value/glamour effect was nonexistent among firms where expectations in price were aligned with the strength of the firm’s recent fundamentals.
They concluded that firms with low BtM ratios and low F-Scores (weak fundamentals) were persistently overvalued, and firms with high BtM ratios and high F-Scores (strong fundamentals) were persistently undervalued. In these subsets, the pricing errors were strongest.
The authors also noted that, while both the traditional value/glamour strategy (relying solely on BtM rankings) and the incongruent value/glamour strategy produce consistently positive annual returns, the frequency of positive returns was higher for the incongruent value/glamour strategy. It generated positive returns in 35 of 39 years over the sample period (versus 27 of 39 years for the traditional value/glamour strategy).
They also found that annual returns to the incongruent value/glamour strategy were larger than the traditional value/glamour strategy in all but six years, with an average annual portfolio return of 20.8% versus 10.5% for the traditional value/glamour strategy.
Role Of Fundamental Accounting Information
Piotroski and So’s findings are supported by those of H.J. Turtle and Kainan Wang, authors of the study “The Value in Fundamental Accounting Information,” which appears in the Spring 2017 issue of the Journal of Financial Research. Using Piotroski’s F-Score, they examined the role of fundamental accounting information in shaping portfolio performance over the period 1972 through 2012, and concluded there was a behavioral-based explanation for the value premium.
Similar to Piotrowski and So, Crawford, Gray, Vogel and Xu sorted stocks on EM and 11 well-documented anomalies (and a combination score) that are proxies for the fundamental value of the stock:
- Net Stock Issues: Net stock issuance and stock returns are negatively correlated. It has been shown that smart company management issues shares when sentiment-driven traders push prices to overvalued levels.
- Composite Equity Issues: Issuers underperform nonissuers with “composite equity issuance,” defined as the growth in a firm’s total market value of equity minus the stock’s rate of return. It’s computed by subtracting the 12-month cumulative stock return from the 12-month growth in equity market capitalization.
- Accruals: Firms with high accruals earn abnormally lower average returns than firms with low accruals.
- Net Operating Assets: The difference on a firm’s balance sheet between all operating assets and all operating liabilities, scaled by total assets, is a strong negative predictor of long-run stock returns.
- Asset Growth: Companies that grow their total assets more earn lower subsequent returns.
- Investment-to-Assets: Higher past corporate investment predicts abnormally lower future returns.
- Distress: Firms with high failure probabilities have lower subsequent returns.
- Momentum: High (low) recent past returns forecast high (low) future returns.
- Gross Profitability: More profitable firms have higher expected returns than less profitable firms.
- Return on Assets: More profitable firms have higher expected returns than less profitable firms.
- Ohlson O-Score: Stocks with a high risk of bankruptcy have lower returns than stocks with a low risk of bankruptcy.
Crawford, Gray, Vogel and Xu then created two test portfolios. First, a “high mispricing” portfolio (where there is misalignment between valuations), which goes long value firms with high expectation errors (low EM, high fundamental value), and goes short glamour firms with high expectation errors (high EM, low fundamental value).
Second, a “low mispricing” portfolio (where there is alignment of valuation), which goes long value firms with low expectation errors (low EM, low fundamental value), and short glamour firms with low expectation errors (high EM, high fundamental value).
Their empirical tests examined the spread between the two portfolios. The risk-based hypothesis predicts no difference between the returns of these portfolios, and the mispricing-based hypothesis predicts a positive spread in returns. Following is a summary of their findings:
- Portfolios with high investor expectation errors earn higher returns than portfolios with low investor expectation errors.
- Consistent with the hypothesis that the EM effect is likely explained by mispricing, in the low-mispricing portfolios, the reported four-factor (beta, size, value and momentum) alpha estimates are not statistically different from zero, while the four-factor alpha estimates for the high-mispricing portfolios are positive and significant at the 5% level in every instance.
- Examining earnings announcement returns, forecast errors and forecast revisions, evidence supported the notion that the EM effect is driven at least partially by mispricing associated with predictable investor expectation errors. For example, the average announcement return for the value portfolio with high expectation errors (i.e., cheap, with high fundamental value) is 0.65%, which is larger than the -0.22% average announcement return for the glamour portfolio with high expectation errors (i.e., expensive, with low fundamental value). The difference in these two returns is the announcement return in the high-mispricing portfolio of 0.87%, which is significant at the 1% level. On the other hand, the returns to firms in the two portfolios with low expectation errors are similar.
- Returns to the high-mispricing EM strategy are significantly higher during periods of high investor sentiment relative to times of low investor sentiment. The same pattern doesn’t hold for the low-mispricing portfolio, providing further support to the mispricing-based hypothesis.
These results were robust to various tests, including exploring results against various time periods and asset pricing models. Thus, the authors concluded there “is strong evidence in favor of the mispricing-based hypothesis and little evidence that the EM effect is a proxy for higher discount rates,” which would provide a risk-based explanation.
Effects Of The Limits To Arbitrage
The authors also examined the question, if the results are behavioral driven, hasn’t the market arbitraged away the anomaly? To address this question, they examined the limits to arbitrage associated with exploiting the EM portfolio strategy. The limits-of-arbitrage hypothesis predicts the abnormal expected returns associated with the high-mispricing portfolio will be driven by the short leg of the portfolio, where arbitrage costs are highest.
In support, they found that, on average, 62% of the alpha, and more than 80% of the alpha associated with the best EM portfolio, is generated by the short leg of the portfolio. They note: “To the extent that managing short positions is costly, these results suggest that the mispricing associated with the high-mispricing EM portfolio is difficult to profitably exploit. In addition, if costly market frictions continue to exist and investor expectation errors persist, we can expect that the EM effect may continue in the future.”
Crawford, Gray, Vogel and Xu concluded: “The evidence supports the hypothesis that the excess returns associated with EM sorted portfolios is driven by mispricing and not increased systematic risk exposure.”
In summary, taken together, the three studies, along with Piotroski and So’s work, provide support that both the value premium’s existence and behavioral explanations are at least partly responsible for the premium. The limits to arbitrage that exist (and the fact that human behavior has a strong tendency to persist) suggest the premiums are likely to persist in the future.
Larry Swedroe is the director of research for The BAM Alliance, a community of more than 140 independent registered investment advisors throughout the country.