Eugene Fama and Kenneth French’s 1992 paper, “The Cross-Section of Expected Stock Returns,” resulted in the development of the Fama–French three-factor model. This model added the size and value factors to the market beta factor.
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, and 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 in other definitions. For example, in the United States for the period 1952 through 2015, the annualized value premium was 4.1% (t-stat = 2.4) as measured by BtM, 4.7% (t-stat = 2.9) as measured by the cash flow-to-price ratio, and 6.3% (t-stat = 3.4) as measured by the earnings-to-price ratio.
Ray Ball, Joseph Gerakos, Juhani T. Linnainmaa and Valeri Nikolaev contribute to the literature on the value premium in their February 2017 paper, “Earnings, Retained Earnings, and Book-to-Market in the Cross Section of Expected Returns.” Their study, using Center for Research in Security Prices and Compustat data, covers the period July 1963 through December 2015. The sample excluded the bottom 20% of stocks (microcap stocks) as ranked by NYSE market capitalizations, as well as financials.
Note that excluding these financials is typical in academic papers. Omitting the smallest stocks is done so that these stocks don’t dominate results—they make up more than half the number of stocks but have much less market-cap weight. It’s also important to note that, typically, one sees stronger results for factors in these smaller-cap stocks. For example, using BtM, the value premium is much larger in small stocks than in large stocks.
Ball, Gerakos, Linnainmaa and Nikolaev begin by noting that book value of equity consists of two main parts: retained earnings (earnings less dividends) and contributed capital (the value of subsequent net share issuance). Following is a summary of their findings:
- Retained earnings-to-market subsumes book-to-market’s predictive power in the cross section of stock returns despite comprising, on average, only 42% of the book value of equity (book-to-market predicts the cross section of returns only because it contains retained earnings).
- The accumulated dividends component of retained earnings is uninformative of the cross section of average returns (another example of dividend policy not being relevant to expected returns). The authors state: “This result is consistent with book-to-market’s explanatory power arising only because it provides a good proxy for expected earnings yield.”
- Retained earnings-to-market is a signiﬁcant predictor of future earnings yield, and thus returns, while contributed capital has no predictive power.
- The value premium (the return on the high minus low portfolio) is greater using retained earnings-to-market than using book-to-market (43 versus 35 basis points per month) and has stronger statistical significance. Using only contributed capital, the value premium falls to 7 basis points per month and is no longer statistically significant (t-stat = 0.54).
- Retained earnings predicts the cross section of stock returns out to seven years, while book-to-market predicts returns only as far out as three years.
Relative to the value premium, this last finding provides an argument against the mispricing (behavioral-based) theory, and for the risk-based theory. As the authors point out: “Why would the correction of mispricing occur gradually over a horizon that extends at least as far as seven years, especially bearing in mind that all this accounting information is made publicly accessible at essentially zero cost for all ﬁrms and all years?”
Finally, the fact that book-to-market’s explanatory power comes from the earnings component highlights the potential importance of including price-to-earnings and cash flow metrics instead of just BtM (as some investment firms do, such as Bridgeway Capital Management and AQR).
Another alternative is to include the new profitability factor in construction of value portfolios (as Dimensional Fund Advisors does), as retained earnings contain information about future expected earnings/profitability. It will be interesting to see if these findings will be incorporated into portfolio construction rules of funds that seek to capture the value premium.
Larry Swedroe is the director of research for The BAM Alliance, a community of more than 140 independent registered investment advisors throughout the country.