According to conventional financial theory, competition among diverse investors and arbitrage should keep stock prices close to their “fair” value. But a growing body of research indicates that psychological factors, “noise” trading and fads in investment styles can cause stock prices to deviate from value, and that such departures can be significant and long-lasting. Furthermore, there is now substantial evidence that return regularities are associated with equity attributes. In a market that is not price-efficient, value as measured by a dividend discount model (DDM) is but a small part of the security pricing story.
An examination of the relation between DDM expected returns and 25 equity attributes, using multivariate regression and data from over 1000 stocks during the mid-1982 to mid-1987 period, reveals whether certain equity attributes tend to be favored by DDM models. Low P/E, sales/price, yield, zero yield, beta, sigma and trends in analysts’ earnings estimates were all positively correlated with DDM expected return. Coskewness, small size, residual-return reversal and high book/price were negatively associated. Neglect and earnings surprise were uncorrelated with DDM expected return, although they were associated with anomalous returns. A DDM strategy cannot be replicated with equity attributes alone, nor can all anomalous returns be captured with a DDM.
An examination of actual security returns over the five-year period shows that a DDM strategy would have produced positive but insignificant returns. At times, the DDM was perverse, with DDM expected returns negatively correlated with actual returns. In a bivariate regression pitting the DDM against low P/E, low P/E provided a higher payoff and was significant in more quarters than the DDM; low P/E appears to subsume some portion of the DDM, rather than the reverse. A full multivariate regression considering the DDM simultaneously with 25 equity attributes showed the DDM to be insignificant, while many equity attributes—sales/price, neglect, relative strength, residual-return reversal, trends in analysts’ estimates and earnings surprise—provided statistically significant abnormal performance.
Equity attributes such as P/E are not mere proxies for value. Many attributes are better predictors of return than the DDM. In fact, the results suggest that DDM value is nothing more than an additional equity attribute and, like other attributes, may be amenable to prediction.
