The ability of a V/P ratio to predict cross-sectional returns has been confirmed by
other studies. For example, Herzberg (1998) shows the Frankel and Lee (1998) results
can be improved using somewhat more refined model estimation procedures. Dechow
et al. (1999) implements the RIM using a number of different time-series models for
predicting future ROEs. They also find that a V/P ratio estimated in this manner predicts
cross-sectional returns. In both studies, firms with higher (lower) V/P ratios earn
higher (lower) future returns, particularly over horizons of three to five years.
Lee et al. (1999) examine the question of how value estimates based on accounting
numbers should be evaluated when the stock market price itself is also a noisy measure
for the true intrinsic value. Rather than assuming price is always equal to intrinsic
value, they model price and value as a cointegrated system—in other words, both the
observed price and the accountant's estimate of value are constantly converging to the
true (but unobservable) intrinsic value. They show that in this framework, under fairly
general conditions, better value estimates will not only be more correlated with contemporaneous
returns, but will also yield better predictions of future returns. The authors
compute a RIM-based value estimate for the 30 stocks in the Dow Jones Industrial
Average (DJIA). They find that an aggregate V/P ratio for the 30 Dow stocks has
significant predictive power for overall market returns in the U.S.