Abstract:
Using monthly South African data for 1990:01-2009:10, this paper, to the best of our knowledge, is the first to examine the predictability of real stock prices based on valuation ratios, namely, price-dividend and price-earnings ratios. We cannot detect either short-horizon or long-horizon predictability; that is, the hypothesis that the current value of a valuation ratio is uncorrelated with future stock price changes cannot be rejected at both short- and long-horizons based on bootstrapped critical values constructed from linear representations of the data. We find, via Monte Carlo simulations, that the power to detect predictability in finite samples tends to decrease at long horizons in a linear framework. Though Monte Carlo simulations applied to exponential smooth-transition autoregressive (ESTAR) models of the price-dividend and price-earnings ratios, show increased power, the ability of the non-linear framework in explaining the pattern of stock price predictability in the data does not show any promise both at short- and long-horizons, just as in the linear predictive regressions.
