When Are Stocks Less Volatile in the Long Run?

Abstract

Pastor and Stambaugh (2012) demonstrate that from a forward-looking perspective, stocks are more volatile in the long run than they are in the short run. We investigate how the economic constraint of non-negative equity premia aspects predictive variance. When investors expect non-negative returns in the market and thus impose the constraint on predictive regressions, they find that stocks are less volatile in the long run, even after taking account of estimation risk and uncertainties on current and future expected stock returns because the constraint provides additional parameter identification condition and prior information for future returns. Thus, it substantially reduces uncertainty on future stock returns. This fact, combined with the mean reversion property of stock return dynamics, leads to lower predictive variance in the long run.

Keywords: Bayesian method, predictive variance, non-negative equity premium

JEL Classification: G1, C11