On the Economic Value of Stock Market Return Predictors
54 Pages Posted: 10 May 2019
Date Written: November 17, 2018
Kandel and Stambaugh (1996) demonstrate that forecasting variables with weak statistical support in predictive return regressions can exert considerable economic influence on portfolio decisions. Using a Bayesian vector autoregression framework with stochastic volatility in market returns and predictor variables, we assess the economic value of return predictability and reach a complementary conclusion. Statistically strong predictors can be economically unimportant if they tend to take extreme values in high-volatility periods, have low persistence, and/or follow distributions with fat tails. Several popular predictors exhibit these properties such that their impressive statistical results do not translate into large economic gains for investors.
Keywords: Market return predictability, Bayesian investors, Stochastic volatility, Multiperiod horizons
JEL Classification: G10, G11, G12
Suggested Citation: Suggested Citation