Predicting international stock returns with conditional price-to-fundamental ratios

Taking the perspective of international asset allocation, this paper tests if predictive regressions conditional on time-series and cross-sectional information can improve forecasts of stock index returns. We use different current price-to-fundamental ratios as predictors and condition the sample on the indicator if time-series and cross-section deliver consistent versus opposing signals. Using panel regressions, we find that only consistent ratios (i) display significant mean-reverting behavior, (ii) provide strong in-sample as well as out-of-sample evidence for return predictability, and (iii) yield economic gains in a Bayesian asset allocation framework.