...And Nothing Else Matters? On the Dimensionality and Predictability of International Stock Returns

Abstract

We explore the dimensionality of stock returns in North America, Europe, Japan, Pacific, and Emerging Markets on the basis of 240 cross-sectional predictors. Our approach allows us to identify those predictors that are most consistently related to nonmicro-cap stock returns (i.e., independent of other predictors, adjusted for data mining, existent in different time periods, and across regions). There is a large geographic heterogeneity in the significance of individual characteristics and in time trends, which leads to substantial out-of-sample diversification gains for global multidimensional hedge portfolios. Our results are most consistent with the mispricing hypothesis for anomalies.

Keywords: Anomalies, international stock markets, multidimensionality, market efficiency, return predictability

JEL Classification: G11, G12, G14