This paper studies the relation between 36 firm-level characteristics and stock returns in 48 countries using instrumented principal components analysis (IPCA). A non-U.S. country-neutral conditional factor model performs well in describing risk and returns and generates small and statistically insignificant anomaly intercepts when allowing for three or more latent factors. The non-U.S. model performs better in emerging than in developed markets, while showing substantial differences across countries. On average, only ten characteristics contribute significantly to the models' performance. Market beta, momentum and firm size characteristics instrument for systemic exposure in U.S. and non-U.S. models, while investment and book-to-market do not.     «

This paper studies the relation between 36 firm-level characteristics and stock returns in 48 countries using instrumented principal components analysis (IPCA). A non-U.S. country-neutral conditional factor model performs well in describing risk and returns and generates small and statistically insignificant anomaly intercepts when allowing for three or more latent factors. The non-U.S. model performs better in emerging than in developed markets, while showing substantial differences across coun...     »