Is the size premium really driven by firm size?

Abstract

By decomposing firm size into horizon-based components, the authors find that the changes in firm size in prior years, instead of its recent level, drive the size premium. Specifically, size five years ago explains 80% of the current firm size but has little predictive power for the size premium. In contrast, the change in size over the prior two to five years explains only 18% of the size but almost completely captures the size premium. Their decomposition indicates that not all small stocks earn a size premium. Only small stocks that had significant losses in market value in the prior two to five years earn a premium. This analysis also offers new insights into the disappearance of the siz..