The disappearing size effect

Abstract

In this paper, we use average monthly returns and monthly cross-sectional regressions to investigate the relation between returns and firm size. During the period 1963-1981, we find an annualized return difference between small and large firms over 13% compared to a negative 2% return differential since 1982. Removal of the smallest firms (less than $5 million market value) eliminates any statistically significant size effect during the sample period using a regression framework. Several explanations are proposed for the disappearance of the size effect. Our results imply that size should not be considered as a systematic proxy for risk. © 2000 University of Venice.