A Spline Analysis of the Small Firm Effect: Does Size Really Matter?

Abstract

This paper uses average monthly returns and linear spline regressions to investigate the relation between expected return and firm size during 1980-1994. We find that the average monthly returns are approximately constant across size deciles. The estimated spline regressions vary substantially from year-to-year. Our analysis of the year-by-year estimates suggests that the annual regression function is essentially flat, except possibly for the smallest two deciles. The results are similar for the January and non-January months. Hence, the evidence does not support the prevalent use of size as an explanatory variable for returns during the 1980-1994 period.

JEL Classification: C1, C2, C3, C4, C5, C8