In preparation for Keim, D. and W. Ziemba, editors, Security Market Imperfections and World Wide Equity Markets, Cambridge University Press.
Abstract
I present the view that the so called ``size effect" (the relation between realized returns and variables such as market value, book-to-market, earnings-to-price, etc.) should not even be classed as a puzzle, let alone an anomaly. Furthermore, while I argue that a derivative of this research, the development of sized-based factor models, might be useful in helping to explain the cross-sectional variation of expected returns, such ``models'' provide no information on the underlying economic cause of this variation. Finally, I argue that the size variables nevertheless have a useful role to play in tests of asset pricing theories.