Portfolio management using a factor‐analytic stock selection strategy

Abstract

This study takes an integrated look at six widely documented price‐related CAPM anomaly variables. Using maximum likelihood factor analysis, we extract factors common to these variables. We find that portfolios formed according to the first extracted factor alone exhibit abnormal performance. Further, the performance of firms ranked on the basis of extracted factor scores for this factor is superior to that of firms selected on the basis of any one of the six variables. Our results have implications for (1) isolating missing factors in the CAPM specification and (2) designing dynamic portfolio strategies aimed at jointly exploiting more than one anomaly.