Evaluating the effectiveness of Common-Factor Portfolios

In this paper we use the standard factor models to compose common-factor portfolios by a novel linear transformation extracted from large data sets of asset returns. Although the transformation proposed here retains the basic properties of the usual common factors, some interesting new properties are further included in them. The advantages of using common-factor portfolios in asset pricing are: (i) they produce a dimension reduction in the asset- pricing data-base while preserving the usual restrictions imposed by the asset-pricing equation, and (ii) from the empirical perspective, their performance is better than those of standard factor models. The practical importance is confirmed in two applications: the performance of common-factor portfolios is shown to be superior to those of the asset returns and factors commonly used in the finance literature.