Factoring Characteristics into Returns: A Clinical Approach to Fama-French Portfolio Decomposition
58 Pages Posted: 29 May 2018 Last revised: 25 Jun 2018
Date Written: June 20, 2018
This paper demonstrates the biases introduced by unequal allocation of stocks into portfolios when building risk factors under the framework of an independent sort. We show further that these biases make a significant difference in producing mean-variance efficient portfolios. The biases induced by an independent and asymmetric sorting on-NYSE breakpoints can become very severe in specific market conditions. Mean-variance factor efficiency proves to be highly sensitive to the number of stocks into portfolios, which itself depends on the economic context. Our quasi-clinical investigation examines three methodological choices that have an impact on portfolio diversification: the (in)dependence and the (a)symmetry of the stock sorting procedure, and the sorting breakpoints. To minimize biases coming from unpriced risks, the optimal combination of these methodological criteria is unambiguously a sequential sort that starts with the control variables (pre-conditioning) with whole sample breakpoints.
Keywords: asset pricing, characteristics, attribute-matched portfolios, size, value
JEL Classification: G10, G11, G12
Suggested Citation: Suggested Citation