Are Macroeconomic Factors Adequate Proxies for Systematic Influences in Stock Returns?

Abstract

This paper investigates whether pre-specified macroeconomic factors can adequately proxy for the pervasive influences in stock returns, within the context of macroeconomic linear factor models motivated by the multifactor Arbitrage Pricing Theory (APT). Variation in stock returns can be attributed to systematic and idiosyncratic sources of variation. As idiosyncratic factors can be diversified away, systematic variation will remain and the only factors that will be relevant will be those representative of systematic influences. In this study, systematic influences are quantified by statistically derived factor scores which are then related to a set of carefully selected macroeconomic factors. The identification of macroeconomic factors that proxy for systematic influences in returns is a challenge in itself. Once identified, macroeconomic factors are found to be poor and unstable proxies for systematic influences. The use of a residual market factor, an often-applied solution the factor omission problem in linear factor models motivated by the APT, does not significantly improve the approximation of factor scores. Macroeconomic factors are unlikely to provide a complete representation of the return generating process. Researchers should recognize that macroeconomic linear factor models are likely to be underspecified, even if a residual market factor is included.