Where are the Factors in Factor Investing?

Abstract

In this article, I advocate for the use of causal graphs to modernize the field of factor investing, and set it on a logically-coherent foundation. In order to do that, first I must introduce the concepts of association and causations. Second, I explain the use of causal graphs and the real (causal) meaning of the “ceteris paribus” assumption that is so popular among economists. Third, I explain how causal graphs help us estimate causal effects in observational (non-experimental) studies. Fourth, I illustrate all of the earlier concepts with Monte Carlo experiments. Fifth, I conclude that the field of factor investing must embrace causal graphs in order to wake up from its associational slumber.

Keywords: Association, causation, causal inference, causal discovery, causal mechanism, confounder, factor investing, backtest overfitting

JEL Classification: G0, G1, G2, G15, G24, E44