The Performance of Characteristic-Sorted Portfolios: Evaluating the Past and Predicting the Future

Abstract

We study the performance of characteristic-sorted portfolios through the lens of a statistical model that allows for persistent variation in expected returns. Allowing for the possibility of time varying expected returns substantially weakens the evidence for long-run outperformance compared to standard approaches: the value, investment, and profitability portfolio returns have p-values above 9% in our maximum likelihood tests. We also use Bayesian analyses to examine the predictability of characteristic portfolio returns.  With relatively agnostic priors, our Bayesian posterior estimates exhibit large fluctuations in expected returns over time.

Keywords: Portfolio returns, time-variation, autocorrelation, value strategies, return predictability, Bayesian, standard errors

JEL Classification: G10, G11, G12, G14