Moments of Truth: Forecasting the Mean, Variance, and Skewness of Asset Returns

Abstract

Predicting the nature of future returns is fundamentally about what one believes about the future. And the tools we have for forming beliefs are deduction and induction. Rather than naively using historical return moments as our estimates for future moments, we ought to also use these tools. Three examples for how one might do this include (1) broadening the scope of data one considers, (2) being cognizant of the tendency of the central limit theorem, with portfolios of assets, to drive third and higher order moments towards zero, and (3) noticing how certain properties of second and third order moments can affect the first moment because of investors' behavioural biases.

Keywords: expected returns, inductive and deductive reasoning

JEL Classification: G12