Equity Premium Prediction with Bagged Machine Learning
AFA 2020
AsianFA 2019, AMES 2019, FMND 2019
56 Pages Posted: 8 Jan 2019 Last revised: 5 Dec 2020
Date Written: December 5, 2020
Abstract
We introduce a variation of Yu(2011)'s weighted bagging estimation method and show it substantially improves the predictability of the equity premium and other economic variables. This new machine learning method sharply improves equity premium predictability of many models with significant monthly out-of-sample R2 up to almost 3% and annual utility gains of more than 3.5%. The improved predictive performance stems from better performance during periods of economic recession and market turbulence and downturns, as well as increased diversity and built-in shrinkage of our weighted bagging method. Interest rate related variables show the strongest predictive ability for the equity premium.
Keywords: Equity premium, Out-of-sample prediction, Instability, Machine learning, Weighted bagging
JEL Classification: G17, G12, G02, C58
Suggested Citation: Suggested Citation