Macro Factor-Mimicking Portfolios

Abstract

The estimation of risk factors and their replication through mimicking portfolios are of critical importance for academics and practitioners in finance. Starting from a general optimization framework that encompasses existing factor-mimicking approaches, we incorporate machine learning estimation improvements to mitigate the impact of estimation errors in the observed factors. The empirical application for the primary major macroeconomic factors (growth and inflation) over the period 1974-2022 proves the superiority of the approach.

Keywords: Macroeconomics, factor-mimicking portfolios, portfolio optimization, machine learning

JEL Classification: G11, D81, C60