Nowcasting in a Pandemic using Non-Parametric Mixed Frequency VARs

TL;DR

This paper introduces Bayesian non-parametric mixed frequency VARs using additive regression trees, enhancing macroeconomic nowcasting accuracy during extreme events like COVID-19 by effectively modeling outliers.

Contribution

It develops a novel Bayesian non-parametric approach with regression trees for mixed frequency VARs, improving nowcasting during extreme macroeconomic shocks.

Findings

Significant improvement in nowcasting accuracy for euro area countries.

Regression tree models better handle outliers caused by pandemic shocks.

Demonstrates robustness of the proposed method in extreme conditions.

Abstract

This paper develops Bayesian econometric methods for posterior inference in non-parametric mixed frequency VARs using additive regression trees. We argue that regression tree models are ideally suited for macroeconomic nowcasting in the face of extreme observations, for instance those produced by the COVID-19 pandemic of 2020. This is due to their flexibility and ability to model outliers. In an application involving four major euro area countries, we find substantial improvements in nowcasting performance relative to a linear mixed frequency VAR.