Nowcasting in a Pandemic using Non-Parametric Mixed Frequency VARs
This work addresses the challenge of accurate economic forecasting during crises like the COVID-19 pandemic, which is crucial for policymakers and economists, though it is incremental as it builds on existing VAR methods with a novel adaptation.
The paper tackles the problem of macroeconomic nowcasting during the COVID-19 pandemic by developing Bayesian non-parametric mixed frequency VARs using additive regression trees, resulting in substantial improvements in nowcasting performance for four major euro area countries compared to a linear baseline.
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.