Tail Forecasting with Multivariate Bayesian Additive Regression Trees

RePEc Short ID: pcl55
ORCID iD: 0000-0002-4985-1709

Clark, Todd E.; Huber, Florian; Koop, Gary; Marcellino, Massimiliano; Pfarrhofer, Michael

doi:10.26509/frbc-wp-202108r

Working Paper Revision

Tail Forecasting with Multivariate Bayesian Additive Regression Trees

Abstract: We develop multivariate time series models using Bayesian additive regression trees that posit nonlinearities among macroeconomic variables, their lags, and possibly their lagged errors. The error variances can be stable, feature stochastic volatility, or follow a nonparametric specification. We evaluate density and tail forecast performance for a set of US macroeconomic and financial indicators. Our results suggest that the proposed models improve forecast accuracy both overall and in the tails. Another finding is that when allowing for nonlinearities in the conditional mean, heteroskedasticity becomes less important. A scenario analysis reveals nonlinear relations between predictive distributions and financial conditions.

Keywords: Nonparametric VAR; regression trees; macroeconomic forecasting; scenario analysis;

JEL Classification: C11; C32; C53;

https://doi.org/10.26509/frbc-wp-202108r