A0204
Title: Bayesian modeling of time-varying parameters using regression trees
Authors: Florian Huber - University of Salzburg (Austria) [presenting]
Niko Hauzenberger - University of Strathclyde (United Kingdom)
James Mitchell - Federal Reserve Bank of Cleveland (United States)
Gary Koop - University of Strathclyde (United Kingdom)
Abstract: In light of widespread evidence of parameter instability in macroeconomic models, many time-varying parameter (TVP) models have been proposed. A nonparametric TVP-VAR model is proposed using Bayesian Additive Regression Trees (BART). The novelty of this model arises from the law of motion driving the parameters being treated nonparametrically. This leads to great flexibility in the nature and extent of a parameter change, both in the conditional mean and in the conditional variance. In contrast to other nonparametric methods that are black box, structural inference using our model is straightforward. Parsimony is achieved by adopting nonparametric factor structures and the use of shrinkage priors. In an application to US macroeconomic data, we illustrate the use of our model in understanding both the evolving nature of the Phillips Curve and how the effects of business cycle shocks on inflationary measures vary nonlinearly with movements in uncertainty.
