Professor Florian Huber and Dr Niko Hauzenberger, University of Salzburg

"Coarse Bayesian VARs"
"Bayesian Modeling of Time-varying Parameters Using Regression Trees"
Friday, 17 March. 3 pm
Room 305 Main Buiding

Speaker: Professor Florian Huber
Title:  Coarse Bayesian VARs
Abstract: We propose a simple method for addressing issues related to model mis-specification in multivariate time series models. Our method, called coarse Bayesian VARs (cBVARs), replaces the exact likelihood with a coarsened likelihood that takes into account that our model might be mis-specified along important dimensions. Coupled with a conjugate prior, this results in a computationally simple model. As opposed to more flexible models, our approach avoids overfitting, is simple to implement, and estimation is fast. The resulting cBVAR performs well in simulations using highly non-Gaussian data generating processes. Applied to US data, cBVARs improve point and density forecasts compared to standard BVARs. Additionally, we demonstrate the use of cBVARs for analyzing the effects of uncertainty shocks.

 

Speaker: Dr Niko Hauzenberger
Title: Bayesian Modeling of Time-varying Parameters Using Regression Trees
Abstract: In light of widespread evidence of parameter instability in macroeconomic models, many time-varying parameter (TVP) models have been proposed. This paper proposes a nonparametric TVP-VAR model 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 parameter change, both in the conditional mean and in the conditional variance. In contrast to other nonparametric and machine learning methods that are black box, inference using our model is straightforward because, in treating the parameters rather than the variables nonparametrically, the model remains conditionally linear in the mean. Parsimony is achieved through adopting nonparametric factor structures and use of shrinkage priors. In an application to US macroeconomic data, we illustrate the use of our model in tracking 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.

 

Professor Huber’s  Bio

Florian Huber is a Professor of Economics at the University of Salzburg. His main research interest is on Bayesian econometrics, with a particular focus on large dimensional multivariate time series models, nonparametric methods and time-varying parameter regressions. His research has been published in top international journals such as the Journal of Business & Economic Statistics, the Journal of Econometrics, the International Economic Review, the Journal of Applied Econometrics, and the European Economic Review, among others.

Dr Niko Hauzenberger’s Bio

Niko Hauzenberger is currently a post-doctoral researcher in the Economics Department at the University of Salzburg, and he will start a new position as a Senior Lecturer in the Economics Department at the University of Strathclyde in fall 2023. His research focuses on the development of novel econometric methods for the efficient use of Big Data in macroeconomics. In doing so, he combines modelling techniques from the machine and Bayesian learning literature with multivariate time series models that macroeconomists commonly work with. His work has been published in the Journal of Business and Economic Statistics, the Journal of Applied Econometrics, and the International Journal of Forecasting, among others.


For further information, please contact business-school-research@glasgow.ac.uk

We foster a positive and productive environment for seminars through our Code of conduct.

First published: 7 March 2023

<< 2023