Efficient Selection of Hyperparameters in Large Bayesian VARs Using Automatic Differentiation

24 Pages Posted: 25 Jun 2019

See all articles by Joshua C. C. Chan

Joshua C. C. Chan

University of Technology Sydney (UTS) - UTS Business School

Liana Jacobi

University of Melbourne - Faculty of Business and Economics; IZA Institute of Labor Economics

Dan Zhu

Monash University - Department of Econometrics & Business Statistics

Date Written: June 25, 2019

Abstract

Large Bayesian VARs with the natural conjugate prior are now routinely used for forecasting and structural analysis. It has been shown that selecting the prior hyperparameters in a data-driven manner can often substantially improve forecast performance. We propose a computationally efficient method to obtain the optimal hyperparameters based on Automatic Differentiation, which is an efficient way to compute derivatives. Using a large US dataset, we show that using the optimal hyperparameter values leads to substantially better forecast performance. Moreover, the proposed method is much faster than the conventional grid-search approach, and is applicable in high-dimensional optimization problems. The new method thus provides a practical and systematic way to develop better shrinkage priors for forecasting in a data-rich environment.

Keywords: automatic differentiation, model comparison, vector autoregression, factor models

JEL Classification: C11, C53, E37

Suggested Citation

Chan, Joshua C. C. and Jacobi, Liana and Zhu, Dan, Efficient Selection of Hyperparameters in Large Bayesian VARs Using Automatic Differentiation (June 25, 2019). CAMA Working Paper No. 46/2019. Available at SSRN: https://ssrn.com/abstract=3409550 or http://dx.doi.org/10.2139/ssrn.3409550

Joshua C. C. Chan (Contact Author)

University of Technology Sydney (UTS) - UTS Business School ( email )

Sydney
Australia

Liana Jacobi

University of Melbourne - Faculty of Business and Economics ( email )

Victoria, 3010
Australia

IZA Institute of Labor Economics

P.O. Box 7240
Bonn, D-53072
Germany

Dan Zhu

Monash University - Department of Econometrics & Business Statistics ( email )

Wellington Road
Clayton, Victoria 3168
Australia

Register to save articles to
your library

Register

Paper statistics

Downloads
5
Abstract Views
85
PlumX Metrics