Bayesian Compressed Vector Autoregressions

35 Pages Posted: 26 Mar 2016 Last revised: 6 Jun 2017

See all articles by Gary Koop

Gary Koop

University of Strathclyde, Glasgow - Strathclyde Business School - Department of Economics

Dimitris Korobilis

affiliation not provided to SSRN

Davide Pettenuzzo

Brandeis University - Department of Economics

Date Written: June 5, 2017

Abstract

Macroeconomists are increasingly working with large Vector Autoregressions (VARs) where the number of parameters vastly exceeds the number of observations. Existing approaches either involve prior shrinkage or the use of factor methods. In this paper, we develop an alternative based on ideas from the compressed regression literature. It involves randomly compressing the explanatory variables prior to analysis. A huge dimensional problem is thus turned into a much smaller, more computationally tractable one. Bayesian model averaging can be done over various compressions, attaching greater weight to compressions which forecast well. In a macroeconomic application involving up to 129 variables, we find compressed VAR methods to forecast better than either factor methods or large VAR methods involving prior shrinkage.

Keywords: multivariate time series, random projection, forecasting

JEL Classification: C11, C32, C53

Suggested Citation

Koop, Gary and Korobilis, Dimitris and Pettenuzzo, Davide, Bayesian Compressed Vector Autoregressions (June 5, 2017). Available at SSRN: https://ssrn.com/abstract=2754241 or http://dx.doi.org/10.2139/ssrn.2754241

Gary Koop

University of Strathclyde, Glasgow - Strathclyde Business School - Department of Economics ( email )

100 Cathedral Street
Glasgow G4 0LN
United Kingdom

Dimitris Korobilis

affiliation not provided to SSRN

Davide Pettenuzzo (Contact Author)

Brandeis University - Department of Economics ( email )

Waltham, MA 02454-9110
United States
781.736.2834 (Phone)
781.736.2269 (Fax)

Register to save articles to
your library

Register

Paper statistics

Downloads
415
rank
67,935
Abstract Views
1,254
PlumX Metrics