Forecasting with Medium and Large Bayesian VARS

Gary Koop

Research output: Working paper › Discussion paper

Abstract

This paper is motivated by the recent interest in the use of Bayesian VARs for forecasting, even in cases where the number of dependent variables is large. In such cases, factor methods have been traditionally used but recent work using a particular prior suggests that Bayesian VAR methods can forecast better. In this paper, we consider a range of alternative priors which have been used with small VARs, discuss the issues which arise when they are used with medium and large VARs and examine their forecast performance using a US macroeconomic data set containing 168 variables. We find that Bayesian VARs do tend to forecast better than factor methods and provide an extensive comparison of the strengths and weaknesses of various approaches. Our empirical results show the importance of using forecast metrics which use the entire predictive density, instead of using only point forecasts.
LanguageEnglish
Place of PublicationGlasgow
PublisherUniversity of Strathclyde
Pages1-34
Number of pages35
Volume2011
Publication statusPublished - Feb 2010

Fingerprint

Bayesian VAR
Factors
Empirical results
Macroeconomics
Point forecasts
Predictive density
Forecast performance

Keywords

  • bayesian
  • Minnesota prior
  • stochastic search variable selection
  • predictive likelihood

Cite this

Koop, G. (2010). Forecasting with Medium and Large Bayesian VARS. (17 ed.) (pp. 1-34). Glasgow: University of Strathclyde.
Koop, Gary. / Forecasting with Medium and Large Bayesian VARS. 17. ed. Glasgow : University of Strathclyde, 2010. pp. 1-34
@techreport{4398d54527054b1883e26b0294e1bdf1,
title = "Forecasting with Medium and Large Bayesian VARS",
abstract = "This paper is motivated by the recent interest in the use of Bayesian VARs for forecasting, even in cases where the number of dependent variables is large. In such cases, factor methods have been traditionally used but recent work using a particular prior suggests that Bayesian VAR methods can forecast better. In this paper, we consider a range of alternative priors which have been used with small VARs, discuss the issues which arise when they are used with medium and large VARs and examine their forecast performance using a US macroeconomic data set containing 168 variables. We find that Bayesian VARs do tend to forecast better than factor methods and provide an extensive comparison of the strengths and weaknesses of various approaches. Our empirical results show the importance of using forecast metrics which use the entire predictive density, instead of using only point forecasts.",
keywords = "bayesian, Minnesota prior, stochastic search variable selection, predictive likelihood",
author = "Gary Koop",
note = "Discussion paper.",
year = "2010",
month = "2",
language = "English",
volume = "2011",
pages = "1--34",
publisher = "University of Strathclyde",
edition = "17",
type = "WorkingPaper",
institution = "University of Strathclyde",

}

Koop, G 2010 'Forecasting with Medium and Large Bayesian VARS' 17 edn, University of Strathclyde, Glasgow, pp. 1-34.

Forecasting with Medium and Large Bayesian VARS. / Koop, Gary.

17. ed. Glasgow : University of Strathclyde, 2010. p. 1-34.

Research output: Working paper › Discussion paper

TY - UNPB

T1 - Forecasting with Medium and Large Bayesian VARS

AU - Koop, Gary

N1 - Discussion paper.

PY - 2010/2

Y1 - 2010/2

N2 - This paper is motivated by the recent interest in the use of Bayesian VARs for forecasting, even in cases where the number of dependent variables is large. In such cases, factor methods have been traditionally used but recent work using a particular prior suggests that Bayesian VAR methods can forecast better. In this paper, we consider a range of alternative priors which have been used with small VARs, discuss the issues which arise when they are used with medium and large VARs and examine their forecast performance using a US macroeconomic data set containing 168 variables. We find that Bayesian VARs do tend to forecast better than factor methods and provide an extensive comparison of the strengths and weaknesses of various approaches. Our empirical results show the importance of using forecast metrics which use the entire predictive density, instead of using only point forecasts.

AB - This paper is motivated by the recent interest in the use of Bayesian VARs for forecasting, even in cases where the number of dependent variables is large. In such cases, factor methods have been traditionally used but recent work using a particular prior suggests that Bayesian VAR methods can forecast better. In this paper, we consider a range of alternative priors which have been used with small VARs, discuss the issues which arise when they are used with medium and large VARs and examine their forecast performance using a US macroeconomic data set containing 168 variables. We find that Bayesian VARs do tend to forecast better than factor methods and provide an extensive comparison of the strengths and weaknesses of various approaches. Our empirical results show the importance of using forecast metrics which use the entire predictive density, instead of using only point forecasts.

KW - bayesian

KW - Minnesota prior

KW - stochastic search variable selection

KW - predictive likelihood

M3 - Discussion paper

VL - 2011

SP - 1

EP - 34

BT - Forecasting with Medium and Large Bayesian VARS

PB - University of Strathclyde

CY - Glasgow

ER -

Koop G. Forecasting with Medium and Large Bayesian VARS. 17 ed. Glasgow: University of Strathclyde. 2010 Feb, p. 1-34.

Access to Document