Large Bayesian VARMAs

Joshua C C Chan, Eric Eisenstat, Gary Koop

Research output: Working paperDiscussion paper

Abstract

Vector Autoregressive Moving Average (VARMA) models have many theoretical properties which should make them popular among empirical macroeconomists. However, they are rarely used in practice due to over-parameterization concerns, difficulties in ensuring identification and computational challenges. With the growing interest in multivariate time series models of high dimension, these problems with VARMAs become even more acute, accounting for the dominance of VARs in this field. In this paper, we develop a Bayesian approach for inference in VARMAs which surmounts these problems. It jointly ensures identification and parsimony in the context of an efficient Markov chain Monte Carlo (MCMC) algorithm. We use this approach in a macroeconomic application involving up to twelve dependent variables. We find our algorithm to work successfully and provide insights beyond those provided by VARs.
LanguageEnglish
Place of PublicationGlasgow
PublisherUniversity of Strathclyde
Pages1-41
Number of pages42
Volume14
Publication statusPublished - 25 Sep 2014

Fingerprint

Markov chain Monte Carlo
Autoregressive moving average model
Vector autoregressive
Bayesian approach
Inference
Multivariate time series
Macroeconomics
Parsimony
Time series models

Keywords

  • varma identification
  • markov chain monte carlo
  • bayesian
  • stochastic search variable selection

Cite this

Chan, J. C. C., Eisenstat, E., & Koop, G. (2014). Large Bayesian VARMAs. (09 ed.) (pp. 1-41). Glasgow: University of Strathclyde.
Chan, Joshua C C ; Eisenstat, Eric ; Koop, Gary. / Large Bayesian VARMAs. 09. ed. Glasgow : University of Strathclyde, 2014. pp. 1-41
@techreport{f48a86a6c3784d1cbb2fc4a9464fda40,
title = "Large Bayesian VARMAs",
abstract = "Vector Autoregressive Moving Average (VARMA) models have many theoretical properties which should make them popular among empirical macroeconomists. However, they are rarely used in practice due to over-parameterization concerns, difficulties in ensuring identification and computational challenges. With the growing interest in multivariate time series models of high dimension, these problems with VARMAs become even more acute, accounting for the dominance of VARs in this field. In this paper, we develop a Bayesian approach for inference in VARMAs which surmounts these problems. It jointly ensures identification and parsimony in the context of an efficient Markov chain Monte Carlo (MCMC) algorithm. We use this approach in a macroeconomic application involving up to twelve dependent variables. We find our algorithm to work successfully and provide insights beyond those provided by VARs.",
keywords = "varma identification, markov chain monte carlo, bayesian, stochastic search variable selection",
author = "Chan, {Joshua C C} and Eric Eisenstat and Gary Koop",
note = "Published as a paper within the Discussion Papers in Economics, No. 14-09 (2014)",
year = "2014",
month = "9",
day = "25",
language = "English",
volume = "14",
pages = "1--41",
publisher = "University of Strathclyde",
edition = "09",
type = "WorkingPaper",
institution = "University of Strathclyde",

}

Chan, JCC, Eisenstat, E & Koop, G 2014 'Large Bayesian VARMAs' 09 edn, University of Strathclyde, Glasgow, pp. 1-41.

Large Bayesian VARMAs. / Chan, Joshua C C; Eisenstat, Eric ; Koop, Gary.

09. ed. Glasgow : University of Strathclyde, 2014. p. 1-41.

Research output: Working paperDiscussion paper

TY - UNPB

T1 - Large Bayesian VARMAs

AU - Chan, Joshua C C

AU - Eisenstat, Eric

AU - Koop, Gary

N1 - Published as a paper within the Discussion Papers in Economics, No. 14-09 (2014)

PY - 2014/9/25

Y1 - 2014/9/25

N2 - Vector Autoregressive Moving Average (VARMA) models have many theoretical properties which should make them popular among empirical macroeconomists. However, they are rarely used in practice due to over-parameterization concerns, difficulties in ensuring identification and computational challenges. With the growing interest in multivariate time series models of high dimension, these problems with VARMAs become even more acute, accounting for the dominance of VARs in this field. In this paper, we develop a Bayesian approach for inference in VARMAs which surmounts these problems. It jointly ensures identification and parsimony in the context of an efficient Markov chain Monte Carlo (MCMC) algorithm. We use this approach in a macroeconomic application involving up to twelve dependent variables. We find our algorithm to work successfully and provide insights beyond those provided by VARs.

AB - Vector Autoregressive Moving Average (VARMA) models have many theoretical properties which should make them popular among empirical macroeconomists. However, they are rarely used in practice due to over-parameterization concerns, difficulties in ensuring identification and computational challenges. With the growing interest in multivariate time series models of high dimension, these problems with VARMAs become even more acute, accounting for the dominance of VARs in this field. In this paper, we develop a Bayesian approach for inference in VARMAs which surmounts these problems. It jointly ensures identification and parsimony in the context of an efficient Markov chain Monte Carlo (MCMC) algorithm. We use this approach in a macroeconomic application involving up to twelve dependent variables. We find our algorithm to work successfully and provide insights beyond those provided by VARs.

KW - varma identification

KW - markov chain monte carlo

KW - bayesian

KW - stochastic search variable selection

M3 - Discussion paper

VL - 14

SP - 1

EP - 41

BT - Large Bayesian VARMAs

PB - University of Strathclyde

CY - Glasgow

ER -

Chan JCC, Eisenstat E, Koop G. Large Bayesian VARMAs. 09 ed. Glasgow: University of Strathclyde. 2014 Sep 25, p. 1-41.