Model Uncertainty in Panel Vector Autoregressive Models

Gary Koop, Dimitris Korobilis

Research output: Working paperDiscussion paper

Abstract

We develop methods for Bayesian model averaging (BMA) or selection (BMS) in Panel Vector Autoregressions (PVARs). Our approach allows us to select between or average over all possible combinations of restricted PVARs where the restrictions involve interdependencies between and heterogeneities across cross-sectional units. The resulting BMA framework can find a parsimonious PVAR specification, thus dealing with overparameterization concerns. We use these methods in an application involving the euro area sovereign debt crisis and show that our methods perform better than alternatives. Our findings contradict a simple view of the sovereign debt crisis which divides the euro zone into groups of core and peripheral countries and worries about financial contagion within the latter group.
LanguageEnglish
Place of PublicationGlasgow
PublisherUniversity of Strathclyde
Pages1-25
Number of pages26
Volume14
Publication statusPublished - Aug 2014

Fingerprint

Vector autoregressive model
Vector autoregression
Model uncertainty
Bayesian model averaging
Sovereign debt crises
Euro zone
Financial contagion
Euro area
Interdependencies
Model selection

Keywords

  • bayesian model averaging
  • stochastic search variable selection
  • financial contagion
  • sovereign debt crisis

Cite this

Koop, G., & Korobilis, D. (2014). Model Uncertainty in Panel Vector Autoregressive Models. (08 ed.) (pp. 1-25). Glasgow: University of Strathclyde.
Koop, Gary ; Korobilis, Dimitris. / Model Uncertainty in Panel Vector Autoregressive Models. 08. ed. Glasgow : University of Strathclyde, 2014. pp. 1-25
@techreport{0cbbf0c2fe26460cb1274dc6a2f6dff1,
title = "Model Uncertainty in Panel Vector Autoregressive Models",
abstract = "We develop methods for Bayesian model averaging (BMA) or selection (BMS) in Panel Vector Autoregressions (PVARs). Our approach allows us to select between or average over all possible combinations of restricted PVARs where the restrictions involve interdependencies between and heterogeneities across cross-sectional units. The resulting BMA framework can find a parsimonious PVAR specification, thus dealing with overparameterization concerns. We use these methods in an application involving the euro area sovereign debt crisis and show that our methods perform better than alternatives. Our findings contradict a simple view of the sovereign debt crisis which divides the euro zone into groups of core and peripheral countries and worries about financial contagion within the latter group.",
keywords = "bayesian model averaging, stochastic search variable selection, financial contagion, sovereign debt crisis",
author = "Gary Koop and Dimitris Korobilis",
note = "Published as a paper within the Discussion Papers in Economics, No. 14-08 (2014)",
year = "2014",
month = "8",
language = "English",
volume = "14",
pages = "1--25",
publisher = "University of Strathclyde",
edition = "08",
type = "WorkingPaper",
institution = "University of Strathclyde",

}

Koop, G & Korobilis, D 2014 'Model Uncertainty in Panel Vector Autoregressive Models' 08 edn, University of Strathclyde, Glasgow, pp. 1-25.

Model Uncertainty in Panel Vector Autoregressive Models. / Koop, Gary; Korobilis, Dimitris.

08. ed. Glasgow : University of Strathclyde, 2014. p. 1-25.

Research output: Working paperDiscussion paper

TY - UNPB

T1 - Model Uncertainty in Panel Vector Autoregressive Models

AU - Koop, Gary

AU - Korobilis, Dimitris

N1 - Published as a paper within the Discussion Papers in Economics, No. 14-08 (2014)

PY - 2014/8

Y1 - 2014/8

N2 - We develop methods for Bayesian model averaging (BMA) or selection (BMS) in Panel Vector Autoregressions (PVARs). Our approach allows us to select between or average over all possible combinations of restricted PVARs where the restrictions involve interdependencies between and heterogeneities across cross-sectional units. The resulting BMA framework can find a parsimonious PVAR specification, thus dealing with overparameterization concerns. We use these methods in an application involving the euro area sovereign debt crisis and show that our methods perform better than alternatives. Our findings contradict a simple view of the sovereign debt crisis which divides the euro zone into groups of core and peripheral countries and worries about financial contagion within the latter group.

AB - We develop methods for Bayesian model averaging (BMA) or selection (BMS) in Panel Vector Autoregressions (PVARs). Our approach allows us to select between or average over all possible combinations of restricted PVARs where the restrictions involve interdependencies between and heterogeneities across cross-sectional units. The resulting BMA framework can find a parsimonious PVAR specification, thus dealing with overparameterization concerns. We use these methods in an application involving the euro area sovereign debt crisis and show that our methods perform better than alternatives. Our findings contradict a simple view of the sovereign debt crisis which divides the euro zone into groups of core and peripheral countries and worries about financial contagion within the latter group.

KW - bayesian model averaging

KW - stochastic search variable selection

KW - financial contagion

KW - sovereign debt crisis

M3 - Discussion paper

VL - 14

SP - 1

EP - 25

BT - Model Uncertainty in Panel Vector Autoregressive Models

PB - University of Strathclyde

CY - Glasgow

ER -

Koop G, Korobilis D. Model Uncertainty in Panel Vector Autoregressive Models. 08 ed. Glasgow: University of Strathclyde. 2014 Aug, p. 1-25.