Bayesian Analysis of ARMA Models

Frank R. Kleibergen, Henk Hoek

Research output: Working paperProfessional

Abstract

Root cancellation in Auto Regressive Moving Average (ARMA) models leads tolocal non-identification of parameters. When we use diffuse or normal priorson the parameters of the ARMA model, posteriors in Bayesian analyzes show ana posteriori favor for this local non-identification. We show that the priorand posterior of the parameters of an ARMA model are the (unique)conditional density of a prior and posterior of the parameters of anencompassing AR model. We can therefore specify priors and posteriors on theparameters of the encompassing AR model and use the prior and posterior thatit implies on the parameters of the ARMA model, and vice versa. Theposteriors of the ARMA parameters that result from standard priors on theparameters of an encompassing AR model do not lead to an a posteriori favorof root cancellation. We develop simulators to generate parameters fromthese priors and posteriors. As a byproduct, Bayes factors can be computedto compare (non-nested) parsimonious ARMA models. The procedures are appliedto the (extended) Nelson-Plosser data. For approximately 50% of the seriesan ARMA model is favored above an AR model.
Original languageEnglish
Place of PublicationAmsterdam
PublisherTinbergen Instituut
Publication statusPublished - 2000

Publication series

NameDiscussion paper TI
No.00-027/4

Fingerprint

Autoregressive Moving Average Model
Bayesian Analysis
AR Model
Cancellation
Roots
Bayes Factor
Conditional Density
Autoregressive Moving Average
Simulator
Imply

Cite this

Kleibergen, F. R., & Hoek, H. (2000). Bayesian Analysis of ARMA Models. (Discussion paper TI; No. 00-027/4). Amsterdam: Tinbergen Instituut.
Kleibergen, Frank R. ; Hoek, Henk. / Bayesian Analysis of ARMA Models. Amsterdam : Tinbergen Instituut, 2000. (Discussion paper TI; 00-027/4).
@techreport{d7582c2299894aecba24d0ee52a8a70c,
title = "Bayesian Analysis of ARMA Models",
abstract = "Root cancellation in Auto Regressive Moving Average (ARMA) models leads tolocal non-identification of parameters. When we use diffuse or normal priorson the parameters of the ARMA model, posteriors in Bayesian analyzes show ana posteriori favor for this local non-identification. We show that the priorand posterior of the parameters of an ARMA model are the (unique)conditional density of a prior and posterior of the parameters of anencompassing AR model. We can therefore specify priors and posteriors on theparameters of the encompassing AR model and use the prior and posterior thatit implies on the parameters of the ARMA model, and vice versa. Theposteriors of the ARMA parameters that result from standard priors on theparameters of an encompassing AR model do not lead to an a posteriori favorof root cancellation. We develop simulators to generate parameters fromthese priors and posteriors. As a byproduct, Bayes factors can be computedto compare (non-nested) parsimonious ARMA models. The procedures are appliedto the (extended) Nelson-Plosser data. For approximately 50{\%} of the seriesan ARMA model is favored above an AR model.",
author = "Kleibergen, {Frank R.} and Henk Hoek",
year = "2000",
language = "English",
series = "Discussion paper TI",
publisher = "Tinbergen Instituut",
number = "00-027/4",
type = "WorkingPaper",
institution = "Tinbergen Instituut",

}

Kleibergen, FR & Hoek, H 2000 'Bayesian Analysis of ARMA Models' Discussion paper TI, no. 00-027/4, Tinbergen Instituut, Amsterdam.

Bayesian Analysis of ARMA Models. / Kleibergen, Frank R.; Hoek, Henk.

Amsterdam : Tinbergen Instituut, 2000. (Discussion paper TI; No. 00-027/4).

Research output: Working paperProfessional

TY - UNPB

T1 - Bayesian Analysis of ARMA Models

AU - Kleibergen, Frank R.

AU - Hoek, Henk

PY - 2000

Y1 - 2000

N2 - Root cancellation in Auto Regressive Moving Average (ARMA) models leads tolocal non-identification of parameters. When we use diffuse or normal priorson the parameters of the ARMA model, posteriors in Bayesian analyzes show ana posteriori favor for this local non-identification. We show that the priorand posterior of the parameters of an ARMA model are the (unique)conditional density of a prior and posterior of the parameters of anencompassing AR model. We can therefore specify priors and posteriors on theparameters of the encompassing AR model and use the prior and posterior thatit implies on the parameters of the ARMA model, and vice versa. Theposteriors of the ARMA parameters that result from standard priors on theparameters of an encompassing AR model do not lead to an a posteriori favorof root cancellation. We develop simulators to generate parameters fromthese priors and posteriors. As a byproduct, Bayes factors can be computedto compare (non-nested) parsimonious ARMA models. The procedures are appliedto the (extended) Nelson-Plosser data. For approximately 50% of the seriesan ARMA model is favored above an AR model.

AB - Root cancellation in Auto Regressive Moving Average (ARMA) models leads tolocal non-identification of parameters. When we use diffuse or normal priorson the parameters of the ARMA model, posteriors in Bayesian analyzes show ana posteriori favor for this local non-identification. We show that the priorand posterior of the parameters of an ARMA model are the (unique)conditional density of a prior and posterior of the parameters of anencompassing AR model. We can therefore specify priors and posteriors on theparameters of the encompassing AR model and use the prior and posterior thatit implies on the parameters of the ARMA model, and vice versa. Theposteriors of the ARMA parameters that result from standard priors on theparameters of an encompassing AR model do not lead to an a posteriori favorof root cancellation. We develop simulators to generate parameters fromthese priors and posteriors. As a byproduct, Bayes factors can be computedto compare (non-nested) parsimonious ARMA models. The procedures are appliedto the (extended) Nelson-Plosser data. For approximately 50% of the seriesan ARMA model is favored above an AR model.

M3 - Working paper

T3 - Discussion paper TI

BT - Bayesian Analysis of ARMA Models

PB - Tinbergen Instituut

CY - Amsterdam

ER -

Kleibergen FR, Hoek H. Bayesian Analysis of ARMA Models. Amsterdam: Tinbergen Instituut. 2000. (Discussion paper TI; 00-027/4).