Bayesian nonparametric sparse VAR models

Monica Billio, Roberto Casarin, Luca Rossini

Research output: Contribution to JournalArticleAcademicpeer-review

Abstract

High dimensional vector autoregressive (VAR) models require a large number of parameters to be estimated and may suffer of inferential problems. We propose a new Bayesian nonparametric (BNP) Lasso prior (BNP-Lasso) for high-dimensional VAR models that can improve estimation efficiency and prediction accuracy. Our hierarchical prior overcomes overparametrization and overfitting issues by clustering the VAR coefficients into groups and by shrinking the coefficients of each group toward a common location. Clustering and shrinking effects induced by the BNP-Lasso prior are well suited for the extraction of causal networks from time series, since they account for some stylized facts in real-world networks, which are sparsity, communities structures and heterogeneity in the edges intensity. In order to fully capture the richness of the data and to achieve a better understanding of financial and macroeconomic risk, it is therefore crucial that the model used to extract network accounts for these stylized facts.
Original languageEnglish
JournalJournal of Econometrics
DOIs
Publication statusAccepted/In press - 2019

Fingerprint

Vector autoregressive model
Coefficients
Clustering
Stylized facts
Prediction accuracy
Vector autoregressive
Overfitting
Community structure
Macroeconomics

Keywords

  • Bayesian nonparametrics
  • Bayesian model selection
  • Connectedness
  • Large vector autoregression
  • Multilayer networks
  • Network communities
  • Shrinkage

Cite this

@article{4142368996424855b3fb772e9363dca7,
title = "Bayesian nonparametric sparse VAR models",
abstract = "High dimensional vector autoregressive (VAR) models require a large number of parameters to be estimated and may suffer of inferential problems. We propose a new Bayesian nonparametric (BNP) Lasso prior (BNP-Lasso) for high-dimensional VAR models that can improve estimation efficiency and prediction accuracy. Our hierarchical prior overcomes overparametrization and overfitting issues by clustering the VAR coefficients into groups and by shrinking the coefficients of each group toward a common location. Clustering and shrinking effects induced by the BNP-Lasso prior are well suited for the extraction of causal networks from time series, since they account for some stylized facts in real-world networks, which are sparsity, communities structures and heterogeneity in the edges intensity. In order to fully capture the richness of the data and to achieve a better understanding of financial and macroeconomic risk, it is therefore crucial that the model used to extract network accounts for these stylized facts.",
keywords = "Bayesian nonparametrics, Bayesian model selection, Connectedness, Large vector autoregression, Multilayer networks, Network communities, Shrinkage",
author = "Monica Billio and Roberto Casarin and Luca Rossini",
year = "2019",
doi = "10.1016/j.jeconom.2019.04.022",
language = "English",
journal = "Journal of Econometrics",
issn = "0304-4076",
publisher = "Elsevier BV",

}

Bayesian nonparametric sparse VAR models. / Billio, Monica; Casarin, Roberto; Rossini, Luca.

In: Journal of Econometrics, 2019.

Research output: Contribution to JournalArticleAcademicpeer-review

TY - JOUR

T1 - Bayesian nonparametric sparse VAR models

AU - Billio, Monica

AU - Casarin, Roberto

AU - Rossini, Luca

PY - 2019

Y1 - 2019

N2 - High dimensional vector autoregressive (VAR) models require a large number of parameters to be estimated and may suffer of inferential problems. We propose a new Bayesian nonparametric (BNP) Lasso prior (BNP-Lasso) for high-dimensional VAR models that can improve estimation efficiency and prediction accuracy. Our hierarchical prior overcomes overparametrization and overfitting issues by clustering the VAR coefficients into groups and by shrinking the coefficients of each group toward a common location. Clustering and shrinking effects induced by the BNP-Lasso prior are well suited for the extraction of causal networks from time series, since they account for some stylized facts in real-world networks, which are sparsity, communities structures and heterogeneity in the edges intensity. In order to fully capture the richness of the data and to achieve a better understanding of financial and macroeconomic risk, it is therefore crucial that the model used to extract network accounts for these stylized facts.

AB - High dimensional vector autoregressive (VAR) models require a large number of parameters to be estimated and may suffer of inferential problems. We propose a new Bayesian nonparametric (BNP) Lasso prior (BNP-Lasso) for high-dimensional VAR models that can improve estimation efficiency and prediction accuracy. Our hierarchical prior overcomes overparametrization and overfitting issues by clustering the VAR coefficients into groups and by shrinking the coefficients of each group toward a common location. Clustering and shrinking effects induced by the BNP-Lasso prior are well suited for the extraction of causal networks from time series, since they account for some stylized facts in real-world networks, which are sparsity, communities structures and heterogeneity in the edges intensity. In order to fully capture the richness of the data and to achieve a better understanding of financial and macroeconomic risk, it is therefore crucial that the model used to extract network accounts for these stylized facts.

KW - Bayesian nonparametrics

KW - Bayesian model selection

KW - Connectedness

KW - Large vector autoregression

KW - Multilayer networks

KW - Network communities

KW - Shrinkage

U2 - 10.1016/j.jeconom.2019.04.022

DO - 10.1016/j.jeconom.2019.04.022

M3 - Article

JO - Journal of Econometrics

JF - Journal of Econometrics

SN - 0304-4076

ER -