pvars: pvars: VAR Modeling for Heterogeneous Panels

In pvars: VAR Modeling for Heterogeneous Panels

pvars: VAR Modeling for Heterogeneous Panels

Description

This package implements (1) panel cointegration rank tests, (2) estimators for panel vector autoregressive (VAR) models, and (3) identification methods for panel structural vector autoregressive (SVAR) models as described in the accompanying vignette. The implemented functions allow to account for cross-sectional dependence and for structural breaks in the deterministic terms of the VAR processes.

Details

(1) The panel functions to determine the cointegration rank are:

• pcoint.JO panel Johansen procedures,

• pcoint.BR panel test with pooled two-step estimation,

• pcoint.SL panel Saikkonen-Luetkepohl procedures,

• pcoint.CAIN correlation-augmented inverse normal test.

(2) The panel functions to estimate the VAR models are:

• pvarx.VAR mean-group of a panel of VAR models,

• pvarx.VEC pooled cointegrating vectors in a panel VECM.

(3) The panel functions to retrieve structural impact matrices are:

• pid.chol identification of panel SVAR models using Cholesky decomposition to impose recursive causality,

• pid.grt identification of panel SVEC models by imposing long- and short-run restrictions,

• pid.iv identification of panel SVAR models by means of proxy variables,

• pid.dc independence-based identification of panel SVAR models using distance covariance (DC) statistic,

• pid.cvm independence-based identification of panel SVAR models using Cramer-von Mises (CVM) distance.

Supporting tools, such as the specification functions (speci.VAR, speci.factors) and the panel block bootstrap procedure (sboot.pmb), complement the panel VAR functions and complete this coherent approach to VAR modeling for heterogeneous panels within the vars ecosystem. The provided data sets further allow for the exact replication of the implemented literature.

Author(s)

Lennart Empting [e-mail] (ORCID: 0009-0004-5068-4639)

See Also

Useful links:

• https://github.com/Lenni89/pvars

• Report bugs at https://github.com/Lenni89/pvars/issues

pvars documentation built on Nov. 5, 2025, 6:57 p.m.