BayesianMS-VAR-GC

Bayesian Estimation of Markov-Switching VARs for Granger Causal Inference in R

by Matthieu Droumaguet, Anders Warne, & Tomasz Woźniak

A block Metropolis-Hastings algorithm for the Bayesian estimation of the Markov-switching Vector Autoregressive models with restrictions for Granger noncausality is provided, as well as an appropriate estimator for the marginal data density.

Keywords: R, Markov-switching VARs, Block Metropolis-Hastings Sampler, Marginal Data Density

To refer to the codes in publications, please, cite the following paper:

Droumaguet, M., Warne, A., Woźniak, T. (2017) Granger Causality and Regime Inference in Markov-Switching VAR Models with Bayesian Methods, Journal of Applied Econometrics, 32(4), pp. 802--818, DOI: 10.1002/jae.2531.

The project's file structure includes:

• BayesianMSVAR.pdf - a document presenting the model, main functions, and their application
• BayesianMSVAR-example.R - a file presenting code application for a simple example
• BayesianMSVAR - a folder containing the functions for the estimation of the considered models
• ReproductionScripts - a folder containing scripts for the reproduction of all the results contained in the JAE paper
• data.csv and data.RData - data used in the paper