SVAR-MSH-ID

R Codes for Bayesian Inference for Structural Vector Autoregressions Identified with Markov-Switching Heteroskedasticity

by Tomasz Woźniak

The project provides source codes and reproduction files for the paper mentioned below. Utility functions for the Gibbs sampler for Bayesian Structural Vector Autoregressions with Markov-Switching Heteroskedasticity, Savage-Dickey density ratio for uniqueness conditions and homoskedasticity hypothesis, and a marginal data density estimator are provided.

Keywords: SVAR-MSH, identification through heteroskedasticity, Savage-Dickey density ratio for uniquenss conditions, Gibbs sampler

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

Lütkepohl, H., Woźniak, T. (2020) Bayesian Inference for Structural Vector Autoregressions Identified with Markov-Switching Heteroskedasticity, Journal of Economic Dynamics and Control, 113, DOI: 10.1016/j.jedc.2020.103862.

The project includes:

• folder codes with the whole source code
• R files for the reproduction of most of the results from the paper
• file dataBI2015.RData with data used in the paper