An R package for Bayesian Estimation of Structural Vector Autoregressive Models
This package provides efficient algorithms for Bayesian estimation of Structural Vector Autoregressive (SVAR) models via Markov chain Monte Carlo methods. A wide range of SVAR models is considered, including homo- and heteroskedastic specifications and those with non-normal structural shocks. All models include three-level equation-specific local-global hierarchical prior distribution for the estimated level of shrinkage for autoregressive and structural parameters. Additionally, the package facilitates predictive and structural analyses with a variety of tools and methods.
The heteroskedastic SVAR model setup is similar as in Woźniak & Droumaguet (2015) and Lütkepohl & Woźniak (2020). The sampler of the structural matrix follows Waggoner & Zha (2003), whereas that for autoregressive parameters follows Chan, Koop, Yu (2022). The specification of Markov switching heteroskedasticity is inspired by Song & Woźniak (2021), and that of Stochastic Volatility model by Kastner & Frühwirth-Schnatter (2014).
A
and error terms E
, and the structural equation with structural shocks U
Y = AX + E (VAR equation)
BE = U (structural equation)
A
and the structural matrix B
are feature a three-level local-global hierarchical prior that estimates the equation-specific level of shrinkagespecify_bsvar_*()
functions, for instance, specify_bsvar()
estimate()
methodpredict()
methodcompute_impulse_responses()
, compute_variance_decompositions()
, compute_historical_decompositions()
, and compute_structural_shocks()
respectivelycompute_fitted_values()
, compute_conditional_sd()
, and compute_regime_probabilities()
respectively**************************************************|
bsvars: Bayesian Structural Vector Autoregressions|
**************************************************|
Gibbs sampler for the SVAR-SV model |
Non-centred SV model is estimated |
**************************************************|
Progress of the MCMC simulation for 1000 draws
Every 10th draw is saved via MCMC thinning
Press Esc to interrupt the computations
**************************************************|
0% 10 20 30 40 50 60 70 80 90 100%
[----|----|----|----|----|----|----|----|----|----|
*************************************
The beginnings are as easy as ABC:
library(bsvars) # upload the package
data(us_fiscal_lsuw) # upload data
spec = specify_bsvar_sv$new(us_fiscal_lsuw, p = 4) # specify the model
burn_in = estimate(spec, 1000) # run the burn-in
out = estimate(burn_in, 50000) # estimate the model
Starting from bsvars version 2.0.0 a simplified workflow using the |>
pipe is possible:
library(bsvars) # upload the package
data(us_fiscal_lsuw) # upload data
us_fiscal_lsuw |>
specify_bsvar_sv$new(p = 4) |> # specify the model
estimate(S = 1000) |> # run the burn-in
estimate(S = 50000) -> out # estimate the model
Now, you’re ready to analyse your model!
Just open your R and type:
install.packages("bsvars")
The developer’s version of the package with the newest features can be installed by typing:
devtools::install_git("https://github.com/bsvars/bsvars.git")
The package is under intensive development. Your help is welcome! Please, have a look at the roadmap, discuss package features and applications, or report a bug. Thank you!
Tomasz is a Bayesian econometrician and a Senior Lecturer at the University of Melbourne. He develops methodology for empirical macroeconomic analyses and programs in R and cpp using Rcpp.