R package for Bayesian Vector Autoregression
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Failed to load latest commit information.
R
man
.Rbuildignore Bayesian VAR for Minnesota Prior and independent Normal-Wishart Prior Mar 3, 2017
.gitignore
DESCRIPTION
NAMESPACE
bvar.Rproj
readme.md

readme.md

bvar

Overview

bvar is a collection of R routines for estimating Linear and Nonlinear Bayesian Vector Autoregressive models in R. The R code is based on the Matlab Code by Blake and Mumtaz (2012) and Koop and Koribilis (2009)

Models and functionalities include:

  • Linear VAR-models:
    • Minnesota and independent Normal-Wishart Prior
  • Nonlinear VAR-models:
    • Threshold VAR with generalized impulse response-functions
  • Other Models:
    • Factor-Augmented VARs
    • Regime Switching Models
  • Prior for VARs
    • Independent Normal-Wishart
    • Minnesota Prior
    • Natural Conjugate prior
    • Uninformative prior
  • Identification of structural shocks
    • Cholesky decomposition
    • Sign restrictions

Installation

To install the package you need the devtools package. If you don't have the devtools package, you can install it with

install.packages("devtools")

Once you have installed the devtools package you can install the bvar package with

library('devtools')
devtools::install_github('joergrieger/bvar')

To-do list

  • improve numerical stability of Threshold-models
  • speed up generalized impulse-response functions
  • Documentation
  • add functions for plotting impulse-respone functions, summary of inference, diagnostics, forecasting
  • regime switching models with time-varying transition probabilities
  • dummy observation prior
  • ssvs - prior

Known issues and bugs

  • generalized impulse functions work only for lags>1