Mixed-Frequency Bayesian VAR Models

Functions and tools for estimation of mixed-frequency Bayesian vector autoregressive (VAR) models. The package implements a state space-based VAR model that handles mixed frequencies of the data as proposed by Schorfheide and Song (2015) , and extensions thereof developed by Ankargren, Unosson and Yang (2020) , Ankargren and Joneus (2019) , and Ankargren and Joneus (2020) . The models are estimated using Markov Chain Monte Carlo to numerically approximate the posterior distribution. Prior distributions that can be used include normal-inverse Wishart and normal-diffuse priors as well as steady-state priors. Stochastic volatility can be handled by common or factor stochastic volatility models.

mfbvar

Overview

The mfbvar package implements Bayesian mixed-frequency VAR models.

Installation

The current development version of the package can be installed with the help of devtools:

devtools::install_github("ankargren/mfbvar")

Usage

See the vignette for details and examples.