Structural Scenario Analysis for Bayesian Structural Vector Autoregression Models

Implements the scenario analysis proposed by Antolin-Diaz, Petrella and Rubio-Ramirez (2021) "Structural scenario analysis with SVARs" .

APRScenario (version 0.0.4.0)

Version 0.0.4.0 fixes a critical bug in the impulse-response recursion of big_b_and_M()/mat_forc() (results at forecast horizons of 3+ periods were incorrect in earlier versions and should be recomputed; see NEWS.md) and adds cross-platform parallelization over posterior draws (n_cores argument of scenarios()/big_b_and_M(); works on Windows via PSOCK clusters and on Unix/macOS via forking).

What it does

This package is a stab implementing Structural scenario analysis with SVARs by Juan Antolín-Díaz, Ivan Petrella and Juan F. Rubio-Ramírez, JME (2021) in R

It depends on the output of the package bsvarSIGNs although it could be adapted to the output of many other packages by changing the function get_mats.R.

See the application code APRScenario.Rmd.

Installation

devtools::install_github("giannilmbd/APRScenario", ref = "master")

See also the forked bsvarSIGNs with parallelized draws of the rotation matrix

devtools::install_github("giannilmbd/bsvarSIGNs", ref = "master")

You can also download the tar.gz package from the tar-package branch and install in in R with

install.packages("<path>APRScenario_XXXX.tar.gz", repos = NULL, type = "source");

For XXXX see the latest version Use APRScenario.Rmd as a template for the application.

Feedbacks are welcome.