BEARX Toolbox

The BEARX Toolbox is a Matlab toolbox for Bayesian estimation, analysis, and reporting of vector autoregressive (VAR) models. BEARX is an extended version of the original BEAR Toolbox (version 5), adding a new command-line interface (CLI) and a new graphical user interface (GUI) while keeping the original BEAR available as before.

Features

• A comprehensive range of Bayesian VAR estimators: plain (OLS, Minnesota, Normal-Wishart, …), time-varying, panel, factor-augmented (FAVAR), threshold, and mixed-frequency
• Structural identification via Cholesky decomposition, zero restrictions, sign restrictions, and generalized restrictions
• Forecasting (unconditional and conditional), impulse response functions, historical shock decomposition, and forecast error variance decomposition
• A transparent GUI that auto-generates editable Matlab scripts — nothing runs behind the scenes

Quick start

Step 1. Clone this repository and, optionally, the companion BEARX GUI Examples repository from within Matlab:

!git clone https://github.com/OGResearch/BEARX-Toolbox
!git clone https://github.com/OGResearch/BEARX-GUI-Examples

Step 2. Add the toolbox to the Matlab path:

addpath BEARX-Toolbox -end; bearPaths

Step 3. Create or open a model folder and launch the GUI:

gui.start    % fresh session (resets configuration)
gui.resume   % resume a previous session

See the wiki page Setting things up for detailed instructions.

Repository layout

bearPaths.m        – adds the toolbox subfolders to the Matlab path
tbx/
  bear/            – original BEAR Toolbox (version 5) code
  bearing/         – BEARX extended estimation and model engine
  gui/             – GUI source (HTML interface, form definitions, script generator)

Documentation

Full documentation is available in the BEARX Toolbox Wiki:

• Setting things up — cloning the repositories, adding paths, opening the GUI
• GUI step-by-step guide — walkthrough of every GUI tab, from estimator selection to script generation
• Reduced-form estimators — reference material on the available estimation methods

License

See the BEAR End User Licence Agreement included in the repository (tbx/bear/BEAR End User Licence Agreement.pdf).