Here you can find a toolbox with Matlab functions which I wrote together with Filippo Ferroni at the Chicago Fed. The toolbox covers, for the moment, VARs, FAVARs, panels of VARs, mixed frequencies VARs, local projections, point and density predictions. There are a number of prior specifications available for estimation and the identification of the shocks covers traditional procedures ( Cholesky, zero, long run, sign) and newer ones (narrative, quantity, IV, variance decomposition and mixed types of restrictions).
Keep watching: this is work in progress and other stuff will be added in the future.
Fixed a few bugs and added :
1) Trend/cycle, permanent/transitory decompositions (univariate and multivariate).
2) Dating turning points and computing recessions/expansions statistics.
3) Regularization and shrinkage methods for large scale models
NEW VERSION!! OCTOBER 2021.
Fixed a few bugs and added :
5) Static and dynamic factor models
NEW VERSION!! MARCH 2022.
Fixed a bug in VARX model. Added :
6) VAR identification via heteroskedasticity .
7) Historical decomposition in VARX models.
NEW VERSION!! MARCH 2024.
8) VAR identification via higher moments (non-gaussian VARs).
9) Robust impulse response bands.
10) Cospetra and power cohesion calculations.
- Priors on impulse responses
- Joint estimtion of large systems with SVARs and LPs
Keep this page bookmarked! New stuff may be coming soon.
The newest version of the manual here
The codes are freely available at https://github.com/naffe15/BVAR_
The presentation slides are here
Video of the tutorials are here
Slides of the tutorial are here.
YOU TUBE channel: @empiricalmacrotoolbox.