Matlab code for inference on impulse responses using local projection or VAR, with or without lag augmentation
Reference: Montiel Olea, José Luis, and Mikkel Plagborg-Møller (2021), "Local Projection Inference is Simpler and More Robust Than You Think", Econometrica 89(4), 1789-1823 (also available as arXiv:2007.13888)
Tested in: Matlab R2020a on Windows 10 PC (64-bit) and Matlab R2019a on Macbook Pro 2019
functions: Matlab routines
- ir_estim.m: main function for impulse response inference
examples: Empirical examples
- gk.m: example based on Gertler & Karadi (2015)
simulations: Simulation studies
- run_sim_ar1.m: AR(1) study
- run_sim_var.m: bivariate VAR study
- run_sim_var_calib.m: empirically calibrated VAR study
- create_figs.m: create figures of AR(1) and bivariate VAR results
- create_table.m: create table of AR(1) and bivariate VAR results
- create_fig_var_calib.m: create figure of empirically calibrated VAR results
Plagborg-Møller acknowledges that this material is based upon work supported by the National Science Foundation under Grant #1851665.