Toolkit for the estimation of hierarchical Bayesian vector autoregressions. Implements hierarchical prior selection for conjugate priors in the fashion of Giannone, Lenza & Primiceri (2015). Allows for the computation of impulse responses and forecasts and provides functionality for assessing results.
There are some significant speed improvements possible. I implemented some already, cutting down runtime to about half. The big things that remain is a bit of solveing, rmvnorm and avoiding the companion matrix.
There's quite some potential in the multivariate normal draws - (a) we wouldn't need to check for symmetry, (b) the Cholesky is quite a bit faster, (c) in draw_post() we actually already have the eigendecomposition of sigma.
There are some significant speed improvements possible. I implemented some already, cutting down runtime to about half. The big things that remain is a bit of
solve
ing,rmvnorm
and avoiding the companion matrix.There's quite some potential in the multivariate normal draws - (a) we wouldn't need to check for symmetry, (b) the Cholesky is quite a bit faster, (c) in
draw_post()
we actually already have the eigendecomposition of sigma.