# Smoothed variables, smoothed shocks, ghx and ghu **URL:** \ **Category:** Stochastic simulations\ **Created:** [13 April 2026 04:49 UTC](https://forum.dynare.org/t/smoothed-variables-smoothed-shocks-ghx-and-ghu/28338 "2026-04-13T04:49:23Z")\ **Posts on this page:** 2\ **Page:** 1
**Author:** ![stanislav](https://forum.dynare.org/letter_avatar_proxy/v4/letter/s/e5b9ba/32.png) [@stanislav](https://forum.dynare.org/u/stanislav)\ **Post date:** [13 April 2026 04:49 UTC](https://forum.dynare.org/t/smoothed-variables-smoothed-shocks-ghx-and-ghu/28338/1 "2026-04-13T04:49:23Z")
Dear all, I have the following question. I have a DSGE model, and I need to work with its VAR-like representation. So I have oo\_.SmoothedVariables, oo\_.SmoothedShocks, which are essentially the trajectories of endogenous variables and shocks according to the model, if I understand correctly. I also have matrices oo\_.dr.ghx and oo\_.dr.ghu, and the vector of steady states oo\_.steady\_state. There is a (stylized) formula for the first-order approximation of the model: y\_t = y^s + A y^h\_{t-1} + B u\_t. Do I understand correctly that this formula must, after the corresponding variable order rearragements (DR etc), hold as identity with y being oo\_.SmoothedVariables, u being oo\_.SmoothedShocks, A and B being oo\_.dr.ghx and oo\_.dr.ghu? I am asking because I have quite significant discrepancies. ---
**Author:** ![jpfeifer](https://forum.dynare.org/user_avatar/forum.dynare.org/jpfeifer/32/5044_2.png) [@jpfeifer](https://forum.dynare.org/u/jpfeifer)\ **Post date:** [13 April 2026 15:29 UTC](https://forum.dynare.org/t/smoothed-variables-smoothed-shocks-ghx-and-ghu/28338/2 "2026-04-13T15:29:57Z")
Yes, that should be the case unless you have stochastic singularity. Have a look at [tests/kalman\_filter\_smoother/compare\_results\_simulation/fs2000.mod · master · Dynare / dynare · GitLab](https://git.dynare.org/Dynare/dynare/-/blob/master/tests/kalman_filter_smoother/compare_results_simulation/fs2000.mod?ref_type=heads)