# Shock decomposition after Occbin calib\_smoother

**URL:** <https://forum.dynare.org/t/shock-decomposition-after-occbin-calib-smoother/27113>\
**Category:** Occbin\
**Created:** [30 January 2025 13:37 UTC](https://forum.dynare.org/t/shock-decomposition-after-occbin-calib-smoother/27113 "2025-01-30T13:37:44Z")\
**Posts on this page:** 5\
**Page:** 1

<div class="post-metadata">

**Author:** ![BDN](https://forum.dynare.org/letter_avatar_proxy/v4/letter/b/c37758/32.png) [@BDN](https://forum.dynare.org/u/BDN)\
**Post date:** [30 January 2025 13:37 UTC](https://forum.dynare.org/t/shock-decomposition-after-occbin-calib-smoother/27113/1 "2025-01-30T13:37:44Z")

</div>

Dear Prof Pfeifer

I am running this implementation of an NK model subject to the ZLB:

> **[tests/occbin/filter/NKM\_calib\_smoother.mod ·...](https://mnemosyne.ithaca.fr/stephane/dynare/-/blob/e1a5a4e1d9e1b2de24e0d6342c1f3a455ee12002/tests/occbin/filter/NKM_calib_smoother.mod)**
>
> GitLab Community Edition

I noticed that if I run shock\_decomposition after either of the calib\_smoother commands, the resulting shock decompositions show large contributions from “initial values” during ZLB periods. These appear necessary to fit the data during such periods.

My question is: what do these “initial values” represent? They do not decay over time starting from the beginning of the simulation, as you would expect / typically find when ignoring OBCs and estimating a purely linear model.

More generally, is shock\_decomposition the correct command to use after the Occbin smoother has been run, or is there an Occbin-specific version of the shock decomposition command?

Thank you

---

<div class="post-metadata">

**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:** [31 January 2025 14:18 UTC](https://forum.dynare.org/t/shock-decomposition-after-occbin-calib-smoother/27113/2 "2025-01-31T14:18:12Z")

</div>

Shock decomposition is not compatible with OccBin yet. The results should be wrong.

---

<div class="post-metadata">

**Author:** ![BDN](https://forum.dynare.org/letter_avatar_proxy/v4/letter/b/c37758/32.png) [@BDN](https://forum.dynare.org/u/BDN)\
**Post date:** [31 January 2025 14:35 UTC](https://forum.dynare.org/t/shock-decomposition-after-occbin-calib-smoother/27113/3 "2025-01-31T14:35:28Z")

</div>

Thank you very much for your reply. In what sense would they be wrong at this stage?

---

<div class="post-metadata">

**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:** [31 January 2025 14:59 UTC](https://forum.dynare.org/t/shock-decomposition-after-occbin-calib-smoother/27113/4 "2025-01-31T14:59:58Z")

</div>

They won’t take the constraint into account.

---

<div class="post-metadata">

**Author:** ![BDN](https://forum.dynare.org/letter_avatar_proxy/v4/letter/b/c37758/32.png) [@BDN](https://forum.dynare.org/u/BDN)\
**Post date:** [1 February 2025 07:32 UTC](https://forum.dynare.org/t/shock-decomposition-after-occbin-calib-smoother/27113/5 "2025-02-01T07:32:14Z")

</div>

Thank you very much