# Optimal parameters leads IRFs were explosive

**URL:** <https://forum.dynare.org/t/optimal-parameters-leads-irfs-were-explosive/13102>\
**Category:** Optimal Policy\
**Created:** [17 January 2019 02:21 UTC](https://forum.dynare.org/t/optimal-parameters-leads-irfs-were-explosive/13102 "2019-01-17T02:21:20Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![jiang](https://forum.dynare.org/letter_avatar_proxy/v4/letter/j/3e96dc/32.png) [@jiang](https://forum.dynare.org/u/jiang)\
**Post date:** [17 January 2019 02:21 UTC](https://forum.dynare.org/t/optimal-parameters-leads-irfs-were-explosive/13102/1 "2019-01-17T02:21:20Z")

</div>

Dear dynare, I used the global optimizer cmaes to find optimal policy parameters, and the parameters’ LowBounds is -500, UpBounds is 500, and I got optimal parameters and run dynare file, the IRFs were explosive, I saw the paper use the restrictions :

(max(-oo\_.irfs.R\_eps\_s+R)\>1 | max(oo\_.irfs.R\_eps\_s+R)\>1 | min(-oo\_.irfs.R\_eps\_s+R)\<0 | min(oo\_.irfs.R\_eps\_s+R)\<0)

and use this restriction to get optimal parameters then would not lead explosive, I’m confused,  
Please give me some help. Thanks so much.

---

<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:** [18 January 2019 05:53 UTC](https://forum.dynare.org/t/optimal-parameters-leads-irfs-were-explosive/13102/2 "2019-01-18T05:53:54Z")

</div>

I would need more details here. Are you using a second order approximation? If yes, did you use `pruning` when generating the IRFs? If yes, then the IRFs cannot explode if you properly controlled for violation of Blanchard-Kahn conditions during your parameter search.