# Bad gradient error by changing shock

**URL:** <https://forum.dynare.org/t/bad-gradient-error-by-changing-shock/18163>\
**Category:** Uncategorized\
**Created:** [25 May 2021 19:50 UTC](https://forum.dynare.org/t/bad-gradient-error-by-changing-shock/18163 "2021-05-25T19:50:25Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![Madihaatif](https://forum.dynare.org/letter_avatar_proxy/v4/letter/m/c0e974/32.png) [@Madihaatif](https://forum.dynare.org/u/Madihaatif)\
**Post date:** [25 May 2021 19:50 UTC](https://forum.dynare.org/t/bad-gradient-error-by-changing-shock/18163/1 "2021-05-25T19:50:25Z")

</div>

Dear all,  
I am working on IRF-matching technique. whenever, I change shock in my model like I add price shock instead of technology shock. I got error as "bad gradient’ and my estimation returns no results.

> %set impulse to 1 percent of GDP  
> shock\_vector=zeros(n\_augmented\_vars,1);  
> shock\_pos=y\_pos;  
> shock\_vector(shock\_pos)=1/Q(shock\_pos,shock\_pos); %set to 1 percent of G

if I add

> %set impulse to 1 percent of GDP  
> shock\_vector=zeros(n\_augmented\_vars,1);  
> shock\_pos=pie\_pos;  
> shock\_vector(shock\_pos)=1/Q(shock\_pos,shock\_pos); %set to 1 percent of G

I got error as ‘bad gradient’

can anyone help me where I am wrong  
[finaldata.xlsx](https://forum.dynare.org/uploads/short-url/vFRQsjk5F8PC5WitFVFQldATZW.xlsx) (105.7 KB)  
[IRFNEW.mod](https://forum.dynare.org/uploads/short-url/dt85bVP84OygOEox0qsk87ixQYR.mod) (6.7 KB)  
[build\_companionform.m](https://forum.dynare.org/uploads/short-url/8QkYKLvLTZ3ZUyOiwXoQl3sjPuW.m) (752 Bytes)  
[get\_empirical\_IRFs.m](https://forum.dynare.org/uploads/short-url/hH8ZFwtvHcC5nXMzgNaikwkL7On.m) (6.3 KB)  
[IRF\_matching\_objective.m](https://forum.dynare.org/uploads/short-url/qLcWCcJp0ucz6ezVnsx7IJIMqH1.m) (1.4 KB)