Shock is much more volatile than other macro variables

Hi, Dynare team

Now I am trying to modify a typical DSGE model (Iacoviello 2015 RED)

and I want to add some shocks i.e.) house construct shock.

When I use quaradic detrended house supply data and match the data with the shock, it is highly volatile.(-0.2 to 0.2)

However other shocks’ prior and estimates are quite modest and above all, other macro variables’ data (also quadratic detrended and I used other data set) is less volatile relative to my new shocks.

As a result shock decompostion results are quite strange.

If I change the other shocks prior as volatile as house shock, the prior-posterior graphs are fine, but shock decompostion graphs are messy.

Is there any advice to deal with too big shocks? I think that I should not scale the house supply shock like multiply 0.1 to make it small…

Welcome to the Forum!
Could you provide more information on how you “match the data with the shock”? Also posting your mod file is always helpful as people can have a direct look and find other issues as well.
With that said, having big shocks in a stochastic model solved by perturbation is problematic…