Counterfactual analysis with shock_decomposition

Estimation ML/Bayesian estimation
1

Dear all,

I am studying a scenario by playing with a deep parameter. In terms of economics, this makes sense because each deep parameter is assumed to be independent of any other deep parameter (but in practice fixing one deep parameter would change all the estimated deep parameters).

Back to Dynare. I am successful in changing a deep parameter before the stoch_simul command (by modifying M_ structure). However, shock_decomposition function specifically asks for what parameter set to use (e.g. posterior mean or median), hence I cannot plug in the modified M_ structure in this function.

Is there any neat way to achieve this, or should I rely on changing the code base?

Thanks.
Yakup.

2

shock_decomposition has a parameter_set=calibration option that will use the parameters set in M_.params

3

Prof. Pfeifer,

Thank you very much.

Best.