Dynare forums

Hello everybody,

I attach below the .mod file relative to a model I am currently trying to estimate.
The model is provided to Dynare in its Non-linear version and, therefore, the solution of the steady state is required for the estimation (and simulation).

I set out the .mod file so as an external matlab function is called in the ''Calibration block'' (e.g right before ''Model block''): this function takes as
input the parameters of the model and returns steady state solution for each variable (I attach it below).
Finally, using the 'Initval' block, I set the SS values for each variables found above.

As far as model simulation is involved this codes works fine.
Now, my question is: does this arrangement work for the estimation as well or do I necessarily need to rely on a 'Steady_State_Model' block?

Thanks for your time,

Mattia

If you are able to compute the exact steady state, I would always write a proper steady state file for estimation. It assures better speed and that no draws are discarded just because Dynare did not find the steady state. But your file is different. You must use a steady state file. The reason is that your file features parameter dependence. Some parameter of the model change depending on the value of other parameters in order to satisfy calibration targets. Only a steady state file assures that this parameter dependence is satisfied.