Estimating parameters that change steady state and other parameters

[mhesam_ahmdn]

Hi,

I am estimating a model in which I have also included parameters to be estimated that slightly change the steady state and also other parameters that are based on them directly. For example, I am adding habit formation to the set of parameters to estimate. But, habit formation changes my steady states and with the steadystates changing then other parameters which are not being estimated and are set directly at the start of the mode file should also be changed. Not only does habit formation change my ideal parameters indirectly by changing steady states, it also does it directly by being in their formulation when calibrating with hand based on FOCs. Now imagine 3 or 4 of such parameters that their estimation indirectly (via change in steady state) and directly (by being in the other parameters ss formulation) requires change in other parameters, are included in the list of parameters to be estimated. Can you please help me on how I should update other parameters in the mode file while dynare is constantly changing the said parameters to find the best one? I tried to include the affected parameters as endogenous variables in the model with steady_state()s in it and a degree of freedom with one variable as endogenous, and it works for 2 of my parameters that are affected, but when I add the equations of the other affected parameters it just gets too complicated and dynare cant even find the steady state at that point. So that method is not viable. Is their any efficient way to make dynare change the affected parameters by giving it the required constraints?
So, in summary, I want habit formation parameter to be estimated. But it being changed also changes other parameters that are calibrated based on its initial value. Even if habit formation parameter wasnt present in their ss formula, it would still affect them by changing the steady states. And there are also 3 other parameters having a similair situation to habit formation in the parameter estimation section. I dont know what to do with these 4 estimating parameters requiring changes in other parameters based on their direct change and also them causing changes in the steady states.

Thanks.

[jpfeifer]

You can either do that in a steady_state_model-block if you have closed-form expressions, see

Setting parameter value to match steady state target

That is the right way to go. See e.g. https://github.com/JohannesPfeifer/DSGE_mod/blob/master/RBC_baseline/RBC_baseline.mod The warning that you get should be one that says If these parameters are not initialized in a steadystate file or a steady_state_model-block But obviously you are setting it there.

The alternative is a user-written steady state file that can also solve numerical equations to achieve calibration targets. See

Problem with finding steady state

I am not sure I can follow. Why don’t you write a steady state file that sets the parameter values to satisfy a calibration target? An example would be https://github.com/JohannesPfeifer/DSGE_mod/blob/master/Jermann_Quadrini_2012/Jermann_Quadrini_2012_RBC/Jermann_Quadrini_2012_RBC_steadystate.m