Bayesian estimation of parameters that influence steady states

Dear all,

I am currently working on a model where some of the parameters that I want to estimate are present in the steady states of certain variables. Luckily, I have closed-form solutions for all the steady states as a function of these parameters. My question is: is it enough to write the steady states of these variables as a function of the parameters in the steady_state_model block? Or should I additionally create a steady_state .m file? In other words, as long as I declare my steady states as a function of estimated parameters in the steady_state_model block, is this enough to make sure that at each iteration the steady state is modified according to the estimated parameters?

Thank you very much in advance!

Yes. The steady_state_model-block will create the necessary steady_state2.m-file.

Dear professor Pfeifer, thank you very much for your response!

Dear professor Pfeifer, I just have one last question. Imagine that I have a parameter in my model \bar{b}, that determines the debt elasticity of the interest rate (as in the RBC models by e.g. Uribe and Yue). We know that in steady state, b_t is equal to \bar{b}. Hence, we can recover the steady state debt level \bar{b} as a linear function of the steady state trade balance, \bar{TB}. The steady-state trade balance, however, depends on the parameters that I am trying to estimate. Hence, one parameter from the model, \bar{b}, depends on the steady-state level of another variable, \bar{TB}, which indeed depends on a parameter that I am trying to estimate. Is it possible to do this by creating a steady-state file? Because it is not very straightforward to define at the very beginning of the code a parameter value with a steady state value that depends on an estimated parameter. Thank you very much in advance!

Yes, that is one of the standard cases where you need such a file.

Dear professor Pfeifer, many thanks for your fast reply. Do you happen to know about the existence of such a file? Likewise, do you think it is feasible to write such a routine?

Many thanks for everything

What exactly are you looking for? DSGE_mod/Jermann_Quadrini_2012_RBC_steadystate.m at master · JohannesPfeifer/DSGE_mod · GitHub is an involved example that calibrates parameters to a target.