I’m trying to do the Bayesian estimate following the one done in a paper. I managed to get some results, but they are not the same as those obtained in the reference search. First of all, some error messages appear but, reading other posts, I don’t understand whether I can leave them out or not, like “WARNING: in the ‘steady_state_model’ block, variable ‘i’ is not assigned a value”, “Some of the parameters have no value (Gss, Yss) when using steady. If these parameters are not initialized in a steadystate file or a
steady_state_model-block, Dynare may not be able to solve the model. Note that simul, perfect_foresight_setup, and perfect_foresight_solver do not automatically
call the steady state file”, “Warning: Matrix is close to singular or badly scaled. Results may be inaccurate”.
Furthermore, only the observed parameters appear in the identification analysis, how can I also obtain all the other parameters?
Finally, I would like to specify that the data used are slightly different from the original ones and above all I did not proceed with detrending, so the difference in the estimate could also be due to this. Furthermore, the value of some parameters is also different, otherwise the estimate would not start. Ultimately, in an equation referring to the observed variables it introduces a new variable that I don’t understand how to insert and what it refers to, I_hat. invobs,t = I_ha,t + sk,t - x*se,t
I attach the codes and data and thank you in advance even for just a few suggestions.
Those are too many questions at once. The most important things for now:
- Your estimation will fail due to the incorrect/missing handling of the trend in the data.
- You need to deal with parameter dependence.
- You can get identification results for all parameters if you call
identificationbefore declaring estimated parameters.