Measurement errors, stoch_simul and estimation

[Aldo]

My final goal is estimation, and I have next questions:

(1) Inclusion of measurement errors. The variances of these are not part of the parameters to be estimated, so I am including them within the block of shocks. Is this the correct way to include measurement errors?

(2) The role of stoch_simul before estimation. The .mod runs fine when I precede the estimation command with stoch_simul. Without stoch_simul appear errors:

Error using print_info (line 100)
Likelihood is a complex number

(3) Where can I find output of use Monte-Carlo based optimization routine (mode_compute=6).

I attach my files.

Thanks.

Aldo

[jpfeifer]

1. Yes, that is the correct way of calibrating measurement error
2. In Dynare 4.5 I cannot replicate this issue. This seems to have been due to a bug in Dynare 4.4.3 related to the qz_criterium as your model has a unit root. With the

Code: Select all

diffuse_filter

option in Dynare 4.5, it is gone.
3. Which output are you looking for?

[Aldo]

Thanks a lot dear jpfeifer. Your work is great.

About my last question, I am looking for result showed in waitbar, Acceptance rate.

[jpfeifer]

That information is not saved as it is pretty meaningless. The acceptance rate does not provide any indication on whether the mode-finding was successful