Page 1 of 1

"[D]eclar[ing] endogenous variables after the estimation..."

PostPosted: Sat Jul 06, 2013 11:34 pm
by Dr. En Ami
What is the syntax for doing so? For example, if I want to only forecast y in RBC_est.mod, what would I add to the model file? I tried adding "var y;" to the end of the model to no avail. Thanks, --En

Re: "[D]eclar[ing] endogenous variables after the estimation

PostPosted: Sun Jul 07, 2013 8:11 am
by jpfeifer
Write the variable names after the estimation command:
Code: Select all
estimation(...) y;

Re: "[D]eclar[ing] endogenous variables after the estimation

PostPosted: Mon Oct 06, 2014 11:44 pm
by AbrahamVela
Dear Johannes: I do not understand why the estimate command is required me to list the endogenous veriables I want to see the results. My question is whether I shall list the same veriables in the varobs command (such a y_obs) or as they are specified in the model (just "y"). May be it does not matter because I have an expression in the model stating y_obs=y and in the datafile as well. Of course y_obs is in my datafile. The bottom line is when this is needed and what names I shall use after estimate. My intuition is that htese must be the same as in varobs command.
Thanks in advance for your support.

Re: "[D]eclar[ing] endogenous variables after the estimation

PostPosted: Tue Oct 07, 2014 8:33 am
by jpfeifer
With the unstable version, you have the option
Code: Select all
consider_all_endogenous

and
Code: Select all
consider_only_observed

options to do this task. The reason Dynare asks is that posterior moment computation can be really time intensive and doing it for only selected variables saves time.

Regarding which you should choose: it depends what you are interested in. For stationary models, people are usually interested in the endogenous variables. However, for nonstationary models, people are sometimes interested in the observed growth rates entered as observables.