Dear all,
the preprocessor currently allows setting ```constant``` and ```noconstant```-options (although they are not fully documented yet).
But in ```dynare_estimation_init.m``` any such setting is automatically overwritten by:
if all(abs(oo_.steady_state(bayestopt_.mfys))<1e-9) options_.noconstant = 1; else options_.noconstant = 0; end
I don't really see the point of keeping the option. Rather the user should specify a correct observation equation. In that case, the current treatment should assure correct results.
Best,
Johannes
Hi Johannes,
I do not think that we still need these options in the |estimation| command. Dynare adds a constant if and only if the steady state is nonzero, we may safely remove these options. I found a related issue in the same matlab routine. In the first part of the file (line 98 in master), we define |noconstant| depending on the value of option |prefilter|:
|if options_.prefilter == 1 options_.noconstant = 1; end|
and |noconstant| is rewritten at the end... We should check that the user does not prefilter the data if the model has a nonzero steady state. We just need to set option |noconstant| before line 98 and issue an error message if options |noconstant| and |prefilter| are nonzero.
Best, Stéphane.
On 29/01/2014 17:15, Johannes Pfeifer wrote:
Dear all,
the preprocessor currently allows setting ```constant``` and ```noconstant```-options (although they are not fully documented yet).
But in ```dynare_estimation_init.m``` any such setting is automatically overwritten by:
if all(abs(oo_.steady_state(bayestopt_.mfys))<1e-9) options_.noconstant = 1; else options_.noconstant = 0; end
I don't really see the point of keeping the option. Rather the user should specify a correct observation equation. In that case, the current treatment should assure correct results.
Best,
Johannes _______________________________________________ Dev mailing list Dev@dynare.org https://www.dynare.org/cgi-bin/mailman/listinfo/dev