Page 1 of 1

Error computing likelihood; This case shouldn't happen.

PostPosted: Tue Jun 23, 2009 1:46 pm
by jtsoukalas
Hi all

Has anyone enounter this error before?

Error in computing likelihood for initial parameter values
??? Error using ==> print_info at 67
This case shouldn't happen. Contact the authors of Dynare

The code works perfectly fine with v3 but not with the latest v4. It seems to me that it does not read the mode file I provide in the estimation command. I don't have any NaN for parameters. Any ideas?

JT

Re: Error computing likelihood; This case shouldn't happen.

PostPosted: Tue Jun 23, 2009 4:22 pm
by bigbigben
Have you checked your parameter values? Are all of them in the permissible region?

Re: Error computing likelihood; This case shouldn't happen.

PostPosted: Tue Jun 23, 2009 8:12 pm
by jtsoukalas
Hi

Thanks. mode file is computed after I have carried out estimation with v3, so the likelihood can indeed be evaluated at these parameter values.

Re: Error computing likelihood; This case shouldn't happen.

PostPosted: Thu Jun 25, 2009 8:33 am
by StephaneAdjemian
Hi, You may have a problem with the steady state of your model. Do you use a matlab steadystate file to compute the steady state of your
model?

Best, Stéphane.

Re: Error computing likelihood; This case shouldn't happen.

PostPosted: Thu Jun 25, 2009 11:26 am
by jtsoukalas
Hi Stephane

Thanks for the reply. I set up a linear model so don't use any matlab function inside the mod file. It is just weird that it works perfectly fine with v3 but not 4.

Best

John

Re: Error computing likelihood; This case shouldn't happen.

PostPosted: Thu Jun 25, 2009 1:01 pm
by StephaneAdjemian
John, Can you post your mod file and the data file.
Best, Stéphane.