Page 1 of 1

Possible bug in DSGE-VAR estimation in Dynare 4.1.3

PostPosted: Wed Nov 10, 2010 11:26 am
by George
Dear all,

I am trying to estimate a DSGE-VAR model using Dynare 4.1.3 and using mode_compute=5 as an optimizer. After properly specifiyng the "dsge_prior_weight" in my code and starting the code it get the following error message:

??? Error using ==> feval
Undefined function or method 'DsgeVarLikelihood_hh' for input arguments of type 'double'.

Error in ==> mr_hessian at 57
[f0, ff0]=feval(func,x,varargin{:});

Error in ==> newrat at 67
[dum, gg, htol0, igg, hhg]=mr_hessian(0,x,func_hh,flagit,htol,varargin{:});

Error in ==> dynare_estimation_1 at 443
[xparam1,hh,gg,fval,invhess] =
newrat('DsgeVarLikelihood',xparam1,hh,gg,igg,crit,nit,flag,gend);

Error in ==> dynare_estimation at 62
dynare_estimation_1(var_list,varargin{:});

Error in ==> swfadsgevar at 695
dynare_estimation(var_list_);

Error in ==> dynare at 132
evalin('base',fname) ;

Could it be that there is a potential bug in in the DSGE-VAR code in the context of using mode_compute=5? (the other optimization routines seem to work fine)

Thank you very much in advance for your feedback!

George

Re: Possible bug in DSGE-VAR estimation in Dynare 4.1.3

PostPosted: Mon Nov 15, 2010 7:58 am
by StephaneAdjemian
Dear George,

Sorry this is a bug in the manual. This optimization routine does not work with DSGE-VAR models. You have to choose another optimization routine.

Best, Stéphane.