Error Message: "POSTERIOR KERNEL OPTIMIZATION PROBLEM!"

[mbov]

Hi there,

I am having an issue running baynesian estimation of a DSGE model. Dynare is displaying the following messages and I am unsure how to solve this issue. I have tried changing the mode_compute and priors, but doesn't seem to offer a solution.

Any help on this issue is greatly appreciated. Thank you

POSTERIOR KERNEL OPTIMIZATION PROBLEM!
(minus) the hessian matrix at the "mode" is not positive definite!
=> posterior variance of the estimated parameters are not positive.
You should try to change the initial values of the parameters using
the estimated_params_init block, or use another optimization routine.
Warning: The results below are most likely wrong!
> In dynare_estimation_1 at 694
In dynare_estimation at 89
In TRIAL_JP3_MATT at 351
In dynare at 180

Error using chol
Matrix must be positive definite.
Error in metropolis_hastings_initialization (line 68)
d = chol(vv);
Error in random_walk_metropolis_hastings (line 62)
[ ix2, ilogpo2, ModelName, MetropolisFolder, fblck, fline, npar, nblck, nruns, NewFile, MAX_nruns, d ] = ...
Error in dynare_estimation_1 (line 782)
feval(options_.posterior_sampling_method,objective_function,options_.proposal_distribution,xparam1,invhess,bounds,dataset_,options_,M_,estim_params_,bayestopt_,oo_);
Error in dynare_estimation (line 89)
dynare_estimation_1(var_list,dname);
Error in TRIAL_JP3_MATT (line 351)
dynare_estimation(var_list_);
Error in dynare (line 180)
evalin('base',fname) ;

[jpfeifer]

One obvious problem is that your observation equations are wrong. You cannot estimate a model with mean 0 variables with data that is not mean 0. Please see Pfeifer(2013): "A Guide to Specifying Observation Equations for the Estimation of DSGE Models" https://sites.google.com/site/pfeiferecon/Pfeifer_2013_Observation_Equations.pdf.

[mbov]

Thank you for you help and quick reply. I have now demeaned the series, however still obtain the same error messages. Would you know what could still be causing this error?

Thank you

[jpfeifer]

As always, please provide the updated files.

[mbov]

Sorry- please find attached the latest files. Thanks for your help and reply.

[jpfeifer]

Please have a look at the

Code: Select all

mode_check

plots. There seems to be a big stability/uniqueness issue as indicated by the red dots.

[mbov]

Thank you very much for your reply. Seeing these red dots, would you suggest that the model is not correctly specified? Model calibrations work though?
I appreciate your time and response.

[jpfeifer]

You need to understand where these dots come from, i.e. which economic mechanism gives rise to the model moving to the boundary of the uniqueness/stability region. It might be that there is a mistake in specifying the model, but there could also be an economic reason. Maybe you should start with a simpler model and then add features to it and see what drives the problem

[mbov]

Thank you very much. I have really appreciated all your time and help.