erroe message"MJDGGES returns the following error code"

[iona]

I am trying to estimate the model attached below but I receive the following error:

??? Error using ==> print_info at 36
MJDGGES returns the following error code: 13

Error in ==> check at 50
print_info(info, options_.noprint);

Error in ==> iona at 163
check;

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

dat.xls

the data

(23.5 KiB) Downloaded 200 times

I would be very grateful if someone could give me a clue. I do not even know if the code is correct. Thanks.

dat.xls

the data

(23.5 KiB) Downloaded 200 times

[jpfeifer]

Always move steady before check. As Dynare then says

Warning: Some of the parameters have no value (sigma, omega, lb, lf, k, csi, theta, sigma_alpha, rho_c, fi, rho_yf, rho_a, rho_g,
rho_tau, g_y, g_b, tau_y, tau_b, rho_r, r_pi, r_y) when using steady. If these parameters are not initialized in a steadystate
file, Dynare may not be able to solve the model...

When using check, initializing the parameters in the estimated_params block is not sufficient. Either remove check or define those parameters.

[iona]

thank you very much for your answer.
If I still receive "Error in computing likelihood for initial parameter values"and "numerical initial values incompatible with the following equations...." this means that the initial values are wrong? or it is possible that the code model to be entirely wrong?

[jpfeifer]

"Error in computing likelihood for initial parameter values" means that your starting values for estimation specified in the estim_params-block are wrong. If you did not specify starting values, Dynare takes the prior mean. This could be the source of the problem.

"numerical initial values incompatible with the following equations...." means that the steady state could not be computed, usually because the initvalb-block sets something to 0 and a division by 0 occurs. If no initval-block is present, everything is initialized to 0. Given that your model should be linear, this error message suggests a mistake in your model.

[iona]

I obtain the first message if I dont specify the command steady; in my code and the second error if I write steady; after the block model.
Thank you.

[jpfeifer]

As Dynare says

Some of the parameters have no value (omega, sigma_alpha) when using steady.

Your definition

Code: Select all

sigma_alpha=(sigma/((1-alpha)+alpha*omega));
omega=(sigma*gamma+(1-alpha)*(sigma*eta-1));


is circular. You define sigma_alpha using omega and omega using sigma_alpha. That cannot work.

[iona]

Thank you very much.
Your answers helped me to improve my code. I dont receive anymore those errors, but they were replaced by new ones.
"Blanchard Kahn conditions are not satisfied: indeterminacy". The rank conditions isn't verified.
As I read on forum, this is a problem with the calibration parameters rather than with the model itself, isn't it?

[jpfeifer]

No, most of the time, people actually set correct parameters, but have a timing error in their model. It could be parameters, but this is rather rare for the typical user.

[iona]

Hello again,
I corrected the timing error (indeed it was a timing error) but now I have another problem. I receive the following message:
"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..."

I don't know what to do, because the same code (attached here) ran perfectly yesterday but today it gives me this error.
Could you help me? Thank you.

[jpfeifer]

Try a different mode-finder. Use for example mode_compute=6 or 9.