Hi,

I have run a standard new Keynesian model:

xhat(+1)=xhat+sigma*ihat-sigma*pihat(+1)-sigma*rnhat; (is curve)

pihat=k*xhat+beta*pihat(+1); (phillips curve)

rnhat=ro_rn*rnhat(-1)+s_rn; (process for the natural rate of interest rate)

where monetary policy is derived through the minimisation of the loss function with additional constraint on the nominal interest variability, i.e.

L(t)=pihat(t)^2+lambda_x*xhat(t)^2+lambda_i*(ihat(t)-im)^2

So you get additional FOC in order to derive monetary policy:

pihat=beta^(-1)*sigma*gama1(-1)-gama2+gama2(-1);

lambda_x*xhat=-gama1+beta^(-1)**gama1(-1)+k*gama2;

lambda_iihat=lambda_i*im-sigma*gama1;

The model works fine however for some values of parameters of lambda_i and im I obtain the following error:

??? Error using ==> print_info

MJDGGES returns the following error code7

Error in ==> stoch_simul at 50

print_info(info);

Error in ==> woodford_searchLM at 97

stoch_simul(var_list_);

Error in ==> dynare at 26

evalin(‘base’,fname) ;

I just wonder what this error message means.

I would be grateful for any comment.

Regards,

Anna

Dear Anna,

this very rare error message means that something wrong happend during the Real Generalized Schur Decomposition used to solve the linear rational expectation problem.

Usually this means that the matrix of coefficients obtained after linearization contains complex numbers.

Could you post or send me your example (*.mod file), because I think that Dynare should catch the problem earlier and give a more explicit error message.

Kind regards

Michel

Hi,

I send you the code that I am working with.

The problem that I have described appears when I work with loops trying to find first for a given lambda_i a specific value of im so that one constraint holds (loop while) and then find such lambda_i that the loss function is minimised (loop for).

In the code there are values of lambda_i and im now for which the error appears.

Regards,

Anna

woodford_searchLM.mod (1.03 KB)

Dear Anna,

I just got hold of your example. I will try to answer you in the next 10 days.

Best

Michel

In your example, the loops are commented out and there is only one value for the parameters. Was it the values that were causing problems at your end? These values don’t trigger a problem on my machine.

Best

Michel