Ramsey Policy Error

[lah89]

Getting the following error

Code: Select all

Configuring Dynare ...
[mex] Generalized QZ.
[mex] Sylvester equation solution.
[mex] Kronecker products.
[mex] Sparse kronecker products.
[mex] Local state space iteration (second order).
[mex] Bytecode evaluation.
[mex] k-order perturbation solver.
[mex] k-order solution simulation.
[mex] Quasi Monte-Carlo sequence (Sobol).
[mex] Markov Switching SBVAR.

Starting Dynare (version 4.4.0).
Starting preprocessing of the model file ...
Ramsey Problem: added 14 Multipliers.
Found 14 equation(s).
Found 29 FOC equation(s) for Ramsey Problem.
Evaluating expressions...done
Computing static model derivatives:
 - order 1
Computing dynamic model derivatives:
 - order 1
 - order 2
Computing static model derivatives:
 - order 1

??? Error using ==> dynare at 156
DYNARE: preprocessing failed

any ideas whats going wrong ?
I dont only get that error but also Dynare crashes, stating "dynare_m.exe does not work anymore"

[costa]

Try to look at row 156, or share the code.

[jpfeifer]

Also, use Dynare 4.4.2

[lah89]

I dont only get that error but also Dynare crashes, stating "dynare_m.exe does not work anymore"

This problem comes from you using model local variables (those defined with a #) in planner_objective. Future Dynare versions will yield an error. Try to get of those and it should run.

Now that you mention it...
Thank you very much !

[lah89]

I dont wanna open a new topic for this questions althought it is not related to the error above (however it is related to Ramsey optimal policy):


Is there a way to constraint the instrument used for the optimal policy exercise to be positive?
I am trying to replicate a paper's IRFs (which works quite well for the basecase) but so far not for the optimal policy part.

One source for the differences might be that Dynare suggests a path for the instrument that involves negative values (althought this does not make sense from an economic point of view). How can I prevent that ? (or cant I since I am in a stochastic model and I cant use e.g. a max{}-operator due to local approximations ?)

Thanks!

[jpfeifer]

You could put the instrument in exp() in the model. Up to first order, it will be constrained to be positive that way.

[lah89]

You could put the instrument in exp() in the model. Up to first order, it will be constrained to be positive that way.

thanks! problem is that the instrument has steady state 0 in the base case by construction (and only becomes positive under an optimal policy)
hmm

[lah89]

any other ideas?


(btw:
-since I used exp() in my model in order to loglinearize, I enter the variables as well with exp() into the objective function, no?
-including an endogenous variable that is following an AR(1) which innovation is getting shocked into the planner's objective should not be an issue, should it?)

[jpfeifer]

- If the planner cares about absolute deviations instead of log-deviations/percentage deviations, then the planner objective must contain exp()
- That should not be an issue.

I must say that I find it strange to have a planner using an instrument that can always go in only one direction. What is the economic intuition?