Page 1 of 1

Simulated data from model

PostPosted: Thu Jan 15, 2015 3:56 am
by abman10
Hi,
I am trying to simulate data (endogenous variables) from my DSGE model i implemented by using Dynare. It's clear that Dynare is able to provide impulse response of one variable following a given shock.
But, my question is the following: Is Dynare able to hit the model by several shocks simultanousely ( not separately ) , and then get impulse responses of endogenous variables?

Thanks a lot for help!!

Manadir

Re: Simulated data from model

PostPosted: Thu Jan 15, 2015 8:26 am
by jpfeifer
For simulations, use the periods-option of stoch_simul.

Re: Simulated data from model

PostPosted: Tue Jan 20, 2015 3:20 pm
by abman10
jpfeifer wrote:For simulations, use the periods-option of stoch_simul.


Thanks...I will try this option.

Re: Simulated data from model

PostPosted: Wed Feb 11, 2015 1:48 am
by abman10
Hi again,
I have tried your last advice to get simulated data from model, but Dynare gives me the following warning message: " Matrix is close to singular or badly scaled. Results may be inaccurate..."
What should i do in this case?
Thank you for help!!

Re: Simulated data from model

PostPosted: Wed Feb 11, 2015 12:06 pm
by jpfeifer
Please provide the mod-file

Re: Simulated data from model

PostPosted: Sun Feb 15, 2015 9:57 pm
by abman10
Here are two files attached. MSF.mod is simulation file.

I remember you that i try to hit my economy model by several shocks simultanously and at every time for a fixed period, then get artificial data ( for endogenous variables).

Thanks a lot for your help.

Re: Simulated data from model

PostPosted: Mon Feb 16, 2015 1:15 pm
by jpfeifer
Dynare 4.4.3 is unable to solve your model. Please try the unstable version. With that version, your file runs.

Re: Simulated data from model

PostPosted: Tue Feb 17, 2015 4:10 pm
by macroresearch123
abman10 wrote:Here are two files attached. MSF.mod is simulation file.

I remember you that i try to hit my economy model by several shocks simultanously and at every time for a fixed period, then get artificial data ( for endogenous variables).

Thanks a lot for your help.

I've found this often happens due to mis-specification. Recheck your equations and make sure you can solve your steady state with another solver. Sometimes big shocks, or too many shocks, cause problems in nonlinear systems.

Re: Simulated data from model

PostPosted: Tue Feb 17, 2015 4:51 pm
by abman10
Your suggestions are really appreciated.