Page 1 of 1

VAR, Identification problem

Posted: Tue Jul 06, 2010 3:06 am
by shipoopi
Hi,

Currently, I am trying to estimate a small macroeconomic model consisting of three equations: IS equation, Taylor rule and Phillips equation (see picture). I want to do an impulse response analysis.
A first step I did was Quick->Estimate VAR and specified the endogenous and exogenous variables and got the output.

Before I can do the IR analysis, I have to specify how the equations look like, right? But which steps do I have to follow? Do I need Proc->Estimate Structural Factorization ?

Thanks a lot!

Image

Re: VAR, Identification problem

Posted: Tue Jul 06, 2010 7:22 am
by EViews Gareth
Unfortunately you will not be able to put those equations into a VAR in EViews, since EViews only supports symmetric VARs (i.e. same number of lags for each variable). You can estimate your equations in a System object, but then you will not have the impulse responses.

Re: VAR, Identification problem

Posted: Tue Jul 06, 2010 8:11 am
by shipoopi
Thanks for your help. I'll try to program it in Matlab.
cheers!