VAR, SVAR and IRF

For econometric discussions not necessarily related to EViews.

Moderators: EViews Gareth, EViews Moderator

Nemster92
Posts: 3
Joined: Thu Jul 28, 2016 3:39 am

VAR, SVAR and IRF

Postby Nemster92 » Fri Aug 05, 2016 2:49 am

Hey guys,

I have estimated this VAR:

X_t = A_0 + B(L)X_t-1 + C(L)d_t + e_t

A_0: column vector of constants, X_t-1 : column vector of endogenous variables, d_t: column vector of exogenous variables
I have used lag length criterion etc. to obtain lag lengths.
I am trying to use Impulse Response Functions to estimate the effect of a shock in the exogenous series to the rest of the system. However I am struggling to see how I incorporate the exogenous series as the shock. Even through User Specified shocks it seems to only consider the endogenous variables.
Should I consider SVAR?

I am using EViews 8.

Cheers lads

dakila
Posts: 489
Joined: Tue Nov 24, 2015 4:57 pm

Re: VAR, SVAR and IRF

Postby dakila » Fri Aug 05, 2016 9:35 pm

No built-in procedure. The trick is to make model from VAR model.(proc/make model). Then simulate it.

Nemster92
Posts: 3
Joined: Thu Jul 28, 2016 3:39 am

Re: VAR, SVAR and IRF

Postby Nemster92 » Mon Aug 08, 2016 2:37 am

Thanks for the response dakila!

Have you done this before?
I make the model with my exogenous series as the exogenous variable and it does not appear in my model as a variable. I make the model with my exogenous series as an endogenous variable and it does appear. I assume I'd have to alter the variable just before I make the model i.e. reduce by 1% etc, to see the changes. By simulate, do you mean solve (and go through the succeeding options)? All I seem to get is some text output but ideally I'd want some graphs? Can you shed any light? Really appreciate this!


Return to “Econometric Discussions”

Who is online

Users browsing this forum: No registered users and 2 guests