Page 1 of 1

implement Generalized IRF from system object (VAR estimated with GMM)

Posted: Mon Sep 11, 2017 8:50 pm
by yunqingMa
Hi, there,

Two questions actually.

1. For EViews 9 enterprise version, is there a way to simulate generalized impulse response function from system object? The reason I want to do this is because I hope to use GMM to estimate my VAR, the only way I can do is to estimate from system object. But after estimation, I couldn't find option for simulating Generalized impulse response function and all other options (e.g. cumulative IRF, monte carlo simulation. etc.)

2. Is there a way in EViews to implement a ONE UNIT shock of generalized impulse response function? EViews is giving one-standard deviation shock for generalized impulse response function, according to my knowledge.

Thank you very much :D !

Re: implement Generalized IRF from system object (VAR estimated with GMM)

Posted: Tue Sep 12, 2017 6:05 am
by dakila
2. Is there a way in EViews to implement a ONE UNIT shock of generalized impulse response function? EViews is giving one-standard deviation shock for generalized impulse response function, according to my knowledge.


Try the sirf add-in.

Re: implement Generalized IRF from system object (VAR estimated with GMM)

Posted: Tue Sep 12, 2017 3:34 pm
by yunqingMa
Try the sirf add-in.[/quote]



Thanks. Actually I tried before and I had two problems.
1. how to make the shock as the generalized IRF in this add-in?
2. Seems this add-in doesn't provide Monte Carlo standard error, which I need.

Could you please provide further guidance?

Re: implement Generalized IRF from system object (VAR estimated with GMM)

Posted: Tue Sep 12, 2017 3:34 pm
by yunqingMa
dakila wrote:
2. Is there a way in EViews to implement a ONE UNIT shock of generalized impulse response function? EViews is giving one-standard deviation shock for generalized impulse response function, according to my knowledge.


Try the sirf add-in.



Thanks. Actually I tried before and I had two problems.
1. how to make the shock the generalized IRF in this add-in?
2. Seems this add-in doesn't provide Monte Carlo standard error, which I need.

Could you please provide further guidance?

Re: implement Generalized IRF from system object (VAR estimated with GMM)

Posted: Tue Sep 12, 2017 5:14 pm
by dakila
First, estimate GIRF using eviews built-in command or dialog box. Then use the sirf add-in.

Re: implement Generalized IRF from system object (VAR estimated with GMM)

Posted: Wed Sep 13, 2017 4:12 pm
by yunqingMa
dakila wrote:First, estimate GIRF using eviews built-in command or dialog box. Then use the sirf add-in.



Thank you so much! But........after I tried yesterday....

1. The graph seems not correct.. what should I put in the field of scale factor for IRF? No understanding about this..
2. When I impulse two variables in the add-in, why it shows me "0 is not a valid index for vector-series-coefficient FAC_DIAG02"...Is this because I put a "1" in the scale factor field?