EViews.com

Hi,

First of all english is not my main language so sorry for my writing mistakes. I am writing my thesis (which consists in the comparison between two SVARs with LR restrictions) and I need to obtain the IRs from the structural decomposition but for a given shock. The problem is that EVIEWS predefines the shock as a standard deviation and since the samples are different (with different standard deviations) a simple contrast between IR graphs is not useful.

Reading through the forum I found an add-in called 'sirf' that I think may be useful for my purposes, but I have two complications. First, I do not know if this will allow me to contrast two IRs (could I assign the same shock for the two different SVAR's?). Second, I do not know how to use the add-in (do I use it after applying the structural factorization of the VAR or after getting the impulse response under structural decomposition?)

I forget to mention that the SVAR is calculated imposin long-run restrictions.

Thanks in advance for your help, im kind of worried about this,

PRG

1. Yes, you can use the sirf add-in.
2. After getting the impulse response under structural decomposition

I really appreciate your help! Thank you very much..
¿Is there any way that i can boostrap the results after using sirf add-in?

no

Dear all,

I am using the SIRF add-in in SVAR model.
the add-in seems not working within the SVAR. I estimate the VAR, I estimate the structural factorization by imposing restirctions and finally i apply the IRF by setting "stractural decomposition" as impulse definition. After that, I try to implement the SIRF add-in, but eviews stops working without give me any rusults. However, the SIRF Add-in works only without estimating the structural identification, simply in the VAR when a IRF is estimated.
Could you help me? where can I find some explanation about the implemetation of SIRF add-in?

best,
M.

What is the version you use?

Hi, everyone,

Could you please clarify, sirf provide response to 1% shock in % or original units of response variable.

Thanks.