Page 1 of 1

Question on Generalized impulse response function

Posted: Sun Mar 31, 2019 8:57 pm
by econonick
Dear Sir

I am estimationg generalized impulse response function(Pesaran and Shin, 1998).

But I'm quite confused because the estimated results are differ from EViews.
I am attaching the code and data.
I would really apprecite your help.

Regards

Re: Question on Generalized impulse response function

Posted: Sun Mar 31, 2019 9:31 pm
by TomDoan
Note that the RATS output is blocked by shocks and the EViews output is blocked by target variable. Once you allow for that, there appears to just be a difference in scaling of the covariance matrix. You'll have to check the EViews documentation to see how they are doing that.

Re: Question on Generalized impulse response function

Posted: Sun Mar 31, 2019 10:25 pm
by econonick
TomDoan wrote:Note that the RATS output is blocked by shocks and the EViews output is blocked by target variable. Once you allow for that, there appears to just be a difference in scaling of the covariance matrix. You'll have to check the EViews documentation to see how they are doing that.

Dear Tom.

Thank you for your reply.
I find that the different results are "d.f. adjustment" issue.

Sincerely.