Page 1 of 1

How do I estimate Newey West Standard Errors for a VAR in Eviews?

Posted: Tue Sep 17, 2019 11:24 am
by titzaaa
I have heteroskedasticity issues and also some autocorrelation issues in my VAR and wanted to include Newey West Standard Errors. However, I do not find a way to conduct this in Eviews.
I am a super beginner and prefer the click options, which means I did not work with coding so far.
Is there a step-by-step way for me to include Newey West errors in a VAR?

Thanks a lot for your support!

Best,
Franziska

Re: How do I estimate Newey West Standard Errors for a VAR in Eviews?

Posted: Tue Sep 17, 2019 11:38 am
by EViews Gareth
There is nothing built in that will do it.

Re: How do I estimate Newey West Standard Errors for a VAR in Eviews?

Posted: Tue Sep 17, 2019 12:59 pm
by startz
Since a VAR is just a set of ols regressions, you can just run each of the regressions and choose Newey-West standard errors.

Re: How do I estimate Newey West Standard Errors for a VAR in Eviews?

Posted: Tue Sep 17, 2019 1:06 pm
by EViews Gareth
startz wrote:Since a VAR is just a set of ols regressions, you can just run each of the regressions and choose Newey-West standard errors.


But you'll miss out on all the juicy VAR goodness, like impulse responses

Re: How do I estimate Newey West Standard Errors for a VAR in Eviews?

Posted: Tue Sep 17, 2019 1:09 pm
by startz
EViews Gareth wrote:
startz wrote:Since a VAR is just a set of ols regressions, you can just run each of the regressions and choose Newey-West standard errors.


But you'll miss out on all the juicy VAR goodness, like impulse responses


First you run using the VAR object. That gets you the juicy stuff. Then run OLS to get the standard errors. That gets you [some clever metaphor here].

Re: How do I estimate Newey West Standard Errors for a VAR in Eviews?

Posted: Wed Sep 18, 2019 4:27 am
by titzaaa
Hi everyone,
thanks a lot for your help!
So step-by-step:
I first estimate my VAR and define my lag length there right?
Then I go into the single regressions.
But where do I perform all my heteroscedasticity, non-normality and autocorrelation tests? Meaning, how do I get my robust standard errors then back into the VAR to do all these tests?
And when do I run my Impulse response and error variance?

Hope not to be confusing :D

Best,
Franziska

Re: How do I estimate Newey West Standard Errors for a VAR in Eviews?

Posted: Wed Sep 18, 2019 4:59 am
by startz
You can run your tests in the ols regression. I don't think there is a way to get them back into the VAR.

But let me point out that we usually put enough lags in a VAR that there isn't any serial correlation left, which is the main reason for using Newey-West standard errors.