Prais Winsten

For econometric discussions not necessarily related to EViews.

Moderators: EViews Moderator, EViews Gareth

Prais Winsten

Postby ThomasHazenberg on Wed Sep 23, 2009 4:59 am

Hi all,

I'm trying to solve a question from the 'Wooldridge - introductory econometrics' book (4th edition).
In question (12,10) we are asked to estimate the static Philips curve model by iterative Prais-Winsten.
My question is: How can I do this is Eviews??

What I tried is:
ls inf c unem (where inf and unem are the variables of interest)
genr uhat=resid
ls uhat c uhat(-1)
this gives me the rhohat, which is 0.572969.

Now I try to do a Prais-Winsten transformation;
genr infnew=(1-0.572969^2)^(1/2)*inf
genr unemnew=(1-0.572969^2)^(1/2)*unem

and run a OLS on the transformed variables.
when regressing the new residuals on residuals-1 the rho is not what it should be.

I attached the workfile. The prais-winsten rho the book finds is 0.781, I can only find 0.572969

Can anyone help me with this?

Thanks in advance
Attachments
phillips imported.wf1
Dataset needed
(13.96 KiB) Downloaded 82 times
ThomasHazenberg
 
Posts: 1
Joined: Wed Sep 23, 2009 4:39 am

Re: Prais Winsten

Postby EViews Glenn on Wed Sep 23, 2009 10:55 am

The Prais-Winsten transform should only be performed on the first observation. The remainder should use the traditional Cochrane-Orcutt transform.
EViews Glenn
EViews Developer
 
Posts: 2004
Joined: Wed Oct 15, 2008 9:17 am


Return to Econometric Discussions

Who is online

Users browsing this forum: No registered users and 5 guests