Page 1 of 1
Windmeijer finite sample correction (EViews 7)
Posted: Tue Mar 15, 2011 1:47 pm
by jadegruen
I am currently tring to estimate a dynamic GMM model based on an unbalanced panel with large N and small T. Literature suggests for that case to apply a two-step procedure and to perform a Windmeijer finite sample correction. Given the structure of my data is this possible in EViews (in the user guide I only found a reference to this correction method in the section non-panel data GMM options)?
Thanks for any help!
Re: Windmeijer finite sample correction (EViews 7)
Posted: Tue Mar 15, 2011 2:41 pm
by EViews Gareth
Currently we don't offer Windmeijer corrected standard errors for panel GMM, sorry!
Re: Windmeijer finite sample correction (EViews 7)
Posted: Wed Mar 16, 2011 2:37 am
by jadegruen
Thanks for your quick reply. Would it make any difference to the bias in standard errors if I alternatively applied the "iterate to converge" procedure?
(Unfortunately, I cannot use the one-step estimation with respect to the result of the Sargan-Hansen test.)
Re: Windmeijer finite sample correction (EViews 7)
Posted: Fri Apr 01, 2011 2:16 am
by rdellanno
I bought few months ago Eviews 7.1 and I see that I can not apply neither Windmeijer finite sample correction for two-step procedure nor system-GMM estimator.
I would know if Eviews technical staff has planned to update GMM estimation for Dynamic Panel in the next months.
To my great regret, I point out as your main competitor (STATA) done this several years ago.Thanks