It appears to me that areg/xtreg, in addition to not adjusting the observations to account for singletons, is estimating the within regressions with a bunch of zeros in X and Y (demeaning groups with only one observation should yield zeros in X and Y). This affects both the slope and intercept of the regression line and can lead to coefficients (not just S.E.) substantially different from the ones obtained by explicitly dropping single-occurring observations from the fixed effects regression.
Can someone please confirm if I am right in my thinking about how Stata is treating single observation groups? Thank you!
-----Original Message-----
From: [email protected] [mailto:[email protected]] On Behalf Of Austin Nichols
Sent: Tuesday, October 07, 2008 6:56 AM
To: [email protected]
Subject: Re: st: fixed effects for unbalanced panel
See
http://www.stata.com/statalist/archive/2006-07/msg00535.html
On Tue, Oct 7, 2008 at 1:29 AM, Deepak Hegde <[email protected]> wrote:
> I have an unbalanced panel (for most firms, I have two or more years of data, and for some firms only one year).
>
> I want to know what stata does to observations (firms) with only one year of data when I run a fixed effects model with either the xtreg, fe(firm_id) or areg, absorb (firm_id) command. It seems like the correct thing to do is to drop observations with only one year of data, but I am getting different results when I explicitly exclude firms with only one-year of data and when I use Stata's fixed effects commands. This suggests that Stata is using data on firms without time-variation in a way that I don't understand...
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
