Re: st: xtreg fixed effects

[Christopher Baum <kit.baum@bc.edu>]

<>
On May 24, 2012, at 8:33 AM, Nick wrote:

> 1. Each panel contributes at least one data point to the -xtreg-.
> That's information!
> 
> 2. Each singleton panel contains no information about within-panel
> variability, or rather implies that within-panel variability is zero
> in some cases. But that is only part of the data's properties.

I beg to differ.

webuse grunfeld
drop if year>1935 & company>1
xtreg invest mvalue kstock,fe
reg invest mvalue kstock if company==1

illustrates that the slope coefficients, in both point and interval form, from the one complete panel are the same as those computed from that panel and nine singleton panels. The R^2s are identical, as are the F-statistics, with the same d.f.  sigma_e from the xtreg is equal to RMS Error from the OLS regression. So I fail to see any value in considering these singleton panels. They do give you a different constant term, but one could argue that a FE regression doesn't really have a constant term anyway.

Kit


Kit Baum   |   Boston College Economics & DIW Berlin   |   http://ideas.repec.org/e/pba1.html
                             An Introduction to Stata Programming  |   http://www.stata-press.com/books/isp.html
  An Introduction to Modern Econometrics Using Stata  |   http://www.stata-press.com/books/imeus.html


*
*   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/