Stata The Stata listserver
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

Re: st: xtreg


From   Ulrich Kohler <ukohler@sowi.uni-mannheim.de>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: xtreg
Date   Tue, 3 Dec 2002 09:09:59 +0000

Lee Chuntao  wrote
>      Can i run the following panel regression with fixed effect ?
>
>      y(i,t)=constant+ w(i)' * alpha + x(i,t)'*beta+u(i,t)
>
>
>      where w(i) is a vector of individaul specific but time invariant
> variables
>      x(i,t) is individual & time variant
>      u(i,t) is the residual term
>
>     the command i used is : xtreg y w x, fe
>     stata dropped the w(i) automatically in the regression, why?


You can't. The fixed-effects model allows different intercepts for each 
individual, that is 

   y_it = constant_i + x_(it)b + u_it


All variables which are constant within persons (all time invariate 
variables) are collinear with the individual intercepts. There is no way to 
use them in a fixed effects model.

Another way to look at this is by looking at an alternative formulation of 
the fixed effects model:

         _               _                   _
  y_it - y_i  =  (x_it - x_i)* b_1 + (z_it - z_i) * b_2 + u_it

                                                                  _ 
If z is constant within time for each individual the term (z_it - z_i) 
becomes zero for all observarions. 

regards
uli



-- 
ukohler@sowi.uni-mannheim.de
http://www.sowi.uni-mannheim.de/lesas

*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index