On Sun, Feb 17, 2013 at 1:52 PM, Swanquist, Quinn Thomas <qswanqui@utk.edu> wrote:
> It seems the variable limit for IC that won't allow me to create the necessary binary variables for i.FEvar.
Yes, IC has a variable limit.
> Is there a more efficient way to estimate this similar to -xtreg-? Also, is there a way to remove the coefficients on i.FEvar similar to -areg-?
The relevant projection in L2 involves sweeping out the group means and then modeling the individual level residuals. Because this is an orthogonal projector it has certain desirable properties that make FE work the way its advertised. Switching to group medians would be the L1 analog but what are its properties? I'm not sure.
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/faqs/resources/statalist-faq/
* http://www.ats.ucla.edu/stat/stata/
