<>
Apart from one constraint, you are describing a model fit with
individual fixed effects (xtreg, fe). That constraint is that if you
used SUR, you would estimate a separate sigma^2 for each equation,
whereas FE will estimate a single sigma^2. But if you use robust SEs
that should not matter. There really is no advantage to setting it up
as a SUR and then constraining away all variation in slopes, as that
variation is the rationale for using SUR in the first place.
Kit
On Oct 31, 2009, at 2:33 AM, Valeriano wrote:
I want to run a SUR regression for a system of equations. I have the
same
equation for different years and I want equal coefficients for all
regressors except for the incercept that will be year-specific.
I have a panel dataset, do I have to rename Y and X´s for each year
differently in order to run SUR? What is the correct syntax for the
model
and the "constancy restriction"?
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/
