st: re: how to do parameter restrictions in fe, re be models

[Kit Baum <baum@bc.edu>]

<>
Rup said

suppose y is a linear function of a & b. i want to estimate y such that the coefficients of 
a and b adds to one in panel data methods. can anyone suggest some commands.

The same way you would do it in OLS models if you did not have a -constraint- command:

y = beta0 + beta1 a + beta2 b     s.t. beta1+beta2 = 1

(y - beta2) = beta0 + beta1 (a-b)

So regress the variable (y-beta2) on the variable (a-b) using fixed effects, or random effects, or whatever you would like.

If you want to know beta2 in point and interval form, use lincom 1 - beta1.

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/