Statalist The Stata Listserver


[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

st: re: fixed vs random effects


From   Marcello Pagano <pagano@hsph.harvard.edu>
To   statalist@hsphsun2.harvard.edu
Subject   st: re: fixed vs random effects
Date   Sun, 04 Feb 2007 10:02:51 -0500

From: Kit Baum <kitbaum@mac.com>
Date: February 3, 2007 7:59:54 AM EST
To: statalist@hsphsun2.harvard.edu
Subject: re: fixed vs random effects

Shams said

Kit Baum: Presence of more than one endogenous
variable complicates estimating the equation using
xtivreg2 (which is considered to be single equation
model).

Not so. xtivreg and xtivreg2 do not assume that there is a single endogenous variable among the regressors, just as standard IV regression allows for any number of endogenous regressors. You merely have to be able to satisfy the equation via the rank and order conditions, meaning that to estimate this single equation you must list the instruments that *would appear* as regressors in the other structural equations but *do not appear* in this one.

It is a fallacy to think that system estimation (reg3) solves any problems. It offers an efficiency gain over single equation (ltd. info) estimation, period (and allows you to impose across-equation constraints). If you cannot write down the single IV equation satisfying conditions for identification, you cannot estimate it with reg3 either. And in the panel setup you describe, you probably need a FE or FD estimator in any case to deal with unobserved heterogeneity.


Kit Baum, Boston College Economics
http://ideas.repec.org/e/pba1.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/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