[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: re: xtiverg2

From   Kit Baum <>
Subject   st: re: xtiverg2
Date   Sun, 12 Jul 2009 06:49:04 -0400

Menale said
I have learned the application of xtiverg2 when we have unbalanced data ( singletons)from previous posts. After reading previous posts and running xtiverg2 , I would like to ask the following questions. 1) What happen to the constant term when we use xtiverg2. The constant term is not shown in xtiverg2 although the slope coefficients are the same as xtreg, fe. 2) Can I use xtivreg2 even if I do not have endogenous regressor. The reason I want using xtiverg2 is to avoid deleting single observations manually for the reasons mentioned by Mark in his reply to the question, xtiverg2 and singleton-cases, in 2006.

Mark Schaffer's -xtivreg2- (SSC) is the routine in question.
(1) A standard fixed effects estimation (the LSDV model) wipes out the constant term, as does a standard estimation in first differences. For consistency between these two approaches, -xtivreg2- does not report the constant term. -xtreg- and -xtivreg- do so because they represent the coefficients on the panel fixed effects as the deviations from their grand mean. But the within transformation leads to a model without a constant term. As you note, -xtivreg2- will automatically detect and remove singleton cases. That is no big deal, though;
bysort panelid: drop if _N==1
will do that.
(2) -xtivreg2- is a 'wrapper' for Baum-Schaffer-Stillman -ivreg2- (SSC). -ivreg2- can handle estimation of models without endogenous regressors (so it can do OLS, OLS with HAC VCE, etc.) Therefore - xtivreg2- can handle such models, and provides a way to implement the first difference (FD) transformation in an OLS model. Strangely, official -xtreg- does not handle FD, but -xtivreg- does. Perhaps that strange inconsistency will be rectified in Stata 11.

Kit Baum   |   Boston College Economics & DIW Berlin   |
An Introduction to Stata Programming |
   An Introduction to Modern Econometrics Using Stata  |

*   For searches and help try:

© Copyright 1996–2017 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index