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

st: re: xtabond 2

From   Kit Baum <>
Subject   st: re: xtabond 2
Date   Fri, 16 Nov 2007 09:43:18 -0500

Bill asked

1. Are 800 observations large enought to run xtabond2?
2. How can I deal with the warnings: relative large instruments than
the observations?

First of all, abdata (the classic dataset used by Arellano-Bond: webuse abdata) has 1031 observations, and it is certainly large enough. What matters more perhaps is the relative size of T and N. The technique is suitable for large N, relatively small T (such as single digits). It will work with larger T, but N should be much larger than T.

To keep from generating 'too many' instruments, use the lags( ) option when specifying gmmstyle() instruments. And read through Roodman's "How to do xtabond2" (you can find it on RePEc, http://

Kit Baum, Boston College Economics and DIW Berlin
An Introduction to Modern Econometrics Using Stata:

* For searches and help try:

© Copyright 1996–2015 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index