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

st: re: time dummies

From   Kit Baum <[email protected]>
To   [email protected]
Subject   st: re: time dummies
Date   Tue, 22 Jan 2008 06:57:47 -0500

Grant said

I am estimating my panel data model using xtabond2 with time dummies as follows;
yr70_74, yr75_79, yr80_84, yr85_89, yr90_94, yr95_99, yr00_04. Thus, my T=7 ( where t=1...7). However, before including the time dummies the model starts estimating at period 3. Does it mean that I can only include my time dummies starting with yr80_84 or yr75_79?

It is standard with Arellano-Bond difference estimator to lose the first two obs. of your level series. It is a model in first differences, so you lose one obs. defining the dependent variable, and as the model is dynamic you include the lagged difference as a regressor. Defining the lagged difference of a level requires two observations.

Through sleight of hand, the GMM-System estimator of Blundell-Bond (which is in fact the default in xtabond2) only loses one observation, as it can be thought of in levels with a lagged level as a regressor. Thus with GMM-System you should be able to include time periods 2..7.


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

* For searches and help try:

© Copyright 1996–2024 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index