Stata The Stata listserver
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

Re: st: Panel data, missing values and estimation in differences.

From   David Jacobs <>
Subject   Re: st: Panel data, missing values and estimation in differences.
Date   Thu, 06 May 2004 10:30:33 -0400

Check out the recent Wooldridge book on pooled and cross sectional modeling. My recollection is that first differencing and using dummies will give identical results only when you have just two cross sections. Results diverge with more cross sections than two due to differences in the way these two procedures respond to serial correlation.

Wooldridge, Jeffrey. 2002. Econometric Analysis of Cross Section and Panel Data. MIT Press.

Dave Jacobs

At 11:54 AM 5/6/2004 +0200, you wrote:

Hello to all,
I have two questions regarding panel data estimation.
I am running two regressions with the within estimator, once inserting the dummy variables, the other with the differenced series. Since I get different results for certain coefficients, I was trying to understand if this can depend, beside some theoretical problems (endogeneity, errors of measurment etc), from the changed size of the sample. I have 18 countries in my sample, and the differences in the number of observations between levels and differences is 18. Intuitively Stata should eliminate the first year. But if I run the regression in levels excluding the first year (if != option)the number of observations does not change!So,:
- Is there a way to understand what observations does Stata run the regression on in the difference equation and run the levels one on the same ones?
Apart from this, I have a probably elementary question that is troubling me.
-If a variable is constant in levels for a certain country(and thus captured by the dummy) , hence assuming values 0 in differences, does this country participate in the estimation of the coefficient of that variable in the difference equation?

Thanks a lot for any help you may be able to give me, and sorry for the lenght of the message.
Have a good day

* For searches and help try:

*   For searches and help try:

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