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

Re: st: A Question on Panel Data


From   "V. Sarafidis" <vs242@cam.ac.uk>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: A Question on Panel Data
Date   01 Mar 2005 22:44:13 +0000

Hi Ricardo,

The xtgls command is strictly speaking not asymptotic in panels - rather, it requires a large time dimension. This is essentially because you are estimating variance parameters for each panel and the estimates require many time-periods per panel for consistency.

You may attempt to use the command
regress, cluster()
This command gest you OLS parameters (as the xtpcse command) but the standard errors are consistent for N large rather than T large.

Vasilis
On Mar 1 2005, Ricardo Gonçalves Silva wrote:


Dear Stata users

My question concerns on panel data estimation. I have cross section of 124 countries observed over 19 to 33 years. Additionally, I suspect my dependent variable is heterosckedastic (income growth rate) and correlated with one independent variable (the income level) and a time-trend. Finally, I want to get specific country effects. In this way, my model is something like that:

gr_y = x1 time i.idd

The question is: what estimation procedure should I use. In my view, the basic ones are xtgls or xtpcse. I know that xtgls are asymptotic in panels, and xtpcse are asymptotic in years. So, given my setup, the direct choose would be xtgls. But I also know that the Standard Deviations of FGLS are too optimistic.

So, the question remains? What one?

Any help will be very appreciated.

Cheers

Ricardo




*
*   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