Statalist


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

st: Arellano Bond with correlated panels


From   "Massimiliano Tani Bertuol" <mtani@efs.mq.edu.au>
To   <statalist@hsphsun2.harvard.edu>
Subject   st: Arellano Bond with correlated panels
Date   Fri, 13 Jun 2008 17:57:52 +1000

I am working with a panel of European regions where N=177 and T=10, and estimate a dynamic panel model explaining employment growth in region i as a function of that region's previous employment growth and other variables. The model to be estimated is:

EG(t+1) = aEG(t) + bEG(t)*X(t) + cEG(t)*Y(t) + dEG(t)*X(t)*Y(t) + Z(t) + error terms

where:
EG = employment growth in region i
X, Y, Z = other variables related to region i
a, b, c, d = parameters

Given the dynamic panel, I use Arellano-Bond estimator (AB).  
However, from my understanding of AB, the error term is assumed to be uncorrelated across regions.  
This cannot be ruled out in my case: it is likely that shocks in a region affect neighboring regions in the same  
country, though perhaps not regions located in other member states. Is there a way to control for this issue within Stata (9.0)? Is anyone using an alternative approach for this type of problem? Your comments are much appreciated. Thanks in advance
max



*
*   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