re:st:Bundell and Blond: Sargan Test

[Christopher Baum <kit.baum@bc.edu>]

<>
I have a panel data model of international trade, and im regressing the
quantity of imports, for a set of products, using as regressors numbers of
importers, number of exportes, trend, etc.

I´ve estimated the model using POLS, Fixed Effects and GMM (Arellano Bond
and Bundell Blond). The problem that I found is that ar(1) and ar(2) test
perform well, but I reject always the Sargan Test. 

What did you recommend? Forget GMM and use Fixed Effects? 

The Sargan test is not robustified, as it assumes iid errors. You should consult the Hansen J test statistic available, for instance, from Roodman's xtabond2 (on SSC).

BTW the authors of SYS-GMM are Richard Blundell and Steve Bond.

Kit



Kit Baum   |   Boston College Economics & DIW Berlin   |   http://ideas.repec.org/e/pba1.html
                              An Introduction to Stata Programming  |   http://www.stata-press.com/books/isp.html
   An Introduction to Modern Econometrics Using Stata  |   http://www.stata-press.com/books/imeus.html


*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/