st: FW: xtabond2 chi squares

["Bromiley, Philip" <bromiley@uci.edu>]

I'm using xtabond2 with about 40,000 observations and 3500 panels.  What
I think are normal lags in the gmm give significant Hansen statistics.
 
In covariance structure analysis (Lisrel etc.), with large samples you
can almost always reject the hypothesis that the model fits the data.
The large sample results in the test rejecting the hypothesis of
equality between the estimated covariance matrix from the maximum
likelihood and empirical matrix even if the two differ trivially (and
they will always differ somewhat in practice).  This has resulted in
researchers in the area using a variety of fit indices instead of
testing whether they can reject the hypothesis that the estimated and
empirical matrices are equal.
 
Is this problem possible with the Hansen and Sargan tests in xtabond2?
As I understand Hansen (1982, lemma 4.1,  pp. 1049), the test tests
whether a particular matrix "converges in distribution to a normal
random vector with mean zero and covariance matrix..."   That is, would
extremely large sample sizes cause the Hansen and Sargan tests to reject
the null hypothesis of validity of instruments even if the differences
are trivial?  
 
Phil
 
Philip Bromiley
Dean's Professor of Strategic Management
Merage School of Business
University of California, Irvine
Irvine, CA 92697-3125
(949) 824-6657
(949) 725-2898 (fax)

*
*   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/