st: Specification tests and EC2SLS

[Damon Cann <dcann@uga.edu>]

I  have a set of panel data with endogenous variables.  Executing an 
error components two-stage least squares model poses no problem, but I 
want to show that I've met the assumptions of the method.  I'm not 
certain that the standard specification tests for 2SLS are valid after 
ec2SLS.

Particularly, I want to know if I can proceed with the following two 
tests as though my EC2SLS results were 2SLS results.

*Hausman test for exogeneity (should I compare the EC2SLS to a random 
effects model instead of to OLS?)
*Test of Overidentifying restrictions (a la Davidson and MacKinnon)


Many thanks for any comments,

Damon Cann
dcann@uga.edu

*
*   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/