Re: Re: st: Hausman test after xtnbreg and alternatives?

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

<>
On Feb 12, 2013, at 2:33 AM, David wrote:

> 
> Date: Mon, 11 Feb 2013 18:11:07 -0500
> From: David Greenberg <dg4@nyu.edu>
> Subject: Re: st: Hausman test after xtnbreg and alternatives?
> 
> Your model is not overidentified, so an overidentification test is
> irrelevant. When you say that you do not get "a right statistic" by
> reversing the order of the models in the command, what do you mean by
> that? If you want help you have to show us what you get. David
> Greenberg
> 
> On Mon, Feb 11, 2013 at 5:18 PM, Heewon Chae <heewon.chae@gmail.com> wrote:
>> Thanks for the reply but I don't think the two can be good
>> alternatives. 1) xtoverid doesn't work for xtnbreg. 2) doing it with a
>> reverse order doesn't give a right statistic.
>> (http://www.stata.com/statalist/archive/2004-08/msg00251.html)
>> 
>> Is there anyone who can help this out?
> 

Re "an overidentification test is irrelevant":

In the context of linear or IV regression, an overidentification test is indeed appropriate for testing the 'hausman' null that RE is suitable, as the difference between FE and RE is that RE imposes k additional orthogonality conditions on the model (that each column of X is distributed independently of the random effect). That is the rationale for -xtoverid- (Schaffer and Stillman, SSC).

This point is discussed at some length in Baum, Schaffer, Stillman, Stata Journal 3:1, 1-31, section 5.3, freely downloadable from stata-journal.com, IDEAS, EconPapers, etc.

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
                                                                                                   | http://www.crup.com.cn/Item/111779.aspx	


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