Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down on April 23, and its replacement, statalist.org is already up and running.


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: performance of Hansen's J test of overid with number of instruments = N


From   Caterina Astarita <lasta01@students.bbk.ac.uk>
To   statalist@hsphsun2.harvard.edu
Subject   st: performance of Hansen's J test of overid with number of instruments = N
Date   Sun, 19 Sep 2010 11:35:22 +0100 (BST)

Dear all, 

I'm running a GMM system one step estimator with xtabond2 with 
a dataset where N = 19 and T = 15. I'm using the rule of thumb 
of taking the number of instruments less or equal to N. 
Furthermore I'm using the command collapse. With respect to 
Sargan and Hansen's J test of overidentification is 
respectively p = 0.502 and p = 1.00. The fact that I'm using a 
very low number of instruments is sufficient to rely on Hansen 
test? If I undestrand well the answer should be "no". 
According to Roodman "A note on the theme too many 
instruments"  in the  paragraph dedicated to "weak hansen test 
of instrument validity" is written that "a high p.value as, 
say 0.25 should be viewed with concerns". Thus I deduce that 
having a p = 1.00 is a bad signal and that a better scenario 
would be a p-value between 0.10 and 0.25 (or lower) . Is that 
correct? I know that is difficoult to establish a treshold but 
I would like to be certain I have understood Roodman's 
suggestion

Thank you 

Caterina 

*
*   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/


© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   Site index