st: Arellano - Bond GMM

[Ömer İskenderoğlu <omeri@cu.edu.tr>]

Hi everybody
 
I hope someone could help me some.
 
I have a Balanced Panel data of 127 companies for 10 years. The expectation
model is a dynamic model which can be written as follows : 
Yi(t) = a + bYi(t-1) + e(t)  Where Y(t) denotes to log(labors) of 
company i on a time t. 
 
I use stata Arellano Bond estimation and autoregression for 1 and 2 periods.
While the results seems satisfactory up to now but there seems to have the
Sargan test either. Neverthless Sargan test rejects the 
Ho:overidentifying restrictions.
 
So the result of Sargan test implements me the results are biased and not
valid. I find out on some of the studies time dummies are used. I employ
time dummies for each year and i receive valid results. While Baltagi (2003)
page 120 shows that time dummies are used in similar approaches.

The question is : am i an on a straight path on this estimation and do you
have any criticism

 
Thank you 

  
Omer Iskenderoglu 

University of Cukurova 

F.E.A.S.




*
*   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/