st: Re: Re: GMM or 2SLS with cluster adjustment or others

[Kit Baum <baum@bc.edu>]

Again I would reference Roodman's paper "how to do xtabond2".  The  
Arellano-Bond approach, applied mechanically, can create hundreds of  
instruments. On p.10 of his paper, he discusses how estimates of the  
variances can be very poor if too many instruments are used. If used  
wisely, the A-B estimator or its friends in the DPD stable can be  
relatively efficient, but if used improperly, a huge 'weak  
instruments' problem can occur. The standard IV-GMM estimator (e.g.  
xtivreg2, gmm2s robust bw()) will not be as susceptible to these  
problems (and to the extent there are weak instruments issues, they  
can be diagnosed with the built-in tests in -xtivreg2-).


Kit Baum, Boston College Economics and DIW Berlin
http://ideas.repec.org/e/pba1.html
An Introduction to Modern Econometrics Using Stata:
http://www.stata-press.com/books/imeus.html


On May 15, 2008, at 02:33 , Yuluen wrote:


> Is there a citation that suggests Arellano-Bond is not the best  
> option (e.g. not efficient, etc.) when running a model without a  
> LDV? Thanks.
*
*   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/