st: ivreg2 with weights

[Kit Baum <baum@bc.edu>]

<>
Jeff Pitblado's message earlier today indicated that a coding error in  
Stata's _robust routine was to blame for -ivreg2- results changing  
under Stata 11. StataCorp have now released an update which corrects  
that error. With that update in place, Baum/Schaffer/Stillman's - 
ivreg2- now produces the correct results when used with weights. You  
can verify that with the certification script distributed with - 
ivreg2- (ssc desc ivreg2), which now runs to completion on Stata 11,  
and contains three tests for the equivalence of -ivreg2- and - 
ivregress- results when weights are employed.

Thanks to Mark Schaffer for tracking down the source of the problem  
and documenting it to StataCorp, whose developers very rapidly  
squashed this bug. As they noted, it did not affect any official Stata  
commands, but did adversely affect -ivreg2- and perhaps other user- 
written commands that make use of _robust.

If you are running Stata 11, I recommend that you -update all-; the  
update contains this fix and a number of other ado-file changes that  
have been discussed on Statalist. It does not contain an executable  
update, which we have been told will be forthcoming later this month.

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



*
*   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/