re: st: ivreg2 cannot produce first-stage result

[Kit Baum <baum@bc.edu>]

<>
Phil said


I used the following ivreg2 to estimate an IV-regression, but could
not produce the first-stage result.

.  ivreg2 my_y $xfirm $year_dummies choice_dummy (= choice_sic3),   first

Unable to display first-stage estimates; macro e(first) is missing

Unable to display summary of first-stage estimates; macro e(first) is missing


I browsed the Statalist, found a similiar posting, but the suggested
answer is that if OLS is used in the second stage, there is no
first-stage results.

But I actually used the same command syntax in another setting some
months ago, it worked.

I would appreciated any suggestions on how to obtain the first stage result.


The current versions of ivreg2, ivreg29 and ivreg28 from SSC all give these same two messages. The FSRs that 
you're considering are the regressions of what lies on the left of the equals sign above on all exogenous variables.
As there is nothing there, there are no such regressions to be estimated. You are using the 'heteroskedastic OLS' (HOLS)
estimator, in which the excluded instruments on the right of the equals sign are being used to improve the efficiency 
of the estimates in the presence of heteroskedasticity of unknown form. However, their presence does not change
the fact that there are no FSRs in this kind of model.

Kit
coauthor ivreg2*

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/