Hi,
I am trying to estimate caste pay gaps in India using the Oaxaca-Blinder
decomposition. This works fine for OLS regression models using the "regress"
command in Stata. I tried the suggestion of Ben Jann in statalist archives but
still run into problems when doing this after controlling for potential
selection bias using the "selmlog" command. What I am trying to do is the
following:
e.g. OLS regression:
regress lnwage `xvarlist' if hhgrp == 1 , robust
est sto sc
regress lnwage `xvarlist' if hhgrp == 2 , robust
est sto gc
oaxaca sc gc
e.g. selection-corrected regression:
selmlog lnwage `xvarlist' if hhgrp == 1, sel(status = `xvarlist' `zvarlist')
dmf(2) bootstrap(100) gen(cprob1)
est sto sc
>>> error message -- last estimation results not found, nothing to store
r(301);
/*selmlog lnwage `xvarlist' if hhgrp == 2, sel(status = `xvarlist' `zvarlist')
dmf(2) bootstrap(100) gen(cprob2)
est sto gc
cap drop _m1 _m2 _m3
gen _m1 = .
gen _m2 = .
gen _m3 = .
oaxaca sc gc
*/
I am using selmlog version 1.3 and Stata version 9.
The error message suggest that no estimates are stored despite the fact that
e(b) and e(V) matrices are stored after running selmlog. Please could someone
help me understand how to store the estimated results post-selmlog so that I can
apply the oaxaca command?
Thanks very much,
Puja
*
* 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/
