Hello,
I've been using the cmp command to estimate a 2-equation system in which one of the variables is endogenous. To be concrete, my estimation looks thusly:
--------------------------
svy: cmp
(D12: MRJMON = UNEMPL UNEMR_NSA MRJPRICE COCPRICE HERPRICE METHPRICE
RACE MAR IRAGE IRAGE2 IREDUC NCHILD CITY RURORURB NONWAGE)
(E12: UNEMPL=H_MEAN UNEMR_NSA MRJPRICE COCPRICE HERPRICE METHPRICE RACE MAR IRAGE IRAGE2 IREDUC NCHILD CITY RURORURB NONWAGE),
ind($cmp_probit $cmp_probit)
--------------------------
So, in my model, UNEMPL (unemployment) is endogenous and is instrumented for by H_MEAN. My question is why is it that the "first-stage" coefficients do not stay the same when I only change the dependent variable in the "second-stage"? For example, if I change the above equation to
--------------------------
svy: cmp
(D13: MRJMONF = UNEMPL UNEMR_NSA MRJPRICE COCPRICE HERPRICE METHPRICE RACE MAR IRAGE IRAGE2 IREDUC NCHILD CITY RURORURB NONWAGE)
(E13: UNEMPL=H_MEAN UNEMR_NSA MRJPRICE COCPRICE HERPRICE METHPRICE RACE MAR IRAGE IRAGE2 IREDUC NCHILD CITY RURORURB NONWAGE),
ind($cmp_probit $cmp_probit)
--------------------------
That is, I only changed MRJMON to MRJMONF in the D13 equation, but I get that the coefficients for the E13 equation, e.g. the coefficients on H_MEAN UNEMR_NSA . . . etc. are not the same as the coefficients in the E12 equation. Why is that?
Thank you for any help you can offer,
Vuong
*
* 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/
