Dear Statalisters,
I am trying to perform the -selmlog- command with the -dhl- option
(Dahl 2002). Like Dahl, I want to estimate a wage equation taking into
account the selectivity biais in the return to education due to free
migration within a country (Roy model).
According to the command of selmlog
selmlog depvar varlist [if exp] [in range],select (depvar_m=varlist_m)
[lee dmf dmf2 dhl(# [all]) showmlogit wls
bootstrap(number_of_replications [sample_size]) mloptions(mlogit
options)]
I want to estimate the wage equation with depvar "wageperhour" and the
selection equation in form of mlogit with "region" as depvar_m and
"skills" etc as varlist_m,
so that my command looks p.ex like this
selmlog wagehour x1 x2 x3, select (region_destination=skills x4 x5) dhl
The structure of my data is the following;
individual 1: wagehour x1 x2 x3 region-origin region-destination x4 x5 ...
individual 2: wagehour x1 x2 x3 region-origin region-destination x4 x5
My problem is that the help file tells me the following:
"The outcome variable depvar is observed for only one value of
depvar_m. It is important that depvar should have missing values for
any other value of depvar_m."
Since all my individuals exist only once in my dataset, I have have
only one "wage" for each and I wonder where I should have the missing
values and what is meant by "any other value of depvar_m". Do I have
to create a dataset where the observations per individual are repeated
for each possible outcome like in a conditional logit model?
If you can help me with my little problem or if you see that I
understood this command completely wrong or give me any examples of
the datastructure and dofiles when you used the selmlog command, I
would be very greatful!
Thank you very much in advance,
Laura Hering
Sorbonne-University, Paris 1
UFR 02 Economics
Maison des Sciences Economiques
75013 Paris
France
*
* 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/
