devcon

[andrea wurzel]

i am doing a paper about the descomposition of public and private sector wage gap.
To estimate this; I am using "oaxaca" command. I used a comand similiar to the one recomended by Ben Jann in : http://www.stata.com/statalist/archive/2007-04/msg00948.html.
The series of command i used were:
-gen byte privado = publicoprivado==1 | publicoprivado==8 -gen byte publico1 = publicoprivado==2
-reg lnw edad primaria ciclobasico secundobachiller bachitec enseñanzatec magisoprof universidad terciarionouniver sexo dregion rama0 rama1 rama2 rama3 rama4 rama5 rama6 rama7 rama8 rama9 tt0 tt1 tt2 tt3 tt4 tt5 tt6 tt7 tt8 tt9 [ fweight=pesoano] if privado -devcon , groups( rama0 rama1 rama2 rama3 rama4 rama5 rama6 rama7 rama8 rama9 , tt0 tt1 tt2 tt3 tt4 tt5 tt6 tt7 tt8 tt9 )
- estimates store privado
- reg lnw edad primaria ciclobasico secundobachiller bachitec enseñanzatec magisoprof universidad terciarionouniver sexo dregion rama0 rama1 rama2 rama3 rama4 rama5 rama6 rama7 rama8 rama9 tt0 tt1 tt2 tt3 tt4 tt5 tt6 tt7 tt8 tt9 [ fweight=pesoano] if publico1
- devcon , groups( rama0 rama1 rama2 rama3 rama4 rama5 rama6 rama7 rama8 rama9 , tt0 tt1 tt2 tt3 tt4 tt5 tt6 tt7 tt8 tt9) estimates store publico1 -reg lnw publico1 edad primaria ciclobasico secundobachiller bachitec enseñanzatec magisoprof universidad terciarionouniver sexo dregion rama0 rama1 rama2 rama3 rama4 rama5 rama6 rama7 rama8 rama9 tt0 tt1 tt2 tt3 tt4 tt5 tt6 tt7 tt8 tt9 [ fweight=pesoano]

- estimates store pooled
- oaxaca publico1 privado, ref(pooled) detail

when i used this in stata appear the problem: "option by() required"

how can i resolved this problem?, and why this problem doesn´t appear in the one recomenden by Dr Jann in :http://www.stata.com/statalist/archive/2007-04/msg00948.html?

[FernandoRios]

Hi Andrea,
I think the problem comes from using an old syntax of oaxaca command.
For instance, specific for your problem, it would be easier if you would just do the following
oaxaca lnw publico1 edad primaria ciclobasico secundobachiller bachitec enseñanzatec magisoprof universidad terciarionouniver sexo dregion rama0 rama1 rama2 rama3 rama4 rama5 rama6 rama7 rama8 rama9 tt0 tt1 tt2 tt3 tt4 tt5 tt6 tt7 tt8 tt9 [ fweight=pesoano] , by(privado)
That should give you the standard three fold decomposition. I suggest you to consult the help on that command for a more specific case.
Hope this helps
Fernando