Regression-based inequality decomposition following Fields (2003) and Heckman selection

[Aleksandra Anic]

Dear all,

Is it possible to use regression-based inequality decomposition (user-written command ineqrbd) where dependent variable (i.e. lnwage) is corrected for sample selection. I need to implement Heckman model with inequality decomposition.

Hope to read your answer soon.

Thank you.

Best,
Aleksandra

[Weiwen Ng]

Aleksandra,

I have never used -ineqrbd- before, so I cannot answer your question directly. However, the Oaxaca decomposition command written by Ben Jann and available from SSC will support Heckman's selection correction. I don't know how -ineqrbd- differs from Oaxaca decomposition, but I get the sense that the Oaxaca decomposition is a generally accepted technique - if you decide to switch to Oaxaca, then I doubt that reviewers will complain.

Perhaps I am wrong and someone more familiar with -ineqrbd- can explain.

[Aleksandra Anic]

Thank you for your answer, I know for oaxaca command, but it is not what I need. Oaxaca is for "discrimination", and ineqrbd is for inequality decomposition.

Best,
Aleksandra

[Stephen Jenkins]

More correctly, -oaxaca- is for decomposition of (differences in) means; -ineqrbd- is for decomposition of variances of log outcomes where outcome is typically an "income" variable. The short answer to the original question is, as far as I know, "no".