Variations of an Oaxaca Decomposition

[david chong]

Hi I'm wondering whether anyone from the community can help me. I'm trying to understand the extent of occupational segregation on the racial income gap and I have decided to use a variation of oaxaca decomposition. I'm just wondering does stata allow this variation?

I have attached a file below that shows the kind of decomposition that I'm interested in learning how to code on STATA. Hope to hear from the community

[Weiwen Ng]

David,

This parameterization of the Oaxaca/Blinder decomposition looks like something that is available with the user-written command -oaxaca- by Ben Jann. His Stata Journal article has a full explanation, including equations and worked examples from a sample dataset. You can install the command from within Stata by typing:

ssc install oaxaca
help oaxaca

This is a bit like installing an R package, if you're familiar with those. Jann is quite well-regarded by the Stata community. I have only used this command for a linear decomposition (i.e. running OLS models), which should suffice for your purposes. It can handle a logistic decomposition as well, but it would be restricted to stating the decomposition in terms of log odds (but the PDF does link to other approaches).