Announcement

Collapse
No announcement yet.
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • Wage decomposition for migrant-native using the code oaxaca

    I study wage differential between migrant and natives and I use oaxaca as below

    Code:
    oaxaca depvar [indepvars] [if] [in] [weight], by(groupvar) [options]
    for the wage decomposition. I read the STATA journal by Ben Jann (2008) for a guide on how to use Blinder-Oaxaca decomposition framework in STATA. The problem is that in Jann paper the regressors are the same for males and females but in my work immigrants have additional regressors (e.g., year of immigration, migrant cohort, country of origin and etc). How can I specify different regressors for groups in Oaxaca framework?

  • #2
    Hi Taghi,
    You simply add all of them for both immigrants and non immigrants.
    For nonimmigrants, the variables like year of immigration should have a value of zero, (or any other constant for that matter). That way, it will be omitted due to collinearity. Other than that, you just need to specify the option -relax-.
    That being said. While the program allows you to do that, Im not sure to what extent such decomposition is valid. After all, one of the assumptions is that the OB, or as far as I remember, is that the same factors affect both groups' wages.
    Fernando

    Comment


    • #3
      Hi Fernando,

      Thank you for your reply.
      As I understood, I should consider the specification for migrants for both groups and since there would be several zero variables in the native equation I will need to relax collinearity. Is that right? please correct me if I am wrong.

      Taghi
      Last edited by Taghi Ghadiri Abkenar; 09 Nov 2019, 02:19.

      Comment


      • #4
        How can I use option -relax in oaxaca framework? I couldn't find it by help oaxaca in STATA.

        Comment


        • #5
          Here is an example:
          Code:
          use http://fmwww.bc.edu/RePEc/bocode/o/oaxaca.dta
          oaxaca lnwage educ exper tenure, by(female)
          replace single=0 if female==0
          oaxaca lnwage educ exper tenure single, by(female) relax
          As you can see, I made it so single is always zero for men. If I do not add the relax option, it would stop with an error.
          But adding the option relax, it would proceed without any other problems.

          Also, if you do not see the option "relax", perhaps you have an older version of the oaxaca command. Try re-installing it.
          HTH

          Comment


          • #6
            Thank you so much. It works now. I had an older version of oaxaca command.

            Comment

            Working...
            X