Announcement

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

  • Inequality decomposition

    I want to decompose the inequality with subgroup. More specifically, within group and between group. My inequality index is GE (2). I use various command ineqdecgini, ginidesc,ineqdeco, and ineqdec0. these gave me the decomposition result. All of these give me the residual. I want to see, how total 100% distributed between within and between group, not the residual components.
    Click image for larger version

Name:	decomposition.png
Views:	1
Size:	65.1 KB
ID:	1752015

  • #2
    I do not understand: -ineqdeco- and -ineqdec0- do not provide "residual" terms. All Generalized Entropy indices are additively decomposable by population subgroup, i.e., total inequality can be expressed as Within-Group Inequality + Between-Group Inequality. W-G inequality is a weighted sum of the inequalities for each subgroup. B-G inequality is the inequality arising if each obs is attributed with the average income of the subgroup to which they belong. The Gini index is not decomposable thus (cf. -ineqdecgini- on SSC) except in the case where the subgroups do not overlap along the income range. In general, there is an additional term (in addition to W-G and B-G), a 'residual' that reflects the overlapping. Please may I advise reading some of the standard literature about this stuff. Good luck

    Comment


    • #3
      Thanks Stephen Jenkins please do not take it otherwise. I am very new in this fields. Please look at my attachment that I have shared with my main post. they allocate total contribution into W-g and B-G. they didn't add any residual or overlapping term. Is it possible? or are they correct?

      Comment


      • #4
        Nadim: If you're new to the field, it's important to do background reading regarding the methods before rushing to the data, however attractive that is. Your table refers to the GE(2) index, i.e., half the squared coefficient of variation. It is a member of the GE family. Now reread my earlier post to understand why there is no residual.

        Comment


        • #5
          Thanks Stephen Jenkins : Can you please share some reading. It would be great for me. As a learner, I need your help. Please do sir!

          Comment


          • #6
            You could start by looking at materials cited in the help file for -ineqdeco-. As a learner, you do need to take some initial initiatives yourself!

            Comment

            Working...
            X