Regression-based decomposition of inequality: “Levels” and “Differences” questions

[Ahmed Zohirul ISLAM]

I want to use Fields' (2003) regression-based decomposition approach to answer the following questions:

• Levels question: How much income inequality is accounted for by different explanatory factors?
• Differences question: How much of the differences in income inequality between the two periods is accounted for by each explanatory factor?

Which user-written command do I need to use to compute "Factor Inequality Weights" and the "Relative share of each factor in income inequality" to answer the Levels question?
Which command may I use to decompose income inequality in terms of the Differences question?

[Stephen Jenkins]

Check out -ineqrbd- on SSC for the 'levels' question. How you address the "differences" question is something I don't have time to discuss (sorry), but the problem is analogous to the question of how to use Shorrocks's decomposition of inequality by factor components to. For one relatively old paper doing this, see Jenkins, ‘Accounting for inequality trends: decomposition analyses for the UK, 1971–86’, Economica,62 (1), February 1995, 29–63.

[Ahmed Zohirul ISLAM]

Dear Professor Stephen Jenkins,
Thank you so much for being there when I need support in my research. You are really awesome.
I am grateful.