Description
Income inequality
Income inequality in the U.S. and in other countries is a major research topic in the social sciences.
A well-regarded collection of data is maintained by Branko Milanovic of the Stone Center on SocioEconomic Inequality. We will primarily deal with two data sets that we’ll call WPID and Ventile.
WPID is the Lakner-Milanovic World Panel Income Distribution data set (Stata file LM WPID web 2.dta.)
Potentially useful variables include:
• country: country name
• contcod: 3-letter country code
• bin year: the year for which incomes are estimated (1988 to 2008.) For technical reasons,
use this rather than the year variable.
• group: the income decile group the estimate is for, where a decile is 10% of the population.
“1” means the 10% of individuals in the country with the lowest income, while “10” means
the 10% of the individuals in the country with the highest income.
• RRinc: the per capita income of that decile in that country, in 2005 US dollars.
• RRmean: the mean per capita income of the whole country, in 2005 US dollars.
Ventile (Stata file ventile 2011 for release LCU.dta) contains more recent data (circa 2011.)
Some important variables are:
• contcod: 3-letter country code
• ventile: the income ventile group the estimate is for, where a ventile is 5% of the population.
“1” means the 5% of individuals in the country with the lowest income, while “20” means
the 5% of the individuals in the country with the highest income.
1
• ventile income: the per capita income of that ventile in that country. Important: This
number is NOT adjusted for currency or inflation, so it is not directly comparable to the
incomes in the WPID data set (or to other countries in the Ventile data set.)
One further data set, the World Income Distribution (WYD, Stata file wyd 88 05 for release.dta)
has been uploaded for your interest/reference, but you are not required to use this.
Full descriptions of the data sets can be found on Milanovic’s website.
Questions to answer
1. How has per capita income for each decile in the U.S. changed since 1988?
Use the WPID data set to answer this. Give both graphical and numerical results, and
describe in words both the differences in income level between the deciles and the differences
in changes in income for each decile.
2. How does the present distribution of income, relative to a country’s mean, differ
between selected countries?
Use the Ventile data set for the most recent data. For this question, compare the United
States (USA), the United Kingdom (GBR), Germany (DEU), and TWO other countries of
your choice using the most recent available data. We know that (for example) relatively lowincome people in the U.S. have higher incomes than relatively high-income people in Ghana,
so you’ll need to find some way of removing this effect. Find a clear way or ways to graphically
display your results, and briefly write about each country in turn, explain how its distribution
differs (or doesn’t differ) from the others.
3. Is the percentage of income earned by the top 5 percent in a country related to
mean income in that country? What about the percentage of income earned by
the bottom 5 percent? If so, what’s the relationship, and does the relationship have a
simple explanation, such as regional differences? Are there any outliers that require special
explanation?
Again, the Ventile data set has the most recent data.
Write a PDF report of no more than six pages, including graphs, addressing these
questions. The body of the report should not include code — it should be readable to someone
who has never used R (generally you should not just copy-paste output.) Additional graphs for
model checking can be placed in an appendix that does not count toward the six page limit and
which probably no one will read.
2
Notes and hints
• There is not necessarily any single objectively right answer to any question in the project.
However, some answers are better than others.
• Whenever you present something that isn’t just raw data, carefully explain what it is and
how it’s calculated.
• Think about whether it’s more appropriate to describe differences in absolute terms or in
percentage terms (or both.)
• You’ll be penalized for any P-value that appears in your report.
What to submit
• A PDF or other file containing your report.
• A .Rmd or other file containing your code.
• Any other supplementary files required to reproduce your work.
Grading
• Question 1: 5 points
• Question 2: 5 points
• Question 3: 10 points
• Communication: 10 points. Full credit for presentation requires a readable, informative,
comprehensive, clearly labeled set of graphs, and a comprehensible write-up with few glaring
spelling and grammatical errors that makes the main points of the analysis clear.
3