slides_02_measures_continued.pdf

# slides_02_measures_continued.pdf - Econ 473 Class 2...

• 58

This preview shows pages 1–15. Sign up to view the full content.

Econ 473 Class 2 Numerical inequality measures

This preview has intentionally blurred sections. Sign up to view the full version.

Today Numerical inequality measures I Gini coefficient I Equalizing transfers Desirable properties of inequality measures Stata introduction Next: Link measurement to theory
Numerical inequality measures

This preview has intentionally blurred sections. Sign up to view the full version.

Common inequality measures variance coefficient of variation Gini coefficient log variance quantile-based measures Later: theory-based measures
The variance V = 1 N N X i =1 ( y i - ¯ y ) 2 An obvious candidate. Problem: It rises with average income. For example, doubling all incomes, V 0 = 1 N N X i =1 (2 y i - y ) 2 = 1 N N X i =1 4( y i - ¯ y ) 2 = 4 V . Two remedies.

This preview has intentionally blurred sections. Sign up to view the full version.

The coefficient of variation cv = sd The previous problem disappears: cv 0 = sd 0 0 = 2 sd / (2 μ ) = cv .
The log variance is actually the variance of log income: v = 1 N N X i =1 (ln y - ln y ) 2 . Again, this is insensitive to scale: v 0 = 1 N N X i =1 (ln(2 y ) - ln(2 y )) 2 = 1 N N X i =1 (ln 2 + ln y - ( ln 2 + ln y )) 2 = 1 N N X i =1 (ln 2 + ln y - ln 2 - ln y ) 2 = v .

This preview has intentionally blurred sections. Sign up to view the full version.

Sensitivity to equalizing transfers A desirable feature of an inequality measure is that it decreases if an equalizing transfer is made: Reduce y i by 1 and raise y j by 1, i > j in ordered data. Where necessary, assume that this does not affect rankings. Δ V = V y j - V y i ¯ y constant = 1 N [2( y j - ¯ y ) - 2( y i - ¯ y )] = 2 N ( y j - y i ) < 0 . Δ c = 1 sd 1 N y j - y i μ < 0 . Both of these depend only on y i - y j , not on the position of i and j in the distribution. This appears desirable.
Sensitivity to equalizing transfers – 2 The log variance: Δ v = 2 N ln y j y j - ln y y j - ln y i y i - ln y y i Given y i - y j , this increases in y i , which is probably desirable, but: it can turn > 0! That is undesirable. (It happens at very high incomes. Try it out. ) Since Δ v increases in y i , redistribution in the lower tail affects this inequality measure most.

This preview has intentionally blurred sections. Sign up to view the full version.

Other aspects V ignores the skewness of empirical income distributions. The log variance captures the distribution perfectly if income is actually log-normally distributed. A skewness statistic is an interesting additional aspect, but not a measure of inequality per se .
The Gini coefficient Probably the most popular “fancy” inequality measure. Developed by Corrado Gini, 1912. Defined as (twice) the area between the 45 degree line and the Lorenz curve. Ranges between 0 (full equality) and 1 (one person has it all). (If all incomes are positive.) There are alternative definitions of what it captures.

This preview has intentionally blurred sections. Sign up to view the full version.

The Gini coefficient 1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 Proportion of population Fraction of income Gini coe ffi cient
The Gini coefficient A definition based on the graph: G = 1 - 2 Z Φ( y )d F ( y ) There are also alternative definitions. For example, in the discrete case with N individuals, G = 1 2 N 2 N i =1 N j =1 | y i - y j | ¯ y , the relative mean difference over mean income.

This preview has intentionally blurred sections. Sign up to view the full version.

The Gini coefficient: Sensitivity to transfers Reduce y i by 1 and raise y j by 1, i > j in ordered data, without changing the ranking.
This is the end of the preview. Sign up to access the rest of the document.
• Fall '17
• MarkusPoschke

{[ snackBarMessage ]}

### What students are saying

• As a current student on this bumpy collegiate pathway, I stumbled upon Course Hero, where I can find study resources for nearly all my courses, get online help from tutors 24/7, and even share my old projects, papers, and lecture notes with other students.

Kiran Temple University Fox School of Business ‘17, Course Hero Intern

• I cannot even describe how much Course Hero helped me this summer. It’s truly become something I can always rely on and help me. In the end, I was not only able to survive summer classes, but I was able to thrive thanks to Course Hero.

Dana University of Pennsylvania ‘17, Course Hero Intern

• The ability to access any university’s resources through Course Hero proved invaluable in my case. I was behind on Tulane coursework and actually used UCLA’s materials to help me move forward and get everything together on time.

Jill Tulane University ‘16, Course Hero Intern