ECON 426 Lecture Notes - Lecture 18: Percentile, Gini Coefficient, Absolute Difference
It has a long right tail (right skewed)
○
The mean is > than the median
○
Some individual earn very large salaries and relative salaries at the top increase rapidly
○
The density of earnings isn't symmetric
•
Symmetric with mean = median
○
95% of observations within 2 s.d. of the mean
○
90% within 1.6 s.d. of the mean
○
Log earnings are close to normal
•
Describing Earnings Distributions
•
Skew: the Very Top
Similar: P80 - P20, P50 - P10, P90 - P50
▪
P90 - P10 ratio of wages = ratio of 90th percentile wage to 10th percentile wage
○
Easy to understand and interpret
▪
Ability to control what part of earnings distribution one focuses on
▪
Advantages of using Percentile Ratios:
○
Disadvantage: percentile ratios compare just two points, but don't describe inequality along the whole distribution
○
Percentile Ratios
•
Describing Inequality
Lecture 18 - Earnings and Income Inequality
Tuesday, March 27, 2018
3:19 PM
ECON 426 Page 1
Perfect equality is a straight line
▪
Perfect inequality is an "inverted L"
▪
Lorenz Curves show the share of total income earned by all workers up to the amount indicated on the horizontal
○
Also: Gini is 1/2 the "relative mean absolute difference" = 1/2(|difference in incomes between random pair|/
average income)
▪
Gini Coefficient:
○
Below 0.3 - quite equal
▪
Above 0.5 - quite unequal
▪
Interpretation?
○
Canada ~ 0.34
▪
US ~ 0.37
▪
France ~ 0.29
▪
Mexico ~ 0.48
▪
After taxes and transfers:
○
•
○
Gini and P90 - P10 of Wages
•
Trends in Inequality
Another Way to Look at Wage Inequality
ECON 426 Page 2