Gini coefficient

The Gini coefficient summarizes inequality in a distribution, ranging from zero for perfect equality to one for maximal inequality. It is commonly used to describe income and wealth distributions.

Lorenz curve: A graph comparing a distribution’s cumulative share with the cumulative share of its population. The coefficient is geometrically related to the area between this curve and the equality line.
Income inequality: Unequal distribution of income among individuals or households. The Gini coefficient is widely used to summarize differences in income distributions.
Theil index: An inequality measure from the generalized entropy family that is decomposable across population groups. It supports exact between-group and within-group decomposition, which the standard Gini does not.
Relative inequality: Inequality measured by proportional differences rather than absolute gaps. The Gini is generally invariant to multiplying every value by the same positive factor.
Equivalence scale: A method for adjusting household resources to account for differences in household size and composition. Household Ginis can change substantially depending on how income is equivalized.
Gini mean difference: The average absolute difference between all pairs of values in a distribution. For nonnegative data, the coefficient can be expressed as this difference divided by twice the mean.
Wealth inequality: Unequal distribution of assets and liabilities among people or households. Wealth Ginis often exceed income Ginis because wealth is more concentrated.
Atkinson index: An inequality measure that incorporates an explicit sensitivity to inequality at different parts of a distribution. Its inequality-aversion parameter makes value judgments more explicit than the Gini does.
Absolute inequality: Inequality measured by differences in units rather than by relative proportions. Two distributions can share a Gini while having very different absolute gaps.
Top income measurement: Methods for estimating incomes at the upper end of a distribution, where surveys often undercount them. Missing high incomes can bias measured inequality downward.

Linked from 10 pages