Gini coefficient calculator with Lorenz curve, Palma ratio, Theil and Atkinson indices, from incomes or income shares

What it does

How unequal is a set of incomes, wages, wealth or any other amount shared among people? This tool draws the Lorenz curve and works out the most used inequality measures, all in your browser:

• the Gini coefficient (0 is perfect equality, 1 is one person with everything), with the small-sample correction for unweighted data;
• income shares of each fifth (quintile), of the bottom 40 %, the top 10 % and the top 1 %;
• the Palma ratio, the top 10 %‘s share divided by the bottom 40 %‘s;
• Theil T and Theil L (mean log deviation), the two “generalised entropy” indices that can be split into within-group and between-group parts;
• the Atkinson index for any inequality aversion ε, the share of total income society would give up to have the rest shared equally;
• the P90/P10 ratio of the incomes at the 90th and 10th percentiles.

It works well with the biodiversity indices, the publication plot maker and the sample size and power calculator.

How to use it

1. Incomes: paste one income per row, or open a CSV or Excel file. Add a second column of weights when each row stands for several people (household size) or for a survey weight.
2. Grouped (shares): when you only have published shares, such as the share of income going to each fifth of households, give one row per group with its population (count or %) and its income (total or %). Everyone inside a group is treated as equal, so the results are lower bounds.
3. Set the Atkinson ε values you want, such as 0.5, 1 and 2.
4. Download the measures, the quintile table and the Lorenz curve points as CSV, or the report as an image, PDF or Word file.

Worked example (made-up incomes)

The 18 made-up incomes (in thousands) have a Gini of 0.399 (0.422 with the sample correction). The poorest fifth gets 6.1 % of all income and the richest fifth 46.8 %. The Palma ratio is 1.803: the top 10 % earn 1.8 times as much in total as the bottom 40 %. Theil T is 0.264, and the Atkinson index with ε = 1 is 0.233: the same welfare could be had with 23.3 % less income shared equally.

With the made-up quintile shares (6.5, 11, 15.5, 22 and 45 %), the Gini is 0.352 and the Palma ratio 1.286.

Tips

• Use the same income concept when comparing (before or after tax, per household or per person, with or without an equivalence scale).
• The Gini is most sensitive to the middle of the distribution, the Palma and top shares to the top, and the Atkinson index with high ε to the bottom.
• Zero incomes are fine for the Gini, Theil T and Palma, but Theil L is not defined and the Atkinson index for ε ≥ 1 becomes 1.

FAQ

Why is the corrected Gini higher? For a sample of n people the largest possible Gini is (n − 1)/n, so multiplying by n/(n − 1) lets small samples reach 1 and removes the downward bias.

Can I use wealth or other amounts? Yes, any amount that is zero or more: wealth, land, sales by firm, or citations by author.

Last updated 2026-11-09.