Global Inequality Through History
In 1820, most inequality was within countries — global rich and poor lived next to each other. By 1980, most inequality was between countries — rich countries vs poor countries. Since 1990 the picture has shifted back: emerging-market growth (China, India) has compressed between-country inequality while within-country inequality has risen in many advanced economies.
Key insights
The elephant curve, explained
Branko Milanovic and Christoph Lakner's 1988–2008 'elephant curve' showed real income growth by global income percentile. The body: high growth around the 40th–60th percentile (mainly Asia's rising middle classes). The dip where the trunk begins: very low growth at the 75th–85th (Western middle classes). The raised trunk: very high growth at the 99th+ (the global elite). Extended through 2018, the curve has flattened but the basic shape — emerging-market middle growing, advanced middle stagnant — remains.
Between vs within decomposition
Global inequality (the variance of world incomes) decomposes into between-country (variance of country means) and within-country (variance within each country) components. In 1820 within-country was dominant. In 1980 between-country was dominant. Today they are roughly equal — China's rise has narrowed between-country gaps; rising US, UK and Chinese top-shares have widened within-country gaps.
Top shares: U-shape in many advanced economies
The income share of the top 10% in the US fell from ~50% (1928) to ~33% (1970) — the 'Great Compression' — and has risen back above 45%. The pattern is similar but milder in the UK; flatter in continental Europe (especially France); the opposite in much of Latin America (top shares fell post-2000 then stabilized). 'Inequality' is not a single global trend — it depends entirely on which country and which decade.
World Gini coefficient 1820–2020
Gini of individual world incomes; higher = more unequal
Show data table
| World Gini | |
|---|---|
| 1820 | 0.56 |
| 1870 | 0.59 |
| 1910 | 0.61 |
| 1950 | 0.65 |
| 1980 | 0.69 |
| 1990 | 0.72 |
| 2000 | 0.69 |
| 2010 | 0.66 |
| 2020 | 0.62 |
Key Finding: World inequality rose continuously from 1820 to roughly 1990, then began to fall as China and other emerging markets caught up.
Top 10% income share — selected countries
Pre-tax national income share of richest 10%
Show data table
| USA | UK | France | World | |
|---|---|---|---|---|
| 1900 | 40 | 47 | 51 | 55 |
| 1930 | 44 | 44 | 48 | 55 |
| 1950 | 33 | 28 | 32 | 55 |
| 1970 | 32 | 27 | 30 | 55 |
| 1990 | 38 | 36 | 32 | 57 |
| 2010 | 45 | 39 | 33 | 55 |
| 2024 | 46 | 36 | 33 | 52 |
Key Finding: Most advanced economies show a U-shape; emerging markets show more variation. The US is on the high end among advanced economies.
Methodology & caveats
Income vs wealth inequality
Income inequality measures the flow of earnings in a year; wealth inequality measures the stock of accumulated assets. Wealth is always more concentrated than income — top 1% wealth shares are typically 2–3× top 1% income shares. Trends can diverge: a country can become more equal in income while becoming more unequal in wealth if asset prices rise faster than wages.
Pre-tax vs post-tax
Pre-tax income includes wages, capital income and self-employment. Post-tax income subtracts direct taxes and adds public transfers (pensions, benefits). For most advanced economies, taxes-and-transfers reduce the Gini by 10–20 percentage points. The choice of pre- or post-tax matters more than the choice of inequality metric.
Historical reconstructions
Pre-1980 inequality data relies on tax records (top shares — Piketty, Saez, Zucman methodology), household budget surveys (post-WW2), and historical wage and price reconstructions (pre-1900). The World Inequality Database harmonizes these sources. Pre-1900 figures should be read as orders of magnitude, not point estimates.
Why long-run Ginis differ
The 1820 and 1910 figures rest on national averages plus guesses about the distribution inside each country. Estimates by Milanovic and by Bourguignon and Morrisson, and the World Inequality Database, share the shape but not the level, partly because they use different PPP benchmarks. When the 2005 PPP revision cut China and India's real incomes, global inequality estimates jumped overnight with no change in anyone's life. Tax-based series also find higher top shares than surveys, which is why WID tends to show less decline since 1990.
What would change the picture
The fall since 1990 was driven by China, and China is now richer than the global median, so its further growth begins to add to world inequality. The next decline depends on India and Africa. If African population growth outruns its income growth, the world Gini may stop falling.
Cite this page
GlobeChart. “Global Inequality Through History.” Last updated 2026-10-31. https://globechart.com/economy/inequality-history/
Please also credit the original publisher named under each chart: World Inequality Database, Milanovic 'Global Inequality' historical reconstructions; World Inequality Database 2024. Spotted an error? Tell us.