Global Inequality
Global inequality — inequality among all individual humans on Earth — has fallen meaningfully since ~2000. The decline was driven entirely by between-country convergence (mostly Asia rising). Within-country inequality has been roughly stable in aggregate but with major variation. The 'elephant curve' visualizes the income growth pattern: emerging-market middle gained substantially, advanced-economy middle stagnated, global top grew fastest.
Key insights
The elephant curve, briefly
Lakner and Milanovic (2016) plotted real income growth 1988-2008 by global percentile. The shape: very low growth at the 75-85th percentile (Western middle classes), high growth at the 40-60th (Chinese middle), and very high growth at the 99th+ (global elite). Updated through 2018, the curve has flattened — China's contribution to the body is smaller as Chinese growth slowed and started benefiting different segments — but the basic pattern persists.
Between-country was the bigger contributor — once
In 1988, between-country differences (rich countries vs poor countries) explained ~65% of total global inequality. Within-country differences explained ~35%. By 2024, the split is roughly 45/55 — within-country has overtaken between-country. China's rise narrowed between-country gaps; widening US, UK and Chinese top-shares widened within-country gaps. The total has fallen because between-country reduction was larger than within-country increase.
Future depends on Africa
Continued global inequality decline requires sub-Saharan African economic convergence — and so far, it has not happened at the same scale as Asian convergence. Sub-Saharan African per-capita growth has been 1-2% per year for most of the past two decades, vs 4-6% for Asia. If African convergence accelerates, global inequality will continue to fall. If it stalls, global inequality could rise again — especially with within-country inequality rising in many countries.
Global Gini coefficient 1820–2024
Inequality across all individual humans
Show data table
| Global 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 | 0.63 |
| 2024 | 0.62 |
Key Finding: Continuous rise through the 19th and 20th centuries; reversal post-2000 driven by Asian growth.
Stylized elephant curve — real income growth by global percentile
Cumulative income growth 1988–2018 by global income percentile
Show data table
| % real income growth 1988-2018 | |
|---|---|
| P10 | 100 |
| P20 | 120 |
| P30 | 130 |
| P40 | 140 |
| P50 | 140 |
| P60 | 90 |
| P70 | 40 |
| P75 | 15 |
| P80 | 30 |
| P85 | 50 |
| P90 | 70 |
| P95 | 80 |
| P99 | 120 |
| P99.9 | 140 |
Key Finding: Top 1% and emerging-market middle gained most; Western middle classes stagnated.
Methodology & caveats
Theil decomposition
Theil index of inequality is additively decomposable into between-group and within-group components — useful for between-country / within-country splits. Gini is not strictly decomposable but can be approximated. The decomposition depends on group definitions; using countries as groups is conventional but treating regions or income deciles produces different shares.
PPP-adjusted
Global inequality measures use PPP-adjusted incomes (otherwise market exchange rates would make poor-country incomes appear smaller). Most analyses use World Bank PIP data plus household surveys reconciled to national accounts. The 2017 ICP revision and the 2021 ICP revision both produced significant updates to global inequality estimates.
Top tail challenges
Household surveys systematically undersample the rich — top 1% wealth is poorly captured. Lakner-Milanovic and others combine survey data with tax records to better estimate the top tail. Critical for elephant-curve analysis: the 'tip' of the elephant trunk (top 1%) is most uncertain. Recent improvements (WID combining surveys and tax data globally) have made top-share estimates more reliable.
Why estimates of global inequality disagree
Studies differ on level and on the between-country share for three reasons. The index matters: the Gini is most sensitive to the middle of the distribution, the Theil to the top, so a Theil-based decomposition gives a different split from the figures here. The data matter: survey-only estimates such as the World Bank's record lower inequality than estimates like the World Inequality Database's, which add tax records and tend to show a much larger within-country share. Prices matter: each new PPP round changes the relative incomes of China and India, and those two countries alone can shift the global Gini. Treat the values on this page as one reasonable series, not a measured fact. The direction since 2000, falling, is the robust finding across studies; the size of the fall and the 45/55 split are less certain. Growth in Africa and India will shape what happens next.
Cite this page
GlobeChart. “Global Inequality.” Last updated 2026-10-31. https://globechart.com/poverty/global-inequality/
Please also credit the original publisher named under each chart: World Inequality Database, Milanovic 'Global Inequality' historical reconstructions; Lakner-Milanovic 2016, updated 2024. Spotted an error? Tell us.