Poverty Headcount Ratio at $3.00 a Day

Population below $3.00 a day, 2021 PPP (%)·Higher is worse — rank 1 is worst affected·Published by World Bank Group

We hold 21 editions of this indicator, from 2002 to 2022 — 1,677 rows in total, and 73 countries in the newest.

21 reference years (2002–2022), covering 69 to 93 countries, from the World Bank’s own Poverty and Inequality Platform via its indicator API (SI.POV.DDAY). $3.00 a day at 2021 purchasing power parities is the Bank’s international extreme-poverty line. ⚠ THE LINE MOVED IN 2025, from $2.15 at 2017 PPPs. Poverty did not jump that year; the yardstick was re-based, so these figures are not comparable with published $1.90 or $2.15 counts from earlier reports. ⚠ THE COUNTRY SET CHANGES EVERY YEAR, exactly as it does for the Gini index and for the same reason: each figure comes from whichever household survey that country happened to run, so a country missing from an edition did not escape poverty — it did not run a survey. One country’s line over time is meaningful; comparing the shape of one year against another is not. ⚠ A zero here is real, not a gap: several high-income countries genuinely have nobody below this line. ⚠ Years outside 2002–2022 fall below 60 countries and are not imported. ⚠ The Bank publishes no rank, so none is stored; the POSITION shown on the statistics page is derived here by ordering the score, and is labelled as ours. Opted in 2026-09-10, once PIP’s two decimals thinned the ties: the deepest is now five countries in one edition, and those are genuine zeros — countries with nobody below the line, which share position 1 for a real reason — where the Bank’s one-decimal copy tied eight by rounding real readings such as Belgium’s 0.02% down to nothing. Read from the Bank’s Poverty and Inequality Platform (PIP), not its indicator API: the API re-publishes PIP rounded to one decimal, which turned Belgium’s 0.02% into 0. Re-sourced 2026-09-09. ⚠ Nineteen countries ran an income AND a consumption survey in the same year; only one can be an edition row, and the two differ by up to 8 index points. The choice is not made here — it is read out of the Bank’s own published figure for that country-year (income in 83 cases, consumption in 5).