Bangladesh reduced extreme poverty mainly through work: garment jobs
for rural women, farm and rural jobs, and money sent home by migrants.
About one person in two lived below $3.00 a day in 1991 and about one in
17 in 2022, about as fast a fall as the rest of South Asia; its gains in
health and schooling were faster. Poverty has risen since 2022, and the
size of the recent fall is disputed.
Work, wages and money sent home
Bangladesh reduced extreme poverty mainly through work. Garment factories hired millions of women from the countryside, farm and other rural jobs paid more, and migrants abroad sent money home; together, earnings drove more than half of the fall in poverty between 2010 and 20228.
In 1991 about one Bangladeshi in two lived on less than $3.00 a day, the World Bank’s extreme-poverty line (adjusted for what money buys locally). By 2022 it was about one in 177.
On income alone, that fall matched South Asia’s and India’s rather than beating them7. What sets Bangladesh apart is health, family size and girls’ schooling, far ahead of countries with similar incomes14. Since 2022 poverty has risen, on World Bank estimates10, and a review commissioned by the interim government questions how much of the recent fall is real12.
That's the fall in people living on less than $3.00 a day between 1990 and 2022, by our arithmetic from World Bank data
47 million fewer poor, on three different yardsticks
The long fall shows up on every measure, but its exact size depends on which one you use. On the World Bank’s international line, 51% of Bangladeshis were in extreme poverty in 1991 and 5.9% in 20227. The 2022 survey was run differently from earlier ones, with interviewers using tablets and asking about a longer list of household spending, so the drop across it is approximate7.
Bangladesh also has two national poverty lines: a lower one for extreme poverty, based on the cost of a basic food basket, and a broader upper one. The cleanest comparison uses the lower line, recalculated so the years match. On that basis extreme poverty fell from 12.2% in 2010 to 5.6% in 2022: about 9 million fewer people, with 9.3 million still below the line8.
Those 47 million are about one in every 30 people the world lifted out of extreme poverty over those years, from a country with one in 50 of its people, by our arithmetic7. Part of that outsized share is where Bangladesh started: half its people were below the line in 1990, against 43% worldwide7.
The world’s share in extreme poverty, on the same $3.00-a-day line.
Share of the world's population in extreme poverty
Living on less than $3.00 a day, right now
Garment jobs, farm work and remittances
The garment industry helped by putting poor, mostly rural women into paid work. Between 2010 and 2016 industry, above all garments, accounted for 42% of the fall in poverty in towns and cities8. After 2016 industry jobs shrank, poverty among garment-worker households rose from 15% to 18.7%, and farm and other rural jobs took over the lead8.
The industry employs around 4 million people, 60.5% of them women in 2018, and earns about 82% of the country’s export revenue5. It came at a cost: the collapse of the Rana Plaza factory building in 2013 killed over 1,100 people, mostly garment workers, and 3,780 export factories were then inspected for structural, electrical and fire safety5.
It also changed more than wages. Girls living near garment factories married and had children later1.
The factories did more for girls’ schooling than the government’s own programme paying families to keep their daughters in school1.
Money from Bangladeshis working abroad reached $33.9 billion in 2025, 7.4% of the economy11. The World Bank simulated taking that money out of household budgets, assuming nothing else changed, and found 3.8 million more people would have been poor in 2022; without government social assistance, about 1.5 million more8.
Income fell as fast as India’s; health outran the country’s income
Bangladesh’s income poverty did not fall unusually fast. On the $3.00 line, South Asia went from 49.7% in 1990 to 6.1% in 2022 and India from 49.4% to 6.3%, almost the same path as Bangladesh’s 51.1% to 5.9%7.
What stands out is everything else. A woman in Bangladesh had 6.8 children on average in 1971 and 2.1 in 2024, and the share of children dying before five fell from 147 per 1,000 in 1990 to 31 in 202411 (see how the world’s child deaths fell). Development economists call this the “Bangladesh surprise”: health, schooling and gender outcomes well ahead of countries with similar incomes, not explained by income or government spending alone14.
No one has modelled a Bangladesh without the garment boom or its aid groups, so how much of the fall each one caused is unknown.
Workers, migrants and village programmes, more than the state
Government spending played a smaller part than wages: social transfers account for about 5% of the fall in poverty between 2016 and 20228. Most of the work was done by the women who moved to factory towns, the workers who went abroad and the farmers whose incomes rose. The government’s own review of the economy credits a “development surplus” produced by non-state groups and informal institutions, which delivered “low cost solutions in agriculture, public health, educations, and infrastructure”12.
Some of those programmes have been rigorously tested. A nationwide programme that gave the poorest women livestock and training was evaluated in an experiment across 1,309 villages; seven years later the women were still less poor2.
Microcredit’s record is weaker than its fame
Small loans to poor women are the Bangladeshi idea best known abroad, but there is little evidence that they reduced poverty much. The best-known study finding that they did, based on Bangladeshi data, lost its poverty result when two economists reanalysed it and dropped outliers3.
The original study’s author has disputed the reanalysis, and the argument is unresolved3. Six randomised trials of microcredit in other countries found “modestly positive, but not transformative, effects”4. The programme that did work for the poorest, the asset transfer above, gave women something to earn with rather than a loan.
The direction is reliable; the size of the recent fall is not
The direction isn’t in doubt: poverty fell on every line and in every survey, and even the committee that audited the figures found it fell between 2016 and 2022 “irrespective of the poverty line”12. How big the recent fall was is less certain, because the 2022 survey changed method and the committee reports allegations of pressure on the poverty statistics.
The committee was appointed by the interim government in 2024 to review the country’s economic record12.
The real one is the social progress. The statistical one, the committee says, is the claim of exceptionally fast growth, which it calls “a figment of statistical manipulation”12. It reports allegations of “strong pressure on poverty statistics in 2016 and 2022”, and notes that the change in survey method increased the measured fall between 2010 and 202212. The World Bank says the official growth figures for 2015 to 2022 “should be interpreted with caution”8.
The lower line itself is low. It rests on a basket of 11 foods that now make up less than 40% of what poor households eat8. Raise the line by 20%, the committee finds, and the poverty rate nearly doubles and extreme poverty more than doubles12.
Poverty has risen since 2022, on World Bank estimates
Yes: on World Bank estimates, poverty has risen since the last household survey in 2022. On Bangladesh’s upper national line, the broader of its two (the 5.6% above is on the lower one), poverty rose from 18.7% in 2022 to an estimated 21.4% in 2025, adding 1.4 million poor people in 202510. On the $3.00 line, the estimates go from 6.3% in the 2023 financial year to 9.0% in 20259.
Prices rose by 10% in 2025 while growth slowed to 3.5%, so wages bought less9, and employment fell by nearly 2 million between 2023 and 20248. The share of women working or looking for work fell from 43.7% in 2022 to 38.4% in 2024119.
A third of the country just above the line
The rise since 2022 shows how thin the margin is. In 2022, 36% of Bangladeshis, 62 million people, lived just above the poverty line8, so a year of 10% inflation and lost jobs didn’t need to reach the very poorest to push 1.4 million people back below it9810.
The World Bank’s model has the $3.00 rate easing slightly, to 8.7%, in 20269; the first household survey since 2022 will be the real test. Bangladesh is due to leave the UN’s list of least developed countries in November 2026, at the end of a five-year preparation period6, and climate change could cut farm output by a third by 20508, which would hit the rural work that led the fall after 2016.
Our reading is that the fall was real and built on work, and that is also what makes it fragile: an economy whose exports are 82% garments, with a third of its people just above the line, can lose ground fast.