Escaping the Pollution–Poverty Trap

The cooking-fuel cycle is the most direct link between poverty and environmental damage. In 47 countries most people still cook with wood, charcoal, dung, crop waste or kerosene. The smoke is a leading cause of disease, and collecting the fuel takes hours, mostly from women and girls. Where demand for wood and charcoal outruns regrowth, it strips woodland around towns, so the fuel must be gathered from ever further away. Median extreme poverty in these countries is 40%, against under 1% elsewhere.

The water and sanitation cycle runs through the body. Where most people lack even basic sanitation, as in 34 countries, faecal contamination spreads diarrhoea and environmental enteric dysfunction. These diseases stunt children's growth, cut learning and lower lifetime earnings. Median poverty in these countries is 44%. The land cycle runs through the farm: degraded soils lower yields, which push cultivation onto steeper and drier land. The wealth cycle runs through the national balance sheet. When forest and mineral rents are consumed rather than reinvested in people and infrastructure, total wealth per person falls and so do future incomes.

The air cycle is different. It affects 34 countries and 3 billion people, most of them in South Asia, the Sahel and the Gulf, where dust from the Sahara and Arabia combines with smoke from coal, vehicles, brick kilns and crop burning. Its link to poverty runs through health and productivity rather than through household survival. Its median poverty rate (17%) is lower than for the household cycles, because many of these countries have grown out of extreme poverty without escaping the smoke.

If the cycles were independent, a country's position on one would say little about the others once income is accounted for. The table shows correlations among developing countries, both raw and after removing the influence of income (the correlation of residuals from a regression on log GDP per head). The household cycles cluster strongly. Polluting cooking fuel and lack of basic sanitation are correlated at 0.85, and still at 0.56 among countries at the same income. Lack of sanitation is linked to forest loss (0.33 at the same income) and to extreme poverty (0.44). Negative adjusted net savings, the wealth drawdown, is linked to poverty (0.31).

These correlations are consistent with reinforcing cycles, though they do not prove that one causes another. They are what one would expect if the same households, short of cash, infrastructure and secure rights, depend on the same degrading practices. They also mean that a country in one household cycle is likely to be in the others. A programme that tackles only one, such as a stove programme without fuel supply or toilets without water, will often see its gains eroded by the others.

Air pollution does not follow this pattern. At the same income, fine-particle pollution is not correlated with cooking fuel or sanitation, and it is negatively correlated with extreme poverty (−0.29). Among countries at the same income, the more industrialised and urbanised, and the dustier, are both less poor and more polluted. Land degradation as reported for SDG 15.3.1 is uncorrelated with everything, which probably reflects how differently countries measure it as much as any real independence. The trap at the core is a household trap. Air is the cycle that follows growth.

Counting the cycles each country is trapped in gives a simple measure of the trap's depth. In 2024, 72 countries with 2.1 billion people were trapped in none, and their population-weighted extreme poverty rate was 2.5%. Thirty-nine countries with 4.4 billion people, including India, China and Pakistan, were trapped in one, usually air or land, and their poverty rate was 3%. Then the picture changes sharply. Twenty-three countries in two cycles had 34% of their people in extreme poverty, 12 countries in three cycles 34%, and 13 countries in four cycles 51%.

The 48 countries in two or more cycles hold 1.6 billion people but 639 million of the world's 823 million extremely poor people, or 78%. Thirty-six of them are in sub-Saharan Africa. The 13 countries trapped in four cycles are Nigeria, the DR Congo, Uganda, Madagascar, Cameroon, Guinea, Benin, Rwanda, Burundi, Sierra Leone, Congo, the Gambia and Guinea-Bissau. Child stunting follows the same gradient: a median of 7% in countries with no cycle, 27% with two and 31% with four.

The gradient is steep because the cycles compound. A household that cooks with wood, lacks a toilet and farms degraded land faces several health and productivity losses at once. Each loss makes the clean alternatives less affordable. Countries in this position also tend to have the weakest public finances, so the government cannot fill the gap. That is what distinguishes a trap from a set of separate problems: the way out of each cycle is blocked partly by the others.

Among low- and lower-middle-income countries, those where most people still cook with polluting fuels lost a median 9.5% of their forest area between 2000 and the early 2020s, while the rest gained 0.9%. Across these countries, a higher share of polluting fuel goes with faster forest loss (r = 0.29, and 0.24 at the same income). Firewood and charcoal are not the only drivers of deforestation, since agricultural expansion and logging dominate in many places. But they degrade woodland near cities and roads, and they tie household energy to the forest's future. Burkina Faso, Benin, Chad, the Gambia, Côte d'Ivoire, Malawi and Uganda all lost between a quarter and a half of their forest area.

The sanitation cycle shows how quickly a cycle can be broken. In 2000, 73% of Indians practised open defecation, and by 2024 the figure was 7%, after the Total Sanitation Campaign and the Swachh Bharat Mission built household toilets at scale and campaigned on use. Nepal went from 68% to under 1% through district-by-district open-defecation-free campaigns, Cambodia from 88% to 5%, Laos from 61% to 9% and Ethiopia from 72% to 16%. Globally, 354 million people still practised open defecation in 2024, and 3.4 billion lacked safely managed sanitation. Open defecation remains above 60% in Niger, South Sudan and Chad.

Forest recovery is the third escape route. Rwanda's forest area grew by more than half after 2000, Burundi's by 44%, China's by 28% and Viet Nam's by 26%, through state reforestation programmes, community management and, in Viet Nam, the 1998–2010 programme to plant five million hectares. In Niger, farmers have regrown trees on more than five million hectares by protecting the stumps and roots already in their fields. This practice, farmer-managed natural regeneration, spread through changes to tree tenure and the forest code, not tree planting.

Comparing each country's position in 2000 with today shows how often cycles are broken. Of 68 developing countries where most people cooked with polluting fuels in 2000, 22 crossed to majority clean cooking: India, China, Indonesia, Pakistan, the Philippines, Viet Nam, Peru and Cambodia among them. Of 51 where most lacked basic sanitation, 19 crossed. Of 42 with fine-particle pollution above 35 µg/m³, 15 got below it, led by China and Viet Nam. Of 28 with negative adjusted net savings in 2000–05, 16 turned positive.

Fifty-two developing countries escaped at least one cycle. Their extreme poverty fell by a median 26 percentage points between 2000 and 2024, against 11 points for the 33 countries trapped in 2000 that escaped none. Escape and poverty reduction are intertwined. Growth paid for clean fuels and sanitation, and those investments in turn raised health and productivity. Indonesia escaped three cycles (fuel, sanitation and wealth drawdown) while its extreme poverty fell from 66% to 5%. Viet Nam escaped air, fuel and sanitation while poverty fell from 41% to 1%, and Cambodia escaped fuel and sanitation as poverty fell from 63% to 8%.

Few countries moved the other way. Mongolia entered the air and wealth cycles, the Gambia fell below half basic sanitation coverage, conflict pushed most Yemenis back to polluting fuels, and Brazil, South Africa, Tunisia and Libya slipped into negative adjusted net savings. The bigger concern is the countries that never left: Nigeria, the DR Congo, Madagascar, Uganda, Niger, Burundi and Chad remain in three or four cycles. Extreme poverty in these countries still ranges from about 40% to over 80%, and their populations are among the fastest-growing in the world.

Fine-particle pollution is the cycle that growth has done least to break. Of the 42 developing countries above 35 µg/m³ in 2000, 27 remain above it. In India annual exposure rose from 51 to 54 µg/m³, in Pakistan from 57 to 58, in Bangladesh from 65 to 69, in Nepal from 42 to 53 and in Niger from 58 to 65, according to the satellite-based estimates in the World Development Indicators. China is the main exception: after its 2013 air-pollution action plan, exposure fell from a peak above 50 µg/m³ to 32. Viet Nam fell from 38 to 25.

Two different problems share the cycle. In the Sahel and the Gulf, much of the burden is desert dust, which policy can reduce only at the margins (by restoring vegetation and controlling dust from construction and roads). In South Asia and much of urban Africa it is combustion: coal power and industry, brick kilns, vehicles, burning of crop stubble and waste, and household fuels, which also pollute outdoor air. Combustion is the part that clean energy, cleaner industry and waste management can tackle, and where the health gains are largest.

The economic case is strong. The World Bank valued the health damage from air pollution at $8.1 trillion, or 6.1% of global GDP, in 2019, and the Lancet Commission attributes more than 6.5 million deaths a year to air pollution. Because the same coal, kerosene and biomass drive both local pollution and climate change, cleaning the air also cuts emissions. Our Clean Cooking Transition and Pollution–Development Trap reports analyse the air cycle in more detail.

Adjusted net savings, sometimes called genuine savings, measures whether a country is adding to or subtracting from its total wealth. It starts with national saving, subtracts the depreciation of produced capital, the depletion of forests, minerals and energy, and the damage from air pollution, and adds spending on education. A negative figure means the country is getting poorer in total wealth even if GDP is growing. Averaged over 2015–2021, 23 countries were negative, including Timor-Leste, Burundi, Sierra Leone, Congo, Guinea, South Sudan, Zimbabwe, the DR Congo, Laos, Mozambique and Rwanda.

Across developing countries, lower adjusted net savings goes with higher extreme poverty (r = 0.31 at the same income). Median poverty in the drawdown countries is 36%, against 1% in the rest. The causes are familiar: dependence on extracted resources whose rents are consumed or captured rather than invested, weak tax bases, and debt service that crowds out investment. Forest depletion is a large component in several African economies. When the forest goes, so does the fuel, food and income that the poorest households drew from it.

The wealth cycle is the most policy-driven of the five, and many countries have escaped it. Sixteen of the 28 countries with negative adjusted net savings in 2000–05 turned positive by 2015–21, including Indonesia, Kenya, Côte d'Ivoire, Burkina Faso, Kazakhstan and Benin, through higher investment, better resource management or both. Fiscal rules that save part of resource rents, transparent concessions, community forest rights and investment in education and infrastructure are the escape route.

The cycle-count layer is the clearest picture of the trap: a band of three- and four-cycle countries across West and Central Africa and East Africa's Great Lakes, with Madagascar, Mozambique and Malawi to the south. Two-cycle countries ring them, and most of Asia and Latin America are trapped in at most one cycle, usually air in South Asia and land in parts of Latin America and China. The poverty layer almost overlays the cycle count.

The cooking-fuel and sanitation layers show the household cycles directly, nearly universal across sub-Saharan Africa outside the south and north. The forest layer shows change since 2000: loss across West Africa, the Sahel, the Horn, Madagascar and Central America, and recovery in China, Viet Nam, Rwanda, Burundi, Ghana and parts of Central Asia and the Caribbean. The air layer shows the dust and smoke belt from the Sahara to the Gangetic plain, and the wealth layer shows adjusted net savings.

Hover over or tap a country for its card. It shows which cycles the country is trapped in today, which it has escaped since 2000, and its cooking fuel, sanitation, open defecation, forest, air, land, savings and poverty figures for 2000 and the latest year. Selecting a country also loads it into the escape clock in section 9. The sortable table lists every country.

The escape clock projects a country's recent trend on four measured cycles: clean cooking, basic sanitation, fine-particle pollution and forest cover. The pace is the average annual change over the most recent ten years of data, fitted by least squares. The chart extends that pace to 2050 and shows the year the country would cross the trap threshold (half the population with clean cooking or basic sanitation, PM2.5 below 35 µg/m³) and the universal or health-based goal (95% access, or WHO's interim target of 15 µg/m³).

The acceleration slider multiplies the recent pace. The target-year slider shows the annual gain required to reach the universal goal by that year, and the benchmark shows the pace achieved by the fastest tenth of countries that were trapped in 2015. Indonesia added five to six points of clean cooking a year at the height of its LPG conversion, and India, Nepal and Cambodia each added about three points of basic sanitation coverage a year for two decades. A pace that seems impossibly fast for one country has often been achieved by another.

The projections are straight lines, not forecasts. Real transitions follow S-curves, slow at first, fast in the middle, slow near the end, and pollution trends swing with weather and dust. The clock does not apply to land and wealth, which lack consistent annual series. Its purpose is to show the gap between the current pace and the pace that escape requires.

Because the household cycles share a root, the most effective programmes attack several at once. Clean cooking with LPG, electricity or biogas cuts smoke, saves women's time and relieves pressure on forests. Sanitation campaigns that pair toilets with water supply and behaviour change cut disease and stunting. Secure land and tree tenure, as in Niger's farmer-managed regeneration and Nepal's community forests, lets households profit from restoring land rather than mining it. Each programme succeeds more easily when the others are under way. Electrification helps cooking, piped water helps sanitation, and healthier children learn more and later earn more.

The escapers point to the ingredients. They set national targets with deadlines and political ownership, such as India's 2019 open-defecation-free goal and Indonesia's kerosene-to-LPG conversion. They subsidised the entry point (the first stove, the first toilet, the first cylinder) and then the use. They measured outcomes, not outputs: use rather than toilets built, refills rather than stoves distributed, forest cover rather than seedlings planted. They also worked through local institutions, such as village sanitation committees and community forest user groups.

For the countries still in three or four cycles, mostly in Africa, money and stability are the binding constraints. Clean cooking, sanitation and land restoration all have high returns but need upfront investment that poor households and indebted governments cannot make alone. Results-based finance, concessional climate and development finance, and fiscal rules that turn forest and mineral rents into schools and infrastructure rather than consumption can break the wealth cycle that keeps the others in place. Clean air needs a different toolkit: clean power and industry, emission standards, waste and crop-residue management, and air-quality monitoring in the many African cities that still have none.

Coverage: 160 countries with at least one million people. Cycles and thresholds:

• Air: population-weighted annual mean PM2.5 of 35 µg/m³ or more (WDI, 2023; 2000).
• Cooking fuel: under 50% primarily relying on clean fuels (WHO via WDI, latest 2018–2024; high-income countries set to not trapped).
• Water and sanitation: under 50% with at least basic sanitation (WHO/UNICEF JMP via WDI).
• Land: 20% or more of land degraded (SDG 15.3.1, UN SDG database, 2019 or 2015; no 2000 baseline).
• Wealth: adjusted net savings including particulate emission damage below zero (WDI, 2015–2021 average; 2000–2005 for the baseline).

Poverty: World Bank PIP lineups at $3.00 a day (2021 PPP), 2000 and 2024. Forest change: % change in forest area (km²) from 2000 to the latest year. Open defecation, stunting, GDP per head (PPP 2021), cooking fuel and sanitation series 2000–2024: WDI.

Correlations are across low- and middle-income countries with data on both measures. Partial correlations use residuals from regressions on log GDP per head. Escape means trapped in 2000 and not trapped in the latest year, for air, fuel, sanitation and wealth. The escape clock fits a least-squares trend to the last ten years of data and extends it linearly.

Caveats: thresholds are conventions, and countries near them can switch class on small changes. PM2.5 estimates include desert dust. SDG 15.3.1 is self-reported with heterogeneous methods (Mexico reports 72% of land degraded), so the land cycle is the least reliable. Adjusted net savings depends on modelled depletion and damage values. Poverty figures after the latest survey are projections. None of the associations establishes causation.