The Green Development Dividend

A green development dividend is the set of returns that accrue when the same policy advances both ledgers: fewer people in poverty and fewer tonnes of carbon, or more resilience, than a conventional path would deliver. Measuring it means separating three things that are usually blurred together. The first is the goal, meaning the poverty line chosen and so the number of people who must cross it. In 2024 about 823 million people lived below $3.00 a day, 1.48 billion below $4.20 and 3.75 billion below $8.30, in the 157 countries covered here.

The second term is the route, which determines how much total income must rise for each dollar that reaches poor households. A perfectly targeted transfer moves one dollar to close one dollar of poverty gap. Distribution-neutral growth must raise everyone's income in proportion, so most of each new dollar goes to people already above the line. The third term is the energy system: how many kilograms of CO₂ are emitted for each dollar of new income. Across the countries here this ranges from about 0.03 kg per dollar in hydro-powered, low-consumption economies such as the DR Congo to 0.5 kg in coal-heavy South Africa and more than 0.7 kg in Mongolia.

Because the three terms multiply, a modest improvement in each compounds into a large difference in the total. The joint returns flow through four channels. Clean power is now often the cheapest source of new electricity. Cleaner combustion saves lives through cleaner air. Reforming fossil subsidies, which mostly benefit richer households, frees money for the poor. Adaptation and social protection keep climate shocks from undoing progress: the World Bank's Unbreakable report estimated that natural disasters push about 26 million people into poverty each year. The rest of this dashboard measures each term and channel.

The poverty gap measures how far, on average, people's incomes fall below the line. Multiplied by the line and the population, it gives the income transfer that would lift every poor person exactly to the line. In 2024 that came to about $312 billion a year at $3.00 (2021 PPP dollars), $811 billion at $4.20 and $4.9 trillion at $8.30. Spending that income at each country's 2024 carbon intensity of GDP would add 28 Mt of CO₂ a year at $3.00 (0.07% of the 38.6 Gt of fossil CO₂ emitted in 2024), 84 Mt at $4.20 (0.22%) and 720 Mt at $8.30 (1.9%). This is close to the lower bound found in the literature. Bruckner and colleagues estimated, with a global supply-chain model, that lifting more than a billion people above poverty lines would raise global emissions by 1.6–2.1% or less.

Poverty is rarely ended by transfers alone, however. If it is ended by distribution-neutral growth, every income in a country rises in proportion until only 3% of people remain below the line. The required growth then depends on how deep and how unequal poverty is. Nigeria would need average incomes to rise 2.9-fold, Kenya 4-fold and Mozambique almost 15-fold. At today's carbon intensities, that route would add 2.9 Gt of CO₂ a year at $3.00 (7.6% of global emissions), 8.3 Gt at $4.20 (21.5%) and 42 Gt at $8.30, more than today's global total. The gap between routes is a factor of about 100 for extreme poverty and about 60 for the $8.30 line.

These estimates bracket the World Bank's own. Its 2024 Poverty, Prosperity, and Planet Report found that eradicating extreme poverty would raise emissions by less than 5%, but that lifting everyone above $6.85 a day (in 2017 prices) would add nearly 50% to 2019 emissions by mid-century. The policy lesson is the same in both. Ending extreme poverty is not a climate problem. Ending poverty at higher lines through growth powered by today's energy mix would be one, and the remedy lies in the route (more inclusive growth and targeted transfers) and in the carbon intensity of growth, not in lowering the ambition for the poor.

A historical test of the trade-off is to ask how much carbon countries were emitting when they escaped mass poverty. Between 1991 and 2024, 37 countries brought the share of people below $3.00 a day under 10% for good. The median country did so at 1.9 tonnes of CO₂ per person, less than half today's world average of 4.7 tonnes, and 57% did so below 2 tonnes. Nepal crossed the line in 2014 at 0.27 tonnes, Sri Lanka in 2008 at 0.56, Bangladesh in 2020 at 0.58, Viet Nam in 2010 at 1.6 and India in 2018 at 1.9. China, whose growth ran on coal, crossed in 2013 at 7.2 tonnes, and Turkmenistan, Iran and Equatorial Guinea, all fossil-fuel producers, crossed at between 5 and 11.5 tonnes.

A second test looks at change rather than level. Of 57 countries where poverty at $4.20 a day fell by 20 percentage points or more between 2000 and 2024, 44 added less than one tonne of CO₂ per person over the period and 34 added less than half a tonne. Ethiopia's poverty rate at $4.20 fell from 84% to 57% while its emissions per person rose by 0.08 tonnes. Bangladesh's fell from 67% to 18% with a rise of 0.4 tonnes. At the other end, Mongolia added 9.8 tonnes per person and China 5.8, and in Ukraine, Uzbekistan and South Africa emissions per person fell while poverty fell.

Low emissions at the point of escape partly reflect low incomes. Escaping extreme poverty at $3.00 a day does not require a rich economy. But the spread in emissions at the point of escape, from 0.27 to 11.5 tonnes per person, a factor of more than 40, shows that the energy system and the structure of growth matter as much as income. Countries that escaped poverty through labour-intensive manufacturing, agriculture and services, powered partly by hydropower, emitted far less than those that grew through heavy industry and coal. The next generation of escapes, mostly in Africa, will happen when clean power is cheaper than ever.

Between 2015 and 2024, 49 of the 150 countries with complete data grew their economies while cutting fossil CO₂ in absolute terms. Another 53 decoupled relatively, with emissions growing more slowly than GDP, and 39 saw emissions grow as fast as or faster than GDP. Nine had shrinking economies. Absolute decoupling is concentrated at the top. It includes 36 of 49 high-income countries and 11 upper-middle-income ones, among them Brazil, Mexico and South Africa, but only two lower-middle-income countries and no low-income country.

That pattern is expected, and not a failure. Where energy use per person is a tenth of rich-country levels, development means using more energy, and the relevant test is relative: whether each dollar of growth carries less carbon than before. On that test, the record is mixed. High-income countries cut the carbon intensity of GDP by a median 3.5% a year. Low- and lower-middle-income countries held theirs roughly flat (median −0.04% a year). India (−1.5% a year), China (−3.0%), Bangladesh (−2.2%) and Ethiopia (−3.1%) cut intensity, while Viet Nam, Indonesia, the Philippines, Cambodia, Laos and Nepal raised it as industry, cement and coal power grew.

The countries where intensity rises fastest are mostly fast-growing Asian manufacturers, and they illustrate the central risk of the coming decades. Growth that ends poverty is being built on new fossil capacity that will operate for 30 to 40 years. In Nepal, emissions nearly tripled between 2015 and 2024 as oil and coal use grew, even though the country's electricity is almost entirely hydropower. The decoupling task in developing countries is less about cutting emissions now than about ensuring that the factories, power plants and vehicles built during the escape from poverty are the low-carbon ones.

Combining the two ledgers gives four kinds of developing country for 2015–2024. A country counts as making progress on poverty if the share below $4.20 a day fell by at least a quarter of a point a year or was already below 3%. It counts as making progress on climate if the carbon intensity of its GDP fell. Forty-four countries, with 4.2 billion people, made joint progress. They include India, China, Brazil, Bangladesh, Mexico, Türkiye, Thailand, Tanzania, South Africa, Colombia, Uganda, Morocco and Uzbekistan.

Twenty-two countries, with 859 million people, cut poverty while their carbon intensity rose. Indonesia, the Philippines, Viet Nam, Myanmar, Peru, Nepal and Cambodia are the largest, mostly economies where coal power, cement and manufacturing expanded fast. Sixteen countries, with 824 million people, cut carbon intensity without reducing poverty. Pakistan, Ethiopia, Egypt, Iran and Kenya lead this group, where poverty at $4.20 was flat or rising amid rapid population growth, debt stress or inflation. Nineteen countries, with 617 million people, made progress on neither. Nigeria, the DR Congo, Sudan, Afghanistan and Madagascar lead this group, which is dominated by conflict, macroeconomic crisis or both.

The classification is deliberately simple and uses only two indicators, both drawn from modelled data. It does not prove that any policy caused the joint gains. It does show that joint progress is the most common outcome among developing countries, not the exception, and that the countries missing out on both are mainly those where the binding constraint is instability, not a trade-off between climate and poverty. For them, the dividend begins with peace and macroeconomic stability.

The third term of the identity, the carbon intensity of new income, is set largely by the power system. Here the poorest countries hold an unexpected advantage. The median low-carbon share of electricity generation is 39% in low- and lower-middle-income countries, against 25% in upper-middle-income countries and 65% in high-income ones. Twenty-two low- and lower-middle-income countries generate 70% or more of their electricity from low-carbon sources, mostly hydropower. They include Ethiopia, Nepal, Uganda, the DR Congo, Malawi, Zambia, Mozambique, Kenya (91%, much of it geothermal) and Laos.

The catch is access. Many of these clean grids reach only a fraction of the population: 55% in Ethiopia, 51% in Uganda, 22% in the DR Congo and 16% in Malawi. The development task is therefore to extend a clean system, not to clean up a dirty one, and the cost of doing so has collapsed. IRENA found that 91% of utility-scale renewable capacity commissioned in 2024 produced power more cheaply than the cheapest new fossil-fuel alternative. Solar PV was on average 41% cheaper and onshore wind 53% cheaper, at global average costs of 4.3 and 3.4 US cents per kWh.

Since 2015, the low-carbon share of electricity has risen in 72 of 106 developing countries with data. Where it has fallen, as in Cambodia and Laos, new coal plants were built to meet fast-rising demand. The cost advantage of renewables is largest where capital is cheap, and capital is most expensive in exactly the countries with the largest access gaps. Lowering the cost of capital for clean power in Africa and South Asia is therefore one of the highest-yielding climate and poverty investments available.

The IMF estimates that explicit fossil-fuel subsidies, meaning fuel sold below supply cost, totalled about $725 billion in 2024, or 0.6% of world GDP. Implicit subsidies, the unpriced costs of air pollution and climate damage, added $6.7 trillion, bringing the total to about $7.4 trillion. Explicit subsidies are regressive. An earlier IMF review found that the richest fifth of households in low- and middle-income countries captured 43% of fuel subsidies, because they buy the most fuel.

Valued at market exchange rates, the transfers needed to close every extreme-poverty gap come to about $113 billion a year at $3.00 and $285 billion at $4.20, in 2021 US dollars across the 150 countries with price data. Explicit fossil subsidies are therefore about six times the extreme-poverty gap and two and a half times the gap at $4.20. Real transfer programmes cannot target perfectly and have administrative costs, and subsidies are concentrated in different countries from the poor. The comparison is not a budget plan, but it shows the scale of resources tied up in the opposite of a green development dividend.

Reform has worked where its proceeds were visibly shared. In 2015 Indonesia removed the gasoline subsidy and capped the diesel subsidy, saving IDR 211 trillion (about $15.6 billion, more than 10% of state spending). The budget increases that followed went to transfers to regions and villages, health insurance, low-income housing, clean water and infrastructure. The other half of the fiscal dividend is health. Air pollution caused 8.1 million deaths in 2021, more than 700,000 of them children under five, and cutting fossil and household combustion is the main way to reduce it.

The poverty layer shows extreme poverty concentrated in sub-Saharan Africa, with smaller pockets in Yemen, Afghanistan, Haiti, Papua New Guinea and parts of South and South-East Asia. The emissions layer is close to its mirror image. The highest emissions per person are in the Gulf, North America, Russia, Central Asia and Australia, and the lowest in the same African countries where poverty is deepest. The joint-progress layer shows the four classes from section 5, with high-income countries in grey.

The decoupling layer shows which countries grew while cutting emissions (absolute), grew emissions more slowly than GDP (relative) or grew emissions as fast as or faster than GDP (coupled) in 2015–2024. The clean-power layer maps the low-carbon share of electricity generation and highlights the hydro-rich grids of East and Southern Africa, the Andes, Nepal and Central Asia. The growth layer shows how many times average income would have to multiply to bring extreme poverty to 3% by distribution-neutral growth alone. Where the multiple is large, targeted transfers and inclusive growth are indispensable.

Hover over or tap a country for its card: poverty at the three lines since 2000, emissions per person since 1990 as a sparkline, carbon intensity and its trend, the decoupling and joint-progress classes, low-carbon power and access, the poverty gap and the carbon cost of each route. The sortable table below the map lists every country with the same measures.

The calculator applies the identity from section 1 to every country and adds up the results. The poverty line sets the goal. The route slider blends the two bounds from section 2. At 100% targeted, the whole poverty gap is closed by transfers that reach only the poor. At 0%, poverty falls to 3% through distribution-neutral growth. Real strategies lie in between, with growth doing most of the work and social protection reaching those it misses.

The carbon-intensity slider scales the carbon emitted per dollar of new income. At 100%, new income carries each country's 2024 intensity. At 50% it is half as carbon-intensive, roughly where intensity would be after two decades at the 3.5% a year decline achieved by high-income countries since 2015. At 10% it is close to fully clean power, transport and industry. The gap between the chosen setting and the high-carbon reference (growth at today's intensity) is the green development dividend in tonnes.

The outputs are static, annual and illustrative. They show the extra emissions once poverty has been ended under the chosen combination, not a path through time, and they ignore the emissions avoided by cleaner cooking, transport and power among the poor themselves. The transfer cost shows the annual cost of the targeted share at market prices. The scope menu limits the calculation to a region, an income group or a single country.

The first lever is the route. Social protection systems that can identify poor households and pay them reliably are the cheapest way to end extreme poverty, and they are also the lowest-carbon. Digital identification and payments, cash transfers indexed to prices, and employment programmes that build climate-resilient assets all reduce the income that must be generated to lift each person out of poverty. Inclusive growth works the same way. Growth that raises the incomes of farmers, informal workers and small firms lifts more people per dollar than growth captured by capital and the already well-off.

The second lever is the carbon intensity of growth. Clean power is now the cheapest new electricity in most markets, and many of the poorest countries already have low-carbon grids that need extending rather than replacing. The highest returns come from lowering the cost of capital for clean generation, grids and mini-grids in Africa and South Asia. They also come from avoiding new coal in the fast-growing Asian economies whose intensity is rising, and from clean cooking, electric two- and three-wheelers and efficient appliances that cut household energy costs and pollution together.

The third lever is fiscal. Explicit fossil subsidies of about $725 billion a year, mostly captured by richer households, exceed the cost of closing the extreme-poverty gap several times over. Reform succeeds when compensation comes first and is visible, as with Indonesia's 2015 reallocation to regional transfers, health insurance and infrastructure. Carbon-pricing revenue can be used the same way. The fourth lever is resilience. Adaptation investment and adaptive social protection that scales up after droughts, floods and heatwaves keep climate shocks from pushing tens of millions back into poverty each year, and they protect the dividend once it is earned.

Coverage: 157 countries with at least 1 million people and both poverty and emissions data (Argentina, which has only urban poverty data, is excluded). Poverty: World Bank Poverty and Inequality Platform lineup estimates (fill-gaps), headcount and poverty gap at $3.00, $4.20 and $8.30 a day in 2021 PPP, 1990–2026. Values after the latest survey are projections or nowcasts. Emissions: fossil and industrial CO₂ (Global Carbon Budget via Our World in Data), excluding land use, 1990–2024. GDP: WDI GDP in constant 2021 PPP dollars. Low-carbon share of electricity: Ember and Energy Institute via Our World in Data. Access to electricity: WDI. Market-price conversion: WDI PPP conversion factor for private consumption divided by the official exchange rate, 2021.

Targeted route: poverty gap × line × 365 × population, times the country's 2024 CO₂ per PPP dollar of GDP. Growth route: a lognormal income distribution with dispersion from the PIP mean and median, shifted to match the actual headcount at each line and then scaled up until 3% remain poor. The growth factor times 2024 CO₂ gives the added emissions at constant intensity. Escape year: first year from which the $3.00 headcount stays below 10% (or 3%) through 2024, having been above it in 1990. Decoupling (2015–2024): absolute if GDP rose and CO₂ fell, relative if CO₂ grew more slowly than GDP, coupled otherwise. Joint progress (developing countries): poverty at $4.20 falling by at least 0.25 points a year or below 3%, and carbon intensity of GDP falling.

Caveats: the carbon-cost estimates are static and hold intensity constant. Poor households' spending need not carry their country's average intensity, the lognormal approximation is rough in the tails, and poverty projections may be revised. Territorial emissions differ from consumption-based footprints. The classes depend on thresholds and on modelled data, and none of the associations is causal. Subsidy and poverty-gap comparisons differ in year, coverage and price basis. Literature figures are quoted as published.