The Human Dividend of Polycrisis Prevention

The two largest channels are almost equal. Air pollution, outdoors and in the home, caused about 6.6 million deaths in 2021 according to WHO's joint-effects estimates, 5.5 million of them from heart disease, stroke, lung cancer and chronic lung disease, mostly in adults. About 4.9 million children died before their fifth birthday in 2024, mainly from newborn complications, pneumonia, diarrhoea and malaria. Unsafe water, sanitation and hygiene caused about 1.0 million diarrhoeal deaths in 2019, 760,000 of them at ages five and over. Some 250,000 women died from pregnancy and childbirth in 2023.

The shock channels are smaller in a typical year but far more volatile. Organised violence killed an average of 231,000 people a year in 2021–2025, more than twice the 2015–2019 average, driven by the wars in Ukraine, Ethiopia, Sudan and Gaza. Disasters killed an average of 38,000 a year in 2015–2024, with peaks in the 2023 Türkiye–Syria earthquake and European heatwaves. Malaria killed about 610,000 people in 2024, mostly young African children.

Rates of preventable death fall steeply with income, from a median of about 285 per 100,000 people a year in low-income countries to 33 in high-income ones, but the spread at any given income is wide. That spread is the first source of the human dividend: countries at the same income level that manage risk better already show what is possible. The scatter shows each country's ledger rate against income; countries well below the trend at their income, such as Malawi, Burundi, Nicaragua, El Salvador, Jordan and Rwanda, are the existence proofs.

The best evidence that prevention works is how much it has already achieved. If each country's under-five mortality rate had stayed at its 2000 level, about 105 million more children would have died between 2001 and 2024, given the births that actually occurred. The annual dividend has grown every year, reaching 6.9 million in 2024, because rates kept falling and births shifted toward countries where mortality had been highest. Falling maternal mortality protected about 3.4 million mothers over 2001–2023.

Malaria shows how targeted prevention adds up. Applying each country's 2000 malaria death rate to its later population gives 13.9 million deaths averted between 2001 and 2024, almost exactly the 14 million that WHO's World Malaria Report 2025 attributes to malaria control since 2000, with about 1 million in 2024 alone. Bed nets, rapid tests, artemisinin therapies and seasonal chemoprevention did this work. Most of these lives are children, so they sit inside the under-five dividend, not on top of it.

Disaster deaths tell the same story for shocks. EM-DAT recorded about 38 deaths per million people a year in the 1970s and under 7 in the 2010s, even as more people came to live in exposed places. Bangladesh is the emblem. The 1970 Bhola cyclone killed 300,000 to 500,000 people and the 1991 cyclone about 138,000; Cyclone Sidr in 2007 killed a few thousand, and Amphan in 2020 killed 26 in Bangladesh, after decades of investment in warnings, volunteers and shelters.

The levers work differently by channel. For conflict, the best quarter of countries at every income level has no conflict deaths, so risk reduction, meaning peace, could in principle protect all 231,000 lives a year. For disasters, risk reduction is almost everything: matching peers would avoid 93% of deaths, but development alone would avoid only 8%, because high-income countries have higher median disaster death rates than middle-income ones, largely from heatwaves. Richer is not automatically safer; it has to be built.

Air pollution is the opposite case. Matching the best quarter at the same income would avoid only 29% of air-pollution deaths, because countries at a similar stage of industrialisation tend to burn similar fuels. Moving to the next income group's typical rate would avoid 58%, and coordination 70%. The lever that matters is a different development path, one that skips the dirtiest stage through clean power, clean cooking and transport electrification. China's large potential dividend, about 1.5 million lives a year, is almost entirely in this channel.

For children and mothers, both levers matter about equally. Matching peers would avoid 42% of under-five deaths and 59% of maternal deaths, and coordination 74% and 86%. Unsafe water and sanitation deaths are among the most avoidable at every income level: 77% with risk reduction alone and 94% with coordination. By income group, lower-middle-income countries hold the largest dividend, 4.4 million lives a year, and low-income countries the highest rate, about 177 lives per 100,000 people a year.

India could protect about 1.7 million lives a year under coordination: 860,000 from air pollution, 420,000 children under five and 400,000 from unsafe water and sanitation. China's 1.5 million is almost entirely air pollution. Nigeria's 970,000 is dominated by child survival: about 790,000 under-five deaths a year could be prevented, and its gap to income peers alone accounts for 810,000 lives, the largest risk-reduction dividend of any country. Pakistan (510,000) and the DR Congo (320,000) follow.

The mixes call for different packages. In Nigeria, Pakistan and the DR Congo, primary health care, vaccination, nutrition, water and malaria control come first. In India and China, the decisive lever is energy: clean power, clean cooking and cleaner transport. Ethiopia shows the shock channel: about 61,000 conflict deaths a year in 2021–2025 sit alongside 72,000 avoidable child deaths. Ukraine, Sudan and Gaza carry the other large conflict tolls.

Some countries are already at the frontier. Rwanda's death rates in every channel are close to the best quarter of low-income countries, so its risk-reduction gap is almost zero, and its remaining dividend comes from development. Malawi, Nicaragua and Jordan sit well below the ledger trend for their income. Their records show that the benchmarks used here are achievable, not aspirational.

If preventable deaths were scattered, each channel could be handled by its own programme. They are not. Of the 160 countries, 55 do worse than the median of their income peers in at least four of the six channels. They hold a third of the people covered (2.7 billion) but 42% of the coordinated dividend. Seven, among them Nigeria, the DR Congo, Indonesia and the Philippines, fall short in all six. In these countries, siloed programmes compete for the same scarce clinics, budgets and administrators. Coordinated packages can share them.

The burdens also share drivers. The matrix shows partial correlations between each driver and each channel's death rate across countries, after removing what income per head explains. Universal health coverage is associated with lower death rates in five of seven channels, and basic sanitation in four. Government effectiveness is associated with fewer air-pollution, conflict and child deaths. Child stunting is associated with higher child, maternal and water-related deaths. Clean cooking is associated with fewer malaria, water-related and maternal deaths, but only weakly with air-pollution deaths once income is controlled for, because outdoor pollution also matters.

No driver predicts disaster deaths well at equal income, which fits the evidence that disaster mortality depends on specific preparedness: warnings, shelters and building codes. These correlations are not causal. Richer countries within an income group have more of everything, and drivers move together. But they point to the investments with the widest reach: health coverage, sanitation, nutrition and capable institutions do double or triple duty, so they belong at the centre of a coordinated agenda.

Conflict is the channel moving the wrong way. UCDP's event data record about 106,000 deaths a year in organised violence in 2015–2019, rising to 318,000 in 2022 and 245,000 in 2025. Wars also drive the other channels: they destroy clinics and water systems, displace millions and cut off food, so their indirect deaths exceed the direct toll. In this ledger, peace is the purest form of risk reduction: the best quarter of countries at every income level records no conflict deaths.

Disasters show what deliberate preparation can do. Bangladesh's Cyclone Preparedness Programme, started after the 1970 cyclone, now relies on tens of thousands of volunteers who spread warnings, along with thousands of cyclone shelters. UNDRR and WMO found in 2022 that countries with limited early-warning coverage have disaster mortality about eight times higher than countries with substantial coverage. Yet about half of countries still lack multi-hazard warning systems, and the gaps are largest in the least-developed countries and small islands.

This is why disaster deaths are the channel where risk reduction is nearly the whole answer. Development alone would protect 8% of disaster deaths in this model, but matching the best quarter of income peers would protect 93%. Heat is the growing exception. EM-DAT records heat deaths mainly where they are counted, as in Europe in 2022 and 2024, and largely misses them elsewhere, so the true disaster toll and dividend are likely higher than shown.

Projecting the ledger forward with the UN's 2024 population and birth projections, deaths at today's rates would total about 330 million between 2025 and 2050, as populations grow fastest where rates are highest. If every country closed its coordinated gap in a straight line by 2040 and held it, about 172 million of those deaths would be avoided. Risk reduction alone would avoid about 94 million and development alone about 139 million.

The baseline is a counterfactual, not a forecast. Some of the gap would close under existing trends: under-five mortality has been falling by a median of about 2.6% a year since 2015. But progress has slowed since 2015, conflict deaths have doubled, and aid budgets have been cut. A study in The Lancet by Cavalcanti and colleagues estimates that programmes supported by USAID prevented about 91 million deaths between 2001 and 2021, and that the 2025 cuts could cause more than 14 million additional deaths by 2030, including 4.5 million children under five, if not reversed.

The projection treats air-pollution deaths as growing with population and ignores ageing, which will raise the number of older people vulnerable to heart and lung disease. The true baseline is probably larger, and so is the dividend. Climate change will add heat, flood and disease risks that the 2015–2024 benchmarks do not capture. Coordinated prevention is how countries keep the dividend they have earned and claim the rest.

The dividend layer shows lives a year that coordinated prevention could protect per 100,000 people. It is highest in Nigeria, Somalia, Niger, the Central African Republic, Guinea and Chad, and in the war zones of Ukraine and Gaza. The share layer shows how much of each country's ledger that represents, which is high almost everywhere outside the rich world.

The channel layers show under-five mortality, age-standardised air-pollution deaths, water and sanitation deaths and conflict deaths. Age-standardised air-pollution death rates peak where households still burn solid fuels, across West and Central Africa, Afghanistan, Haiti and the Philippines, while conflict deaths are concentrated in a handful of countries. The burdens layer counts the channels in which each country is worse than its income peers and picks out the multi-burden countries of section 5.

Hover over or tap a country for its card: its ledger by channel, the lives protectable under each lever, the cumulative dividend to 2050 and its main drivers. Clicking a country loads it into the calculator in section 9. The sortable table below lists all 160 countries.

The calculator starts from the country's ledger in each channel. The risk-reduction slider sets how much of the gap to the best quarter of its income peers is closed. The development slider sets how far the country moves toward the typical rates of the next income group. The two combine multiplicatively, as in section 3, so closing both fully gives the coordinated estimate.

The outputs show lives protected a year at today's population, the share of the ledger that represents, the channel that contributes most, and the cumulative lives protected over 2025–2050 if the chosen gaps close in a straight line by 2040 and the country's population follows the UN's medium projection. The chart compares each channel's ledger with the lives protected.

Try a country at the frontier, such as Rwanda, where only the development slider moves the result, against Nigeria, where risk reduction alone protects most of the dividend, and China, where development matters most because the remaining gap is air pollution. These are benchmarks of what peers achieve, not predictions of what a given policy will deliver.

Start with the investments that reach the most channels. Primary health care that delivers vaccines, skilled birth attendance, malaria treatment and nutrition protects children and mothers and makes populations more resilient to every shock. Water and sanitation cut diarrhoeal deaths at every age and reduce the toll of floods and displacement. Social protection that reaches poor households quickly prevents hunger turning into child deaths when prices or harvests fail. These are the drivers that correlate with lower deaths in several channels at once.

Make the development path a clean one. Air pollution is the largest channel and the one where matching today's peers achieves least. Clean power, clean cooking and electrified transport let countries grow without passing through the dirtiest stage, which is where most of the 3.8 million air-pollution lives in the coordinated dividend lie. Build preparedness deliberately, because development alone does not lower disaster deaths: early warnings for all, shelters, building codes and heat action plans.

Protect the dividend already earned. Conflict prevention and humanitarian access are life-saving investments in their own right, and aid cuts reverse gains that took decades to build. Finance should follow the dividend: low- and lower-middle-income countries hold two-thirds of it, and 55 multi-burden countries hold 42%. Measuring progress as lives protected across channels, not outputs within silos, is how governments and donors can see what coordination is worth.

Coverage: 160 countries with at least one million people and a World Bank income group. Ledger: under-five deaths 2024 (SH.DTH.MORT) and maternal deaths 2023 (SH.MMR.DTHS), UN IGME and MMEIG via WDI; air-pollution deaths 2021, WHO joint ambient and household effects (GHO AIR_35) minus lower respiratory infections; WASH-attributable diarrhoea deaths 2019 (GHO WSH_10) minus under-fives; conflict deaths, mean of 2021–2025, UCDP GED 26.1 (state-based, non-state and one-sided, with Gaza and Ukraine attributed by event location); disaster deaths, mean of 2015–2024, EM-DAT via Our World in Data. Malaria deaths 2024 (GHO) are shown but excluded from totals.

Benchmarking rates: under-five mortality per 1,000 births, maternal mortality per 100,000 births, age-standardised air-pollution death rate (GHO SDGAIRBODA), and crude rates per 100,000 people for the other channels. For each channel and income group, the risk gap is 1 − (group 25th percentile ÷ country rate) where positive; the development gap is 1 − (next group's median ÷ own group's median), which is zero for high-income countries. Coordinated = 1 − (1 − risk gap)(1 − development gap). Earned dividend = births × (2000 rate − actual rate) for under-five and maternal, and population × change in rate for malaria. Projections scale each channel by UN WPP 2024 medium-variant births or population, with gaps closed linearly between 2025 and 2040.

Drivers: clean cooking, basic water, basic sanitation and electricity access (WDI), the WHO UHC service coverage index, government effectiveness (WGI) and child stunting, latest year. Partial correlations control for log GDP per head (PPP). Caveats: estimates come from different years and models; benchmarks describe what peers achieve, not causal effects of policy; peer percentiles of zero (conflict, malaria) imply full avoidance; disaster data undercount heat deaths; projections ignore ageing and climate change; residual overlaps (for example, conflict deaths among children) are small.