Clean Air Has No Vaccine

Air pollution has a funding problem. We know it's harmful and takes over 8 million lives each year, which is higher than the death toll of HIV, malaria and tuberculosis combined. Yet, the funding for improving air quality remains limited, despite there being enough philanthropic funding to invest into this issue.

So what is the funding problem? In our opinion, this is related to how air pollution and clean air interventions are evaluated by funders. In our field, it is difficult to tell a funder with confidence: “The USD 1m you gave, saved 10.000 people’s lives.” For some areas of global health, that link is much clearer.

A vaccine is given to a specific person, against a specific disease, with decades of evidence connecting the intervention to the outcome. We don't need to prove that every individual vaccination prevented a death.

Intervention (e.g. Vaccine) → lower mortality → lives saved

Air pollution is different.

Clean-air intervention's success is dependent on changes in environments and systems. You might distribute face masks, improve building ventilation, install air filters, reduce burning practices, change industrial emission policies, or build a monitoring network that helps a community understand and protect themselves from what they are breathing.

The chain is longer:

Direct or indirect intervention → cleaner air → lower exposure → lower disease risk → better health

And every step becomes harder to attribute to a clear cause and effect relationship. This doesn't mean the impact is smaller. It only means the impact is harder to measure.

When Impact Doesn't Come With a Receipt

We’ve seen this in Kenya with one of our long-term partners, the Demography Project, led by Richard Muraya. Starting as a project focused on freshwater resource advocacy, the citizen group in Kihumo village quickly noticed that the nearby stone quarries, construction in the village and use of biomass fuel also impact the quality of the air. Local clinic records showed high amounts of respiratory issues. Soon, they began distributing protective face masks to construction workers and children, a subset of the population who are most exposed to ambient air pollution. In addition, they encouraged families to ventilate their homes, since there are limited alternatives to biomass cooking fuel. Lastly, in order to evaluate whether the air was indeed getting cleaner inside homes, and to know when to encourage wearing face masks among all residents, Richard installed air quality monitors. These are now particularly helpful for his future plans; petitioning the nearby quarries and road construction companies to reduce excessive particulate matter emissions.

Clearly, clean air interventions can have direct positive health outcomes. Similar to the distribution of malaria nets, face masks protect people from excessive exposure to particulate matter, reducing their chance of developing respiratory issues. At the same time, education about healthy practices to reduce air pollution exposure, such as ventilation, using cleaner fuel sources, or avoiding agricultural burning are arguably comparable to the education of safe sex practices that prevent HIV and AIDS.

These two direct health interventions show that air quality has a right to be considered a prospective field for effective altruism. Most importantly, monitoring the current air quality is an indispensable tool in these interventions, enabling when, where and who these interventions target.

But can we measure the deaths that these projects prevent? We cannot responsibly give a precise number in the same way as a vaccine project could. But surely that doesn't make identifying a serious pollution problem, distributing protective equipment, educating an affected population, and installing cleaner technology a bad public-health investment.

It just shows the limits of attribution, the limits of clear, easily quantifiable cause and effect relationships.

Evidence-based philanthropy is a good thing.

We should measure what we can, challenge our assumptions and stop funding things that don't work. But we should not confuse what is easy to measure with what is important.

Air pollution is very complex. Exposure happens across homes, schools, workplaces and entire cities. Health effects accumulate over years. The benefits of cleaner air can be spread across thousands or millions of people. But there may never be a neat causal receipt saying: Your $100,000 saved exactly 212 lives.

This uncertainty, I believe, makes clean-air projects less attractive to fund.

Here are some specific examples. GiveWell explicitly models interventions such as insecticide-treated malaria nets and preventive malaria medication in terms of expected lives saved and cost-effectiveness; it recommended a $41 million grant for mosquito nets in the DRC partly because it estimated the program to be around 11 times as cost-effective as unconditional cash transfers. The Gates Foundation has similarly committed enormous sums to vaccines and infectious-disease prevention, including $1.6 billion to Gavi over five years, where established evidence allows vaccination coverage to be translated into expected disease and mortality reductions.

These are extremely valuable investments, but they illustrate the challenge for clean air: when philanthropic decision-making rewards interventions with a relatively short and quantifiable path from dollars spent to health outcomes, systemic environmental interventions can be disadvantaged simply because their causal chain is longer and harder to measure. But it does not necessarily make them less impactful.

Globally, air pollution currently kills more people each year than routine vaccination saves in an average year, but the comparison needs careful framing because one is an annual death burden and the other is a modeled counterfactual benefit accumulated over decades.

WHO-led research published in The Lancet estimates that vaccination against 14 major diseases averted about 154 million deaths from 1974–2024. That works out to an average of roughly 3.1 million lives saved per year over those 50 years. About 101 million of those saved lives were infants, and measles vaccination was responsible for the largest share.

For air pollution, WHO’s global data portal estimates about 6.6 million deaths in 2021 from the combined effects of ambient and household air pollution. WHO commonly describes the burden as roughly 6.7–7 million premature deaths per year. A different major assessment, the State of Global Air, estimated 8.1 million air-pollution-attributable deaths in 2021, reflecting differences in methodology and risk factors included.

We should also be careful not to overstate our own case. Air quality impacts can be calculated, and some of the numbers are remarkably good. When US embassies installed PM2.5 monitors and published the readings in cities that previously had little or no public data, annual PM2.5 levels in those cities fell by more than 10 µg/m³ over six years - enough, if sustained, to add roughly a year of life expectancy for everyone breathing that air.

Energy Policy Institute at the University of Chicago (EPIC) 2025 Clean Air Investment Update goes further, identifying 83 countries where $50,000 to $100,000 a year routed to local organisations could plausibly unlock national-level clean air impacts, in places that are home to 2.9 billion people and 4.9 billion life years lost to particulate pollution.

These are not vague figures; they sit comfortably next to the cost-per-life-saved estimates GiveWell publishes. The point is not that every clean air project can be modelled this precisely. It is that the field's supposed unmeasurability is partly a funding artefact. Without monitors there is no data, without data there is no evidence base, without an evidence base there is no case to make, and so the money goes elsewhere and the data gap stays open.

Why We Focus on Open Public Infrastructure

This problem has shaped how we think about the Open Air Foundation.

We focus on building open public infrastructure for airborne environmental health: open monitoring tools, open-source hardware and software, shared data, practical knowledge, local capacity and the systems that help turn measurements into action.

These are not always easy things to fund.

A shared dataset doesn't come with a number of lives saved. Neither does an open-source monitoring platform, a calibration method or the knowledge gained when a project discovers that an intervention doesn't work.

But these are the foundations on which better interventions are built.

So, Does PM2.5 Need a Vaccine?

Probably not.

A vaccine is almost the perfect intervention for evidence-based funding: a defined intervention, a defined recipient, a defined disease and a strong causal relationship.

Clean air is almost the opposite. It is environmental, distributed and systemic.

We should absolutely keep improving the evidence around what works. But we should not force every clean-air project to behave like a vaccine before we consider it worth funding.

Some of the most important work may never produce a perfectly attributable number of lives saved. That does not make it less impactful. It makes it harder to measure.

And that is exactly why philanthropy has an important role to play.

Not only in funding interventions with the cleanest causal receipts, but also in building the open infrastructure that makes better decisions, better evidence and better protection possible.

Clean air may never look like a vaccine.

It should not have to in order to deserve the same ambition from global health philanthropy.