In 2013, nine-year-old London resident Ella Adoo-Kissi-Debrah died after a severe asthma attack. Having suffered from a long history of respiratory illness, a public inquest ruled that air pollution exposure was a major contributing factor to her death.

This unprecedented announcement marked the first time that air pollution was explicitly listed as a cause of death, prompting Ella’s mother to take highly publicized legal action against the British government and to start a foundation for spreading awareness about the overlooked threats posed to children by toxic air.

Why weren’t existing regulations enough to protect people like Ella? One reason is that environmental regulation often gives short shrift to a relatively small number of places that bear the heaviest load of pollutants – typically places where poor people and victims of racial discrimination live.

Indeed, questions of environmental justice have only recently become high priorities. Ella’s case galvanized an ongoing cross-party initiative in the UK Parliament to establish clean air as a human right in what would be a groundbreaking acknowledgement of distributional environmental inequities.

But there is another widely overlooked reason why air pollution is more dangerous than generally understood. Governments continue to regulate air quality as though each contaminant exists in isolation, when in reality people breathe in a complex cloud of pollutants. And mounting evidence shows that assaying the damage caused by individual pollutants understates public health risk because many chemicals interact synergistically. So single-substance rules can miss damage only evident when exposures are combined.

To address the mismatch between law and lived experience, we propose a shift toward a multipollutant cap-and-trade regime that treats air pollution as a cumulative public health burden and aims to reduce it more holistically.

Limits of One Pollutant at a Time

In the 16th century, the Swiss physician Paracelsus famously declared, “The dose makes the poison.” And that principle has provided a powerful framework for assessing risk that informed generations of environmental and public health regulation.

By the 1970s, the proliferation of industrial chemicals and urban air contaminants led governments to codify the historic dose principle in law. In the United States, the separate National Ambient Air Quality Standards for each “criteria pollutant” – small particulate matter, ground-level ozone, nitrogen dioxide (NO₂), sulfur dioxide (SO₂), carbon monoxide (CO) and lead – include maximum permissible load for each enforced independently.

These pollutant-specific limits were intended to simplify compliance and enforcement, but they had the drawback of cementing fragmented regulation. Regulators could declare “attainment” for ozone, for example, even if particulate matter or nitrogen dioxide levels remained dangerously high in the same community.

Other statutes carved out responsibility for distinct chemical categories such as the 1947 Federal Insecticide, Fungicide, and Rodenticide Act for pesticides, the 1976 Toxic Substances Control Act for industrial chemicals, the 1976 Resource Conservation and Recovery Act for hazardous waste and the 1974 Safe Drinking Water Act for waterborne contaminants. Each of these frameworks evolved in isolation, overseen by different offices within the EPA or other federal agencies, creating what experts now call “regulatory silos.”

The patchwork allows little institutional space for addressing how pollutants interact in real-world environments. For instance, communities near major highways are simultaneously exposed to fine particulates, nitrogen oxides, and volatile organic compounds from vehicle exhaust, with each still judged against its own threshold without regard for their combined cardiovascular or respiratory effects.

In water systems, a community’s supply might test below the legal limits for arsenic, nitrates and per- and polyfluoroalkyl substances (PFAS) individually, but the toxic load of these contaminants together can still pose significant risks to human health.

Occasional early warnings about gaps in public health coverage did emerge. In the 1960s, Rachel Carson’s Silent Spring raised alarms about pesticide mixing, but such concerns were sidelined for decades. Technical challenges made testing every possible chemical combination impractical, and institutional inertia favored the simplicity of one-pollutant rules.

Market-based environmental policies that effectively put a price on pollution, while often more flexible and cost-effective than prescriptive regulation, have so far done little to address these cumulative risks. Current emissions trading systems that incentivize reductions, such as the federal SO₂ trading program, regional NO₂ trading markets and global CO₂ cap-and-trade schemes, are designed for single pollutants. They often succeed in lowering the targeted pollutant but can inadvertently worsen conditions for vulnerable communities if other co-pollutants remain uncontrolled. For example, a power plant might reduce its CO₂ output through fuel-switching but increase local pollution of particulate matter or toxic metals.

Regulators in Europe and the United States have begun exploring ways to integrate cumulative risk considerations into environmental policy. Yet progress remains slow. And closing the gaps left by fragmented and single-pollutant rules requires more than administrative reform. It demands a reckoning with the science itself. Below, we dig a bit into these issues, offering evidence that pollutants combine in real-world settings to produce harms far greater than the sum of their parts – and why this makes it impossible for one-pollutant safety limits to ever fully protect communities.

The Science of Mixtures and

Why It's Time to Rethink the Model

Modern toxicology is moving decisively beyond the old assumption that chemicals act independently. When substances co-occur, they can combine their effects additively – each contributing to a growing burden on the same biological pathway – or interact synergistically so that the combined harm exceeds the collective individual total.

A useful analogy is medicine: two mild drugs taken together can produce side effects far worse than either alone. The same logic applies to tiny amounts of disparate pollutants that converge on the same hormone system or cellular 9oprocess. For example, laboratory reconstructions of maternal blood composition have shown that low-dose mixtures can disrupt thyroid signaling in developing neural tissue – outcomes that no single chemical’s profile would have predicted. Epidemiological cohorts increasingly tie routine, low-level mixtures to higher risks of certain cancers, endocrine dysfunction and neurodevelopmental impairments even though each constituent sits below its regulatory limit.

The technical problem of getting a handle on these interactions is daunting, with thousands of chemicals and millions of possible blends to consider. But it is no longer intractable: high-throughput bioassays (laboratory tests that expose living cells or tissues to chemicals to measure biological responses such as toxicity, hormone disruption or DNA damage), mass-spectrometry imaging (which maps chemicals inside tissues) and computational models now let scientists triage which combinations are most likely to be hazardous and therefore worthy of regulatory attention. Indeed, policymakers can now see that additive and synergistic effects are common enough to change how we set standards.

Drinking water illustrates the gap between law and reality. Utilities routinely meet contaminant- by-contaminant standards while delivering blends of arsenic, nitrates, pesticides, disinfectant byproducts and PFAS to households every day. The 2014 Flint, Michigan, water crisis painfully demonstrated how lead exposure in conjunction with corrosive treatment chemistry and disinfectant byproducts produced health outcomes no single standard predicted.

For regulators, the lessons are strong: drinking water rules must require cumulative-risk assessment, utilities must model and report combined exposures, and remediation funding and legal-technical support should be prioritized for overburdened watersheds and communities.

The stakes are highest for the very young and very old. Fetuses and infants are exposed to maternal chemical mixtures – phthalates, bisphenols, flame retardants and numerous other compounds – that readily cross the placenta and act on precisely timed development pathways. Laboratory reconstructions of real-world maternal mixes have disrupted thyroid-mediated neurodevelopment at concentrations commonly found in pregnant women, and longitudinal human studies link higher combined prenatal loads to lower IQ, increased ADHD risk and other persistent deficits.

With the elderly, multipollutant exposure further increases already elevated base risks of respiratory and cardiovascular issues in addition to geriatric neurological conditions like Alzheimer’s disease. Because these effects accumulate during narrow windows, regulating chemicals one by one fails to protect the most vulnerable: a dozen individually “safe” exposures together can alter developmental trajectories in irreversible ways. It thus makes sense to require mixture-based testing for chemicals that cross the placenta and to build on our knowledge with long-term birth-cohort studies that measure combined exposures and life-course outcomes.

Urban air is the paradigmatic cocktail. A single breath in a city can deliver fine particulates laced with metals and organics, nitrogen oxides from traffic, ozone produced in sunlight, sulfur compounds from industry and a cloud of volatile organic compounds. Epidemiology shows that days with elevated multipollutant burdens produce larger spikes in asthma attacks, heart attacks and hospital visits than single-pollutant metrics predict. Research has shown that particulates act as carriers, ferrying adsorbed toxins deep into the lung and into systemic circulation.

In practice, pivoting means embedding the assessment of mixtures into standard-setting:

• Introducing conservative mixtures assessment factors where co-exposure is likely
• Redesigning economic tools so pollution externality taxes and tradable allowances reflect cumulative toxicity rather than single-pollutant volume
• Creating multipollutant cap-and-trade systems that reward reductions in broadly defined toxic portfolios

It also means upgrading monitoring systems, so air, water, product and biomonitoring data are interoperable and discoverable across agencies that require manufacturers and registrants to submit mixture-toxicity modeling or testing when products are authorized – not to mention reforming statutes and interagency cooperation to permit cross-program action. The next step is to consider how these scientific insights can be translated into a market design that creates efficient incentives to cut the entire portfolio of emissions.

The Case for Multipollutant Cap-and-Trade

How do we bend market incentives so firms reduce the mix of pollutants that do the most harm? A multipollutant cap-and-trade system answers that question by converting the single-permit market into a bundled-permit system that manages firms’ pollutant portfolios rather than regulating isolated emissions.

Before going into the details, it helps to recall the logic of conventional cap-and-trade so the contrast is clear. A traditional cap-and-trade program sets a hard limit – an overall cap – on emissions of a single pollutant, then distributes allowances up to that cap, and lets firms buy and sell those allowances so the pollution control is undertaken by emitters who can manage the process at least cost. The multipollutant variant, by contrast, bundles rights across several pollutants into a single permit and is explicitly grounded in mixture toxicology and health-impact evidence.

Regulators set an aggregate cap informed by those scientific and public health priorities, issue a fixed number of bundled permits and allow firms to trade those multipollutant permits. A company that reduces its portfolio of emissions below its allotment can sell unused permits to other firms or perhaps sell them back to the regulator. When the regulator permanently cancels a returned multipollutant permit, the supply shrinks and the overall cap tightens.

Because each permit covers multiple pollutants and prices embed health-weighted values, the market rewards investments that reduce several harmful emissions at once. An important design point here is firm heterogeneity: firms differ dramatically in the pollutants they emit. A coal-fired power plant typically releases SO₂, NO₂ and PM2.5 (ultra-small particulates), while a metal smelter emits toxic metals and PM2.5, and a chemical plant releases volatile organics. So a one-size-fits-all bundled permit would not be efficient in achieving public health goals informed by the science of chemical mixture.

Instead, regulators should design several types of bundled permits that reflect the market’s mix of firm profiles and dominant emissions portfolios. A heterogeneity-aware permit structure allows a matching mechanism to allocate bundles that fit firms’ technical capabilities and abatement costs, improving allocative efficiency and nudging capital toward holistic abatement technologies that cut the chemical mixtures that matter most for health.

Yes, this isn’t simple, but practical building blocks already exist. Integrated multipollutant management programs and multipollutant monitoring networks offer the data backbone needed for designing bundled markets. Additionally, a Resources for the Future policy brief lays out concrete schematics for institutional relationships and permit structures that are directly usable in pilot settings.

Several factors make the design especially attractive. First, it preserves market flexibility: firms retain the ability to choose the lowestcost path to compliance. But now that choice must account for the health impacts of the whole emissions mix. Second, it channels investment toward holistic abatement technologies – filters and process changes that cut particulates, toxic metals and volatile organics together – because such technologies earn greater returns in a market that values pollutant portfolios. Third, the approach can be tailored to equity goals: permits can be weighted or phased so that reductions concentrate in neighborhoods with the highest cumulative exposures rather than only where abatement is least expensive.

There are positive dynamic effects at work here, too. Bundled-permit markets could reshape incentives so firms innovate broadly to cut the total health burden of their emissions. Unlike narrow single-pollutant programs that encourage end-of-pipe fixes, multipollutant markets reward system-level solutions – advanced catalytic systems, integrated biofilters and process redesigns. Those market signals accelerate the development and deployment of technologies that reduce multiple harms at once, delivering larger and more durable public health gains.

Equally important, bundled markets can spur cross-firm partnerships and industrial redesigns that treat pollution as a collaborative design challenge. For example, if neighboring plants each emit elements of the same harmful mixture, they can co-finance a shared technology – say, a centralized filtration unit or a heat-recovery system – that reduces several pollutants at once and lowers permit exposure for both. By adjusting the permit limits in a constructive manner, these market-based approaches create a self-perpetuating scheme that continues to improve air quality over time while fostering key industrial innovations that reshape the energy and manufacturing sectors.

All that said, implementation requires careful sequencing and interdisciplinary work. Constructing bundles requires toxicological and epidemiological inputs so the market price genuinely reflects health harms. Monitoring and verification systems must be robust so trades correspond to real reductions, and additional oversight will likely be needed to ensure the effectiveness of the bundles. Market design must explicitly prevent volatility and perverse incentives – and agencies would likely need new statutory authority to issue bundled permits and coordinate across jurisdictions. An RFF working paper formalizes much of this architecture, drawing on mathematical economics tools to match heterogeneous firms to permit bundles in a way that maximizes social welfare while meeting environmental goals.

How to begin? A pragmatic rollout is the sensible path. Start with pilots in well-bounded sectors or regions – an industrial cluster, a power-generation region or a metropolitan airshed – where monitoring is already strong and the mix of pollutants is well understood. Pair pilots with independent evaluation, transparent data sharing and direct engagement with affected communities so adjustments are based on evidence and equity, not conjecture. Funding interdisciplinary teams that link economists, toxicologists and public health experts would help translate pollutant reductions into health gains and use those health metrics to set permit weights and price signals.

Multipollutant cap-and-trade is not a silver bullet, but it is a realistic mechanism to bring market discipline to the messy consequences of the chemical brews generated by modern economies. Policymakers have the opportunity to move from idea to evidence by funding pilots, building up monitoring capacity and commissioning the interdisciplinary research needed to design permit bundles that reflect real public health stakes.

From Concept to Implementation

The shift from single-pollutant to multipollutant regulation requires more than a technical adjustment. It represents a structural change in the way governments conceive of environmental risk. Existing regulatory systems were built around pollutant-by-pollutant statutes and agency divisions, and this fragmented architecture resists reform. As multipollutant research leaders George Hidy and William Pennell argued in 2010, reform requires not only technical expertise but also explicit statutory direction – legal mandates that acknowledge multipollutant assessments must be conducted as a matter of course, rather than as exceptions.

Even with strong political will, multipollutant regulation confronts the inherent difficulty of measuring and modeling chemical mixtures. Science has advanced dramatically in recent years. Epidemiological evidence now shows that mixtures drive health outcomes that no single-pollutant assessment would predict. New statistical and computational tools help make sense of this complexity. Artificial neural networks – machine learning systems designed to capture nonlinear relationships – have been applied to pollution datasets in São Paulo, Brazil, to improve predictions of mortality risk under multipollutant exposures. Quantile-based g-computation, pioneered in recent epidemiological work, enables researchers to assess how incremental changes across multiple pollutants affect health, even when the changes in a variety of pollutants are statistically correlated. Both methods represent a leap beyond traditional statistical regression analysis, which falters in the face of collinearity and synergy.

Yet these advances highlight the need for robust infrastructure. The new modeling techniques demand high-frequency monitoring data, large sample sizes and computing capacity that many agencies still lack. Multipollutant regulation will thus require scaling up monitoring networks like the EPA’s NCore stations, expanding coverage to disadvantaged neighborhoods and ensuring data interoperability across agencies. Without this empirical backbone, even the most sophisticated models cannot reliably guide policy.

Industry, too, faces uncertainty regarding multipollutant reform. Firms are accustomed to pollutant-specific regulations: a carbon price here, an SO₂ allowance there. A bundled-permit system that prices pollutant portfolios is unfamiliar terrain. And from a corporate perspective, uncertainty about future compliance costs can deter investment in abatement or clean technology. For example, a company considering retrofitting a refinery must know whether multipollutant reductions will be rewarded in future markets or if only carbon emissions savings will count.

Evidence from prior cap-and-trade programs shows that stable and credible regulatory signals are critical. Without them, firms delay investment or focus on short-term, pollutant- specific fixes. A 2007 Georgia case study found that clear multipollutant frameworks enabled firms to adopt technologies that reduced both particulates and chemical precursors to ozone, an outcome unlikely under fragmented rules. For industry, the challenge is not a lack of capacity, but a lack of predictable incentives.

Policymakers can take immediate steps that do not require a wholesale rewrite of environmental law. Governments can adopt multipollutant health-based indices like China’s proposed health-risk-based air quality index, which integrates multiple pollutants into a single measure of risk. Doing so improves communication with the public and ensures that regulators evaluate progress against health outcomes, not arbitrary pollutant- by-pollutant thresholds. Multipollutant cap-and-trade systems could bundle existing permits across pollutants, ensuring firms are able to profit the most from holistic abatement strategies. Regulators can further tailor these markets to equity goals by weighing permits toward pollutants disproportionately affecting vulnerable communities.

The path forward will not be easy. Institutional inertia, scientific complexity and industry uncertainty are real obstacles. In the United States, moreover, environmentalists and environmental science are facing a potent political backlash fed by special interest opposition.

But pollution and its health consequences aren’t going away on their own, and it’s hard to imagine a long-term equilibrium in which the political constituency for clean air and water is unable to reassert its clout. When it does, it is immensely important to recognize that the tools now exist to overcome the technological, administrative and economic challenges to multipollutant management, including advanced models that capture mixture risks, monitoring systems that measure multiple pollutants and policy designs that align financial incentives with health outcomes.

The signs are there for policymakers, industry leaders, scientists and the public to recognize air pollution as a holistic health crisis demanding integrated solutions. If Ella’s legacy teaches us anything, it is that the price of delay is great and that a cleaner, fairer future is possible if we make it a priority.