What Are the Long-Term Health Consequences of Air Pollution?
Deryugina and Reif (2026) examine how long-term exposure to air pollution affects mortality and life expectancy. They analyze approximately 18 million U.S. death records from 1972 to 1988, combined with daily pollution and weather data. Using changes in wind direction to identify pollution’s effects, they find that a 10% increase in sulfur dioxide pollution raises daily mortality by 0.30%. Their demographic model predicts that permanently reducing pollution by approximately 10% would increase life expectancy by 1.18 years. About 90% of these survival benefits occur after age 50, suggesting that cleaner air produces substantial health benefits decades after exposure declines.
Why This Article Was Selected for The Policy Scientist
Air pollution remains a major public health concern, with consequences for life expectancy, healthcare expenditures, and economic productivity. Understanding its cumulative effects is particularly important as governments assess the long-term benefits of environmental regulations. This study contributes to that discussion by demonstrating that conventional estimates may substantially understate pollution’s lifetime health consequences. Its historical dataset is unusually comprehensive, encompassing approximately 18 million deaths. The use of wind direction as an instrumental variable strengthens causal inference, although lifetime projections depend on modeling assumptions. While the findings are most directly applicable to the United States, the underlying mechanisms may extend to other industrialized countries.
Full Citation and Link to Article
Deryugina, T., & Reif, J. (2026). The long-run effect of air pollution on survival. American Economic Review. Advance online publication (forthcoming).
Central Research Question
How does long-term exposure to air pollution affect mortality and life expectancy, and how much can sustained reductions in pollution improve survival over an individual's lifetime?
The research examines whether conventional estimates of air pollution's health consequences substantially underestimate its cumulative effects. Most studies measure mortality shortly after pollution exposure, often within days or months. However, pollution may initiate biological processes that gradually damage health and increase mortality decades later. Consequently, short-term studies may capture only a fraction of the total health burden.
The authors distinguish between two mechanisms: mortality displacement, in which pollution accelerates deaths among individuals already near the end of life, and accelerated biological aging, in which pollution causes lasting health deterioration. Their central objective is to determine how these mechanisms influence lifetime survival and estimate the potential gains in life expectancy from permanently reducing air pollution.
Previous Literature
The research builds on several influential contributions to environmental economics, epidemiology, and demographic modeling.
Currie and Neidell (2005) examined the relationship between air pollution and infant mortality, while Chay, Dobkin, and Greenstone (2003) investigated how pollution reductions affected infant health. These studies established important connections between environmental conditions and mortality but focused primarily on outcomes early in life.
Deschênes and Greenstone (2011) developed methods for translating short-term mortality estimates into changes in life expectancy. Deryugina et al. (2019) subsequently used changes in wind direction to identify the causal effects of fine particulate pollution on mortality and healthcare expenditures among elderly Americans.
Other research examined longer exposure periods. Chen et al. (2013) and Ebenstein et al. (2017) investigated the long-term health consequences of air pollution in China. Anderson (2020) provided additional quasi-experimental evidence concerning mortality effects over several years.
The present study also draws on Lleras-Muney and Moreau (2022), whose demographic model describes how individual health deteriorates with age and how these changes influence survival.
Collectively, this literature demonstrates that pollution affects mortality but leaves substantial uncertainty about its cumulative lifetime consequences. The principal contribution of the current research is integrating credible short-term causal estimates with a demographic model capable of projecting survival over an entire lifetime.
Data
The researchers construct a comprehensive dataset combining approximately 18 million death records from the United States between 1972 and 1988 with daily measurements of air pollution and weather conditions.
Mortality information comes from the National Vital Statistics system and includes the date, location, and underlying cause of death. These records allow the researchers to distinguish cardiovascular disease, cancer, external causes, and other medical conditions.
Pollution measurements come from the Environmental Protection Agency's Air Quality System. The primary pollutant is sulfur dioxide (SO₂), which originates largely from burning coal and oil. Sulfur dioxide also contributes to the formation of fine particulate matter (PM₂.₅), making it relevant to broader questions about atmospheric pollution.
The average sulfur dioxide concentration during the study period was approximately 9.1 parts per billion.
Weather information includes daily temperature, precipitation, humidity, wind speed, and wind direction. These observations allow the researchers to distinguish pollution-related mortality changes from those associated with meteorological conditions.
The historical dataset is particularly valuable because exact dates of death are publicly available for the study period, permitting unusually detailed examination of mortality following daily pollution exposure.
Methods
The study employs a two-stage analytical strategy combining causal inference with demographic modeling.
First, the researchers use instrumental variables estimation to identify the causal relationship between daily pollution exposure and mortality. Wind direction serves as the instrument because changing winds transport pollution into or away from particular geographic areas, creating variation in exposure that is largely independent of individual health characteristics.
The statistical models incorporate detailed weather controls, county-by-month fixed effects, and month-by-year fixed effects. These adjustments account for geographic differences, seasonal patterns, and broader changes affecting mortality. The researchers also examine mortality over different periods following exposure and distinguish effects by age and cause of death.
Second, they incorporate the causal estimates into a demographic model describing how health deteriorates throughout life. The model distinguishes immediate mortality among vulnerable individuals from persistent biological damage that increases future mortality risk.
The researchers calibrate the model using observed survival patterns and evaluate its predictions against mortality estimates from different age groups, follow-up periods, and previous quasi-experimental research.
This approach allows them to project the lifetime consequences of sustained pollution reductions without requiring decades of individual-level follow-up data.
Findings/Size Effects
The findings indicate that air pollution has both immediate and persistent effects on mortality, with substantial implications for life expectancy.
A one-part-per-billion increase in daily sulfur dioxide concentrations, approximately 10% of the historical average, increases same-day mortality by 0.071 deaths per million people, equivalent to a 0.30% increase.
The cumulative effect grows considerably after the initial exposure. Within one week, the estimated increase reaches 0.17 deaths per million. After approximately one month, it rises to 0.25 deaths per million, more than three times the immediate effect.
Mortality patterns differ substantially by cause of death. Cancer-related mortality increases immediately following exposure but subsequently declines, suggesting that pollution accelerates deaths among individuals already experiencing serious illness.
In contrast, mortality associated with cardiovascular disease and other medical conditions continues increasing during the following month. These findings are consistent with pollution producing lasting biological damage rather than simply advancing deaths by several days.
The most consequential results concern lifetime survival.
The demographic model predicts that permanently reducing sulfur dioxide concentrations by one part per billion, together with associated reductions in fine particulate pollution, would increase life expectancy at birth by approximately 1.18 years.
The estimated 90% confidence interval ranges from 0.42 to 2.01 additional years. Alternative model specifications produce estimated gains between 0.80 and 1.25 years.
These projections substantially exceed the 0.12-year improvement implied by simply extrapolating the observed one-month mortality effects.
The timing of these benefits is particularly important. Approximately 90% of the projected improvement in survival occurs after age 50, while more than 75% occurs after age 65.
Consequently, reductions in pollution exposure during childhood and early adulthood may produce their largest measurable mortality benefits several decades later.
Additional analyses suggest that the observed mortality effects reflect exposure to both sulfur dioxide and associated fine particulate matter. The researchers therefore interpret their estimates as capturing the combined consequences of these pollutants rather than the isolated effect of sulfur dioxide alone.
Conclusion
The research demonstrates that evaluating air pollution primarily through immediate mortality outcomes can substantially underestimate its long-term health consequences.
By combining quasi-experimental estimates with demographic modeling, the researchers provide evidence that pollution affects survival through both short-term mortality displacement and persistent deterioration in health.
Their central finding is that sustained reductions in pollution can produce substantially larger improvements in life expectancy than conventional short-term estimates suggest.
The results also have important implications for evaluating environmental regulations. Because most projected survival benefits emerge after age 50, policies that reduce pollution today may generate measurable mortality improvements across several generations.
The findings are particularly relevant to understanding the long-term consequences of pollution reductions following the 1970 Clean Air Act.
Nevertheless, the lifetime estimates are model-based projections rather than directly observed causal effects of decades-long exposure. Their accuracy depends on assumptions about biological aging, mortality displacement, and the persistence of pollution-related health damage.
Despite these limitations, the study provides a framework for estimating the cumulative health consequences of environmental exposure and demonstrates why short-term mortality studies may provide an incomplete assessment of the benefits associated with cleaner air.
Comments