Introduction

Air pollution remains a persistent global threat to human health. While extreme episodes often attract the most attention, most individuals experience pollution as steady, low-to-moderate exposure over many years. Existing literature has documented that even short-run fluctuations harm health and productivity, yet far less is known about whether sustained exposure produces lasting physiological damage. Distinguishing temporary effects from cumulative harm is essential for assessing welfare costs and the adequacy of regulatory standards. This paper provides causal evidence on how long-term exposure to fine particulate matter (PM2.5) impacts the onset of chronic disease in a young, initially healthy population.

Estimating long-term impacts is challenging due to the endogenous sorting of people to locations based on income and health. This study addresses the challenge by focusing on U.S. Army soldiers. After basic training, service members are assigned to bases across the country according to operational needs rather than personal preference. This institutional feature implies conditional random assignment, where pollution exposure at the initial posting is as good as random. The researchers use this initial assignment as an instrument for average annual pollution exposure over a soldier’s career.

Institutional setting

The U.S. Army provides a unique setting for studying long-term morbidity. Recruits are exceptionally healthy at baseline, as they must pass stringent physical and medical screenings. Only about 23% of Americans of prime recruiting age meet military eligibility standards. Consequently, any health conditions observed during service represent new diagnoses rather than preexisting conditions. Furthermore, Army personnel have uniform, generous access to health care, allowing researchers to distinguish persistent conditions from transient episodes while controlling for socioeconomic disparities in care-seeking.

Data

The researchers assembled a dataset tracking personnel records for roughly 150,000 soldiers who served at least five years. These records are linked to high-resolution annual measures of PM2.5 derived from satellite data, providing consistent nationwide coverage. Health outcomes are measured using administrative files of annual healthcare encounters for circulatory and respiratory conditions between 2015 and 2020. To ensure the capture of persistent illness rather than isolated episodes, morbidity is measured using repeated diagnoses of the same condition; the preferred measure requires at least three claims for the same outcome.

Econometric strategy

The study employs a two-stage least squares (2SLS) system. The first stage uses the PM2.5 level at a soldier’s initial assignment location to instrument for their average exposure over their military career. The model conditions on a full set of occupation, skill, rank, and quarter-of-assignment indicators. This approach exploits variation in exposure induced by the Army’s personnel system. The first-stage F-statistic is 1,201, indicating a strong instrument. The second stage relates morbidity to the instrumented average pollution exposure. Models also consider the number of service years in which PM2.5 exceeds the current National Ambient Air Quality Standards (NAAQS) of 9 µg/m³.

The effect of long-term pollution exposure on morbidity

The results reveal that instrumented long-term PM2.5 exposure leads to pronounced increases in circulatory morbidity. A 1 standard deviation increase in PM2.5 results in a statistically significant 5.5% increase in circulatory diagnoses relative to the mean of 112 cases per 1,000 soldiers. Disaggregated results suggest this effect is driven primarily by hypertension. One unit of PM2.5 exposure increases hypertension rates by 2.46 cases per 1,000 soldiers when climate controls are included, representing a 6.4% increase relative to the mean.

For respiratory conditions, the estimates show a smaller, 1.3% statistically insignificant increase. At the diagnosis level, however, there is evidence of increases in asthma and influenza/pneumonia, though these are sensitive to climate specifications. A one-unit increase in average PM2.5 exposure implies an increase of 1.23 asthma cases per 1,000 soldiers. Notably, exposure over shorter durations (less than two years) has little detectable effect, which is consistent with the cumulative nature of pollution’s health burden. The mean PM2.5 exposure in the sample is 7.5 µg/m³, which is below the current U.S. EPA annual standard of 9 µg/m³.

Conclusion

The findings demonstrate that even moderate long-term exposure to air pollution can erode health capital in ways invisible to studies focused on short-run variations or mortality alone. Sustained exposure can produce measurable chronic disease even in a population of relatively young adults who begin from an unusually healthy baseline. These results highlight the importance of regulatory standards that account for chronic exposure risks. Since the civilian population is generally less healthy and has worse access to healthcare than the military sample, the corresponding burden in the broader population is likely to be larger.