Air pollution: The invisible effects on productivity, health and economic output

Last year, the WHO revised its global air quality guidelines, recognising air pollution as “the single biggest environmental threat to human health” and noting that “[t]he burden of disease resulting from air pollution also imposes a significant economic burden” (WHO 2021). But the burden of disease is not the only economic cost arising from poor air quality. An increasing body of work shows that exposure to air pollution, even at the relatively low concentrations found in developed countries, can have meaningful impacts on both physical and cognitive performance. Though not diagnosable as diseases, these effects of air pollution have meaningful impacts on our economic output and well-being.

Some of the ‘non-health’ effects of air pollution are nearly contemporaneous with the pollution exposure itself and represent short-term reductions in performance measured relative to days or areas with less pollution. Other impacts are detectable many years after exposure and manifest as accrued physiological damage significant enough to affect behaviour, but below the threshold that would lead to a medical diagnosis. From a policymaking perspective, the findings from this literature imply that even small improvements from relatively low baseline levels of air pollution exposure may have substantial economy-wide implications.

Immediate impacts of air pollution include reduced productivity at work. Investigations of the effects of air pollution on workers in industries as disparate as professional sports (Lichter et al. 2017, Archsmith et al. 2018), garment production (Adhvaryu et al. 2019), agriculture (Graff et al. 2012), call centres (Chang et al. 2019) and the court system (Kahn and Li 2020) have found that exposure to air pollution reduces worker productivity (though He et al. 2019 finds negligible effects in a severely polluted factory setting). These studies show that the level of air pollution on a given day impacts worker productivity even for healthy workers; the harm caused by air pollution is not solely the result of medical conditions caused by prolonged exposure. This effect on the productivity of workers on the job is separate from the effect of air pollution on labour supply; several studies find that increases in air pollution reduce labour hours (Aragón et al. 2017, Hanna and Oliva 2015, Holub et al. 2021). Studies that look at production at larger scales, rather than focusing on particular worksites or cities, also find meaningful impacts of air pollution on overall output (Decelerate et al. 2019, Fu et al. 2018).

The finding that air pollution reduces productivity in physically demanding jobs is consistent with longstanding evidence on the deleterious impact of air pollution on the circulatory system. But recent biomedical research also suggests a link between air pollution and neuroinflammation in the brain. This link may explain repeated findings of cognitive and behavioural effects of air pollution that are hard to attribute to cardiovascular impairments.

A small but growing literature studies how pollution makes it harder to think, in both academic and non-academic settings. Ebenstein et al. (2016) find that an increase in fine particulate pollution¹ reduces scores on high-stakes tests in Israel (with a standard deviation increase in fine particulates lowering scores by 1.7% of a standard deviation), while Roth (2021) finds that particulate matter in classrooms reduces the scores of London-area university students. Zhang et al. (2018), examining a nationally representative survey in China, find a statistically significant effect on test scores of verbal ability.

Some evidence also suggests that pollution may not merely dim our mental acuity, but also lead to altered decision-making, sometimes in quite unexpected contexts. Several studies (Huang et al. 2020, Dong et al. 2021, Meyer and Pagel 2017) find that higher air pollution is associated with an increase in behavioural biases among investors and analysts in the financial sector, which appear to be linked to constrained attention and affect alterations. These findings potentially extend to a myriad of high-stakes situations and skilled service professions. Perhaps even more troubling is the possible link between air pollution and criminal activity. Herrnstadt et al. (2021) find that higher particulate matter concentrations lead to increases in violent crime across areas of Chicago (a one standard deviation increase in particulate matter raises violent crime by 2.9%); Burkhardt et al. (2019) estimate that a 10% reduction in daily PM2.5 and ozone in the US could save $1.4 billion in crime costs per year. Across the pond, Bondy et al. (2020) find that increases in the Air Quality Index (a measure of pollution) are associated with increased crime in London.

The effects of air pollution mentioned so far are caused by short-term fluctuations in air quality. But repeated exposure to air pollution can also reduce our capabilities in the longer term through accumulated damages. These consequences may be particularly severe when air pollution is encountered during gestation or early in life, a period of rapid physical development in which the effects of environmental insults are magnified.

One example of such a finding is from Isen et al. (2017), who use the differential impact of the Clean Air Act across different areas to examine the labour market outcomes of children born in more- and less-polluted counties. An increase in birth-year total suspended particles (TSP) by 10 microgrammes per cubic metre (about 1/7 of the then-current EPA air quality standard) was associated with a 1.4% decline in income and a 2.8% decline in the number of quarters employed. Voorheis (2017) and Colmer and Voorheis (2020) link in-utero exposure to TSP with later earnings and college attendance, finding that a 10 microgrammes/m3 increase in TSP is associated with approximately $250 a year in earnings reductions and a decline in college attendance rates of several percentage points. These findings are consistent with other evidence that links in-utero air pollution exposure with reduced academic performance, as measured by standardised test scores (Sanders 2012, Bharadwaj et al. 2017).

As we discuss in a recent overview of the literature (Aguilar-Gomez et al. 2022), great care is needed both when identifying the effects of air pollution (in the sense of determining what part of a correlation represents a causal impact) and when interpreting the resulting estimates. The most convincing evidence in the ‘immediate effects’ literature surmounts problems of sorting (which arises if, for instance, poorer people live or work in more polluted places, which makes it hard to disentangle the effect of poverty form the effect of pollution) and endogeneity (lower output may reduce pollution, even as higher pollution reduces output) by exploiting short-term shocks to exposure that are plausibly unrelated to human choices (such as exposure driven by weather patterns).

Even when we believe the study is well-designed to tell us about a causal connection, we must be careful in interpreting the resulting estimates. People do not stick to prearranged schedules and allow air pollution to take its toll. They may take action to avoid indoor air pollution when they are aware of it (by buying masks and air filters, by running inside on smoggy days); they also take action to deal with its accumulated effects on their body, whether or not they recognise air pollution as the culprit (perhaps by studying harder to compensate for reduced mental acuity). Some of these responses (such as buying air filters or masks) are measured as economic output but should properly be considered as economic costs of pollution; others, such as increased study time, are hard to measure but represent real costs. Estimates in the literature typically correspond to net impacts of these ‘avoidance’ and ‘amelioration’ behaviours; a full accounting of air pollution’s costs must attempt to measure these impacts and add them back in.

The growing concern over exposure to air pollution on the grounds of health is well justified, but there is more to the picture. Further research is needed, especially on air pollution’s cognitive effects and on the costs people incur in trying to avoid or cope with it (Barwick et al. 2020, Khanna et al. 2021). A growing body of evidence suggests that while many people suffer from diseases caused by air pollution, a far greater number may be negatively affected by air pollution even if they never see the inside of a doctor’s office. In our review (Aguilar et al. 2022), we conclude that virtually no aspect of human life is unaffected by the quality of our air. While we propose some future avenues of research and some unanswered questions that will improve policy design, it should be clear that, to a significant degree, we are what we breathe.