In case you are in a hurry, here is the short version.
I discovered that our health and cognitive ability immediately worsen when we are exposed to even low levels of air pollution (i.e., below or at WHO-established thresholds). These effects are so large that we can quantify them in percentage points of GDP lost. They are also very local and short-run. Hence, a city/region investing in reducing air pollution will immediately benefit economically (in terms of higher GDP).
If we sum up its short-run and long-run negative effects, we see that the cost of air pollution is gigantic. It should be a central concern in Europe, where the race to find energy sources alternative to Russian gas is becoming frantic. But unfortunately, nobody in the policy circles seems to care. For example, subsidizing the burning of wood and other biomass is completely idiotic: it is sold as a “green option” on the rather disputable ground that it is carbon neutral. At the same time, it generates loads of air pollution (without mentioning the destruction of forests).
In case you have time, here is the longer version.
This summer, I went through a bit of a rabbit hole: I read two books about air pollution (“Choked” by Beth Gardiner and “Cleaning the air” by Tim Smedley) and several research papers (especially from economists but not only). I was motivated by the sudden realization that my views on air pollution were deeply wrong.
Mine was essentially the mainstream view based on scientific evidence produced until ten years ago. In short, I thought air pollution was a problem only:
(i) in places with extremely high levels of air pollution---for example, in India, China, and other cities in the developing world.
(ii) in the long run. That is: non-extreme pollution levels (like the ones commonly experienced in cities across the developed world) may nonetheless have a detrimental effect after 20 or 30 years of exposure.
But about ten years ago, several researchers (including economists studying health, labor, and education) started wondering whether low pollution levels can have short-run negative effects. Rather quickly, various studies showed that these negative effects exist and are very large. The recent pandemic provided even more evidence for this because the various lockdowns led to a decrease in air pollution and an immediate improvement in several health outcomes. Air pollution is not only a problem when it is extreme or in the long run (as I thought). It is a big problem also at low levels and in the short run.
The academic literature
Before discussing some of these recent studies, two comments. First, they mainly consider one specific pollutant: particulate matter with a diameter smaller than 2.5 (PM 2.5), measured in μg/m3 (micrograms per cubic meter of air). To give you a sense of the numbers, before the pandemic, the average annual concentration of PM 2.5 in some European and North American cities was: 23 μg/m3 in Milan; 15 μg/m3 in Paris; 13 μg/m3 in London, Los Angeles, and Chicago; 7 μg/m3 in New York (compare this with 100 μg/m3 in Dehli; 110 μg/m3 in Hotan, China; 90 μg/m3 in Lahore; 80 μg/m3 in Dhaka: source here).
This is not to say that PM 2.5 is the most dangerous pollutant. In fact, several recent studies show that ultrafine particles (those with a diameter smaller than 0.1 microns) can pass directly into the bloodstream, reaching various organs, including the brain, and are probably the most dangerous pollutant. Nonetheless, the concentration of PM2.5 and ultrafine particles correlates strongly, as they are both produced via combustion.
Second, the vast majority of these studies look at relatively small variations in pollution levels. For example, the difference in pollution nearby a highway exit before and after the installation of some E-Zpass (an automatic electronic toll collection system). Or normal variations in pollution levels due to weather changes. Or those due to the installation of a cheap indoor air filter. The message is, therefore, that a small reduction in air pollution leads to measurable improvements in several outcomes---we can extrapolate what happens when we fully eliminate air pollution, but that is not what those studies do.
Alright, what do those studies show? I put a detailed list at the end of the post, but to summarize: they consider pollution levels that are “normal” in the US and Europe and find that:
air pollution immediately worsens our cognitive ability: on more polluted days, stock traders earn less (lose more), chess players make more mistakes, judges take longer to decide on a verdict, baseball referees make more mistakes, and there are more traffic accidents.
air pollution negatively affects learning and school performance: kids tests’ scores are lower on more polluted days; kids learning outcomes are lower in more polluted years. Also, the effects are very large! For example, Gilraine (2020) finds that installing air purifiers in school increased student performance almost as much as if class sizes were reduced by a third.
air pollution has measurable instantaneous and short-run effects on our health. For example, on more polluted days, people have more heart attacks. Also, introducing an automatic toll system (E-ZPass) in some American cities reduced pollution in the surrounding areas, decreasing premature births and children born with low weight. Finally, exposure to higher levels of air pollution over 4-6 years increases the chance of developing dementia, Alzheimer's disease, ASD, Parkinson's disease, of having a stroke, as well as overall mortality.
Is eliminating air pollution cost-effective?
Given the above evidence, you may wonder whether we can quantify the cost of air pollution in terms of GDP; after all, people make more mistakes, are less attentive, and are more likely to be sick on more polluted days/years. This is exactly what a team from the OECD found. Focusing on air pollution variations due to weather events, they establish that “a 1 μg/m3 increase in PM2.5 concentration [...] causes a 0.8% reduction in real GDP that same year. Ninety-five percent of this impact is due to reductions in output per worker, which can occur through greater absenteeism at work or reduced labor productivity.” A similar study in China found similar effects: a “1% decrease in PM2.5 nationwide increases gross domestic product by 0.039%” (Fu et al., 2021).
Taken at face value, these effects are gigantic. To give a sense, suppose that a city like Paris (where I live) was to eliminate air pollution. Also, consider the effect found in the OECD study and assume that they are linear, in the sense that the effect of reducing pollution from, say, 15 µg/m3 to 14 µg/m3 is equal to the effect of reducing pollution from say, 7 µg/m3 to 6 µg/m3 (which is probably not true but is nonetheless the best guess we can make). Paris’ GDP would jump by approximately 12% (!!). Or take Milan, a city where I used to live. If Milan were to eliminate air pollution, its GDP would jump by 18% (!!!!). Similarly, London's GDP would jump by approximately 10%, and New York's GDP would jump by approximately 6%. These numbers are huge, and they only consider short-term (i.e., year-on-year) benefits without including economic benefits accruing over a longer period, such as better overall health and educational outcomes. The overall effect on GDP is likely to be much larger.
All right, this is the benefit side. How about the cost side? This is a tough question for which I could not find a precise answer. But I will try to figure out some numbers to get a sense of the ballpark. First, note the policies that eliminate air pollution are, more or less, the same policies that also eliminate greenhouse gas emissions: electrification of the means of transportation, clean energy production, and so on (more below). To get a sense of the cost of eliminating air pollution, we can therefore look at the cost estimate of the green transition, which a recent McKinsey report puts at 6.8/8.8 percent of global GDP between now and 2030 (of which approximately 2.8 percent of global GDP in new investment, and the rest investments already planned).
Can we compare the McKinsey number with the benefit of eliminating air pollution derived earlier? Well, we have to be careful with this comparison because:
The McKinsey number is an investment. Some of these investments will generate economic activity and should not be considered a cost. The overall cost is likely to be much, much smaller.
The McKinsey number is the investment required to achieve the green transition (and eliminate air pollution) over time. While the back-of-the-envelope calculations I made earlier are supposed to capture the benefit of instantaneously eliminating air pollution.
Nonetheless, I find it striking that the McKinsey number is in line with the benefit of eliminating air pollution I derived earlier. To me, it suggests a rather amazing possibility: that the green transition may pay for itself once we include in our cost-benefit analysis the short-run benefit of reducing air pollution!
Air pollution and climate change
Another fascinating angle is the relationship between air pollution and climate change. At a macro level, the two issues almost completely coincide because the primary cause of both problems is the combustion of fossil fuels.
(as an aside, I initially thought intensive animal farming was a problem for climate change because of methane emissions, but not for air pollution. It turns out that intensive animal farming is also an important source of air pollution, causing significant problems in the surrounding communities. Similarly, I thought that the shipping industry might be an issue for climate change but not for air pollution. I was wrong on this one, too: commercial ships burn one of the worst types of fuel available, and by doing so, they cause a very high level of pollution around big ports. Finally, airplanes emit greenhouse gasses, but I could not figure out their impact on air pollution. End of the long aside.)
In my opinion, it is politically easier to push for the green transition by arguing that it will reduce air pollution rather than arguing that it will contrast climate change. The point is that the net benefit of contrasting climate change is very uncertain and probably negative for a single country (of course, it is positive and large if we look at the world as a world, but very few decisions are taken at this level). For example, the US may invest in reducing its GHG emissions. But if China continues building coal power plants, the benefit of these investments (for the US) may be zero. If positive, it may not accrue in the US but in some other country --- say because droughts in Iran are less severe than if the US had not invested. On top of that, these benefits accrue very far in the future. Of course, some people may be willing to make the required sacrifices out of an ethical principle, but the fraction of the population that only looks at cost and benefits will not be convinced.
Instead, when the same policies are justified in terms of reduction of air pollution, then the benefit of these investments are local (i.e., nobody cares anymore about what China does) and accrue in the short term. You can expect cold-blooded business people, bankers, and economists (!) to join the protest on Friday!
But when we move away from the macro level to look at specific cases, we discover instances where there may be tension between fighting climate change and reducing air pollution. The most relevant at the moment is the burning of biomass (woods and pellets). The reason is that policymakers worldwide consider the burning of wood and pellet as a “climate friendly” source of energy, on the ground that if we re-plant another tree in place of the one that was burned. But this is not as clear as it looks because (i) the promise of re-planting could be hollow, (ii) the tree's age determines the ability to absorb CO2, with older trees absorbing more CO2 (iii) the soil’s ability to absorb COs is also reduced (see the discussion here). Overall, carbon-neutrality depends on the age of the tree that was cut relative to the age at which the replacing tree will be cut, and it is very difficult to calculate, let alone verify. On the other hand, it is quite clear that wood burning is a leading cause of air pollution. Furthermore, because of its touted green credentials, wood burning is a rapidly increasing source of pollution in developing countries. To add to the problem, European policymakers decided to subsidize the burning of wood and other biomass to compensate for the shortfall from the lack of Russian gas. Currently, the most visible outcome of this policy is the large-scale destruction of forests. This winter, it may cause unbreathable air in Europe.
What to do about it
The most effective way to reduce air pollution is via political activism: communities of various sizes greatly impact the levels of air pollution on the street. In this respect, the book “cleaning the air” by Tim Smedley nicely summarizes how various cities worldwide were able to reduce air pollution and proposes a point-by-point blueprint that any city can follow. BTW, the central recommendation is to ban diesel cars because even the most modern ones pollute 20 to 50 times more than an equivalent gasoline car
(Another aside: the history of how European policymakers came to subsidize diesel cars is one of those gut-wrenching stories mixing bureaucratic ineptitude, corporate greed, cheating, and the death of millions due to air pollution, second only to the story of how lead ended up in gasoline and then in our bodies).
But this is not to say that individual behavior cannot make a difference; on the contrary! Remember that small variations in air pollution exposure can have a measurable effect on cognitive ability and health. At the same time, the most dangerous type of pollution (ultrafine particles) decays rather rapidly as we move away from its source. The two facts together imply that our behavior matters tremendously. Walking along a polluted street or a side street; walking close to the cars or far away from them (if the sidewalk is large), installing air filters, choosing which widow to open (preferably one not giving on a busy street); planting trees and other plants that can filter ultrafine particles; all these choices do make a difference on the level of pollution you (and your loved ones) are exposed, and ultimately on your (and their) health. Again, I think Tim Smedley does a wonderful job explaining this, so I recommend reading his book.
Epilogue
At the end of this journey, I was left wondering: why are these facts not better known? Why don’t they feature more prominently in the policy debate? Part of the explanation is that air pollution is a rather complex topic. For example, despite my best efforts, it is still quite unclear to me how different pollutants transform into other pollutants once they enter the atmosphere (for example, somehow there may be no PM2.5 emissions, and nonetheless, some of it is produced in the atmosphere from the interaction of various other gases that are emitted).
But I think complexity is, at best, only part of the story. I believe the root cause is our innate reluctance to consider the “normal state of affairs” as extremely dangerous. To say it differently, you may see a newspaper article claiming that “pollution levels are three times the average and this is causing serious health issues”. You will never see a newspaper article claiming that “pollution levels are close to the average and this is causing serious health issues”. But this is exactly the point: what we have considered “normal” up until now is causing severe damage to our cognitive ability and health. How to make this message worth a newspaper title? I think answering this question is key to improving the air we breathe.
Appendix: some notable studies mentioned in the post
Major-league baseball umpires: Archsmith et al. (2016) find that the number of incorrect calls increases by 2.6% when PM2.5 increases by 10g/m3
NYSE returns: Heyes et al. (2016) find that a 7g/m3 increase in PM2.5 in New York causes a same-day fall of 12% in NYSE returns.
Individual investor's performance: Huang et al. (2020) “find a negative relation between air pollution and trade performance.”
Chess players: Künn et al. (2019) find that ``an increase of 10 µg/m³ raises the probability of making an error by 1.5 percentage points and increases the magnitude of the errors by 9.4%.’’
Trial judges: Kahn & Li (2020) studied court cases in China. They find that a 1% increase in PM2.5 leads to a 0.182% increase in case duration (i.e., judges take longer to decide)
Traffic accidents: Sager (2019) looks at data from the UK and finds that “an increase of 0.3–0.6% in the number of vehicles involved in accidents per day for each additional 1 μg/m3 of PM2.5.”
Students’ test scores are lower on more polluted days (Ebenstein et al. 2016, Roth, 2021, Zhang et al. 2018, Gilraine and Zheng 2022);
Students’ school achievements are lower in more polluted years (Ebenstein et al., 2016; Persico and Venator, 2021; Gilraine, 2020; Duque and Gilraine, 2020; Heissel et al., 2020; Mullen et al., 2020; Marcotte, 2017);
The lockdown decreased air pollution in most US cities, which caused an instantaneous reduction in heart attacks (Aung et al. 2022)
Dominici et al. (2022) studied the impact of pollution exposure well below WHO thresholds over four years. They find “a 6% to 8% increased risk of mortality per 10 micrograms per cubic meter (μg/m3 ) increase in PM 2.5 exposure across the different analyses, with stronger associations at exposure levels below the current annual national standard of 12 μg/m3.”
Fu et al. 2019 “Short- and long-term PM2.5 exposure was associated with increased risks of stroke [...] and mortality [...] of stroke. Long-term PM2.5 exposure was associated with increased risks of dementia [...], Alzheimer's disease [...], ASD [...], and Parkinson's disease [...]
The relationship between pollution and premature births and low birth weight is so strong that the introduction of an automatic toll system (E-ZPass) in American cities reduced both problems in areas close to toll plazas (by 10.8 % and 11.8 %, respectively, Currie and Walker, 2011)