Do you know or have you heard of anyone killed by air pollution?
Is there an International Classification of Disease Code for air pollution?
Have any polluters been prosecuted for murder by air pollution?
Age at death or reduction in life expectancy is a more important mortality metric because it accounts for age differences. With a loss of 2 years, the U.S. ranks 2nd in the world in COVID mortality, sandwiched between Russia and Bulgaria at about twice the global median . Wallace-Wells cites AP-related losses of up to 10 years in South Asia that have little relevance to the developed world. Typical AP loss estimates range up to several months in the U.S., ostensibly because only the frail elderly are at risk. 
Another metric favored by environmental economists is the monetary cost of premature death, currently set at $7 million based on lost wages and productivity. This would amount to about $6 trillion for COVID deaths thus far, with the loss of 1.6 million life-years. At $7 million each, 10 million global excess AP deaths would amount to about 80% of global gross domestic product! These economic estimates are highly unrealistic.
Wallace-Wells cites a U.S. estimate of 370,000 AP-related deaths annually or about 10% of all deaths, slightly more than annual COVID-19 deaths and the 3rd leading cause of death. However, there is an important difference between these two categories. COVID-19 deaths occur to identified individuals; by contrast, AP-related deaths are statistical estimates subject to the vagaries of statistical regression analysis. Here are some of them.
- Exposure estimates are based on being outdoors 24/7. Indoor exposures have not been considered.
- Long-term relationships are typically based on concurrent AP exposures rather than lifetime accumulations of deposits or cumulative effects.
- AP exposures are not limited to one pollutant at a time; the probability of a co-pollutant effect depends on the relative precision of measurements.
- Health status before lethal exposure is seldom considered.
- Animal or human clinical studies have not confirmed epidemiological findings.
- EPA only regulates pollutants that have been measured. Monitoring networks are only established for pollutants that are regulated. (EPA likes to look where the light is best.)
- Routine monitoring networks have not been established for some of the most hazardous pollutants, such as arsenic, benzene, benzo(a)pyrene, formaldehyde, elemental or organic carbon, mercury, nickel, vanadium, nanoparticles.
- Intervention studies involving new sources and changes in fuels or traffic patterns have not reported long-term health effects. There are few long-term studies supporting AP relationships with hospitalization or morbidity.
- Long-term studies of populations necessarily lack individual data on potentially confounding personal characteristics like education, smoking, body-mass index, income. Cohort studies lack data on individual income because of privacy concerns. No AP mortality study has considered personal exposures.
- No long-term mortality study has adjusted for coincident short-term effects.
We all enjoy clean air, and no one doubts that extensive exposure to air pollution may have deleterious effects. But habeas corpus applies; studying mortality requires finding the bodies.
 Ten Million a Year David Wallace-Wells on polluted air
 Effects of covid-19 pandemic on life expectancy and premature mortality in 2020: time series analysis in 37 countries BMJ DOI: 10.1136/bmj-2021-066768
 A new time-series methodology for estimating relationships between elderly frailty, remaining life expectancy, and ambient air quality Inhalation Toxicology DOI: 10.3109/08958378.2011.638947