By Olivier Chanel
This text is adapted from the chapter “Air pollution and health: economic implications”, in preparation for the Springer Handbook of Labor, Human Resources and Population Economics.
In December 2020, for the first time ever, a court recognized that air pollution (AP) was the cause of the tragic death of a black 9-year-old girl with asthma who had lived 30 meters from a major ring road, in London. This ruling is “one giant leap” for public health, for several reasons. It makes an invisible killer like ambient AP (AAP) visible, by shifting from “association with adverse health effects in a population” to “causality at the individual level.” It confirms that long-term exposure to AP can provoke the onset of chronic diseases like asthma, and opens the door to other complaints likely to bring about change. Finally, it draws attention to the fact that AP disproportionately affects minorities and deprived areas, and that the public health issue is not restricted to older people or developing countries.
Anthropogenic AP emissions are conditioned by our political, energy and consumer choices, by the increasing movement of goods and people and by population growth, and represent a negative externality. First, through environmental effects on water, forests, biodiversity and ecosystem services, visibility impairment, contributions to climate change, agricultural yield losses and effects on buildings. Second, through the health burden imposed on society, the monetary valuation of which needs to take into account both market-related costs for the health care system and firms (production or productivity losses) and non-market-related effects on the population (loss of well-being including premature mortality, suffering, psychological effects, restrictions on activity and quality of life). AAP is considered the primary environmental risk to health worldwide in terms of number of premature deaths. Overall, associated welfare losses worldwide expressed in terms of gross domestic product amount to about 4.2% for mortality, 0.6% for morbidity and 0.07% for production losses. Without additional pollution control measures, these figures are expected to increase by about 50% by 2050, in particular due to urban population growth in Asia.
The following briefly focuses on three of the many health topics related to AAP: inequities, uncertainties and possible ways to improve knowledge.
INEQUITIES ISSUES AT NATIONAL AND TRANSNATIONAL LEVELS
Vulnerabilities associated with exposure to AP manifest themselves in several ways. First, some people are more exposed to AAP, for professional reasons (bus drivers, people in contact with traffic, road users), or because of their location (houses or other places close to major emission sources). Second, individuals of low socioeconomic status are at higher health risk, for two reasons. First, they may be more exposed to AP than more advantaged populations. Second, they may be more vulnerable to the effects of AAP than more advantaged populations, due to a comparatively worse health status and less access to health care. As a result, disadvantaged populations may suffer more from the health effects of AAP.
Besides higher vulnerability, individuals of low socioeconomic status also experience other negative environmental externalities at home (water pollution, toxic wastes, noise, natural hazards), worse working conditions (jobs requiring few or no qualifications, involving shift work) and poor housing conditions. Moreover, it is harder for them to combat negative environmental externalities through costly behaviors like prevention, avoidance or moving away from AAP exposure. In low - and middle - income countries, the consequences of AAP on health are worsened by the lower socioeconomic status on average: higher incidence of comorbidities due to chronic, infectious and non-communicable diseases, higher rate of social deprivation, lower health care quality. In addition, once the health effects have occurred, the lack of access to (affordable) treatments, with high out-of-pocket fees, implies a stop to working (or studying), which intensifies poverty.
At the international level, two sources of inequity are particularly worth noting.
First, high-income countries escape some of the AAP associated with production processes by having their goods manufactured in less developed and more polluted countries. Multinational polluting firms tend to move there because of lower land and labor costs as well as the weaker local capacity to mobilize against pollution, further weakened by poor local environmental regulations. The production of batteries for electric vehicles in China, for instance, is a case in point. The processes consume a huge amount of heavily coal-generated electricity and require the extraction of rare metals, both contributors to AAP in less developed countries, in order to improve air quality for the urban population of the richest countries.
Second, it needs to be borne in mind that country-specific monetary values are used by supranational agencies to assess AAP-related mortality effects. Thus, because the monetary value of a premature death in high-income countries is more than one hundred times larger than in low-income countries, we should be on the lookout for any potential plutocratic international environmental policy that appears promising but would actually tend to shift premature deaths avoided from high- to low-income countries.
DEALING WITH UNCERTAINTIES IN AN INTERDISCIPLINARY APPROACH
The economic evaluation of the health-related effects of AP adds uncertainties from the two main upstream disciplines to its own, which calls for an interdisciplinary approach.
Firstly, there are uncertainties in the characterization of population exposure, mainly from measurement of concentrations, from modeling to cover the geographical area under study and from assessment of both initial exposure and the exposure in the scenario under consideration. The quality of the modeling depends on the quality of the input data, the topography of the studied area, the availability of measurement data, etc., making the uncertainty spatially heterogeneous. This type of uncertainty can be examined, in particular by comparing the measured concentrations with the modeled values.
Secondly, the epidemiological uncertainties pertain to the quality of health data, the choice of an exposure-risk function and their transposability to the population studied, which depends on lifestyle, climate or the nature of the emission sources. Part of this uncertainty is provided by the confidence interval, which reflects the statistical variability specific to the relationship between exposure and health effect.
Finally, the quantification of economic uncertainties differs, since the underlying knowledge is more subjective than scientific, leading to an approach more normative than positive. The differences relate to the choice of the effects to be taken into account (market and non-market, direct and indirect, short- or long-term), the valuation method, the unit monetary values used (in particular for mortality, the greatest fraction of health effects) and technical parameters like discount rate, etc.
These three types of uncertainty are generally considered either independently or jointly, by including their respective sources in a methodologically preferable but more complex Monte Carlo simulation approach.
Reduced uncertainties are one explanation for the rising economic valuation of AAP-related health effects over time despite the decreasing concentrations of most pollutants experienced for the last 20 years. Firstly, the quality of measurements is improving: the pollution indicators adopted better reflect the compounds most harmful to health, the number of ground monitoring stations has increased and satellite data are being used more frequently. Secondly, epidemiological knowledge is evolving, thanks to more complex modeling, a growing number of cohort studies, the use of causal inference methods and better knowledge of the effects of low-level exposure. The scope of the effects accounted for is also widening, from short - to long-term effects, first for mortality and then for morbidity. Finally, monetary evaluation is evolving too, with increasing consideration of non-market components or higher values used to assess mortality. Based on the last decade of research and the classification of AAP in 2013 as carcinogenic to humans, the World Health Organization decided in September 2021 to strengthen its air quality guidelines for the protection of health, which suggests that the threat to health was substantially underestimated.
WHAT WILL THE FUTURE BRING?
Regarding data, better knowledge of the composition of particulate matter would be useful, as its nature seems to be a factor in health impacts. Measuring the size and chemical mass loading for sub-micron aerosol particles and black carbon would help determine their exact contribution to health impacts, so as to design the best policies from a regulatory and public health perspective. Increasing the number of ground-level measurement stations in developing countries would provide more accurate data for health impact assessments. Improvements are also expected on personal exposure, with individual monitoring to provide accurate real exposure data.
The exposome (which tackles all nongenetic factors) and epigenetics are beginning to be explored within the AAP framework, and that will shed light on individual susceptibility. A better understanding of in utero exposure and pregnancy outcomes is also important, given the huge potential loss in life years. Likewise the impacts of exposure on mental health and degenerative illnesses, involving major loss of well-being.
Regarding methodological advances, causal inference methods are evolving very rapidly. Over the past 20 years, interventional and quasi-experimental studies have been used to tackle very diverse health (and non-health) issues at a city, county, regional or national level, thanks to wider access to data. They usually proceed by exploiting exogenous variations in pollution due to weather (changes in wind, atmospheric temperature inversions, oceanic weather events), wildfires, variations in traffic (on air, road or sea), lockdowns or disruptive events (volcanic eruption, strikes, bans etc.). Combined with machine learning, they can help to reevaluate previous analyses, include more data, disentangle complex relationships (multi-pollutant mixtures, multiple exposures, multiple health effects) via intermediate variables (principal stratification and causal mediation), accounting for long-lasting and long-term effects. They can also tackle the complexity of the relationships between energy, transport, climate change, indoor pollution and (indirect) health impacts. For instance, higher temperatures and heatwaves are likely to have synergistic effects, especially on cardiovascular disease, and ozone seems to cause a fraction of heat-related deaths.
From an economic perspective, all the above advances will obviously improve economic assessments, which come at the final stage. In addition, standard health impact assessments frequently ignore the fact that some chronic diseases may not only be exacerbated but also be caused by AAP exposure. Acknowledging this would greatly increase the overall burden of disease found to be attributable to AAP. Re-evaluating policies or regulations (such as emissions trading schemes or transportation choices) in the light of methodological advances might improve decision-making. A more integrated approach jointly considering urbanism, AP and health would certainly be an interesting step toward tackling climate change issues (increase in temperature, flash floods) simultaneously with issues of health and well-being. A better understanding of the relationship between socioeconomic status and AAP exposure is also relevant. Disadvantaged populations, in addition to suffering higher exposure to AAP, might well be more sensitive or susceptible to the related health effects than the better-off.
To conclude, with improved knowledge, an interdisciplinary approach adopting the widest possible perspective on the issues at stake could be expected to offer a picture closer to the real AP-health relationship.
→ This article was issued in AMSE Newletter, Summer 2022.
© Yuriy Vahlenko, toa555, Viacheslav on Adobe Stock
