Full Citation
Title: Estimating Avoidance Behavior From Human Mobility
Citation Type: Dissertation/Thesis
Publication Year: 2023
ISBN:
ISSN:
DOI:
NSFID:
PMCID:
PMID:
Abstract: Air quality in the United States has improved significantly over the past few decades. According to the Environmental Protection Agency (EPA), total emissions of the six principal air pollutants (carbon monoxide, lead, nitrogen oxides, ground -level ozone, particulate matter, and sulfur oxides) fell by 78% between 1970 and 2020 (US EPA, 2016). However, compared to countries like New Zealand, Sweden, Australia, and Canada, the United States still has much to do with regard to pollution management or reduction (McCarthy, 2020). As pollution control has increasing marginal costs, the cost of reducing pollution in a country like the United States, where pollution levels have been brought down significantly, becomes more and more expensive with each additional effort aimed at reducing pollution. Despite the level of spending on air pollution, exposure to pollution in the US is still quite high. The American Lung Association reported in a 2021 report that four in ten Americans live where pollution levels frequently make the air too dangerous to breathe (American Lung Association, 2021). The amount of spending on air pollution mitigation pales is negligible compared to the cost of the effects of air pollution. The effects of air pollution are very costly, as they directly affect human wellbeing and environmental conditions (Santana et al., 2020; Chen & Chen, 2021; Delucchi et al., 2002; Gouveia & Fletcher, 2000). Between 2010 and 2020, mortality from exposure to ozone, and from exposure to PM.2.5 among adults were expected to increase by about 65% and 44%, respectively (US EPA, 2011). Adverse health conditions are not the only negative effect of air pollution. Other effects of exposure to air pollution can be seen in reduced visibility, restricted activity, and increased social misbehavior (Lu, 2020; Hyslop, 2009; Yan et al., 2019). 2 It is important to know the full costs of air pollution so that governments (federal and state) can make policies whose benefits will exceed those costs. It is hard to measure the full cost of air pollution as some of the effects of air pollution are not as easily observable and quantifiable as health effects. One such effect of air pollution is the change in behavior to mitigate exposure to air pollution. These changes in behavior are important of themselves as they represent a loss in utility because people give up enjoyable activities in order to avoid being exposed to pollution. Behavioral changes are also important because they indirectly affect the cost of health effects of pollution by reducing the probability of getting sick from exposure to pollutants. This type of behavior is usually referred to as avoidance behavior or averting behavior. Avoidance behavior with regards to air pollution constitutes any behavior that enables a person to stay clear of or protect themselves in environments that could potentially expose them to air pollutants. It may include staying indoors on days with high levels of air pollution, avoiding highly polluted areas, using protective coverings such as face masks, and in the longer term, migrating from highly polluted areas to less polluted areas. Estimates of the cost of the health effects (and the total costs) of air pollution that do not take into account the cost of these behavioral adjustments are biased downwards because these actions affect the frequency and severity with which people’s health are affected by pollution. The estimates from this study measure the causal effects of air quality and mobility. Measures of air quality have been found to be endogenous in the existing body of research about the relationship between air quality and human mobility. To control for this endogeneity, I employ a control function model which corrects the endogeneity problem by modelling the endogeneity in the error term. Though similar in many ways to a two-stage least squares estimation, the control function method is more advantageous as it is compatible with more model specifications (including non-linear models) and allow for heterogeneous effects. Thus, the control function method allows me to test the robustness of my results to different model specifications without changing the underlying statistical method. To the best of my knowledge, this is the first research work that employs the use of the control function method in correcting for endogeneity in the relationship between air quality and human mobility. This study finds evidence of avoidance behavior in the United States as shown by a 1.4 percentage point increase in time spent at home, a 4.2 percentage point decrease in time spent away from home, and a 4.5 percent reduction in distance traveled from home by a device on average on days of unhealthy air quality. These translate to an 8-minute increase in time spent at home, a 10- minute reduction in time spent away from home and a 385-meter reduction in distance traveled from home on days that the EPA reports unhealthy air quality compared to days on which good air quality is reported. The study uses observed data from across the entire country and thus, findings are more reliable and generalizable than others in the literature. Results from heterogeneity analyses done as part of the study show variations in response to air pollution among different demographic groups. Minority communities, for example, are shown to be more responsive to changes in air quality than non-minority communities. Knowledge of the extent to which avoidance measures are effective in reducing exposure to air pollution is important to policy makers, as it could serve as the basis for the development of policies geared towards making individuals play more active roles in reducing the adverse effects of pollution. If avoidance behavior reduces exposure to pollution, then it might be prudent to encourage people to practice these behaviors while the government finds more efficient ways to reduce pollution levels. It must be noted, however, that avoidance behavior has many costs aside from the cost of implementing information systems to make people aware of the levels of pollution. As there has been a general failure to limit pollution (in the form of greenhouse gases, for instance) globally, finding a way to measure these costs is important for governments to make decisions on whether to make environmental policy more focused on mitigation and adaptation practices. Avoidance behavior, which is the focus of this paper, is a type of mitigation practice. As such, the findings from this paper contribute valuable knowledge to the decision-making process in matters concerning environmental policy. Also, the finding that different groups respond differently to air pollution should emphasize to policymakers the importance of considering vulnerable populations when making policies related to encouraging avoidance behavior.
User Submitted?: No
Authors: Addai, Nana Boakyewaa
Institution: The University of North Carolina at Greensboro
Department: Economics
Advisor:
Degree:
Publisher Location:
Pages: 1-111
Data Collections: IPUMS NHGIS
Topics: Health, Natural Resource Management
Countries: