Conference Agenda

Despite extensive evidence on air pollution's health effects, surprisingly little is known about whether individuals actively avoid pollution exposure. This constitutes a critical open question, since behavioral responses both attenuate estimated health impacts and represent distinct welfare costs. In this paper, I study avoidance behavior in response to short-term air pollution shocks using high-frequency mobility data from Spain. Combining a unique anonymized mobile phone–based mobility dataset for more than 2,700 municipalities with ground-level PM10 measurements and detailed weather controls, I estimate the causal effect of air pollution on mobility patterns. To address endogeneity, I exploit exogenous variation in PM10 generated by Saharan dust intrusions (calima) as an instrumental variable. The unique characteristics of the data allow me to disaggregate trips by distance and destination type (home, work, frequent vs. infrequent locations). The results show that increases in PM10 lead to modest reductions in short-distance mobility, partly offset by increases in medium-distance trips. Heterogeneity by destination type shows particularly strong increases in trips to infrequent locations, suggesting adjustments consistent with avoidance behavior rather than purely health-driven constraints. Importantly, my results imply smaller effects than reported in the previous literature, which exploits salient variation in air pollution in China or due to wildfires. My findings highlight the importance of accurately accounting for behavioral responses when estimating the welfare costs of air pollution and its health impacts.