Causal effects of air pollution on child health : evidence from a low-pollution setting

Abstract

To what extent does air pollution in low-pollution settings affect children's health? Which children benefit most from further reductions, and what factors moderate this relationship? We address these questions using the universe of administrative medical records from the universal public healthcare system in Catalonia (Spain) between 2013 and 2017. We combine these data with spatio-temporal kriging techniques to construct complete time-by-location data on several air pollutants and environmental confounders. We then instrument for local PM10 concentrations-the main reference pollutant for air quality policies at the time-using variation in local wind direction in a multiple fixed effects model. Our primary outcome is respiratory-related healthcare visits, a measure of child morbidity. We find that even at relatively low ambient levels, increases in PM10 concentrations raise the incidence of respiratory-related visits. Our preferred instrumental variables estimate indicates that a 1 µg/m³ (or 4.5%) increase in PM10 leads to a 0.5% increase in overall respiratory visits, driven mainly by lower respiratory illnesses, which carry more serious health implications than other respiratory illnesses. We also find evidence of heterogeneous effects, with the youngest children (ages 0-5) and those exposed during hot or drier months being most affected. We estimate that the observed decline in PM10 concentrations during our sample period may have prevented approximately 16 million respiratory-related visits and saved around €800 million in direct healthcare costs. The results highlight the value of targeted public health interventions, particularly for young children and during periods of elevated environmental risk.