Abstract
I study how smoke from agricultural fires affects children’s health in India. Conventional instruments proxy pollution exposure with wind direction and distance to fires, but in agrarian districts both burning and winds are highly predictable, allowing families to avoid smoke and pushing causal estimates downward. I introduce NetFires, which tracks deviations from long-run fire–wind patterns to capture unexpected upwind exposure that is less vulnerable to anticipatory behavior. I link nationally representative NFHS-5 (2019–21) birth records, geocoded to mothers’ residence and conception timing, with NASA FIRMS fire detections, ERA5 wind fields, and high-resolution PM2.5 concentration maps. One additional upwind NetFire in the final trimester reduces birth weight by 0.14 grams (95% CI: −0.26g, −0.02g). The effect is strongest in the third trimester and remains empirically robust across alternative exposure definitions and fixed-effect structures. Heterogeneity tests show larger impacts among rural households, poorer wealth quintiles, and female infants, highlighting environmental inequality. Compared with conventional fire counts, the NetFires instrument delivers larger and more precise estimates and emphasizes the importance of modeling exposure uncertainty. Placebo exercises using non-upwind reclassifications, along with alternative IV designs and falsification checks, support internal validity and rule out confounders. Overall, results show that accounting for behavioral predictability matters for exposure modeling and for estimating pollution’s health impacts. NetFires provides a scalable, policy-relevant tool for identifying high-risk pollution shocks in agrarian settings. The findings inform climate-health policy, motivate development of early-warning systems, and support protective interventions targeted to pregnant women living in fire-prone regions.
Presenters
Hieu Thi Hoang NguyenStudent, PhD Candidate, CERGE-EI, Praha, Hlavní město, Czech Republic
Details
Presentation Type
Paper Presentation in a Themed Session
Theme
Human Impacts and Responsibility
KEYWORDS
Agricultural fires; Children's health; Air pollution; Instrumental Variable; India
