The Asian Games, Air Pollution and Birth Outcomes in South China: An Instrumental Variable Approach

Thumbnail Image
Penn collection
Population Center Working Papers (PSC/PARC)
Degree type
ambient air pollution
preterm births
instrumental variable
Demography, Population, and Ecology
Environmental Public Health
Family, Life Course, and Society
Maternal and Child Health
Place and Environment
Social and Behavioral Sciences
Liu, Behrman and Hannum acknowledge funding for this project from the National Science Foundation (No:1756738) and the China Research and Engagement Fund and School of Arts and Sciences Dean’s Integrative Global Inquiries Fund, both at the University of Pennsylvania.
Grant number
Copyright date
Related resources

We estimate the causal effects of air pollution exposure on low birthweight, birthweight, and prematurity risk in South China, for all expectant mothers and by maternal age group and child sex. We do so by exploiting exogenous improvement in air quality during the 2010 Guangzhou Asian Games, when strict regulations were mandated to assure better air quality. We use daily air pollution levels collected from monitoring stations in Guangzhou, the Asian Games host city, and Shenzhen, a nearby control city, between 2009 and 2011. We first show that air quality during the Asian Games significantly improved in Guangzhou, relative to Shenzhen. Further, using birth-certificate data for both cities for 2009 to 2011 and using expected pregnancy overlap with the Asian Games as an instrumental variable, we study the effects of three pollutants (PM10, SO2, NO2) on birth outcomes. Results show that 1) air pollutants significantly reduced average birthweight and increased preterm risk; 2) for birthweight, late pregnancy is most sensitive to PM10 exposure, but there is not consistent evidence of a sensitive period for other pollutants and outcomes; 3) for birthweight, babies of mothers who are at least 35 years old show more vulnerability to all three air pollutants; and 4) male babies show more vulnerability than female babies to PM10 and SO2, but birthweights of female babies are more sensitive than those of male babies to NO2.

Date Range for Data Collection (Start Date)
Date Range for Data Collection (End Date)
Digital Object Identifier
Series name and number
Publication date
Volume number
Issue number
Publisher DOI
Journal Issue
Recommended citation