The Effect of Air Pollution on Infant Health
The passage of the landmark Clean Air Act in 1970 launched an era of tightening standards for air pollutants. This process continues today and can be quite contentious. For example, the Environmental Protection Agency (EPA) issued new standards for ozone and particulate matter in 1997, but they were held up by legal challenges until a Supreme Court decision in 2001. Just last month, the EPA moved towards enforcing the new standards by announcing the names of 500 counties -- containing more than half the US population -- that violate or contribute to violations of the new ozone standards.
One of the primary motivations for stricter standards is to prevent pollution-related illness and premature mortality. Yet there is still much to learn about the effects of air pollution on health. Researchers Janet Currie and Matthew Neidell explore one aspect of this subject in Air Pollution and Infant Health: What Can We Learn from California's Recent Experience? (NBER Working Paper 10251).
The authors concentrate on infants because there is significant scientific uncertainty about the health effects of pollution for infants and a strong interest in protecting these vulnerable members of society. The authors focus on the recent experience of California because the pollution levels are lower than those examined in many past studies and thus more relevant to the current debate over appropriate pollution levels.
The authors use the California Birth Cohort files and the California Ambient Air Quality Data for 1989 to 2000. From these sources, they construct a data set containing information on infant outcomes such as low birth weight or mortality, on the mother's background and use of prenatal care, and on the level of four criteria pollutants in the mother's zip code -- ozone, carbon monoxide, particulate matter, and nitrogen dioxide - for 70% of the births in the state over this period.
This rich data set allows the authors to rely on changes in pollution levels in a given zip code over time to identify the effects of pollution on health. The authors include the mother's zip code in the model to control for time-invariant factors such as poverty that are geographically concentrated and may be associated with poor infant outcomes. They also include variables such as the mother's education to control for individual differences between mothers that may affect birth outcomes.
The authors' findings differ from some of the previous epidemiological literature. For example, the authors find little average effect of prenatal pollution exposure on the probability of low birth weight, short gestation, or fetal death after including the mother's zip code in the model. However, they do find that living in a very high-pollution area is associated with a higher risk of fetal death, suggesting that pollution may be harmful above a certain threshold level.
By contrast, the authors find significant effects of carbon monoxide and particulate matter levels on infant mortality. In the authors' preferred estimates, reductions in these two pollutants during the 1990s are estimated to have saved over 1,000 infant lives in California. Assuming a $4.8 million value per life saved, these health benefits would be valued at $5.1 billion. As the authors note, these estimates do not incorporate other potential improvements in infant health, such as reduced respiratory disease, and so are lower-bound estimates of the total health benefits of pollution reduction to infants.
This research was supported by the Center for Health and Well-Being at Princeton University and by the Center for Integrating Statistical and Environmental Science at the University of Chicago. It was summarized by Courtney Coile.