Association for Public Policy Analysis & Management

The impacts of pollution on health outcomes, and in particular infant health outcomes, have drawn the attention of economists. Studies have shown that infant health, typically measured by birth weight, can predict educational outcomes such as test scores, and there is some evidence that birth weight can also predict socioeconomic status and labor market outcomes in adulthood. Moreover, there is much evidence that infant health can be affected by conditions while in utero, such as ambient pollution due to local economic activity. Most studies of pollution and infant health have focused on the effects of air quality, with only limited research on the effects of water quality. To fill this gap in the literature, we study the relationship between drinking water contamination and infant health in Pennsylvania, using data on births and water quality from 2003 through 2014.

Our methodology builds on the within-mother approach pioneered by Currie et. al. (2013), who focused on New Jersey over the period 1997-2007. First, we spatially link maternal addresses from birth records to public water systems and assign water quality measures during the gestation period to each birth. We use the resulting data to estimate the relationship between drinking water violations and birth outcomes. We then expand our water quality measures beyond Safe Drinking Water Act Maximum Contaminant Level (MCL) violations to Pennsylvania Drinking Water Reporting System data that contain water quality sampling results. This allows us to observe water contamination levels that do not reach the regulatory threshold, and also allows us to assess whether contaminants detected reach health-based guidelines, which are often less than the MCLs, and therefore, more stringent. We also link births across mothers and assess whether moms are moving in response to drinking water contamination.

We find that drinking water contamination, when measured using public water system water quality violations, has effects on low birth weight that are similar to those in Currie et. al. (2013). However, our estimated effects on maternal mobility and pre-term birth are different from their findings. We find that maternal mobility is much more responsive to drinking water violations, and we uncover a larger and statistically significant negative effect of violations on pre-term birth. Moreover, in contrast to Currie et. al. (2013), births that are exposed to an MCL violation in our sample are more likely to be from mothers that are less disadvantaged by multiple demographic measures. We find that models using direct water quality sampling data yield more precise estimates and the richness of this data allows us to explore effect heterogeneity by trimester and type of contaminant.