Association for Public Policy Analysis & Management

Recent public health crises have heightened attention to the adverse consequences of lead exposure and that the burden of lead exposure is shared disproportionately by socio-economically disadvantaged children. Children are exposed to lead in a variety of ways, one of which is through contaminated drinking water, which, in turn contributes to heightened blood lead levels (BLLs). The Environmental Protection Agency (EPA) has a water lead "enforcement level" set at 15 parts per billion (ppb); the Centers for Disease Control and Prevention (CDC) recommends action when children's BLLs exceed 5 micrograms per deciliter. Lead exposure and heightened BLLs have been linked to cognitive impairment and increased behavior problems in the short-term, and a higher risk of dropping out of high school and criminal arrests in the longer term. However, the existing literature is limited in two ways. First, much of the work relies on observational data where lead exposure may be confounded with other considerations. Second, it is unclear as to how high water lead levels need to be to contribute to BLLs higher than the CDC limit, as the threshold at which BLLs can negatively affect students' behavior is not well understood. Studies have found that children who consumed water at school that was contaminated with lead higher than the EPA "enforcement level" did not have blood lead levels (BLLs) that exceeded the CDC action level. In other words, a worryingly high level of lead in schools was not associated with a worryingly high BLL in children. Yet, other work has shown that consuming water with lead levels below the EPA action threshold is significantly associated with adverse outcomes. To contribute to the literature on how lead exposure might affect school achievement, I examine the link between exposure to slightly elevated levels of lead in school water fountains on school suspensions and absences. To do so I exploit within-school variation in lead exposure that is unlikely to be affected by selection biases. I link student-level administrative data with water-fixture level lead data recorded during school testing in Portland, Oregon in 2016 (n=56,994). I find that, on average, students consume water (from classroom or nearest hallway water fountains) with lead levels of 5.88 ppb (SD=8.18). Further, I find that level of lead in water is significantly positively associated with number of days suspended from school, and this finding is robust to controlling for receipt of free or reduced-priced lunch. I do not find a statistically significant relationship between water lead level and days absent. The current research suggests that low levels of lead consumed in school water fountains can be predictive of maladaptive behavior in school. One key limitation of this work is that it does not take into account children's BLLs. Future work should link school-level lead data, with children's BLL and behavioral outcomes.