Assessing the Impact of PFAS Water Regulation

This paper evaluates the drinking water quality impacts of stricter state regulations for PFAS contaminants. A key challenge of this analysis is the wide prevalence of interval censoring in PFAS testing. I first show that without further assumptions, treatment effects are not generally identified and applying standard approaches that ignore this issue lead to qualitatively different results. I overcome this problem using a parametric approximation to recover the latent cumulative distribution function through censored maximum likelihood. I implement the changes-in-changes approach using the recovered distribution of concentration to estimate the impact of tightening regulatory standards on water quality. On the intensive margin, I find that notification level changes had no impact on water quality across the distribution. On the extensive margin, I find qualitative evidence of investment in the form of new or retrofitted treatment plants to address PFAS after notification and response levels became more stringent. I find these new treatment plants tend to be located in urban and above average income counties. The methods implemented allow for policy evaluation under interval censoring in a variety of contexts.