Beyond Violations: Measuring Compliance in Drinking Water Regulation

Ensuring safe drinking water depends not only on infrastructure but on effective regulatory compliance. This paper studies compliance with drinking water quality regulations through a three-layer framework that distinguishes among whether providers report monitoring data at all, whether those that do report collect the required number of samples, and whether the samples collected reveal violations of regulatory quality thresholds. Because each layer is a necessary condition for the next, failures in reporting and sampling can prevent official statistics from accurately reflecting underlying water quality risks. Using comprehensive administrative data from Brazil covering more than 5,500 municipalities between 2010 and 2022, we show that failures at each compliance layer are widespread, non-random, and systematically correlated with municipal characteristics. Non-reporting and under-sampling are concentrated in poorer and lower-capacity municipalities, while over-sampling is prevalent among larger providers. Over-samplers report fewer violations than municipalities that comply exactly with the prescribed sample count, a pattern that admits several interpretations and that the administrative record alone cannot adjudicate. The reported violation share is therefore jointly determined with sampling behavior, and this joint determination is systematically related to the municipal characteristics that also predict contamination risk. Treating reported violations as the sole measure of regulatory performance yields a misleading picture of drinking water safety and obscures the importance of monitoring compliance with reporting and sampling requirements alongside quality thresholds.