ABSTRACT
The paper examines the relationship between the supervision carried out in the municipalities by the main Brazilan supervisory institution (Controladoria Geral da União-CGU, in portuguese) and the incidence of dengue cases in them. Since the audited municipalities were randomized, this allows the identification of a control group that adequately represents the counterfactual of the treated group. The sample was composed of all municipalities that could be selected for that CGU inspection cycle, that is, 1520 municipalities, of which 70 were drawn and therefore belong to the study treatment group. We identified a negative effect of the policy on the incidence of the disease. However, when we consider a model with lags, we note that this initial impact from the drawing did not persist throughout the year of the inspection. Our analysis suggests that when federal resources are monitored, municipalities reduce irregularities, which may contribute to a decrease in the number of dengue cases.