This article aims to examine the existence of rule manipulation and moral hazard in the Brazilian Unemployment Insurance Program. For empirical analysis, the rule manipulation test by Cattaneo, Janson and Ma (2016) was used, as well as fuzzy and sharp regression discontinuity. The data was built using data from the National Employment and Unemployment Survey from January 2008 to June 2014 due to the greater homogeneity of the rules for benefit access. Based on the results, the program is an influence on the length of employment of Brazilian workers given the existence of rule manipulation, assessed by the length of stay in the last job. Furthermore, it was found that heads of families and their children were less likely to search for employment. This findings were corroborated when data from the program beneficiaries only was assessed, showing a lower job search probability, between-21.80 pp and-15.08 pp for the children, and between-39.40 pp and-28.50 pp for the heads of families. Thus, it is possible to confirm the existence of both rule manipulation the access of the program, as well as moral hazard, which points to the need to restructure the program, and above all, have less influence on the national labor market.