We analyze the impact of a recent initiative by the Australian Government to reduce disadvantage and improve children's welfare in Aboriginal communities. The policy - known as income management - quarantines 50 percent of welfare payments to be spent on priority goods (e.g., food, housing, education) and not on socially harmful goods (drugs, pornography, gambling). Our focus is on childrens school attendance, which is a precise, high-frequency measure of community functionality and a key policy objective. We identify the causal impact of income management on attendance rates by exploiting exogenous variation in its staggered rollout across communities. We find no evidence that income management increased attendance. Rather, the introduction of income management reduced attendance by 2.7 percentage points (4 percent) on average in the first five months after which attendance eventually returned to its initial level. The attendance penalty is similar for boys and girls, but is larger for secondary school students and students with a tendency to attend school regularly. Exploring the potential mechanisms, we show that income management did not significantly affect student enrollments or mobility patterns into and out of Aboriginal communities. Nor are our results explained by confoundedness with other policy initiatives. Instead, we find that the attendance penalty associated with the introduction of income management is virtually zero after the adoption of more exible administrative arrangements suggesting that implementation issues may be responsible for the temporary reduction in school attendance that we observe.