In the late 1960s, the United States Federal Government resorted to publically funded insurance systems to deal with two quite different problems: floods and riots. Both programs were administered by the same agency, both relied heavily on the spatial mapping of risk, and both were haunted by problems of moral hazard. Curiously, and most importantly, however, riots as well as floods were viewed as “environmental hazards” by the insurance industry and the government agencies involved. The underlying assumption was that social problems could be treated as quasi-natural hazards, i.e. as a homogeneous and unpredictable force that could be contained by actuarial means. Yet uprisings, civil commotions, and riots are not “acts of god” that are located outside of society (and neither are floods). This article discloses the origins of both programs, it describes their communalities and differences, and it reveals the views of those who were subject to racist steering practices.