Since 1970, mixed-income (inclusionary) housing projects have proliferated in the United States. In a community of this sort, only some of the dwelling units, perhaps as few as 10 to 25 percent, are targeted for delivery of housing assistance. Eligible households that successively occupy these particular units pay below-market rents, while the occupants of the other units do not. This article situates this innovation within the broader history of U.S. housing policy and evaluates its merits. I contend that the mixed-income project approach, while superior to the traditional public housing model, is in almost all contexts distinctly inferior to the provision of portable housing vouchers to needy tenants. Although prior commentators also have touted the voucher approach, I enrich their analyses by addressing more fully the social consequences of various housing policies that might be used to economically integrate neighborhoods and buildings. It has traditionally been thought that enhancing socioeconomic diversity within a neighborhood has unalloyed social benefits. Many recent social-scientific studies present a more complex picture and weaken the case for government support of mixed-income projects.