We develop a multi-community urban land use framework to investigate the implications of increasing school choice opportunities on educational and residential choices of a city's residents. When deciding on the location and the size of land, the households care about the distance to the business district, and a local public good: education. There is a private education alternative that breaks the link between choosing a residence area and choosing a school. The households differ in their incomes and preferences for education. In five models that differ in various aspects of choice and financing, we study the housing and education choices of the city residents, and the endogenously determined education provision levels in equilibrium. The results of the article support reformist arguments: We ¯rst show that the presence of a private alternative benefits every household, whereas school district consolidation hurts everyone. We then examine two policies that aim to increase choice. An untargeted local government support (financed by property taxes) that can be used at the private school can improve things for talented poor. A policy that supports the talented poor (using city income taxes) with funds that can be used for public as well as private schools can also improve welfare of all talented students, rich or poor.