We build a general equilibrium, heterogeneous agent, incomplete markets version of a classic urban residential choice model. House prices and rents are non-linear functions of location and size. We examine why families with young children move and buy homes in suburban areas even though observationally similar rentals are relatively cheap.