We use data from the Moving to Opportunity (MTO) experiment to estimate a model of neighborhood choice. The experimentally generated data is used for both estimation and out of sample validation. The experimental variation, which effectively randomizes the rents that households face, allows us to identify the model's structural parameters. The model replicates well the behavior of a random subset of households not used in estimation and experimentally exposed to different incentives. We then use the estimated model to simulate a set of alternative constraints on voucher use. We find that, relative to MTO, tighter location constraints substantially reduce take-up. As a result, average exposure to poverty actually increases under these restrictive vouchers. We also explored further limiting where households can move based on the racial composition of the destination neighborhood. Compared with the MTO experimental voucher, this policy further reduces take-up and, on average, ends up exposing households to the same neighborhood characteristics.