My dissertation used address-level data on housing voucher holders in Baltimore to examine the relationships between residential sorting patterns and program design. I used multilevel models to look at how residential trajectories differed among household who received different types of housing assistance, finding large differences between recipients of different types of vouchers, and also among different racial groups. Following, I used disaggregate discrete choice models to study residential sorting patterns, finding that voucher holders are sensitive to both dwelling unit characteristics, and neighborhood characteristics, but that the strongest sorting factor, by far, is neighborhood racial composition.