This paper quantitatively evaluates the hypothesis that the housing bust in 2007 decreased geographical reallocation and increased the dispersion and level of unemployment during the Great Recession. We construct an equilibrium model of multiple locations with frictional housing and labor markets. When house prices fall, the amount of home equity declines, making it harder for homeowners to afford the down payment on a new house after moving. Consequently, the decline in house prices reduces migration and causes unemployment to rise differently in different locations. The model accounts for 90 percent of the increase in geographical dispersion of unemployment and the entire decline in net migration. However, despite large effects on migration and geographical dispersion of unemployment, the effect on aggregate unemployment is moderate: Our findings suggest that, absent the housing bust, aggregate unemployment would have been 0.5 percentage point lower.