Why Don’t Poor Families Move? A Spatial Equilibirum Analysis of Parental Decisions with Social Learning

In the United States, less-educated parents tend to choose lower levels of parental inputs, they reside in bad neighborhoods and allocate little time to parent-child activities. I propose a spatial overlapping generation model of parental decisions about time and neighborhood quality with imperfect information and social learning. Specifically, young agents learn about the relevance of parental inputs through observing their neighbors. Crucially, however, they are prone to misinferences as they may not be able to perfectly correct for selection induced by income segregation. I calibrate the model using several United States representative datasets. The calibrated model matches targeted and non-targeted parental behavior moments across socioeconomic groups. I find a relatively modest level of parental delusion that increases inequality by 3% (income Gini index) and social immobility by 12% (intergenerational rank-rank). A housing voucher policy improves the neighborhood quality of eligible families, raising children’s future earnings. When scaling up the policy, long-run and general equilibrium responses in parental beliefs amplify the policy effects. Inequality reduces, and intergenerational mobility improves.