Statistical Discrimination with Neighborhood Effects: Can Integration Eliminate Negative Stereotypes?

We introduce neighborhood effects in the costs of human capital acquisition into a model of statistical discrimination in labor markets. This creates a link between the level of segregation and the likelihood and extent of statistical discrimination. As long as negative stereotypes persist in the face of increasing integration, skill levels rise in the disadvantaged group and fall in the advantaged group. If integration proceeds beyond some threshold, however, there can be a qualitative change in the set of equilibria, with negative stereotypes becoming unsustainable and skill levels in both groups changing significantly. This change can work in either direction: skill levels may rise in both groups, or fall in both groups. Which of these outcomes arises depends on the population share of the disadvantaged group, and on the curvature of the relationship between neighborhood quality and the costs of human capital accumulation.