A striking negative correlation exists between an area's residential racial segregation and its population characteristics, but it is recognized that this relationship may not be causal. I present a novel test of causality from segregation to population characteristics by exploiting the...