Whether people of differing types can live happily together is one of the most important social and political questions concerning urban areas. From a variety of theoretical perspectives, such mixing seems extremely unlikely. While the theoretical result seems well supported in the context of race, the evidence for income mixing is much less stark. Compared to the strict segregation predicted by the models (and embodied in the context of race), Americans live in economically diverse neighborhoods. While this has lead to some further theoretical experiments, the stability of this mixing has never been addressed as an empirical matter. It would be naïve to look at cross-sectional snapshots of income mixing as representing stable situations, since neighborhood change is a prevalent feature of American urban economies. This paper sketches out the empirical implications of slow transition towards the predicted equilibrium, and tests those implications. It is the first paper to directly evaluate the persistence and stability of mixed-income communities. The results are supportive of the three models of income segregation: income mixing appears to be unstable, although the adjustment process is slow. This work is of especial importance due to the focus mixed-income communities receive in the urban planning and policy.