Yann Algan, Camille Hémet, and David D. Laitin have a piece from the latest JPE, based on what appears to be very good data:
Relying on diversity measures computed at the apartment block level under conditions of exogenous allocation of public housing in France, this paper identifies the effects of ethnic diversity on social relationships and housing quality. Housing Survey data reveal that diversity induces social anomie. Through the channel of anomie, diversity accounts for the inability of residents to sanction others for vandalism and to act collectively to demand proper building maintenance. However, anomie also lowers opportunities for violent confrontations, which are not related to diversity.
A sentence from the conclusion explains that last bit more clearly: “…fractionalization has no effect on public safety, diversity being associated with social anomie within the housing blocks rather than violent confrontations among neighbors — helped as well by an increase in municipal policing in municipalities of high diversity.”
The paper also offers a useful but brief survey of what we know about ethnic diversity and social capital. Here is an earlier ungated copy (pdf).