A possible cost of diversity?

Professor Joanna Shepherd (Clemson) presented "Racial Diversity, Residential Segregation, and Crime: An Industrial Organization Analysis of Racial Competition" at N. C. State. She finds that, other things equal, more racially diverse areas have more crime. From her conclusion:

My econometric analysis of counties from 1990-1999 and metropolitan areas in 1980, 1990, and 2000 finds that both diversity and segregation increase crime. Moreover, tests of the combined effect of diversity and segregation reveal that segregation worsens diversity’s effect on crime. My results are robust to many alternative specifications. Moreover, tests confirm that my diversity measures are not proxying for racial groups that disproportionately commit crime. Nor are the results caused by any potential endogeneity between crime and diversity. Finally, my estimations are designed so that my racial diversity measure is not picking up other types of diversity, such as income diversity or religious diversity, which could increase crime.

The results from my econometric analysis confirm the predictions of my industrial organization theory of racial competition. The theory suggests that diversity increases both inter and intra-racial crime as racial groups compete. Segregation sharpens the competition.

My results in no way establish that diversity is bad; diversity provides countless benefits such as awareness of other racial groups and an intermingling of cultures. However, this paper, for the first time, focuses on one of diversity’s costs: increased crime. To determine the optimal amount of diversity in a region, one must weigh diversity’s numerous benefits against the costs.

She hypothesizes that a primary means by which diversity affects crime is provision of local goods. Areas that are more diverse tend, other things equal, to provide less public goods.