Abstract
We explore the role of social capital in explaining patterns of rural larceny and burglary crime rates. We find consistent evidence that higher levels of social capital tend to be associated with lower levels of rural property crime rates. We also find that there is significant spatial heterogeneity in the underlying data-generating process. This spatial heterogeneity suggests that relying on global estimates from classical statistical methods, such as least squares, may lead to erroneous policy recommendations at the local level. We suggest that some of the inconsistencies in the ecological empirical criminology literature might be explained by spatial heterogeneity.