Capturing Crime at the Micro-place: A Spatial Approach to Inform Buffer Size

The current study develops a methodology to identify spatially relevant buffer sizes for micro-place evaluation research. It applies this methodology in an examination of the causal impact of demolitions on crime in Detroit, Michigan. We utilize Ripley’s bivariate K-function to guide our choice of buffer size. We select a buffer size as the distance at which the examined spatial features exhibit significant sustained attraction prior to the introduction of the intervention. We argue that buffers that are identified in this way capture the spatial relationship between environmental features and are therefore better-suited to capture the actual impact of the treatment on crime. We apply this knowledge in a synthetic control design that estimates the citywide effect of demolitions on disaggregated crime outcomes. With the exception of burglaries, we find fairly limited evidence of a strong, consistent effect of demolitions on crime. The largest negative effects were observed in the immediate months following demolition. Overall, the considerable uncertainty of our estimates suggests that the effect of demolitions may not be consistent across all neighborhoods. At the very least, demolition programs may help temporarily reduce burglaries in areas immediately around demolition sites. However, additional crime reductions gains may be possible if demolition efforts are coupled with complementary crime prevention approaches that focus on the restoration of vacant land. We hope future micro-place evaluation research will use and expand upon our buffer size selection protocol to help improve how places are understood and captured.