Displacement of crime or diffusion of benefit
Evidence from the New Deal for Communities Programme
This report looks at the effect on local crime rates of the New Deal for Communities (NDC). Specifically, is crime displaced from the NDC location into the surrounding area; or conversely is there a diffusion of crime reduction benefits into the surrounding area. The authors tentatively suggest that diffusion of benefits is far more likely than displacement of crime.
Title: Displacement of crime or diffusion of benefit: Evidence from the New Deal for Communities Programme
Authors: David McLennan & Adam Whitworth for Department for Communities and Local Government
Number of pages: 54
Date published: May 2008
Availability: Download full report 
PDF 500Kb
Key findings
There are many more instances of possible diffusion of benefit than of possible displacement of crime. This finding is consistent across the four different time comparison periods and the four different crime types examined.
Seventy-seven instances of a measurable reduction in crime were observed within NDC Partnerships across the Programme as a whole. In total, 383 individual buffer rings were therefore eligible to be tested for possible displacement or diffusion effects. It is apparent that 23% of these eligible buffer rings experienced possible diffusion of benefit compared to just 2% which experienced possible displacement of crime. The remaining 75% of eligible buffer rings exhibited changes in crime levels which were not suggestive of either possible displacement or diffusion.
There is remarkable consistency between the four crime types in the propensity for possible displacement/diffusion. Across each of the four crime types, between 21% and 25% of eligible buffer rings indicate possible diffusion while between 0% and 5% indicate possible displacement.
One or more instances of possible diffusion of benefit were observed around 24 different NDC Partnerships. Certain NDC areas appear to be associated with diffusion of benefit across multiple crime types. The findings from Tower Hamlets and Sandwell NDC areas suggest possible diffusion of benefit to surrounding buffer rings in three of the four crime types. Evidence from a further ten NDC areas suggested possible diffusion to surrounding areas in two of the four crime types and findings from a further 12 areas suggested possible diffusion on a single crime type.
There is evidence of a distance-decay function associated with instances of possible diffusion of benefit: diffusion is most likely to be observed in the areas immediately surrounding an NDC area and the likelihood of observing diffusion decreases with distance away from the partnership boundary. This finding is consistent across all four crime types and with criminological theory.
There is a degree of consistency in findings of potential diffusion when looking across different buffer rings within the same year and within the same crime type. In other words, an NDC area is more likely to be associated with diffusion in a number of buffer rings in the same year and for the same crime type than be associated with diffusion in a particular buffer ring over different time periods or between different crime types.
There is considerable variation in findings across the different time comparison periods indicating a lack of apparent temporal consistency in possible diffusion/displacement effects
Getting a copy
Last update: Monday, May 19, 2008