Abstract:
How can crime be disrupted effectively without increasing resources? To answer this question, we develop a spatial network model of crime diffusion, using London as a case study. Moving beyond traditional hot spot policing, we identify key player neighborhoods-highly connected areas in the network. Counterfactual analysis shows that targeting top 10% of key players reduces crime by 10.7% (5.8 percentage points) more than targeting top 10% of hot spots, resulting in potential annual savings exceed- ing £130 million. Examining the underlying mechanisms, we find that while hot spots attract crime locally, key players facilitate its propagation across areas.