We demonstrate that police effectiveness can be improved without additional investment of resources by targeting the "key players'' -- that is, the most influential nodes that drive the diffusion of crime in the network.

We apply the method developed by Ballester et al (2006)  to the case of London, focusing separately on property and violent crime. We first estimate a spatial correlation fixed effect model using panel data on crime at the neighborhood level from 2013 to 2019 and use the prediction of the model to derive the neighborhoods that are key players in crime. We then simulate the potential crime reduction achievable by targeting key player neighborhoods and compare it with the alternative policy that focuses on hot spot neighborhoods. 

Our simulations demonstrate that the key player policy could result in an 11 percentage point reduction in total crime. Finally, by utilizing existing estimates of the cost of crime, we quantify the financial benefit of this policy shift.

For online participation, please register at the link below!