ABSTRACT
We develop a theory of optimal networks in the context of criminal organizations. In this framework the criminals choose their network links with others according to a set of specified costs and benefits to participation. The optimal number and configuration of links within each network is solved for a set of 10,000 parameter simulations specifying the direct cost of links between agents, the benefit to connections, and the cost of being in the network with others. In addition, agents determine the size of the optimal network. This framework allows consideration of a variety of crime policy scenarios. In particular, removing the ‘key player’, the best strategy when the network is exogenous, may not be the optimal strategy in an environment in which the agents can change the size and structure of the network endogenously. More generally, optimal crime policy may be different if the criminals are aware of the policing strategy and can alter their network.
