Application of Agent-based Simulation to Policy Appraisal in the Criminal Justice System in England and Wales

This paper describes an agent-based approach for constructing a model of criminal justice system operations in England and Wales. The primary purpose of the model is to assess the impact of policy variants across the entire criminal justice system. Because of the structure of this system, three separate government departments interact and deliver services. Decisions in one area of the criminal justice system can be crucial in determining what happens in another area. Our purpose was twofold. First, we needed to contribute to the Treasury’s spending review by working with different groups in criminal justice agencies to reach a consensus on how things actually occur (i.e., linking behavior and actions of one group with another and with resources). Second, we needed to produce a model of the entire criminal justice system that would provide insights into questions related to capacity, case flow, and costs. We also needed to model the ways in which individuals go through the system. The result is a hybrid model that combines a simple system dynamics approach with an agent-based model. The distinctive approach used in this work integrated modeling with practical ways of enabling people to engage in strategic policymaking, while taking into account the complexities of the criminal justice system. The agent-based framework developed to meet these needs models the criminal justice system, provides the ability to assess policy across the system, and allows sharing of model output to improve cooperative efforts among departments.