Simulating Change: a Systematic Literature Review of Agent-Based Models for Policy-Making

Social phenomena emerge from agent-environment interactions, rendering many statistical models unsuitable. Agent-based Models (ABMs) offer a viable alternative for exploring policy implications. While recent crises like the COVID-19 pandemic may have increased ABM awareness, their use in policy-making has a long history. To better understand the potential challenges and opportunities of using ABMs to inform policy-making, we conducted a systematic literature review and identified 34 articles describing the use of ABMs involving policymakers. This review revealed that ABMs have been implemented to support policymakers across a range of policy areas, but also identified low levels of model traceability and formal communication. Moreover, the review showed that the model's purpose and type tend to influence how validation is performed. The review concludes that models that have undergone little validation and lack proper documentation, while being informally communicated, may hinder policymakers from effectively motivating their decision-making.