Agent-based Modeling of Tax Evasion

Chapter 4Using Agent-Based Modeling to Analyze Tax Compliance and Auditing

Nigar Hashimzade and Gareth Myles

4.1 Introduction

The economic analysis of tax compliance has the objectives of explaining and predicting compliance behavior. The achievement of these objectives is essential for the design of beneficial interventions that increase the level of compliance and raise revenue. Several different research methodologies can contribute to this program of research. Theoretical analysis can develop models that are evaluated by empirical studies and tested using laboratory and field experiments. The potential value of applying these methods is constrained by the limited detail that can be incorporated within a theoretical model if useful insights are to be obtained, the potential lack of external validity of laboratory experiments, and the legal (and cost) constraints on field experiments.

Agent-based modeling provides a way of circumventing these difficulties. It makes it possible to obtain insights from a much richer model than can be used for a purely theoretical analysis, and permits experiments to be conducted that would not be possible in the field. This does not imply that the other methods are not equally valuable. Without economic theory we would not know how to specify the agent-based model; and without empirical and experimental research we would not know how to calibrate the model. It is our belief that the real benefits of agent-based modeling will only be realized ...