Title:Income Tax Evasion Dynamics: Evidence from an Agent-based Econophysics Model

Abstract:We analyze income tax evasion dynamics in a standard model of statistical mechanics, the Ising model of ferromagnetism. However, in contrast to previous research, we use an inhomogeneous multi-dimensional Ising model where the local degrees of freedom (agents) are subject to a specific social temperature and coupled to external fields which govern their social behavior. This new modeling frame allows for analyzing large societies of four different and interacting agent types. As a second novelty, our model may reproduce results from agent-based models that incorporate standard Allingham and Sandmo tax evasion features as well as results from existing two-dimensional Ising based tax evasion models. We then use our model for analyzing income tax evasion dynamics under different enforcement scenarios and point to some policy implications.