Tax Compliance as an Evolutionary Coordination Game: An Agent-Based Approach

Abstract

Tax reporting compliance by small business owners is modeled in an agent-based framework using concepts and methods based on evolutionary dynamics. A business owner’s ‘‘fitness’’ is a function of net after tax (and post-audit) income. Business owners exhibit heterogeneous tax morale and compliance propensity following four stochastically assigned behavioral ‘‘archetypes’’: Honest, Strategic, Defiant, and Random. The model is calibrated to observations from laboratory experiments and taxpayer random audits. The calibrated model is used to simulate evolutionary changes in a static population of 10,000 small business owners. A simulation using realistic parameters for the probability of audit and penalty rate finds that after fifteen time periods, the initial number of Honest business owners declines by approximately one-third and are displaced by proprietors having either Defiant or Strategic compliance behaviors.