Taxpayer’s Behavior: From the Laboratory to Agent-Based Simulations

Recalling the materiality thesis by Guala (2002) and extending the mediation approach by Guala and Mittone (2005), we address the issue of external validity of tax experiments. The gap between experimental systems and the real domain of application is bridged by simulations involving human-calibrated agents: laboratory observations make simulated agents resemble real decision makers at a material level; simulations enrich the experimental environment with a more formal correspondence to reality. In order to analyze the synergy between tax experiments and agent-based simulations, three approaches are identified: (i) models focusing on the macro dynamics among heterogeneous behavioral types identified in the laboratory; (ii) models analyzing micro behavioral patterns, observed in the laboratory and uncovering psychological drivers; (iii) models involving both micro and macro perspectives. This analysis provides a possible guidance for the adoption of a human–agent combination to a policy purpose.