Framing and Salience Effects in Tax Evasion Decisions – An Experiment on Underreporting and Overdeducting

Abstract

Tax evasion costs governments worldwide trillions of U.S. dollars. Thus, a comprehensive understanding of tax evasion behavior is needed to ensure that tax evasion is combatted effectively. Using different controlled and incentivized experiments, we analyze whether taxpayers are more willing to evade taxes by underreporting positive income (e.g., business or nonbusiness income) than by overdeducting negative income (e.g., deductions, credits, or losses). We robustly observe an asymmetric tax evasion behavior. Specifically, subjects are less compliant in case of positive income. We argue that this result can be explained by an asymmetric evaluation of tax payments and tax refunds in accordance with prospect theory and the income tax withholding phenomenon. However, in an experimental environment in which the interaction of positive and negative income reporting is made very saliently and in which we consequently expect that subjects decide on both tax evasion decisions jointly, the effect vanishes. We therefore provide evidence that 1) tax evasion behavior is asymmetrically in case of positive and negative income reporting and that 2) the salience of income interaction plays an important role in tax evasion decisions.

Keywords: Tax evasion, cheating, prospect theory, behavioral taxation, experimental economics

JEL Classification: C91, D14, H26