A Competitive Audit Selection Mechanism with Incomplete Information

The experimental tax and regulatory compliance literature has shown the effectiveness of competitive audit selection mechanisms (ASMs) based on declarations and a signal of the taxpayers' actual income. However, collecting information about actual income prior to audit selection is costly. In this article, we test the effectiveness of an endogenous ASM based solely on declared income. We show theoretically and in a laboratory experiment that this new endogenous ASM significantly increases compliance in comparison with an ASM where all taxpayers face audit with equal probability. However, a further consequence of conditioning solely on declared income is that poorer taxpayers are audited more frequently, reducing the effectiveness of this ASM in generating revenue and reducing inequality. We further compare the new mechanism with an ASM that also uses a noisy signal of actual income and show that it is a significant improvement over the other two ASMs in terms of compliance, revenue, and inequality. Our results suggest that ASMs that condition only on reported income can increase compliance but should be implemented with caution, and investing in acquiring information before audit selection can have substantial benefits.