Deterrence Effects of Auditing Rules: An Experimental Study

This paper experimentally examines a new auditing rule termed the bounded rule, which takes into account the budget constraint of the auditor (e.g., a tax authority). Compared to a traditional rule that audits income reports with a constant probability, the bounded rule can induce the same deterrence e¤ect with a smaller budget. The basic setting follows a classic tax-compliance game in which each taxpayer receives either high or low income with certain probability. On knowing an auditing rule, the taxpayers have to decide simultaneously and independently whether to report their income truthfully to the auditor. The traditional rule audits every low-income report with a constant probability. The bounded rule audits a randomly selected sample of low-income reports whenever the number of these reports exceeds the maximum number of audits allowed by the budget, or otherwise all of the low-income reports. The experimental evidence suggests that, as predicted, the two auditing rules have the same deterrence e¤ect. The bounded rule needs a smaller budget ex-ante, and conducts fewer audits ex-post. The results provide support for the bounded rule as a more cost-e¤ective alternative to the traditional rule.