Estimate of bunching at the disclosed Sector Study threshold
June 1, 2021
with E. Di Gregorio and E. Sartori
Tax authorities typically concentrate enforcement resources on large businesses, yet small business evasion generates tax revenue losses that often exceed those from multinational profit shifting. We show that authorities can improve small business compliance by strategically disclosing audit-relevant information. We study audit rules that inform taxpayers that audit risk drops discontinuously above a threshold based on predicted revenues. Under empirically plausible conditions, our model shows that the tax base is concave in the size of this discontinuity. Consequently, if widening an existing discontinuity raises the tax base, the pre-reform disclosed rule must already outperform a flat, undisclosed benchmark. We test this implication using more than 26 million tax files (2007–2016) from Italy’s Sector Studies. Taxpayers bunch sharply at the threshold, and bunching correlates with evasion proxies and evasion technologies. Exploiting a staggered reform that increased the audit risk discontinuity, we find that compliance rises below the threshold and falls above it, yet average reported profits, the relevant tax base, grow by 16.2% in treated sectors over six years. Our theoretical result therefore implies that disclosure outperforms nondisclosure. Structural estimates further show that strengthening incentives at the threshold could more than halve the audit budget without reducing compliance, freeing resources to target larger firms, and that the reform brought the policy close to the optimum.