Daniel Reck is an Assistant Professor at the London School of Economics.
Daniel Reck will be presenting in person. Viewers may also attend via Zoom.
This paper studies tax evasion at the top of the U.S. income distribution using IRS micro-data from (i) random audits, (ii) targeted enforcement activities, and (iii) operational audits. Drawing on this unique combination of data, we demonstrate empirically that random audits underestimate tax evasion at the top of the income distribution. Specifically, random audits do not capture most tax evasion through offshore accounts and pass-through businesses, both of which are quantitatively important at the top. We provide a theoretical explanation for this phenomenon, and we construct new estimates of the size and distribution of tax noncompliance in the United States. In our model, individuals can adopt a technology that would
better conceal evasion at some fixed cost. Risk preferences and relatively high audit rates at the top drive the adoption of such sophisticated evasion technologies by high-income individuals. Consequently, random audits, which do not detect most sophisticated evasion, underestimate top tax evasion. After correcting for this bias, we find that unreported income as a fraction of true income rises from 7% in the bottom 50% to more than 20% in the top 1%, of which 6 percentage points correspond to undetected sophisticated evasion. Accounting for tax evasion increases the top 1% fiscal income share significantly