Up Next

Market-Based Solutions to Vital Economic Issues


Building bridges between tax scholars, policymakers and practitioners
News & Media

New Research Suggests IRS Underestimates Tax Evasion by Wealthiest Filers

Write-Off: The Tax Blog

At a recent hearing of the Senate Budget Committee styled “Ending a Rigged Tax Code,” Chairman Bernie Sanders opened proceedings with a list of familiar economic “absurdities” and “obscenities,” as he called them, stating among other things that Jeff Bezos and Elon Musk alone own as much wealth as the poorest 40% of Americans combined. There was one new obscenity for the list: the claim that the wealthiest 1% of Americans are evading about 20% of their tax obligation, amounting to hundreds of billions annually. This figure comes from a recently published paper by economists John Guyton, Patrick Langetieg, Daniel Reck, Max Risch, and Gabriel Zucman. Zucman in particular has been a prominent voice in academic discussions around the precise scale of various measures of income and wealth inequality, most notably with his UC Berkeley colleague Emmanuel Saez, and Zucman was on hand to testify before the Committee.

The ambitious and prolific Zucman has chosen a set of elaborate problems significant for forming socioeconomic and tax policy and requiring the parsing of varied and technically complex data. Wading into these areas has occasionally brought Zucman onto shaky ground. A recent article by scholars of accounting Jennifer Blouin (UNC Ph.D., 2004)  and Leslie Robinson (UNC Ph.D, 2006), for example, indicates that economists writing on the subject of base erosion through foreign profit shifting, among whose efforts Zucman’s are well represented, have significantly overestimated the extent to which multinational corporations have siloed their profits in tax havens due to a misunderstanding of the technical aspects of financial reporting and the intricacies of Bureau of Economic Analysis data.

Ironically, one of the more important and secure contributions of Zucman’s work is in pointing out the blind spots of established methodologies in seeking to make sense of intentional wealth and income obfuscation. In this most recent paper, Zucman and his co-authors point out that official IRS and academic estimates of tax evasion have utilized the results of a large-scale program of random audits without adequately considering these audits’ efficacy. According to the traditional approach, observed tax evasion, even when adjusted for an internally developed measure of auditor skill, declines sharply within the top 0.5%, particularly as one approaches the wealthiest taxpayers, especially the top .1% and .01%. The embedded assumption here is that the randomized audit program, when adjusted for auditor skill, leads to discovery of substantially all evasion, or is equally deficient across the population of taxpayers. Guyton, Langetieg, Reck, Risch, and Zucman challenge this assumption.

Using audit data collected just prior to a crackdown on unregistered foreign holdings approximately a decade ago that involved a partial amnesty leading some taxpayers to declare holdings for the first time, Guyton et al show that the limited approach utilized in random audits was highly ineffectual in uncovering non-compliance by especially wealthy individuals. (In only a small percentage of cases did individuals register accounts that had been caught by the audit process.) They infer that this episode is instructive of the present, and that the FATCA regime has been substantially insufficient on the basis of estimates of evasion found in previous studies. Specifically, on the basis of Zucman’s earlier work, individual and with Annette Alstadsæter and Niels Johannesen, that utilized varied inputs including macroeconomic statistics and inadvertently disclosed financial records, the authors provide an estimate of the scope of the international evasion. Additional evasion via pass-through entities, which the authors find to be more significant in magnitude, is estimated by analogy to sole proprietorship income and general investment income by individuals as well as reference to a brief audit program conducted between 2003 and 2004.

In general, the explanation the authors offer for evasion among the wealthiest reduces to auditors’ inability to adequately counter sophisticated evasion techniques involving foreign jurisdictions and complex corporate structures. The best evidence of this comes very early in the paper as pretext, when the authors note the magnitude of the gap in tax recoveries between various groups of specialized audits, which far exceed actual recoveries or even official estimates of evasion based on the less exacting random audits.

This brings us to the complicated takeaway of these figures. Taking these results, which will of course need to be scrutinized by the wider community of economists, at face value, the IRS could potentially reap significant returns by investing further resources into deconstructing the tangled financial webs of the wealthiest individuals. Yet to do so, they will need to fully explore complex entity structures involving many jurisdictions, and, crucially, also verify these firms’ taxable income figures. This is exceedingly difficult and expensive on a systematic scale, potentially prohibitively so.

Noticeably, the authors settle for describing in rather general terms the nature of the evasion and assume that additional enforcement would yield positive returns so long as some share of random audits could be made more thorough. Given what they acknowledge to be the sophisticated nature of the evasion and evaders, though, the authors do not consider just how quickly these returns might diminish. While it would be a mistake to infer too much from this omission, deemphasizing the finer points of enforcement cost could reflect a belief that evasion should be pursued as a matter of equity as much as practicality, that to yield to large-scale tax evasion by sophisticated actors would in itself be obscene.


By Simon Belokowsky

You may also be interested in: