This week, Jeff Gordon (Vanderbilt, Google Scholar) reviews a new work by Michael Love (Columbia, Google Scholar), Taxing Complexity (2026):

Tax scholars may be generally familiar with the difficulty the IRS faces in auditing sophisticated taxpayers with complex tax positions, but the scale of the problem is shocking even to this cynical reader. As Michael Love reports, drawing on a recent GAO review, about 0.2% of large partnerships (those with over $100 million in assets and over 100 partners) were audited in 2019. Almost all of those audits ran out of time: the average audit concluded 2.99 years after filing, against a statute of limitations of just 3 years. The problem has likely become even worse given significant workforce cuts at the IRS. Thanks to a combination of surrendering enforcement resources (on the government side) and ever-more sophisticated planning (on the taxpayer side), we have reached a point where large businesses, and especially large partnerships, are effectively un-auditable. Much of the work we tax scholars do—proposing legal reforms that could theoretically tax income or wealth more equitably and efficiently—is shouting into a void in light of the system’s practical inability to enforce the law.

In this week’s feature article, Professor Love explores how the government could reclaim an edge in enforcement even without increasing IRS funding. His central insight is that transactional complexity is a social harm, all else equal, and should be managed with a Pigouvian tax instrument. The argument starts by distilling the essence of complexity’s harmful effect. Love observes that complexity—defined functionally as anything that increases the IRS’s resource burden in audits—is harmful in two distinct ways. First, complexity allows taxpayers to shield their aggressive tax positions from IRS oversight, leading to less tax collected than if the IRS could understand the positions taken. But second, complexity dilutes enforcement resources. Because the IRS must spend so much time parsing complex structures, audit resources get pulled away from other taxpayers, causing revenue loss even if the complex taxpayer’s return turns out to be fully compliant. It’s worth noting that the dilution effect will be muted if the IRS doesn’t even try to audit complex returns, which appears close to the reality today. In that case, the shielding effect is the one that dominates.

Professor Love proposes two types of solutions to mitigate the harms of complexity. First, he proposes that deficiency assessments and penalties should be scaled up based on taxpayers’ degree of complexity. This approach is motivated by a deterrence logic: complex taxpayers would respond by making sure they are compliant, or possibly by becoming less complex. But scaled-up penalties would not have any effect on taxpayers who know they are actually compliant, but whose complexity nonetheless poses social costs through dilution. To address that problem, Love proposes a Pigouvian fee on complexity itself. The fee should be an excise tax on some measure of complexity, not a complexity-adjusted income tax—which would be ill-targeted, because a complex taxpayer might have low income (or at least report low income).

This raises the central question that readers have undoubtedly been waiting for: how should we measure complexity? Professor Love explores this question through the lens of large partnerships, among the most challenging business structures to bedevil IRS auditors. He argues that rather than attempt to build a comprehensive complexity index, we should impose many excise taxes, each on a distinct “base” that reflects complexity. Examples include number of higher-tier partnerships, number of unidentified partners, number of foreign partners, number of distributions of property (especially encumbered property), extent to which liability allocations deviate from fixed profit shares, and so on. A useful feature of these indicators is that they should be fairly visible to the IRS on the face of the partnership’s tax return. It would not be much use if the indicia of complexity were only visible after conducting an audit—the very thing that complexity makes unlikely.

But I suspect that some drivers of complexity are not immediately visible before conducting an audit. Examples might include opacity in the beneficial ownership of layered trusts, uncertainty in valuation-model assumptions, legal characterizations that depend on side agreements, and the presence of circular cash flows. We might refer to these as sources of “opaque complexity” or as “unknown unknowns.” If the IRS can’t identify certain forms of complexity without embarking on an audit, we might worry that taxpayers will structure into exactly these forms—if the fee were large enough to be worth avoiding. Still, it is probably worth deterring the more visible versions of complexity even if such deterrence is not comprehensive.

The prospect of avoiding the complexity tax surfaces a central question lurking throughout the paper. To what extent is complexity driven by business purpose rather than tax minimization? And can we tell the difference in individual cases? To be clear, Professor Love’s proposals don’t necessarily require us to make this distinction in order to achieve their intended effects. When complexity is necessary for business reasons, the Pigouvian tax would at least raise revenue which can be used to offset the enforcement-dilution problem. In that light, the IRS might simply become like many other federal agencies that charge user fees to support their work. And when complexity is driven by tax motives, the complexity fee would raise the cost of shielding income and perhaps deter it. But it would be useful to get a sense of why, and how much, taxpayers value their complexity in order to anticipate responses to a complexity tax. When complexity is tax-motivated, we might expect taxpayers to switch into other complex structures that are not counted by the complexity tax. That is, for a tax-motivated firm, different versions of complexity are substitutes. But when a specific transactional form is valuable in its own (non-tax) right, an alternative, equally complex form is not a good substitute.

Tax scholars should welcome this article as an invitation to begin the difficult work of measuring transactional complexity, studying how taxpayers respond to making it costly, and perhaps distinguishing between more and less socially wasteful variants. Professor Love is surely right that we should treat transactional complexity as endogenous to tax law rather than a fact of life that tax administration must accept.

Here’s the rest of this week’s SSRN Tax roundup:

Zhimin Chen (Nanyang Technological University), Martin Jacob (University of Navarra), & Xiang Zheng (Nanyang Technological University), Environmental-Unfriendly Tax Avoidance (2026)

Hayat Daghay (Chicago School of Professional Psychology), Yahya Aissaoui (Independent), & Mohammed Belkasseh (Hassan First University of Settat, Morocco), The Taxpayer’s Mind: Unpacking the Behavioral and Psychological Determinants of Compliance (2026)

Fabio B. Gaertner (Wisconsin, Dept. Acct.) & Jeffrey Hoopes (UNC, Kenan-Flagler Bus. Sch.), A Price-Based Primer on Carbon Taxation (2026)

Assaf Harpaz (Georgia), Taxing AI (2026)

Nora Ijeoma (Independent), Navigating the 2025 Nigerian Tax Reforms: Implications for International Commercial Investors (2026)

Aryan Khandelwal (Independent) & Piyush Jain (IPU Delhi), From Competition to Collaboration: The Global Corporate Tax Transformation (2026)

Mary Lee (Inst. of Chartered Acct. in England & Wales), Tax Architecture in Travel: Strategic Impacts for Hotels and Agencies (2025)

Oliver Zhen Li (NUS), Yupeng Lin (NUS), & Keyuan Zhang (Shanghai Jiao Tong University), The 2003 U.S. Dividend Tax Cut, Small Business Loan Supply, and the Real Economy, Acct. Rev. (2026)

Roberta F. Mann (Oregon) & Tracey M. Roberts (Samford), The Long and Winding Road: The Inflation Reduction Act’s Energy and Environmental Tax Credits, 78 Nat. Tax J. 223 (2025)

Antonio Lopo Martinez (University of Coimbra), IFRS 18 (CPC 51) and Tax Accounting Evidence (2026)

Victoria Plekhanova (University of Auckland), The Integrity of a Tax System as a Governing Tool (2026)

Prasad MK (Independent), Taxing the AI Agents: A Proposal for Corporate Taxation of Autonomous AI Systems That Displace Human Employment (2026)

Mansi R. Rai (N.Y. St. Dep’t of Tax’n & Fin.), From Physical Presence to Digital Participation: Toward a Participation-Based Nexus Doctrine in State Taxation (2026)

Benjamin Silver (Yale), Tax-tualism, 111 Minn. L. Rev. (forthcoming 2027)

Pramod Kumar Siva (Texas A&M) & William Byrnes (Texas A&M), To Tariff or Not to Tariff? The Supreme Court’s Answer and Its Structural Consequences (2026)