Abstract

This Article identifies and theorizes a neglected element of tariffs: “shadow tariffs,” the tariff-like burdens created by executive volatility rather than by durable border collection. When an administration repeatedly announces, delays, conditions, escalates, and reverses tariffs, it changes the expected distribution of future trade costs. Corporations respond ex ante by repricing risk, rewriting contracts, pulling forward inventory, reconfiguring supply chains, investing in compliance and origin planning, and delaying hard-to-reverse investments. Those defensive adjustments impose real costs, even when threatened tariffs are never implemented or are withdrawn before meaningful revenue is collected. The Article uses President Trump’s second term as a case study because it combines unusually frequent threats with aggressive testing of statutory and constitutional boundaries. It makes three contributions. First, it distinguishes shadow tariffs from ordinary incidence analysis and explains why policy reversal does not undo the harm. Second, it maps the private-law and operational channels through which volatility “collects” burdens. Third, it proposes a measurement strategy based on threat windows and observable proxies, and it shows how the Constitution’s separation-of-powers design is economic policy in this setting.

Introduction

Tariffs can be “hidden” in one sense and visible in another. Tariffs are generally visible to importers and customs-facing businesses, but often invisible to consumers because the burden is embedded in final prices rather than charged as a separate line item.^[2] Yet tariffs are visible in formal terms: They rest on identifiable statutory authority, are assessed and collected at the border, and appear in published tariff schedules and customs revenue data.^[3] This standard view misses a second system, one that can impose large and persistent burdens, even when formal tariffs are never collected or when threatened tariffs are later withdrawn. I refer to that second system as “shadow tariffs.”

Shadow tariffs are not metaphors; they are a describable incidence mechanism. When a government repeatedly announces, threatens, delays, conditions, escalates, and partially reverses tariffs, it changes the expected distribution of future trade costs. Corporations and states react by re-pricing risk, rewriting legal contracts, shifting inventory timing, paying for compliance and litigation, and delaying investments. All of these reactions come with a hefty economic price. Therefore, in effect, shadow tariffs operate like tariffs because they force private actors to shift resources from production and exchange to defensive adaptation. Furthermore, they operate like the tariff system because they persist across episodes and compound across supply chains. Last, they are “shadow” only in the narrow sense that they are not line items in the Harmonized Tariff Schedule and do not appear as a tax enacted by Congress.

President Trump’s second term is a unique case study for this because it combines two features that create shadow tariffs at a scale the global economy has not experienced before. First, it created a tariff enforcement environment that is unstable and unpredictable, with frequent threats and sudden changes.^[4] Second, this term tests the institutional boundary between Congress’s power over tariffs and the executive branch’s role to act quickly through delegated authorities and emergency powers, including the International Emergency Economic Powers Act (IEEPA).^[5] The Supreme Court’s decision in Learning Resources narrows this issue, but certainly does not eliminate the problem. The Court held that IEEPA does not authorize the President to impose tariffs, making IEEPA a failed attempt rather than the end to this institutional problem. Executive tariff volatility can migrate to other delegated authorities, including Sections 122, 201, and 301 of the Trade Act of 1974; Section 232 of the Trade Expansion Act of 1962; and Section 338 of the Tariff Act of 1930.^[6] Each authority has different predicates, procedures, and limits. The shadow-tariff analysis still shows not that every statutory theory will succeed, but that firms may continue to face credible tariff risk while the executive shifts between statutory justifications as courts test these different theories.

This Article offers three contributions. First, it defines “shadow tariffs” not merely as hidden, incidental costs, but as a distinct instrument in effect: a volatility regime in tariff policy that imposes tariff-like burdens in advance (ex-ante) through private risk pricing and adjustment on corporations and governments on both sides of the border, even when no additional duty is ultimately collected. Legal analysis often misses that this threat-and-reversal cycle can impose real and uneven burdens without any increases listed in the tariff schedule. The key historical benchmark is the 2018–2019 U.S.-China trade war, when the United States imposed multiple rounds of tariffs on Chinese imports, China retaliated against U.S. exports, and the tariff program was administered through product lists, exclusions, revisions, deadlines, and partial adjustments.^[7] This experience made later tariff threats credible: Firms already saw that large tariffs could be imposed, modified, and sustained long enough to affect pricing, contracts, inventory, supply chains, and investment decisions. The empirical economic literature on trade policy uncertainty provides the missing bridge. For example, using an economy-wide statistical model tracing how trade policy uncertainty shocks affect investment over time, one study estimates that the rise in trade policy uncertainty in 2018 reduced U.S. business investment by about 1% to 2%, reflecting a macroeconomic effect from uncertainty itself rather than from the statutory tariff rate alone.^[8] The uncertainty effects also extend beyond trade and into financial markets and credit. A growing empirical literature shows that trade policy volatility affects finance, not just trade. One study finds that higher tariff uncertainty can weaken the U.S. dollar by raising risk premia.^[9] Another study, using bank supervisory loan data from the 2018–2019 trade war, finds that banks more exposed to trade uncertainty reduce lending, shorten loan terms, and raise borrowing costs, with corporations that heavily rely on bank credit suffering the largest negative effects.^[10]

Second, it explains how tariff threats create real costs. Shadow tariffs show up in higher risk pricing,^[11] rewritten legal contracts,^[12] inventory pull-forwards,^[13] supply-chain shifts,^[14] added compliance and classification work,^[15] and delayed investment.^[16] Therefore, frequent tariff reversals do not mean de minims or low harm, but a real and meaningful economic burden; by the time the tariff threat is withdrawn, corporations and governments have often already incurred the price to hedge, reorganize, and adapt to the potential tariff threat. Hence, even when a threatened tariff never takes effect, the sunk costs have already been incurred through private markups and wasted resources.

Third, it explains why this matters for institutions and for legitimacy. Tariffs are both a way to raise money and a way to regulate trade. The key problem is not only who has the power to set tariffs, but what happens when fast executive action turns tariff volatility into bargaining leverage. Volatility pushes costs onto U.S. corporations, workers, and consumers, and onto foreign governments coerced into bargaining. It also weakens the credibility of trade commitments and can spill into finance, exchange rates, and supply chains. These burdens are hard to track and hard to challenge. They do not look like taxes, but they act like taxes that are not transparent, nor clearly assigned, nor easily undone.

The core claim in this Article is simple. Tariff harm is not limited to the posted tariff schedule. A Bloomberg Economics review of 49 tariff threats since November 2024 suggests that most of these threats were not ultimately implemented,^[17] yet these repeated threats and reversals still create a shadow tariff system imposing real costs without new collection.

I. Defining Shadow Tariffs

1. The Concept

Shadow tariffs are economically costly obligations and behavioral changes created by tariff threats and tariff volatility, even when the threatened tariff is never implemented or is later reduced or is withdrawn. The key feature of shadow tariffs is that they are incurred ex ante, when corporations respond to credible tariff risk by repricing, renegotiating, hedging, and reconfiguring sourcing and logistics. These effects can persist even if the threatened tariff is never imposed or is later rolled back.

However, shadow tariffs are not just uncertainty, because corporations face uncertainty in every market and under many legal regimes. Shadow tariffs are the private costs created when an administration uses tariff volatility, including threats and reversals, as a policy tool to gain leverage, including in trade negotiations. They arise when tariff outcomes are meaningfully plausible across a wide range of rates, the timing is hard to predict, and the downside of being unprepared is large. Corporations then respond predictably by treating the tariff as a real contingency and spending resources on hedging, repricing, renegotiation, or reconfiguration of their operations. Even if the administration later withdraws the threat (as it did more often than not^[18]) or reaches a deal, those costs have already been incurred, and many cannot be fully unwound.

This is consistent with the theory of irreversible investment and real options.^[19] When entry costs are sunk, policy uncertainty increases the option value of waiting, so corporations rationally delay irreversible commitments until the policy environment clarifies. Applied to trade, this implies that tariff uncertainty can reduce investment and entry into export markets even when current tariffs are low. Furthermore, that delay itself is costly because it decreases trade volume and can reduce welfare even without an immediate tariff increase.^[20]

1. Distinguishing Shadow Tariffs from Ordinary Tariff Incidence

Tariff incidence is about who pays the tariffs once those are imposed. Evidence from the 2018 trade war suggests many tariffs were largely passed through into duty-inclusive import prices, so domestic importers and consumers bore much of the cost.^[21] That is classic incidence analysis. A recent shipment-level study of the 2025 tariff shock reports the same basic pattern on a larger scale, estimating that foreign exporters absorbed about 4 percent of the burden—passing roughly 96 percent through to U.S. buyers.^[22]

Shadow tariffs operate differently because they impose costs through uncertainty and the protective steps corporations take in response. Corporations can be harmed without ever paying a duty: When tariff outcomes become harder to predict, corporations may implement certain buffers into prices, incur higher legal and compliance costs, shift supply chains, and delay investment.^[23] Those responses shift the burden to others, raising prices for consumers and creating inflationary pressure in exposed goods and inputs,^[24] while also weighing on hiring and capital spending.^[25] Also, small and medium-sized enterprises are often more impacted because they cannot diversify supply chains or spread fixed compliance costs,^[26] which can weaken economic competition and leave consumers with higher prices and fewer choices.^[27]

1. Why Reversal Does Not Undo the Harm

It is tempting to treat tariff reversals as evidence that threats were harmless. Yet this conclusion fails for two reasons. First, shadow tariffs incur costs when a threat is issued and treated as credible. Once corporations begin taking protective measures, the resources committed are sunk, and a later rollback does not recoup those costs.

Second, a rollback can eliminate the specific tariff threat, or even remove an imposed tariff, but it does not restore stability. One deal can settle the immediate dispute while still signaling that tariffs remain available as a rapid, discretionary lever under adjacent executive authorities. After the Supreme Court rejected IEEPA tariff authority, the administration promptly invoked Section 122 of the Trade Act of 1974, reminding firms that the volatility problem is not limited to one statute.^[28] Corporations therefore keep pricing and planning around renewed tariff risk. The expected tariff level may fall, but the range of credible future tariff outcomes can still be broad, and this kind of policy volatility is widely understood to deter investment and reduce output.^[29]

II. How Shadow Tariffs Are Created

1. Pricing and Precautionary Margins

In a stable tariff regime, corporations can price around known customs duties. In a volatile tariff regime, pricing must also account for downside risk. A threatened tariff is not a forecast, but a plausible adverse state that can arrive quickly. Because supply chains are slow to adjust, underpricing that risk can be costly. Corporations thus build precautionary margins into prices, which can look like inflationary pressure even if the tariff is never collected.

The tariff incidence literature documents substantial pass-through of imposed tariffs into domestic prices. Evidence from the 2018 trade war indicates that many duties were incorporated into duty-inclusive import prices, so the burden fell largely on United States importers and consumers and produced measurable real-income losses.^[30] The shadow tariff point is that credible tariff threats can produce similar pricing effects to inflation even without collection, because corporations build the risk of a high-tariff outcome into prices through precautionary margins and contractual repricing. In 2025, that credibility is reinforced by the 2018 trade war experience, which naturally showed that large tariffs can, in fact, be imposed and sustained, so corporations rationally treat renewed threats as plausible rather than hypothetical.^[31]

1. Contracting and the Private Law of Volatility

Shadow tariffs inspire parties to shift tariff risks within private contracts. Even when an administration uses tariff threats mainly as a negotiation lever, the threat does not happen in a vacuum and thus does not stay in the negotiation room. Once the threat is considered credible, corporations and their counterparties must protect themselves in real time, and they do so by rewriting deals, changing pricing terms, and reallocating risk in their legal contracts. The result is that a public bargaining tactic becomes a private cost that shows up in contract terms, prices, and business decisions.

This changes contracting in different ways. First, contracts allocate tariff risk expressly through tariff-adjustment clauses, reopener provisions, and indexation, which increases drafting, negotiation, and monitoring costs.^[32] These clauses do more than increase transaction costs: They determine whether the importer, foreign supplier, downstream buyer, or final customer bears the tariff-risk premium. Second, contract duration often shortens as a protection mechanism.^[33] Long-term pricing becomes harder when tariff terms can change quickly, so parties rely more on shorter agreements, which raises transaction costs and can discourage relationship-specific investment. Third, parties adjust the legal framework. They may choose governing law, forum, and dispute terms with greater care, or shift sourcing and counterparties toward jurisdictions that appear more predictable.^[34] They may also use change-in-law clauses, termination rights, minimum-volume commitments, indemnities, delivery terms, and documentation obligations to determine who must absorb a potential tariff change.

The normative point is that volatility changes how the costs are borne. Stable tariffs impose visible border costs. Executive tariff volatility pushes a large share of the burden into private contracting, where it is embedded in pricing and allocation terms and largely “disappears” into ordinary commercial pricing. Also, because contracting outcomes frequently follow bargaining power, smaller corporations typically have less ability to protect themselves. In practice, larger counterparties can push tariff risk upstream to suppliers or (more often) downstream to buyers, while smaller importers and suppliers are more likely to absorb the risk through lower margins, shorter commitments, higher compliance costs, or price increases where pass-through is possible.

1. Inventory Distortion and Pull-Forward Behavior

When the threat of a new tariff is time-sensitive, as it often is, corporations have a predictable incentive to accelerate imports to clear goods before the effective date. That produces short-term pull-forward surges as corporations spend real resources to win the timing race: expedited freight, additional warehousing, and working-capital financing, along with internal compliance and logistics effort.^[35] The main economic cost is not a tariff payment, because in many instances no new tariff is ever collected,^[36] and the cost is the deadweight resource waste from threat-driven timing behavior: corporations spend to rush, store, and finance inventory earlier than economically necessary, rather than in response to consumer demand, productive efficiency, or ordinary inventory needs.^[37] Furthermore, bottlenecks and system strain can add to the harm in some cases through delays and higher logistics prices, but they are secondary to the corporate resource cost.

These incentives also distort the data. Customs receipts can rise before any tariff takes effect, simply because corporations accelerate entries to beat a possible effective date. If the tariff is later delayed or withdrawn, that temporary spike can be mistaken for a real increase in trade or revenue. But the costs remain sunk: inventory was pulled forward, working capital was tied up, and supply chains absorbed avoidable disruption. Overall, this whole process damages the natural flow of business activity and does not benefit the economy.

1. Supply-Chain Reconfiguration and Duplication

Tariff volatility can push corporations to build redundancy into their supply chains because it makes a single, optimized network fragile.^[38] When a tariff can be announced and take effect on short notice, reliance on one country or one supplier exposes the corporation to sudden margin compression and delivery failure mid-contract. Switching is not frictionless. Inputs are often relationship-specific and cannot be substituted overnight without search, qualification, and renegotiation.^[39] Therefore, redundancy is the defensive response: add suppliers, qualify substitute inputs, duplicate tooling or capacity where needed, and shift sourcing or assembly so they have an operational fallback if the tariff state changes.^[40] The economic cost is the real resource waste of building and maintaining this hedging strategy, including sunk engineering, testing, and compliance costs to qualify substitutes and underused duplicate capacity, rather than a tariff payment.

Though some of this duplicity may be genuine resilience, when the driver of this is policy volatility rather than bona fide business growth reasons, it can lead to inefficiency that eventually raises costs.

1. Compliance, Classification, and Origin Planning

Tariffs function as taxes on importation, so they predictably generate some planning to reduce liability.^[41] Volatility amplifies that planning because uncertainty over future rates increases the expected payoff from flexibility, designs with contingencies, and reclassification strategies.

Tariff compliance has three recurring fixed-cost components: classification, valuation, and country-of-origin determination.^[42] Volatility increases those costs for two reasons. First, it forces corporations to plan for multiple tariff rates rather than a single, stable schedule. Second, it raises the payoff to legal and operational redesign aimed at reducing tariff exposure.

Corporations can respond by modifying product attributes to fit a different tariff classification,^[43] shifting where value is added to alter origin outcomes, or restructuring transactions to satisfy exemption criteria.^[44] Some of this activity is unavoidable because lawful importation requires correct classification, valuation, and origin reporting. However, volatility-induced planning often functions as avoidance through private redesign rather than productive (or bona fide) business investment, and though the private costs are real, the economic, social, and welfare gains are little or nothing, because resources go toward changing labels or routes to lower duties rather than producing more goods or services.

1. Investment Delay and Uncertainty

Investment delay under tariff uncertainty is the most important mechanism, and the trade policy uncertainty literature supports this conclusion.^[45] When trade policy creates uncertainty, corporations delay or scale back sunk investments to enter or expand,^[46] even if effective tariff rates are lower than expected.^[47] On the other hand, binding trade commitments, especially tariff bindings in trade agreements, can increase trade and new market entry because they reduce the likelihood of future tariff increases, therefore reducing downside risk for corporations contemplating irreversible entry and expansion investments.^[48]

Market entry is just the first place this shows up. The same point applies whenever corporations must make hard-to-reverse commitments. When tariff policy is volatile, corporations face more risk when they commit to fixed supply chains, specialized equipment, long-term legal contracts, or market-specific distribution, and they often respond by delaying investment, shortening commitments, and favoring choices that are easier to unwind.

The effect is not limited to who pays the tariff if it is imposed; it also arises when the tariff never takes effect, because the threat creates uncertainty, and uncertainty makes delay rational when investments are hard to reverse. Therefore, higher trade policy uncertainty is associated with lower business investment and weaker real activity.^[49] Perhaps more importantly in fiscal terms, the burden shows up as delayed or foregone output, rather than collected customs revenue.

Furthermore, this matters for stated “reshoring” and domestic-investment goals. If the administration’s stated goal is to attract more domestic investment, tariff volatility can work against that goal because it increases uncertainty about future input costs, retaliation, and market access, creating uncertainty that makes corporations less willing to commit hard-to-reduce capital. Instead of rushing to build, corporations postpone projects, scale them back, or choose more flexible arrangements.^[50] This effect is exactly what standard investment theory predicts when uncertainty rises and investments are difficult to undo, matching the macro evidence linking higher trade policy uncertainty to weaker investment.^[51]

III. Distributional and Cross-Border Incidence

1. The Domestic Incidence in the United States

Formal tariffs are legible in law and accounting: The rate schedule is public, and collections show up as customs revenue.^[52] But their economic incidence is still partly opaque because the burden can be transmitted through pass-through, margin compression, input costs, and supply-chain adjustments.^[53] Shadow tariffs add another layer of opacity because a large share of the burden arises through behavior under uncertainty, such as precautionary pricing, redundant sourcing, compliance redesign, and delayed investment, even when no new tariff is ultimately collected. The result is a less transparent burden, and whether it is more regressive depends on the context, so I argue it should be framed as a likely outcome rather than a certain one.

For consumers, the burden often takes the form of higher prices, reduced variety, and higher supply-chain costs, especially in sectors that rely on imported final goods or imported intermediates. The evidence from the 2018 tariff episode is consistent with substantial pass-through and real income losses borne domestically.^[54]

For workers, the burden is less direct and shows up more through slower investment and weaker labor demand.^[55] When corporations delay their investment or reconfigure their supply chains, labor demand can fall in exposed regions and industries. Also, the risk of retaliation can further increase this effect by reducing demand in export-facing industries and sectors.

For corporations, volatility changes the relative burden depending on the corporation’s size.^[56] Large corporations can diversify suppliers, shift sourcing, and spread fixed compliance and adjustment costs across higher volumes. Small and medium-sized corporations face higher per-unit costs and weaker bargaining power, and some may reduce trade exposure or exit importing altogether. The result is a competitive distortion risk: When participation requires costly and uncertain adjustments, smaller importers are more likely to be constrained or fail, which can weaken entry, expansion, and competition.^[57]

The 2018 trade war has useful evidence on this point. The economic literature finds large welfare losses from the policy shift toward protectionism and retaliation, along with measurable declines in real income tied to tariffs and their price effects.^[58] Shadow tariffs can deepen these same harms because, as discussed earlier, uncertainty itself imposes meaningful costs.

1. Burdening Other Nations Through Volatility Rather Than Collection

Tariffs can burden foreign nations by restricting market access through border duties that raise the cost of their exports.^[59] Shadow tariffs can impose costs on foreign corporations because, even when the tariff schedule does not change, the relevant mechanism is credible tariff risk rather than duties collected. In other words, the economically meaningful response happens ex ante. By the time a tariff is imposed, foreign corporations have often already incurred adjustment costs to reduce exposure, whether by diversifying export markets, reconfiguring production and sourcing, reallocating inventory and shipping routes, renegotiating supply contracts and pricing terms, shifting assembly or final processing to alternative jurisdictions, or delaying and discounting investment tied to United States demand.^[60]

The point is not that threatened tariffs replicate the effects of an enacted tariff. The point is that credible volatility changes decision-making conditions. Foreign corporations and their counterparties have no choice but to treat tariff exposure as a real risk when they set prices, choose legal contract terms, and decide whether to commit sunk capital. This leads to three predictable effects.

First, tariff uncertainty can deter foreign corporations from investing in U.S. capacity or from forming relationship-specific supply arrangements with U.S. buyers. When the probability, timing, and magnitude of a tariff shock are hard to forecast, foreign corporations rationally discount projects based in the United States.^[61] That discount will be greatest for projects with high sunk costs, long lead times, relationship-specific tooling, regulatory approvals, or dependence on stable long-term U.S. demand. The result can be less cross-border integration even if market access remains formally open.^[62]

Second, volatility can function as bargaining leverage. When foreign corporations and their U.S. counterparties treat rapid executive tariffs as a credible downside risk, they have incentives to offer concessions that reduce exposure—even if the threatened tariff is never imposed. The mechanism operates through the threat and the risk premium it creates, not through the later collection of duties.^[63] The price is often relational, as repeated tariff threats strain long-term commercial relationships by undermining trust, increasing suspicions of opportunism, and pushing parties to rely on protective contract terms and alternative suppliers rather than cooperative, relationship-specific investments.

Third, volatility can trigger defensive responses abroad. States may subsidize domestic substitutes, redirect supply chains toward alternative markets, accelerate “friend shoring” and regional blocs, or design contingency arrangements to reduce exposure to the discretion of the U.S. government. Those adjustments can fragment the trading system and raise global costs, again without any change in the posted tariff rate.^[64]

The trade policy uncertainty literature supports the general idea that trade policy uncertainty can depress trade and investment.^[65] What is distinctive about the Trump-era approach is the scale and frequency of discretionary tariff threats.^[66] Existing statutory delegations, as interpreted and used by the executive, allow tariffs to be announced, revised, and withdrawn with limited ex ante procedure.^[67] That institutional flexibility allows the executive to create credible tariff risk for virtually any trading partner and for any corporation whose business materially depends on U.S. market access. Foreign corporations and foreign governments then incur mitigation, pricing, and bargaining costs in advance, even when no tariff is ultimately imposed, or when it is imposed only briefly and generates little revenue.

IV. The Institutional Engine of Shadow Tariffs

1. Speed, Delegation, and Emergency Powers

Shadow tariffs are created at scale when the executive branch can impose tariff risk quickly and with minimal checks and balances. Traditional trade statutes often require investigation, findings, and procedural steps,^[68] which also slow the policy cycle and make abrupt changes harder to make. Put differently, traditional trade policy provides some stability.

The recent controversy over IEEPA illustrates a different mechanism for the executive branch to impose tariffs.^[69] President Trump invoked IEEPA in early 2025 to impose a broad tariff program, including a baseline tariff and then higher “reciprocal” tariffs layered by country.^[70] That move was a significant departure from ordinary practice and triggered litigation over whether an emergency sanctions statute can support a general tariff schedule.^[71] The resulting cases reached the Supreme Court through different procedural paths, with Learning Resources coming from a decision by the District of Columbia Circuit and V.O.S. Selections from the Federal Circuit.^[72] On February 20, 2026, the Court held that IEEPA does not authorize the President to impose tariffs.^[73]

For the shadow tariff mechanism, however, the key dispute is not the merits of this specific case, as the Trump administration had already publicly stated its intention to continue to impose its prior tariff policy through other statutory means.^[74] Justice Kavanaugh’s dissent makes the same point: Although the Court rejected the argument that IEEPA granted the president tariff powers, there are other statutory authorities, including Section 232, Section 122, Section 201, Section 301, and Section 338, that function as possible alternative routes for presidential tariff action (though sometimes with additional procedural steps).^[75] Hence, Tariffs would still be announced, revised, and withdrawn within days, and corporations would not be able to rely on the current state of trade agreements. Unfortunately, the Supreme Court’s questioning about reliance interests and the practical consequences of broad executive use of tariffs did not fully consider this volatility mechanism.^[76]

The remedial point is part of the same shadow-tariff mechanism. Even when a tariff is later held unlawful, importers do not immediately return to the status quo ante. They may have already paid the duties, financed the cash outflow, passed part of the cost through prices, renegotiated contracts, filed protests or refund claims, and waited for liquidation, reliquidation, or a special administrative refund process. The IEEPA refund litigation illustrates the point: After Learning Resources, the Court of International Trade had to address how CBP would liquidate or reliquidate entries without IEEPA duties, and CBP developed CAPE to process refunds at scale rather than entry-by-entry.^[77] That remedial lag reinforces defensive behavior as a rational response. If corporations know that even an unlawful tariff may require months of financing, documentation, litigation, and refund uncertainty, they have reason to respond at the threat stage rather than wait for later judicial relief. Illegality may end the government’s authority to retain the duty, but it does not erase the interim costs of payment, financing, compliance, and uncertainty.

1. The Constitutional Allocation and Nondelegation Anxiety

Tariffs sit at the intersection of taxation and the regulation of foreign commerce.^[78] The Constitution assigns those fiscal choices to Congress,^[79] but modern trade law contains wide delegations that can be activated quickly, especially when framed as “national security” or “emergency” measures.^[80]

During the Learning Resources argument, the Justices repeatedly returned to the separation-of-powers problem created by treating emergency authority as an implied tariff statute,^[81] and Justice Gorsuch pressed the government on the limiting principle: What stops Congress from handing over “all responsibility” for trade or other core powers if the Court accepts the government’s theory?^[82] That exchange captures the institutional concern that drives shadow tariffs: Broad delegation substantially reduces ex ante procedural checks and allows the executive to generate tariff volatility at a speed and scale that ordinary trade statutes do not. The implications are also broader than the Court’s discussion suggested because the Court did not fully address either the real economic impact of tariff volatility or the practical mechanism by which the federal government would refund importers after unlawful tariffs had already been collected.^[83]

V. Measuring Shadow Tariffs

The following proposes a structured measure of shadow tariffs. The first task is conceptual, because we need to identify the phenomenon: the costs are imposed by credible tariff risk even when the posted tariff schedule does not change, or changes only briefly. Only after the phenomenon is recognized does measurement become useful. Measurement will also rarely be reduced to a single number. The object is not “the tariff rate,” and more likely requires a combination of observable proxies that reflect information, legal contracting, and operational responses.

A reasonable measurement strategy is therefore incremental and mixed. Begin with observable indicators that track the mechanisms described earlier, test whether they move during clear “threat windows” (announcements, deadlines, reversals, and implementation dates), and then refine the framework as the evidence accumulates. One can think of the result as an index, but the more important point is triangulation: multiple imperfect measures pointing in the same direction.

To see the logic, consider a simple stylized example. Assume the executive announces on Day 1 that a 25 percent tariff on a major input from Country X will take effect on Day 30, signals on Day 20 that implementation may be delayed, and withdraws the tariff on Day 29. The posted tariff schedule may never meaningfully change, and revenue may be near zero, but the shadow tariff can still be substantial because behavior adjusts during the risk window. In those thirty days, uncertainty measures can spike; buyers may insist on tariff-adjustment clauses or reopeners in new contracts; importers may pull shipments forward to beat the deadline; customs brokers may see higher volume and more classification and origin questions; and corporations may pause, re-route, or diversify orders rather than commit to long-lead inputs tied to Country X. These are observable responses to risk, not to collection, and are what the measurements must capture.

With that in mind, a measurement framework can draw on at least five observable components:

1. Information and uncertainty – Trade policy uncertainty measures, including the series developed by Caldara and coauthors from newspaper coverage, earnings-call transcripts, and tariff actions,^[84] can proxy for how salient and variable tariff risk is at a given time.

2. Contract adaptation – The private-law response can be proxied by tracking changes in tariff-related contract terms, including tariff-adjustment provisions, renegotiation triggers, origin flexibility, change-in-law clauses, and termination rights tied to tariff changes. This proxy is imperfect and may lag the threat itself: parties may add tariff clauses before a tariff is imposed, and those clauses may remain in contracts after a threat subsides. For that reason, the relevant measure is not the raw presence of tariff clauses alone, but changes in clause prevalence, amendment timing, renegotiation frequency, and sector-specific clustering around identifiable threat windows.

3. Inventory and shipment timing – If corporations fear a near-term change in tariff rates, they may pull purchases forward. Industry inventory-to-sales ratios, import timing at major ports, and shipment bunching around threat windows can proxy for this precautionary behavior.

4. Compliance and dispute activity – Volatility can raise the practical stakes of classification and origin decisions. Changes in customs brokerage demand, ruling requests, protests, and trade-related litigation can track the compliance burden generated by tariff risk.

5. Investment and entry – If tariff exposure becomes harder to predict, corporations may delay, scale down, or re-route long-horizon projects. Corporation-level investment patterns and entry decisions in trade-exposed industries can serve as a proxy for this “wait or shift” response.

Note that these measures are not decisive on their own. If shadow tariffs are operating, several of these indicators should move in the same direction around the same period. That is also where legal scholarship can add value. Economists can measure uncertainty and correlate it with macro-outcomes; legal analysis can help specify when a threat is institutionally credible, why it is credible, and which episodes should count as the relevant threat windows for measurement and, later, for reform.

Conclusion

A key mistake in contemporary tariff debate is to treat the posted tariff rate as the entire instrument because that focus misses the volatility regime, which can operate as a parallel system. Shadow tariffs are burdens collected through precautionary behavior: corporations change prices, legal contracts, inventories, sourcing, and investment plans in response to credible tariff risk, often before any duty is collected and sometimes even when no tariff is ultimately imposed.

This matters because the standard defense of tariffs assumes a stable tool that can be evaluated in familiar cost-benefit terms. Volatility changes the nature of the instrument. It moves tariffs away from a rule-like policy and towards a threat technology. A threat strategy may sometimes gain bargaining leverage, but it is also costly in the aggregate because it forces defensive behavior, generating real resource waste, not merely transfers. Whether an administration intends those costs is not the point; the mechanism operates once the risk is credible.

Volatility also tends to compound. As tariff risk rises, corporations invest more in defensive planning. Defensive planning increases pressure for exceptions and carve-outs. Exceptions and carve-outs increase complexity. Complexity expands planning opportunities and deepens uncertainty. The result can be a recursive complexity loop in which the shadow tariff system grows, even if the formal tariff schedule does not. If the goal is predictable trade and investment, this is the wrong direction.

Instability is therefore not only costly for business and the economy; it is a problem of institutional design as much as of trade policy. The legal stakes follow directly. Shadow tariffs scale when credible tariff risk can be generated cheaply and quickly. Boundary questions about emergency tariff authority, statutory delegation, and separation of powers are therefore not technical disputes; they set the institutional cost of creating tariff risk. The Constitution’s separation-of-powers design is therefore economic policy: it determines how cheaply the executive can create tariff volatility, and thus how much shadow-tariff capacity the legal system permits.

1. * Associate Professor of Tax and Business Law and Director of the MTax Program, Steininger Faculty Fellow at the University of Akron, College of Business, Daverio School of Accountancy. ↑
2. . Mary Amiti, Stephen J. Redding & David E. Weinstein, The Impact of the 2018 Tariffs on Prices and Welfare, 33 J. Econ. Persp. 187, 207 (2019); Reuven S. Avi-Yonah, Doron Narotzki & Tamir Shanan, From Relic to Relevance, The Resurgence of Tariffs, 77 U.C. L.J. 497, 512 (2026) [hereinafter Avi-Yonah et al., From Relic to Relevance]; Doron Narotzki, Tamir Shanan & Reuven S. Avi-Yonah, The U.S. Tax Paradox, 14 Tex. A&M L. Rev. (forthcoming Fall 2026) (manuscript at 9) (on file with the author). ↑
3. . U.S. Const. art. I, § 8, cl. 1; 19 U.S.C. §§ 1202, 1505(a); U.S. Int’l Trade Comm’n, Harmonized Tariff Schedule of the United States, https://hts.usitc.gov [https://perma.cc/ZQK3-R8LZ]; U.S. Dep’t of the Treasury, Bureau of the Fiscal Serv., Monthly Treasury Statement of Receipts and Outlays of the United States Government (Apr. 5, 2026), https://fiscaldata.treasury.gov/datasets/
   monthly-treasury-statement/summary-of-receipts-outlays-and-the-deficit-surplus-of-the-u-s-government [https://perma.cc/RS44-2BHS] (listing “Customs Duties” receipts). ↑
4. . Doron Narotzki, Bluff Without Purpose: Rethinking Trump’s Tariff Policy and the Misuse of Game Theory 2–4 (May 26, 2026) (unpublished manuscript), https://ssrn.com/abstract=5220905 [https://perma.cc/4JT5-C9QV]; Brendan Murray, TACO Tracking: Trump Carries Out One in Four Tariff Threats, Bloomberg (Jan. 27, 2026), https://www.bloomberg.com/news/articles/2026-01-27/
   bluffing-or-not-counting-trump-s-tariff-threats-versus-actions [https://perma.cc/D24D-CDUP]; Chloe Taylor, Trump’s Latest Tariffs U-Turn Is Sparking a Global Market Rally – and Reviving Talk of the ‘TACO Trade’, CNBC (Jan. 22, 2026), https://www.cnbc.com/2026/01/22/trump-tariffs-greenland-europe-deal-taco-trade-sell-america.html [https://perma.cc/4U4Z-L23J]. ↑
5. . Doron Narotzki, Hidden Taxation and the Rise of the Shadow Fiscal State, 7 Corp. & Bus. L.J. 59, 92–106 (2026), [hereinafter Narotzki, Hidden Taxation]; Doron Narotzki, Tariffs Are Taxes, Congress Must Levy Them, 80 U. Mia. L. Rev. Caveat 29, 35–38 (2026) [hereinafter Narotzki, Tariffs Are Taxes]. ↑
6. . See Learning Res., Inc. v. Trump, 146 S. Ct. 628, 691 (2026) (Kavanaugh, J., dissenting); Megan Messerly, Doug Palmer, Daniel Desrochers & Ari Hawkins, The White House’s Plan A Is Winning Its Supreme Court Tariff Case. It Also Has a Plan B., Politico (Sept. 4, 2025) https://www.politico.com/news/2025/11/04/trumps-trade-wall-faces-a-legal-stress-test-the-whitehouse-isnt-panicking-yet-00633773. ↑
7. . Nurullah Gur & Serif Dilek, US–China Economic Rivalry and the Reshoring of Global Supply Chains, 16 Chinese J. Int’l Pol. 61, 63 (2023). ↑
8. . Dario Caldara, Matteo Iacoviello, Patrick Molligo, Andrea Prestipino & Andrea Raffo, The Economic Effects of Trade Policy Uncertainty, 109 J. Monetary Econ. 38, 48 (2020). To be clear, the impact of trade policy uncertainty is not limited to U.S. corporations. See, e.g., Felipe Benguria, Jaerim Choi, Deborah L. Swenson & Mingzhi (Jimmy) Xu, Anxiety or Pain? The Impact of Tariffs and Uncertainty on Chinese Firms in the Trade War, J. Int’l Econ., Mar. 20, 2022, at 22. ↑
9. . Ṣebnem Kalemli-Özcan, Can Soylu & Muhammed A. Yildirim, Global Trade, Tariff Uncertainty and the U.S. Dollar 10 (Nat’l Bureau of Econ. Rsch., Working Paper No. 34728, 2026). ↑
10. . Ricardo Correa, Julian di Giovanni, Linda S. Goldberg & Camelia Minoiu, Trade Uncertainty and U.S. Bank Lending 23 (Nat’l Bureau of Econ. Rsch., Working Paper No. 31860, Nov. 2023, rev. May 2024). ↑
11. . Alberto Cavallo, Gita Gopinath, Brent Neiman & Jenny Tang, Tariff Pass-Through at the Border and at the Store: Evidence from US Trade Policy, 3 Am. Econ. Rev.: Insights 19, 26–31 (2021). ↑
12. . Brooks E. Allen et al., Navigating the Impact of the Trump Tariffs on Commercial Contracts, Skadden Insights (June 2025), https://www.skadden.com/insights/publications/2025/06/insights-june-2025/navigating-the-impact-of-the-trump-tariffs [https://perma.cc/7KAC-7HVK]. ↑
13. . See, e.g., George Alessandria, Shafaat Yar Khan & Armen Khederlarian, Taking Stock of Trade Policy Uncertainty: Evidence from China’s Pre-WTO Accession, J. Int’l Econ. Apr. 2024, at 1 (finding “the costs associated with the trade policy induced stockpiling reduce entrants’ incentive to operate in a market with tariff uncertainty”); Cavallo et al., supra note 10, at 30. ↑
14. . Kyle Handley, Fariha Kamal & Ryan Monarch, Supply Chain Adjustments to Tariff Shocks: Evidence from Firm Trade Linkages in the 2018-2019 U.S. Trade War 29–30 (Nat’l Bureau of Econ. Rsch., Working Paper No. 31602, 2023); Gene M. Grossman, Elhanan Helpman & Stephen J. Redding, When Tariffs Disrupt Global Supply Chains, 114 Am. Econ. Rev. 988, 1026 (2024). ↑
15. . See U.S. Customs & Border Prot., What Every Member of the Trade Community Should Know About: Tariff Classification 3, 8–9 (May 2004), https://www.govinfo.gov/content/pkg/GOVPUB HS4_100-PURL-LPS74209/pdf/GOVPUB-HS4_100-PURL-LPS74209.pdf [https://perma.cc/9QAT-9US3]; U.S. Customs & Border Prot., What Every Member of the Trade Community Should Know About: Recordkeeping 3, 12–17, 20–23 (Jan. 2005), https://www.cbp.gov/sites/default/files/2025-07/Recordkeeping.pdf [https://perma.cc/KP96-THN6]; U.S. Customs & Border Prot., What Every Member of the Trade Community Should Know: Reasonable Care 3, 8–13 (Sept. 2017), https://
    http://www.cbp.gov/sites/default/files/assets/documents/2020-Feb/icprescare2017revision.pdf [https://perma.cc/894G-5Q54]. ↑
16. . Caldara et al., supra note 7, at 45. ↑
17. . Murray, supra note 3. ↑
18. . Id. ↑
19. . Avinash Dixit & Robert Pindyck, Investment Under Uncertainty 3–25 (1994); Robert S. Pindyck, Irreversibility, Uncertainty, and Investment, 29 J. Econ. Lit. 1110, 1111–13 (1991); Kyle Handley & Nuno Limão, Trade and Investment Under Policy Uncertainty: Theory and Firm Evidence, 7 Am. Econ. J.: Econ. Pol’y 189, 214 (2015). ↑
20. . See Handley & Limão, supra note 18, 190–91, 211 tbl. 6, 214. ↑
21. . Amiti et al., supra note 1, at 207–08. ↑
22. . Julian Hinz, Aaron Lohmann, Hendrik Mahlkow & Anna Vorwig, America’s Own Goal: Who Pays the Tariffs?, Kiel Pol’y Brief, Jan. 2026, at 2. ↑
23. . See supra notes 6–13. ↑
24. . Amiti et al., supra note 1, 207–08. ↑
25. . Daniel Fried, The Effects of Tariffs and Trade Barriers in CBO’s Projections, Cong. Budget Off. Blog (Aug. 22, 2019), https://www.cbo.gov/publication/55576 [https://perma.cc/38GS-NS42]; Juan M. Londono, Sai Ma & Beth Anne Wilson, Costs of Rising Uncertainty, FEDS Notes 3 (Apr. 24, 2025), https://www.federalreserve.gov/econres/notes/feds-notes/costs-of-rising-uncertainty-20250424.html [https://perma.cc/YB8X-RECH]. ↑
26. . See Philippe Andrade et al., Small and Medium-Sized Businesses’ Expectations Concerning Tariffs, Costs, and Prices, Current Pol’y Persps., Apr. 10, 2025, at 2, https://www.bostonfed.org/
    publications/current-policy-perspectives/2025/smb-expectations-concerning-tariffs-costs-prices.aspx [https://perma.cc/G838-EALJ] (“[F]irms planned to pass along the expected tariff-induced changes in unit costs to their customers through price increases.”). ↑
27. . Eliot G. Disner, Barrier Analysis in Antitrust Law, 58 Corn. L. Rev. 862, 862–63 (1973). ↑
28. . Keigh E. Hammond & William F. Burkhart, Cong. Rsch. Serv., R48549, Presidential 2025 Tariff Actions Timeline and Status 2–5 (2026) (summarizing repeated tariff actions followed by negotiations, framework statements, and temporary tariff truces), https://www.congress.gov/crs_external_products/R/PDF/R48549/R48549.12.pdf [https://perma.cc/6B
    HN-7VZ7]; Michael Lowell et al., Trump 2.0 Tariff Tracker, Trade Compliance Resource Hub, https://www.tradecomplianceresourcehub.com/2026/01/27/trump-2-0-tariff-tracker/ [https://perma.cc/
    84AS-VPDJ]. Even after the use of Section 122 for these tariffs was struck down by the Court of International Trade, the Trump “administration is already working on its next plan for tariffs.” Tony Romm & Ana Swanson, Trade Court Rules Trump’s 10% Global Tariff Is Illegal, N.Y. Times (May 7, 2026), https://www.nytimes.com/2026/05/07/business/economy/trump-global-tariff-ruled-illegal.html [https://perma.cc/Y9UB-NL8J]. ↑
29. . See, e.g., Caldara et al., supra note 7, at 45–46; Londono, Ma & Wilson, supra note 24, at 3. ↑
30. . Amiti et al., supra note 1, at 208. ↑
31. . Reuven S. Avi-Yonah & Doron Narotzki, The Tariffs Are Coming! The Tariffs Are Coming!, 116 Tax Notes 1577, 1577–79 (2024); Doron Narotzki, Tariffs: Back to the Future, 116 Tax Notes 565, 567–69 (2024). ↑
32. . Flavio Peter, Clio Mordivoglia & Carolina Solis Villares, The Impact of U.S. Tariffs on International Sales Contracts, P&K Insights, Apr. 16, 2025, https://peterandkim.com/uploads/2025/
    04/Peter-Kim-Insights_Impact-of-US-tariffs-on-International-Sales-Contracts.pdf [https://perma.cc/9L
    LR-9C83]; see also Robert E. Scott & George G. Triantis, Anticipating Litigation in Contract Design, 115 Yale L.J. 814, 823–39 (2006) (describing how front- and back-end adjustments to contracts can result in different costs to the parties). ↑
33. . See Peter et al., supra note 31. ↑
34. . Id. ↑
35. . See, e.g., Academy Sports + Outdoors, Inc., Current Report (Form 8-K) (June 10, 2025), https://www.sec.gov/Archives/edgar/data/1817358/000181735825000116/exhibit991pressreleaseq120.htm [https://perma.cc/D245-X2BT]; Aditya Kalra & Munsif Vengattil, Apple Airlifted iPhones Worth a Record $2 Billion From India in March as Trump Tariffs Loomed, Reuters (Apr. 15, 2025), https://www.reuters.com/technology/apple-airlifted-iphones-worth-record-2-billion-india-march-trum
    p-tariffs-loomed-2025-04-15/ [https://perma.cc/B74C-KAVH]; Evolus, Inc., Current Report (Form 8-K) (Aug. 5, 2025), https://www.sec.gov/Archives/edgar/data/1570562/000157056225000094/q22025
    eolsex991.htm [https://perma.cc/JSD8-CS8H]; Helios Technologies, Inc., Annual Report (Form 10-K) (Feb. 2, 2025), https://www.sec.gov/Archives/edgar/data/1024795/000095017025026703/hlio-2024
    1228.htm [https://perma.cc/WF78-HH3R]; Mercedes-Benz Stocking Inventory in U.S. Ahead of Tariffs, Analyst Note Says, Reuters (Apr. 1, 2025), https://www.reuters.com/business/autos-transportation/
    mercedes-benz-stocking-inventory-us-ahead-tariffs-analyst-note-says-2025-04-01/ [https://perma.cc/9
    4NS-N4S3]. ↑
36. . Murray, supra note 3. ↑
37. . See, e.g., sources cited supra note 34. ↑
38. . Jafar Namdar, Sachin Modi & Jennifer Blackhurst, Diversify or Concentrate? Supply Chain Responses to Policy Uncertainty, 61 J. Supply Chain Mgmt. 62, 66 (2025). ↑
39. . Grossman et al., supra note 13, at 988, 1024. ↑
40. . See, e.g., Tennant Co., Quarterly Report (Form 10-Q) (Nov. 4, 2025), https://www.sec.gov/Archives/edgar/data/97134/000009713425000036/tnc-20250930.htm [https://perma.cc/WQ55-BLHE] (describing “targeted supply chain initiatives, including . . . dual sourcing, and logistics shifts . . . to mitigat[e] tariff-related cost inflation”); Omnicell, Inc., Quarterly Report (Form 10-Q) (Nov. 5, 2025), https://www.sec.gov/Archives/edgar/data/926326/0000926326250
    00031/omcl-20250930.htm [https://perma.cc/T6X8-5TR2] (stating that, “[i]n an effort to address” fluctuating tariffs, the company “implemented various mitigation measures, including dual-sourcing of components and nearshoring manufacturing”); Universal Elecs. Inc., Annual Report (Form 10-K) (Mar. 11, 2025), https://www.sec.gov/Archives/edgar/data/101984/000010198425000038/ueic-20241231
    .htm [https://perma.cc/F8Z3-8W4Y] (explaining a strategy to “de-risk [the company’s] reliance on a [Chinese]-based supply chain,” to evaluate “additional third-party manufacturers and sources of supply,” while noting that it previously “maintained duplicate tooling for certain of our products”). ↑
41. . See, e.g., Ford Motor Co. v. United States, 254 F. Supp. 3d 1297, 1303 (Ct. Int’l Trade 2017), rev’d, 926 F.3d 741 (Fed. Cir. 2019); Kevin J. Fandl, 2019 International Trade Law Decisions of the Federal Circuit, 69 Am. U. L. Rev. 1301, 1302 (2020); Andrew O’Brien-Penney, The Changing Landscape of Tariffs and Taxation, and What Companies Can Do, Tax Notes, Oct. 23, 2025, at 2, 7–10, https://www.taxnotes.com/special-reports/tariffs/changing-landscape-tariffs-and-taxation-and-what
    -companies-can-do/2025/10/22/7t6hb [https://perma.cc/ZKQ2-VTW2]. ↑
42. . See Fandl, supra note 40, at 1302, 1314–16. ↑
43. . Id., at 1302. ↑
44. . Id. ↑
45. . Handley & Limão, supra note 18, at 190–91. ↑
46. . Id. ↑
47. . In this context, it is important to distinguish headline announced tariff rates from realized effective tariff incidence. The Penn Wharton Budget Model estimated that the average effective U.S. tariff rate, measured as customs duties as a share of import value, was 9.8 percent through December 2025 and 8.9 percent through February 2026. See Effective Tariff Rates and Revenues, Penn Wharton Budget Model, https://budgetmodel.wharton.upenn.edu/p/2026-04-15-effective-tariff-rates-and-revenues-updated-april-15-2026/ [https://perma.cc/224D-NHCY]. Bloomberg’s reporting points in the same direction, but on a different metric: Rather than measuring tariff incidence, it tracked tariff threats and trade actions, reporting in January 2026 that only about one quarter of 49 tariff threats or trade investigations since November 2024 had been carried out in full. Murray, supra note 3. Bloomberg’s updated tracker in June 2026 similarly stated that, since November 2024, Trump had followed through on some major threats, but that most had been watered down or never implemented. See Bloomberg, Tracking Trump’s Tariffs Across the Global Economy (June 8, 2026), https://www.bloom
    berg.com/graphics/trump-tariffs-tracker/ [https://perma.cc/D6TB-HRXN]. ↑
48. . Kyle Bagwell, Chad P. Bown & Robert W. Staiger, Is the WTO Passé?, 54 J. Econ. Lit. 1125, 1169 (2016); Alberto Osnago, Roberta Piermartini & Nadia Rocha, Trade Policy Uncertainty as Barrier to Trade 3 (World Trade Org. Working Paper No. ERSD-2015-05, 2015), https://www.wto.org/
    english/res_e/reser_e/ersd201505_e.pdf [https://perma.cc/SS24-WVTD]. ↑
49. . Caldara et al., supra note 7, at 39. ↑
50. . See, e.g., Tasmin Lockwood, Novartis’ U.S. Deal Could Shield It from Tariffs, CEO Tells CNBC, CNBC (Jan. 21, 2026), https://www.cnbc.com/2026/01/20/possible-us-deal-to-shield-novartis-from-tariffs-ceo-tells-cnbc.html [https://perma.cc/YM66-PY7Y]. ↑
51. . Sasatra Sudsawasd & Robert E. Moore, Investment Under Trade Policy Uncertainty: An Empirical Investigation, 12 Rev. Int’l Econ. 316, 323–25 (2006); Sangyup Choi, Davide Furceri & Chansik Yoon, Policy Uncertainty and Foreign Direct Investment, 29 Rev. Int’l Econ. 195, 200 fig.1, 220 tbl.10, 221–22 (2019). ↑
52. . See, e.g., Bureau of the Fiscal Serv., U.S. Dep’t of the Treasury, Monthly Treasury Statement of Receipts and Outlays of the United States Government 4 Fig.1 (May 2026), https://fiscaldata.treasury.gov/datasets/monthly-treasury-statement/summary-of-receipts-outlays
    -and-the-deficit-surplus-of-the-u-s-government [https://perma.cc/K8B9-HJ3M]. ↑
53. . See, e.g., Avi-Yonah et al., From Relic to Relevance, supra note 1, at 512. ↑
54. . Cavallo et al., supra note 10, at 26–31. ↑
55. . Press Release, Bd. of Governors of the Fed. Rsrv. Sys., Federal Reserve Issues FOMC Statement (Dec. 10, 2025), https://www.federalreserve.gov/newsevents/pressreleases/monetary2025
    1210a.htm [https://perma.cc/3PZ9-2BT7] (“Available indicators suggest that economic activity has been expanding at a moderate pace. Job gains have slowed this year, and the unemployment rate has edged up through September.”). ↑
56. . Veronika Penciakova, Valerie Nguyen, Camelia Minoiu & Lauren Taylor, Federal Reserve Bank of Atlanta, Are US Importers Ready for the New Tariff Landscape?, Policy Hub: Macroblog (Aug. 26, 2025), https://www.atlantafed.org/research-and-data/publications/policy-hub-macroblog/
    2025/08/26/are-us-importers-ready-for-new-tariff-landscape [https://perma.cc/4MRW-CEVE]. ↑
57. . Edward J. Markey, U.S. Senate Comm. on Small Bus. & Entrepreneurship, The Trump Tariffs: A Small Business Crisis 4–7 (Apr. 17, 2025), http://www.sbc.senate.gov/public/_cache/files/a/7/
    a760a470-5f3b-40ca-b728-634359e07490/DB068FE37D91B435A2A5129F16312542E67F495E3EE6
    8150167EF9FCC14ADE1A.the-trump-tariffs-a-small-business-crisis.pdf [https://perma.cc/2VYG-5Z
    7C]; Penciakova, supra note 55. ↑
58. . Pablo D. Fajgelbaum, Pinelopi K. Goldberg, Patrick J. Kennedy & Amit K. Khandelwal, The Return to Protectionism, 135 Q.J. Econ. 1, 1 (2020). ↑
59. . It should be noted that, though the duty is typically remitted by the importer at the point of entry, legal incidence is not the same as economic incidence. A tariff increases the domestic price of the imported good relative to substitutes; that price wedge can reduce quantity demanded and shift part of the burden to foreign exporters through lower received prices, lost sales, or both, depending on demand and supply elasticities and the availability of alternative sources. ↑
60. . See supra notes 36–37. ↑
61. . See Handley & Limão, supra note 18, at 191. ↑
62. . See, e.g., VW Chief Blume: New Audi Plant in US Not Feasible with Tariffs, Yahoo Fin. (Jan. 25, 2026), https://finance.yahoo.com/news/vw-chief-blume-audi-plant-102835297.html [https://
    perma.cc/L3MA-FMBN]. ↑
63. . See, e.g., David Ferris, Hyundai Pledges $21B in US Investment to Avoid Tariffs, E&E News by POLITICO (Mar. 25, 2025), https://www.eenews.net/articles/hyundai-pledges-21b-in-us-investment
    -to-avoid-tariffs/ [https://perma.cc/2VX6-AQBN]. ↑
64. . See, e.g., Peggy Corlin & Maria Tadeo, What Is the EU’s Anti-Coercion Instrument, and How Does It Work?, Euronews (Jan. 18, 2026), https://www.euronews.com/my-europe/2026/
    01/18/what-is-the-eus-anti-coercion-instrument-and-how-does-it-work [https://perma.cc/SX6Y-76HF]; Didi Tang & Sam McNeil, Pushed by Trump, U.S. Allies Are Resetting Relations with China, Associated Press (Jan. 30, 2026), https://apnews.com/article/china-trump-europe-allies-trade-greenland-ff445f3c6e2c52d6229d05441aae23f0 [https://perma.cc/AUW4-AXAD]; Jakub Krupa, EU-India Deal ‘Accelerated’ over Past Six Months Amid Trump’s Tariff Threats – Europe Live, The Guardian (Jan. 27, 2026), https://www.theguardian.com/world/live/2026/jan/27/european-union-india-trade-deal-ukraine-greenland-holocaust-news-updates-europe-live [https://perma.cc/UM9F-4RJ2]. ↑
65. . See Caldara et al., supra note 7, at 38–39; Handley & Limão, supra note 18, at 214; Osnago et al., supra note 47, at 15. ↑
66. . Trade Compliance Resource Hub, Trump 2.0 Tariff Tracker, https://www.tradecompliance
    resourcehub.com/2026/01/27/trump-2-0-tariff-tracker/ [https://perma.cc/55DE-KVKV]. ↑
67. . See Narotzki, Hidden Taxation, supra note 4, at 84. ↑
68. . Id., at 72–74; Christopher T. Zirpoli, Cong. Rsch. Serv., R48435, Congressional and Presidential Authority to Impose Import Tariffs 9–24 (2025). ↑
69. . See Narotzki, Hidden Taxation, supra note 4, at 94–96; Narotzki, Tariffs Are Taxes, supra note 4, at 32–34. ↑
70. . Narotzki, Hidden Taxation, supra note 4, at 89. ↑
71. . Id. at 90. ↑
72. . Learning Res., Inc. v. Trump, 607 U.S. 229, 238–39 (2026). ↑
73. . Id. at 255. ↑
74. . See, e.g., Bessent Expects Supreme Court to Uphold Legality of Trump’s Tariffs but Eyes Plan B, CNBC (Sep. 1, 2025), https://www.cnbc.com/2025/09/01/bessent-expects-supreme-court-to-uphold-legality-of-trumps-tariffs-but-eyes-plan-b.html [https://perma.cc/Y9LA-54DG]. ↑
75. . Learning Res., 607 U.S. at 354–55 (Kavanaugh, J., dissenting). ↑
76. . See generally Transcript of Oral Argument, Learning Res., Inc. v. Trump, 607 U.S. 229 (2026) (No. 24-1287) and Trump v. V.O.S. Selections, Inc., 607 U.S. 229 (2026) (No. 25-250). ↑
77. . Atmus Filtration, Inc. v. United States, No. 1:26-cv-01259, 2026 WL 679285, at *1–2 (Ct. Intl. Trade Apr. 8, 2026); U.S. Customs & Border Protection, CSMS #68340863, UPDATE – Consolidated Administration and Processing of Entries (CAPE) for IEEPA Refunds, April 20, 2026 Deployment (Apr. 13, 2026). ↑
78. . See Narotzki, Hidden Taxation, supra note 4, at 71. ↑
79. . U.S. Const. art. I, § 8. ↑
80. . Adam Looney & Elena Patel, Why Does the Executive Branch Have So Much Power over Tariffs?, Brookings Inst. (Jan. 15, 2025), https://www.brookings.edu/articles/why-does-the-executive-branch-have-so-much-power-over-tariffs [https://perma.cc/KLT2-XM7H]; see also John H. Jackson, The General Agreement on Tariffs and Trade in United States Domestic Law, 66 Mich. L. Rev. 249, 254 n.29 (1967). ↑
81. . See Narotzki, Hidden Taxation, supra note 4, at 104–106. ↑
82. . Transcript of Oral Argument at 65, Learning Res., Inc. v. Trump, 607 U.S. 229 (2026) (No. 24-1287) and Trump v. V.O.S. Selections, Inc., 607 U.S. 229 (2026) (No. 25-250). ↑
83. . See Narotzki, Hidden Taxation, supra note 4, at 105–06. ↑
84. . See generally Caldara et al., supra note 7. ↑

Abstract

This Essay accepts strict scrutiny as the constitutional terrain within which race-conscious policy must operate and builds a refined framework for equal protection analysis: the Historical Racial Nexus (“HRN”). HRN offers a narrow tailoring theory designed to survive contemporary judicial scrutiny by charting a third path between colorblind formalism and categorical race-based remedialism. Rather than treating race as a present-day identity classification, HRN reconceptualizes race as historical evidence of specific, documented government-inflicted injury. Under this approach, constitutionally permissible race-conscious policy is grounded in temporally and geographically defined acts of state action and ties eligibility for redress to demonstrable causal connections to that harm.

In a legal landscape marked by increasing judicial hostility to race-based remedies—Students for Fair Admissions v. Harvard is a notable example—HRN provides a constitutionally grounded reframing that uses race as an evidentiary lens to trace and remedy state-sponsored discrimination. Unlike policies that rely on broad identity categories or generalized claims of social inequality, HRN anchors reparative legislation in historically specific government conduct.

Through retrospective analysis of the Evanston, Illinois municipal reparations initiative and the federal Civil Liberties Act of 1988, this Essay demonstrates how policies structured around an identifiable historical nexus can satisfy contemporary formulations of strict scrutiny by articulating a compelling governmental interest, remedying past constitutional injury, and employing narrowly tailored means. The Essay situates HRN within equal protection jurisprudence and argues that the framework aligns with the Supreme Court’s doctrinal requirements while preserving space for democratically enacted, historically accountable policymaking.

HRN also addresses a longstanding tension between social scientific understandings of race as fluid and socially constructed and the law’s tendency to treat race as fixed and essential. By distinguishing race as identity from race as historical evidence, HRN offers a more contextually grounded and normatively coherent model for race-conscious policy. In doing so, it reimagines how equal protection can meaningfully fulfill its remedial promise, grounding race-conscious legislation in history, causation, and accountability while expanding the constitutional imagination.

Introduction 

The use of racial criteria in governmental policy is subject to a great deal of conceptual confusion, which often leaves it unlinked to social and historical context.^[1] This Essay accepts strict scrutiny (bracketing a great deal) as the terrain to be navigated and develops a refined framework for equal protection analysis grounded in a historical nexus between governmental harm and racial specificity. The Historical Racial Nexus (“HRN”) offers an alternative basis for redress, distinct from race as a present-day identity category, by anchoring reparative policy in specific, documented instances of state-sanctioned racial discrimination. Unlike broader race-conscious measures that rely on generalized claims of inequality or present racial status, HRN-based policies tie eligibility to a direct evidentiary link between identifiable government action and the individuals or descendants affected by that harm. ^[2]

Questions about the use of race in policy frameworks which have emerged following the Supreme Court’s interpretation of Strict Scrutiny in Students For Fair Admission v. Harvard (2023) (“SFFA”) raise important provocations to prevailing judicial understandings of the relationship between race, equal protection, and strict scrutiny.^[3] This Essay retrospectively examines the use of race in two reparations^[4] policy frameworks—the recent municipal housing reparations framework developed in Evanston, Illinois,^[5] and the policy framework developed in the federal Civil Liberties Act of 1988^[6]—to argue for a refined conceptual use of race in reparatory legislative policy as one of HRN. It also addresses a considerable source of confusion in law around the construction of race: Social scientists have often stated that while race is not a fixed category,^[7] the effects of racism are profound and intersectional.^[8] To contrast, the law often presumes that race is a fixed identity category instead of a social construction with porous and ever-changing boundaries, and it often fails to account for intersectional complexities.^[9]

The HRN framework realizes a conception of race that reframes the use of race not as providing advantage vis-a-vis racial identity^[10] or as a proxy^[11] for generalized disadvantage, but as an evidentiary marker of government inflicted past harm.^[12] Under such a model, race is not the basis on which a benefit is conferred at the exclusion of others, but rather it is one aspect of identification of those who were situated in the path of deliberate and documented government action that caused measurable injury.^[13]

For example, race-based harm resulted from explicit governmental choices such as those related to discriminatory zoning,^[14] redlining,^[15] deed restrictions,^[16] financing,^[17] inheritance,^[18] and property seizure.^[19] Remedial and reparatory policies would benefit by being able to reference the racial aspects of this harm, which are readily identifiable. Identifying the racialized nature of harm in such circumstances does not confer advantage upon individuals based on racial identity,^[20] but acknowledges past government misconduct or mistakes and its lingering, systemic effects. By linking redress to documented governmental action, HRN offers a path for government redress that is distinct from the basis of the affirmative action policies that have fallen into disfavor in the Supreme Court.^[21]

The Equal Protection Clause of the Fourteenth Amendment prohibits racial classifications in policy unless they are justified by a compelling government interest and narrowly tailored to achieve that interest.^[22] In practice, this has meant that nearly all race-conscious policies that have come before the Supreme Court have been determined to be unconstitutional.^[23] However, the Supreme Court continues to say that the application of strict scrutiny is not an impossible standard, even when the pathways for constitutional policies are indeed ever narrower.^[24] The HRN framework sheds new light on constitutional arguments around narrowly tailored, harm-based reparative measures and whether they can withstand equal protection scrutiny.

Thus, HRN is a critical refinement of the legal understanding of how race can inform remedial policy without violating constitutional constraints. It suggests a framework in which race is not used to provide an abstract or merely categorical basis for benefits, but as evidence of a traceable state-driven injury. This refinement matters. Under traditional Equal Protection doctrine, policies that make distinctions based on race are presumed suspect and functionally “dead on arrival.”^[25] The use of HRN, however, uses race to mark a harm that is anchored in time, place, and government action. Conceptually different from current or contextual identification with “race,” the HRN allows policy makers to address the harms of past racial discrimination in a way that aligns with constitutional principles, emphasizing causation, documentation,^[26] and specificity as opposed to broad racial generalizations. In doing so, HRN offers an objective approach to racial redress, which may be more resilient in the face of judicial scrutiny.

I. The Historical Racial Nexus: Two Retrospective Explorations.

Put simply, the HRN serves as a framework through which policymakers can design reparative measures for past injustices with racial contours without triggering the constitutional pitfalls associated with race-based classifications.^[27] While traditional remedial policies often rely on present-day racial identity (i.e., race) as a proxy for systemic harm,^[28] this Essay advocates for the use of the HRN as a more precise and legally resilient alternative.^[29] Rather than asking whether an individual is currently a member of a racial group, the HRN inquires whether that individual or their ancestors was harmed by specific, identifiable government actions that targeted racial classifications within a particular historical and social context.

By grounding redress in documented instances of state-inflicted racial harm, the HRN allows for the development of policies that are anchored in history and context rather than identity. To put it another way, the use of HRN would be used as evidence of past harm, rather than race in the present sense being used as the basis for a benefit. The policy, then, is structured around historical harm, using factors like time, place, ancestry, and documented policy records to define who was affected and who should be eligible for redress. In doing so, the framework offers a constitutionally sound, narrowly tailored approach to racial redress that aligns with equal protection doctrine while preserving the possibility of meaningful historical accountability.

Presently, equal protection analysis hinges on a level of scrutiny determination based on what particular kind of personal characteristics that claims of discrimination are formulated in.^[30] Race is one of the categories that the court has determined should be subject to the highest level of scrutiny.^[31] As a result, a government that must defend laws with classifications based on race faces a heavier scrutiny burden than, for example, if a law classified based on gender, which requires only intermediate scrutiny.^[32] Other categories and differentiations face the much lower burden of rational basis review.^[33]

Race is subject to a higher burden in equal protection because policies which include race are thought to also be more likely to be rooted in the type of prejudice that is not supportable in American democracy. Although identity is central to the origins of equal protection and the development of identity categories within that jurisprudence,^[34] the use of identity categories also often undermines the objective of equal protection.^[35]

Nevertheless, there is useful information in identity designations, which is one reason for their continued reliance. When understood as indicators of relational power imbalance, which are historically and contextually situated, identity categories can help reveal patterns of harm. Recently, the use of race as a static category has failed under a formalist, anticlassification regime of interpretation which finds fault with the unmoored use of race—used in terms of present identity—to find entitlement to remedy for systemic injustice.^[36] The HRN framework shifts the use of static racial classifications to racial identity as historical evidence, tied to specific policies which resulted in de facto racial discrimination and, thus, racialized harm. Such a reframing addresses both the critique that categories cannot be used in universalizable ways and that refraining from acknowledging demographic contours of harm will lead to further subordination.^[37] One of HRN’s key contributions is in pointing out that there is a difference between policy which hinges on static identity categories like race and one that hinges on contextual relationships with history and power.

Historical identity categories are able to do work in this context because they are anchored in objective harm–systemic harm can be shown through data analysis, and there are verifiable events, policies, or governmental actions that produce measurable harms.^[38] Thus, the point that the Supreme Court makes in SFFA, for example, that identity categories as currently used are unconstitutional because they are broad,^[39] interminable,^[40] unmeasurable,^[41] and un-linkable with any kind of verifiable specific harm^[42] are avoided. In situating identity within specific frames based in time, policy or geography, what otherwise might be seen as a racial category subject to fail under strict scrutiny, to one of an evidentiary link between state action and injury. Using HRN, as opposed to race, in this way preserves the usefulness of races as a marker for understanding harm while also avoiding the present-day pitfalls of overly broad classifications.

This section explores two examples successful legislation that were framed in terms of ‘race’ at the time of their passage, but which can analyzed used retrospectively to demonstrate the meaning and potential power of HRN^[43]: The recent municipal reparations program created in Evanston, IL, and the Federal Civil Liberties Act of 1988. Both programs described below employ race as a contextual, historically grounded evidentiary marker of government-inflicted harm.

Though these programs are wildly different in scope, scale, design and context, both illustrate how governments might structure reparative policies around a documented historical nexus of state-inflicted racial harms. Rather than using race as a categorical basis for benefit, each program links eligibility to specific historic events in which specific government’s specific actions produced measurable injury. Through grounding such policies in temporally and administratively identifiable instances of harm, the policies model how an HRN may be used to withstand the critiques associated with Constitutional scrutiny.

A. Evanston’s Reparations Program

Beginning in 2020, several municipal governments across the United States began the process of atoning for past racialized harm in response to a renewed interest in racial justice brought by the Black Lives Matter movement and various instances of police violence.^[44] One of the ways that cities focused on ameliorating harm was through reparations. Evanston, Illinois’s initiative, for example, provides $25,000 grants for homeownership, mortgage assistance, or home improvement to eligible recipients.^[45] Evanston’s policy, one of the most successful of its time, has brought legal challenges under the Equal Protection Clause which are ongoing.^[46]

Evanston’s reparations program offers a strong practical illustration of the possibilities of the HRN framework in action. Unlike policies that rely on broad racial categories or current identity, Evanston ties eligibility for reparations to specific, documented acts of city-sponsored racial discrimination in housing between 1919 and 1969.^[47] The program focuses on harms inflicted directly by the municipality—such as redlining, exclusionary zoning, and displacement—which are temporally and geographically defined, well-documented, and explicitly racialized in their historical context.^[48] These harms were identified through dedicated research of the historical record. ^[49]

There are three ways in which one may qualify for the benefits of the Evanston reparations. First, one could have been personally a Black or African American citizen of Evanston during the time in which Evanston identified that the discriminatory policies took place.^[50] Second, one could have a Black or African American ancestor who lived in Evanston at the time that the discriminatory policies took place.^[51] Third, one could show that discrimination occurred in Evanston against them personally or against their ancestor during the time identified by the city, without having to show relation to a Black or African American ancestor.^[52]

By requiring applicants to demonstrate either personal experience of this discrimination or ancestral ties to those harmed as the focus of discriminatory policies, Evanston’s reparations policy centers historical evidence rather than present-day race, aligning closely with HRN’s goal of using race as a marker of state-inflicted injury rather than as a basis for distributing benefits. In doing so, the program may avoid the constitutional vulnerabilities of affirmative action-style remedies and instead creates a narrowly tailored, compelling model that can be meaningfully reviewed and supported.^[53] Evanston thus tees up HRN exceptionally well, offering both a legally defensible and administratively feasible approach to reparations that could inform future reparations programs.^[54]

Evanston’s program requires applicants to prove that their Black or African American ancestors resided in Evanston at the time the city enacted and enforced racially discriminatory housing policies.^[55] However, the policy accounts for the systemic discrimination, which the city government supported, by assuming that any Black individual in the city at that time would have been impacted. Because the program assumes that all Black residents at the time faced systemic discrimination, the need to prove individual instances of harm is eliminated. This is the essence of HRN and the way that it links equal protection to consideration of systemic harm along a historical nexus.

Despite offering a race-neutral path to eligibility for reparations, Evanston’s program has been subject to an ongoing lawsuit in the form of an equal protection challenge.^[56] The Plaintiffs in the challenge have stipulated that the requirement of having a Black or African American ancestor, applicable to two of three eligibility categories (a third being race neutral but with a higher evidentiary burden), constitutes a constitutionally impermissible racial classification that excludes non-Black individuals who may have experienced comparable harm.^[57] They further argue that this increases the burden placed on non-Black claimants which raises further concerns of disparate treatment in violation of Equal Protection. ^[58] They claim that

At no point in the application process are individuals in the first and second groups required to present evidence that they or their ancestors experienced housing discrimination or otherwise suffered harm because of an unlawful Evanston ordinance, policy, or procedure, or some other unlawful act or series of acts by Evanston between 1919 and 1969. In effect, Evanston is using race as a proxy for having experienced discrimination during this time period.^[59]

The fallacy in this argument may also be explained through the lens of HRN. Evanston has identified a compelling government interest in remediating past government-inflicted harm (in the form of racial discrimination), and this interest is well-established as being compelling.^[60] The HRN is a tool through which this interest may be described–there is a historical nexus between the government harm and the systemic contours of the harm which fall along the lines of race.^[61] They have also narrowly tailored a response to those who were harmed or whose ancestors were harmed through specific discriminatory government policies, along with systemic harm that resulted specifically from them. This is not done in terms of a ‘category’ of ‘race’, which might include any Black individual who happens to have settled in Evanston. The HRN is quite different in that it invokes a limiting principle in the historical nexus, only those individuals who were affected by those policies and their ancestors who were in turn affected by those policies may qualify for reparations. The historical nexus between race and government action definitively shows the coordinates of repair.

Evanston’s reparations program is one example of how the use of an HRN can operate as a constitutional bridge between race conscious reparatory policy and equal protection doctrine. The program grounds eligibility in a documented history of local municipal racially discriminatory policy, as opposed to present racial identity. Though the program’s use of HRN has still invited legal challenges on equal protection grounds, its structure reflects a shift away from race as a categorical preference and towards the use of race as an evidentiary marker of government-inflicted harm. Thus, the use of race is used as a record of injury rather than one that confers benefit to the exclusion of others just as deserving. The judicial fate of the Evanston initiative remains to be seen, but the use of HRN remains open in the narrow pathway through strict scrutiny.

B. The Federal Civil Liberties Act of 1988^[62]

While the Evanston reparations program offers a contemporary paradigm for the use of HRN, there are other examples where conceptualizing HRN, as opposed to race, is useful. Nearly four decades before Evanston created its reparations program, the federal government of the United States enacted the Civil Liberties Act of 1988 (CLA)^[63], which also, I again identify in retrospective analysis, grounded its reparative framework through a specific instance of historical, government-perpetrated racial harm. Although the scale and context of the program is different than that in Evanston, the underlying logic remains consistent. Both policies tie eligibility for the redress offered by the policy to the direct relationship between state action and racialized injury.^[64] This section examines the CLA as an early embodiment of the framework identified in this Essay as one of HRN.

The CLA is one of the most explicit acknowledgments by the Federal Government of its direct role in racial injustice. Enacted over 40 years^[65] following the internment of Japanese Americans during the second world war, the CLA authorized $20,000 in reparations to surviving individuals^[66] who were interned, as well as a formal apology issued by the government of the United States.^[67] The passage of the CLA meant that the US government formally conceded to the racially discriminatory nature of its own conduct, and also linked that to a narrowly tailored policy scheme which provided redress.

The CLA only occurred after years of intensive political pressure, including public hearings and findings from Congressionally established commissions.^[68] The reports issued from these commissions noted that individuals were interned because of “racial prejudice, war hysteria, and a failure of political leadership”.^[69] In response, the CLA was grounded in the official historical record of racial animus and financial loss.^[70]

Crucially, the CLA’s racial classification was not viewed as presumptively unconstitutional. Courts reviewing the law upheld the racial classification under equal protection principles finding that it satisfied strict scrutiny because of its clear remedial purpose and narrow tailoring. The CLA focused on a documented, state-inflicted, race-based harm. The CLA thus set both legal and political precedent for policy frameworks that seek redress for historical racial harm through legislative action.

In one case which challenged the CLA on equal protection grounds, Obadele v. United States,^[71] Black plaintiffs argued that African Americans similarly deserved redress for slavery and other systemic discrimination. The court rejected their claim, stating that the plaintiffs did not meet the statutory eligibility criteria.^[72] More specifically, they lacked the Japanese ancestry and had not been subject to the specific government actions addressed by the CLA.^[73] The court emphasized that while the plaintiffs may have suffered historical injustice, redress for that harm had not yet been won through the political process.^[74] The decision underscored that the reparations under the CLA were not based on a general entitlement to justice, but on a narrowly defined, legislatively recognized class of victims tied to a specific government action. The court effectively affirmed that reparations are possible and lawful, but only when grounded in clear legislative intent and historical specificity, a point that reinforces the utility of an HRN framework.

HRN moves away from using race as a present identity marker and instead uses specific, documentable acts of discrimination by the government to define eligibility for redress. By rooting reparations in provable historical government wrongdoing, HRN mirrors the successful logic of the CLA and avoids the constitutional pitfalls that doomed broader claims like those in SFFA. In short, Obadele shows that reparations can be constitutional when they are tightly tied to historical governmental acts even when they are limited by considerations of race, exactly what the HRN is designed to do.

While the Evanston program and the CLA differ in many ways, the two policies share a common constitutional logic. Both frameworks operate within the contours of what this Essay identifies as the HRN. As with the Evanston program, the CLA illustrates that race, when not used as a mere proxy for identity but rather as a marker for injury caused by state action, can form a constitutional basis for race-conscious remedial policy. Such policy does not suggest racial preference, but rather explicit accountability achieved through constitutional and democratic legislative processes.

II. From Race To The Historical Racial Nexus

There is excellent critical literature on the quest of ‘making equal protection protect’, and the various problems and inconsistencies implicit in the major views on equal protection.^[75] For example, Lauren Sudeall argues that the equal protection doctrine relies too heavily on identity-based categories such as race or gender, and that doing so flattens and oversimplifies complex experiences.^[76] If identity is used as a proxy for discrimination, it is seemingly impossible for equal protection law to address the substantive harm of discriminatory policies.^[77] Similarly, HRN is one practical way in which the strict and judicially unpopular categories of race might be discarded in favor of categories that better address substantive harm.

One issue further implicit to the use of strict categories in interpreting and applying equal protection is how race is treated in the social sciences versus how it is treated in law. In social science, race is not understood as biologically determined or as part of a fixed identity, but as socially constructed by a variety of historical and contextual forces.^[78] Of course, race continues to have salience in terms of the consequences of racism and discrimination despite its fluidity, though legal doctrine often treats race as objective and stable. Such a mismatch hearkens to the mismatch which Sudeall explores –systemic discrimination is simply not well-reflected within insular identity categories.

The HRN framework addresses this mismatch, at least in part, by reframing the use of race in certain policies not as static classification, but as historical evidence of state harm. Using race in this way acknowledges that race is socially constructed while still recognizing the concrete nature of racial discrimination. One divisive issue in reparations policy is the question of deservedness, and HRN also helps to address this by allowing a shift away from deservedness based in static identity categories, and towards asking questions concerning, for example, which individuals were harmed by state actions that targeted people racialized as Black at a particular time in history.

Through using HRN—a historically situated indicator of injustice unconnected to an individual’s present identification with racial categories–as opposed to an individual trait, HRN offers a more nuanced way of including race in policy making. This refocuses policy on the relationship between race and state power, further opening a space for addressing racial harm through law without reproducing the same inequality the equal protection clause is meant to ameliorate.

III. Advantages and Critiques of The Historical Racial Nexus

While the HRN framework offers a promising doctrinal innovation, it is not without limitations. Advantages and critiques of HRN, at least as it exists in theory, are discussed in further detail below.

A. Advantages of HRN

Situating the HRN framework within equal protection jurisprudence shows how it might address Constitutional concerns that have previously led to invalidation of more generalized race-conscious policies like affirmative action. HRN satisfies strict scrutiny by directly aligning with the constitutional requirements established in key Equal Protection jurisprudence. In Adarand Constructors v. Peña, it is iterated that race-conscious policies must serve a compelling governmental interest and be narrowly tailored to that interest.^[79] HRN meets this standard through the grounding of reparative efforts in specific, documented acts of government-inflicted racial harm, unlike claims that are grounded in more generalized claims of societal discrimination. This distinction is critical, as Regents of the University of California v. Bakke^[80] and City of Richmond v. Croson^[81] both emphasize that policies aimed at remedying vague, broad patterns of discrimination fail strict scrutiny. Rather, HRN identifies precise historical actions, such as those related to redlining, urban renewal, or exclusionary zoning, carried out by specific governmental entities, and which ties eligibility to individuals demonstrably affected by that conduct.^[82]

For example, Evanston’s reparations program, which limits benefits to those harmed by the city’s housing policies between 1919 and 1969, mirrors the kind of concrete, historically rooted redress that courts have previously upheld, such as in Obadele v. United States, where reparations for Japanese American internment were deemed constitutionally sound, even against an equal protection challenge, because they were based on specific, documented government actions and resulted from a legitimate democratic practice.^[83]

The HRN framework also fulfills the narrow tailoring requirement by using historically defined criteria, such as time, place, ancestry, and government records, to identify beneficiaries, rather than relying on present-day racial identity. This approach avoids the overbreadth that the Supreme Court rejected in Students for Fair Admissions v. Harvard, where race-conscious admissions policies were struck down for lacking measurable goals and for relying on imprecise racial categories.^[84] By using HRN to define a class of harmed individuals through historical evidence rather than race per se, reparatory policies are both more legally durable and more consistent with constitutional principles. In this way, HRN offers a doctrinally sound model for racial redress that remains viable even under the increasingly constrained interpretations of Equal Protection.

The HRN framing re-invigorates the role of race in law not as means of categorizing or privileging, but as a historical signal of state-inflicted harm. In so doing, HRN responds to specific doctrinal constraints in equal protection jurisprudence. At the same time, the framework preserves the possibility of legislative redress.

Beyond its constitutional utility, HRN offers a deeper, more accurate way of understanding the relationship between race and equal protection. One of the persistent challenges in legal and policy discourse is the reliance on race as a static identity category which does not adequately reflect inequality.^[85] Courts have increasingly rejected the use of race, particularly when it is used as a stand-in for disadvantage or harm without clear evidentiary grounding. HRN addresses this issue by shifting the focus from who someone is now to what happened to them or their ancestors through state action. It also further reinforces the legal principle that remedies must be rooted in injury, not inferred from group membership, which is also one of the central barriers to enaction of reparatory policies.^[86]

Finally, HRN helps to reconcile tensions between legal formalism and the social complexity of identity, systemic discrimination, and state action.^[87] A historical lens not only makes policies more legally defensible but also acknowledges the ways in which the state has used race to distribute harm. In turn, HRN opens the door for race-conscious but constitutionally grounded policymaking that is rooted in accountability rather than abstraction. It allows governments to repair past injustices without relying on present-day race as a proxy^[88], creating a stronger foundation for democratic legitimacy and social repair.

B. Anticipating Critiques of HRN

There are many possible critiques of HRN, some of which are enumerated here. Perhaps most relevantly, although HRN avoids using race as a basis for benefits, critics may argue that it still implicitly relies on race. Because applicants must demonstrate that harm occurred due to their or their ancestor’s racial classification, some courts may interpret HRN as a subtler form of race-conscious policymaking. Additionally, in terms of, e.g., the Evanston reparations, courts may question whether post-hoc rationalizations of policy design, particularly those not explicitly framed in HRN terms at the time of enactment, can satisfy strict scrutiny’s demand for narrow tailoring now. Thus, the conceptual distinction between race as evidence and race as classification may prove unstable in practice, especially if courts are inclined to treat ancestry-based qualifications as racial classifications.

Furthermore, administrability challenges are a central consideration in remedial policymaking.^[89] Records which demonstrate ancestral race, for example, may be difficult to find or may not exist, and may greatly increase the time and cost of administering a program of redress.

Additionally, HRN’s focus on specific, limited acts of harm may be seen as too narrow and potentially excluding broader systemic injustices that may be harder to trace to a single government policy. The arguably technocratic focus of HRN may be insufficiently responsive to the full moral and structural scope of racial injustice. Finally, there is the risk that HRN could justify only modest or symbolic reparations, such as Evanston’s $25,000 housing grants, which, while perhaps legally defensible, may fall far short of addressing generations of wealth deprivation and state-sanctioned exclusion. Thus, while HRN offers a constitutionally grounded path forward, its implementation must be carefully designed to ensure that it delivers meaningful redress to systemic harm.

HRN acknowledges the central critiques of critical race theory by rejecting both static racial categories and race-neutrality as an adequate framework for redress and instead centers race contextualized as a marker of historically specific harm. By grounding remedies in state-inflicted injury, HRN offers an approach that retains historical accountability while navigating the narrow doctrinal space currently yet available. However, under the CRT framework, criticism might find issue with the accommodation of judicial hostility to race-conscious policy that HRN essentially makes.^[90]

Future scholarship should explore how HRN can be operationalized in legislation and public discourse, and how it might be integrated with race-neutral policy instruments that nonetheless aim to repair racialized harms.

Conclusion 

As the narrow path for race-conscious policy continues to narrow, the call for new legal frameworks becomes urgent. HRN, developed here only in its most core theory, should be considered by the judiciary as part of equal protection analysis. It should also be considered by legislatures in designing programs that they wish to survive judicial review. Although this Essay articulates the HRN as a framework that may traverse such doctrinal ground, it may also address some of the key moral and social concerns around justice and fairness that have been articulated in the equal protection jurisprudence. The HRN framework reframes the use of race not as one of categorical identity upon which generalized remedies may be awarded, but as an evidentiary aspect of government and policy-inflicted discrimination. Thus, HRN distinguishes itself from the broad, identity-based approaches that have failed to pass strict scrutiny in the past.

Both the Evanston municipal reparations program and the CLA exemplify how HRN may function in practice. These initiatives provided redress to beneficiaries not because beneficiaries are members of specific racial groups, but because their eligibility is tied to historical state action which itself caused racially associated harm. Neither of these initiatives deploy race as a proxy for disadvantage, rather they use race as a tool for historical understanding. This specifically enables narrowly tailored reparative measures that correspond directly to governmental wrongdoing. They are able to then satisfy the core demands of strict scrutiny.

Decisions such as those in Obadele and Students for Fair Admissions^[91] delineate a doctrinal boundary between race-conscious policies which rely on abstracted notions of diversity or general social inequality, and those that are created in response to specific, documented government-inflicted harm. Obadele further points out that it is through the legislative process that these are legitimated–this is democracy at work and to deprive governments of the capability of making redress would be to curtail legitimate democratic processes.^[92] HRN offers a principled approach to remedial policy that is both historically rigorous and legally resilient.

Some programs which are functionally created around concepts of redress have taken a race neutral approach. For example, the Chicago police reparations^[93], though aimed at redress to communities of color, is a functionally race neutral policy. In considering how to proceed with remedial policy, this tactic should be considered alongside HRN. However, HRN may still be useful to establish historical causation, and other proxies (like zip code or socioeconomic data points) might be otherwise insufficient at capturing governmental harm.

The HRN framework elaborated here has the distinct potential to guide equitable policy in other domains marked by government malfeasance. As state and local governments confront historical legacies, the HRN framework is able to provide vocabulary and constitutional architecture for doing so within the narrow bounds of constitutional permissibility as it is currently (even if not ideally) defined. The HRN framework allows race to be acknowledged without being reified. As such, HRN presents the possibility of bridging the gap between past injustice and accountability, clearing the way for reparative justice that is grounded, lawful, and long overdue.

1. . See generally, Benjamin Eidelson & Deborah Hellman, Unreflective Disequilibrium: Race-Conscious Admissions After SFFA, 4 Am. J.L. & Equal. 295 (2024). ↑

2. . Several scholars have pointed out that reparations can or should be grounded in state-inflicted harm. See, e.g., Eric Yamamoto, Charles Ogletree & Reva Siegel. ↑

3. . See, e.g., Issa Kohler-Hausmann, What Did SFFA Ban? Acting on the Basis of Race and Treating People as Equals, 66 Ariz. L. Rev. 305, 312–14 (2024); Avital Fried, Equal Standards for Equal Protection: Revisiting Race Discrimination in Jury Selection After SFFA, 134 Yale L.J. 709 (2025); Justin Driver, The Cure as Disease: The Conservative Case Against SFFA v. Harvard, 2023 Sup. Ct. Rev. 1, 10 (“SFFA’s approach to the strict scrutiny framework that the Court embraced in Grutter nonetheless represented a sea change from the earlier era.”). ↑

4. . Reparations are legislative actions which seek to provide repair in relation to past government actions. They are the result of a political and democratic process and as such may discriminate against other racial groups in whatever specific atonement the reparations are addressing. See Obadele v. U.S., 52 Fed. Cl. 432, 443 (2002), aff’d, 61 F. App’x 705 (Fed. Cir. 2003). ↑

5. . City of Evanston, 126-R-19 A Resolution: Establishing a City of Evanston Funding Source Devoted to Local Reparations (2019). ↑

6. . H.R. 442, 100th Cong. (1988). ↑

7. . See generally, Khiara M. Bridges, The Dangerous Law of Biological Race, 82 Fordham L. Rev. 21 (2013). ↑

8. . See generally, Mara Loveman & Jeronimo O. Muniz, How Puerto Rico Became White: Boundary Dynamics and Intercensus Racial Reclassification, 72 Am. Socio. Rev. 915 (2007). ↑

9. . See, e.g., Students for Fair Admissions, Inc. v. President & Fellows of Harv. Coll., 600 U.S. 181, 201–210 (2023). In SFFA, race is viewed as an easily identifiable and fixed characteristic as opposed to one that is socially constructed through relations with social power. Id. ↑

10. . Again, this is how the Supreme Court characterizes the affirmative action program in SFFA, for example. See Students for Fair Admission, 600 U.S. at 201–210. ↑

11. . See generally, Lauren Sudeall Lucas, Identity as Proxy, 115 Colum. L. Rev. 1605 (2015) (providing an expanded explanation of how categories are used as proxies in equal protection). ↑

12. . See generally Kimberlé Williams Crenshaw, Race, Reform, and Retrenchment: Transformation and Legitimation in Antidiscrimination Law, 101 Harv. L. Rev. 1331 (1988) (providing a critique of formal equality, upon which historical context is essential in crafting constitutionally permissible remedies). ↑

13. . For example, when the federal government creates aid packages for the victims of natural or man-made disasters, eligibility is limited by the individual’s proximity to a specific event, such as 9/11 or a hurricane. For a description of the 9/11 Victim’s Compensation Fund, see Kyle D. Logue & Kenneth S. Abraham, The Genie and the Bottle: Collateral Sources Under the September 11th Victim Compensation Fund, 53 DePaul L. Rev. 591 (2003) (“… the Fund can be viewed as an innovative alternative to tort liability, analogous to workers’ compensation or auto no-fault insurance, that pays more generous benefits than these systems but (of course) to a more narrowly defined set of beneficiaries.”) ↑

14. . Note, Addressing Challenges to Affordable Housing in Land Use Law: Recognizing Affordable Housing as a Right, 135 Harv. L. Rev. 1104, 1104 (2022) (“[T]he zoning movement also buoyed efforts to separate neighborhoods by race, income, and social class. In Village of Euclid v. Ambler Realty Co., the U.S. Supreme Court justified zoning’s burden on property rights in part by pointing to the necessity of sequestering the new apartment—‘a mere parasite’—for fear that higher-density housing would infect American social values and instigate ‘race suicide.’”). ↑

15. . Danielle Stokes, From Redlining to Greenlining, 71 UCLA L. Rev. 628, 631 (2024) (“Although the Fair Housing Act of 1968 marked the beginning of the end of redlining as a formal practice, scholars have analyzed the many ways in which its legacy of injustice persists today. For instance, research has shown redlining’s impact on education and community development. It has shown the associated lack of public transportation. And it has documented the prevalence of food deserts and other sources of food insecurity.”). ↑

16. . Charles M. Lamb, Housing Discrimination and Segregation in America: Problematical Dimensions and the Federal Legal Response, 30 Cath. U. L. Rev. 363, 402–04 (1981). ↑

17. . Id. ↑

18. . See, e.g., Brenda D. Gibson, The Heirs’ Property Problem: Racial Caste Origins and Systemic Effects in the Black Community, 26 CUNY L. Rev. 172 (2023). ↑

19. . Jill M. Fraley, Eminent Domain and Unfettered Discretion: Lessons from a History of U.S. Territorial Takings, 126 Penn St. L. Rev. 609 (2022). ↑

20. . This is how the majority in SFFA characterized race and affirmative action benefits. See Students for Fair Admissions, Inc. v. President & Fellows of Harv. Coll., 600 U.S. 181, 220 (2023) (“The point of respondents’ admissions programs is that there is an inherent benefit in race qua race—in race for race’s sake.”). ↑

21. . Cara McClellan argues that the same race-based justifications used to strike down affirmative action policies in SFFA are used to maintain existing racial hierarchy while relying on formal principles of colorblindness. See Cara McClellan, Evading a Race-Conscious Constitution, 25 U. Pa. J. Const. L. Online (2023). ↑

22. . See Students for Fair Admissions, Inc., 600 U.S. at 206–07 (“Any exception to the Constitution’s demand for equal protection must survive a daunting two-step examination known in our cases as ‘strict scrutiny.’ Under that standard we ask, first, whether the racial classification is used to ‘further compelling governmental interests.’ Second, if so, we ask whether the government’s use of race is ‘narrowly tailored’—meaning ‘necessary’—to achieve that interest.” (internal citations omitted)). ↑

23. . See generally, James E. Fleming & Linda C. McClain, The Myth of Strict Scrutiny for Fundamental Rights, 12 Dartmouth L. J. 1 (2014) (explaining the relationship between strict scrutiny and equal protection). ↑

24. . See Students for Fair Admissions, Inc., 600 U.S. at 206–07. ↑

25. . See generally, Adam Winkler, Fatal in Theory and Strict in Fact: An Empirical Analysis of Strict Scrutiny in the Federal Courts, 59 Vand. L. Rev. 793 (2006) (showing that strict scrutiny is not necessarily fatal but very strict). ↑

26. . There are administrability problems with documentation, though many recent reparations focus on more documentable harms, such as housing discrimination within the past 100 years. This is a downside. For example, the Evanston reparations program focused on harm between 1919 and 1969. See City of Evanston, supra note 5. ↑

27. . At least as they are currently enumerated by, for example, SFFA. See generally, Students for Fair Admissions, Inc., 600 U.S. at 181. ↑

28. . Reparations, for example for slavery, may consider reliance on present-day racial identifications for Black Americans for redress. In part, this is due to the “sacred intentions” of reparations, which invoke a need for repair and transformation after generations of discrimination and bias. See Cicely L. Fleming, Op-Ed: Why I Voted Against Evanston’s Reparations Program, Chi. Trib., Mar. 27, 2021, [https://perma.cc/EY3N-8NUU]. There are also practical reasons, such as a lack of documentation of those who were enslaved. Who is “deserv[ing]” of this redress—for example, only those whose ancestors were enslaved (impossible to prove), all Black Americans, including recent immigrants? These issues are difficult to transcend in what is necessarily the political process of reparations. See, e.g., Who Should Be Eligible to Receive Detroit Reparations?, Nat’l Afr. Am. Reparations Comm’n (Mar. 7, 2022), [https://perma.cc/W93G-7RB6]. ↑

29. . This concept was initially developed in previous work in Lisa Lucile Owens, An Argument for Housing Reparations, 77 Me. L. Rev. 243, 284–⁠87 (2025). ↑

30. . Winkler, supra note 25, at 821. ↑

31. . Id. at 834. ↑

32. . There are various reasons for this. Much of the jurisprudence around equal protection comes from the civil rights era, which was heavily focused on ameliorating racial inequality. ↑

33. . See generally, Note, When Rational Basis Review Bit, 138 Harv. L. Rev. 1843 (2025). ↑

34. . As Lauren Sudeall Lucas shows, “[e]qual protection’s reliance on identity is understandable in light of its origins. The institution of slavery, from which the Reconstruction Amendments arose, and more modern legal regimes enforcing or permitting overt discrimination, used the law to explicitly oppress individuals sharing certain identity traits. These regimes gave rise to group-based identities, which, in turn, gave rise to group-based social movements. The group-based exclusionary nature of the law solidified the social identities of these movements. Given the centrality of identity to both legal and social frameworks, it was logical for those aiming to change the law to focus on altering the perception of certain identities and, later, on rendering identity an illegitimate basis for legal differentiation.” Sudeall Lucas, supra note 11, at 1616. ↑

35. . Such categories are seen as overly broad or they are only proxies for harm. See e.g., Sudeall Lucas, supra note 11, at 1616. ↑

36. . See Students for Fair Admissions, Inc. v. President & Fellows of Harv. Coll., 600 U.S. 181, 220 (using phrases like “on the basis of race alone,” to indicate present-day racial identification). ↑

37. . For additional discussions on antisubordination versus anticlassification, see, generally, Reva B. Siegel, Equality Talk: Antisubordination and Anticlassification Values in Constitutional Struggles over Brown, 117 Harv. L. Rev. 1470 (2004). ↑

38. . A recent wave of municipal reparations has focused initially on research which heavily documents racialized harm because of specific, municipal government action. See, e.g., Providence Mun. Reparations Comm’n, Report of the Providence Municipal Reparations Commission § 8 (2022). ↑

39. . Students for Fair Admissions, Inc., 600 U.S. at 216 (“For starters, the categories are themselves imprecise in many ways. Some of them are plainly overbroad.”) ↑

40. . Id. at 313 (citation omitted). They “must be limited in time.” Id. ↑

41. . Id. at 214 (“Because ‘[r]acial discrimination [is] invidious in all contexts,’ we have required that universities operate their race-based admissions programs in a manner that is ‘sufficiently measurable to permit judicial [review]’ under the rubric of strict scrutiny. . . Respondents have fallen short of satisfying that burden.”). ↑

42. . The majority’s argument, essentially. Students for Fair Admission, 600 U.S. at 260 (Thomas, J., concurring) (“But, neither have even attempted to explain how their current racially discriminatory programs are even remotely traceable to their past discriminatory conduct.”). ↑

43. . The nature of the use of race as HRN is identified post-hoc. While the HRN framework was not explicitly adopted by policy makers at the time the policies were written, constitutional frameworks (i.e. rational basis review or even strict scrutiny were articulated by courts in response to practice as well). HRN is an emerging practice that should be formalized into doctrine. ↑

44. . See, e.g., Shira Stein, Is Sorry Enough? What Black Americans are Owed in Reparations, Wash. Post (Mar. 25, 2024) https://www.washingtonpost.com/nation/2024/03/25/reparations-san-francisco-apology/. ↑

45. . Blair Paddock, Evanston City Council Approves $25K Direct Cash Payments in Expansion of Reparations Program, WTTW News (Apr. 1, 2023) [https://perma.cc/XC3B-AQP7]. ↑

46. . See Class Action Complaint at ¶ 27, Flinn v. City of Evanston, No. 1:24-cv-04269 (N.D. Ill. May 23, 2024). ↑

47. . Memorandum from Kimberly Richardson, Interim Assistant City Manager for the City of Evanston, Ill., to the Mayor and Members of City Council (Mar. 22, 2021), [https://perma.cc/FY5F-E5K9]. ↑

48. . See Evanston Local Reparations, City of Evanston, (last visited April 23, 2026) [https://perma.cc/8TU4-6XVG] (“Reparations, and any process for restorative relief, must connect between the harm imposed and the City. The strongest case for reparations by the City of Evanston is in the area of housing, where there is sufficient evidence showing the City’s part in housing discrimination as a result of early City zoning ordinances in place between 1919 and 1969, when the City banned housing discrimination.”). ↑

49. . Id. ↑

50. . Memorandum from Kimberly Richardson, supra note 47. ↑

51. . Id. This attenuated assumption of harm (i.e. that the harm was upon an entire group without having to show specific harm) sets the Evanston initiative apart from, e.g., the following analogy to the CLA. Aggregate presumptions of harm do have precedent in the form of class action settlements or other forms of compensation such as that related to asbestos exposure. These rely on an assumption that harm was functionally ubiquitous, and proof of individual harm would make justice functionally impossible. See generally, Georgene Vairo, Lessons Learned by the Reporter: Is Disaggregation the Answer to the Asbestos Mess?, 88 Tul. L. Rev. 1039 (2014). Similarly, the eligibility of ancestors (intergenerational harm) for redress sets this program apart from the CLA. There are several ways that intergenerational harm can be measured, such as through longitudinal data analysis or other models. For examples of methods, see Andrea E. Willson & Kim M. Shuey, A Longitudinal Analysis of the Intergenerational Transmission of Health Inequality, 74 J. Gerontol. B Psychol. Sci. & Soc. Sci. 181 (2018) and M.K.M. Lünnemann et al., The Intergenerational Impact of Trauma and Family Violence on Parents and Their Children, 96 Child Abuse & Neglect 104134 (2019). ↑

52. . Memorandum from Kimberly Richardson, supra note 47. ↑

53. . There are many aspects of the programs like the one in Evanston, IL, that, in terms of equal protection, set it apart from the affirmative action programs like those discussed in SFFA, which may indicate the possibility of them surviving strict scrutiny at the Supreme Court. For example, reparations which are tied to specific government actions and given by the government that conducted that action are considered to be different from university boards trying to ameliorate all social inequality. For a full analysis of this claim, see Lisa L. Owens, An Argument for Housing Reparations, 77 Me. L. Rev. 243 (2025). See also, Memorandum from Kimberly Richardson, supra note 47, at 2–⁠3 (acknowledging historical harm and outlining participant eligibility as being related to historically harmed groups). ↑

54. . The program is nonetheless vulnerable to legal challenges under equal protection doctrine, particularly if it is perceived as using present-day race as a proxy for harm rather than relying solely on historically grounded evidence. ↑

55. . See Memorandum from Kimberly Richardson, supra note 47. ↑

56. . See Class Action Complaint at ¶ 27, Flinn v. City of Evanston, No. 1:24-cv-04269 (N.D. Ill. May 23, 2024). ↑

57. . Id. at ¶ 15. ↑

58. . See id. at ¶ 36. ↑

59. . See id. at ¶ 15. ↑

60. . The Federal government identified a similar compelling interest in the CLA of 1988, for example. See Civil Liberties Act of 1988, 50 U.S.C. § 4215 (1988). ↑

61. . If a contemporary Black resident of Evanston could only demonstrate that they had a non-Black ancestor in Evanston at the time period identified, they would ostensibly only be able to qualify through the third race neutral category with the higher evidentiary burden as well. ↑

62. . Id. ↑

63. . Id. ↑

64. . See id. (defining who is an “eligible individual” under the Act). ↑

65. . The Long Road to Redress, Hist., Art & Archives: U.S. House of Representatives, [https://perma.cc/796V-ABFS]. ↑

66. . 50 U.S.C. § 4215(a). Notably, not to families or ancestors of those who had already died. Id. ↑

67. . Id. ↑

68. . The Long Road to Redress, Hist., Art & Archives: U.S. House of Representatives, [https://perma.cc/796V-ABFS]. ↑

69. . S. 1009, 100th Cong. §I(4) (1st Sess. 1987). ↑

70. . S. 1009, 100th Cong. §I(5) (1st Sess. 1987). ↑

71. . 52 Fed. Cl. 432, 433–35 (2002). Jurisdiction for administering policy claims was granted here under the Civil Liberties Act of 1988.  Id. at 434; 50 U.S.C. § 4215(h)(1). ↑

72. . Id. at 435. ↑

73. . Id. ↑

74. . Obadele v. United States, 52 Fed. Cl. 432, 443 (2002) (“[T]he Plaintiffs are really only saying that the victims and heirs of one historical wrong have succeeded in persuading the public and Congress of their entitlement to redress, whereas Dr. Obadele and like-minded advocates are in the process of trying to achieve that same result for another historical wrong.”). ↑

75. . See e.g., Making Equal Protection Protect, 138 Harv. L. Rev. 2113 (2025); Justin Driver, The Strange Career of Antisubordination, 91 U. Chi. L. Rev. 651, 652–53 (2024). ↑

76. . See generally, Sudeall Lucas, supra note 11, at 1605. ↑

77. . Id. ↑

78. . See id. at 1627 n.93. ↑

79. . Adarand Constructors, Inc. v. Pena, 515 U.S. 200, 227 (1995) (“All racial classifications, imposed by whatever federal, state, or local governmental actor, must be analyzed by a reviewing court under strict scrutiny.”). ↑

80. . Regents of the Univ. of Cal. v. Bakke, 438 U.S. 265, 291 (1978) (“Racial and ethnic distinctions of any sort are inherently suspect and thus call for the most exacting judicial examination.”). ↑

81. . City of Richmond v. J.A. Croson Co., 488 U.S. 469, 493 (1989) (The means of accomplishing the compelling interest must be chosen “so closely that there is little or no possibility that the motive for the classification was illegitimate racial prejudice or stereotype.”) ↑

82. . Such as, for example, in Providence where “[d]uring the 20th century, race-based discrimination continued as municipal, state, and federal government-sanctioned redlining, discriminatory employment and housing practices, urban renewal, and interstate highway improvements deconstructed largely African heritage and limited-income neighborhoods including Fox Point, Upper South Providence, West Elmwood, College Hill and Lippitt Hill.” Report of the Providence Mun. Reparations Comm’n (Aug. 2022). ↑

83. . Obadele v. United States, 52 Fed. Cl. 432, 433 (2002) (“The statute established within the Justice Department Civil Rights Division the Office of Redress Administration (ORA) to identify individuals eligible for relief in connection with those wrongs associated with the World War II emergency action.”). ↑

84. . See Students for Fair Admissions, Inc. v. Presidents & Fellows of Harvard Coll., 600 U.S. at 214, 226 (2023). ↑

85. . See e.g., Sudeall Lucas, supra note 75, at 1646. ↑

86. . Jennifer Ludden, Cities May Be Debating Reparations, but Here’s Why Most Americans Oppose the Idea, NPR (Mar. 27, 2023) [https://perma.cc/W9GJ-X79L] (“‘You can’t take what we know now and try to superimpose yourself onto 150 years ago,’ says Jeff Bernauer, visiting from Huntsville, Alabama. He calls racism a sin and says of course slavery was wrong. But to try and make amends at this point makes no sense. ‘The generation that would be paying for it have nothing to do with what was done in the past,’ he says. ‘And then you’re paying people that have nothing to do with it in the past.’”). ↑

87. . Lisa L. Owens, Revisiting Gender in Public Accommodations, ___ N.Y.U. J. Legis. & Pub. Pol’y (forthcoming) (Further describing the relationship between formalism and social complexity in terms of gender and race). ↑

88. . See id.; see generally Sudeall Lucas, supra note 11. ↑

89. . For example, administrability challenges undergirded the CLA, which was the second federal reparations program to attempt redress for Japanese internment. Also, administrability challenges were a key reason for changes in the way awards were made in the Evanston reparations program.  See American-Japanese Evacuation Claims Act, 50 U.S.C. § 1981 (1948); Comm’n on Wartime Relocation and Internment of Civilians, Personal Justice Denied 461 (1983) (“The Act did not allow claims for lost income or for pain and suffering.  Approximately $37 million was paid in claims, an amount far below what would have been full and fair compensation for actual economic losses.  Awards were low because elaborate proof of loss was required, and incentives for settling claims below their full value were built into the Act.”); see also Gina Castro, City Council Approves Direct Cash Payment Reparations, Evanston RoundTable (Mar. 27, 2023) [https://perma.cc/8EXU-43EC]. ↑

90. . See Derek Bell, Racial Realism, 24 Conn. L. Rev. 363 (1992). From this view, HRN might very well risk reinforcing the very structures that CRT seeks to dismantle by translating claims for justice into terms palatable to the hostile judiciary rather than demanding doctrinal transformation. Critics from the CRT tradition may also view HRN as overly constrained by formal legal categories and insufficiently responsive to the structural, intergenerational nature of racial injustice. As well, while HRN may appeal to courts or legislatures, it may additionally risk alienating grassroots reparations movements by narrowing the permissible kind of claims to cases with the strongest evidentiary foundation, contributing to access to justice issues. ↑

91. . Students for Fair Admission, Inc. v. Presidents & Fellows of Harvard Coll., 600 U.S. 181 (2023); Obadele v. United States, 52 Fed. Cl. 432 (2002). ↑

92. . Students for Fair Admission, 600 U.S. at 181. ↑

93. . See The Reparations Ordinance, CHI. TORTURE JUST. MEM’LS, [https://perma.cc/6AFW-LYAG]. ↑

Introduction

Local governments rely on property taxes to fund many essential public services. Owners who do not pay the taxes impose an increased burden on their taxpaying neighbors. To what extent can local governments interfere with the property rights of delinquent owners in order to safeguard the interests of ordinary taxpayers? In particular, can the government use the foreclosure process in ways that threaten the equity of delinquent property owners? The Supreme Court confronted that issue in Tyler v. Hennepin County,^[2] and will confront it again this term in Pung v. Isabella County.^[3]

In Tyler, the Court held that when a municipality forecloses on property for nonpayment of taxes, the Federal Constitution’s Takings Clause precludes the municipality from retaining any surplus between what the successful bidder pays for the property and the amount the delinquent owner owes; those surplus proceeds

are the owner’s property, and must go to the owner.^[4] In Pung, the federal district court, even before the Supreme Court’s decision in Tyler,^[5] ordered the county to return the surplus to the owner, but the delinquent owner seeks more. The owner contends that as a result of the foreclosure sale, the county took the owner’s property and is required to compensate the owner based on the property’s market value which, according to the owner, was in excess of the successful bid at the foreclosure sale.^[6]

In both cases, the foreclosed property owner has received significant support both from the political left and the political right. Organizations filing briefs from the left have expressed concerns that the targets of foreclosure tend to be disproportionally poor, old, or disabled.^[7] On the right, the concern has been the sanctity of property rights against government interference.^[8]

In focusing on the rights of delinquent owners, however, these strange bedfellows ignore the rights of taxpayers who must make up the difference when delinquent owners fail to pay their share of the local tax burden. In Tyler, the Court attempted to balance those interests with Chief Justice Roberts’ pithy remark that “[t]he taxpayer must render unto Caesar what is Caesar’s, but no more.”^[9] But what is Caesar’s? Collecting taxes from a recalcitrant taxpayer is not cost-free. To what extent can a municipality use processes that reduce the cost of collection, even if those processes have an adverse impact on delinquent taxpayers? That is the issue the Court will have to address in Pung.

In both Tyler and Pung, the municipality conducted a foreclosure sale to recover delinquent taxes. If Pung holds that, upon foreclosure, tax delinquent owners are entitled to compensation based on fair market value even when that value exceeds the price realized at a foreclosure sale, municipalities may be forced to abandon use of foreclosure sales, to the detriment of taxpaying residents. Of course, municipalities might use different processes to ensure timely collection of property taxes and to reduce the cost of collection, as municipalities in many states do.^[10] But the Court has yet to consider how the constitutional analysis it developed in Tyler would apply to those processes.

This essay explores the challenges facing the Court as it attempts to reconcile the interests of taxpayers and property owners, and the challenges facing municipalities as they attempt to comply with the Court’s constitutional commands. Part I focuses on the Court’s opinion in Tyler. Part II lays out the alternatives facing the Court in Pung. Part III expands the discussion to explore additional issues raised by constitutionalizing the tax foreclosure process.

I. Tyler v. Hennepin County

In 2010, Geraldine Tyler, past age 80, moved from her one-bedroom condominium into a senior community.^[11] Beginning in 2011, she stopped paying real estate taxes on the condominium.^[12]

Under Minnesota law, once real estate taxes become delinquent, penalties and interest begin to accrue.^[13] Then, in the year following a missed payment, the county auditor brings a lawsuit against the delinquent property.^[14] The county provides notice to the owner,^[15] and if the owner does not file an answer, the court enters a judgment, establishing a lien on the property.^[16] In May of that year, the county auditor transfers limited title to the state, subject to the owner’s right to redeem by paying back taxes, interest, and penalties.^[17] The owner has three years to redeem.^[18] The county auditor provides notice to the owner 120 days before expiration of the time for redemption.^[19]

Apparently, Geraldine Tyler did not respond to the auditor’s lawsuit, and did not exercise the right to redeem.^[20] As a result, in 2015, under Minnesota law, absolute title vested in the state.^[21] By that time, about $2,300 in unpaid taxes had accrued, and the total due, including interest and penalties, totaled $15,000.^[22] A year later, the county sold the Tyler condominium for $40,000. Pursuant to Minnesota statutes, the government kept all $40,000.^[23]

Tyler then brought suit, alleging that the county’s retention of the surplus (the difference between the $40,000 sale price and the $15,000 due in taxes, interest, and penalties) constituted an unconstitutional taking of her property.^[24] She also alleged that retention of the surplus constituted an excessive fine in violation of the Eighth Amendment’s prohibition on excessive fines.^[25] The District Court dismissed for failure to state a claim, and the Eighth Circuit affirmed, reasoning that because Minnesota’s statutory scheme recognized no property interest in surplus proceeds, no property was taken, unconstitutionally or otherwise.^[26] The District Court also concluded that retention of the surplus did not violate the Excessive Fines Clause because retention was designed to remedy the state’s tax losses, not to punish the delinquent taxpayer.^[27]

A unanimous Supreme Court reversed in an opinion by Chief Justice Roberts. The Court held that Minnesota could not extinguish an owner’s property interest by enacting a law providing for forfeiture of a property interest by falling behind on tax obligations. While conceding that the Takings Clause itself does not define “property,” the Court reaffirmed the principle that, for taking purposes, state law cannot be the only source of property rights.^[28] Looking to history and precedent, the Court held that use of tax debt to confiscate more property than was due effected a “classic taking in which the government directly appropriates private property for its own use.”^[29] Because relief under the Takings Clause would remedy the harm to Geraldine Tyler, the Court did not have to decide whether her excessive fines claim should have been dismissed. In a concurring opinion, Justice Gorsuch, joined by Justice Jackson, concluded that the Eighth Circuit erred in dismissing that claim.^[30]

In holding that the Minnesota statute effected a taking of Geraldine Tyler’s property, the Court distinguished Nelson v. City of New York,^[31] an earlier case in which the Court had rejected constitutional claims of landowners who received no surplus after the city had foreclosed on their property for unpaid water bills. The Tyler Court emphasized that the New York procedures at issue in Nelson would have entitled the owners to recover the surplus if they had “filed a timely answer in [the] foreclosure proceeding, asserting [the] property had a value substantially exceeding the tax due.” ^[32] By contrast, the Minnesota statute provided no opportunity for the owner to receive surplus value once the redemption period expired.

In its takings analysis, the Court’s focus was on the taking Tyler identified: the county’s retention of surplus moneys resulting from sale of the apartment for more than the amount she owed. Tyler did not challenge, and the Court did not address, the interest and penalties assessed as a result of her delinquency, although the interest and penalties dwarfed the amount of the tax due. ^[33]And Tyler did not allege that, because the county’s auction sale brought less than the fair market value of her apartment, her compensation should be measured by the difference between fair market value and the tax obligation she has failed to satisfy. Finally, because her claim was focused on the Minnesota scheme, neither her complaint nor the Court’s opinion addressed other tax collection mechanisms used by states and municipalities, some of which could result in loss of equity by defaulting owners.

II. Pung v. Isabella County

In Pung, the Court will have the opportunity to opine on one of those issues: the measure of compensation when the market value of a delinquent owner’s property exceeds the price realized at a foreclosure sale. The argument for basing compensation on market value is straightforward: when the government exercises dominion over the defaulting owner’s property by holding a foreclosure sale, the government has taken property; the standard measure of just compensation for a taking is fair market value. Awarding that measure, however, would seriously complicate the ability of local governments to collect delinquent taxes.

A. Background

Pung is an appealing case for the property owner because there were no delinquent taxes and there should have been no foreclosure. Michigan affords a homeowner an exemption from some local property taxes if the owner submits an affidavit that the home is a primary residence. Timothy Pung applied for and received this Principal Residence Exemption.^[34] After Timothy’s unexpected death in 2005, his wife lived in the home until she died in 2008.^[35] After her death, the couple’s son continued to live in the home.^[36] Timothy’s uncle, Michael Pung, served as executor of Timonthy’s estate and did not resubmit an affidavit that the home was a primary residence. As a result, in 2010, the township assessor retroactively revoked the exemption for the years 2007–09^[37].

When Michael challenged the revocation, the Michigan tax court held that resubmission of the affidavit was unnecessary and that the Pungs were entitled to the exemption for the tax years 2007–2011^[38]. Although the initial 2012 tax bill included the exemption, the assessor ignored the tax court’s decision and revoked the exemption for the 2012 tax year.^[39] When Michael paid the amount due on the initial 2012 bill, the assessor reported the property as delinquent and the county brought foreclosure proceedings while simultaneously appealing the tax court’s holding that the Pungs were entitled to the exemption.^[40] The county sent several notices of the foreclosure proceedings to Timothy Pung at Michael’s address, but Michael inexplicably did not respond.^[41] The county made no mention of the foreclosure proceedings to Michael or his lawyer when they came into contact at the January 2015 oral argument in the Michigan Court of Appeals on the county’s appeal of the tax court’s decision.^[42]

On February 10, 2015, the Michigan Court of Appeals affirmed the tax court’s determination that the Pungs were entitled to the exemption for the years 2007–2011.^[43] Ten days later, the county obtained a judgment of foreclosure based on the alleged deficiency in 2012 taxes, but did not send notice of the judgment to Michael Pung until April 30, after the estate’s redemption rights had expired.^[44] The county took title to the property. Michael Pung then brought a proceeding challenging the foreclosure, and the trial court invalidated the foreclosure, holding that Pung had not received constitutionally adequate notice of the foreclosure proceedings.^[45] The trial court reasoned that the county had constructive notice that Pung had not received adequate notice because Pung would not have been arguing against the county’s appeal of the 2010–2011 taxes if the appeal would have been moot a month later when the foreclosure was concluded.^[46] That constructive notice, the trial court held, obligated the county to make additional efforts to ensure that Pung had notice of the foreclosure proceedings.^[47] However, the Michigan Court of Appeals reversed, holding that the notice was constitutionally adequate.^[48] Pung did not seek review of that determination. On remand, in June 2018, the county obtained a final judgment of foreclosure, taking absolute title to the Pung home.^[49]

In short, the Michigan courts upheld a foreclosure for a deficiency in 2012 taxes because Pung did not respond to the county’s foreclosure notices—even though the courts had established that the Pungs were entitled to an exemption for the years 2007–2011 on grounds that would have been equally applicable to the 2012 tax year.

Subsequently, the county conducted an auction sale which brought a sale price of $76,008.^[50] The assessed fair market value of the property was $194,400, and the auction sale purchaser resold the property for $195,000.^[51] Michael Pung then brought an action in federal district court, alleging an unconstitutional taking and a violation of the Excessive Fines Clause. The District Court awarded summary judgment to Pung on the claim for the surplus between the auction sale price and the taxes allegedly due, but denied his claim for loss in the property’s equity.^[52] The Sixth Circuit affirmed.^[53]

B. The Court’s Options

The county’s treatment of the Pungs was inexcusable. The question for the Court is what constitutional remedy the Pungs are entitled to receive—the surplus proceeds awarded by the District Court and the Sixth Circuit, or the difference between fair market value and the amount of the tax delinquency. This Part examines the case for each of the options available to the Court.

1. Affirm the Sixth Circuit

The Court’s first option in Pung is to affirm the Eighth Circuit’s decision, holding that Tyler entitled defaulting taxpayers to surplus proceeds derived from the sale of tax-delinquent property, and no more. Although the Court in Tyler did not explicitly address whether an owner might obtain recovery based on fair market value when that value exceeds the proceeds of a foreclosure sale, its analysis indicates that recovery is limited to surplus proceeds.^[54]

First, the Court relied on history to establish the contours of the owner’s property right. The Chief Justice started by citing English statutes and Blackstone for the proposition that when a tax collector seized a taxpayer’s property, the collector was obligated to restore the “overplus” to the taxpayer when the property is sold.^[55] He then turned to a Civil War-era statute imposing a nationwide tax but providing that if the tax were not paid the property would be sold and the “surplus” would be paid to the owner, and to a later case entitling an owner to surplus even though a subsequent statute did not explicitly mention the surplus.^[56]

Second, the Court focused on contemporary Minnesota law in other contexts. Minnesota law permitted seizure and sale of property to satisfy income tax debts, but provided for refund to the owner of any surplus proceeds from the sale.^[57] The Chief Justice also cited Minnesota law with respect to mortgages, which, like the law of every other state, entitles the property owner to any surplus proceeds of a foreclosure sale.^[58] The Court did not even suggest that in these instances, the taxing authority or the foreclosing mortgagee would be required to restore the fair market value of the property to the defaulting taxpayer or borrower.

If affirmance seems unfair to the Pungs, it is probably because the foreclosure in the Pung case was not the garden variety tax foreclosure at issue in Tyler. Unlike most defaulting property owners, the Pungs were consistently willing and able to pay all taxes due; the dispute was about whether taxes were in fact due. The problem was a process failure: if the Pungs had known about the impending foreclosure, they certainly would have successfully challenged the foreclosure, as they had successfully challenged rejection of the tax exemption for prior years. If the county failed to provide adequate notice of the foreclosure sale, the Supreme Court’s opinion in Jones v. Flowers^[59] establishes that the Pungs would have a due process claim for invalidating the foreclosure sale.

For two reasons, however, the Court cannot invoke due process to provide the Pungs a more complete remedy. First, it appears that the county provided Michael Pung, who held legal title to the property as executor, with notice reasonably calculated to apprise him of the impending foreclosure. The county sent notices to his home address and those notices were not returned undelivered. The county also posted a notice on the door of the tax delinquent property. Second, even if the notice to Michael Pung was constitutionally inadequate, Pung raised those issues in the earlier Michigan state court proceeding, persuading the trial court, but not the Michigan Court of Appeals. Because Pung did not appeal the Michigan Court of Appeals decision, res judicata principles would bar the Supreme Court from considering those issues anew.

        2. Reverse and Hold the County Liable for a Taking of Fair Market Value

The case for basing an award to the Pungs on fair market value of their property is straightforward. The first step is identifying the taking, which occurred when the county took title to the property in the foreclosure proceeding. The Sixth Circuit said as much in Hall v. Meisner,^[60] a case cited by the Court in Tyler. Other courts have indicated, in somewhat different contexts, that once a state officer authorizes a permanent physical occupation, a taking occurs.^[61] The second step is measuring compensation, and the Court has said, in Brown v. Legal Foundation of Washington, that when the government takes property, just compensation is measured by the owner’s loss, not the government’s gain.^[62]

This approach avoids questions that arise if the government elects to retain the property rather than selling it to a third party,^[63] or if the property’s value increases or decreases before the government conducts the sale. And if the taking occurs at the moment the municipality forecloses, it avoids questions about the defaulting owner’s rights between that moment and the time of the foreclosure sale. ^[64]

From the perspectives of fairness and efficiency, however, awarding compensation based on fair market value has a huge downside: it would cripple government efforts to collect delinquent property taxes—at the expense of taxpaying residents. If a municipality faced a lawsuit seeking fair market value every time it foreclosed on delinquent property, foreclosures would become impractical. Rather than providing the municipality with a reliable mechanism for collecting unpaid taxes, foreclosure would leave the municipality with negative revenue whenever a foreclosure sale brought less than market value (or when the cost of litigation over fair market value exceeded the taxes collected through the foreclosure process). One or both of these circumstances are the norm. In that environment, no sensible municipality would use tax foreclosures. Of course, government could turn to other mechanisms for collecting delinquent taxes, as many governments do, but those mechanisms, too, could engender taking claims (as Part IIIB illustrates).

Moreover, the scope of a holding that requires government to reimburse a defaulting property owner for the market value of property upon foreclosure could threaten basic understandings of property law. Such a holding would seem to apply equally to federal efforts to execute judgments for unpaid income taxes against the property of delinquent taxpayers and to mortgage foreclosures in those states that require judicial supervision of those foreclosures. If a municipality or state were required to reimburse a property owner for the difference between fair market value and taxes owed even though the municipality did not benefit from that market value, why wouldn’t a bank (or the state that authorized the foreclosure) be required to reimburse a mortgagor for the difference between fair market value and amounts owed on the mortgage, even if the bank’s foreclosure sale brought less than market value?

The Tyler opinion’s citation to Blackstone suggests a way out of the dilemma: owners hold property subject to an implied contract that allows the government to treat the property as security for tax obligations.^[65] Blackstone focused on the flip side of the implied contract—the government’s obligation “to render back the overplus” upon sale—but that obligation arises only if there is an implied understanding that the government may sell the property in the first place. On the implied contract theory, the government’s obligations with respect to a tax sale parallel the obligations of a lender foreclosing on a security interest: ensure that any sale incorporates procedural safeguards for the defaulting owner and restore any surplus to that owner.^[66]

        3. Reverse and Hold that the County Imposed an Excessive Fine on the Pungs

Another option for providing relief to the Pungs would be to hold that the forfeiture of their property constituted an excessive fine in violation of the Eighth Amendment. Justices Gorsuch and Jackson endorsed that approach in Tyler.^[67] The excessive fines approach would avoid the need to determine when a taking occurred and precisely what property interest the county took. Unfortunately for the Pungs, the case for finding an excessive fines violation when government denies the owner surplus from a foreclosure sale is far stronger than when the owner demands fair market value of the property.

Unlike other Eighth Amendment protections, the excessive fines provision applies in civil actions as well as criminal proceedings, and it applies in particular to civil forfeitures.^[68] But the Court has indicated that the provision applies only to forfeitures that are intended, at least in part, to punish.^[69] Applying that standard, the court has invalidated a statute providing for forfeiture of real property used to commit drug crimes,^[70] and a statute mandating forfeiture of funds transported abroad in violation of a reporting violation.^[71] But in United States v. Bajakian,^[72] the Court distinguished traditional in rem forfeitures involving forfeiture of goods imported without payment of duties, concluding that that those forfeitures were not considered punishment, but instead served “the remedial purpose of reimbursing the government for the losses accruing from the evasion of customs duties,”^[73] even though the property forfeited far exceeded the customs duties the government was owed. In light of cases permitting forfeiture for nonpayment of duties (a form of tax),^[74] forfeiture of land to reimburse government for nonpayment of property tax might also be conceptualized as remedial rather than punitive.

Even if forfeiture of land were treated as punishment for nonpayment of taxes, it would be much more difficult to treat as punishment a foreclosure process in which the foreclosed owner is paid the surplus resulting from the government’s sale of the tax-delinquent property. Although the prospect of foreclosure may have a deterrent effect, when government commits to paying the surplus to the former owner, it is quite clear that the animating purpose behind the foreclosure is collection of tax due.

Finally, even if a tax foreclosure were treated as a fine for purposes of the excessive fines clause, it would be difficult to characterize the fine as excessive when the government retains only so much money as necessary to satisfy the tax obligations associated with the property, returning the balance to the owner. For the Pungs, the most plausible path to reversal on an excessive fine theory would be to focus on what the county did before a court held that they were entitled to the surplus from the foreclosure sale. That is, until court decisions established that the county was obligated to return any surplus to property owners, the county was prepared to retain all of the proceeds from the foreclosure sale. Perhaps the Court could treat that as an excessive fine, and then conclude that the remedy for the excessive fine was restoration of the Pung property or its value. That approach, however, would be inconsistent with the Court’s opinion in Bajakian, where upon concluding that forfeiture of all funds transported in violation of a reporting violation constituted an excessive fine, the Court then upheld decisions of the courts below ordering forfeiture of a lesser amount.^[75]

      4. Summary

As Chief Justice Roberts observed in Tyler, paying property taxes is a reasonable condition on property ownership.^[76] Those who fail to satisfy that condition impose greater burdens on those who do. Tax foreclosures have provided a time-honored mechanism for protecting taxpaying members of the public. Nothing in Tyler suggests that states and municipalities must abandon tax foreclosures; indeed the most sensible reading of the Court’s opinion suggests an implicit bargain between municipalities and taxpayers: government may treat property as security for tax obligations, but government must hold procedurally fair foreclosure sales and turn over any surplus to the delinquent taxpayer. A reversal in Pung could threaten that bargain at the expense of taxpaying members of the public.

      5. Beyond Pung: What Constitutional Limits on Property Tax Collection?

Pung presents an attractive case for property rights advocates on the right and foreclosure opponents on the left. The behavior of local officials showcased government at its worst; the town and county treated the Pungs outrageously. Pung, however, is not the typical case of a delinquent owner whose failure to pay taxes imposes a greater burden on other taxpayers. And the tax sale process at issue in Tyler and Pung is only one of the mechanisms states and municipalities have developed for dealing with those delinquencies. Many government entities collect taxes on delinquent properties by selling tax liens to private parties, who acquire the right to collect from the defaulting taxpayer, a process that raises constitutional questions not addressed in Tyler.^[77] Other states have recently revised, or are in the process of revising, their statutes in light of Tyler.^[78] As a consequence, Pung is unlikely to be the last word in the conflict between taxpayers and delinquent owners. This Part sketches some of the unresolved issues.

A. Procedural Issues

Geraldine Tyler sought surplus proceeds five years after the county sold the property to a third-party buyer. Because Minnesota law held that she was not entitled to the proceeds in any event, timing and process were irrelevant to the Court’s taking analysis. But suppose Minnesota gave a delinquent owner 30 days, or six months, to make a claim to surplus proceeds and provided that the county could keep the proceeds if no claim were made. Would the statute have worked a taking? Put another way, must the foreclosing government entity arrange for automatic payment to the delinquent taxpayer, or may government impose an affirmative burden on the taxpayer to make a claim within a reasonable period?

The Tyler Court’s citation to Nelson v. City of New York^[79] suggests that so long as the foreclosing entity provides a mechanism for recovering the surplus, a defaulting taxpayer who fails to use the mechanism would lose any taking claim. In Nelson, the city ordinance conditioned recovery of the surplus on answering the foreclosure proceeding with an assertion that the property’s value exceeded the tax due.^[80] The Tyler court concluded that because the New York procedure “simply defined the process through which the owner could claim the surplus, we found no Takings Clause violation.”^[81]

In many cases, the foreclosing entity will not know precisely who has what interests in the defaulted property. For instance, in Tyler itself, there were allegations that mortgages were outstanding on Tyler’s condominium unit.^[82] A mortgagee’s interest would generally enjoy priority over the fee owner’s interest. But without costly investigation, the foreclosing entity would have no way to know how much was outstanding on a mortgage. Reflexively paying out surplus proceeds to the fee owner might constitute a taking of the mortgagee’s interest.

Even if the Court were to hold in Pung that a foreclosing municipality’s liability to the defaulting owner must be based on fair market value rather than surplus proceeds, questions would arise about the state’s power to limit the time during which a defaulting owner must assert the claim for compensation. Every state already gives the defaulting owner a redemption period during which the owner can free the property from government claims by paying the taxes due. When the owner does not redeem during that period, how long must the state give the owner to claim just compensation before the municipality can acquire title free of a taking claim? Inverse condemnation claims, even when styled as federal claims under 42 U.S.C. §1983, are subject to statutes of limitations.^[83] Even if the state could not specify a short statute of limitations targeting claims by defaulting taxpayers, could it accomplish the same result by shortening the redemption period so that the taking that that triggers compensation comes earlier in the collection process?

For example, the redemption period at issue in Tyler was three years; much longer than the period in Pung. If Minnesota shortened its redemption period from three years to three months and the Court concluded that the taking occurred at the expiration of the redemption period, triggering a claim to fair market value, would a defaulting taxpayer lose any just compensation claim at the expiration of the state’s statute of limitations for personal injuries?

Precisely what processes will survive scrutiny after Tyler (and potentially Pung) will remain an open question,^[84] especially as states amend their statutes in light of Supreme Court precedent.

B. Tax Lien Sales

The tax foreclosure processes at issue in Tyler and Pung resemble the familiar mortgage foreclosure process. When the government entity forecloses, it sells the property at auction and transfers a deed to the auction sale purchaser. Many states, however, use an entirely different process to collect delinquent property taxes: they sell tax liens. Sale of tax liens raises constitutional questions the Court did not consider in Tyler and that are not at issue in Pung, but that have started to arise in lower courts.

Although the process for selling tax liens differs materially from state to state, the basic principle is that the taxing entity, which holds a lien against delinquent property for taxes due, sells that lien, usually at a public auction.^[85] The purchaser acquires the right to collect the taxes due, together with interest, from the delinquent owner, and to foreclose on the property after a specified period if the delinquent owner does not pay the taxes.^[86] In some states, the lien is awarded to the highest bidder at the auction.^[87] In others, bidding is by interest rate, and the lien goes to the bidder who accepts the lowest interest rate on the delinquent owner.^[88] Sometimes, the municipality avoids the public sale of tax liens and instead engages in a bulk sale to a third party trust that issues bonds backed by tax lien receivables.^[89]

From the municipality’s perspective, selling the tax lien generates immediate revenue; the purchase price fills the revenue gap caused by the tax delinquency.^[90] For the investor who purchases tax liens, the primary incentive is the interest that accrues on the tax lien, which typically exceeds market rates.^[91] If, however, the defaulting owner does not pay the taxes and interest within a statutory redemption period, the investor has the right to foreclose, and in some states statutory foreclosure provisions authorize the tax lien purchaser to obtain a tax deed without a public sale.^[92] In an increasing number of states, especially since the Tyler decision, foreclosure of the tax lien requires a public auction.^[93]

If a delinquent taxpayer loses title through the tax lien process, when (if at all) does a taking occur? When the government sells the tax lien, title to the property does not change hands. The delinquent taxpayer still holds title to the property and has a right to redeem by paying the back taxes. If, at the expiration of the redemption period, the municipality transfers title to the lien purchaser without a public sale, there is no surplus, raising questions about what the delinquent taxpayer is entitled to recover. And if there is a public sale, it will be because the lien purchaser—typically not a government entity—has foreclosed. Finally, who is liable to provide just compensation to the delinquent owner—the municipality, or the lien purchaser?

These issues have begun to percolate through the lower courts. In Edmundson Community Organization v. Mayor and City Council of Baltimore,^[94] a federal district court rejected the argument advanced by the city and tax lien purchasers that tax delinquent owners receive adequate compensation when the receive the surplus between the amount the lien purchaser pays at the tax lien sale and the amount of the tax delinquency. In denying a motion to dismiss the taking claim, the court concluded that the delinquent owners had adequately alleged that by providing only the value of the lien, rather than the value of the property, the city and the purchasers had failed to provide just compensation.^[95] At the motion to dismiss stage, the court expressed no opinion on the ultimate calculation of just compensation.^[96]

In Continental Resources v. Fair,^[97] the Nebraska Supreme Court held that a tax delinquent owner stated a taking claim when the county treasurer issued a tax deed to a tax lien purchaser, but as in the Edmundson case, the court did not indicate how compensation should be computed. The court also held that the tax lien purchaser—not the county—was liable for just compensation to the tax delinquent owner, reasoning that it was the purchaser’s request for a tax deed that extinguished the tax delinquent owner’s title, and that in the circumstances of tax lien foreclosures, a private entity can qualify as a state actor.^[98] By contrast, other courts have held that a municipality may be held liable for a taking when it issues a tax deed even if the municipality receives no excess equity from the property.^[99] They did not decide how compensation should be measured because the cases held only that the municipal defendants’ motions to dismiss should be denied.

Because the tax lien sale process is market driven—unlike the government foreclosure followed by sale of a tax deed process at issue in Tyler and Pung—finding that a state’s tax lien sale process works a taking may disadvantage defaulting owners in other ways. In jurisdictions that entitle a lien holder to a tax deed upon expiration of the defaulting owner’s redemption period, that entitlement is one of the financial incentives that lead investors to purchase tax liens. In jurisdictions where investors bid for tax liens in dollars, municipalities might seek to compensate for that lost incentive by raising the interest rate and penalties on unpaid tax—to the detriment of defaulting owners who want to redeem. Similarly, in jurisdictions where the successful bidder at a tax lien sale is the investor who bids the lowest interest rate, the interest rates bidders are willing to accept may rise, increasing the defaulting owner’s tax burden and reducing the value of the owner’s equity.

Of course, reducing the incentive to bid on tax liens might have other effects. It might leave some tax liens unpurchased, resulting in an increased burden on non-defaulting taxpayers. The actual effect a particular legal rule might have is uncertain, but if markets do respond, increasing the takings protection for defaulting owners is likely to have an adverse effect, either on taxpaying property owners or on the defaulting owners themselves.

C. Interest and Penalties

Neither Tyler nor any of the post-Tyler decisions has challenged the right of a municipality or a private tax-lien purchaser to retain statutorily authorized interest or penalties on taxes due. Currently, Connecticut authorizes municipalities to collect 18% interest on delinquent taxes;^[100] other states authorize less.^[101] But money is fungible. If a state or municipality responds to Tyler (or to the prospect of limitations on tax lien sales) by increasing the interest or penalties on overdue taxes, would that increase be subject to a takings analysis? Surely the Court would not want to impose a constitutional maximum on statutory interest rates. If a local government were not permitted to authorize penalties and interest that substantially exceed market interest rates, the government would be forced, in effect, to provide long-term loans to defaulting owners at the expense of taxpaying neighbors. But at some point, in sorting out the limits on local government’s tax collection mechanisms, the Court may have to confront the claim that interest and penalties constitute takings or excessive fines.

CONCLUSION

It is easy to portray the disputes in Tyler and Pung as David v. Goliath battles between a small property owner and the government. But when tax collection is at issue, Goliath is just an amalgam of other Davids who, unlike the defaulting owners, have paid their share of the costs of local government. The balance the Court struck in Tyler does not unfairly impair the interests of taxpaying property owners because requiring disgorgement of surplus proceeds does not significantly interfere with tax collection processes. By contrast, holding that a delinquent taxpayer is entitled to fair market value whenever a municipality requires a transfer of title to remedy tax defaults could force local governments to abandon a traditional mechanism for property tax collection while calling into question some prominent alternatives, leaving the debt uncollected. One hopes that as the Court approaches the Pung case it focuses at least as much on the taxpaying public as on the plight of property owners who have defaulted on their tax obligations.

1. * Mack Professor of Law, Benjamin Cardozo School of Law and Distinguished Professor, Yeshiva University. Thanks to Michael Herz and Michael Pollack for helpful comments on earlier drafts. ↑

2. . 598 U.S. 631 (2023). ↑

3. . No. 25-95 (U.S. cert. granted Oct. 3, 2025). ↑

4. . 598 U.S. at 642–45. ↑

5. . Pung v. Cnty. of Isabella, 632 F. Supp. 3d 743 (E.D. Mich. 2022), aff’d sub nom. Pung v. Kopke, No. 22-1919, 2025 WL 318222 (6th Cir. Jan. 28, 2025), cert. granted sub nom. Pung, No. 25-95. ↑

6. . Brief for Petitioner, Pung, No. 25-95, 2025 WL 3489419. ↑

7. . See, e.g., Brief for New Disabled South and Emory Law School Disabled Law Students Association as Amici Curiae in Support of Petitioner, Pung, No. 25-95, 2025 WL 3555109. ↑

8. . See, e.g., Brief of Citizen Action Defense Fund, Manhattan Institute, Reason Foundation, and Oregon Property Owners Association as Amici Curiae in Support of Petitioner, Pung, No. 25-95, 2025 WL 3669487. ↑

9. . Tyler v. Hennepin Cnty., Minn., 598 U.S. 631, 647 (2023). ↑

10. . See Section III.B, infra, concerning tax lien sales used in many states. ↑

11. . See Brief for Petitioner at 2, Tyler, 598 U.S. 631 (No. 22-166), 2023 WL 2339362. ↑

12. . Id. ↑

13. . Minn. Stat. Ann. § 279.01(1) (West 2025) (detailing penalties for late payment); id. § 279.03 (provision for interest on delinquent property taxes and penalties). ↑

14. . Id. § 279.05. ↑

15. . Id. § 279.091. ↑

16. . Id. § 279.16. ↑

17. . Id. § 280.01. ↑

18. . Id. § 281.17(a). ↑

19. . Id. § 281.23. ↑

20. . Brief for Respondents, at 8, Tyler v. Hennepin Cnty., Minn., 598 U.S. 631 (2023) (No. 22-166), 2023 WL 2759804. ↑

21. . Id. ↑

22. . Tyler v. Hennepin Cnty., Minn., 598 U.S. 631, 635 (2023). ↑

23. . Id. ↑

24. . Id. at 635-36. ↑

25. . Id. ↑

26. . Tyler v. Hennepin Cnty., 26 F.4th 789, 793 (8th Cir. 2022), rev’d sub nom. Tyler, 598 U.S. 631 (2023). ↑

27. . Tyler v. Hennepin Cnty., 505 F. Supp. 3d 879, 896 (D. Minn. 2020), aff’d, 26 F.4th 789 (8th Cir. 2022), rev’d sub nom. Tyler, 598 U.S. 631 (2023) (affirming dismissal of excessive fines claim based on District Court opinion). ↑

28. . 598 U.S. at 638 (citing Phillips v. Washington Legal Foundation, 524 U.S. 156, 167 (1998)). ↑

29. . Id. at 639 (quoting Tahoe-Sierra Preservation Council, Inc. v. Tahoe Regional Planning Agency, 535 U.S. 302, 324 (2002)). ↑

30. . Id. at 648 (Gorsuch, J., concurring) ↑

31. . 352 U.S. 103 (1956). ↑

32. . 598 U.S. at 644 (citing 352 U.S. at 110). ↑

33. . See infra Section III.C (discussing interest and penalty charges). ↑

34. . Brief for Petitioner, Pung v. Isabella County, No. 25-95 (U.S. cert. granted Oct. 3, 2025), 2025 WL 3489419. ↑

35. . Id. at 4. ↑

36. . Id. at 3–4. ↑

37. . Id. at 4. ↑

38. . Id. ↑

39. . Id. at 5. ↑

40. . Id. at 5–6. ↑

41. . Joint Appendix at JA-86, JA-90–91, Pung v. Isabella countyCnty., No. 25-95 (U.S. cert. granted Oct. 3, 2025), 2025 WL 3516487 (Michigan Court of Appeals opinion). ↑

42. . Id. at 6, Pung v. Isabella County, No. 25-95, 2025 WL 3489419. ↑

43. . Id. ↑

44. . Id. ↑

45. . Joint Appendix, at JA-94–106., Pung, No. 25-95, 2025 WL 3516487. ↑

46. . Id. at JA-104. ↑

47. . Id. at JA-105105–106. ↑

48. . Id. at JA-84–85. ↑

49. . Pung v. Cnty. of Isabella, 632 F. Supp. 3d 743, 744 (E.D. Mich. 2022), aff’d sub nom. Pung v. Kopke, No. 22-1919, 2025 WL 318222 (6th Cir. Jan. 28, 2025), cert. granted sub nom. Pung, No. 25-95. ↑

50. . Id. ↑

51. . Brief for Petitioner at 8, Pung, No. 25-95, 2025 WL 3489419. ↑

52. . Pung, 632 F. Supp. 3d at 751–752. ↑

53. . Pung v. Kopke, No. 22-1919, 2025 WL 318222 (6th Cir. Jan. 28, 2025), cert. granted sub nom. Pung, No. 25-95. The Sixth Circuit also dismissed the excessive fines claim because the Michigan statute was not designed to be punitive. ↑

54. . In addition to the Sixth Circuit in Pung itself, the Sixth Circuit reached the same conclusion in Freed v. Thomas, 81 F.4th 655 (6th Cir. 2023), as did a district court in Reado v. City of New Orleans, No. CV 24-2451, 2025 WL 3250901 (E.D. La. Nov. 21, 2025), and the Michigan Supreme Court in Yono v. Cnty. of Ingham, No. 166791, 2025 WL 1957960 (Mich. July 16, 2025). In Yono, there were no surplus proceeds because the public auction brought no bidders, at which point the county treasurer sold the property to a land bank for $1. Id. ↑

55. . 598 U.S. at 639–40. ↑

56. . Id. at 642–43. ↑

57. . Id. at 645. ↑

58. . Id. ↑

59. . 547 U.S. 220 (2006) (holding that due process requires reasonable steps to provide owner with notice of impending tax sale). ↑

60. . 51 F.4th 185, 196 (6th Cir. 2022) (concluding that the county’s taking of absolute title was the taking, because “[b]efore that event, the plaintiffs held equitable title; after it, they held no title at all.”). ↑

61. . See Kidd v. Pappas, No. C 7061, 2025 WL 3507374, at *18 (N.D. Ill. Dec. 8, 2025) (citing Loretto v. Teleprompter Manhattan CATV Corp., 458 U.S. 419 (1982) and Cedar Point Nursery v. Hassid, 594 U.S. 139 (2021)), (in the context of invalidating Illinois tax lien sale procedures, court concludes that by authorizing permanent physical occupation by tax lien purchaser, county treasurer authorized a taking). ↑

62. . 538 U.S. 216, 235–36 (2003). ↑

63. . See Jackson v. Southfield Neighborhood Revitalization Initiative, No. 166320 (Mich. July 16, 2025) (holding that delinquent taxpayer has a taking claim when government exercises statutory right of first refusal rather than holding public sale). ↑

64. . See Ramsey v. City of Newburgh, No. 23-CV-8599, 2024 WL 4444374 (S.D.N.Y. Oct. 8, 2024) (holding taking claim unripe upon foreclosure because city had not yet sold the delinquent property). ↑

65. . 598 U.S. 631, 640 (2023). ↑

66. . The implied contract theory would presumably also impose on the foreclosing government entity an obligation to sell the property of the defaulting owner; government could not simply retain the property and avoid payment of any surplus. See, e.g. Jackson, slip op. ↑

67. . 598 U.S. 631 at 648 (Gorsuch, J., concurring). ↑

68. . Austin v. United States, 509 U.S. 602 (1993). ↑

69. . Id. at 609–11. ↑

70. . Id. ↑

71. . United States v. Bajakian, 524 U.S. 321 (1998). ↑

72. . Id. ↑

73. . Id. at 342. ↑

74. . See, e.g., Stockwell v. United States, 80 U.S. 531 (1871). ↑

75. . Bajakian, 524 U.S. 321 at 326 (noting that trial court had ordered forfeiture of $15,000, and that the property owner had not appealed from that order). ↑

76. . Tyler v. Hennepin Cnty., 598 U.S. 631, 646 (2023). ↑

77. . See generally Lawrence Ponoroff, The Curious Case of Tyler v. Hennepin County, 48 Va. Tax Rev. 181 (2023); Cameron M. Bartlett & Chrisopher G. Bradley, Property Tax Privateers, 41 Va. Tax Rev. 89 (2021); Michelle Z. Marciony, Making Debt Pay: Examining the Use of Property Tax Delinquency as a Revenue Source, 62 Emory L.J. 217 (2012); Frank S. Alexander, Tax Liens, Tax Sales, and Due Process, 75 Ind. L. J. 747, 761–62 (2000); Georgette C. Poindexter, Lizabethann Rogovy & Susan Wachter, Selling Municipal Property Tax Receivables: Economics, Privatization, and Public Policy in an Era of Urban Distress, 30 Conn. L. Rev. 157 (1997). ↑

78. . See, e.g. S.B. 1431, 56th Leg., 2d Sess. (Ariz. 2024); Or. Rev. Stat § 77.3 (mandating study to ensure compliance with Tyler); H.B. 2062, 102nd Gen. Assemb., 2d Sess. (Mo. 2024). ↑

79. . Tyler, 598 U.S. 631 at 643-44, citing Nelson v. City of New York, 352 U.S. 103 (1956). ↑

80. . Nelson, 352 U.S. 103 at 110. ↑

81. . Tyler, 598 U.S. 631 at 644. ↑

82. . Id. at 636 (public records indicated that the condominium might be subject to a mortgage and a lien for unpaid homeowners’ association fees) ↑

83. . Section 1983 includes no statute of limitations, but the Court has held that section 1983 claims should be treated as personal injury claims for statute of limitations purposes, subject to the general statute of limitations applicable to those claims. Wilson v. Garcia, 471 U.S. 261 (1985). ↑

84. . See Wright v. Rollyson, No. 2:24-cv-00474, 2025 WL 835040 (S.D. W. Va. March 17, 2025) (upholding statute that gives delinquent owner two years to seek surplus value before surplus is forfeited to the state). ↑

85. . See generally, Alexander, supra note 74, at 770–75 (noting variation of procedures from state to state, and outlining major approaches). ↑

86. . Id. at 775–76 (describing general processes). ↑

87. . See, e.g. Colo. Rev. Stat. § 39-11-115 (2024). ↑

88. . See, e.g. Ariz. Rev. Stat. § 42-18114 (2024). ↑

89. . See, e.g. N.Y. Pub Auth Law § 2345-b (authorizing municipality to enter into contract for sale of tax liens to municipal bond bank or its tax lien entity). See generally Poindexter et al, supra note 74, at185–86. ↑

90. . See Marchiony, supra note 74, at 222 (noting that tax delinquencies create revenue uncertainty). Marchiony notes that traditional foreclosure procedures generate unpredictable cash flow for municipalities, who might have to use debt financing to even out cash flow. Id. at 227–28. ↑

91. . See, e.g. Ariz. Rev. Stat § 42-18053(A) (providing for 16% interest on all delinquent taxes). ↑

92. . See, e.g. Ind. Code § 6-1.1-25-4.6. ↑

93. . See, e.g. Fla. Stat. § 197.502. In some states, a public sale is required only if the defaulting taxpayer requests it and a court deems the request reasonable. See, e.g. Ariz. Rev. Stat § 42-18204 (2024) (amended in 2024 in light of the Tyler decision). ↑

94. . 797 F. Supp. 3d 497 (D. Md. 2025). ↑

95. . Id. at 524, 528. ↑

96. . Id. at 525. ↑

97. . 10 N.W.3d 510 (Neb. 2024). ↑

98. . Id. at 407. See also 257-261 20th Ave., Realty, LLC v. Roberto, 327 A.3d 1177 (2025) (holding that lienholders can be treated as state actors and are not entitled to surplus equity that exceeds the value of certificates they purchased plus interest and related costs). ↑

99. . See Grady v. Wood Cnty. W. Va., 786 F. Supp. 3d 984, 994 (S.D.W. Va. 2025) appeal dismissed, No. 25-1606, 2025 WL 3488765 (4th Cir. Aug. 28, 2025); Kidd v. Pappas, No. 22 C 7061, 2025 WL 3507374 (N.D. Ill. Dec. 8, 2025); Top Metal Buyers Inc. v. Lopinot, No. 3:24-CV-1073, 2025 WL 2780046 (S.D. Ill. Sept. 30, 2025). ↑

100. . Conn. Gen. Stat. § 12-146 (2024). ↑

101. . Minnesota statutes, for instance, authorize interest based on the prime rate. See Minn. Stat. §§ 279.03(1a), 270C.40(5) (2025). The current rate set by the Minnesota Department of Revenue is seven percent. See IntErest Rates for Minnesota Counties, Minn. Dep’t of Revenue https://www.revenue.state.mn.us/interest-rates-minnesota-counties [https://perma.cc/U8FV-HDLY]. In addition to interest, Minneosta authorizes penalties on unpaid property taxes, but the aggregate penaltiesmay not exceed eight percent in the case of homestead property. Minn. Stat. § 279.01(1) (2025). ↑

Introduction

Several months ago, I wrote in Law360 that, given the tenor of the questioning in Trump v. CASA—the birthright citizenship case—I did not expect the court to address the issue of universal injunctions.^[1] That prediction plainly turned out to be wrong. But I was not wrong that Republican and Democratic administrations alike would continue to be troubled by the forum shopping that has given individual district court judges authority to issue rulings of nationwide impact. And I continue to believe, as I said then, that a simple legislative reform could mitigate that concern. For nearly forty years, when several parties appeal the same federal administrative agency decision to different circuit courts of appeal,

the Multi-District Panel randomly assigns venue to a single circuit court. Expanding the Multi-District Panel’s authority to do random selection in district as well as circuit court cases brought under the Administrative Procedure Act (“APA”)^[2] would greatly reduce forum-shopping. A single footnote in Trump v. CASA underscores the continued importance of that suggested reform.“Nothing we say today resolves the distinct question whether the Administrative Procedure Act authorizes federal courts to vacate federal agency action.”^[3] That was the largely overlooked, but immensely significant footnote in Justice Barrett’s opinion in Trump v. CASA—an opinion that itself announced the qualified conclusion that “‘universal injunctions’ . . . likely exceed the equitable authority that Congress has granted to federal courts” under the Judiciary Act of 1789.^[4]

This Article examines the significance of the footnote—a footnote suggesting that the federal courts’ jurisdiction in cases arising under the APA is likely broader than the equitable authority they possess under the Judiciary Act, as construed in Trump v. CASA. This Article’s focus on a singular footnote is not intended to understate the power of the Court’s central holding. That ruling—granting a partial stay^[5] of the universal injunction ordered by a lower court to halt implementation of President Trump’s executive order purporting to end birthright citizenship—was itself only nominally preliminary (hence Justice Barrett’s qualification that universal injunctions “likely” exceed the authority given federal courts). To be sure, the circuit courts have cautioned that stays and preliminary injunctions are not rulings on the merits.^[6] The Supreme Court, however, has made quite clear that its rulings in what has been called its “shadow docket”^[7] put a decidedly heavy thumb on the scales. “Although our interim orders are not conclusive as to the merits,” the Court recently advised, that such orders “inform how a court should exercise its equitable discretion in like cases.”^[8] Justices Gorsuch and Kavanaugh have gone a step further, rebuking judges who they said had flouted the Court’s interim directives, even though “not necessarily ‘conclusive as to the merits.’”^[9]

I.Remedial Authority of the Federal Courts Under the Administrative Procedure Act

The Court’s opinion in Trump v. CASA involved lawsuits challenging the constitutionality of President Trump’s executive order claiming to end birthright citizenship. It did not, however, decide any constitutional issue, turning instead on the Court’s interpretation of the remedial authority given federal courts by statute. Since the lawsuits were challenges to the actions of the President,^[10] and since the President is not an “agency” within the meaning of the APA,^[11] the Judiciary Act of 1789, not the APA, governed the scope of the federal courts’ remedial powers. But with the exception of purely ministerial acts by agencies to carry out executive orders,^[12] most executive orders are implemented by federal agencies whose actions are then presumptively subject to review under the APA.^[13]

The distinction between executive order and the actions of administrative agencies is relevant to the issue of universal judicial remedies because the APA grants federal courts two additional remedial powers akin to, but distinct from, preliminary injunctive relief under the Judiciary Act: the power to stay and the power to set aside agency action. In both cases, as several of the Justices have noted, relief, almost by definition, is not confined solely to the plaintiffs bringing the challenge.

A. Stays of Agency Action Under Section 705

First, section 705 of the APA permits courts to “postpone”, i.e., stay “the effective date of an agency action . . . pending conclusion of the review proceedings.”^[14] While the standards for stays under APA section 705 are the same as those governing requests for preliminary injunctions, i.e., likelihood of success on the merits, irreparable harm, balancing of interests and the public interest,^[15] the relief provided by a preliminary injunction after Trump v. CASA is potentially different. To be sure, a preliminary injunction under the Judiciary Act and a stay under the APA are similar.^[16] “Both,” the Supreme Court observed in Nken v. Holder, “can have the practical effect of preventing some action before the legality of that action has been conclusively determined.”^[17] But a preliminary injunction is directed against a particular party (as relevant here, a government agency) and, under Trump v. CASA, provides direct^[18] relief only to the complainants. By contrast, a stay, while a “kind of an injunction,”^[19] “operates upon the judicial proceeding itself[,] . . . either by halting or postponing some portion of the proceeding, or by temporarily divesting an order of enforceability.”^[20]

Justice Kavanaugh’s concurring opinion in Trump v. CASA makes essentially the same observation, noting that “in cases under the Administrative Procedure Act, plaintiffs may ask a court to preliminarily ‘set aside’ a new agency rule.”^[21] His meaning is plain enough, even if his reference to “set aside,” which appears in section 706, not section 705,^[22] is imprecise. “In my view,” he adds, “there often (perhaps not always, but often) should be a nationally uniform answer on whether a major new federal statute, rule, or executive order can be enforced throughout the United States during the several-year interim period until its legality is finally decided on the merits.”^[23] A stay, which temporarily maintains the status quo ante, means that the challenged rule or regulation will not be in effect, essentially making its import nationwide and not simply applicable to those parties who have sought the stay. Over the decades, many agency rules and regulations have been stayed pending appeal.^[24]

B. Vacatur of Agency Actions Under Section 706

Second, a reviewing court has express authority under the APA to “set aside” final agency action that it finds arbitrary and capricious or otherwise not in accordance with law.^[25] “On its face,” observes Washington University law professor Richard Levin, “the phrase ‘shall . . . set aside” in § 706 seems to mean that a court not only may, but must, “set aside” a rule that it considers unlawful.””^[26] “Thus,” he added, “the normal remedy in this situation is for the court to order that the rule be vacated and remanded to the agency for further consideration. The rule has to be either remanded or not remanded; it cannot be remanded only with respect to an individual plaintiff.”^[27]

Justice Kavanaugh’s concurrence in Corner Post v. Federal Reserve System is to similar effect:

When a federal court sets aside an agency action, the federal court vacates that order—in much the same way that an appellate court vacates the judgment of a trial court. … Over the decades, this Court has affirmed countless decisions that vacated agency actions, including agency rules. … Those decisions vacated the challenged agency rules rather than merely providing injunctive relief that enjoined enforcement of the rules against the specific plaintiffs.^[28]

To be sure, several scholars have maintained that the “set aside” language in the APA should be read more narrowly, as in putting to the side rather than to nullify.^[29] And at least a few of the Supreme Court justices (Alito, Gorsuch and Barrett) have voiced their support for that view.^[30] But Justices Roberts, Kavanaugh and Jackson—all of whom once sat on the D.C. Circuit bench (a group Justice Kagan jokingly referred to as the “DC Cartel”)—“have exhibited strong hostility to that interpretation.”^[31] Indeed, if this view of “set aside” were adopted, it would invalidate the circuit courts’ “ordinary APA remedy” of vacatur of unlawful or arbitrary and capricious agency action.^[32]

As the foregoing discussion suggests, it is at the very least arguable, if not mostly settled that the APA gives federal courts power both to grant preliminary stays of agency orders and regulations and to vacate them in rulings on the merits—both with universal effect. Trump v. Casa’s footnote thus assures that the objections of those troubled by the power this vests with a single judge to issue orders with nationwide application will not have disappeared with the Supreme Court’s decision in that case.

II. The Real—and Bipartisan—Concern About Universal Injunctions is about Forum-Shopping

In the years leading up to Trump v. Casa, judges, legal commentators and representatives of the Biden and Trump Administrations have expressed misgivings about the power of single federal district court judges to grant nationwide injunctions that bring administrative agency actions to a halt. But their objection—at least in the APA context—is not really to the notion that so much power should be vested in a single decisionmaker. After all, as the “DC cartel” observed—drawing from their own prior experiences as circuit court judges— single panels of only three appellate judges can, and do, overturn nationwide agency regulations issued by major agencies like the FTC, FCC, NLRB, Surface Transportation Board, OSHA and others using the APA’s authorization to “set aside” unlawful agency actions.^[33] Rather the concern about district court-issued nationwide injunctions seems to be twofold. Multiple district courts might reach conflicting decisions.^[34] And, probably more importantly, with over seven hundred district court judges to choose from (some being the only judge in a federal district or one of its divisions),^[35] plaintiffs can often select the decisionmakers they believe most likely to rule in their favor.

The latter was the sentiment voiced by Justice Kagan during May 2025 oral arguments in Trump v. Casa. “In the first Trump administration, it was all done in San Francisco, and then in the next administration, it was all done in Texas, and there is a big problem that is created by that mechanism.”^[36] Statistics suggest that from the challengers’ perspective this has been a highly successful strategy.^[37]

The broader underlying problems of conflicting court rulings and forum-shopping will not disappear—there are scores of legal challenges to the actions of the Trump Administration that are already in the courts and many of them involved requests for nationwide injunctive relief for violations of the APA.^[38] But the problems these requests pose are also not beyond the ability of Congress to address. It has addressed the issue before.

A. The Multi-District Panel Has Long Been Used to Reduce Forum-Shopping

One need look no farther than the history of the current treatment of appeals of administrative agency decisions that, by statute, go straight to the various circuit courts of appeal. Congress first addressed the issue of forum-shopping among the circuit courts decades ago, when it enacted 28 U.S.C. § 2112(a), more commonly known as Multi-Circuit Petition Statute.

That 1958 law was intended to consolidate venue for multiple petitions for review of the same agency decision in a single circuit court—the first circuit court in which a petition for review had been filed. But, despite its good intentions, the law soon engendered races to be the first to file. The races themselves became a precision science:

On the afternoon of December 14, 1978, Stephen L. Grossman, an FERC Administrative Law judge, stood poised, stopwatch in hand, waiting for an employee of the Commission to insert a document into a time-stamp machine located in the Commission’s Office of Public Information. Nearby, a lawyer for a natural-gas producer also stood poised, one hand raised skyward. Watching the lawyer intently was another lawyer employed by his firm, positioned, arm likewise raised, in the office doorway; two other employees of the same firm completed a human chain to a colleague standing at a public telephone on the second floor of the same building.^[39]

That is how Nicholas Fels opened his 1980 law review article describing a hearing ordered by the Fifth Circuit to ascertain which of dueling petitions for review of a FERC order—one filed in the D.C. Circuit and the other filed in the Fifth Circuit—came first. His article, another law review article^[40] and a report from the Administrative Conference of the United States (ACUS)^[41]—all published that same year—had a common message. It was time to reform the “race to the courthouse” spectacle that had been fostered by the 1958 statute.

“Courts,” recalled George Washington University law professor Jonathan Siegel, “were required to conduct rather absurd investigations into which of multiple petitions for review had been filed first [like the hearing Mr. Fels described in his article], and the priority given to the first-filed petition favored wealthier parties, who could, for example, afford to pay agents to wait in an agency’s file room and in a courthouse to detect the exact moment an order was issued and to file a petition for review immediately thereafter.”^[42]

The common solution suggested by the two authors and by ACUS was for the courts to adopt a random selection process. It took a while, but in 1988 Congress amended U.S.C. § 2112(a). Under the modified statute, when petitions for review of the same agency action are both served on the agency and filed in multiple circuits within 10 days of the agency’s decision, the matter goes to the Judicial Panel on Multidistrict Litigation (JPML), which then conducts a lottery to determine venue and all related cases get transferred to the circuit court selected in the lottery.^[43] Parties still have the right to seek transfer to “any” other circuit court on grounds that transfer would serve “the convenience of the parties in the interest of justice.”^[44] But even in those limited circumstances where transfer occurs, the case will still be heard by a single court.

This system of consolidation and random selection both addresses the concern that different circuit courts might issue conflicting opinions and mitigates the problem of forum shopping. To be sure, losing parties still file their petitions for review in the circuit courts they perceive as most favorable to them. But major rulemaking proceedings typically involve multiple issues, so even largely prevailing parties will often find some issue to appeal.^[45] And these parties will seek review in the circuit courts they perceive to be most favorable.

Thus, random selection in major agency cases has become commonplace.^[46] But while it works quite well where the statutes contemplate direct review of agency action in the circuit courts, the “vast majority” of challenges to administrative agency actions must first be raised in the federal district courts.^[47] It would not be complicated to utilize the random selection process in the district court context.

Since 1968 the Judicial Panel on Multidistrict Litigation—the same panel that handles random selection of circuit courts—has had authority to transfer civil cases involving “common questions of fact” to a single district court judge or a panel of district court judges for pretrial proceedings.^[48] It has exercised this authority thousands of times.^[49] One of the key purposes of the transfer and consolidation is to avoid inconsistent discovery rulings.^[50]

Consolidation of civil cases brought in multiple district courts for pretrial purposes, however, would serve no practical purpose in the APA context. “[W]hen a party seeks review of agency action under the APA,” the D.C. Circuit noted in its 2001 decision in American Bioscience, Inc. v. Thompson, the district judge “sits as an appellate tribunal” and “[t]he ‘entire case’ on review is a question of law.”^[51] It is this very attribute of APA cases brought in federal district court, however, that makes the random venue selection process contemplated under § 2112(a) so appealing (no pun intended).

As in both the multidistrict pretrial process and the random selection process in circuit court APA cases, deploying a random selection process to establish venue in federal district court APA cases would avoid conflicting court decisions on the same questions of law. And, by creating uncertainty among litigants about a case’s ultimate venue, forum shopping would be less prevalent.

B. Statutory Obstacles to Reform

But none of this can happen under existing law. Subsection (d) of section 2112 itself expressly states that its provisions “are not applicable to proceedings… to review or enforce those orders of administrative agencies, boards, commissions, or officers which are by law reviewable or enforceable by the district courts.”

Nor is there the same time constraint to bring APA cases in the district courts as apply to petitions for review that, by law, go straight to the circuit courts of appeal. Challenges in the latter category must be brought typically within 30 or sixty days of the agency’s final action.^[52] By contrast the six-year catch-all federal statute of limitations applies to APA cases filed in the federal district courts. And the Supreme Court’s June, 2024 decision in Corner Post further complicates that fact by holding that the six year time limit does not begin to accrue for an individual litigant until that litigant has actually been harmed by the agency’s regulations, even ones that may be decades old.^[53] As Justice Kagan noted in her dissent in that case, “[t]he majority’s ruling makes legal challenges to decades-old agency decisions fair game.” Under it, she adds:

Any established government regulation about any issue—say, workplace safety, toxic waste, or consumer protection—can now be attacked by any new regulated entity within six years of the entity’s formation. A brand-new entity could pop up and challenge a regulation that is decades old; perhaps even one that is as old as the APA itself.^[54]

Conclusion

I’m accordingly suggesting a two-fold statutory solution. First, the time limit to seek review of final agency decisions should be uniform and should follow the Hobbs Act standard of 60 days following issuance of final agency action. There is no logical reason that Congress’s catch-all six-year statute of limitations should apply to an APA case that originates in the federal district court, while parties challenging nearly all other APA cases in the circuit courts of appeal must file within 60 days. APA cases are not like ordinary civil litigation in which an individual plaintiff may not discover a potential cause of action until years after the fact. The errors, if any, in an agency decision are laid bare in the decision itself. More importantly, that reform would enhance stability in the law.

Second, Congress should amend section 2112 by expanding the roll of the MDL Panel to encompass random assignments of APA complaints that challenge the same final agency action in multiple federal district courts. In such cases, as it does now in circuit court cases, it would randomly select the venue for a case from among the district courts in which claims were made within the 10-day window following issuance of the final agency action. Then, as it already does in multidistrict pretrial consolidations, it could select a panel of district court judges from the same circuit in which the case was filed to preside over the matter as an appellate tribunal.

The reforms outlined above are relatively straightforward. At least as important, they are politically and philosophically neutral. Legislators of both major political parties know that political majorities change. Knowing this, there’s at least a chance that they may see the advantages of expanding the successful section 2112 model as a means to reduce forum shopping.

1. * Mr. Reiter is Senior Counsel in the Washington, D.C. office of Stinson LLC. He is also an adjunct professor of law at George Washington University Law School, where he teaches a course on regulated industries. From 2005-2025 he served as Executive Editor and later Editor in Chief of the Energy Law Journal.

   . Harvey Reiter, Address Nationwide Injunction Issues with Random Venues, Law360 (June 3, 2025), https://perma.cc/MW4Y-V2RZ. ↑

2. . As explained later in this Article, several statutes call for petitions for review of the decisions of certain administrative agencies to go directly to the circuit courts of appeal. See, e.g., Federal Power Act, 16 U.S.C. § 825l(b); Securities Exchange Act of 1934, 15 U.S.C. § 78y(a)(1). If multiple parties seek review of the same agency ruling in different courts of appeal within a ten-day window, the Multi-District Panel randomly selects the venue. 28 U.S.C. § 2112(a). But the same statute giving the Panel this authority expressly denies it when multiple appeals of administrative agency cases are filed in different federal district courts. 28 U.S.C. § 2112(d). That is significant because most APA cases begin in the district courts, not the courts of appeal. See Multidistrict Litigation Act of 1968, Pub. L. No. 90-296, 82 Stat. 109 (codified at 28 U.S.C. § 1407). ↑

3. . Trump v. CASA, Inc., 606 U.S. 831, 847 n.10 (2025). ↑

4. . Id. at 837 (emphasis added). ↑

5. . The stay was partial in the sense that the preliminary injunction was still to remain in effect as to the named plaintiffs, including a number of state governments. Trump v. CASA, 606 U.S. at 861–62. ↑

6. . See, e.g., Cook Cnty. v. Wolf, 962 F.3d 208, 234 (7th Cir. 2020) (“There would be no point in the merits stage if an issuance of a stay must be understood as a sub silentio disposition of the underlying dispute.”); CASA de Maryland, Inc. v. Trump, 971 F.3d 220, 282 (King, J., dissenting) (4th Cir.) (same), vacated on grant of rehearing en banc, 981 F.3d 311 (4th Cir. 2020). See also ACA Connects v. Bonta, 24 F.4th 1233, 1249 (9th Cir. 2022) (Wallace, J., concurring) (“[A] disposition of a preliminary injunction appeal is not an adjudication on the merits and . . . the parties should not ‘read too much into’ such holdings.”) (internal citation omitted); Ctr. for Biological Diversity v. Salazar, 706 F.3d 1085, 1090 (9th Cir. 2013) (“We have repeatedly emphasized the preliminary nature of preliminary injunction appeals.”). ↑

7. . The “shadow docket”—a term first coined a decade ago by University of Chicago law professor Willaim Baude—refers to the use of emergency orders and summary decisions by the Supreme Court of the United States without oral argument. Nina Totenberg, The Supreme Court and ‘The Shadow Docket’, NPR (May 22, 2023), https://perma.cc/J4KT-XEHR. See, e.g., Whole Woman’s Health v. Jackson, 141 S. Ct. 2494, 2500 (2021) (Kagan, J., dissenting); see also Louisiana v. Am. Rivers, 596 U.S __ (2022) (Kagan, J., dissenting) (criticizing the use of the shadow docket to establish precedent). ↑

8. . Trump v. Boyle, No. 25A11, slip op. at 1 (U.S. July 23, 2025) (order granting stay of district court injunction that had ordered the President to return Consumer Product Safety Commission Commissioner Boyle to her position). The Court’s unsigned order in that case cited the Court’s earlier ruling in Trump v. Wilcox, 605 U.S. __ (2025), another “interim” ruling in which the Court stated its “judgment that the Government faces greater risk of harm from an order allowing a removed officer to continue exercising the executive power than a wrongfully removed officer faces from being unable to perform her statutory duty.” Boyle, slip. op. at 1. While beyond the scope of this Article, it strikes this author as dubious that a single member of a multimember agency—especially a minority member, like the numerous Democratic commissioners of several agencies whom President Trump has fired to date—has any executive power, much less enough executive power, to somehow harm the government irreparably (a putative prerequisite for emergency relief).

   It is also worth noting the jarring juxtaposition between Justice Barrett’s majority opinion in Trump v. CASA and her questions at oral argument in United States v. Texas, 599 U.S. 670 (2023), which involved a challenge by Texas and Louisiana to DHS guidelines prioritizing immigration enforcement to focus on removal of terrorists and dangerous criminals, a little more than two years earlier. On the one hand, Barrett’s opinion fixes the scope of the federal courts’ remedial powers to those extant at a single point in time—1789, the year the Judiciary Act became law. CASA, 606 U.S. at 841. In stark contrast, during the oral argument in United States v. Texas, Justice Barrett—accepting the Government’s claim that “that [vacatur] was not contemplated at the time of the APA’s enactment” in 1946—asked the following question: “Why can’t remedial authority evolve over time? . . . Remedial authority is a flexible concept, and so maybe the courts of appeals have expanded that concept. Why would that be impermissible?” Ronald M. Levin, Vacatur, Nationwide Injunctions, and the Evolving APA, 98 Notre Dame L. Rev. 1997, 2014 (2023) (quoting Transcript of Oral Argument at 49–50, United States v. Texas, 599 U.S. 670 (2023) (No. 22-58)). Apart from this seeming contradiction, Harvard professor Jack Goldsmith offers a separate quarrel with Justice Barrett’s opinion in CASA: Justice Barrett’s opinion, he notes in an article in the Harvard Law Review, rests on the erroneous assumption that section 11 of the Judiciary Act of 1789 “endowed federal courts with jurisdiction over ‘all suits . . . in equity,’” and “thus [equitable remedies] ‘must have a founding era antecedent.’” Jack Goldsmith, Interim Orders, the Presidency, and Judicial Supremacy, 139 Harv. L. Rev. 86, 114–115 (2025) (emphasis added). But, as Goldsmith then notes, the birthright citizenship case is “based on federal question jurisdiction and suits against the United States,” neither of which became the basis for federal jurisdiction until 1875, “not 1789.” Id. at 115–116 (emphasis added). ↑

9. . Nat’l Insts. of Health v. Am. Pub. Health Ass’n, 145 S. Ct. 2658, 2663–4 (2025) (Gorsuch & Kavanaugh, JJ., concurring in part and dissenting in part). The joint statement by Justices Gorsuch and Kavanaugh prompted an apology from long time district court judge William Young, who said he had never before been accused of defying a Supreme Court order, but who explained that he “simply did not understand that orders on the emergency docket were precedent.” Jan Wolfe & Nate Raymond, Judges Vexed by Supreme Court ‘Shadow Docket’ Rulings in Trump Cases, Reuters (Sept. 10, 2025), https://perma.cc/5BK6-RA9E. ↑

10. . CASA, 606 U.S. at 831. ↑

11. . Franklin v. Massachusetts, 505 U.S. 788, 800–801 (1992). ↑

12. . See, e.g., Orr v. Trump, 778 F. Supp. 3d 394, 420 (D. Mass. 2025). When an agency exercises judgment or discretion, it is still subject to APA oversight, otherwise it “would allow the President and agencies to simply reframe agency action as orders or directives originating from the President to avoid APA review.” AIDS Vaccine Advoc. Coal. v. U.S. Dep’t of State, 766 F. Supp. 3d 74, 83 (D.D.C.), enforced, 768 F. Supp. 3d 1 (D.D.C. 2025) (finding likelihood of success on merits of APA claim challenging agency implementation of Foreign Aid EO). ↑

13. . State v. Su, 121 F.4th 1, 15 (9th Cir. 2024); New York v. Trump, 133 F.4th 51, 70–71 n.17 (1st Cir. 2025); Chamber of Com. of U.S. v. Reich, 74 F.3d 1322, 1326–27 (D.C. Cir. 1996). See also Stephen M. Johnson, Deregulation: Too Big for One Branch, But Maybe Not for Two, 53 Seton Hall L. Rev. 839, 853 n.64 (2023). This point may even turn out to be relevant with respect to birthright citizenship. While the President’s Executive Order purported to end birthright citizenship (Exec. Order No. 14,160, 90 Fed. Reg 8449 (Jan. 29, 2025)), it appears that a number of agencies will be involved in attempts to implement it, should currently effective injunctions be lifted. Devan Cole, John Fritze & Priscilla Alvarez, New Documents Show How Passport and Social Security Rules Would Change to Enforce Trump’s Birthright Citizenship Order, CNN (July 28, 2025), https://perma.cc/3GR2-6MG5 (discussing documents that even U.S. citizens may be required to provide “to verify their own immigration status to obtain a passport or Social Security number for their children”). ↑

14. . 5 U.S.C. § 705 (authorizing federal courts to “issue all necessary and appropriate process to postpone the effective date of an agency action or to preserve status or rights pending conclusion of the review proceedings”). ↑

15. . Nken v. Holder, 556 U.S. 418, 434 (2009). See also Virginia Petroleum Jobbers Ass’n v. FPC, 259 F.2d 921, 925 (D.C. Cir. 1958) (denying stay and preliminary injunction as failing to meet the four standards without distinguishing between stays and injunctions); D.C. v. U.S. Dep’t of Agric., 444 F. Supp. 3d 1, 15 (D.C. Dist. Ct. 2020) (“The factors governing issuance of a preliminary injunction also govern issuance of a § 705 stay” (citing Texas v. EPA, 829 F.3d 405, 424, 435 (5th Cir. 2016); Humane Soc’y of United States v. Gutierrez, 558 F.3d 896, 896 (9th Cir. 2009); Cuomo v. U.S. Nuclear Regulatory Comm’n, 772 F.2d 972, 974 (D.C. Cir. 1985))). ↑

16. . In addition to stays, or postponements of the effective date of agency action, APA section 705 also implicitly authorizes injunctive relief, giving a reviewing court the option to “preserve status or rights” instead of postponing implementation of the agency’s orders or regulations. 5 U.S.C. § 705. Preservation of “status or rights” could be read to limit relief to the parties seeking judicial review. See id. ↑

17. . Nken, 556 U.S. at 428. ↑

18. . As the Court noted in CASA, even though preliminary injunctions provide direct relief only to the parties to the case, such injunctions may indirectly and “incidentally” benefit non-parties. 606 U.S. 831, 851 (2025). ↑

19. . Nken, 556 U.S. at 441–42 (Alito, J., dissenting). ↑

20. . Id. at 428–29. ↑

21. . CASA, 606 U.S. at 869 (Kavanaugh, J., concurring). ↑

22. . Id. Justice Kavanaugh’s interpretation of “set aside” is discussed infra Section I.B. ↑

23. . CASA, 606 U.S. at 872 (Kavanaugh, J., concurring). ↑

24. . See, e.g., Texas Top Cop Shop, Inc. v. Garland, 758 F. Supp. 3d 607, 619 (E.D. Tex. 2024); Wages And White Lion Invs., L.L.C. v. FDA, 16 F.4th 1130, 1135–36 (5th Cir. 2021); D.C. v. U.S. Dep’t of Agric., 444 F. Supp. 3d 1, 46, 48 (D.C. Dist. Ct. 2020) (noting that “federal courts have issued hundreds of injunctions reaching beyond the parties in the lawsuit”); Texas v. EPA, 829 F.3d 405, 411 (5th Cir. 2016); In re GTE Serv. Corp., 762 F.2d 1024, 1026 (D.C. Cir. 1985). ↑

25. . 5 U.S.C. § 706. ↑

26. . Levin, supra note 8, at 2005. ↑

27. . Id. (emphasis added). ↑

28. . Corner Post, Inc. v. Bd. of Governors of Fed. Rsrv. Sys., 603 U.S. 799, 830–831 (2024) (Kavanaugh, J., concurring). ↑

29. . Levin, supra note 8, at 2011 (discussing theory advanced by Professors John Hamilton and Samuel Bray). ↑

30. . See United States v. Texas, 599 U.S. 670, 695–704 (2023) (Gorsuch, J., concurring) (questioning whether “set aside” under APA gives courts power to vacate arbitrary or unlawful agency action). ↑

31. . Professor Levin recounts in his 2023 article the hostile reaction then Solicitor General Elizabeth Prologar encountered when arguing that federal courts lack the authority to vacate final agency action under the APA:

    At argument, Solicitor General Elizabeth Prelogar (SG) took the position that a judicial decree under the APA may not vacate or enjoin an agency rule on a universal basis; normally, she suggested it should only provide relief for the benefit of the prevailing challenger.11 Chief Justice Roberts responded with considerable consternation, as did other members of the Court who, like the Chief Justice, had previously served as judges on the D.C. Circuit. “[Y]our position on vacatur,” Chief Justice Roberts said,

     

    sounded to me to be fairly radical and inconsistent with, for example, you know, with those of us who were on the D.C. Circuit, you know, five times before breakfast, that’s what you do in an APA case. And all of a sudden you’re telling us that, no, you can’t vacate it, you do something different. Are you overturning that whole established practice under the APA?

    When the SG confirmed that she thought that “the lower courts, including the D.C. Circuit, have . . . been getting this one wrong,” Roberts replied with a “[w]ow.” The SG went on to assert that the lower courts had not been paying attention to the text, context, and history of the APA. Justice Kavanaugh, another D.C. Circuit veteran, met her assertion directly. He noted that he had served on that court with very eminent judges who paid a lot of attention to those factors.16 He added that the SG’s claim was “a pretty radical rewrite, as the Chief Justice says, of what’s been standard administrative law practice.” Justice Jackson joined in their criticism.

    In the wake of these unsympathetic, if not hostile, reactions from what Justice Kagan jokingly called the “D.C. Circuit cartel,” it seemed clear that the Court was not likely to accept the SG’s view in this case.

    Levin, supra note 8, at 1999–2000 (internal citations omitted). ↑

32. . Ins. Mktg. Coal. Ltd. v. FCC, 127 F.4th 303, 317 (11th Cir. 2025). See also Long Island Power Auth. v FERC, 27 F.4th 705,717 (2nd Cir. 2022) (noting that “‘vacatur is the normal remedy’ under the APA”); Am. Great Lakes Ports Ass’n v. Schultz, 962 F. 3d 510, 519 (D.C. Cir. 2020) (same); Advocs. for Highway & Auto Safety v. Fed. Motor Carrier Safety Admin., 429 F.3d 1136, 1151 (D.C. Cir. 2005) (explaining that vacatur is “normally required”); SE Alaska Conservation Council et al. v. U.S. Army Corp. of Eng’rs, 486 F. 3d 638, 654-55 (9th Cir. 2007) (referring to vacatur as the “normal remedy for an unlawful agency action”). ↑

33. . Levin, supra note 8, at 1999–2000. ↑

34. . See, e.g., Amanda Frost, In Defense of Nationwide Injunctions, 93 N.Y.U. L. Rev. 1065, 1106 (2018). A third objection voiced by opponents of the nationwide injunction is that it prevents the issue presented from “percolating” through various courts, thereby better fleshing the issue out for likely subsequent review by the Supreme Court. See Charlton C. Copeland, Seeing Beyond Courts: The Political Context of the Nationwide Injunction, 91 Univ. Colo. L. Rev. 789, 808 (2020). But this concern doesn’t logically apply to review of administrative agency decisions. Why would it be important for an issue presented in a single administrative order to percolate among multiple district courts, when other administrative agency decisions that go directly to the circuit courts (and get consolidated) don’t need to “percolate”? ↑

35. . See, e.g., Letter from the Brennan Ctr. for Just. of New York Univ. Sch. of L. to H. Thomas Byron III, Sec’y, Comm. on Rules of Prac. & Proc., Admin. Off. of the U.S. Cts. (Sept. 23, 2023), https://perma.cc/N9GX-7EQN. ↑

36. . Transcript of Oral Argument at 147, Trump v. Casa, Inc., 606 U.S. 831 (No. 24A884). ↑

37. . Justice Kagan’s observation was spot on. As recounted in a 2024 Harvard Law Review report there were 64 injunctions during the first Trump Administration, 59, or 92% of them granted by judges appointed by Democratic Presidents. District Court Reform: Nationwide Injunctions, 137 Harv. L. Rev. 1701, 1705 (2024). During the Biden Administration there were 14 injunctions, all of them ordered by Republican-appointed judges. Id. ↑

38. . See, e.g., Sri Medicherla, Repealing Without Procedure, The Regul. Rev. (May 15, 2025), https://perma.cc/4YMH-6FRB (discussing likely APA challenges to agency repeal of regulations without notice and comment); Harvey L. Reiter, Ending EPA’s “Endangerment” Finding, The Regul. Rev. (Apr. 16, 2025), https://perma.cc/D33J-ANCL (same); Harvey L. Reiter, Could the Major Questions Doctrine Limit Reciprocal Tariffs?, The Regul. Rev. (Apr. 30, 2025), https://perma.cc/KRE2-BYCX (discussing lawsuit challenging agency implementation of Trump executive order on tariffs as violation of APA”s “major questions doctrine”; Bonnie Eslinger, Judge Orders Reinstatement Of Many Fired Federal Workers, Law360 (Mar. 13, 2025), https://www.law360.com/articles/2310225?scroll=1&related=1 (discussing judge’s ruling that unexplained firing of probationary workers was arbitrary and capricious under the APA). ↑

39. . Nicholas Fels, Beyond the Stopwatch: Determining Appellate Venue on Review of FERC Orders, 1 Energy L. J. 35 (1980). ↑

40. . Thomas O. McGarity, Multi-Party Forum Shopping for Appellate Review of Administrative Action, 129 U. Pa. L. Rev. 302, 325 (1980). ↑

41. . Jonathan R. Siegel, Admin. Conf. of the U.S., Sourcebook of Federal Judicial Review Statutes 42–43 (2021) https://www.acus.gov/sites/default/files/documents/revised-sourcebook-combined-copy-to-circ-may282021.pdf (crediting a 1980 ACUS report with the idea of amending 28 U.S.C. § 2112(a) to adopt the random selection process). ↑

42. . Id. ↑

43. . U.S.C. § 2112(a)(1)–(3). ↑

44. . U.S.C. § 2112(a)(5). ↑

45. . One self-described management attorney explained the issue this way: “Many union lawyers on the losing side of Board decisions- even if unsuccessful on only a minor issue- therefore ‘race to the courthouse’ to file petitions for review in the District of Columbia, frequently on the very day the Board decision is released.” Harry L. Browne, A Management View of NLRA Reform, 12 Gonzaga L. Rev. 32, 67 (1976). While his characterization was plainly not neutral, he wasn’t wrong. ↑

46. . Two recent examples are the Multi-Circuit Panel’s random selection, respectively, of the Sixth and Fourth Circuit courts of appeal to hear challenges to (1) the FCC’s restoration of the “net neutrality rule,” In re MCP No. 185, No. 24-7000, 2024 WL 3650468, at *3 (6th Cir. Aug. 1, 2024), and (2) the FERC’s Order No. 1920 governing long-term regional transmission planning, In re MCP 190, Nos. 24-1650, (J.P.M.L. Aug. 8, 2024) (ordering consolidation of cases 24-1748, 24-1751, 24-1756 and 24-1650). ↑

47. . Joseph Mead, Choosing the Court to Review Agency Rulemaking, The Regul. Rev. (Nov. 11, 2024), https://perma.cc/4YMH-6FRB. See also, Admin. Conf. of the U.S. Administrative Conference Recommendation 2024-1, Choice of Forum for Judicial Review of Agency Rules, https://www.acus.gov/sites/default/files/documents/Choice-of-Forum-Adopted-Recommendation-2024.06.13_0.pdf (recommending “direct review of agency rules by a court of appeals” rather than beginning in the district courts). ↑

48. . 82 Stat. 109–110 (1968). ↑

49. . Overview of Panel, U.S Jud. Panel on Multidistrict Litig., https://perma.cc/GB8S-79FJ. ↑

50. . Id. ↑

51. . Am. Bioscience, Inc. v. Thompson, 269 F.3d 1077, 1083 (D.C. Cir. 2001). There are limited circumstances in which a district court will allow discovery in an APA case, such as when there is a question whether the agency has provided the full record on which its decision is based. See, e.g., Bar MK Ranches v. Yuetter, 994 F.2d 735, 739 (10th Cir. 1993) (“An agency may not unilaterally determine what constitutes the Administrative Record”). But “the focal point for judicial review should be the administrative record already in existence, not some new record made initially in the reviewing court.’” Camp v. Pitts, 411 U.S. 138, 142 (1973). ↑

52. . Corner Post, Inc. v. Bd. of Governors of Fed. Rsrv. Sys., 603 U.S. 799, 812 (2024) (noting that the Hobbs Act, which applies to many federal agencies, requires petitions for review to be filed in the applicable circuit court of appeals within 60 days of “a final order.”) The Federal Power and Natural Gas Acts impose the same deadline. See 16 U.S.C. § 825l; 15 U.S.C. § 717r(b). ↑

53. . Corner Post, 603 U.S. at 812. ↑

54. . Id. at 862 (Jackson, J., dissenting). ↑

Introduction

“So in the great universities of the east, . . . Harvard and Yale Universities, and all the kindred universities of the United States, derive their revenue from their landed possessions—possessions which they never part from.”

-Delegate Aaron F. Parker, Idaho Constitutional Convention, July 23, 1889.^[2]

In their groundbreaking High Country News article on “land-grab universities,” Robert Lee and Tristan Ahtone argue that Congress’s 1862 Morrill Act,^[3] which famously granted lands held by the United States to states for the purpose of endowing fifty-two universities, “worked by turning land expropriated from tribal nations into seed money for higher education.”^[4] “In all,” they write, “the act redistributed nearly [eleven] million acres—an area larger than Massachusetts and Connecticut combined”—and these redistributed parcels were “scattered mostly across 24 Western states.”^[5] This land, they claim, was “taken from nearly 250 tribes, bands and communities through over 160 violence-backed land cessions.”^[6] This represented “a massive wealth transfer masquerading as a donation,” which “challenges universities to re-evaluate the foundations of their success”^[7] and presents them with “a chance to acknowledge debts that are impossible to repay but unconscionable to ignore.”^[8] Their intervention, including a forensic accounting of nearly all Morrill Act lands and a New York Times op-ed,^[9] received considerable media coverage,^[10] earned a suite of prestigious scholarly awards,^[11] won a clutch of journalism awards,^[12] galvanized demands for reparations from several Tribes,^[13] and provoked an upwelling of “land-grab” scholarship.^[14] In short, their project has been remarkably influential. But they only told part of the story.

As this essay shows, the eye-popping 10.7-million-acre figure actually underestimates the size of university land grants — by at least 4,079,369.20 acres, collectively a landmass larger than Connecticut.^[15] This is because, by 1862, land-grant universities dotted the American landscape. And land-granting didn’t end with the Morrill Act. The practice has surprisingly persisted right up to today: as recently as 2024, the City of West Palm Beach, Florida, along with Palm Beach County, Florida, granted Vanderbilt University, a Tennessee corporation, seven acres valued at $58.8 million.^[16] This essay shows that at least thirty-eight extant universities received land grants before or after the Morrill Act, and that at least seven universities that received Morrill Act lands received other land grants either before or after 1862. Thus, the United States is home to at least ninety land-grant universities. And, as they received lands from various public authorities, universities in the twentieth century started grabbing land by eminent domain, a development that historians have overlooked.^[17]

This short essay tells the long story of university land-granting and -grabbing, from the seventeenth century to the present, for the first time. In doing so, it challenges us to reconsider even the revisionist histories of our land-grant universities. While those histories helpfully call attention to the ways that public authorities initially acquired lands they granted to universities, they do not pay adequate attention to the many land grants before and after the Morrill Act.^[18] The more than four million acres of pre- and post-Morrill Act land grants, assembled for the first time in this essay, add up to over one-third of the lands granted via the Morrill Act. This history tells a more complete story of our landed universities and suggests that the debts Lee and Ahtone mention are potentially much larger than previously thought. Whereas Lee and Ahtone, along with other land-grant historians, direct our attention westward, this essay contends that understanding land-grant universities requires us to first look east. As detailed below, many universities, including those calling themselves “private,” received lands from public authorities and acquired land by eminent domain. If land-grant universities owe a special duty to the public, then “private” landed universities like Harvard and Yale, two of the earliest land-grant universities, owe that duty as much as public landed universities do. In short, the United States is a land of landed universities,^[19] and has been since its inception. To understand our landed universities, we need to disabuse ourselves of what we might call Morrill-Act exceptionalism, and attend to the many other land grants creating our many other land-grant universities.

Vermont, the state from which the 1862 bill’s namesake Justin Morrill hailed, is a case in point. In 1791, the Vermont Legislature included among the preambulatory statements in the University of Vermont’s charter a note that “several grants of land have already been made by this State,”^[20] totaling 29,000 acres.^[21] Vermont was in no way unique. Six years earlier, one preambulatory statement included in the University of Georgia’s charter observed that,

for the great purpose of internal education, divers allotments of land have, at different times, been made, particularly by the [Georgia] Legislature at their Session in July One thousand seven hundred and eighty three, and February One thousand seven hundred and eighty four, all of which may be comprehended and made the basis of one general and complete establishment.^[22]

As historian John Whitehead wrote in his 1973 book The Separation of College and State, “for nearly fifty years after the American Revolution state governments had made land grants and cash grants resulting from ‘windfalls’ such as bonuses for bank charters or federal refunds for wartime expenses to such colleges as Yale, Harvard, Columbia, and Dartmouth.”^[23] And, as Edward Elliott and Merritt Chambers explained in their 1936 book The Colleges and the Courts, “[t]he first Federal land grant to a state for university purposes was to Ohio in 1802.”^[24] Simply put, by the Morrill Act’s passage, land granting was a “historic practice.”^[25] After the Morrill Act, and even into the twenty-first century, public authorities continued granting lands to and for universities. Decades after historian Eldon Johnson set out to correct several “[m]isconceptions [a]bout the [e]arly [l]and-[g]rant [c]olleges,” many misconceptions remain.^[26]

By recovering the land-grant university’s forgotten history, this essay makes six contributions. First, it assembles for the first time pre- and post-Morrill Act land grants in one place.^[27] Second, it demonstrates that the states and their predecessors, and even the proprietors of cities, counties, towns, and villages, rather than the United States, led the way in furnishing universities with “landed endowment[s],”^[28] especially in Massachusetts. Third, it shows that there are many more land-grant universities than scholars previously recognized. Fourth, it suggests that Harvard was the earliest land-grant university in what would become the United States, having received a land grant in 1638 from what is now the Town of Cambridge, as I explain in Part II.^[29] Fifth, it points out that, in the twentieth century, universities added the power of eminent domain to their collection of land-acquiring techniques. Sixth, it demonstrates that America’s land-hungry universities have used at least nine land-acquiring techniques, at different times, to drive their growth^[30]: the town grant, the royal grant, the state grant, the county grant, the federal grant, eminent domain, individual donation,^[31] the trade, and purchase.

Part I addresses the land grants before the Morrill Act. Part II treats the land grants after the Morrill Act. Part III covers eminent-domain use by land-grab universities.

I. Land-Grant Universities Before the Morrill Act

By the time President Abraham Lincoln signed into law the Morrill Act of 1862, granting land to endow universities was a widespread practice.^[32] By then, the states and their predecessors, along with at least three Tribal Nations, had granted at least 779,150.95 acres^[33]:

• 20,000 acres to the College of William & Mary from the King and Queen in 1693.^[34]
• 500 acres to Harvard from Massachusetts in 1719.^[35]
• 1,500 acres to Yale from Connecticut in 1732.^[36]
• 12,500 acres to Harvard from Massachusetts between 1762 and 1774.^[37]
• 54,000 acres to what became Columbia University from New York around 1770.^[38]
• 8,000 acres to what is now Transylvania University from Virginia in 1780.^[39]
• 12,000 acres to Transylvania University from Virginia in 1783.^[40]
• 412 acres to Hampden-Sydney College from Virginia in 1784.^[41]
• 40,000 acres to the University of Georgia from Georgia in 1785.^[42]
• 23,040 acres, or a township, to Dartmouth College from Vermont (then an independent republic) in 1785.^[43]
• 10,000 acres to Dickinson College from Pennsylvania in 1786.^[44]
• 10,000 acres (actually amounting to 10,831.95 acres) to what is now Franklin & Marshall College from Pennsylvania in 1787.^[45]
• 5,000 acres to what is now Washington & Jefferson College from Pennsylvania in 1787.^[46]
• 5,000 acres to what became the University of Pittsburgh from Pennsylvania in 1787.^[47]
• 42,000 acres to Dartmouth from New Hampshire in 1789.^[48]
• 300 acres to Harvard from Massachusetts in 1790.^[49]
• 29,000 acres to the University of Vermont from Vermont by 1791.^[50]
• 115,200 acres to Bowdoin College from Massachusetts in 1794.^[51]
• 1,200 acres to Hampden-Sydney from Virginia in 1794.^[52]
• 46,080 acres to Williams College from Massachusetts in 1796.^[53]
• 6,000 acres to what is now the University of Louisville from Kentucky in 1798.^[54]
• 6,000 acres to Transylvania from Kentucky in 1798.^[55]
• 3,924 acres to Columbia and Union College from New York in 1802.^[56]
• 46,080 acres to Ohio University from Ohio in 1804.^[57]
• 23,040 acres to Bowdoin from Massachusetts in 1806.^[58]
• 23,040 acres to Vincennes University from Indiana Territory in 1806.^[59]
• 23,040 acres to Dartmouth from New Hampshire in 1807.^[60]
• 23,040 acres to Miami University from Ohio in 1809.^[61]
• 23,040 acres to Harvard from Massachusetts in 1809 (for a natural-history professorship).^[62]
• 29,164 acres to what is now Colby College from Massachusetts in 1815.^[63]
• 1,920 acres from the Chippewa, Ottawa, and Potawatomy Tribes to what is now the University of Michigan in 1817.^[64]
• 40 acres to the University of Pittsburgh from Pennsylvania in 1819.^[65]
• 455 acres to Franklin & Marshall from Pennsylvania in 1819.^[66]
• 46,080 acres to Williams College from Massachusetts by 1820.^[67]
• 14,724 acres to the University of North Carolina (UNC) from North Carolina in 1821.^[68]
• 640 acres to UNC from Tennessee in 1821.^[69]
• 12,320 acres to what became the University of Tennessee and the now-defunct Nashville University from Tennessee in 1838.^[70]
• 16,000 acres to what is now Eastern Michigan University from Michigan in 1849.^[71]
• 14,000 acres to what became Michigan State University (MSU) from Michigan in 1855.^[72]
• 7,000 acres to MSU from Michigan in 1858.^[73]
• 23,040 acres to Colby College from Maine in 1861.^[74]

As the above non-exhaustive list shows, states and their predecessors were already established land-granters by 1862, conveying parcels small and large to universities near and far. Some had developed into major land-granters: by the Morrill Act’s passage, Massachusetts alone had granted at least 295,904 acres to universities.^[75] And, curiously, these pre-Morrill-Act grants lacked the requirements found in that Act regarding the kinds of instruction that must be delivered at universities receiving lands.^[76]

Small liberal arts colleges received large tracts of land. The 138,240 acres that Bowdoin received from Massachusetts by 1806 was more than thirteen universities received individually under the Morrill Act: Colorado State University, the University of Delaware, the University of Florida, the University of Idaho, Kansas State University (which claims to be “the nation’s first operational land-grant college”^[77]), the University of Minnesota, North Dakota State University, the University of Nebraska, the University of Nevada, Oregon State University, the University of Rhode Island, the University of Wyoming, and Washington State University.^[78] The recipient of major land grants more than half a century before the Morrill Act, Bowdoin is no less a land-grant university than any of the thirteen aforementioned universities.

But the size of the grant was not always as important as the location.^[79] Just 20 acres in New York City across from where St. Patrick’s Cathedral now stands^[80] from New York State to Columbia University became, according to Whitehead, “the principal endowment of Columbia University,” valued between $6 and $8 million by 1900.^[81]

For its part, Congress granted lands for universities before 1862, including in 1792, 1802 (as mentioned above), 1803, 1804, 1806, 1811, 1816, 1818, 1819, 1820, 1823, 1826, 1827, 1832, 1836, 1838, 1840, 1845, 1851, 1852, 1853, 1854, 1855, 1856, 1857, 1859, and 1861.^[82] Together, these acts granted at least 279,606 acres. The 1803 act assigned 23,040 acres to Mississippi’s now-defunct^[83] Jefferson College. The 1826 act granted 46,080 acres to the University of Michigan, including, according to an 1839 remonstrance the University of Michigan Regents sent to the Michigan Legislature, a total of 14,666 acres located across what the Regents called the Niles and Nottawassippi Reservations.^[84] The 1851 act granted 46,080 acres to support the University of Minnesota. The 1852 and 1856 acts granted 4,166 and 22,000 acres, respectively, to Indiana University. The 1855 act reserved 46,080 acres for the University of Utah.^[85] The 1857 act granted an additional 46,080 acres to the University of Minnesota.^[86] Like the Morrill Act lands, these lands were meant to generate permanent endowments. Also like the Morrill Act lands, these lands were sometimes far-flung^[87]: the 1856 act, according to historian Samuel Harding, granted Indiana University “land in this and other States.”^[88]

Many of these pre-Morrill Act land grants from Congress supported universities that do not today identify as land-grant universities, including Ohio University and the Universities of Michigan, New Mexico, and Oregon.^[89] Yet nineteenth-century writers understood that these universities benefitted from federal land grants. For example, one Detroit Daily Advertiser columnist wrote in 1854 that

[t]he State did not originate the University [of Michigan], and, in fact, it has never cost the people of the State one cent. It was founded (as the canal of the Saut is dug,) by the proceeds of the sale of public lands belonging to the United States, which money, but for the establishment of the university would have gone into the treasury at Washington, and have been lost to Michigan. The University, therefore, is not the creature of Michigan, but a gift to it.^[90]

As mentioned above, the University of Michigan was indeed a gift from the United States, but it was also a gift from the Sachems, Chiefs, and Warriors of the Chippewa, Ottawa, and Potawatomy Tribes, who granted at least 1,920 acres to the University during its infancy.^[91]

Historians tend to think that “the Land Grant movement” emerged in the second half of the nineteenth century.^[92] The historical record glossed here tells a different story. By 1862, the United States was already home to many land-grant universities. Put differently, the Morrill Act did not “create[] the American land-grant universities.”^[93] Instead, the Act endowed universities according to a practice embraced by pubic authorities long before there was any such thing as the United States. And, as discussed below, public authorities across the expanding nation continued endowing universities with lands in the twentieth and twenty-first centuries.

II. Land-Grant Universities After the Morrill Act

After the Morrill Act, public authorities granted many lands to universities. As shown above, this was nothing new. What was new post-1862 was universities taking land by eminent domain, as discussed below. That is, instead of public authorities acquiring land and then distributing it to universities, universities seized land themselves. This Part reminds us that, because these practices persist, the history of land-grant universities is hardly history at all.

Congress continued granted lands to and for universities after the Morrill Act, including in an 1881 act, an 1884 act, an 1889 act, two 1894 acts, an 1898 act, a 1924 act, and a 1961 agreement. Together, these acts granted at least 532,345.10 acres. The 1881 act granted 46,080 acres for what is now the University of Wyoming^[94] with grants to other universities^[95]; the 1884 act conveyed 46,080 acres for the University of Alabama^[96]; the 1889 act netted 142,080 acres for the University of Washington,^[97] over 50,000 acres more than Washington State University received via the Morrill Act^[98]; one 1894 act conveyed 110,000 acres for the University of Utah^[99]; the other 1894 act assigned 23,040 acres for the support of the University of Mississippi^[100]; the 1898 act conveyed 65,000 acres for the University of New Mexico’s benefit^[101]; the 1924 act granted 100,000 acres for the benefit of what is now the University of Alaska^[102]; and the 1961 agreement ceded 65.10 acres to Dickinson College.^[103]

And Congress wasn’t the only public authority granting lands after the Morrill Act. The People of Texas and their Legislature set aside 2,231,400 acres for the University of Texas (UT) before the turn of the century. In 1839, the then-independent Republic of Texas set aside 231,400 acres for the yet-unborn university,^[104] including 40 acres for what would become UT’s Austin campus.^[105] In their 1876 Constitution, the People of Texas set aside 1,000,000 acres of land for the prospective university, having by that instrument rescinded an 1858 grant reserving about 3,200,000 acres for that purpose.^[106] This 1,000,000 acres alone was more than any university received under the Morrill Act. In 1883, the year of Texas’s birth, the Texas Legislature doubled the already unprecedented gift by granting another 1,000,000 acres for UT.^[107] Thus, UT received more than double the land that Cornell—the university that Lee and Ahtone identify as the one “[w]ho received the most expropriated Indigenous land,” having acquired 977,909 acres through the Morrill Act.^[108] Notwithstanding the fact that UT received more than 2,000,000 acres more than Texas A&M received via the Morrill Act,^[109] the latter apparently does not consider the former a Texas land-grant university.^[110]

UT Tower in March 2024 (photo credit: Michael Banerjee)

In the present century, public authorities continue granting lands to universities. Through acts passed in 2000 and 2005, the Alaska Legislature conveyed 250,000 acres to the University of Alaska to generate funds by sale or use for the university’s endowment.^[111] In a 2009 case, the Alaska Supreme Court held that the acts violated the Alaska Constitution’s prohibition on dedicating tax- or license-proceeds to special purposes, and ordered the lands returned to the State.^[112] The University of Alaska thereby brought into the twenty-first century America’s tradition of failed land grants, a tradition initiated—on the other side of the continent and on other side of three centuries—by Harvard. Massachusetts unsuccessfully granted Harvard 800 acres, 2,000 acres, 2,100 acres, and 1,000 acres in 1652, 1653, 1658, and 1683, respectively.^[113] The following century, in 1770, New Hampshire granted 23,040 acres to Dartmouth College in vain.^[114] Though difficult to detect, these failed grants are part of our land-granting history, and the intensity of the American land-granting impulse—along with the intensity of the contestation over the same—cannot be understood without them.

Importantly, towns, cities, villages, and counties granted many lands to universities, like the 2.66 acres, 120 acres, and 3.5 acres that the Town of Cambridge assigned to Harvard in 1638, 1652, and 1683, respectively^[115]; the 8 acres the proprietors of New Haven, Connecticut, transferred to Yale in 1717^[116]; the 250 acres the proprietors of Rutland, Massachusetts, gave to Harvard in 1718^[117]; the 5,700 acres the Town of Schenectady, New York, conveyed to Union College in 1795^[118]; the 40 acres the citizens of Ann Arbor, Michigan, who organized themselves as the Ann Arbor Land Company, gave to the University of Michigan in 1837^[119]; the 4 acres donated by the Village of Ypsilanti, Michigan, to Eastern Michigan University in 1850^[120]; the 5 acres the people of Greenville, Pennsylvania, granted to Thiel College in 1871^[121]; the 10 acres the citizens of Beaver Falls, Pennsylvania, assigned to Geneva College in 1880^[122]; the 43 acres the City of Cincinnati granted the University of Cincinnati in 1890^[123]; the 12 acres the City of Philadelphia conveyed the University of Pennsylvania in 1894^[124]; the 3.99 acres the Ypsilanti villagers gifted to Eastern Michigan University in 1896^[125]; the 60 acres the City of Akron granted to what is now the University of Akron in 1928^[126]; the 588 acres that Ector County, Texas, conveyed to UT in 1969^[127]; and the aforementioned 7 acres donated to Vanderbilt in 2024. Although they were mostly modest, these grants add up to 6,857.15 acres, and therefore constitute a considerable part of the history of our land-grant universities. This essay does not come close to documenting all grants of this sort, but it nonetheless makes clear that land-granting was (and is) a tool of Americans and their governments, big and small.

Notably, this essay does not address grants to universities from other universities. By the early nineteenth century, UNC, having received from the state many escheated lands, was a major landowner and, surprisingly, itself a land-granter. In 1822, UNC assigned 20,000 acres to the University of Tennessee and 40,000 acres to Nashville University.^[128] Major but not always meticulous. Around the middle of the century, the University sold a 70,000-acre tract of land for only $13,000—about nineteen cents per acre—after underestimating the parcel’s size by 60,000 acres.^[129]

As shown above, the Morrill Act’s passage did not end Americans’ land-granting enterprise. After 1862, public authorities continued granting lands to universities as they had done before the Morrill Act. By the close of the nineteenth century, universities added the power of eminent domain to their complement of land-acquiring techniques. In the twentieth century’s first third, universities began grabbing land by eminent domain. Thus, the twentieth century birthed true land-grab universities, a striking development which university historians have not before identified and to which we turn in the next section.

III. Land-Grab Universities

In the twentieth century, universities newly imbued with the power of eminent domain began taking land themselves in direct university land-grabs.^[130] They seized land with constitutional, statutory, or (in at least one case) no authorization.

The People of Michigan, speaking through the 1908 Michigan Constitution (since amended), ordained that “The regents of the university of Michigan shall have power to take private property for the use of the university, in the manner prescribed by law.”^[131] That same instrument provided that “The regents of the University and their successors in office shall continue to constitute the body corporate known as ‘the Regents of the University of Michigan.’ ”^[132] Thus, the People of Michigan, in their sovereign capacity, delegated this authority directly to the Regents, who are, according to the Michigan Supreme Court, “the highest form of juristic person known to the law, a constitutional corporation of independent authority.”^[133] By the 1920’s, the Regents were grabbing land. Consider two such cases.

First, in 1922, the Michigan Regents used this power to condemn about two blocks in Ann Arbor, on which they built the Law School’s iconic Lawyers Club, a student dorm that still stands today.^[134] In April 1922, Michigan Law alumnus William Cook pledged to donate $1.5 million to build the Lawyers Club.^[135] Specifically, Cook indicated to the Regents, “If agreeable to you, I will erect on the two blocks on South University avenue, between South State street and Tappan avenue, a law students’ combined club and dormitory building.”^[136]

The Lawyers Club in May 2025 (photo credit: Michael Banerjee)

After accepting the proposal, the Regents coordinated with the relevant state agency to disburse funds to purchase about two blocks to the south of campus.^[137] When eleven owners refused the Regents’ overtures, the Regents resolved to seize the land by eminent domain. They “declared the taking of this property a public necessity for use of the University, and directed the Attorney General of the State of Michigan to institute condemnation proceedings.”^[138] A jury trial resulted in a verdict finding that taking the land was a public necessity and awarding the property owners damages.^[139] Six property owners accepted the damages and deeded their respective properties to the University. Five others appealed, asking the Michigan Supreme Court to review the case, which it agreed to do.

Among the questions presented to the Court by the appealing property owners in People v. Brooks was: “The statute under which the proceedings were brought is unconstitutional, to the extent that it attempts to authorize proceedings in behalf of the regents of the University of Michigan, because the title of the statute is not broad enough to authorize enactment of such authority.”^[140] The statute at issue was entitled “An act to authorize proceedings by the state to condemn private property for public use,”^[141] which was important because the 1908 Constitution mandated that “No law shall embrace more than one object, which shall be expressed in its title.”^[142] The appealing property owners argued that the statute did not authorize proceedings by the Regents. Specifically, they claimed that the statute permitted condemnation proceedings by the state only and that the Regents were “a constitutional corporation, independent of the state, separate and distinct in its authority.”^[143] Thus, they argued, the statute did not contemplate proceedings through which the Regents could take private property.

The Court disagreed. After conceding that the Regents were “a separate entity, independent of the state as to the management and control of the University and its property,” the Court concluded that the University is “nevertheless a department of the state government, created by the Constitution to perform state functions.”^[144] More to the point, “the real estate which it holds or acquires is public property belonging to the state, held by the corporation in trust for the purposes of the University, which are public purposes.”^[145]

The appealing property owners supported “their conception of the legal character of this corporation” by reference to the Court’s 1893 opinion in Weinberg v. Regents.^[146] In that case involving payment for building materials used by a contractor, the Court wrote in relevant part that, “Under the constitution [of 1850] the state cannot control the action of the regents. It cannot add to or take away from its property without the consent of the regents.”^[147] The Weinberg Court also clarified that the University “is not under the control of the state when it acts through its executive or legislative departments, but of the regents, who are directly responsible to the people for the execution of their trust.”^[148]

The Court determined that the Weinberg decision rested “on the constitutional right of the regents to the absolute and exclusive control of all University property,” a right that “has been recognized by every judicial decision of this court in which the question has been considered.”^[149] Nonetheless, the Court had never “held that the University was not a state institution, nor that the real estate which the regents are authorized to acquire and hold for University purposes is not property of the state.”^[150] Thus, the law embraces only one purpose—“the condemnation of private property for public use”—and condemning property for the Regents’ use serves that purpose.^[151]

In response to the appealing property owners’ argument that the Lawyers Club was not a public purpose, the Court quoted at length from Cook’s proposal, and concluded that “The claim that the property to be acquired is not for public use is so plainly without merit that we do not deem it necessary to enter into any extended discussion of it.”^[152] (This was a time when courts did not deem arguments meritless as a rhetorical flourish.) The Regents’ seizure of the land for the purpose of constructing the Lawyers Club was accordingly affirmed, with costs.

This case has all but been forgotten. In 2024, Michigan Law celebrated the Lawyers Club centennial without mention of the Lawyers Club’s origins.^[153] A recent monograph on Cook and his philanthropy notes that “the regents voted to acquire the needed land,” but does not discuss how they went about doing so.^[154]

Second, later on in the 1920’s, the Regents condemned 10.5 acres to construct an eighteen-hole golf course.^[155] After a trial conducted under the same statute authorizing the Brooks trial, a jury returned a verdict after finding it necessary to take the land and awarded compensation to the owners of the condemned land.^[156] The owners, seeking compensation exceeding the jury award, opposed the taking because the land was condemned on behalf of the Board in Control of Athletics of the University of Michigan, a corporation they argued lacked the power to take land by eminent domain.^[157] The Michigan Supreme Court agreed to review the case by certiorari.^[158]

The property owners presented the following question to the Court: “Have the Board of Regents of the University of Michigan, acting through the state, the power to condemn land for the use and benefit of a private corporation, whose funds not the state’s or university’s, will pay for the purchase of the property condemned?”^[159]

Answering in the negative, the Court reasoned that “A state agency, vested with power of eminent domain, may not employ the power, directly or indirectly, for the use and benefit of another, unless so authorized by law.”^[160] However, “the answer given to the question does not at all decide this case.”^[161] Why not? Because the Board in Control, “a creature of the Board of Regents” and “a nonprofit corporation,”^[162] was organized under a statute of the Michigan Legislature “for the declared purpose of the furtherance in general of the physical betterment of the students at the University of Michigan; particularly the conduct of intercollegiate athletics.”^[163] The Board’s charter provided for election to the Board by the Regents and directed that its funding would derive from ticket sales and athletic fees.^[164]

Based on these provisions, the Court concluded that the Board, “while a corporate entity, is but an operating agency of the Regents of the University in the management of the designated educational activities and, at all times, under full control of the Regents.”^[165] The Court accordingly affirmed the taking.

After successfully seizing the land, “[t]he regents unofficially opened the course in the fall of 1930 as the first golfers to play the new U-M Golf Course,” according to the golf course’s website.^[166] The title of this second case, People, for Use and Benefit of Regents of University of Michigan v. Pommerening, seems ironically apropos. Like the Law School’s celebration of the Lawyers Club, the golf course website includes no mention of these contested condemnation proceedings.^[167]

In addition to the Michigan Regents’ constitutional authority, turn-of-the-century legislatures imbued universities with the eminent-domain power, including the Universities of Minnesota, Nebraska, New Mexico, Southern California, Texas, and Wisconsin; what is now Connecticut College; what is now Longwood University; and Purdue University.^[168] The corporation called the Trustees of Purdue University used that power to take 32.54 acres to build dorms in 1927.^[169] Longwood University acquired property by eminent domain at least eleven times between 1911 and 1991.^[170] And what became Auburn University took land in the absence of any explicit constitutional or legislative authorization.^[171] Lest one mistake this for an exclusively western or twentieth-century phenomenon, the Trustees of Stony Brook University used the eminent-domain power to acquire 246 acres to build a research campus in 2004.^[172]

Universities also received condemned lands in the twentieth century in more roundabout ways. In 1958, the American Association of University Professors completed a little-known survey of sixteen universities, including the University of California, the University of Chicago, Columbia, Johns Hopkins University, Harvard, the University of Minnesota, Vanderbilt, and Tulane University, that “demonstrated conclusively that the lack of available land for expansion was one of the most acute problems facing each university surveyed.”^[173] The following year, on the strength of the survey, the University of Chicago and other universities supported federal legislation that “was designed to overcome the universities’ inability to acquire land for expansion by allowing them to participate in urban renewal projects.”^[174] Specifically, “[t]he proposal provided that state housing agencies would acquire slum lands near university areas by negotiated sales or condemnations and then turn the land over to the university for academic redevelopment.” In the alternative,

the university could purchase land directly, relying upon state condemnation when necessary, and redevelop it in accordance with a plan approved by the city. In either case, the cost to the university of acquiring the land could be treated as if it were the city’s own contribution to the cost of the project.^[175]

This made it “possible for the city to receive two dollars in federal money for every one dollar spent by the university without contributing anything to the project from its own funds.”^[176] As a result, the university would “finance its own expansion program, [and] . . . also would help clear the slums near campus and raise federal funds for the city in the process.”^[177] The University of Chicago’s representative emphasized to Congress “that the national interest lay in permitting private as well as public universities to rely upon eminent domain.”^[178]

In 1959, Congress passed “the bill as originally drafted by the universities,”^[179] thus “effectively placing eminent domain power behind the university in land acquisitions.”^[180] By 1969, cities had taken about 2,000 acres of “privately owned urban land for university expansion”^[181] on behalf of at least 120 universities.^[182] Once again, Congress joined the party late but made the loudest entrance. The federal government ceased granting money under the statute in 1975.^[183] Nonetheless, universities continued to vicariously acquire urban land by eminent domain into the present century. The Empire State Development Corporation took 17 acres of privately owned land by eminent domain for the purposes of developing Columbia’s West Harlem campus.^[184]

In sum, by the turn of the twentieth century, universities began acquiring land by eminent domain. In the twenty-first century, public authorities continue to grant, and universities continue to grab, land. The conclusion suggests what scholars should do to understand their landed universities.

IV. Conclusion

The Morrill Act represents one of many land grants benefitting American universities. By 1862, legislatures across the country, including Congress, were experienced land-granters. And public authorities continued granting lands after 1862, all the way up to the present century. The “long”^[185] land-grant movement lasted from at least 1638 to 2024. In addition to pushing the land-grant timeline back into the colonial period and pulling it up into to the twenty-first century, this essay invites us to ask what happened to the over 4 million acres granted to universities by acts other than the Morrill Act and to the lands grabbed by or on behalf of universities via eminent domain.^[186]

Three implications flow from this history. First, it suggests that the debts that might be owed for lands granted to universities are likely considerably larger than previously thought. This is in part because public and private universities alike received much more land from public authorities than scholars have indicated and in part because certain of the grants were for lands within Tribal reservations.

Second, it obliges scholars to reconsider and expand our definitions of the land-grant university and the “land-grant movement,” which was well underway by 1862.^[187]

Third, it invites us to further question the distinction between public and private universities. In fact, Americans would have recognized no such distinction before the nineteenth century.^[188] This history troubles the already-specious public-private distinction as applied to universities, whose fundamental publicness was established early on in its one-thousand-year history.^[189] That fundamental publicness persisted in the United States until the myth of the private university emerged in the nineteenth century.^[190] The emphasis that university historians place on the Morrill Act, which mainly granted lands for universities we today consider “public,” has obscured the extent to which universities we now call “private” also amassed wealth through land grants, meaning that they are also land-grant universities. By so obscuring this reliance on public support, the scholarly fixation on the Morrill Act has reified the dubious distinction between public and private universities.^[191]

Much research remains to be done. We need to study the relationship between universities and land unconstrained by the Morrill Act. As noted above, our landed universities acquire land in many ways and their techniques for doing so change over time. Viewing the American university’s relationship to land—a relationship this research suggests is likely more important to the general development of our universities than previously thought—through the Morrill Act has seriously hindered our ability to understand that relationship. Specifically, we need to study this relationship by attending to the manifold ways that universities obtain lands. We also need to research the land grants that failed if we are to fully apprehend the American land-granting impulse. Lastly, we must study the universities’ exercise of eminent domain. As of today, we have no clue how much land universities have taken by eminent domain in the more than 125 years since they began exercising that power or what they have done with the land they acquired in this manner.

If debts arising from the possession and sale of these lands are owed to certain Native American Tribes, then those debts might be significantly larger than Lee and Ahtone indicate. Notwithstanding the diligent work undertaken by them and their colleagues, and others, many university land grants and grabs remain obscure. To understand what debts might be owed, we need to first identify America’s many land-grant and -grab universities.^[192] Doing so requires looking east.

1. * J.D., Harvard Law School. Ph.D. Candidate in Jurisprudence and Social Policy, UC Berkeley Law. Geremia Lizier-Zmudzinski, Amanda Ghibaudo, and Emma Moore provided excellent editorial assistance. The archivists at the University of Vermont’s Silver Special Collections Library and the University of Michigan’s Bentley Historical Library offered indispensable guidance. Bentley staff, students, and volunteers insightfully commented on an earlier draft, and archivist Carolyn Alam helpfully guided me to and through certain of the Bentley’s valuable late-twentieth-century materials. Participants at the November 2025 History of Education Society meeting, UC Berkeley Law’s November 2025 Jurisprudence and Social Policy Forum, and the 2025–2026 UC Berkeley Center for Studies in Higher Education’s Gardner Seminar provided useful and generous feedback. Jonathon Booth, Ming-hsi Chu, Haris Durrani, Jonathan Glater, Bruce Kimball, T.J. Mertikas, Tony Platt, Joseph Rodriguez, Joan Wallach Scott, Brian Soucek, and Christopher Tomlins commented incisively on earlier drafts. All errors are my own. ↑

2. . 1 Proceedings and Debates of the Constitutional Convention of Idaho, 1889, at 705 (I.W. Hart ed., 1912). ↑

3. . See Morrill Act of 1862, ch. 130, § 4, 12 Stat. 503, 504 (codified as amended at 7 U.S.C. § 304 (1988)). I do not address the 1890 Morrill Act universities because that Act granted funds from public-land sales, instead of land itself. See Morrill Act of 1890, ch. 841, § 1, 26 Stat. 417 (codified as amended at 7 U.S.C. §§ 322–23 (1988)). ↑

4. Robert Lee & Tristan Ahtone, Land-Grab Universities, High Country News (Mar. 30, 2020), https://www.hcn.org/issues/52-4/indigenous-affairs-education-land-grab-universities/ [https://perma.cc/HUM6-2JUQ]. ↑

5. . Id. ↑

6. . Id. ↑

7. . Id. ↑

8. Tristan Ahtone & Robert Lee, Ask Who Paid for America’s Universities, N.Y. Times (May 7, 2020), https://www.nytimes.com/2020/05/07/opinion/land-grant-universities-native-americans.html [https://perma.cc/R4KT-7YZA]. ↑

9. . Id. ↑

10. . See, e.g., Bennett Leckrone, Why New Research Calls Some Flagships ‘Land-Grab Universities’, Chron. Higher Educ. (Apr. 8, 2020), https://www.chronicle.com/article/why-new-research-calls-some-flagships-land-grab-universities/ [https://perma.cc/P4PV-S2YR]. ↑

11. . See Dr Robert Lee Wins Awards for ‘Land-Grab Universities’, Univ. Cambridge Fac. Hist. (Nov. 16, 2021), https://www.hist.cam.ac.uk/news/dr-robert-lee-wins-awards-land-grab-universities-0 [https://perma.cc/Z37P-XA44]. ↑

12. . See id. ↑

13. . See Sara Weissman, Tribes Call for Reparations from University of Minnesota, Inside Higher Ed (July 13, 2023), https://www.insidehighered.com/news/quick-takes/2023/07/13/tribes-call-reparations-university-minnesota [https://perma.cc/S79J-EFJB]. ↑

14. . A 2021 Native American and Indigenous Studies “Intervention” featured twelve commentaries and an interview with Lee and Ahtone. See K. Tsianina Lomawaima, Kelly McDonough, Jean M. O’Brien & Robert Warrior, Editors’ Introduction: Reflections on the Land-Grab Universities Project, 8 Native Am. & Indigenous Stud. 89, 90 (2021). Several scholarly articles have been written in this vein. See, e.g., Richard Finlay Fletcher, Not About U: Social Dissonance at a Land-Grab University, Future Human., June 21, 2024, at 1; Caitlin Harvey, University Land Grabs: Indigenous Dispossession and the Universities of Toronto and Manitoba, 104 Canadian Hist. Rev. 467 (2023). At least one doctoral dissertation and two master’s theses have since been written on land-grab universities. See Isabelle Soulay Soifer, “This Is Not a Campus, This Is New York City.”[:] Colonial University Land Grabs and the Fight for Homefulness (Sept. 19, 2025) (unpublished Ph.D. dissertation, University of California, Irvine) (on file with eScholarship, University of California); Georgina E. Mullins, Wyoming’s Land-Grab: Education for All, Geographies of Responsibility, and Land-Grant Revenue for Indigenous Students (2025) (unpublished M.A. thesis, University of Wyoming) (on file with ProQuest); Jenny Sullivan, Contested Landscapes: The Origins and Consequences of Rhode Island’s Land-Grant Institutions (1862–1892) (2024) (unpublished M.A. thesis, University of Rhode Island) (on file with Digital Commons at University of Rhode Island). ↑

15. . See State Area Measurements and Internal Point Coordinates, U.S. Census Bureau (2010), https://www.census.gov/geographies/reference-files/2010/geo/state-area.html [https://perma.cc/B347-AC3M]. ↑

16. . See Ellen Rosen, Can a University From Tennessee Help Accelerate Growth in West Palm Beach?, N.Y. Times (Oct. 19, 2025), https://www.nytimes.com/2025/10/19/business/vanderbilt-university-expansion.html [https://perma.cc/F9XD-M7BP]; Stephen Ross and Cody Crowell, BS’07, Fuel Momentum for West Palm Beach Campus, Vanderbilt Univ., https://giving.vanderbilt.edu/impact-stories/ross-crowell-west-palm-beach-campus/ [https://perma.cc/C7FM-5EQ7]. ↑

17. . For example, neither John Thelin’s classic history of American higher education, nor Christopher Loss’s more recent monograph, mentions eminent domain at all. See generally John R. Thelin, A History of American Higher Education (2004); Christopher P. Loss, Between Citizens and the State: The Politics of American Higher Education in the 20th Century (2012). Historian Davarian Baldwin’s 2021 book is the exception, although that text does not identify the American university’s acquisition of eminent-domain powers as a turn-of-the-century development, as this essay does. See Davarian L. Baldwin, In the Shadow of the Ivory Tower: How Universities are Plundering Our Cities 9–10, 29 (2021). This historical oversight might have resulted from the tendency to view the American university’s relationship to land almost exclusively through the prism of the Morrill Act. One example of recent work focused on other instances of “landing” universities is the 2024 Grist article on lands held by states in trust for universities by Lee, Ahtone, and their collaborators. See Tristan Ahtone, Robert Lee, Amanda Tachine, An Garagiola, Audrianna Goodwin et al., Misplaced Trust: Stolen Indigenous Land Is the Foundation of the Land-Grant University System. Climate Change Is Its Legacy., Grist (Feb. 7, 2024), https://grist.org/project/indigenous/land-grant-universities-indigenous-lands-fossil-fuels/ [https://perma.cc/QM54-PA4X]. But even there, the authors analyzed the lands held in trust for fourteen universities that also received Morrill-Act lands. See id. On Arizona’s trust lands, see generally Teresa Miguel-Stearns, Samantha Ginsburg, & Kristen Cook, More Than Morrill: The Intertwined History of Indian Land Dispossession, Arizona Statehood, and University Enrichment, 16 Ariz. J. Env. L. & Pol’y 1 (2025). ↑

18. . See generally Sharon Stein, Unsettling the University: Confronting the Colonial Foundations of US Higher Education (2022); Margaret A. Nash, Entangled Pasts: Land-Grant Colleges and American Indian Dispossession, 59 Hist. Educ. Quart. 437 (2019). ↑

19. . This turn of phrase modifies a chapter title in James Axtell’s excellent history of the American research university. See James Axtell, Wisdom’s Workshop: The Rise of the Modern University 147 (2016) (“A Land of Colleges”). ↑

20. . An act for the purpose of founding an University at Burlington (Nov. 3, 1791), University of Vermont Silver Special Collections, Record Group 1, Trustees Minutes, 1791–1811, Book 1. ↑

21. . See Julian Ira Lindsay, Tradition Looks Forward[,] The University of Vermont: A History 1791–1904, at 10 (1954). ↑

22. . See A Charter for a New State University: The University of Georgia Charter (1785), in Essential Documents in the History of American Higher Education 42 (John R. Thelin ed., 2014). ↑

23. . John S. Whitehead, The Separation of College and State: Columbia, Dartmouth, Harvard, and Yale, 1776–1876, at 3 (1973). ↑

24. . Edward C. Elliott & M. M. Chambers, The Colleges and the Courts[:] Judicial Decisions Regarding Institutions of Higher Education in the United States 232 (1936). ↑

25. . Id. Similarly, Sharon Stein notes that, “[b]y the time of the first Morrill Act in 1862, there was already a history of public federal land grants for educational institutions. Between 1796 and 1861, seventeen states received federal land grants for higher education; several states offered their own land grants, as well.” Stein, supra note 17, at 133 (citing Thelin, supra note 16, at 75–76). But this summation adequately captures neither the number nor consistency of the pre-1862 land grants. ↑

26. . See Eldon L. Johnson, Misconceptions About the Early Land-Grant Colleges, 52 J. Higher Educ. 333, 333 (1981). Thank you to Dan Premauden for alerting me to this article. ↑

27. . This essay assembles more of the non-Morrill Act grants than other works on the subject have. For example, historian Roger Geiger mentions only some of the pre-Morrill Act land grants in his two recent books on the history of American higher education. See generally Roger L. Geiger, The History of American Higher Education: Learning and Culture from the Founding to World War II (2015); Roger L. Geiger, American Higher Education since World War II: A History (2019). Geiger’s discussion of the pre-Morrill Act grants does not include acreage, and the many post-Morrill Act grants make no appearance in these books. See Geiger, The History of American Higher Education, supra note 26, at 162, 167, 169, 282. ↑

28. . J. J. Lane, History of the University of Texas 59 (1891). ↑

29. . Eldon Johnson observed that “Harvard, Yale, William and Mary, Dartmouth, and Michigan all [were] the beneficiaries of either colonial or state gifts,” but did not argue that Harvard was the first land-grant university. Johnson, supra note 25, at 335. ↑

30. . See Baldwin, supra note 16, at 6 (“Urban development is higher education’s latest economic growth strategy.”). In 1815, Union College “received in exchange for [a] building 3,000 acres of land in the county,” when the “city and county of Schenectady [sought to use it] for a court house, jail and city offices.” Andrew Van Vranken Raymond, 1 Union University[:] Its History, Influence, Characteristics and Equipment 57 (1907). ↑

31. . It seems to have been forgotten that Johns Hopkins’s munificent gift to John Hopkins University included 330 acres. See Daniel C. Gilman, The Johns Hopkins University (1876-1891), in 9 Johns Hopkins University Studies in Historical and Political Science 41 (Herbert B. Adams ed., 1891). Individual donations were sometimes very large. For example, an individual gifted Middlebury College 30,000 acres in 1915. See David M. Stameshkin, The Strength of the Hills: Middlebury College, 1915–1990 xvi (1996). Middlebury sold 21,858.90 of these acres a couple of decades later to the United States for $435,906.41, see id. at 30, that tract forming part of the famous Green Mountain National Forest. See David Mitchell Stameshkin, The Town’s College: Middlebury College, 1800–1915, at 565–66 (1978) (unpublished Ph.D. dissertation, University of Michigan) (on file with ProQuest). Harvard received 23,040 acres in Maine in 1814 from an individual donor. See Josiah Quincy, 2 The History of Harvard University 311, 416 (1860). Harvard appears to have subsequently sold the land to establish a fund supporting a theology professorship. See id. at 416. Land-hungry universities have not always needed to turn to public authorities to sate themselves. ↑

32. . See Roger L. Williams, The Origins of Federal Support for Higher Education: George W. Atherton and the Land-Grant College Movement 35 (1991) (“Land grants for educational institutions at all levels were common before the Civil War. By 1857, more than six million acres had been set aside for such purposes.”). ↑

33. . This figure does not include failed land grants, discussed infra. ↑

34. . See Royal Charter, William & Mary Special Collections, https://scrc-kb.libraries.wm.edu/royal-charter [https://perma.cc/JYW7-9QRS]. ↑

35. . See Quincy, supra note 30, at 229, 530. ↑

36. . See Whitehead, supra note 22, at 13. ↑

37. . See Quincy, supra note 30, at 229. ↑

38. . See Addison Brown, The Elgin Botanic Garden[,] its Later History and Relation to Columbia College[,] the New Hampshire Grants[,] and the Treaty with Vermont in 1790, at 22 (1908). ↑

39. . See Herbert Baxter Adams, Thomas Jefferson and the University of Virginia 308 (1888). ↑

40. . See Alvin Fayette Lewis, History of Higher Education in Kentucky 36–37 (1899). Transylvania apparently lost these 12,000 acres in 1792. See id. at 37 n. 2. ↑

41. . Alfred J. Morrison, The College of Hampden-Sidney[:] Calendar of Board Minutes[,] 1776–1876, at 109 (1912). ↑

42. . See Whitehead, supra note 22, at 47. ↑

43. . See Adam R. Nelson, Dartmouth College v. Woodward: Colleges, Corporations, and the Common Good 47 (2025) (“Wheelock had told Vermont that Dartmouth would educate Vermont’s youth tuition free if the state gave the college a 23,040 acre ‘glebe’ township that had been taken from the Anglican church during the war. He solicited this land for ‘The Institution, Embracing Dartmouth College and Moor’s Charity School,’ and on June 14, 1785, Vermont granted his request.”); Ahtone et al., supra note 16 (“One 6-by-6 mile square on the grid is known as a township.”). Dartmouth’s 23,040-acre grant almost equaled the 29,000-acre grant that Vermont gave to the University of Vermont, and the former was given in one block, while the latter lands were spread across 120 districts, “mostly in the northern counties,” and therefore “much more difficult to administer.” Lindsay, supra note 20, at 10. The same year of Dartmouth’s grant, Vermont incorporated the granted lands as the Town of Wheelock, named after Dartmouth College’s second president. See Morgan v. Cree, 46 Vt. 773, 774 (1861). ↑

44. . See James Henry Morgan, Dickinson College[:] The History of One Hundred and Fifty Years[,] 1783–1933, at 126 (1933). ↑

45. . See Joseph Henry Dubbs, History of Franklin and Marshall College 24, 28–29, 91 (1903). ↑

46. . See Saul Sack, 1 History of Higher Education in Pennsylvania 72 (1963). ↑

47. . See id. at 79, 308. ↑

48. . See Whitehead, supra note 22, at 35; Nelson, supra note 42, at 48–49. ↑

49. . See Quincy, supra note 30, at 271. ↑

50. . See Lindsay, supra note 20, at 10; An act for the purpose of founding an University at Burlington, supra note 19. ↑

51. . See Louis C. Hatch, The History of Bowdoin College 5–7 (1927) (“The final section of the act of incorporation gave to the college five townships of land, each six miles square to be taken from the unappropriated state lands in Maine.”). ↑

52. . See Morrison, supra note 40, at 109. ↑

53. . See Calvin Durfee, A History of Williams College 82 (1860) (“In January, 1796, the Legislature granted two townships of land . . . .”). ↑

54. . See R. H. Eckelberry, The History of the Municipal University in the United States 23 (1932). ↑

55. . Lewis, supra note 39, at 51. ↑

56. . Raymond, supra note 29, at 527. Apparently, Union College also received lands from New York in 1800, but I have been unable to find the number of acres included in this grant. See Murray Bartlett, The Story of Geneva College, 23 N.Y. Hist. 14, 14 (1942). ↑

57. . See Matheny v. Golden, 5 Ohio St. 361, 362–63 (1856). ↑

58. . 1805 Mass. Acts 684. This might have disappointed the Bowdoin Trustees, who petitioned the Massachusetts Legislature for two townships of land. See id. ↑

59. . Bd. of Trs. for the Vincennes Univ. v. Indiana, 55 U.S. 268, 271–72 (1852). ↑

60. . Nelson, supra note 42, at 65–66. ↑

61. . See Frank W. Blackmar, The History of Federal and State Aid to Higher Education in the United States 218 (1890). ↑

62. . See id. at 91. ↑

63. . See Ernest Cummings Marriner, The History of Colby College 132 (1963). ↑

64. . See Howard Henry Peckham, The Making of the University of Michigan, 187–1992, at 12 (Margaret L. Steneck & Nicholas H. Steneck eds., 1994). The University appears to have later acquired Tribal lands initially granted to the Rector of the Church of St. Anne at Detroit via the Treaty of Fort Meigs, which would bring the total acreage granted by the three Tribes ultimately benefitting the University up to 3,840 acres. See Compl. ¶¶ 10–11, Children of the Chippewa, Ottawa and Potawatomy Tribes v. Regents of Univ. of Mich., Washtenaw County Cir. Ct. (1971), University of Michigan Bentley Historical Library, Box 1, Children of the Chippewa v. U-M Regents. According to former Michigan Chief Justice Thomas Cooley, “[t]he gift . . . was fully equal in positive value and prospectively superior to the gifts for like purposes which made John Harvard and Elihu Yale immortal, and quite as justly entitles [Chief] Tontagini and his associate Chieftains to grateful remembrance among the founders of colleges.” Thomas McIntyre Cooley, Michigan: A History of Governments 312–13 (2d. 1886). Thank you to Carolyn Alam for locating this quotation. Indeed, according to a University of Michigan website, “sale of the original land gift provided a significant part of the basis for Michigan’s permanent endowment.” University Celebrates History of Land Grant from Native Americans, Univ. Mich. News (Nov. 13, 2002), https://news.umich.edu/university-celebrates-history-of-land-gift-from-native-americans/ [https://perma.cc/H62B-HMTC]. ↑

65. . See Sack, supra note 45, at 79–80. ↑

66. . See id. at 308. ↑

67. . According to an 1820 report of the Massachusetts Legislature, “[f]our townships of land have been granted to Williams College, and it has an interest in a certain gore of land,” but further details on the two townships alluded to in the report have thus far eluded me. Durfee, supra note 52, at 422 (quoting The report of the Committee of Both Branches of the Legislature on the Subject of the Removal of the College to Northampton (1820)). At least one of these townships was located in what would become Maine. See Randell v. Mallett, 14 Me. 51, 53 (1836). ↑

68. . Kemp P. Battle, 1 History of the University of North Carolina[:] From Its Beginning to the Death of President Swain, 1789–1868, at 390 (1907). ↑

69. . See Thurston v. Univ. of North Carolina, 72 Tenn. 513, 514 (1880). ↑

70. . State v. Nashville Univ., 23 Tenn. (1 Hum.) 157, 160–61 (1843); see also Aaron D. Purcell, University of Tennessee 7–8 (2007); Bill Carey, Tennessee’s list of defunct colleges is longer than you think, Tenn. Mag. (Nov. 1, 2022), https://www.tnmagazine.org/tennessees-list-of-defunct-colleges-longer-than-you-think/ [https://perma.cc/53JC-EN8G]. ↑

71. . Egbert R. Isbell, A History of Eastern Michigan University[,] 1849–1965, at xi, 8–9 (1971). ↑

72. . See An Act for the Establishment of a State Agricultural School, 1855 Mich. Pub. Acts 279, reprinted in The Agricultural College of the State of Michigan 11–12 (1857); Madison Kuhn, Michigan State[:] The First Hundred Years[,] 1855–1955, at 6 (1955). ↑

73. . Kuhn, supra note 71, at 72 (“The prospective value of the Morrill grant was indicated by the worth of a smaller one of about 7000 acres of swamp land in adjoining townships given by the State Legislature of 1858 in response to a request from President Williams.”). ↑

74. . See Marriner, supra note 62, at 137. ↑

75. . See Frederick Rudolph, The American College and University: A History 186 (2d ed. 1990) (“Massachusetts was generous with townships in Maine.”). ↑

76. . For example, Massachusetts granted Bowdoin a township in 1806 in a single-paragraph resolve, “[s]ubject to such restrictions and Reservations as is usual in Similar Grants.” See 1805 Mass. Acts 684. ↑

77. . See The History of the Land-grant University, AgReport (2021), https://www.k-state.edu/agreport/documents/AgReport_Fall2021_web.pdf [https://perma.cc/M43L-QSJF]. ↑

78. . See Robert Lee, Tristan Ahtone, Margaret Pearce, Kalen Goodluck & Geoff McGhee, et al., Land-Grab Universities, High Country News, https://www.landgrabu.org [https://perma.cc/25PJ-Q4QG]. ↑

79. . The attention paid to smaller land grants, in addition to larger ones, is one important contribution made by this essay. ↑

80. . See Brown, supra note 37, at 1, 6, 14. ↑

81. . Whitehead, supra note 22, at 26–27; see also Sidney Sherwood, The University of the State of New York: History of Higher Education in the State of New York 207 (Herbert B. Adams ed., 1930) (“Thus, solely through the influence of the president of Union [College], Columbia received that magnificent property which to-day forms its principal endowment.”). ↑

82. . See David G. Sansing, The University of Mississippi: A Sesquicentennial History 6 (1999) (“On March 3, 1803, Congress responded to the general assembly’s request and granted Jefferson College a township of land.”); State v. Trs. of Vincennes Univ., 2 Ind. 293, 294 (1850); Sterling v. Regents of the Univ. of Michigan, 68 N.W. 253, 254 (Mich. 1896); McAfee’s Heirs v. Keirn, 15 Miss. 780, 781 (Err. & App. 1846); C. W. Butterfield, History of University of Wisconsin 47–48 (1879); James Gray, 1 The University of Minnesota 1851–1951, at 17 (1951); William A. Rawles, Historical Sketch, in Indiana University[,] 1820–1904, at 12 (Samuel Bannister Harding ed., 1904); John Pillsbury, Address to University of Minnesota Alumni 12 (1893) (discussing Congress’s 1857 Enabling Act, granting additional 46,000 acres); Regents of Univ. of N.M. v. Graham, 264 P. 953 (N.M. 1928); Donald G. Tewksbury, The Founding of American Colleges and Universities Before the Civil War 173 (1932) (“Oregon was admitted as a state in 1859 and received a Congressional land grant for the establishment of a state university in that year.”). For all other acts, see id. at 187. ↑

83. . See Historic Jefferson College, Miss. Dep’t Archives & Hist., https://www.mdah.ms.gov/explore-mississippi/historic-jefferson-college [https://perma.cc/J3PD-S8W3]. ↑

84. . See Charles Kendall Adams, Historical Sketch of the University of Michigan 2–3 (1876); Proceedings of the Board of Regents 1837–1864, at 73 (1915). ↑

85. . See State v. Montello Salt Co., 98 P. 549 (Utah 1908), reversed on other grounds sub nom. Montello Salt Co. v. Utah, 221 U.S. 452 (1911). ↑

86. . See Pillsbury, supra note 81, at 12. ↑

87. . See Lee et al., supra note 77. ↑

88. . Rawles, supra note 81, at 13. ↑

89. . See Matheny v. Golden, 5 Ohio St. 361 (1856); Whitehead, supra note 22, at 49; Edward C. Elliott & M. M. Chambers, Charters and Basic Laws of Selected American Universities and Colleges 573 (1934). Historian Frederick Rudolph wrote that “Ohio University and Miami University . . . were the prototypes of the land-grant-supported state universities.” Rudolph, supra note 74, at 276; see also Ch. LX, Title 1, § 1 (1872), in The Organic and Other General Laws of Oregon 779 (1874) (“A state university is hereby created, and in order to devote to the purposes of education the donation of seventy-two sections of land to the state for the use and support of a state university, made under and by virtue of an act of congress, passed February 14, 1859, entitled ‘An act for the admission of the state of Oregon into the Union.’ Such university is hereby permanently located at the town of Eugene, Oregon.”). ↑

90. . The University of Michigan—Facts, Detroit Daily Advertiser (Jan. 31, 1854). ↑

91. . See Compl. ¶ 4, Children of the Chippewa, Ottawa and Potawatomy Tribes v. Regents of Univ. of Mich., Washtenaw County Cir. Ct. (1971), University of Michigan Bentley Historical Library, Box 1, Children of the Chippewa v. U-M Regents. ↑

92. . Geiger, The History of American Higher Education, supra note 26, at 283. Historian Nathan Sorber begins his 2018 book on land-grant universities as I begin this essay, with the University of Vermont, noting that “[t]he university had been founded in 1792, and under the terms of the Morrill Act, the Vermont legislature designated it a land-grant college in 1864,” with assistance from Morrill, who was then a University of Vermont trustee that Sorber calls “the father of the land-grant college movement.” Nathan M. Sorber, Land Grant Colleges and Popular Revolt: The Origins of the Morrill Act and the Reform of Higher Education 1 (2018). ↑

93. . Nathan M. Sorber, “A History of American Land-Grant Universities and Regional Development,” in Handbook of Universities and Regional Development (Attila Varga & Katalin Erdős eds., 2019). ↑

94. . See Ross v. Trs. of Univ. of Wyo., 228 P. 642, 647 (Wyo. 1881); Elliott & Chambers, supra note 88, at 573. ↑

95. . See State v. Collins, 53 P. 1114, 1114 (Mont. 1898). ↑

96. . See Cox v. Bd. of Trs. of Univ. of Ala., 49 So. 814 (Ala. 1909). ↑

97. . See State v. Callvert, 74 P. 1018 (Wash. 1904). ↑

98. . See Lee et al., supra note 77. ↑

99. . See Montello Salt Co., 98 P. 549 (1908). ↑

100. . See Sansing, supra note 81, at 156. ↑

101. . See Regents of Univ. of N.M. v. Graham, 264 P. 953, 953 (N.M. 1928). ↑

102. . State v. Univ. of Alaska, 624 P.2d 807, 809 (Alaska 1981). ↑

103. . See Charles Coleman Sellers, Dickinson College: A History 594 (1973). ↑

104. . See W.J. Battle, History of the University of Texas at Austin, Tex. State Hist. Assoc. (2023) https://www.tshaonline.org/handbook/entries/university-of-texas-at-austin [https://perma.cc/DM9U-H5KD]. ↑

105. . See Lane, supra note 27, at 254 (“The University, as now constituted, occupies the center of a square of forty acres of ground in the northern portion of Austin.”); Avrel Seale, What’s the Story Behind the Forty Acres?, UT News (Sep. 6, 2018), https://news.utexas.edu/2018/09/06/whats-the-story-behind-the-forty-acres/ [https://perma.cc/294R-7V38]. Thus, not all land grants were used for endowment purposes: some grants were for campuses. ↑

106. . See Interpretive Commentary in 2007 Main Volume, Tex. Const. Art. 7, § 15 (1876). ↑

107. . See Lane, supra note 27, at 81 (discussing UT’s pair of 1,000,000-acre land grants). ↑

108. . Lee & Ahtone, supra note 3. ↑

109. . See Lee et al., supra note 77. ↑

110. . See Texas Land Grants Day, Tex. A&M Univ., https://ieei.tamu.edu/txlgd/ [https://perma.cc/BH7R-NXNX]. ↑

111. . Interestingly, the 2000 act permitted the University to select its own lands, but the Alaska Legislature rescinded this permission in its 2005 act. See Se. Alaska Conservation Council v. State, 202 P.3d 1162, 1165–66 (Alaska 2009). ↑

112. . See id. ↑

113. . See Josiah Quincy, 1 The History of Harvard University 40-41 (1860). And this says nothing of reneged land grants, like the 10 acres promised to what is now Oglethorpe University by the City of Atlanta, Georgia, which the City ultimately refused to deliver. See Allen P. Tankersley, College Life at Old Oglethorpe 143 (paperback ed. 2009). Nor does it say anything about proposed land grants, like the 40 acres offered to Oglethorpe by LaGrange, Georgia. See id. at 123. ↑

114. . See Nelson, supra note 42, at 29–30, 34, 41, 48. ↑

115. . See Quincy, supra note 112, at 511–12. We might understand town grants as the earliest form of “mass giving in higher education,” rather than the Yale Alumni Fund, incorporated in 1890. Bruce A. Kimball & Sarah M. Iler, Wealth, Cost, and Price in American Higher Education: A Brief History 58–65 (2023). Massachusetts also granted Harvard’s head 500 acres in 1639. See Willard Wallace Smith, The Relations of College and State in Colonial America 35 (1949) (unpublished Ph.D. dissertation, Columbia University) (on file with ProQuest). ↑

116. . See Ebenezer Baldwin, History of Yale College: From Its Foundation, A.D. 1700, to 1836, at 307 (1841). ↑

117. . Quincy, supra note 30, at 530. ↑

118. . See Raymond, supra note 29, at 23. ↑

119. . See Peckham, supra note 63, at 18. ↑

120. . See Isbell, supra note 70, at 11. ↑

121. . See Sack, supra note 45, at 138. ↑

122. . See id. at 102. ↑

123. . See University of Cincinnati Record: Annual Reports, 1908, at 36 (1909). ↑

124. . See Edward Potts Cheyney, History of the University of Pennsylvania, 1740-1940, at 352 (1940). ↑

125. . See Isbell, supra note 70, at 149. ↑

126. . See Eckelberry, supra note 53, at 137. ↑

127. . See Calvert v. Hull, 475 S.W.2d 907 (Tex. 1972). ↑

128. . See Battle, supra note 67, at 383, 401; Carey, supra note 69. ↑

129. . See Battle, supra note 67, at 623. ↑

130. . See Robert M. O’Neil, Private Universities and Public Law, 19 Buffalo L. Rev. 155, 183–84 (1970) (discussing eminent-domain power of private universities). ↑

131. . Mich. Const. art. XIII § 4 (1908). ↑

132. . Mich. Const. art. XI § 4 (1908). ↑

133. . Bd. of Regents of Univ. of Mich. v. Auditor Gen., 132 N.W. 1037, 1040 (Mich. 1911). On constitutional corporations, see generally Michael Banerjee, California’s Constitutional University: Private Property, Public Power, and the Constitutional Corporation, 1868–1900, 18 Cal. Leg. Hist. 215 (2023). ↑

134. . See People, for Use of Regents of Univ. of Mich. v. Brooks, 194 N.W. 602 (Mich. 1923); About the Lawyers Club, Mich. L., https://michigan.law.umich.edu/resource-center/about-lawyers-club [https://perma.cc/D6J3-BQ9W]. ↑

135. . Brooks, 194 N.W. at 603; see also Margaret A. Leary, Giving it All Away: The Story of William W. Cook and His Michigan Law Quadrangle 140 (2011). ↑

136. . Brooks, 194 N.W. at 604. ↑

137. . Id. at 603. ↑

138. . Id. ↑

139. . Id. ↑

140. . Id. ↑

141. . Id. ↑

142. . Id. ↑

143. . Id. ↑

144. . Id. ↑

145. . Id. (citing Auditor Gen. v. Regents of the Univ., 47 N.W. 440 (Mich. 1890)). ↑

146. . Id. 604. ↑

147. . Weinberg v. Regents of Univ. of Mich., 56 N.W. 605, 608 (Mich. 1893). ↑

148. . Id. ↑

149. . Brooks, 194 N.W. at 604. ↑

150. . Id. ↑

151. . Id. ↑

152. . Id. at 605. ↑

153. . See Sharon Morioka, 100 Years of the Lawyers Club, L. Quadrangle (2024), https://quadrangle.michigan.law.umich.edu/issues/summer-2024/100-years-lawyers-club [https://perma.cc/D8XE-PFTD]. ↑

154. . See Leary, supra note 134, at 140. An older study of the Law Quadrangle’s architecture does not mention how the Regents acquired the land either. See generally, Ilene H. Forsyth, The Uses of Art: Medieval Metaphor in the Michigan Law Quadrangle (1993). ↑

155. . People, for Use and Benefit of Regents of Univ. of Mich. v. Pommerening, 230 N.W. 194, 194 (Mich. 1930). ↑

156. . Id. ↑

157. . Id. ↑

158. . See id. at 195. ↑

159. . Id. ↑

160. . Id. ↑

161. . Id. ↑

162. . Id. ↑

163. . Id. (internal quotation marks omitted). ↑

164. . Id. at 196. ↑

165. . Id. ↑

166. . History, Univ. of Mich. Golf Course, https://umgolfcourse.umich.edu/history/ [https://perma.cc/74D6-QJA4]. ↑

167. . See id. ↑

168. . See Knapp v. State, 145 N.W. 967 (Minn. 1914) (affirming condemnation for University of Minnesota railway from University farm to University campus); Burger v. State Female Normal Sch., 77 S.E. 489 (Va. 1913) (affirming jury award for taking for what is now Longwood University); Mohler v. Bd. of Regents of Univ. of Neb., 165 N.W. 954 (Neb. 1917) (affirming jury award for taking of 50-by-50-foot parcel for University of Nebraska use); Territory v. Crary, 103 P. 986 (Terr. N.M. 1909) (reversing judgment dismissing New Mexico Regents’ condemnation petition); Univ. of S. Cal. v. Robbins, 1 Cal.App.2d 523 (Cal. Dist. Ct. App. 1934) (affirming condemnation by University of Southern California for construction of library); Cochran v. Cavanaugh, 252 S.W. 284 (Tex. Civ. App. 1923) (dissolving injunction restraining Texas Regents and University land acquisition board from condemning certain land for University of Texas use). On Minnesota, Nebraska, Wisconsin, see Elliott & Chambers, supra note 88, at 360, 371, 565 (quoting Minn. Stat. ch. 187 § 2 (1905); Neb. Comp. Stat. § 83-135 (1905); Wis. Stat. ch. 33 § 1(5) (1919)); on New Mexico and Connecticut College, see Elliott & Chambers, supra note 23, at 464-67, 470-71 (first discussing Conn. Coll. for Women v. Calvert, 88 A. 633 (Conn. 1913), then discussing ch. 137 of the General Laws of Texas (1921), and then discussing § 3693, Laws of New Mexico (1897)). ↑

169. . See Russell v. Trs. of Purdue Univ., 168 N.E. 529 (Ind. 1929). ↑

170. . See W. Taylor Reveley IV Letter to Delores McQuinn at 2 (Jan. 15, 2025), https://s3.amazonaws.com/snwceomedia/rot/0dae3efd-202a-4b3e-b98f-9c48e25e0764.original.pdf [https://perma.cc/A84C-LNPZ] (“Between 1911 and 1991, historical research has revealed that condemnation proceedings were pursued in 11 properties.”); Savannah Bartlett, Nick Hamlett, & Tyler Bagent, “That Was Our Community”: How Longwood Occupied the Triangle, Rotunda (July 24, 2025), https://www.therotundaonline.com/article/2025/07/that-was-our-community-how-longwood-occupied-the-triangle (arguing “Longwood’s use of eminent domain was particularly destructive to the surrounding Black community”). ↑

171. . See Denson v. Ala. Polytechnic Inst., 126 So. 133 (Ala. 1930) (affirming judgment condemning lands for what is now Auburn University in absence of explicit grant of eminent-domain power). ↑

172. . See Gyrodyne Co. of Am., Inc. v. State Univ. of N.Y. at Stony Brook, 794 N.Y.S.2d 87 (N.Y. App. Div. 2005). ↑

173. . The other participating universities were: George Washington University, the University of Illinois, Massachusetts Institute of Technology, New York University, the University of Pennsylvania, St. Louis University, and Washington University in St. Louis. See Charles Fels, N. T. Adams, Richard Carmody, Margaret E. Clark, Randolph H. Lanier et al., The Private Use of Public Power: The Private University and the Power of Eminent Domain, 27 Vand. L. Rev. 681, 707–08 n. 110 (1974). ↑

174. . Id. at 708. ↑

175. . Id. ↑

176. . Id. ↑

177. . Id. ↑

178. . Id. at 709. ↑

179. . Id.; see also 42 U.S.C.A. § 1463. ↑

180. . Fels, supra note 172, at 718. ↑

181. . Id. at 686; see also id. at 686 n.12 (citing Letter from John K. Johnson, Assistant for Legis. Aff., HUD to Congressman Richard Fulton (Mar. 28, 1974)). ↑

182. . See Baldwin, supra note 16, at 31 (“By 1964, there were 154 projects supported by the 112 credits program that involved 120 colleges and universities and 74 hospitals.” (footnote omitted)). ↑

183. . See 42 U.S.C.A § 5316(a). ↑

184. . See Kaur v. N.Y. S. Urb. Dev. Corp., 15 N.Y.3d 235, 244–45 (2010). ↑

185. . See Jacquelyn Dowd Hall, The Long Civil Rights Movement and the Political Uses of the Past, 91 J. Am. Hist. 1233 (2005). ↑

186. . Because of the difficulties attending the study of grants of this sort, especially in the seventeenth and eighteenth centuries, more research will have to be done to confirm that the granted lands were in fact conveyed and, if so, what universities did with these lands. Some of this work is already underway. See Sophie Gao, Undergrad Researchers Find Harvard Owned, Sold Indigenous Land in Maine, Harv. Crimson (Dec. 10, 2025), https://www.thecrimson.com/article/2025/12/10/new-research-harvard-indigenous-land/ [https://perma.cc/S9NL-Y6KV]. Importantly, as Robert Lee said in a 2022 interview, because a university retains their “corporate personhood,” a “university from the 1860s is the university today.” Hayden Royster, This Land is Their Land, Cal. Mag. (June 8, 2022) (quoting Lee), https://alumni.berkeley.edu/california-magazine/2022-summer/this-land-is-their-land/ [https://perma.cc/XFN6-8EGV]. Thank you to Elliott Masouredis for sharing this article with me. ↑

187. . See Earle D. Ross, Democracy’s College: The Land-Grant Movement in the Formative Stage 13 (1942) (“In their origins all the branches of the American system of education from lowest to highest are ‘land grant.’ ”). ↑

188. . See Whitehead, supra note 22, at 6 (arguing that “no clear distinction [between public and private universities] existed before the Civil War.”); see also Christopher Kutz, Publics in Action: The Self-Making of Civic Life 121 (2025) (noting that, when it comes to universities, “our strategies of distinction have failed”). ↑

189. . See Walter Rüegg, Foreword, in 1 A History of the University in Europe[:] Universities in the Middle Ages xix (Hilde De Ridder-Symoens ed., 1992) (noting university was one “[o]f the three acknowledged powers of medieval European society—regnum, sacerdotium, and studium”). ↑

190. . See Rudolph, supra note 74, at 185 (“One great barrier to determining who paid the bills of the American college is the myth of the privately endowed independent college, a myth that was not encouraged until the colleges discovered that they could no longer feed at the public trough and had, in one sense, indeed become private.”); Nelson, supra note 42, at 223 (“[I]t would have been inconceivable in 1769 that Dartmouth College was a completely private corporation, independent of governmental oversight.”). ↑

191. . See Baldwin, supra note 16, at 23 (noting Morrill Act “helped fund public universities to support agricultural research,” among other things). ↑

192. . I do not address all land grants here. For example, the South Carolina Legislature granted “an extensive parcel of land, in an elevated situation” in 1801 to what is now the University of South Carolina. Judge Brevard’s Observations on the Legislative History of South Carolina, in 1 South Carolina Statutes at Large 425, 437 (Thomas Cooper ed., 1836). They then granted the University 10 acres in 1833. See Trs. of Univ. of S.C. v. Trs. of Acad. of Columbia, 67 S.E. 951, 952 (S.C. 1910). I hope university legal historians will study this grant and ones like it. On university legal history, see Michael Banerjee, University Legal History, Cultural Critique Online (2025), https://manifold.umn.edu/read/cc-frame-014-rev-03/section/a01d89fc-2765-4ffc-aadc-88e987c5c897 [https://perma.cc/WNK3-V336]. ↑

Introduction: The Three Kings and Their Domains

The landmark Microsoft Ireland case highlights the limits of U.S. law enforcement authority to compel access to data stored abroad during criminal investigations. In 2013, a magistrate judge in the Southern District of New York ordered Microsoft to produce emails held on a server in Dublin, Ireland under the Stored Communications Act of 1986.^[1] Microsoft released domestic data but withheld the rest, arguing that U.S. courts had no jurisdiction over foreign servers. ^[2] Thus, what started as a routine drug trafficking investigation exposed a deeper jurisdictional tension between cloud computing’s borderless nature and traditional territorial sovereignty.^[3] Microsoft Ireland raised worldwide concerns about data access in criminal investigations, with 289 groups from 37 countries supporting Microsoft’s position.^[4] The tension between borderless cloud infrastructure and territorially bounded authority exists in nearly all technology domains. For global education technology platforms, cross-border data storage raises similar questions as those raised in Microsoft Ireland about who can access student information and under what legal regime.

Educational technology, or “EdTech,” refers to the digital platforms that schools increasingly rely on for instruction and assessment.^[5] Because these platforms handle vast amounts of important student information, they play a central role in today’s debates over privacy and cross-border data laws.^[6] Many operate within global data infrastructures, where private companies collect, process, and analyze student data across borders to enable real-time decision-making at both individual and systemic levels.^[7] Yet legal regimes still struggle to regulate these data flows, balancing privacy rights, cybersecurity, and expanding data usage.^[8]

Because data crosses borders, it is often unclear who controls it and which laws apply. This has resulted in a battle between “three Kings.” EdTech companies (King Innovation) want data to flow freely; national regulators (King Sovereignty) want to keep it at home; and law enforcement agencies (King Authority) seek authority to obtain data regardless of where it is stored. This struggle shapes privacy rights, legal compliance, national security, and investigative powers. Like Lear’s daughters, each King vies for the crown—control over educational data.

This Note focuses on a key flashpoint in the battle: how the United States’ Clarifying Lawful Overseas Use of Data Act (CLOUD Act) challenges domestic and international privacy laws governing EdTech. Section I introduces the three Kings and explains how each competes for control of cross-border data. For example, when student data becomes evidence in a criminal investigation, EdTech companies, domestic privacy regulators, and cross-border law enforcement authorities all demand control. To understand why these entities compete, Section II then examines the foundational legal and philosophical differences underlying three jurisdictions: the United States, the European Union, and China. These differences, reflected in distinct approaches to privacy rights, state power, and data governance, shape how each jurisdiction regulates EdTech platforms. Section III then discusses how these philosophical differences manifest in specific domestic privacy laws, including the United States’ Family Educational Rights and Privacy Act (FERPA), and Children’s Online Privacy Protection Act (COPPA), the European Union’s General Data Protection Regulation (GDPR), and China’s Personal Information Protection Law (PIPL). Section IV presents that when EdTech operates globally, these frameworks may collide with the CLOUD Act, which permits expanded access by U.S. authorities to data stored abroad. Finally, in Section V, I propose reforms to harmonize these laws, such as expanding safe-harbor certification frameworks to provide predictable, government-approved standards for cross-border compliance.

I. The Three Kings

A. King Innovation: EdTech Platforms

EdTechs are rapidly expanding in the U.S. pre-K-12 market.^[9] Because their services rely on student information, EdTech platforms process data about learning information, behaviors, preferences, and performance. This data-intensive model results in privacy concerns, especially when tools are used to target minors and capture personal information, including learning disabilities, emotional responses, behavioral patterns, and biometrics.^[10] Modern analytics can generate sensitive inferences from routine student interactions, creating risks—such as privacy-invasive and discriminatory inferences that can affect students’ opportunities and reputations—that existing data-protection laws do not fully address.^[11] And regulators like the FTC have warned that AI-driven data collection tools may misuse student data without strong safeguards.^[12] The potential sensitivity of this data becomes especially important when law enforcement agencies seek access to it during investigations, requiring EdTech firms to comply with data access requests while protecting student privacy.^[13] Because schools grow more dependent on these platforms, companies also shape the practical rules governing student data.^[14]

For EdTechs, privacy compliance is both a legal duty and a business priority.^[15] They must satisfy parental trust and regulatory requirements while maintaining the data access necessary to deliver their services. This tension is a defining feature of King Innovation’s role in the educational data ecosystem.

B. King Sovereignty: Domestic Privacy Laws

Domestic privacy laws guard personal data, reflecting deep security concerns and often blocking foreign access to educational records.^[16] In the United States, this protection operates primarily through two federal statutes.^[17] FERPA protects student education records and generally requires parental or student consent for disclosure.^[18] Similarly, COPPA safeguards the privacy of children under thirteen by requiring parental consent before online services collect, use, or disclose their personal information.^[19]

Internationally, other national privacy laws have asserted even stronger control. The EU’s GDPR sets strict rules for transferring personal data abroad.^[20] GDPR holds companies accountable with detailed compliance requirements and imposes penalties for violations, up to 20 million euros or 4 percent of global annual revenue.^[21] And China’s PIPL gives authorities broad discretion to decide how information held in China may be collected and shared,^[22] requiring government approval before responding to foreign legal requests.^[23] These national privacy laws establish legal boundaries that restrict the movement of educational data, setting the stage for potential conflicts when foreign law enforcement authorities seek access.

C. King Authority: Law Enforcement Under the Cloud Act’s Extraterritorial Reach

The remaining dynamic concerns law enforcement’s efforts to access student data across borders despite national restrictions.^[24] In the United States, this authority is articulated in the CLOUD Act of 2018, which allows U.S. authorities and partner countries to demand access to digital information relevant to criminal investigations without relying on the slower Mutual Legal Assistance Treaty (MLAT) process.^[25] For EdTechs, this means that even though FERPA and COPPA generally restrict disclosure without consent, the CLOUD Act can still require service providers to disclose educational data directly to U.S. or partner-country authorities.

This statutory assertion of authority can conflict with competing claims of sovereignty—a tension this Note seeks to clarify. The CLOUD Act reflects a unilateral U.S. approach to cross-border data access.^[26] This broad authority is likely to drive countries to adopt stricter localization laws as defensive measures.^[27] The EU’s GDPR specifically regulates foreign access to personal data, and China’s PIPL sets even stricter limitations on foreign access.^[28] Therefore, the CLOUD Act represents King Authority’s claim that U.S. law enforcement may reach educational data held abroad. This conflicts with King Sovereignty’s insistence on maintaining control over information within national borders and King Innovation’s reliance on stable, cross-border data flows.

The struggle plays out in two arenas: domestically, where privacy rules limit or altogether prohibit law enforcement access to student data, and internationally, where foreign laws limit or altogether prohibit the transfer of data to foreign authorities.^[29] These are not separate battles, but two stages of the same struggle over who ultimately controls educational data.

II. The Origins of Rivalry: Differences in Domestic Legal Principles

The three Kings all compete for control over data, but their struggle is complicated—and often escalated—by different countries’ respective approaches to privacy rights and state authority, which shape the power each King can exercise. The United States follows a Fourth Amendment framework that guards against unreasonable searches but allows for numerous exceptions.^[30] The EU elevates privacy to “a fundamental right” under Article 8 of the EU Charter.^[31] And China implements a data sovereignty doctrine that prioritizes national security and state control and requires data localization to maintain government oversight.^[32] These differing foundations complicate the struggle between EdTech platforms, national regulators asserting territorial control, and law enforcement authorities seeking access across them.

A. The United States: Privacy as a Balanced Interest

Under the Fourth Amendment, the United States typically uses a balancing test that flexibly weighs individual rights against other compelling government interests.^[33] This framework is rooted in the “reasonable expectation of privacy,” a standard that ties privacy protection to society’s continually evolving judgments about what counts as reasonable.^[34]

This balancing approach is reflected in two key doctrines that shape how educational data receives protection. The Third-Party Doctrine limits constitutional protections for data shared with service providers, potentially removing student data held by EdTech platforms from Fourth Amendment coverage.^[35] Educational search exceptions further limit protections: New Jersey v. T.L.O. requires only “reasonable suspicion” for school searches^[36] and Board of Education v. Earls permits “suspicionless” drug testing for extracurricular activities.^[37] Together, these context-based doctrines create a framework where students’ data receives context-dependent, often attenuated protection.

In these ways, the U.S. approach—treating privacy as a balanceable interest—has long guided how U.S. or foreign courts assess government claims to access information. The CLOUD Act, which allows U.S. law enforcement to demand access to data stored abroad, is compatible with this tradition.^[38] By placing the executive agreement process entirely within the Department of Justice’s control “without any transparency or input from outside stakeholders or judicial oversight,” the Act prioritizes government access over individual privacy rights.^[39] This differs dramatically from the EU’s rights-based framework and China’s sovereignty-driven approach.^[40]

B. The European Union: Enshrining the Right to Respect for Privacy

Unlike the U.S. model, which treats privacy as a negotiable interest subject to a balancing test, the EU treats privacy as an inalienable human right. Article 8 of the European Convention on Human Rights establishes a broad right to respect for private life that extends across diverse contexts, including public surveillance, corporate data monitoring, and cross-border transfers.^[41] This principle forms the foundation for the GDPR.

In Uzun v. Germany, the ECtHR ruled that, while surveillance may serve legitimate security interests, it must be proportionate, necessary, and subject to judicial oversight.^[42] This contrasts with U.S. courts, which often prioritize law-enforcement interests. In this case, though, the court emphasized that privacy should be the default that deserves presumptive protections, and the burden is on the authorities to prove why that privacy should be compromised.^[43]

C. China: The State as Data Guardian

Data sovereignty is core to China’s approach to data governance. China emphasizes state control over data, regulating how platforms manage consumer information to prevent cybercrime and ensure government oversight.^[44] This model reflects China’s emphasis on public safety,^[45] and its collectivist tradition, where individual rights are shaped by, and exist within, the broader interests of society, as the state defines them.^[46] Whereas the United States provides avenues for government access through balancing tests, and the EU imposes strong privacy obligations on companies, China’s model empowers the state to both regulate private data practices and access digital information itself.

III. The Rivalry Deepens—Different Philosophies Produce Divergent Regulatory and Enforcement Regimes in EdTech

When applied to the EdTech industry, where student data crosses border frequently, these distinct legal frameworks produce divergent regulatory regimes that place EdTech companies, national regulators, and law enforcement agencies in tension, if not direct conflict.

A. The United States: FERPA and COPPA

Enacted in 1974, FERPA is among the most important U.S. laws governing access to student data. It limits the disclosure of education records containing personally identifiable information and generally requires written consent from parents or eligible students.^[47] It applies directly to federally funded schools and places requirements on EdTech companies which handle large amounts of student data through school contracts.^[48]

Due to amendments in 2008 and 2011, FERPA now permits schools, under certain circumstances, to disclose student records to law enforcement agencies, contractors, consultants, and private companies without student or parental consent.^[49] FERPA also requires compliance with judicial orders or subpoenas,^[50] and in such cases schools must make reasonable efforts to notify the student in advance.^[51] In addition, FERPA allows schools to share “directory information” unless students opt out.^[52] Records created and maintained by a school’s law enforcement unit fall outside of FERPA’s protection, meaning they may be shared with EdTech contractors without FERPA restrictions.^[53] Together, these exceptions create a disclosure regime that offers schools considerable flexibility, which can make it easier for EdTech vendors to receive and process student data through school contracts. Furthermore, FERPA’s enforcement system offers limited practical deterrence, as its main penalty—cutting federal funding—exists in theory but is rarely enforced.^[54] This weak enforcement environment leaves EdTech vendors with few consequences for noncompliance.

Enacted in 1998, COPPA extends privacy protections to children under thirteen by regulating how websites and online services, including EdTech platforms, collect, use, and share their personal data.^[55] It requires EdTech platforms to obtain verifiable parental consent before collecting data, such as names, addresses, geolocation, or persistent identifiers;^[56] to publish clear privacy policies outlining their data practices;^[57] to collect only that information which is necessary for a child’s participation in online activities;^[58] and to implement security measures that protect data from unauthorized access or disclosure.^[59] COPPA also allows for safe harbor programs enabling industry groups to create approved self-regulatory guidelines.^[60] These requirements become relevant to cross-border governance when U.S. law enforcement seeks access to children’s data held by EdTech platforms.^[61]

B. The European Union: The General Data Protection Regulation (GDPR)

The GDPR, effective since May 2018, treats privacy as “a fundamental right,” not a mere regulatory issue.^[62] This view of privacy is perhaps best represented by its rules on automated decision-making in education, which require human oversight whenever sensitive data is used or shared.^[63]

The GDPR allows free data flow only to countries with “adequate” protections.^[64] Transfers to inadequate countries like the United States, require safeguards such as standard contractual clauses (contractual commitments ensuring that transferred data receives GDPR-level protection).^[65] Thus, when the CLOUD Act compels U.S. providers to access data stored in Europe, the GDPR allows such transfers only under strict safeguards, making the two regimes uneasy to reconcile.^[66]

C. China: The Personal Information Protection Law (PIPL)

China’s PIPL, enacted in 2021, prioritizes state sovereignty. Its defining feature is strict data localization: unlike the GDPR’s conditional transfer system, China’s PIPL makes cross-border data transfers dependent on explicit government security reviews.^[67] It also explicitly prohibits organizations from providing personal data stored in China to foreign judicial or law enforcement authorities without government approval.^[68] This sovereignty-centered model differs sharply from the CLOUD Act’s access-oriented philosophy.^[69]

For EdTechs, these rules pose unique challenges from those raised by differences between the U.S. and EU. Companies serving Chinese students may be designated “critical information infrastructure operators” and face stricter oversight and mandatory local data storage.^[70]

Taken together, the U.S., EU, and Chinese models form the dominant regulatory triad driving today’s fragmented data-governance landscape. These conflicting frameworks create significant challenges for global EdTech operations.^[71]

IV. Cross-Border Jurisdictional Conflicts

Different national privacy frameworks become highly consequential once the CLOUD Act compels U.S. EdTech companies to disclose data stored overseas, only to encounter foreign privacy and data localization laws that insist the data remain at home.

A. The CLOUD Act puts EdTech Privacy Duties in Conflict with Foreign Law Enforcement

Enacted in March 2018, the CLOUD Act allows U.S. law enforcement to compel providers to disclose data stored overseas. Passed quickly as a part of an omnibus spending bill without formal hearings, it was a direct response to Microsoft Ireland.^[72] Under this framework, approved foreign governments can bypass the slower MLAT process by requesting data directly from U.S. companies through executive agreements, subject to Department of Justice approval and limited judicial oversight.^[73] The U.S.–UK Data Access Agreement of 2019, which allows U.S. service providers to respond directly to legal requests from UK authorities, is an example of this new model.^[74]

While the Act speeds up investigations, it creates a practical tension for EdTechs trying to balance FERPA’s and COPPA’s respective parental consent requirements with law enforcement demands.^[75] FERPA and COPPA reflect a commitment to transparency via consent, but the CLOUD Act authorizes law enforcement to compel data from service providers pursuant to the requesting country’s legal process.^[76] Whether notice is provided to account holders depends on the requesting country’s law.^[77] This means that if the requesting country’s law permits orders without parental consent, schools and families may not learn of data disclosure, bypassing consent and notice requirements.^[78]

As such, EdTechs are trapped between multiple duties. On one hand, they risk CLOUD Act non–compliance. But if they comply with the CLOUD Act, schools and parents may abandon them for platforms that refuse to compromise student privacy.

B. The U.S. CLOUD Act Creates Practical Tensions with the EU’s GDPR and China’s PIPL in Cross-Border-Data Access.

The CLOUD Act’s mechanism of direct provider access creates procedural misalignment with foreign privacy laws. Although the Act permits bilateral executive agreements, these deals often favor U.S. interests, reflecting the United States’ technological dominance and bargaining leverage.^[79] The Act’s core mandate that U.S. providers disclose data within their “possession, custody, or control,” even when stored abroad, creates potential conflicts with foreign data-security and localization laws.^[80]

GDPR overall requires that data transfers meet strict standards of necessity, proportionality, and adequacy.^[81] It restricts cross-border transfers to adequacy decisions, safeguards, or narrow derogations.^[82] It also obliges companies to build privacy protections into their systems, get parental consent for users under sixteen, and conduct Data Protection Impact Assessments.^[83] These steps aim to protect minors from the higher risks of profiling, automated decisions, and large-scale data usage in settings such as IoT-based smart homes.^[84] The same principles apply to EdTech platforms, which also rely heavily on data collection.^[85] Tensions may grow in cross-border cases, where platforms face foreign law enforcement demands, likely without consideration of GDPR’s necessity and proportionality requirements.

In 2020, the Court of Justice of the European Union (CJEU) held that U.S. data protection measures do not provide an adequate level of protection for EU personal data.^[86] The court identified three fundamental deficiencies. First, U.S. surveillance laws—particularly FISA Section 702 and Executive Order 12333— fail to satisfy necessity and proportionality because they permit government access to personal data in excess of that permitted by the GDPR.^[87] Second, these laws lack clear and precise rules about when and how government authorities may access data, failing to ensure that surveillance is limited to what is strictly necessary and proportionate.^[88] Third, EU citizens lack an effective judicial remedy to challenge U.S. government surveillance, undermining the fundamental right to an effective remedy.^[89] Because U.S. law fails on these counts, the CJEU concluded that the adequacy requirement isn’t met.^[90]

Under the CLOUD Act, U.S. law enforcement may compel providers to disclose data stored in the EU.^[91] Such compelled access can constitute an EU-U.S. transfer under the GDPR, a transfer for which it remains unclear whether the GDPR’s specific safeguards can be adequately satisfied.^[92] If the transfer is ultimately found non-compliant, the company could face severe GDPR penalties.^[93]

The GDPR offers strong safeguards against foreign access to personal data. However, it protects only personal data, not corporate information.^[94] This gap highlights Europe’s broader challenge in building sovereign cloud infrastructure and legal safeguards—a vulnerability shared globally.^[95] This challenge has driven a global shift toward data sovereignty, with countries tightening cross-border controls, as seen in GDPR-inspired GCC rules and other new privacy laws.^[96]

Meanwhile, differences between the U.S.’ CLOUD Act (U.S.) and China’s PIPL create an even sharper tension. Most notably, PIPL keeps sensitive data in China without government approval.^[97] This absolute commitment to data localization directly challenges the CLOUD Act’s core objectives.

PIPL reflects China’s broader data sovereignty doctrine, which prioritizes national security over digital information.^[98] Article 38 enumerates four legal bases for transferring personal data outside China, including passing a security assessment conducted by state cyberspace authorities.^[99] Yet, PIPL’s cross-border requirements remain “still quite ambiguous,” unlike the GDPR’s clearer transfer rules, generating additional uncertainty for multinational companies.^[100] Notably, this regulatory framework does not exist in isolation, operating instead alongside China’s other data protection laws, including the Data Security Law (数据安全法) of 2021 and the Cybersecurity Law of 2017 (⽹络安全法).^[101]

EdTech platforms that comply with U.S. demands risk severe penalties in China, including fines of up to 50 million RMB or five percent of annual revenue under the PIPL.^[102] This standoff emphasizes the growing fragmentation of global data rules, the impact of which is significant. Apple’s $1 billion investment in a Guizhou data center, for example, shows the lengths companies must go to meet China’s data localization rules.^[103]

V. Proposals for Harmonization

EdTechs sit at the crossroads of national regulation and law-enforcement access, facing conflicts that their cross-border operations cannot easily resolve.^[104] Addressing this fragmentation requires interoperability that enables cross-border cooperation while still respecting core privacy principles. This section outlines three pathways: adopting international guidelines that provide shared standards, expanding safe-harbor certification frameworks that allow conditional compliance across legal systems, and developing industry best practices to support responsible data governance.

A. International Guidelines and Standards

International guidelines developed by the Organization for Economic Cooperation and Development (OECD) and the United Nations Educational, Scientific and Cultural Organization (UNESCO) might offer a foundation for shared standards.

The OECD Privacy Guidelines, first adopted in 1980 and updated in 2013, established a global baseline that influenced later frameworks such as the EU’s GDPR and Japan’s Act on the Protection of Personal Information.^[105] These guidelines set out eight core principles: collection limitation, data quality, purpose specification, use limitation, security safeguards, openness, individual participation, and accountability.^[106]

UNESCO’s Recommendation on the Ethics of Artificial Intelligence, adopted in 2021, addresses ethical concerns arising from the use of AI systems in educational settings.^[107] It requires that AI systems used in learning environments must be subject to strict requirements, particularly when monitoring or predicting student behavior, and that any data collected must not be misused or commercially exploited.^[108]

These approaches have made an impact in multinational efforts to regulate data governance and ethical AI; the APEC Cross-Border Privacy Rules system, which operationalizes OECD principles, has facilitated data flows among diverse nations, including the United States, Japan, Singapore, and the Philippines.^[109] This demonstrates that international guidelines can evolve into functional interoperability mechanisms when backed by regional cooperation.

Different data laws may create compliance dilemmas that companies cannot resolve through perfect adherence to all frameworks.^[110] International guidelines may influence, but cannot forcefully alter, domestic laws. However, OECD and UNESCO standards offer reliable common ground because they have already helped shape major regulatory frameworks.^[111] As a result, they point to a core set of requirements on which broad consensus is likely to exist across jurisdictions.

Yet reliance on transnational standards faces an implementation challenge: because they are not binding law, they cannot compel participation or compliance.^[112] Indeed, China’s participation in global data-governance systems remains limited, even as its observer status at the OECD suggests potential for future engagement.^[113]

B. Safe Harbor and Cross-Border Frameworks

Likewise, expanding safe-harbor provisions can offer a practical compliance pathway for EdTechs. COPPA already provides a model through its Safe Harbor program, which allows industry groups to create FTC-approved self-regulatory frameworks.^[114] A broader system, such as an expanded safe-harbor certification, could similarly streamline cross-border compliance by giving EdTechs government-endorsed standards to follow. By meeting such standards, companies could reduce legal risk and provide clearer assurance of responsible data practices across borders.^[115] And a multi-jurisdictional safe-harbor program could go further by setting shared benchmarks that help bridge differences across national privacy laws and that reflect core privacy principles.^[116]

A modernized approach could build on the lessons of the EU–U.S. Data Privacy Framework (DPF), adopted in 2023, three years after the CJEU struck down the Privacy Shield in its July 2020 Schrems II decision.^[117] The court had previously invalidated the Safe Harbor Framework in its 2015 Schrems I ruling.^[118] Both agreements were struck down due to concerns over U.S. surveillance and the lack of legal remedies for EU citizens.^[119] While the Safe Harbor Framework and the Privacy Shield were early attempts at bridging regulatory differences, the successive invalidation highlights the challenges of achieving mutual recognition in cross-border data protection.^[120]

In response, the DPF introduced stricter limits on U.S. government data access, requiring that intelligence requests be necessary and proportionate.^[121] This aligns more closely with EU privacy standards.^[122] The DPF also created an independent redress mechanism for EU individuals, including the establishment of a Data Protection Review Court (DPRC) within the U.S. Department of Justice.^[123] This court operates independently from intelligence agencies and can investigate complaints and order corrective actions.^[124] Similar frameworks might be extended to other jurisdictions or tailored to high-risk sectors like EdTech.

In addition, the Global Cross-Border Privacy Rules Forum, launched in 2022, represents initial progress toward interoperable privacy certification across jurisdictions, though it doesn’t yet address the critical issue of law-enforcement access.^[125] Similarly strong multi-jurisdictional certification frameworks might help create supervised channels for sharing data between authorities by offering a common baseline of vetted privacy safeguards.

C. Industry Self-Regulation and Best Practices

Industry-led self-regulation could offer a complementary approach to government regulation. By adopting international standards such as those developed by the International Organization for Standardization (ISO) and other recognized frameworks, EdTechs can establish consistent baseline privacy practices across jurisdictions.^[126] Voluntary measures such as transparency reports, now common among major tech firms like Google, can enhance credibility.^[127] Notably, some EdTech providers have begun adopting these practices. Coursera, for example, maintains an ISO/IEC 27001:2013–certified information security management system and undergoes annual third-party independent audits, including SOC 2 Type II assessments and third-party penetration testing.^[128]

EdTech platforms are not neutral tools but complex ecosystems that mediate social, technical, and political relationships.^[129] Because they encode particular values and priorities, they actively shape students’ educational experiences. Voluntary compliance mechanisms can help reduce the risks of biased algorithms and data misuse by promoting standards for fairness, data minimization, and transparency.^[130] The Student Privacy Pledge, launched in 2014 and signed by nearly 400 companies before its retirement in April 2025, represents an early attempt at such collective self-regulation.^[131]

Implementing meaningful self-regulation requires a multistakeholder approach as a “coalitional project.”^[132] Because no single actor fully understands or controls how educational data systems operate, effective governance depends on incorporating the perspectives of educators, students, administrators, policymakers, and civil society organizations. Self-regulatory efforts benefit from drawing on these diverse viewpoints to identify risks and define practical safeguards. In fact, the Future of Privacy Forum’s work on student data consolidated diverse viewpoints to develop practical guidelines that balance innovation with privacy protection.^[133]

Self-regulation can complement international guidelines and safe-harbor certification frameworks. For EdTechs subject to U.S., EU, and Chinese laws, strong self-regulatory practices signal responsible data handling and good-faith compliance while as broader legal harmonization efforts remain uncertain.

Conclusion

The cross-border conflict over educational data reflects competing approaches to data governance in today’s digital world. The three kings—EdTech companies, national regulators, and law enforcement agencies—all pursue valid goals: companies seek sustainable business operation, regulators aim to protect citizens’ rights, and law enforcement wants to investigate crimes. Nevertheless, as this Note has attempted to show, their respective approaches may create legal tensions that leave platforms caught in compliance uncertainties.

Digital governance requires cooperation, flexibility, and shared infrastructure. International guidelines can provide common languages to bridge legal differences; safe-harbor certification frameworks offer lawful, transparent paths for cross-border data sharing; and strong industry standards can create consistent recognized standards beyond legal requirements.

Getting there will take compromise. EdTech companies can adopt privacy-by-design principles and greater transparency; regulators should respect diverse regulatory systems; and law enforcement agencies should accommodate oversight and operate within frameworks that include clear safeguards. As AI technologies increasingly mediate education, the stakes for student privacy and cross-border data governance will grow higher. This isn’t a choice between privacy and innovation, but a challenge to design systems that uphold rights while enabling the responsible use of data to support modern learning. The curtain isn’t falling on this debate—it’s rising on the next act of global data governance.

1. * I am a J.D. candidate at UCLA School of Law and hold a Ph.D. in Philosophy and Education from Teachers College, Columbia University. I own tremendous thanks to Professor Maximo Langer, and to my editors at the Washington University Law Review—Geremia, Amanda, and Emma—whose support were essential to this piece.

   In Microsoft Corp. v. United States, the Second Circuit held that the Stored Communications Act did not authorize U.S. law enforcement to compel Microsoft to produce emails stored on a server in Ireland. 829 F.3d 197, 222 (2d Cir. 2016). The Supreme Court later deemed the case moot after Congress enacted the CLOUD Act, which amended the Stored Communications Act to require providers to disclose data within their possession, custody, or control, regardless of the data’s physical location. United States v. Microsoft Corp., 584 U.S. ___ (2018). Microsoft ultimately complied with a new warrant issued under the revised law. Id. ↑

2. Microsoft Corp., 829 F.3d at 201. ↑

3. Paul De Hert & Johannes Thumfart, The Microsoft Ireland Case, the CLOUD Act and the Cyberspace Sovereignty Trilemma: Post-Territorial Technologies and Companies Question Regulatory State Monopolies, 21 Jusletter IT 373, 374, 386–87 (2020). ↑

4. Id. at 375. ↑

5. David P. Grosso, Michelle R. Bowling, Starshine S. Chun & Brooke M. Delaney, The Development of AI and Protecting Student Data Privacy, ArentFox Schiff LLP (Feb. 22, 2024), https://www.afslaw.com/perspectives/ai-law-blog/the-development-ai-and-protecting-student-data-privacy [https://perma.cc/9SJP-QEUY]. ↑

6. Id. ↑

7. See Ben Williamson, Digital Education Governance: Data Visualization, Predictive Analytics, and ‘Real-Time’ Policy Instruments, 31 J. Educ. Pol’y 123, 127–29 (2016). ↑

8. See infra, Section III. ↑

9. See, e.g., Natasha Singer, Microsoft and Other Firms Pledge to Protect Student Data, N.Y. Times (Oct. 7, 2014), http://www.nytimes.com/2014/10/07/business/microsoft-and-other-firms-pledge-to-protect-student-data.html [https://perma.cc/47AN-5KDN]; see also U.S. Education Technology Market Size & Outlook, 2025-2030, Grand View Horizon Databooks, https://www.grandviewresearch.com/horizon/outlook/education-technology-market/united-states (last visited Nov. 23, 2025) (“The education technology market in the United States is expected to reach a projected revenue of US$ 90,606.5 million by 2030. A compound annual growth rate of 11.1% is expected of the United States education technology market from 2025 to 2030.”). ↑

10. See, e.g., Benjamin Herold, Schools Are Deploying Massive Digital Surveillance Systems. The Results Are Alarming, Educ. Wk. (May 30, 2019), https://www.edweek.org/technology/schools-are-deploying-massive-digital-surveillance-systems-the-results-are-alarming/2019/05 [https://perma.cc/8KJ4-UD8J] (discussing how platforms like Social Sentinel, Securly, and Gaggle monitor students’ digital content, social media, and even emotional cues); Jason Kelley, Students Are Pushing Back Against Proctoring Surveillance Apps, Elec. Frontier Found. (Sept. 25, 2020), https://www.eff.org/deeplinks/2020/09/students-are-pushing-back-against-proctoring-surveillance-apps [https://perma.cc/TT7Y-UFVW] (discussing student petitions against Honorlock, Proctorio, Respondus, and ProctorU for invasive uses of facial recognition, voice tracking, and other biometric monitoring). ↑

11. Sandra Wachter & Brent Mittelstadt, A Right to Reasonable Inferences: Re-Thinking Data Protection Law in the Age of Big Data and AI, 2019(2) Colum. Bus. L. Rev. 494, 505–12 (2018). ↑

12. Jody Godoy, FTC’s Holyoak Concerned AI Collecting Children’s Data, Reuters (Nov. 15, 2024), https://www.reuters.com/technology/artificial-intelligence/ftcs-holyoak-concerned-ai-collecting-childrens-data-2024-11-14/ [https://perma.cc/39QN-38TZ]. ↑

13. See infra, Section I.C (discussing how EdTechs may have to navigate law enforcement demands while protecting student privacy). ↑

14. T. Philip Nichols & Ezekiel Dixon-Román, Platform Governance and Education Policy: Power and Politics in Emerging Edtech Ecologies, 46 Educ. Eval. & Pol’y Analysis 309, 310, 312 (2024). ↑

15. See infra, Section I.C; see also infra, Section III.A (detailing FERPA and COPPA compliance requirements). ↑

16. See Paul M. Schwartz, Information Privacy in the Cloud, 161 U. Pa. L. Rev. 1623, 1640–42 (2013) (noting that national privacy laws can restrict or block cross-border access to personal data). ↑

17. See infra, Section III.A. ↑

18. See generally 20 U.S.C. § 1232g (1974); 34 C.F.R. § 99.30 (2024). ↑

19. 16 C.F.R. § 312.5 (2025). ↑

20. See infra, Section III.B. ↑

21. Regulation 2016/679 of the European Parliament and of the Council of 27 April 2016 on the Protection of Natural Persons with Regard to the Processing of Personal Data and on the Free Movement of Such Data, and Repealing Directive 95/46/EC (General Data Protection Regulation), art. 44–50, 83, 2016 O.J. (L 119) 1 (EU) [hereinafter GDPR]. ↑

22. See infra, Section III.C; see Rogier Creemers, China’s Emerging Data Protection Framework, 8 J. Cybersecurity 1, 1 (2022); see also Igor Calzada, Citizens’ Data Privacy in China: The State of the Art of the Personal Information Protection Law (PIPL), 5 Smart Cities 1129, 1130 (2022). ↑

23. Zhonghua Renmin Gongheguo Geren Xinxi Baohu Fa (中华人民共和国个人信息保护法) [Personal Information Protection Law of the People’s Republic of China] (promulgated by the Standing Comm. Nat’l People’s Cong., Aug. 20, 2021, effective Nov. 1, 2021), chs. III–IV, 2021 Standing Comm. Nat’l People’s Cong. Gaz. 1117 (China). ↑

24. While the European Union and China have developed their own cross-border access frameworks, this Note focuses on the CLOUD Act’s extraterritorial reach and its conflict with national privacy laws governing educational data. ↑

25. The MLAT process requires formal, government-to-government requests, with the requested state reviewing and executing the request under its own domestic law and procedural requirements—often involving prosecutorial and judicial oversight—a multi-layered process that typically takes months or longer. See U.S. Dept.artment of Just.ice, Promoting Public Safety, Privacy, and the Rule of Law Around the World: The Purpose and Impact of the CLOUD Act 3–4 (Apr. 2019), https://www.justice.gov/criminal/media/999601/dl?inline [https://perma.cc/PG3M-RJEP]. ↑

26. Secil Bilgic, Something Old, Something New, and Something Moot: The Privacy Crisis Under the CLOUD Act, 32 Harv. J.L. & Tech. 321, 335 (2018). ↑

27. Id. ↑

28. See infra, Sections III.B and III.C. ↑

29. See infra, Sections III (discussing domestic privacy framework), IV (analyzing international conflicts over cross-border data access). ↑

30. See Katz v. United States, 389 U.S. 347, 351 (1967) (“[T]he Fourth Amendment protects people, not places.”); United States v. Leon, 468 U.S. 897, 925 (1984) (establishing the good-faith exception to the exclusionary rule). ↑

31. See Charter of Fundamental Rights of the European Union, 2012 O.J. (C 326 art. 8) 391 (“Everyone has the right to the protection of personal data concerning him or her.”); see also Digital Rights Ireland Ltd. v. Minister for Commc’ns, Joined Cases C-293/12 & C-594/12, ECLI:EU:C:2014:238, ¶ 54 (Apr. 8, 2014) (invalidating the EU Data Retention Directive (Directive 2006/24/EC) , which required telecom and internet service providers to retain traffic and location data for law enforcement purposes, for disproportionate surveillance). ↑

32. See Cybersecurity Law of the People’s Republic of China (中华⼈⺠共和国⽹络安全法). (promulgated by the Standing Comm. Nat’l People’s Cong., Nov. 7, 2016, effective June 1, 2017), art. 37, CLI.1.286536(EN) (requiring critical information infrastructure operators to store personal and important data within China); see also Data Security Law of the People’s Republic of China (中华⼈⺠共和国数据安全法) (promulgated by the Standing Comm. Nat’l People’s Cong., June 10, 2021, effective Sept. 1, 2021), art. 2, CLI.1.483071(EN) (establishing China’s data sovereignty framework). ↑

33. U.S. Const. amend. IV. ↑

34. Katz v. United States, 389 U.S. 347, 360 (1967) (Harlan, J., concurring). ↑

35. See Smith v. Maryland, 442 U.S. 735, 743 (1979) (holding that individuals have no reasonable expectation of privacy in information voluntarily conveyed to third parties, thus placing such data outside Fourth Amendment protection); United States v. Miller, 425 U.S. 435 (1976) (holding that a bank depositor has no reasonable expectation of privacy in checks and deposit slips since they are business records of the bank and information voluntarily shared with third parties is outside Fourth Amendment protection). ↑

36. New Jersey v. T.L.O., 469 U.S. 325, 345 (1985) (holding that school officials may search students’ belongings based on reasonable suspicion rather than probable cause) ↑

37. Bd. of Educ. Of Indep. Sch. Dist. No. 92 of Pottawatomie Cnty. v. Earls, 536 U.S. 822 (2002). ↑

38. Supra note 26; see also infra, Section IV.A (discussing the CLOUD Act’s extraterritorial reach). ↑

39. Miranda Rutherford, The CLOUD Act: Creating Executive Branch Monopoly over Cross-Border Data Access, 34 Berkeley Tech. L.J. 1177, 1202 (2019). ↑

40. See infra, Sections II.B, C. ↑

41. See Charter of Fundamental Rights of the European Union, supra note 31. ↑

42. App. No. 35623/05, Eur. Ct. H.R. § 78 (2010). ↑

43. Id. at ¶ 80 (“Against this background, the interference by the applicant’s additional surveillance via GPS thus necessitated more compelling reasons if it was to be justified.”). ↑

44. See, e.g., Cyberspace Admin. of China, Administrative Penalty Decision on Didi Global Inc. (July 21, 2022), translated in Todd Liao, Cyberspace Administration of China Issues Statement on Didi’s $1.2B Fines for Cybersecurity Law Violations Morgan Lewis (July 28, 2022), https://www.morganlewis.com/pubs/2022/07/cyberspace-administration-of-china-issues-statement-on-didis-1-2b-fines-for-cybersecurity-law-violations [https://perma.cc/T477-PPEG] (imposing 8.026 billion yuan [$1.2 billion] fine for violations of the Cybersecurity Law, Data Security Law, and PIPL, citing risks to “the nation’s crucial information infrastructure and data security”).  ↑

45. Liming Liu & Yiming Chen, A Triple-Layered Comparative Approach to Understanding New Privacy Policy Practices of Digital Platforms and Users in China After Implementation of the PIPL, Soc. Media + Soc’y, Oct.-Dec. 2024, at 1. ↑

46. See Id.; Calzada, supra note 22, at 1130. ↑

47. See generally 20 U.S.C. § 1232g (1974); 34 C.F.R. § 99.30 (2024). ↑

48. 34 C.F.R. § 99.31(a)(1)(i)(B). ↑

49. Elana Zeide, Student Privacy Principles for the Age of Big Data: Moving Beyond FERPA and FIPPs, 8 Drexel L. Rev. 339, 360, 371 (2016). ↑

50. § 99.31(a)(9) (allowing disclosure). ↑

51. § 99.31(a)(9)(ii) (requiring “a reasonable effort to notify the parent or eligible student” in advance of compliance). ↑

52. Compare 34 C.F.R. § 99.37 (permitting disclosure of directory information without consent) with GDPR, supra note 21, art. 4(1) (defining personal data to include identifiers such as names and addresses), 6(1) (requiring one of six legal bases for processing, including consent under 6(1)(a)), 7 (setting conditions for valid consent). While GDPR permits processing under legitimate conditions or other bases, educational institutions sharing directory information internationally would typically require consent absent a specific legal obligation or public interest justification. ↑

53. 34 C.F.R. § 99.8. ↑

54. Dylan Peterson, EdTech and Student Privacy: California Law as a Model, 31 Berkeley Tech. L.J. 961, 979–80 (2016). ↑

55. 15 U.S.C. §§ 6501–6506. ↑

56. 16 C.F.R. § 312.5 (2025); see also Joel R. Reidenberg, N. Cameron Russell, Jordan Kovnot, Thomas B. Norton, Ryan Cloutier et al., Privacy and Cloud Computing in Public Schools 9–10 (2013). ↑

57. 16 C.F.R. § 312.4. ↑

58. Id. §§ 312.7, 312.10. ↑

59. Id. §§ 312.3, 312.8. ↑

60. Id. § 312.11. ↑

61. See infra, Section IV.A ↑

62. GDPR, supra note 21, recital 1 (“The protection of natural persons in relation to the processing of personal data is a fundamental right.”).  ↑

63. Id. arts. 15–18, 22. ↑

64. Id. art. 45. ↑

65. Id. arts. 44–46. In 46(2)(c)-(d), The GDPR refers to these as “standard data protection clauses,” though they are often known as standard contractual clauses. ↑

66. The EU Law Enforcement Directive explicitly requires strong safeguards when sharing data with non-EU countries, such as binding agreements or narrow public interest exceptions. Directive (EU) 2016/680 of the European Parliament and of the Council of 27 April 2016 on the Protection of Natural Persons with Regard to the Processing of Personal Data by Competent Authorities for the Purposes of the Prevention, Investigation, Detection or Prosecution of Criminal Offences or the Execution of Criminal Penalties, and on the Free Movement of Such Data, and Repealing Council Framework Decision 2008/977/JHA, arts. 35, 38, 2016 O.J. (L 119) 89. ↑

67. PIPL, art. 40 (“Critical information infrastructure operators and the personal information processors that process personal information up to the amount prescribed by the national cyberspace department shall store domestically the personal information collected and generated within the territory of the People’s Republic of China.”); see also Zhonghua Renmin Gongheguo Wangluo Anquan Fa (中华⼈⺠共和国⽹络安全法) [Cybersecurity Law of the People’s Republic of China] (promulgated by the Standing Comm. Nat’l People’s Cong., Nov. 7, 2016, effective June 1, 2017), art. 37, 2016 Standing Comm. Nat’l People’s Cong. Gaz. [FIRST PAGE] (China) (“Personal information and important business data collected and produced by critical information infrastructure operators during their activities within the territory of the People’s Republic of China, shall be stored within the territory . . . .”). ↑

68. Personal Information Protection Law of the People’s Republic of China (中华人民共和国个人信息保护法) [PIPL] (promulgated by Standing Comm. Nat’l People’s Cong. Aug. 20, 2021, effective Nov. 1, 2021), art. 41. ↑

69. See infra, Section IV.B. ↑

70. See PRC’s Personal Information Protection Law, supra note 68, at art. 40. ↑

71. See infra, Section IV.B. ↑

72. Clarifying Lawful Overseas Use of Data (CLOUD) Act, 18 U.S.C. §§ 2511, 2520, 2523, 2713. ↑

73. Rutherford, supra note 39, at 1184–89. ↑

74. Office of Public Affairs, U.S. and UK Sign Landmark Cross-Border Data Access Agreement to Combat Criminals and Terrorists Online U.S. Department of Justice (Oct. 3, 2019), https://www.justice.gov/archives/opa/pr/us-and-uk-sign-landmark-cross-border-data-access-agreement-combat-criminals-and-terrorists [https://perma.cc/7DCP-NMBB]. ↑

75. Id. ↑

76. See U.S. Dep’t of Justice, CLOUD Act Frequently Asked Questions, at 16, https://www.justice.gov/criminal/media/999616/dl?inline (“CLOUD Act agreements do not create any obligations or restrictions on providers; they simply remove legal restrictions that would otherwise conflict with compliance with covered orders. Providers issued orders covered by a CLOUD Act agreement are subject to the domestic requirements of the issuing country, and the issuing country’s law governs whether or how notice to an account holder by the provider may be prohibited.”). ↑

77. Id. ↑

78. In U.S. practice: FERPA traditionally channels law-enforcement requests through school districts, which provides schools and families an opportunity for notice. COPPA likewise assumes parental involvement. However, CLOUD Act authority allows law enforcement to request data directly from EdTech companies, bypassing schools and potentially families. If the order includes a nondisclosure order, companies cannot notify parents. This means U.S. law enforcement can access U.S.-stored student data—even from students who are U.S. citizens with parents in the U.S.—without the transparency safeguards that FERPA and COPPA require. ↑

79. De Hert & Thumfart, supra note 3, at 376 (“Whilst the bilateral nature of the CLOUD Act is problematic because it infinitely increases the bargaining power of the US, where most tech firms are situated, the included provisions can be regarded as an important step towards the development of a post-territorial legal framework for the obtaining of e-evidence.”). ↑

80. Stephen P. Mulligan, Cong. Rsch. Serv., Cross-Border Data Sharing Under the CLOUD Act, 8 (2018). ↑

81. Case C-311/18, Data Prot. Comm’r v. Facebook Ir. Ltd., 2020 ECLI:EU:C:2020:559 (July 16, 2020). ↑

82. GDPR allows transfers of personal data to third countries through three mechanisms. Adequacy decision under Article 45 allows transfers to countries the European Commission has determined provide protection essentially equivalent to the EU standards such as respect for fundamental rights, limits on public-authority access, and availability of independent oversight and judicial redress. The Commission has so far recognised the United States (commercial organisations participating in the EU-U.S. Data Privacy Framework) as providing an adequate level of protection. Appropriate safeguards under Article 46 permit transfers absent an adequacy decision where controllers implement legally binding mechanisms such as Standard Contractual Clauses. Finally, Article 49 derogations permit transfers in narrow and exceptional circumstances, including explicit consent, important public interest, or the establishment of legal claims. See 2016 O.J. (L 697) arts. 45 (adequacy decisions), 46 (appropriate safeguards), & 49 (derogations); see also Eur. Comm’n, Adequacy Decisions (last visited Dec. 18, 2025), https://commission.europa.eu/law/law-topic/data-protection/international-dimension-data-protection/adequacy-decisions_en. ↑

83. See GDPR, supra note 21, arts. 25, 8, and 35. ↑

84. Stavroula Rizou, Eugenia Alexandropoulou-Egyptiadou, Yutaka Ishibashi, & Kostas E. Psannis, Preserving Minors’ Data Protection in IoT-Based Smart Homes According to GDPR Considering Cross-Border Issues, 17 J. Commc’ns 180, 181-83 (2022). ↑

85. See supra, Section I.A. ↑

86. Supra note 81. ↑

87. Id. at ¶ 166–67. ↑

88. Id. at ¶ 176. ↑

89. Id. at ¶ 197. ↑

90. Id. See also Rizou et al., supra note 84, at 180. ↑

91. Supra note 80. ↑

92. See GDPR, supra note 21; see also supra note 81. ↑

93. Supra note 21. ↑

94. Emmanuelle Mignon, The CLOUD Act: Unveiling European Powerlessness, 1 La Revue Européenne Du Droit 108, 120–25 (2020). ↑

95. Id. ↑

96. Hamad Hamed Alhababi, Cross-Border Data Transfer Between the GCC Data Protection Laws and the GDPR, 13 Glob. J. Of Compar. L. 178 (2024). ↑

97. PIPL, supra note 68. ↑

98. Creemers, supra note 22, at 1. ↑

99. Gulbakyt Bolatbekkyzy, Comparative Insights from the EU’s GDPR and China’s PIPL for Advancing Personal Data Protection Legislation, 11 Groningen J. Int’l L. 129, 138 (2024). ↑

100. Id. at 144. ↑

101. See supra note 32. ↑

102. Bolatbekkyzy, supra note 99, at 139. ↑

103. Apple partnered with Guizhou-Cloud Big Data to construct a data center in China’s Guizhou Province, aligning with China’s Cybersecurity Law, which took effect on June 1, 2017 and requires foreign companies to store Chinese users’ data on servers located within China. See Paul Mozur, Daisuke Wakabayashi & Nick Wingfield, Apple Opening Data Center in China to Comply with Cybersecurity Law, N.Y. Times (July 12, 2017), https://www.nytimes.com/2017/07/12/business/apple-china-data-center-cybersecurity.html [https://perma.cc/8GSW-TJUY]. ↑

104. See supra, Section IV. ↑

105. Organisation for Economic Co-Operation and Development, Recommendation of the Council Concerning Guidelines Governing the Protection of Privacy and Transborder Flows of Personal Data 7 (2013), https://legalinstruments.oecd.org/public/doc/114/114.en.pdf. ↑

106. Id. ↑

107. United Nations Educational, Scientific, and Cultural Organization, Recommendation on the Ethics of Artificial Intelligence (2022), https://unesdoc.unesco.org/ark:/48223/pf0000381137. ↑

108. Id. at 34 (“AI systems used in learning should be subject to strict requirements when it comes to the monitoring, assessment of abilities, or prediction of the learners’ behaviours. AI should support the learning process without reducing cognitive abilities and without extracting sensitive information, in compliance with relevant personal data protection standards. The data handed over to acquire knowledge collected during the learner’s interactions with the AI system must not be subject to misuse, misappropriation or criminal exploitation, including for commercial purposes.”). ↑

109. Global Cross-Border Privacy Rules Declaration, U.S. Dep’t of Com., http://www.commerce.gov/global-cross-border-privacy-rules-declaration [https://perma.cc/3D9H-LDHH]. See also Asia-Pac. Econ. Coop., Benefits of the APEC Cross-Border Privacy Rules 3 (2019). ↑

110. See supra, Section IV. ↑

111. See supra notes 105, 107. ↑

112. See, e.g., supra note 105 (“Recommendations are adopted by Council and are not legally binding. They represent a political commitment to the principles they contain and entail an expectation that Adherents will do their best to implement them.”). ↑

113. China (People’s Republic of) and the OECD, OECD, https://www.oecd.org/en/countries/china-people-s-republic-of.html [https://perma.cc/4M8N-94AY]. ↑

114. See 16 C.F.R. § 312.11 (2023). ↑

115. See supra, Section IV (discussing potential cross-border compliance conflicts that expanded safe-harbor certification could help address). ↑

116. See id; see also supra note 105. ↑

117. Questions & Answers: EU–U.S. Data Privacy Framework, EUR. COM., (July 10, 2023), https://ec.europa.eu/commission/presscorner/detail/en/qanda_23_3752 [https://perma.cc/FR53-FLLA]. See also supra note 81. ↑

118. Shara Monteleone & Laura Puccio, The CJEU’s Schrems Ruling on the Safe Harbour Decision, European Parliament Think Tank (2015), https://www.europarl.europa.eu/thinktank/en/document/EPRS_ATA(2015)569050. ↑

119. See id.; see also Case C-362/14, Schrems v. Data Prot. Comm’r, 2015 ECLI:EU:C:2015:650, ¶¶ 94-95 (Oct. 6, 2015); Case C-311/18, Data Prot. Comm’r v. Facebook Ir. Ltd, supra note 81. ↑

120. Id. ↑

121. See supra note 117. ↑

122. Id. ↑

123. Id. ↑

124. Id. ↑

125. Global Cross-Border Privacy Rules Declaration, U.S. DEP’T OF COM., https://www.commerce.gov/global-cross-border-privacy-rules-declaration [https://perma.cc/3D9H-LDHH]; see also https://www.globalcbpr.org/wp-content/uploads/Global-CBPR-Declaration-2022.pdf. ↑

126. See, e.g., ISO/IEC 27701:2019, Security Techniques — Extension to ISO/IEC 27001 and ISO/IEC 27002 for Privacy Information Management — Requirements and Guidelines, INT’L STANDARD (Aug. 2019), https://www.iso.org/standard/71670.html. ↑

127. Transparency Report: Government Requests to Remove Content, GOOGLE, https://transparencyreport.google.com/government-removals/overview [https://perma.cc/63M5-SFZG]. ↑

128. Coursera, Inc., Annual Report Form 10-K 59 (Dec. 31, 2024), https://www.annualreports.com/HostedData/AnnualReports/PDF/NYSE_COUR_2024.pdf. ↑

129. Nichols & Dixon-Román, supra note 14, at 322. ↑

130. See supra note 128; see also infra notes 131, 133. ↑

131. Future of Privacy Forum & Software & Information Industry Association, About the Pledge, Student Privacy Pledge (Apr. 25, 2025) https://studentprivacypledge.org/ [https://perma.cc/ZBE9-W4LR]. ↑

132. Supra note 129, at 323. ↑

133. Future of Privacy Forum, Future of Privacy Forum Releases Policymaker’s Guide to Student Data Privacy, FUTURE OF PRIV. F. (Apr. 5, 2019), https://fpf.org/press-releases/future-of-privacy-forum-releases-policymakers-guide-to-student-data-privacy [https://perma.cc/LM29-N3R4]. ↑

Introduction

Three relatively recent appellate opinions—McCoy v. ATF,^[2] National Rifle Ass’n v. Bondi,^[3] and Rocky Mountain Gun Owners v. Polis^[4]—upheld modern restrictions on 18-to-20-year-old adults’ ability to purchase firearms. Each does so by choosing the wrong level of generality in its historical analysis. Rather than ask whether there is a tradition of restricting young adults’ arms rights, McCoy^[5] and Bondi^[6] reach for broad, non-firearm-specific infancy rules from contract law; Polis reframes the question as a generic regulation of commercial sales. These choices depart from New York State Rifle & Pistol Ass’n v. Bruen’s^[7] analogical method, which instructs courts to use closely matching analogues where available and to abstract up only when necessary.^[8] When the historical record includes firearm-specific sources about 18-to-20-year-olds’ rights and duties, courts must use those sources rather than attenuated contract law rules designed to protect minors from unscrupulous adults or a rebuttable safe harbor for sales regulations.

I. Bruen’s Analogical Method and the Level-of-Generality Problem

This Part sets forth Bruen’s analogical method, and the level-of-generality constraint it imposes, on courts adjudicating Second Amendment cases. It first explains how Bruen directs courts to begin with the text, and then to look for distinctly similar Founding-era firearm regulations before resorting to higher levels of abstraction. It then applies that method to 18-to-20-year-olds, showing that they fell within the core of the Founding-era arms-bearing population and that the historical record lacks categorical prohibitions on their acquiring, keeping, or carrying arms. Taken together, this Part shows that Bruen’s method fixes the right’s meaning while constraining courts from diluting it through overgeneralization.

A. Bruen’s Analogical Method

Bruen first asks whether the Second Amendment’s text covers the conduct in question; if so, the government must justify its law by showing consistency with the “Nation’s historical tradition of firearm regulation.”^[9] Next, Bruen is clear that the way in which the Founding^[10] generation actually addressed—or not—a potential societal problem is the most “straightforward” evidence of constitutionality.^[11] The presence or absence of a Founding-era firearm regulation “distinctly similar” to the one being challenged by a right-asserting plaintiff is strong evidence under Bruen. In that case, when the societal problem at issue (potential firearm misuse by 18-to-20-year-olds, in this case) is identical, upward abstraction is unnecessary.^[12]

When courts confront a regulatory situation “implicating unprecedented societal concerns” or that was “unimaginable at the founding,” a “more nuanced approach” involving finding analogues that are “relevantly similar” to the challenged regulation is required.^[13] Abstraction is appropriate in this case, bounded “at least”^[14] by Bruen’s “how and why” inquiry: a present-day law must impose a comparable burden for a comparable justification to be constitutional.^[15]

Courts must therefore look first for close historical analogues; only in the absence of such analogues should they generalize upward. Bruen is “neither a regulatory straightjacket nor a regulatory blank check.”^[16] In the appropriate circumstances,^[17] it allows for relying on historical laws that are similar enough to the ones being challenged, as the Court did in United States v. Rahimi.^[18] But it is not a license to swap in non-firearm-specific doctrines when firearm-specific analogues exist, or to abstract higher than necessary. Bruen’s method fixes meaning while allowing application to new circumstances; it does not allow courts to dilute the right by changing the lens of comparison.^[19]

B. In the Founding Era, 18-to-20-Year-Olds Could Keep and Bear Arms

The history is straightforward. 18-to-20-year-olds stood within the core of the Founding-era arms-bearing populace. The Militia Act of 1792^[20] enrolled able‑bodied citizens from 18 to 45 and required each to “provide himself” with a musket.^[21] States sometimes required parents to furnish arms for minors, but that facilitative feature never amounted to a categorical disarmament of 18-to-20‑year‑olds; indeed, several state practices—ranging from armed-community-service mandates to mandating arms carriage when traveling or going to church—assumed that those young adults would be armed.^[22]

Courts striking down age-based restrictions have thus emphasized both the presence of duty at age 18 and the absence of Founding-era prohibitions on 18-to-20-year-olds’ acquiring or carrying arms.^[23] Consistent with that record, modern courts recognizing the right for this cohort have held that 18-to-20‑year‑olds are among “the people” and that the government bears the burden of identifying firearm regulations with a comparable burden and justification.^[24]

Finally, as the Fifth Circuit has explained in Reese v. ATF, the right “to keep and bear” implies a corollary right to acquire.^[25] One cannot keep what one is forbidden to obtain. Treating purchase as outside the Second Amendment invites governments to destroy the right by closing access points.

II. McCoy: Contract Law’s Infancy Doctrine Is the Wrong Analogue

The McCoy majority upholds the federal ban on federal firearm licensee handgun sales to 18-to-20‑year‑olds by analogizing to contract law’s infancy doctrine in the Founding era.^[26] Contracts entered into by persons under the age of 21 were generally voidable at the option of the under-21 party (then considered a minor for contracting purposes). The aim was to protect minors, by giving them the option to avoid the contract. McCoy characterizes the power to void contracts for non-necessaries as a disability “burdening” minors’ ability to purchase firearms, opining that this power made sales on credit risky.^[27] From there, the court reasons that 18 U.S.C. § 922(b)(1)’s ban is “relevantly similar” in its how (deterring sales) and why (youthful judgment) to the common‑law rule.^[28]

This is the wrong level of generality. The infancy doctrine is a general contract rule, not an American gun law. It incidentally affected all non‑necessary purchases; it did not single out arms or reflect a historical judgment that young adults may be disarmed or barred from acquiring firearms. Even on its own terms, the doctrine’s effect is contested and fact‑bound (merchants sometimes extended credit; “necessaries” could be context‑sensitive), which is why the dissent criticized the majority’s economic speculation and emphasized the militia tradition requiring 18‑year‑olds to be armed.^[29]

Bruen requires firearm‑specific analogies when they exist. Here, they do: Founding‑era statutes and practices establish 18 as an age of arms‑bearing responsibility, with no categorical ban on acquisition by 18-to-20‑year‑olds. Abstracting to the contract law of infancy to validate a purchase ban sidesteps Bruen’s “how and why” inquiry and substitutes a non‑firearm, non-disability for a firearm regulation.

III. Bondi: Doubling Down on the Infancy Theory

The Eleventh Circuit’s en banc majority in Bondi similarly leans on contract law’s general infancy principles—minors’ ability to void their own contracts, parental control, and wage rules—to assert that minors “generally could not purchase firearms.”^[30] Its framing of Florida’s under-21 purchase ban as analogous in how and why terms to this legal regime suffers from two doctrinal flaws.

First, this is the same level‑of‑generality error as McCoy’s: a non-firearm-specific potential disability (voidable contracts) is not the relevant tradition under Bruen. Judge Branch’s dissent correctly observes that the majority’s analogue is not a tradition of firearm regulation at all, but a general rule with inferred economic effects, which cannot substitute for the “Nation’s historical tradition of ‘firearm regulation’” in Bruen’s sense.^[31]

Second, the whys don’t match. The purpose of contract law’s infancy doctrine is to protect a young buyer from improvidence. Florida’s gun law is aimed at preventing violent misuse. Those justifications are incommensurable.^[32] And Florida conceded that at the Founding an 18‑year‑old could lawfully purchase a gun with cash—the conduct the statute now forbids—underscoring the mismatch.^[33]

IV. Polis: Commercial “Safe Harbor” in Place of Firearm-Specific History

The Tenth Circuit in Polis takes a different route to the same destination. Recasting Colorado’s across‑the‑board purchase prohibition for 18-to-20‑year‑olds as a “condition[] and qualification[] on the commercial sale of arms,” it treats the law as a “presumptively lawful” sales regulation under District of Columbia v. Heller,^[34] and suggests that purchase may not be covered by the Second Amendment.^[35] That framing lets the court avoid the firearm‑specific analogies that Bruen demands and relies instead on a broad “commercial regulation” presumption (but not a categorical rule, in any event).^[36]

That approach is a category mistake.^[37] As the Fifth Circuit has explained, closing off the ordinary means of acquiring arms to a class of lawful adults is an abridgement of the right and cannot be insulated merely by relabeling it “commercial.”^[38] Purchase is a corollary act practically necessary to “keep,” so such bans must be justified under Bruen’s history‑and‑tradition test rather than swept aside by a rebuttable “safe harbor.”^[39]

Once the correct frame is restored, the better analogues are the same Founding‑era sources Polis sidelined: militia‑enrollment and provisioning laws that treated 18‑year‑olds as arms bearers, and the conspicuous absence of categorical purchase bans for 18-to-20‑year‑olds. Heller’s illustrative list of “presumptively lawful” measures conspicuously did not include age limits.^[40]

Separating “purchase” from “keep and bear” is not doctrinally coherent: a right to possess that denies lawful means of acquisition is illusory. Reese is correct that the constitutional guarantee implies access; allowing a categorical sales ban for an entire class of adults invites the very end‑run Bruen forbids.

V. The Correct Level of Generality

Applied correctly, Bruen’s method proceeds in the following order:

1. Use close firearm-specific analogues first. In this case, courts should use Founding‑era militia and arms‑possession practices for 18-to-20‑year‑olds. Those sources show inclusion within “the people,” an affirmative duty to be armed at 18, and no categorical purchase disqualifications for that cohort.
2. Only if close analogues are lacking and more nuance is warranted should courts abstract. But not to global doctrines (like general contract voidability) that merely had incidental spillover effects on all purchases. Abstraction should preserve Bruen’s “how and why” focus: comparable burdens for comparable reasons.

McCoy and Bondi err by treating contract law’s infancy doctrine as the controlling analogue. Polis errs by elevating a commercial‑regulation safe harbor over firearm‑specific history and by downplaying purchase as a necessary incident of “keep and bear.” Each opinion therefore rests on an incorrect level of generality.

Conclusion

The historical record that Bruen makes decisive is not an empty shelf. It contains firearm‑specific rules and practices about young adults, including a federal statute that presupposed 18‑year‑olds would arm themselves, and no Founding‑era tradition of categorically forbidding 18-to-20‑year‑olds from purchasing arms. When courts bypass those close analogues for distant doctrines about minors’ contracts (McCoy, Bondi) or for general commercial “safe harbors” (Polis), they invert Bruen’s method. The right to keep and bear arms protects 18-to-20‑year‑old adults, and because keeping presupposes acquisition, age‑based purchase bans demand genuine historical justification. On this record, the correct level of generality shows that justification is missing. ↑

1. *Professor of Law, University of Wyoming College of Law; Director, University of Wyoming Firearms Research Center. Fordham University School of Law, J.D., summa cum laude, 2009. University of Rochester Simon School of Business, M.B.A., 1997. I thank Leo Bernabei, Joseph G.S. Greenlee, Nicholas J. Johnson, Jamie G. McWilliam, and Kostas Moros for their valuable insights and feedback.

2. 140 F.4th 568 (4th Cir. 2025). ↑

3. 133 F.4th 1108 (11th Cir. 2025) (en banc). ↑

4.  121 F.4th 96 (10th Cir. 2024). ↑

5. 140 F.4th at 575–77. ↑

6. 133 F.4th at 1118–23. ↑

7. 597 U.S. 1 (2022). ↑

8. See infra notes 8–17 and accompanying text. ↑

9. Bruen, 597 U.S. at 17 (emphasis added). ↑

10. Bruen left it for later whether only the First Founding (when the Second Amendment was ratified in 1791), or that plus the Second Founding (when the Fourteenth Amendment was ratified in 1868), are relevant analogical eras. Id. at 37–38. ↑

11. Id. at 17. ↑

12. Id. at 26 (“[W]hen a challenged regulation addresses a general societal problem that has persisted since the 18th century, the lack of a distinctly similar historical regulation addressing that problem is relevant evidence that the challenged regulation is inconsistent with the Second Amendment.”). ↑

13. Id. at 27–29. ↑

14. Id. at 29. Presumably, the other four members of the “Five Ws”—Who, What, When, and Where—are also relevant. Indeed, the cases discussed herein implicate “Who.” A “sensitive places” case, id. at 30 (citing District of Columbia v. Heller, 554 U.S. 570, 626 (2008)), would implicate “Where.” ↑

15. See id. at 29 (“[W]hether modern and historical regulations impose a comparable burden on the right of armed self-defense and whether that burden is comparably justified are central considerations.” (internal quotation marks omitted)). Bruen also instructs courts to consider whether Founding-era courts rejected similar laws on constitutional grounds. Id. at 27. ↑

16. Id. at 30. ↑

17. See supra notes 12–14 and accompanying text. ↑

18. 602 U.S. 680, 692–700 (2024) (relying on surety and affray laws to uphold the constitutionality of modern bans on arms possession by those under domestic violence restraining orders); see also Jamie G. McWilliam, Second Amendment Principles, 33 Wm. & Mary Bill Rts. J. 1127, 1180 (2025) (explaining how higher levels of generality may be appropriate in some regulatory circumstances under Rahimi); George A. Mocsary, In Denial About the Obvious: Upending the Rhetoric of the Modern Second Amendment, 2024 Cato Sup. Ct. Rev. 201, 212–14 (discussing Rahimi). ↑

19. Cf. Rahimi, 602 U.S. at 740 (Barrett, J., concurring) (“[A] court must be careful not to read a principle at such a high level of generality that it waters down the right.”). ↑

20. Militia Act of 1792, ch. 33, 1 Stat. 271 (1792). ↑

21. Id. ↑

22. George A. Mocsary, Treating Young Adults as Citizens, 27 Tex. Rev. L. & Pol. 607, 621–25 (2023) (discussing practices and citing sources). ↑

23. E.g., Reese v. ATF, 127 F.4th 583 (5th Cir. 2025); Lara v. Comm’r Pa. State Police, 125 F.4th 428 (3d Cir. 2025); Worth v. Jacobson, 108 F.4th 677 (8th Cir. 2024); Firearms Pol’y Coal. v. McCraw, 623 F. Supp. 3d 740 (N.D. Tex. 2022). ↑

24. Reese, 127 F.4th at 589–95; Lara, 125 F.4th at 435–38; Worth, 108 F.4th at 688–92; McCraw, 623 F. Supp. 3d at 748–49. ↑

25. Reese, 127 F.4th at 590. ↑

26. McCoy v. ATF, 140 F.4th 568, 576–77 (4th Cir. 2025). ↑

27. Id. at 577. ↑

28. See id (citing 18 U.S.C. § 922(b)(1)). ↑

29. See id. at 589–90 (Quattlebaum, J., dissenting); see e.g., 1 E. Allan Farnsworth, Farnsworth on Contracts § 4.3 (3d ed. 2004) (“[S]ubstantial areas of commercial activity have developed that could scarcely survive without the patronage of those who are known to be minors.”); Larry A. DiMatteo, Deconstructing the Myth of the “Infancy Law Doctrine”: From Incapacity to Accountability, 21 Ohio N.U. L. Rev. 481, 502 (1994) (“The ‘fact’ is that the ever-growing size of the infant consumer market is an indication that the infancy law doctrine has done little to discourage adults from selling or contracting with minors.”); cf. Wayne Barnes, Arrested Development: Rethinking the Contract Age of Majority for the Twenty‑First Century Adolescent, 76 Md. L. Rev. 405, 444–46 (2017) (arguing that it is unlikely that changing the age of majority to twenty-one would have a meaningful impact on 18-to-21-year-olds’ access to the marketplace). ↑

30. Nat’l Rifle Ass’n v. Bondi, 133 F.4th 1108, 1118 (11th Cir. 2025). ↑

31. Id. at 1164–65 (Branch, J., dissenting) (quoting United States v. Rahimi, 602 U.S. 680 (2024)). ↑

32. See, e.g., id. at 1179 (Brasher, J., dissenting). ↑

33. See id. at 1178. ↑

34.  554 U.S. 570 (2008). ↑

35. Rocky Mountain Gun Owners v. Polis, 121 F.4th 96, 119–20 (10th Cir. 2024). ↑

36. See Mocsary, supra note 17, at 214 n.78. ↑

37. Simply, a category mistake is a semantic or ontological error that occurs when something is treated as though it belongs to a different kind of group than the one to which it actually belongs. For a thorough discussion of category errors, see Ofra Magidor, Category Mistakes, Stan. Encyc. Phil. (Jan. 7, 2025), https://plato.stanford.edu/archives/spr2025/entries/category-mistakes [https://
    perma.cc/5ASN-QA6D]. ↑

38. Reese v. ATF, 127 F.4th 583, 590 (2025). ↑

39. Contra Polis, 121 F.4th at 119 n.5. ↑

40. See District of Columbia v. Heller, 554 U.S. 570, 626–27 (2008). ↑

Introduction

When any one, or more, shall take upon them to make laws,
whom the people have not appointed so to do,
they make laws without authority,
which the people are not therefore bound to obey.

-John Locke^[1]

The Framers of the Constitution deliberately divided federal powers among three branches, ensuring that each branch is held accountable by the others.^[2] The private non-delegation doctrine stands for the common-sense principle that granting such power to a private entity, not vested with such power by the Constitution or subject to the checks of the other branches, would “dash the whole scheme.”^[3] 

Under the private non-delegation doctrine, it is inaccurate to claim that private entities can play no role in exercising federal power. For example, in Sunshine Anthracite Coal Co. v. Adkins, the Supreme Court upheld a statutory scheme that allowed private boards to set wages and hours for coal producers that were then “approved, disapproved, or modified by the [National Bituminous Coal Commission].”^[4] Despite the fact that this sort of regulation “is necessarily a governmental function,”^[5] the Court reasoned that such delegation was permissible because the private boards “function[ed] subordinately” and “operate[d] as an aid to the Commission[,] … subject to its pervasive surveillance and authority.”^[6] To sum up the doctrine, “a private entity may wield government power only if it ‘functions subordinately’ to a[] [federal] agency with ‘authority and surveillance’ over it.”^[7]

In 2024, the Fifth Circuit held, for the second time, that the Horseracing Integrity and Safety Act (“HISA”) violates the private non-delegation doctrine in National Horsemen’s Benevolent and Protective Association v. Black (“Black II”).^[8] A group including the National Horsemen’s Benevolent and Protective Association (“NHBPA”)^[9] challenged HISA as facially unconstitutional, claiming that the statute violated the Constitution in various ways.^[10] Though most of these challenges failed, the Fifth Circuit concluded that “HISA is enforced by private entities that are not subordinate to the [Federal Trade Commission],” and therefore ultimately struck down HISA as unconstitutional under the private non-delegation doctrine due to its enforcement scheme.^[11]

This comment argues that the Fifth Circuit’s holding in Black II is flawed and may lead to disastrous consequences both within and beyond the horseracing industry. Part I begins by outlining the development of HISA. Part II then discusses legal challenges to the current version of the statute. In Part III, I discuss the Fifth Circuit’s holding in Black II before critiquing the ruling and explaining how it may lead to devastating consequences that expand beyond the sport in Part IV. Part V concludes with a discussion of HISA’s future.

I. Background

Horseracing has never faced so turbulent a period as the present. Throughout the 2010s, an average of nearly 600 Thoroughbred racehorses suffered fatal injuries each year.^[12] In 2021, Hall of Fame trainer Bob Baffert began his three-year suspension from Churchill Downs, the home of the Kentucky Derby, resulting from the failed drug test of subsequently disqualified Kentucky Derby winner Medina Spirit.^[13] In 2023, tens of thousands of eyes were on-site at Santa Anita Park, and many more at home, to watch the then undefeated fan-favorite Maple Leaf Mel break down just seconds before crossing the wire in the leading position.^[14] These events, among countless others, led to scathing attacks on the sport from multiple fronts. Commentators from some of the most prominent news outlets questioned its ethics and claimed that the “entire sport is a foul.”^[15] Other, less condemnatory, commentators pointed out the obvious: there was “unquestionably a need for unified oversight.”^[16]

Congress addressed these challenges in 2020 by enacting HISA to “improve the integrity and safety of horseracing by requiring a uniform anti-doping and medication control program to be developed and enforced.”^[17] HISA recognized the Horseracing Integrity and Safety Authority (“Authority”), described as a “private, independent, self-regulatory, nonprofit corporation,”^[18] and tasked it with “developing and implementing a horseracing anti-doping and medication control program and a racetrack safety program.”^[19] To accomplish these tasks, Congress assigned the Authority the responsibility of forming both (i) an Anti-Doping and Medication Control Standing Committee and (ii) a Racetrack Safety Standing Committee to advise and guide the “development and maintenance” of the programs.^[20]

HISA requires the Authority to submit its proposed rules, including those regarding racing standards and enforcement, to the Federal Trade Commission (“FTC”), which must approve the rule before it takes effect.^[21] The FTC “shall approve a proposed rule” if consistent with the Act and rules formerly approved by the FTC.^[22] Once approved, covered persons^[23] must comply with these rules as “a condition of participating in covered races and in the care, ownership, treatment, and training of covered horses.”^[24]

Many industry actors welcomed the legislation, declaring it “the beginning of a new, safer era for our storied sport.”^[25] However, praise was not universal. Others soon challenged the legality of HISA, including the NHBPA.^[26] The NHBPA, comprised of thousands of racehorse owners and trainers,^[27] claimed that HISA was facially unconstitutional due to violations of various constitutional principles, including the private non-delegation doctrine.^[28]

In 2022, the NHBPA’s challenge made it to the Fifth Circuit Court of Appeals, which ruled HISA unconstitutional for violating the private non-delegation doctrine.^[29] In ruling HISA unconstitutional, the Fifth Circuit focused on the Authority’s rulemaking power.^[30] As noted by the court, if the Authority proposed a rule that was consistent with HISA, the FTC was required to approve it.^[31] Moreover, if the FTC found the rule inconsistent with the Act, it could only suggest changes.^[32] The FTC also lacked the power to review policy choices made by the Authority and directed the public to engage directly with the Authority if they disagreed with such choices.^[33] The court noted that “[a]n agency does not have meaningful oversight if it does not write the rules, cannot change them, and cannot second-guess their substance.”^[34] Therefore, the Fifth Circuit held that Congress violated the private non-delegation doctrine by “giv[ing] a private entity the last word over what rules govern our nation’s thoroughbred horseracing industry.”^[35]

The court further rejected HISA’s contention that the statute was constitutional because it was modeled after the Maloney Act, which no court has ever found to violate the private non-delegation doctrine.^[36] The Maloney Act “authorizes a private entity, [the Financial Industry Regulatory Authority], to propose rules that the Securities and Exchange Commission . . . decides whether or not to promulgate into law.”^[37] However, unlike the original version of HISA, the Maloney Act allows the Securities and Exchange Commission (“SEC’) to “‘abrogate, add to, and delete from’ FINRA rules ‘as the [SEC] deems necessary or appropriate[.]’”^[38] This language is meaningful, as it gives the SEC the “final word on the substance of [securities industry] rules” and leaves FINRA subordinate to the SEC.^[39] Because this grants the SEC the power to require changes to FINRA rules, whereas the FTC could only recommend changes to rules, the particular language of the Maloney Act “ma[de] all the difference.”^[40]

On account of the Fifth Circuit’s ruling, Congress quickly amended HISA in 2022.^[41] As amended, the statute now gives the FTC the power to “abrogate, add to, and modify the rules of the Authority promulgated in accordance with [HISA].”^[42] The amendment was heavily inspired by language of the Maloney Act focused on in Black I and the Bituminous Coal Act of 1937, which the Supreme Court upheld as constitutional in Adkins.^[43] By deliberately structuring HISA’s amendment to be “consistent in every material respect with a regulatory configuration long used by Congress and sustained by the courts,” Congress hoped to ensure HISA’s constitutionality.^[44]

II. Post-Amendment Challenges

A. Oklahoma v. United States:

Despite the carefully crafted amendment, HISA soon faced additional legal challenges based on, among other claims, the private non-delegation doctrine. One such challenge appeared before the Sixth Circuit Court of Appeals in 2023, where a group that included numerous industry actors objected to HISA’s constitutionality.^[45] The Sixth Circuit, in Oklahoma v. United States, ultimately held HISA constitutional following its amendment.^[46]

First addressing the private non-delegation challenge in the rulemaking context, the Sixth Circuit held that the additional power granted to the FTC by the amendment “creates ‘a clear hierarchy.’”^[47] In sum, the challenge to the Authority’s rulemaking power failed because:

In full, § 3053(e)’s amended text gives the FTC ultimate discretion over the content of the rules that govern the horseracing industry and the Horseracing Authority’s implementation of those rules. By the same token, ultimate “law-making is not entrusted to the [Authority].” That makes the FTC the primary rule-maker, and leaves the Authority as the secondary, the inferior, the subordinate one.^[48] 

The court then turned to the private non-delegation challenge in the enforcement context. Though the court acknowledged the “extensive” enforcement powers held by the Authority, namely “implement[ing] the Act, investigat[ing] potential rule violations, and enforc[ing] the rules through internal adjudications and external civil lawsuits,” it determined that the rulemaking and revision power granted to the FTC by Congress’ amendment provides the FTC “‘pervasive’ oversight and control” of such enforcement activities, “just as it does in the rulemaking context.”^[49] To make this evident, the court provided various examples of how the FTC could ultimately control the Authority’s enforcement actions, listing numerous rules the FTC could pass to both grant itself more enforcement power and limit that of the Authority.^[50] 

The Sixth Circuit further relied on the FTC’s ability to reverse decisions made by the Authority and review the Authority’s decisions regarding adjudication before they are deemed final, concluding that “the FTC has full authority to review the Horseracing Authority’s enforcement actions.”^[51] Therefore, the court determined that the Authority “is ‘subject to [the FTC’s] pervasive surveillance and authority,’ revealing that the Authority ‘operate[s] as an aid to the [FTC],’ nothing more.”^[52] For these reasons, the Sixth Circuit rejected the private non-delegation challenge and held HISA facially constitutional.^[53]

B. Walmsley v. Federal Trade Commission:

The Eighth Circuit followed suit in 2024, holding the amended HISA facially constitutional in Walmsley v. Federal Trade Commission.^[54] Addressing the Authority’s rulemaking power, the court stated, “[HISA] as amended gives the Commission ‘ultimate discretion over the content of the rules that govern the horseracing industry.’”^[55] Because the FTC has the power to change the Authority’s rules when it disagrees with them, the Eighth Circuit deemed the FTC the “primary rule-maker,” leaving the Authority “inferior” and “subordinate.”^[56] Therefore, the court rejected the private non-delegation challenge to HISA’s rulemaking structure.^[57]

The court then addressed the Authority’s enforcement powers under HISA. Like the Sixth Circuit, the Eighth Circuit illuminated the FTC’s superiority over the Authority by listing numerous ways in which the FTC could utilize its newfound rulemaking and revision powers to control the Authority’s enforcement actions.^[58] This, combined with the FTC’s power to review and reverse the Authority’s enforcement actions, led the court to conclude that the FTC has “‘pervasive oversight and control of the Authority’s enforcement activities.’”^[59] Accordingly, the court rejected the private non-delegation challenge and held HISA facially constitutional.^[60]

However, unlike the Sixth Circuit’s opinion in Oklahoma, Walmsley’s holding was met with dissent. Judge Raymond Gruender disagreed with the majority’s reasoning regarding the Authority’s enforcement power under HISA.^[61] Focusing on “the plain text of HISA,” he viewed the statute as creating “a clear delegation of enforcement power between the FTC and the Authority, ‘each within the scope of their powers and responsibilities under [HISA].’”^[62] Specifically, Judge Gruender reasoned:

HISA expressly provides that the Authority’s “powers” include full investigatory authority and the ability to bring suit against alleged violators for injunctive relief. Therefore, by the plain text of the statute, the FTC cannot impede upon the power granted to the Authority, nor can the FTC compel Authority enforcement action. In this fashion, the Authority does not “function subordinately” to an agency with “authority and surveillance” over it, in violation of the private nondelegation doctrine.^[63]

Judge Gruender further reasoned that the FTC cannot use its newfound rulemaking and revision powers to grant itself greater enforcement powers or limit those of the Authority.^[64] In his view, doing so would “rewrite the statutory scheme that Congress enacted” because “the plain text of HISA empowers the Authority, and not the FTC, with broad enforcement power.”^[65] In sum, he would have held that HISA as amended violates the private non-delegation doctrine.^[66]

III. The Fifth Circuit’s Ruling in Black II

Judge Gruender’s dissenting opinion in Walmsley largely reflects, and indeed relied on, the Fifth Circuit Court of Appeals’ 2024 decision in Black II.^[67] After initially ruling HISA unconstitutional in 2022 for its original rulemaking structure, the Fifth Circuit once again struck down HISA as unconstitutional two years later, despite Congress’ amendment.^[68]

The Fifth Circuit’s opinion in Black II recognized that “in light of Congress’ amendment to HISA [], the Authority’s rulemaking power is subordinate to the FTC’s,” and instead found issue with the Authority’s enforcement powers.^[69] Addressing this issue, the court first listed various enforcement actions that the Authority can perform “without the FTC’s involvement.”^[70] In the court’s view, the FTC’s ability to “review sanctions at the back end” is not enough to save HISA, as “each and every one of those actions is ‘enforcement’ of HISA,” and, therefore, “[a]s far as enforcement goes, the horse was already out of the barn.”^[71] Specifically, the court reasoned:

The Act’s plain terms permit only one conclusion: HISA is enforced by a private entity, the Authority. The Authority decides whether to investigate a covered entity for violating HISA’s rules. The Authority decides whether to subpoena the entity’s records or search its premises. The Authority decides whether to sanction it. And the Authority decides whether to sue the entity for an injunction or to enforce a sanction it has imposed. To be sure, the Authority does not perform these functions itself. Rather, HISA requires the Authority to contract with another private entity, USADA, which undertakes enforcement “on behalf of the Authority.” § 3054(e)(1)(E)(i). The bottom line, though, is that a private entity, not the agency, is in charge of enforcing HISA.^[72]

The court further rejected the argument that the FTC could “rein in” the Authority’s enforcement actions using its rulemaking power under the amended HISA.^[73] In the Fifth Circuit’s view, doing so would “let the [FTC] rewrite the statute.”^[74] Focusing on the same language Judge Gruender highlighted in Walmsley, the court concluded that HISA set out a “definite enforcement scheme.”^[75] In the court’s view, any rule change by the FTC to grant itself more enforcement power under HISA would impermissibly “amend the enforcement scheme delineated by the statute.”^[76]

Despite originally stating that the language absent in the original version of HISA, but present in the Maloney Act, created a “key distinction” that made “all the difference” between the two pieces of legislation,^[77] the Fifth Circuit in Black II held that Congress’ inclusion of nearly identical language to the Maloney Act in the amended HISA^[78] could not save it.^[79] In the court’s view, the division of power outlined in HISA “gives the FTC scant oversight until the enforcement has actually occurred,” which fails to make the Authority subordinate to the FTC.^[80] Because the court also ruled that the FTC’s newfound rulemaking power does not allow it to “amend the plain division of enforcement power laid out in HISA’s text,” the court reasoned that the FTC-Authority relationship strays materially from the SEC-FINRA relationship, as laid out in the Maloney Act, because “the FTC lacks any tools to ensure that the law is properly enforced.”^[81] Therefore, the Fifth Circuit once again held that HISA violates the private non-delegation doctrine.^[82]

IV. Discussion

A. Interpretation:

The Fifth Circuit’s decision in Black I initially focused on the Authority’s rulemaking powers under HISA.^[83] In Black II, the court acknowledged that Congress effectively addressed that issue by amending the statute and instead shifted its focus to the Authority’s enforcement powers, which it entirely failed to address in Black I.^[84] Despite the fact that Congress amended HISA to include nearly identical language to the Maloney Act, which the court originally used to distinguish the two pieces of legislation, the Fifth Circuit held that the amended HISA remains unconstitutional.^[85] The court’s decision in Black II therefore puzzlingly moves the finish line, which is particularly concerning given that it serves as the lone circuit preventing HISA from protecting horses and jockeys across the country.

In my view, the Fifth Circuit’s decision relied on a flawed interpretation of HISA. Under HISA, the Authority only has “subpoena and investigatory authority” over “violations committed under its jurisdiction.”^[86] Put more simply, when considering HISA as a whole, even before the amendment, the Authority could only exercise these powers if they were in accordance with rules approved by the FTC.^[87] The court was also concerned by the Authority’s ability to impose sanctions and commence civil actions under HISA.^[88] However, as amended, the statute allows the FTC to “abrogate, add to, and modify the rules of the Authority.”^[89] Therefore, “the FTC could easily set criteria for the Authority’s enforcement activities,” such as “requir[ing] the Authority to clear enforcement actions [including imposing sanctions and commencing civil actions] in advance.”^[90] Consequently, using the “scope of [its] powers”^[91] under HISA, the FTC can implement rules granting them greater enforcement power, thereby subordinating the Authority’s enforcement powers under HISA.

Doing so would not require the FTC to “rewrite the enforcement scheme Congress enacted” as the Fifth Circuit and Judge Gruender concluded.^[92] “[I]ssu[ing] rules protecting covered persons from overbroad subpoenas or onerous searches,”^[93] for example, does not eliminate the Authority’s power granted under § 3054(h). Rather, it simply imposes a limitation to the exercise of such enforcement power, which the FTC would be entirely within its “scope of [] powers”^[94] under the scheme created by HISA to do.^[95] Still, the FTC’s ability to subordinate the Authority’s power under HISA in this way “gives it ‘pervasive oversight and control of the Authority’s enforcement activities.’”^[96] Because the FTC can work within HISA’s framework to grant itself greater enforcement powers and limit that of the Authority, creating such rules would not rewrite HISA’s enforcement scheme. To this point, the Eighth Circuit stated, “To subordinate the Authority’s enforcement activity, moreover, the Commission need only work within the structure of the Act as designed, not create a new statutory regime.” ^[97]

Crucially, the FTC also has the power to review and reverse any final decision regarding civil sanctions imposed by the Authority.^[98] This power “reveal[s] that the Authority ‘operate[s] as an aid to the [FTC],’ nothing more.”^[99] Combined with its rulemaking power, this allows the FTC to “control the Authority’s enforcement activities and ensure that the FTC, not the Authority, ultimately decides how the Act is enforced.”^[100] 

I also disagree with the Fifth Circuit’s conclusion that the FTC-HISA relationship materially differs from the SEC-FINRA relationship. HISA’s “back-end” language is nearly identical to that of the Maloney Act, which allows the SEC to “‘abrogate, add to, and delete from’ FINRA’s rules ‘as the Commission deems necessary or appropriate.’”^[101] According to the court, the absence of this language in the original version of HISA “ma[de] all the difference” between the FTC-HISA and SEC-FINRA relationships.^[102] Now that Congress amended HISA to address this difference, legislators supporting HISA believe that “there can be no question that the Authority is properly subordinated to the FTC, just as FINRA is to the SEC.”^[103] 

Despite its reasoning in Black I, the Fifth Circuit held that the two relationships still differ materially after HISA’s amendment.^[104] It reasoned that whereas the SEC can initiate its own investigations and seek sanctions, along with other enforcement powers, the FTC’s power is limited to a tail-end review of enforcement actions that already occurred.^[105] In the court’s view, this tail-end enforcement power is insufficient to make the Authority subordinate to the FTC.^[106] However, as previously discussed, the FTC’s enforcement power is not limited to a tail-end review, as it can grant itself greater enforcement powers within HISA’s enforcement scheme.^[107]

Moreover, like the SEC-FINRA relationship, the FTC must approve rules proposed by the Authority before they take effect.^[108] And like the SEC, the FTC can propose its own rules under HISA.^[109] As discussed, the language in the amended HISA granting the FTC this power is nearly identical to that granting the SEC such power.^[110] And like the SEC-FINRA relationship, any enforcement decision handed down under HISA is subject to review by the FTC.^[111] Indeed, Judge Cole’s concurrence in Oklahoma argues that HISA’s enforcement scheme is “even more constitutionally sound” than that found in the Maloney Act.^[112] He reasons, “HISA, unlike the Maloney Act, unambiguously empowers the FTC to obtain additional evidence not in the record below and to review the proceeding de novo.”^[113] Therefore, I agree with the legislators who claim that “the statute as amended is materially indistinguishable from the Maloney Act,”^[114] if not stronger. Since the Maloney Act is universally accepted as constitutional, so should be HISA.^[115]

B. Policy Considerations:

While there is a sound legal foundation for upholding HISA, important policy considerations provide additional support for maintaining the legislation. Whereas other sports are governed by uniform standards put forth by their respective leagues—such as the MLB, NFL, and NHL—horseracing lacks such an entity.^[116] It is important to note that racehorses, too, are athletes that put their bodies on the line to compete. Because horses have noticeably thin legs that carry their roughly half-ton body weight when running at high speeds, just one awkward step can cause a leg to break.^[117] Such an injury nearly always requires euthanasia.^[118] 

The risk for injury is heightened when horses are drugged to mask any sign of pain. Drugging may allow an injured horse to pass veterinarian checks and cause “the horse to overexert itself during periods of intense exercise, which can lead to accidents, broken limbs, or death.”^[119] These accidents harm not only horses, but the jockeys who ride them as well. In fact, hundreds of jockeys have died due to “training or racing accidents in the United States[,] . . . often because of grave injury to a horse.”^[120] Because horses, unlike humans, cannot verbalize their pain, and financial incentives exist to perform such abusive practices, uniform regulation and oversight through HISA is perhaps even more necessary for horseracing than other sports that currently enjoy it.^[121]

While the need for regulation and oversight is apparent, some argue that it should be handled at the state level. As the Sixth Circuit noted in Oklahoma, “no one doubts the imperative for oversight. The question, as is so often the case, is whether the regulation should be national or local.”^[122] Serious problems exist when each state is allowed to set their own standards for the sport. The result of such a scheme is a “patchwork” of rules that allow some states to gain a competitive advantage over others by adopting more lenient regulations that put horses and jockeys at risk.^[123] Furthermore, horses generally race in multiple states throughout their careers. This can cause difficulty for trainers, veterinarians, and other industry actors who need to keep track of the various rules each state puts forth to ensure compliance. Kenny McPeek, trainer of 2024 Horse of the Year, Thorpedo Anna,^[124] recently explained, “every racetrack . . . and every state has different rules.”^[125] Attempting to align these rules without a federal body would be inefficient. As noted in an amicus brief filed by legislators in support of HISA, “each of horseracing’s well-meaning attempts to create uniformity through state-by-state compacts and other agreements failed.”^[126] Instead, McPeek believes that states should use HISA as an opportunity to align their rules.^[127] For these reasons, the need for federal regulation of the sport through HISA is evident.

This need is heightened because horseracing is currently enduring a crisis.^[128] A member of the Board of Stewards of The Jockey Club, attorney William M. Lear Jr., declared that the sport is facing an “existential threat.”^[129] He further emphasized, “If our response to that threat is or even appears to be business as usual, we’re going to lose.”^[130] The long-term consequences of maintaining the status quo would be disastrous, if not terminal, for the sport. Therefore, while HISA has made horseracing slightly more costly for industry actors,^[131] it seems like a fair trade off to help ensure the future of the sport. And as noted by Lisa Lazarus, HISA’s chief executive officer, “any added costs from HISA are a fraction of what honest horse people lose in purse money each year to dishonest ones.”^[132] To make the most of their investment, industry actors would be wise to ensure not only their compliance with the regulations put forth by HISA, but their counterparts’ as well. Failing to do so would risk allowing a few bad seeds to kill the garden they harvest from.

Industry actors opposed to HISA would like people to believe that uses of unethical practices are too rare to justify the restrictions and costs that the Authority’s regulations subject them to.^[133] However, in February 2025, the Authority announced that it would pursue charges against fourteen horse trainers as part of what they alleged was a “widespread conspiracy” to violate rules designed to prevent pre-race pain-masking injections.^[134] Of the horses receiving such alleged injections, 30 percent never ran another race after the race they were injected in preparation for, 10 percent were deemed lame post-race by veterinarians, and three died from injuries sustained during the race.^[135] Four more of these horses died within just months of the injections.^[136] One of the trainers implicated by this investigation criticized HISA less than a year prior, claiming that HISA’s rules were “a nightmare” and “detrimental to the horse.”^[137] Concerningly, the dark side of horseracing extends well beyond this scandal.^[138] These situations bring justifiable criticism to the sport, jeopardizing its future entirely.

The collapse of horseracing would lead to disastrous consequences. For starters, the sport is a popular source of entertainment, as it has been for centuries.^[139] Tens of millions of people tuned into the Kentucky Derby in 2025, setting the highest viewership mark since 1989.^[140] Dedicated fans continue to travel to racetracks across the country to view the horses in action, bet, and spend time with family and friends.^[141] Especially in regions where horseracing plays an important cultural role, losing such a long-standing and valued source of entertainment would be devastating. 

Second, the loss of the sport would cause economic havoc, as the equine industry supports approximately 2.2 million jobs in the United States.^[142] While some careers are quite clearly linked to horses, such as veterinarians, farriers, and breeders, other less-obvious careers are also supported by the equine industry, including truck drivers, chemists, and journalists.^[143] As a whole, the equine industry contributed roughly $177 billion to the United States economy in 2023.^[144] The impact is substantial at the state and local level as well. The Thoroughbred industry has an economic impact of billions of dollars in Kentucky alone, with racing operations contributing over $500 million to that sizable figure.^[145] In 2022, the Breeders’ Cup World Championships were hosted at Keeneland Race Course in Lexington, Kentucky, and brought a nearly $82 million economic impact to the area.^[146] Clearly, losing horseracing would lead to disastrous economic consequences at the national, state, and local levels.

V. The Future of HISA

In June 2024, the Supreme Court denied certiorari in the Sixth Circuit’s ruling in Oklahoma, which upheld HISA.^[147] However, in September 2024, the Court issued an administrative stay, which put a temporary pause on the effects of the Fifth Circuit’s ruling.^[148] Predicting how the Supreme Court would rule should it review HISA based on purely ideological grounds is challenging, given that HISA was passed with bipartisan support^[149] and upheld by two of the three circuit courts that reviewed it—all of which lean conservative.^[150] Additionally, the issuance of the stay does not necessarily mean that the Court will ultimately uphold HISA. That said, the Court’s decision to prevent the Fifth Circuit’s ruling from taking effect just months after it declined to review the Sixth Circuit’s ruling in Oklahoma, indicates that it would be likely to uphold HISA.

Should it decide to review HISA, the Supreme Court would have the opportunity to add clarity to a muddied doctrine that lower courts have struggled to apply for decades.^[151] As noted by the Fifth Circuit, “[k]ey to applying the [private non-delegation] doctrine are two eighty-year-old Supreme Court cases.”^[152] In June 2025, the Court issued a ruling in Federal Communications Commission v. Consumers’ Research, which held that a different statutory scheme—in which a private entity aids the Federal Communications Commission by “manag[ing] the [Universal Service] Fund’s^[153] day-to-day operations and also play[ing] a role in producing the financial projections that end up determining the contribution factor^[154]”—did not violate the private non-delegation doctrine.^[155] As a result, the Court vacated the Fifth, Sixth, and Eighth Circuits’ rulings, ordering them to give HISA “further consideration.”^[156] However, HISA and the statute at issue in Consumers’ Research clearly differ, meaning that the decision may not have a drastic impact on the circuits’ holdings. In fact, it appears that each side plans on making similar, if not the same, arguments the next go-around.^[157] Therefore, it is difficult to predict how the Court would work with its precedent in ruling on HISA’s statutory scheme. 

In the Appointments Clause context, the Supreme Court has recently placed a particular emphasis on the power to issue a final, unreviewable decision in determining whether an officer is deemed inferior or superior. In United States v. Arthrex, Inc., the Supreme Court held that “the unreviewable authority wielded by [administrative patent judges] during inter partes review is incompatible with their appointment by the Secretary to an inferior office.”^[158] To fix this defect, the Court granted the Director of the Patent and Trademark Office the power to “review [such] decisions and, upon review, . . . issue decisions himself on behalf of the Board.”^[159] As the Court declared, this decision is consistent with “the rule from Edmond that the exercise of executive power by inferior officers must at some level be subject to the direction and supervision of an officer nominated by the President and confirmed by the Senate.”^[160] 

Should the Court apply similar principles in determining whether the Authority is subordinate in the context of the private non-delegation doctrine, HISA would almost certainly be upheld as constitutional. The FTC has the power to review de novo decisions of the Authority, and subsequently, an administrative law judge, at its discretion and may “affirm, reverse, modify, set aside, or remand” such decisions.^[161] This, combined with the FTC’s ability to adopt rules granting itself greater enforcement powers and limiting that of the Authority,^[162] should be considered more than enough oversight to make the Authority subordinate to the FTC in satisfaction of the private non-delegation doctrine.

Still, the Court may very well eventually strike down the statute as unconstitutional, in which case Congress would need to once again work to amend HISA. If the recent amendment of the statute was not enough to satisfy the constitutional concerns, Congress will likely need to turn to an extensive re-writing of the statute. Such an outcome would be rather disappointing, however, as a logical interpretation of HISA as it stands satisfies the private non-delegation doctrine.^[163] Forcing Congress to act once again will only serve to unnecessarily delay the implementation of HISA, putting more people and horses at risk in the meantime.

Conclusion

HISA is already making a noticeable impact on the horseracing industry. The statute has led to “year-over-year declines” in horse death rates, reflecting the sport’s improved safety and integrity.^[164] Meanwhile, the death rate at tracks without HISA regulation is nearly double that of HISA-regulated tracks.^[165] Despite the increased responsibilities for industry actors under HISA, trainer Kenny McPeek recognizes the positive potential of HISA and hopes that regulations under the statute will eventually expand into areas such as financial responsibility within the sport.^[166] In his words, “horsemen should support [the Authority].”^[167] 

The private non-delegation doctrine reflects some of this country’s most foundational principles,^[168] and therefore unlawful delegations of federal power should absolutely be struck down as unconstitutional. However, as the Sixth Circuit noted, “[s]ometimes government works”^[169]—in HISA’s case, it does. While the Fifth Circuit’s ruling in Black II currently prevents HISA from crossing the wire, one can hope that the legal challenges will soon be resolved upon further review. Only at that point can HISA operate at its fullest potential, increasing the safety and integrity of horseracing and thereby “preserv[ing] the sport for generations to come.”^[170]

1. * J.D. Candidate, Washington University School of Law (2026); B.S., Southern Illinois University Edwardsville (2023). Thank you to the Washington University Law Review staff—especially Haley Bills and Geremia Lizier-Zmudzinski—for their assistance throughout the publication process. I would also like to express my appreciation for my family and friends who have supported me throughout my academic career. Finally, I would like to offer a special thank you to my grandfather for spurring my passion for horseracing, which made this writing process particularly rewarding.

   Frank D. Garrison, Will the Supreme Court Finally Address the Private Nondelegation Doctrine?, The Federalist Soc’y (Dec. 12, 2024) (quoting John Locke, Two Treatises of Government 426 (1960 ed.)), https://fedsoc.org/commentary/fedsoc-blog/will-the-supreme-court-finally-address-the-private-nondelegation-doctrine [https://perma.cc/GN74-GY6T]. ↑

2. See U.S. Const. art. I, § 1; id. art. II, § 1; id. art. III, § 1; The Federalist No. 51 (James Madison). ↑

3. Dep’t of Transp. v. Ass’n of Am. R.Rs., 575 U.S. 43, 61 (2015) (Alito, J., concurring). ↑

4. Sunshine Anthracite Coal Co. v. Adkins, 310 U.S. 381, 377−78, 400 (1940). ↑

5. Carter v. Carter Coal Co., 298 U.S. 238, 311 (1936). ↑

6. Adkins, 575 U.S. at 388, 399. ↑

7. Nat’l Horsemen’s Benevolent & Protective Ass’n v. Black (Black II), 107 F.4th 415, 423 (5th Cir. 2024) (quoting Nat’l Horsemen’s Benevolent & Protective Ass’n v. Black (Black I), 53 F.4th 869, 881, n. 21 (5th Cir. 2022) (“Black I”)) cert. granted, judgement vacated sub nom. Texas v. Black, No. 24-465, 2025 WL 1787677 (U.S. June 30, 2025). ↑

8. Black II, 107 F.4th at 441. ↑

9. See infra note 27 and accompanying text. ↑

10. Black II, 107 F.4th at 423. ↑

11. Id. at 441. ↑

12. See Shannon Luce, The Jockey Club Releases Data from the Equine Injury Database for 2022, The Jockey Club (Mar. 20, 2023), https://jockeyclub.com/Default.asp?section=Resources&area=10&story=1390 [https://perma.cc/5NNZ-WDSX]. ↑

13. Baffert Is Back. Churchill Downs Puts End to Hall of Fame Trainer’s Three-Year Suspension, Lexington Herald Leader (July 19, 2024 5:54 PM), https://www.kentucky.com/sports/horses/kentucky-derby/article290249994.html [https://perma.cc/8RYG-S5TU]. ↑

14. Tim Layden, Maple Leaf Mel and the Long Arc of a Horse-Racing Tragedy, NBC Sports (Nov. 1, 2023, 1:15 PM), https://www.nbcsports.com/news/maple-leaf-mel-and-the-long-arc-of-a-horse-racing-tragedy [https://perma.cc/2XUP-E55Y]. ↑

15. Sally Jenkins, Forget Maximum Security’s Misstep; The Whole of Horse Racing Is a Foul, Wash. Post (May 5, 2019, 2:43 PM), https://www.washingtonpost.com/sports/2019/05/05/forget-maximum-securitys-misstep-whole-horse-racing-is-foul/ [https://perma.cc/F276-EV7C]; see Jeremy Engle, Is Horse Racing Ethical?, N.Y. Times (Apr. 9, 2019), https://www.nytimes.com/2019/04/09/learning/is-horse-racing-ethical.html [https://perma.cc/DR37-5PMV]. ↑

16. Tim Layden, Horse Racing’s Biggest Problem Will Dominate the 2019 Kentucky Derby, Sports Illustrated (Apr. 29, 2019), https://www.si.com/horse-racing/2019/04/29/kentucky-derby-horses-deaths-racetrack-rule-changes [https://perma.cc/8XEP-6QV6]. ↑

17. H.R. Rep. No. 116-554, at 1 (2020). ↑

18. 15 U.S.C. § 3052(a). The Authority is “governed by a board of directors,” which is comprised of both directors within the equine industry and independent directors. Id. § 3052(b). ↑

19. Id. § 3052(a). ↑

20. Id. §§ 3052(a), 3052(c)(1)(A), 3052(c)(2)(A). ↑

21. Id. § 3053(b)(2). ↑

22. Id. § 3053(c)(2). ↑

23. See Id. § 3051(6) (“The term ‘covered persons’ means all trainers, owners, breeders, jockeys, racetracks, veterinarians, persons (legal and natural) licensed by a State racing commission and the agents, assigns, and employees of such persons and other horse support personnel who are engaged in the care, training, or racing of covered horses.”). ↑

24. Id. § 3054(d)(1); see also id. § 3051(5) (“The term ‘covered horserace’ means any horserace involving covered horses that has a substantial relation to interstate commerce, including any Thoroughbred horserace that is the subject of interstate off-track or advance deposit wagers.”); id. § 3051(6) (“The term ‘covered horse’ means any Thoroughbred horse, or any other horse made subject to this chapter . . . during the period- (A) beginning on the date of the horse’s first timed and reported workout at a racetrack that participates in covered horseraces or at a training facility; and (B) ending on the date on which the Authority receives written notice that the horse has been retired.”). ↑

25. American Association of Equine Practitioners et al., Our Voices: HISA Leads to Safer, More Transparent Sport, Blood Horse (Apr. 21, 2021), https://www.bloodhorse.com/horse-racing/articles/247577/our-voices-hisa-leads-to-safer-more-transparent-sport [https://perma.cc/32EM-8B4K]. ↑

26. See Black I, 53 F.4th 869 (5th Cir. 2022). ↑

27. About HBPA, Nat’l Horsemen’s Benevolent & Protective Ass’n, https://nationalhbpa.com/about-hbpa/ [https://perma.cc/LV7C-XUD3] (“We are Horsemen Helping Horsemen.”). ↑

28. Black I, 53 F.4th at 875. ↑

29. Id. at 890. ↑

30. Id. at 872. ↑

31. Id.; see also 15 U.S.C. § 3053(c)(3). ↑

32. Black I, 53 F.4th at 872; see also 15 U.S.C. § 3053(d). ↑

33. Black I, 53 F.4th at 872. ↑

34. Id. ↑

35. Id. ↑

36. Id. at 887−88. ↑

37. Brief Amici Curiae of Senator Mitch McConnell and Representatives Paul Tonko and Andy Barr in Support of Applicants at 5, Walmsley v. Fed. Trade Comm’n, 117 F.4th 1032 (8th Cir. 2024) (No. 23-2687). ↑

38. Black I, 53 F.4th at 887 (quoting 15 U.S.C. § 78s(c)). ↑

39. Black I, 53 F.4th at 887. ↑

40. Id. at 888. ↑

41. See 15 U.S.C. § 3053(e). ↑

42. Id. ↑

43. Brief Amici Curiae of Senator Mitch McConnell and Representatives Paul Tonko and Andy Barr in Support of Applicants, supra note 37, at 6−7 (citing Sunshine Anthracite Coal Co. v. Adkins, 310 U.S. 381(1940)). ↑

44. Brief Amici Curiae of Senator Mitch McConnell and Representatives Paul Tonko and Andy Barr in Support of Applicants, supra note 37, at 7. ↑

45. Oklahoma v. United States, 62 F.4th 221, 224 (6th Cir. 2023), cert. denied, 144 S. Ct. 2679 (2024) reh’g granted and opinion vacated, No. 23-402, 2025 WL 1787679 (U.S. June 30, 2025), and cert. granted, judgment vacated, No. 23-402, 2025 WL 1787679 (U.S. June 30, 2025). ↑

46. Id. at 225. ↑

47. Id. at 230 (quoting Black I, 53 F.4th 869, 888−89 (5th Cir. 2022)). ↑

48. Oklahoma, 62 F.4th at 230 (quoting Sunshine Anthracite Coal Co. v. Adkins, 310 U.S. 381, 399 (1940)) (internal citations omitted). ↑

49. Oklahoma, 62 F.4th at 231 (quoting Adkins, 310 U.S. at 388). ↑

50. Oklahoma, 62 F.4th at 231 (“To ensure a fair enforcement process, the FTC could issue rules protecting covered persons from overbroad subpoenas or onerous searches. The FTC could require that the Authority provide a suspect with a full adversary proceeding and with free counsel. And the FTC could require that the Authority meet a burden of production before bringing a lawsuit or preclear the decision with the FTC.”). ↑

51. Id. ↑

52. Id. (quoting Adkins, 310 U.S. at 388). ↑

53. Oklahoma, 62 F.4th at 225. ↑

54. Walmsley v. Fed. Trade Comm’n, 117 F.4th 1032, 1036 (8th Cir. 2024) cert. granted, judgment vacated sub nom. No. 24-420, 2025 WL 1789398 (U.S. June 30, 2025). ↑

55. Id. at 1038 (quoting Oklahoma, 62 F.4th at 230). ↑

56. Walmsley, 117 F.4th at 1038 (quoting Oklahoma, 62 F.4th at 230). ↑

57. Walmsley, 117 F.4th at 1039. ↑

58. Id. (“The Commission may, for example, ‘issue rules protecting covered persons from overbroad subpoenas or onerous searches.’ The Commission may choose to create rules that require the Authority to obtain the Commission’s approval before the Authority acts to commence a civil action under § 3054(j). The Commission has power to review the Authority’s enforcement actions and to reverse them.”) (quoting Oklahoma, 62 F.4th at 231) (internal citations omitted). ↑

59. Walmsley, 117 F.4th at 1039 (quoting Oklahoma, 62 F.4th at 231). ↑

60. Walmsley, 117 F.4th at 1039−40. ↑

61. Id. at 1041−42 (Gruender, J., dissenting). ↑

62. Id. at 1042 (citing 15 U.S.C. § 3054(a)). ↑

63. Walmsley, 117 F.4th at 1042 (quoting Sunshine Anthracite Coal Co. v. Adkins, 310 U.S. 381, 399 (1940)) (internal citations omitted). ↑

64. Walmsley, 117 F.4th at 1043. ↑

65. Id.; see also 15 U.S.C. § 3054(h) (“The Authority shall have subpoena and investigatory authority with respect to civil violations committed under its jurisdiction.”); id. § 3054(j)(1) ([T]he Authority may commence a civil action against a covered person or racetrack that has engaged, is engaged, or is about to engage, in acts or practices constituting a violation of this chapter or any rule established under this chapter.”). ↑

66. Walmsley, 117 F.4th at 1043−44. ↑

67. See id. at 1042. ↑

68. Black II, 107 F.4th 415, 435 (5th Cir. 2024). ↑

69. Id. at 426. ↑

70. Id at 429. ↑

71. Id. at 430. ↑

72. Id. at 429. ↑

73. Id. at 431. ↑

74. Id. ↑

75. Id. (citing 15 U.S.C. §§ 3054(e)(2), 3054(c)(1), 3054(e)). ↑

76. Black II, 107 F.4th at 431. ↑

77. Black I, 53 F.4th 869, 887−88 (5th Cir. 2022). ↑

78. Compare 15 U.S.C. § 78s(c) (“The Commission, by rule, may abrogate, add to, and delete from . . . the rules of a self-regulatory organization…as the Commission deems necessary or appropriate to insure the fair administration of the self-regulatory organization, to conform its rules to requirements of this chapter and the rules and regulations thereunder applicable to such organization, or otherwise in furtherance of the purposes of this chapter.”) (emphasis added), with 15 U.S.C. § 3053(e) (“The Commission . . . may abrogate, add to, and modify the rules of the Authority promulgated in accordance with this chapter as the Commission finds necessary or appropriate to ensure the fair administration of the Authority, to conform the rules of the Authority to requirements of this chapter and applicable rules approved by the Commission, or otherwise in furtherance of the purposes of this chapter.”) (emphasis added). ↑

79. Black II, 107 F.4th at 435. ↑

80. Id. ↑

81. Id. ↑

82. Id. ↑

83. Black I, 53 F.4th 869, 887 (5th Cir. 2022). ↑

84. Black II, 107 F.4th at 424, 426 (“Our [Black I] decision did not address this challenge to the Authority’s enforcement powers[.]”). ↑

85. Id. at 435. ↑

86. 15 U.S.C. § 3054(h) (emphasis added). ↑

87. Id. § 3053(b)(2). ↑

88. Black II, 107 F.4th at 429. ↑

89. 15 U.S.C. § 3053(e). ↑

90. Brief Amici Curiae of Senator Mitch McConnell and Representatives Paul Tonko and Andy Barr in Support of Applicants, supra note 37, at 9. ↑

91. 15 U.S.C. § 3054(a). ↑

92. Walmsley v. Fed. Trade Comm’n, 117 F.4th 1032, 1043 (8th Cir. 2024) (Gruender, J., dissenting); Black II, 107 F.4th at 431. ↑

93. Oklahoma v. United States, 62 F.4th 221, 231 (6th Cir. 2023), cert. denied, 144 S. Ct. 2679 (2024) reh’g granted and opinion vacated, No. 23-402, 2025 WL 1787679 (U.S. June 30, 2025), and cert. granted, judgment vacated, No. 23-402, 2025 WL 1787679 (U.S. June 30, 2025). ↑

94. 15 U.S.C. § 3054(a). ↑

95. See id. § 3053(e). ↑

96. Walmsley, 117 F.4th at 1039 (quoting Oklahoma, 62 F.4th at 231). ↑

97. Walmsley, 117 F.4th at 1040. ↑

98. See 15 U.S.C. § 3058(c). ↑

99. Oklahoma, 62 F.4th at 231 (quoting Sunshine Anthracite Coal Co. v. Adkins, 310 U.S. 381, 388 (1940)). ↑

100. Oklahoma, 62 F.4th at 231. ↑

101. Brief Amici Curiae of Senator Mitch McConnell and Representatives Andy Barr

     and Paul Tonko in Support of Defendants-Appellees and Affirmance of Judgment, supra note 37, at 17 (citing 15 U.S.C. § 78s(c)); see also supra note 78. ↑

102. Black I, 53 F.4th 869, 888 (5th Cir. 2022). ↑

103. Brief Amici Curiae of Senator Mitch McConnell and Representatives Andy Barr

     and Paul Tonko in Support of Defendants-Appellees and Affirmance of Judgment, supra note 37, at 17. ↑

104. Black II, 107 F.4th 415, 434 (5th Cir. 2024). ↑

105. Id. at 435. ↑

106. Id. ↑

107. See supra text accompanying notes 92−97. ↑

108. See Response Brief of Authority Defendants-Appellees at 3, Walmsley v. Fed. Trade Comm’n, 117 F.4th 1032 (8th Cir. 2024) (No. 24-420) (citing 15 U.S.C. § 3053(b)). ↑

109. See id. at 17–18 (citing 15 U.S.C. §§ 3053(e), 78s(c)). ↑

110. See supra note 78 and accompanying text. ↑

111. See Response Brief of Authority Defendants-Appellees, supra note 108, at 4–5, 17 (citing 15 U.S.C. §§ 3058, 78s(c)). ↑

112. Oklahoma v. United States, 62 F.4th 221, 244 (6th Cir. 2023), cert. denied, 144 S. Ct. 2679 (2024) (Cole, J., concurring). ↑

113. Id. (citing 15 U.S.C. § 3058(c)(3)(C)). ↑

114. Brief Amici Curiae of Senator Mitch McConnell and Representatives Andy Barr

     and Paul Tonko in Support of Defendants-Appellees and Affirmance of Judgment, supra note 37, at 15. ↑

115. See Black I, 53 F.4th 869, 887 (5th Cir. 2022) (“[E]very court to consider a non-delegation challenge to the Maloney Act has concluded that there is ‘no merit in the contention that the Act unconstitutionally delegates power to’ a private entity.”) (quoting Nat’l Horsemen’s Benevolent & Protective Ass’n v. Black, 596 F. Supp. 3d 691, 717 (N.D. Tex.), rev’d and remanded, 53 F.4th 869 (5th Cir. 2022)). ↑

116. Brief Amici Curiae of Senator Mitch McConnell and Representatives Andy Barr

     and Paul Tonko in Support of Defendants-Appellees and Affirmance of Judgment, supra note 37, at 4. ↑

117. Protecting Racehorses, Am. Museum of Nat. Hist., https://www.amnh.org/exhibitions/horse/an-enduring-bond/protecting-racehorses#:~:text=Fatigued%20bodies%20are%20prone%20to,legs%20can%20twist%20and%20break [https://perma.cc/6E3F-AP5D]; All About the Thoroughbred, The Jockey Club, https://www.thejockeyclub.co.uk/the-racing/all-about-the-thoroughbred/ [https://perma.cc/HKV6-RM6K]. ↑

118. Protecting Racehorses, supra note 117. ↑

119. Brief Amici Curiae of Senator Mitch McConnell and Representatives Andy Barr

     and Paul Tonko in Support of Defendants-Appellees and Affirmance of Judgment, supra note 37, at 9 (quoting Sealed Indictment ¶ 3, at 3, United States v. Grasso, No. 1:20-cr-00163 (S.D.N.Y. unsealed Mar. 9, 2020)). ↑

120. Brief Amici Curiae of Senator Mitch McConnell and Representatives Andy Barr

     and Paul Tonko in Support of Defendants-Appellees and Affirmance of Judgment, supra note 37, at 7 (citing H.R. Rep. No. 116-554, at 17 (2020)). ↑

121. Brief Amici Curiae of Senator Mitch McConnell and Representatives Andy Barr

     and Paul Tonko in Support of Defendants-Appellees and Affirmance of Judgment, supra note 37, at 8−9. ↑

122. Oklahoma v. United States, 62 F.4th 221, 226 (6th Cir. 2023), cert. denied, 144 S. Ct. 2679 (2024). ↑

123. Luke P. Breslin, Reclaiming the Glory in the “Sport of Kings”—Uniformity is the Answer, 20 Seton Hall J. Sports & Ent. L. 297, 315 (2010). ↑

124. Joe Perez, Thorpedo Anna Wins Horse of the Year, Blood Horse (Jan. 23, 2025), https://www.bloodhorse.com/horse-racing/articles/282241/thorpedo-anna-wins-horse-of-the-year [https://perma.cc/5KXP-6P32]/ ↑

125. Breaking Down HISA with Kenny McPeek, Horses Races Now, at 09:10 (Dec. 9, 2024) (downloaded using Apple Podcasts). ↑

126. Brief Amici Curiae of Senator Mitch McConnell and Representatives Andy Barr

     and Paul Tonko in Support of Defendants-Appellees and Affirmance of Judgment, supra note 37, at 10. ↑

127. Breaking Down HISA with Kenny McPeek, Horses Races Now, at 09:15 (Dec. 9, 2024) (downloaded using Apple Podcasts). ↑

128. See supra text accompanying notes 12−16. ↑

129. Brief Amici Curiae of Senator Mitch McConnell and Representatives Andy Barr

     and Paul Tonko in Support of Defendants-Appellees and Affirmance of Judgment, supra note 37, at 8 (quoting Statement by William M. Lear, Jr. at 3, Legislation to Promote the Health and Safety of Racehorses: Hearing on H.R. 1754, the Horseracing Integrity Act of 2019, Before the Subcomm. on Consumer Prot. & Com. of the H. Comm. on Energy & Com., 116th Cong. (2020)). ↑

130. Id. ↑

131. Andrew Cohen, Keeping Pace: The Costs of HISA Come Into View, Paulick Rep. (Jan. 22, 2024), https://paulickreport.com/features/keeping-pace/keeping-pace-the-costs-of-hisa-come-into-view#:~:text=HISA%20costs%2C%20when%20they%20are,any%20per%2Drace%20start%20fees [https://perma.cc/2SHX-MPQ7]. ↑

132. Id. ↑

133. Natalie Voss, ‘Rogue Veterinarian’: Documents Reveal Background of 245 Allegedly Illegal Joint Injections, Investigation Is Ongoing, Paulick Rep. (Feb. 21, 2025), https://paulickreport.com/news/the-biz/rogue-veterinarian-documents-reveal-background-of-245-allegedly-illegal-joint-injections-investigation-is-ongoing [https://perma.cc/C6KD-S2L4]. ↑

134. Id. ↑

135. Id. ↑

136. Id. ↑

137. Id. ↑

138. See, e.g., Cecilia Vega, Aliza Chasan, Sarah Koch & Madeleine Carlisle, Amid Doping Scandals and Deaths, Can a New National Regulator Clean Up Horse Racing?, CBS News (Nov. 12, 2023), https://www.cbsnews.com/news/horse-racing-doping-scandals-deaths-60-minutes/ [https://perma.cc/ZH6L-ULDJ] (“The FBI said horse doping led to broken legs, cardiac issues and, in some cases, death. Richards was the lead FBI agent on the doping investigation that led to the conviction of 29 veterinarians, horse trainers and drug distributors.”). ↑

139. Why Is Horse Racing Still Popular in 2025?, The Plaid Horse (June 25, 2025), https://www.theplaidhorse.com/2025/06/25/why-is-horse-racing-still-popular-in-2025/ [https://perma.cc/J6FM-QB55]. ↑

140. Kentucky Derby 2025 Draws Largest TV Audience Since 1989, Horse Racing Nation (May 5, 2025), https://www.horseracingnation.com/news/Kentucky_Derby_2025_draws_largest_audience_since_1989_123 [https://perma.cc/WD8G-M9FZ]. ↑

141. How Popular is Horseracing?, The Jockey Club, https://www.thejockeyclub.co.uk/the-racing/racing-explained/days-at-races/how-popular-is-horseracing/#:~:text=The%20sport%20was%20popular%20in,popular%20sports%20in%20the%20world [https://perma.cc/P3JQ-5M96]. ↑

142. American Horse Council, 2023 National Survey Reveals Horse Industry Supports 2.2 Million Jobs, Paulick Rep. (Jan. 31, 2024), https://paulickreport.com/horse-care-category/2023-national-survey-reveals-horse-industry-supports-2-2-million-jobs [https://perma.cc/W49Y-HVU7]. ↑

143. Id. ↑

144. Id. ↑

145. Economic Impact of the Equine Industry in Kentucky, Ky. Thoroughbred Ass’n, https://www.kentuckybred.org/kentucky-equine-industry-impact/ [https://perma.cc/5ADG-SMZV]. ↑

146. $82m Economic Impact of Keeneland Breeders’ Cup Revealed, Thoroughbred Daily News (Mar. 17, 2023, 9:55 AM), https://www.thoroughbreddailynews.com/82m-economic-impact-of-keeneland-breeders-cup-revealed/ [https://perma.cc/9C9Q-JYUM]. ↑

147. T.D. Thornton, Supreme Court Won’t Hear HISA Constitutionality Challenge to Sixth Circuit Ruling, Thoroughbred Daily News (June 24, 2024, 8:28 PM), https://www.thoroughbreddailynews.com/supreme-court-wont-hear-hisa-constitutionality-challenge-to-sixth-circuit-ruling/ [https://perma.cc/B85J-Y5Q9]. ↑

148. Andrew Cohen, Keeping Pace: The Supreme Court May Soon Take a HISA Case, Paulick Rep. (Sept. 24, 2024), https://paulickreport.com/features/keeping-pace/keeping-pace-the-supreme-court-may-soon-take-a-hisa-case [https://perma.cc/2ZC4-6A5Z]. ↑

149. Brief Amici Curiae of Senator Mitch McConnell and Representatives Andy Barr

     and Paul Tonko in Support of Defendants-Appellees and Affirmance of Judgment, supra note 37, at 14. (“Indeed, HISA was so uncontroversial and bipartisan that it was selected for inclusion in the annual appropriations bill consistent with this process and consequently enacted by large majorities in both chambers. And Congress’ response to the Fifth Circuit’s decision of November 18, 2022, also came to the floor of the Senate by unanimous consent.”). ↑

150. See Circuit Status, Demand Justice (Aug. 11, 2025), https://demandjustice.org/circuit-status/ [https://perma.cc/372E-G5ZY]. ↑

151. See Garrison, supra note 1. ↑

152. Black I, 53 F.4th 869, 880 (5th Cir. 2022) (citing Carter v. Carter Coal Co., 298 U.S. 238 (1936); Sunshine Anthracite Coal Co. v. Adkins, 310 U.S. 381 (1940)). ↑

153. “[E]very carrier providing interstate telecommunications services [is required] to ‘contribute[]’ . . . to a fund designed to ‘preserve and advance universal service.’ The FCC must use the money in that fund, now known as the Universal Service Fund, to pay for subsidy programs for designated populations and facilities needing improved access.” Fed. Commc’ns Comm’n v. Consumers’ Rsch., No. 24-354, slip op. at 4 (U.S. June 27, 2025) (citing 47 U.S.C. §§ 254(a), (d), (e); Wisconsin Bell, Inc. v. United States ex rel. Heath, No. 23-1127, slip op. at 2 (U.S. Feb. 21, 2025)) (internal citations omitted). ↑

154. “To calculate how much carriers must contribute to the Fund . . . , the FCC has devised a formula, known as the ‘contribution factor.’” Fed. Commc’ns Comm’n v. Consumers’ Rsch., No. 24-354, slip op. at 7 (U.S. June 27, 2025) (citing 47 CFR §54.709(a)). ↑

155. Fed. Commc’ns Comm’n v. Consumers’ Rsch., No. 24-354, slip op. at 7, 37 (U.S. June 27, 2025). ↑

156. HISA Issues Statement on Supreme Court Orders, Horseracing Integrity & Safety Auth. (June 30, 2025), https://hisaus.org/news/hisa-issues-statement-on-supreme-court-orders [https://perma.cc/FF92-BNFA]. ↑

157. See id. (“That [Consumers’ Research] ruling aligns with what every district court has already said in the HISA cases: the private nondelegation doctrine challenges against the Act have no merit.”); HBPA Confident That 5th Circuit Will Rule HISA Unconstitutional a Third Time, Nat’l Horsemen’s Benevolent & Protective Association, https://nationalhbpa.com/hbpa-confident-that-5th-circuit-will-rule-hisa-unconstitutional-a-third-time/ [https://perma.cc/X8SB-VD9J] (“Our well-founded arguments regarding HISA remain unchanged—it is . . . deeply flawed, unconstitutionally delegates governmental authority to a private corporation and places unfair burdens on horsemen.”). ↑

158. United States v. Arthrex, Inc., 594 U.S. 1, 23 (2021). ↑

159. Id. at 25. ↑

160. Id. at 27. ↑

161. See 15 U.S.C. § 3058(c). ↑

162. See 15 U.S.C. § 3053(e). ↑

163. See supra Section IV.A. ↑

164. HISA Communications, Continued Decline in Racing Fatalities Among HISA’s Third Quarter Metrics, Paulick Rep. (Dec. 17, 2024), https://paulickreport.com/news/the-biz/continued-decline-in-racing-fatalities-among-hisas-third-quarter-metrics#:~:text=The%20third%20quarter%20of%202024,began%20on%20May%2022%2C%202023 [https://perma.cc/5FRS-RGXY]. ↑

165. HISA Communications, HISA: Non-Covered States Have Nearly Double Fatality Rate of Covered States, Paulick Rep. (Mar. 17, 2025), https://paulickreport.com/news/the-biz/hisa-non-covered-states-have-nearly-double-fatality-rate-of-covered-states [https://perma.cc/KB84-7S3E]. ↑

166. Breaking Down HISA with Kenny McPeek, Horses Races Now, at 06:30 (Dec. 9, 2024) (downloaded using Apple Podcasts). ↑

167. Id. at 35:45. ↑

168. See supra text accompanying notes 1−3. ↑

169. Oklahoma v. United States, 62 F.4th 221, 225 (6th Cir. 2023), cert. denied, 144 S. Ct. 2679 (2024). ↑

170. See HISA Communications, supra note 164. ↑

Charters have such strong pedigree in corporate law that is easy to take them for granted. But do we need them? If so, do they need to be publicly filed? And why are they so hard to access in Delaware? By examining these questions, an unappreciated function of charters emerges. While charters are poorly designed for broad public transparency, they still provide benefits by functioning like a recording system that reduces investigation costs and produces particularly reliable information. This sharper understanding of charters has important consequences for current doctrinal controversies and the broader debate over Delaware’s future in corporate law.

Introduction

When today’s innovative startups begin corporate existence, they do so by filing an instrument with ancient origins: a charter.^[1] Charters loom large in corporate scholarship, having been characterized as a corporation’s “constitution,”^[2] a vehicle for corporate transparency,^[3] and a site of governance innovation.^[4] Charters are also central to today’s highest profile controversies in corporate law.^[5]

While it is easy to take charters for granted given their pedigree, their peculiar design warrants attention. Most charter content consists of internal arrangements of corporate participants, such as financial and voting rights.^[6] But the most distinctive feature of charters—one that is not shared by any other corporate document—is a requirement to file the charter with the state and thereby make it publicly available.^[7] For startups not yet subject to disclosure requirements under federal securities law and accustomed to operating in stealth mode,^[8] the charter therefore provides a unique glimpse into the inner workings of the company. But why publicly file this document detailing internal rights?

Historically, the dual nature of charters (internal content; public filing) was inevitable because each charter took the form of separate legislation.^[9] Under this “special chartering” system, the charter could only have whatever internal arrangements were set for that corporation by the public act of the legislature, thereby melding the public and the private.^[10]

But special chartering was discarded in the late nineteenth century in favor of a general incorporation system,^[11] under which corporate promoters authored their own “certificate of incorporation” (referred to as a “charter” in corporate law parlance) and filed it with the state.^[12] At the start of general incorporation, maybe it seemed natural to require filing of the charter because this public act brought the new system closer to its legislatively-based predecessor. But this approach was not inevitable. The public function of creating a nexus with the state could have been achieved through a simple registration process (like obtaining a taxpayer ID),^[13] and the private function of delineating internal rights could have been achieved through private record keeping (like the minute book maintained by every corporation).^[14]

All this is not to say that charters, in their current form, play no beneficial role in corporate law. In this essay, I explore benefits of the charter system that scholars have already proposed—such as providing a venue for altering statutory defaults and providing transparency for investors and the broader public.^[15] While these ideas have merit, neither fully aligns with both the content and administration of contemporary charters. The focus on altering statutory defaults does little to explain the public filing requirement, and the assertion of transparency benefits, at least in its broadest form, is at odds with actual charter content.

I therefore explore a previously unacknowledged benefit of charters. Inspired by the historical requirement (now repealed) to file charters in real estate recording offices, I argue that charters improve transactional efficiency by providing especially definitive records of potentially conflicting shareholder rights, much in the way that real estate recording systems reduce investigation costs for property transfers.^[16] This argument is not only novel but also aligns with both charter content and filing practices.

Sharpening our understanding of charters has important implications. First, it may inform doctrinal controversies over charter content, such as determining which governance arrangements should be implemented exclusively in charters. The Delaware Chancery Court’s opinion in Moelis, and the legislature’s controversial decision to undermine that decision, is one example.^[17] Second, this essay’s less than flattering historical account of charter administration may serve as a cautionary tale for Delaware amidst a burgeoning movement by tech companies to reincorporate in other jurisdictions. One reading of the history is that Delaware got too comfortable after dominating the charter business and neglected technical and administrative improvements that today’s entrepreneurs might expect.^[18]

I. Historical Background

In early statehood, Delaware relied on the legislature to form individual corporations through customized charters—a system known as special chartering or special incorporation.^[19] As use of corporations increased in the nineteenth century, this became a corrupting and unwieldy process.^[20] Delaware, like other jurisdictions, therefore experimented with the concept of general incorporation, whereby promoters created their own corporations under a system guided by a statute. 

Delaware’s first general incorporation law, however, was a flop. The process required by that statute was, by today’s standards, laughably cumbersome. Forming a corporation required three incorporators (two of whom had to be Delaware residents), publication of notice in a newspaper 30 days in advance, submission of a certificate of incorporation to a local judge (during a vacation of the court), review by the judge for compliance with statutory requirements and to confirm the corporation would not be “injurious” to the community, further publication of notice in the newspaper initiated by the judge, and filing of the approved certificate with the Secretary of State and the local county recording office.^[21] Under this unwieldy system, it was apparently easier to hire a lobbyist and secure a charter from the legislature.^[22] The general incorporation system languished. 

Eventually opinion turned so sharply against special incorporation that the Delaware constitution prohibited the practice.^[23] At this same time, a group of local boosters believed that Delaware could compete with other states for charter business. ^[24] These boosters included lawyers hoping to profit by forming their own enterprises (which I call “registered agent companies”) to assist with incorporations and serve as registered agents for Delaware corporations. ^[25] The group secured passage of the 1899 General Corporation Law, which eliminated the most problematic features of the earlier general incorporation statutes and marked the true start of general incorporation in Delaware.^[26]

The statute was based on a New Jersey law that had successfully drawn charters to that state.^[27] New Jersey, and then Delaware, were “chartermongering” (or more politely, “liberal”) jurisdictions that actively sought incorporation activity for tax revenues and to create demand for related services.^[28] The effort was a success. By the early twentieth century, Delaware was competing effectively for the charters of the nation’s largest corporations^[29] and featured a significant cottage industry of lawyers and service providers.^[30]

II. What’s in a Charter?

One way to understand the content of charters is to examine who is likely to make use of the information required, or customarily contained, in the document. This exercise reveals three serious candidates: the state, creditors of the corporation, and current and potential shareholders. Among these three constituencies, current and potential shareholders appear to be the primary focus of modern charter practice. Appreciating this point helps pinpoint the essential function of charters, as discussed further in Part IV.

A. Information for the State

The state needs basic information to administer a corporate law system. Several items required in the charter could serve this purpose. Examples include the corporate name and the identity of the corporation’s registered agent for receiving legal notices.^[31] Without obtaining this basic identifying information, the state could not, for example, collect its franchise taxes (fees for being a Delaware corporation).

B. Information for Creditors

Initially, a significant portion of charter content was pitched at creditors of the corporation. Delaware’s 1899 statute, for example, required that a charter include a description of the officers who would conduct corporate affairs, the value of property that the corporation could hold, the maximum indebtedness that the corporation could incur, and the extent to which private property of stockholders would be subject to corporate debts.^[32] All of this would presumably be of interest to a potential lender looking to verify that a potential extension of credit was properly authorized and that the loan would be backed by the expected assets.

These requirements, however, did not survive. The first three were promptly repealed when the 1899 statute underwent its first round of revisions in 1901.^[33] The fourth requirement—to state whether stockholders would be personally responsible for corporate debts—survived a surprisingly long time given the slim chance that a corporation would voluntarily elect for shareholder liability. In 1967, when the Delaware corporate statute underwent a major update, this provision was finally flipped on its head to provide that liability would be limited to corporate assets unless the charter provided otherwise.^[34]

Other forms of creditor protection in the 1899 statute survive, but only in vestigial form. In the original statute, the concept of legal capital loomed large.^[35] The concept was that each share of stock would have a designated “par value,” which was the minimum purchase price the corporation could charge for issuing a share. The aggregate par value of all shares sold to investors would be the company’s legal capital. The corporation would not be permitted to pay dividends out of that legal capital, which would serve as a buffer for creditors. A corporation could not commence business until it had reached a designated level of legal capital through selling initial shares.^[36]

This system proved ineffective, and Delaware abandoned regulating legal capital in any meaningful sense.^[37] By 1917, a Delaware corporation could sell stock without any par value.^[38] By 1967, references to minimum levels of legal capital were deleted from the charter requirements.^[39]

C. Information for Shareholders (Investors)

Most of a contemporary charter is directed at shareholders and potential shareholders considering an investment in a company. Examples include financial rights of shareholders, voting rights of shares, the scope of the board’s fiduciary duties to shareholders, and limits on shareholder litigation.

To add context, I describe these categories of information across the lifecycle of a corporation. Specifically, I focus on the subset of corporations that begin as “startups” seeking venture capital (“VC”) with a goal of growing into large publicly traded companies. This focus on VC-backed startups has advantages and disadvantages. On the one hand, one cannot assume this segment of the corporate environment is representative of all corporations. On the other hand, corporate planning is especially standardized for these startups, producing forms and templates that should reflect actual practice in this setting.^[40]

What this lifecycle approach ultimately shows is that the content of charters fluctuates, with a largely boilerplate document at incorporation, a considerably longer and more consequential document from first preferred stock financing to initial public offering, and a return to a more formulaic document in the public company phase. This ends up being another helpful insight for understanding the function of charters in Part IV because it highlights the relative importance of charters to a particular type of negotiated investment transaction (preferred stock financings).

1. Formation Stage

At formation, the only parties are the startup’s founders (“promoters” in an earlier age).^[41] The goal at this stage is to produce a minimally viable charter. Often, a page or two is adequate.^[42]

Beyond basic identifying information, the statute requires that a charter designate how many shares the corporation is “authorized” to issue.^[43] This number is an outer limit to how much dilution a shareholder might experience without the procedural trappings of a charter amendment. At this point, the authorized shares will likely be the number of shares initially issued to the founders plus a cushion for contemplated stock options.^[44]

Superficially, the statute requires the charter to declare a corporate purpose.^[45] This is a holdover from the special incorporation era during which the declaration of purpose did real work in defining the lawful scope of the entity’s activities.^[46] But early in the general incorporation era, lawyers and corporate promoters pushed boundaries and declared open-ended purposes.^[47] In 1967, the legislature conceded to this practice by specifically allowing it.^[48]

Beyond this required information, the statute offers mechanisms for limiting shareholder litigation, and founders routinely take up the offer. Startups likely include in their charters:

• A provision permitting or committing the corporation to indemnify officers and directors to the full extent permitted by the Delaware statute;^[49]
• An “exculpation clause” providing that directors and officers will not be liable for money damages for certain breaches of the duty of care;^[50] and
• A clause requiring that shareholders file corporate law claims, such as breaches of fiduciary duty, exclusively in Delaware courts.^[51]

It is also common for charters to alter at least two statutory defaults concerning voting. Charters might permit voice votes and extend the power to amend bylaws to the board of directors.^[52]

Altogether, the goal is an unobjectionable starting point for future negotiations with investors. That is not to say the document is inconsequential. The provisions limiting shareholder litigation, for example, have material effect if the company eventually faces litigation. But even these provisions are so ubiquitous that founders and their legal teams can safely include them without scaring off future investors.^[53]

 

2. Venture Capital Financings

If the company attracts interest from VCs, the nature and length of the charter changes significantly. In business terms, this is because VCs negotiate for financial, governance, and information rights. Specifically, they almost always negotiate for preferred stock containing enhanced financial and control rights compared to the common stock held by founders. These terms appear in the charter, rather than private agreements among the company and the investors, because the Delaware statute has, since 1927, required that special rights (preferences) of any class of stock be designated in the charter.^[54]

By word count, preferred stock terms are the primary focus of a modern startup’s charter. Under the statute, simply labeling stock as “preferred” would have no concrete meaning. Instead, all economic and control features of preferred stock must be spelled out in detail. Examples include:

• A “liquidation preference,” meaning a right to receive a payout before common shareholders receive a payout if the company is acquired;^[55]
• The rate at which preferred shares convert into common shares, which usually includes a mathematical formula that adjusts for underpriced share issuances to other investors;^[56]
• The right to elect at least one member of the company’s board;^[57] and
• A right to veto major transactions, such as a sale or dissolution of the company.^[58]

Beyond these preferred stock terms, the charter may narrow the directors’ duty of loyalty to accommodate VC representatives on the board. Since 2000, the Delaware statute has allowed companies to narrow directors’ duty of loyalty by disclaiming specific business opportunities in the charter, and many charters do so.^[59]

After negotiating terms with VC investors, the charter reaches its high-water mark. Though word count is not a perfect proxy for legal significance, it is revealing that the document swells to over 17,000 words based on the widely adopted model from the National Venture Capital Association (“NVCA”).^[60]

It is worth noting, however, that the charter is not the only documentation for a VC financing. Even at this point, when the charter swells in size, a good deal of a VC financing involves contracting around the charter. Examples include a voting agreement, registration rights agreement, investor rights agreement, and others.^[61] Why spread the deal across these other documents?

One possible reason is privacy. Consider, for example, the VC’s right to elect a director. Because the voting rights of preferred stock must be delineated in the charter,^[62] it is not surprising that a section of the charter establishes an exclusive right of the preferred shareholders to control election of one or more board positions.^[63] What is surprising is that these same rights are also spelled out, with additional detail, in a separate voting agreement signed by the company, the founders (as the holders of common stock), and the VC investor.^[64] In this additional agreement, particular holders of the preferred stock (for example, a specific VC fund) may have special rights to control the nomination process for an entire series of preferred stock, and the initial designee may be identified by name.^[65] Why would corporate planners waste time and risk inconsistencies by documenting the right in overlapping documents? Why not spell out the board rights once and for all in the charter?

Although some sources suggest that the charter cannot grant special rights to particular holders because of the traditional principle that all shares within a series must have identical rights,^[66] this is not a meaningful constraint in Delaware, where the voting rights of shares can vary based on shareholder identity.^[67] Doctrinally, a Delaware charter could name names.

Doing so, however, would put more information into the public record because the charter is filed.^[68] Admittedly, this information—basically, the name of a particular fund and VC partner—may not have obvious proprietary value. After all, VC funds often tout their portfolio companies openly, and VC partners sometimes openly discuss their affiliations with those companies.^[69] But we should not dismiss the reflexive instinct of some investors to remain private. A VC could have concerns that are as pedestrian as uninvited solicitations from eager entrepreneurs who cull charters for a funding target list. Or the privacy instinct could be a matter of ego—some VCs not wanting other VCs to know that they are not the specific fund that controls the nomination process. In any event, absent any doctrinal reason to move this information out of the charter, privacy seems like a plausible explanation.

3. Public Companies

Counterintuitively, a more established company with publicly traded stock ordinarily reverts to a slimmer charter. That is because preferred stock ordinarily converts to common stock once the company has an initial public offering (“IPO”) and becomes listed on a stock exchange.^[70]

There are, however, new features that emerge in the charters of mature companies. These features respond to an important contextual change—open trading on stock markets that may introduce adversarial investors. Accordingly, founders may insulate themselves from certain forms of shareholder pressure.

For example, an increasing number of companies adopt a “dual-class” voting structure immediately prior to an IPO.^[71] Under these provisions, founders hold a class of common shares with super voting rights. More traditionally, public companies often adopted staggered boards.^[72] When board members serve staggered terms (i.e., they are not all up for election each year), it takes longer for a hostile acquirer to replace the board with more sympathetic directors.^[73]

While dual-class voting and staggered boards are legally important, what is most striking about public company charters is how generic they are on balance. The Delaware statute gives expansive license to include in a charter “any provision creating, defining, limiting and regulating the powers of the corporation, the directors, and the stockholders . . . if such provisions are not contrary to the laws of this State.”^[74] Yet, compared to the intricate preferred stock rights in the pre-IPO phase, the menu of provisions discussed above (dual-class stock, staggered boards, and limits on shareholder litigation) form a relatively set pattern. For example, in a 2013 essay, Michael Klausner reviewed a sample of over 300 IPO charters in search of idiosyncratic arrangements that would vindicate the dominate contractarian view of corporate law. What he found were almost entirely “plain vanilla” documents, containing staggered boards but not much else.^[75] Resorting again to the imperfect metric of word count, Amazon, with a market capitalization of almost $2.5 trillion, operates with a charter of just over 1,000 words.

To be clear, it is not that innovation has been lacking in corporate governance. Recent decades have seen many meaningful trends such as increasing independence of directors, requirements that directors be elected by a majority vote (not just a plurality), and increased access to the company’s proxy materials.^[76] But these innovations have mostly been enacted outside of the charter, through bylaw amendment (which can be achieved unilaterally by shareholders), pressure from proxy advisors, or regulatory action.^[77]

III. What Does Delaware Do with Charters?

To fully understand charters, it is necessary to look beyond their contents and examine the filing requirements that distinguish charters from all other corporate governance documents. Doing so reveals that Delaware, despite its reputation for attentiveness to corporate interests, maintains a surprisingly archaic system of charter administration.

To illustrate the point, it is helpful to begin with an historical anecdote. Despite streamlining the incorporation process in 1899,^[78] one inconvenient feature of the early incorporation statutes persisted for almost the entire twentieth century. Until 1996, incorporators were still required to file a charter twice: once with the Secretary of State and again with a county recorder office.^[79]

One can imagine why legislatures turned to county recorder offices at the outset of general incorporation. Administering what they hoped would be a brisk business in charters required state capacity. Creditors, shareholders, and potential investors would presumably want access to reliably authentic copies of charters. But going after New Jersey’s incorporation business was a bold gambit, so better to lean on existing institutions than risk a large outlay in building something new. A natural candidate for the task would be county recording offices with experience indexing real estate deeds.

What is surprising is how long it took for Delaware to even partially modernize. By 1967, Delaware undertook a far-reaching effort to update the statute to reflect emerging best practices. The state commissioned an influential report by a corporate law professor named Ernest Folk; in that report, Folk noted that the dual filing requirement was inefficient and presented a potential trap for incorporation mistakes.^[80] His recommendation was for the Secretary of State, which must have had considerable capacity by then, to be the sole recipient of corporate filings.^[81] Folk, however, sensed that the recommendation would be sensitive. Noting that “local recordation may have to be continued for reasons not related to the advantage and convenience of corporations,” Folk recommended both an option for eliminating the recording requirement and, in the alternative, a less ambitious option for streamlining the dual recording requirement.^[82]

Sure enough, the committee tasked with evaluating the proposals quickly rejected Folk’s preferred approach.^[83] Although there is sparse explanation in the committee’s minutes, a contemporaneous commentary provided a cynical account: “[T]o eliminate the Recorders would mean putting people out of jobs and the legislature might not have accepted that.”^[84] Moreover, the committee member who moved to retain the recording system was a representative of a registered agent company,^[85] which might have benefitted from a fragmented system that required professional assistance to navigate.

The recording requirement persisted for almost three more decades, until finally eliminated in 1996. At that point, the process was simplified from the incorporator’s perspective. The Secretary of State became the sole recipient of filings and entered information into a computer system referred to as the Delaware Corporation Information System.^[86] Even then, however, the statute showed continuing signs of parochial interests. To this day, the Secretary of State collects fees on behalf of the county recording offices and remits them to the county recorders. In 1996, these fees were described within the statute as compensation for the expense of transitioning to the new system. By 2022, this rationale was apparently hard to maintain and the legislature deleted the rationale from the statutory language (but nonetheless increased the amount of the fees to the county recorders).^[87]

Turning to the present, Delaware still lags its competitors in making charters and other filings accessible to the public. In California, for example, any interested party can go to a public website, type in the name of a California corporation, and download copies of filings for free.^[88] In Delaware, all that is available online for free is the most basic information: name, incorporation date, and address of the registered agent (almost certainly a registered agent company).^[89] For an extra $10, one can find out if the company is up to date on its franchise taxes.^[90] For an additional $10, the investor can see an index of corporate filings, but not the actual content.^[91] To see the content of any filing, the investor would need to submit an order and provide a check or payment information for potentially significant fees. For a certified copy, those fees will start at $50 plus $2 extra per page (with some charters spanning more than 50 pages).^[92] An expedited request, to be completed in an hour, will set the investor back an extra $1,000.^[93] This process will spit out a paper copy, with extra charges for expedited delivery.^[94] To an extent, commercial service providers make things easier. One can pay a registered agent company to make a document request and retrieve the paper copy at the Secretary of State offices. In addition, there are commercial data providers who assemble copies of some charters and provide access for a fee.^[95] But the help provided by these vendors comes at a cost and does not change the basic reality that charters are relatively expensive and burdensome to obtain in Delaware.

To some, this clunky system may come as a surprise. The level of service provided by the Delaware Secretary of State is sometimes cited as a strength of the Delaware system.^[96] But this is true only for a particular class of corporation for whom $1,000 expedite fees are trivial. In other words, for all its efforts to liberalize incorporation procedures, Delaware is still catering to high-end clientele.^[97] Since 1899, when the legislature eliminated the requirement to post notice of incorporation in a newspaper,^[98] broad and frictionless dissemination of information has not been a priority.

IV. What Do Charters Do Best?

This Part considers what useful functions are performed by the current charter system. After considering candidates already suggested in legal scholarship, I zero in on an unappreciated benefit of charters. Like the real estate recording system to which charters were originally linked, charters enhance transactional efficiency by providing a particularly definitive record of matters within the charter.

A. Contract With the State

A charter has long been described as both a contract among corporate participants and a contract with the state.^[99] Symbolically, filing the charter announces to Delaware that an entity has been created under its laws and subject to its jurisdiction. More practically, at least some information in the charter (corporate name, name of incorporators, registered agent office) is in fact useful or necessary for the state of Delaware to administer its corporate law.^[100]

This is not, however, a complete answer to why corporate charters take their current form. Only a small portion of the charter conveys the type of information the state uses, and that information could easily be collected in the annual report corporations are required to file with the state. In theory, the state could review the whole charter for compliance with Delaware corporate law. But early experiments with that kind of compliance check failed,^[101] and it does not appear that the Delaware Secretary of State now reviews most filings in any substantive way.^[102]

B. A Sticky Altering Rule

When corporate law scholars discuss the role of charters, they mostly focus on charters as a vehicle for opting out of statutory defaults.^[103] In the Delaware statute, a rule is frequently preceded by “unless otherwise provided in the certificate of incorporation.”^[104] Or the Delaware statute suggests that a charter may include a provision that displaces a default rule.^[105] The charter is not the only place where corporate actors can depart from statutory defaults, but it is often the legislature’s preferred location because amending a charter requires the approval of both the board and shareholders, whereas other governance instruments are easier to enact or alter.^[106] Against this backdrop, requiring that a particular default rule be changed only by charter amendment is tantamount to saying it should be hard to change. In more conceptual terms, the charter is legislative shorthand for a demanding “altering rule” that makes the associated default rule “sticky.”^[107]

It is hard to argue with this characterization, so far as it goes. A straightforward reading of the statute reveals that the legislature can and does invoke the charter as a hurdle to altering some statutory defaults.^[108]

So why doesn’t this fully answer the question of why charters have persisted in corporate law? Recall that this essay is concerned with the combination of charter content and administration. The puzzle it seeks to answer is why a general incorporation system should be centered on a publicly filed document memorializing the internal arrangements of corporate actors. The altering rule literature only provides an explanation of charter content. It does not seem to address why charters are filed and made public.

Put differently, if we only think of charters as shorthand for a particular altering rule, then charters are not independently very important. The Delaware legislature could replace “except to the extent the certificate of incorporation otherwise provides” with some slightly clunkier language such as “except as approved in a resolution of the board of directors and at a shareholder meeting called in accordance with this act.”^[109] As Gabriel Rauterberg and Sarath Sanga state in a recent piece, what is important in this literature on altering rules is the approval process and not “the charter per se.”^[110]

C. Transparency

Some commentators suggest a broad public interest in viewing corporate charters.^[111] These commentators point to scandals at high-profile startups that produced public harm or violation of public law.^[112] They note that these companies fall below thresholds for SEC disclosure requirements and that their corporate charters are the only information available about their governance.^[113] The implication is that heightened public knowledge of defective governance structures through charters could discourage future incidents.

This line of reasoning is appealing. It would help explain the unique filing and public availability of charters in a way that the literature on altering rules does not. It aligns with corporate scholarship asserting broad public benefits to disclosure.^[114] It even has some historical pedigree – charters once served a broader audience than just shareholders.^[115]

Yet, the mechanism for advancing broad public interest through charter disclosure is murky if charters take anything like their current form. Assume, for example, that an aggressive corporate planner took every inch that current law allows in drafting the charter for a rogue startup: dual class stock that gave the founder outsized influence over the board, indemnification and exculpation provisions that desensitized the board to fiduciary liability, and provisions allowing board members to take specified corporate opportunities for themselves. What then? Would consumers boycott? Would health and safety regulators launch preemptive interventions based on this information? In fact, the hypothetical charter features just listed are common,^[116] sitting in thousands of corporate charters with copies on offer by the Delaware Secretary of State for those willing to go to the trouble. The problem is that the kinds of information contained in charters is not actionable in the way this argument implies.

One can, however, make a somewhat narrower transparency argument by focusing on employee investors rather than the broader public.^[117] One prominent feature of Silicon Valley is the use of stock options and other forms of stock-based compensation to attract, motivate, and retain employees.^[118] As unicorn startups grow to have headcounts in the thousands, these employees are not exactly the general public but nor are they true insiders or sophisticated investors like VCs.^[119] In theory, these employee investors benefit from public access to charters, which set the fundamental stock rights that determine how gains from an acquisition or other exit will be divided among shareholders.

But even those who advocate for greater access to charters on these grounds concede that the language of charters is impenetrable to most employees.^[120] An emerging literature on pre-IPO equity compensation suggests that most employees already fail to investigate or understand even the terms of their individual stock option agreements.^[121] While some atypical employees may be more motivated to investigate,^[122] even they would likely struggle to use the information in the charter because it is written as a detailed legal document rather than a disclosure tool and it lacks many of the inputs required to effectively value stock.^[123] In academic studies, economists armed with PhDs have sometimes been able to infer common stock values from charter provisions, but this is heady stuff that earns placement in top-tier finance journals.^[124] If there is a need for the average employee to better understand his or her equity compensation, the appropriate vehicle is likely a new form of disclosure under federal securities law tailored to the capabilities and time constraints of employees.^[125]

In short, we could have a charter system that prizes transparency for a wide audience, but it is not the charter system we have. Ultimately, the transparency rationale does not align with charter content.

D. An Unappreciated Role: Transactional Efficiency

A different function of charters emerges by returning to the system’s historical connection to real estate recording because the two systems (charter filing and real estate recording) are more similar than one might expect. A real estate recording system is a species of registration system that encourages property transfers by lowering transaction costs between buyer and seller.^[126] Some real estate transfers would still occur without the recording system because sellers could sometimes satisfy themselves that there are no competing ownership claims. But that would be an expensive process without some official record of previous transfers. By allowing for the recording of real estate transfers and encouraging use of the system by providing legal advantages for recorded instruments,^[127] transaction costs decline and beneficial property transfers increase.^[128]

One can think of charters as providing a similar kind of benefit, though the mechanics differ. In theory, all corporate rights could be defined by private agreements between the corporation and individual investors. In such a system, a new investor could use a variety of tools to satisfy itself that newly offered rights do not conflict with existing rights: due diligence investigations,^[129] representations and warranties,^[130] and certifications from corporate officers.^[131] This job, however, is made easier when certain topics must be addressed in the charter, such as financial rights of preferred stock, voting rights, and litigation procedure. By requiring these rights to be delineated in a single charter, giving the charter precedence over other forums,^[132] and requiring the charter to be filed with the state, investors can place special confidence in this information. From this perspective, the essential role of the charter is not broad-based dissemination of information but instead the production of particularly definitive information for those already committed to a thorough investigation of fundamental corporate matters.^[133] 

Note that terms required to be included in the charter are not only more conspicuous; they may also be more coherent due to being compiled in this single document. The charter serves a “forcing function” that surfaces and avoids potential mistakes as capital structure and governance grow more complicated.^[134] Today’s unicorn startups can have many rounds of financing, starting with the sale of “Series A” preferred stock and eventually reaching Series B, C, D, and so on.^[135] Sometimes a senior series of stock has complicated interactions with prior series.^[136] Defining these interrelated rights within the four corners of a single document forces the drafters to squarely address potential conflicts and produces a more integrated structure than what might result from a series of individual agreements with individual investors.

In a sense, the transactional efficiency argument presented in this Part might feel like only a slight variation on the transparency rationale described earlier. Both ascribe an informational role to the charter. But note that transactional efficiency of the type described here is sometimes in tension with broader public transparency goals. As described above, pre-IPO investors have a revealed preference for keeping investment terms private. This is a likely explanation for why corporate planners push terms into side contracts. ^[137] In theory, making charters more private might best promote transactional efficiency. One could imagine a system in which charters were filed with the state, but companies then controlled access to the document.^[138] Such a system might reduce incentives to move content out of the charter and into side agreements, thereby increasing the informational value of the charter. Currently, there is no groundswell of support for making startups less transparent.^[139] But highlighting this tension has explanatory power. The fact that charters currently sit behind paywalls and bureaucracy might even be a feature and not a bug because a system of mass dissemination might cut against transactional efficiency.

To be clear, there is no evidence that this kind of transactional efficiency was on the minds of legislators when they designed the system. The original reliance on recording offices was inherited from New Jersey and was likely a practical necessity. Its persistence for almost a century was understood as a political capitulation.^[140] Nonetheless, this theory of the current charter system fits the facts on the ground better than any of the alternatives because it aligns with both charter content and administration.

V. Implications

This final Part considers how a better understanding of charters matters for corporate law and broader debates about the future of Delaware as the dominant corporate jurisdiction. Specifically, it considers implications for legal doctrine that determines charter content and potential improvements to the system of charter administration.

A. Determining Charter Content

Courts and legislators are periodically called upon to determine the required and permitted content of charters. For example, in its Moelis decision, the Delaware Chancery Court declared portions of a shareholder agreement between a publicly traded corporation and its controlling shareholder invalid because the court viewed those provisions as granting governance rights to the controlling shareholder that would normally be possessed by the board.^[141] Because the Delaware statute provides that management of a corporation must be vested in its board except as provided otherwise in the certificate of incorporation, the court determined that the provisions in question could only be implemented in the charter.^[142] To bring the point home, the court colorfully explained that everything the controlling shareholder sought to accomplish could have been embodied in a “golden share” created under the charter but was inappropriate for a mere contract between the controlling shareholder and the corporation.^[143]

Practitioners had a negative reaction to the opinion, stating that it would call into question forms of agreements already in use.^[144] The legislature responded by undercutting the court and amending the statute to allow the agreements in question outside of the charter.^[145]

The legislative response drew criticism from corporate scholars. Many scholars focused on the legislative process, expressing concern that Delaware had capitulated too quickly and outside of its usual process.^[146] Others defended the substance of the court’s position, with at least one scholar focusing on the public transparency benefits of including pre-IPO governance arrangements in the publicly available charter.^[147]

The transactional efficiency arguments advanced in this essay are another reason to support the court’s original position. When VCs invest in a startup, they need to understand existing governance arrangements. VCs can and do verify this information through due diligence and representations and warranties.^[148] But the process would be marginally easier if backed by the legal requirement to include atypical board restrictions within the four corners of the charter. By backtracking on the court’s ruling, the legislature whittled away at these informational benefits.^[149]

Improving Charter Administration

Though charter competition has long been a favorite topic of legal scholars,^[150] Delaware has faced very little actual competition for decades. A considerable majority of publicly traded companies, and high-growth startups seeking to obtain that status, choose the state.^[151]

The way Delaware administers charters seems to reflect this comfortable position. The system caters to the top of the market by offering concierge-style service. While the baseline product is technologically primitive and slow compared to other jurisdictions,^[152] a company with a significant legal budget can speed up the process through paying expedite fees and hiring registered agent companies to file and retrieve documents.^[153] The costs of these extra services may represent inefficient rents to parochial interests,^[154] but they seem to be tolerated due to other advantages of Delaware. In short, a clunky filing and retrieval system seems to be a luxury that Delaware has been able to afford.

But what if we enter a new phase of jurisdictional competition? What if the current “Dexit” trend, led by prominent tech CEOs, is real? ^[155] Perhaps it is time for Delaware to think more creatively about how it can attract the next Musk or Zuckerberg. In other words, perhaps Delaware should start paying more attention to the top of the charter funnel – the thousands of new entrepreneurs currently choosing Delaware at incorporation.

In this regard, Delaware might adopt the mindset of the specialized lawyers to these startups. So-called startup lawyers have been recognized as engaging in a distinctive style of legal practice that has grown up right alongside the startup and venture capital system.^[156] These lawyers play the long game by showing a willingness to standardize work product at the expense of current billable hours.^[157] They also fight the legal profession’s stodgy reputation by adopting the technologies their clients develop, sending a signal that they share an innovative mindset.^[158]

Similarly, Delaware might re-evaluate charter administration, even if its approach was sufficient historically. Such an effort would include investment in improved technological interfaces and reducing fees for basic administrative functions like document retrieval.^[159] The filing system, after all, might be the most salient feature of Delaware law to a new entrepreneur. Delaware statutory and case law lurks in the background throughout a startup’s lifecycle, but interactions with the filing system are conspicuous. Each incorporation, preferred stock financing, and document retrieval is a chance for the state to make a favorable or unfavorable impression. Currently, Delaware may be sending the wrong message.

Conclusion

Charters are not fully explained by current scholarship. This essay tried to fill that gap by identifying a new function grounded in minimizing investigation costs and enhancing transactional efficiency. Hopefully this new perspective helps policy makers refine both charter content and administration.

1. * Professor of Law, University of California College of the Law, San Francisco. Thank you to John Crawford, Scott Dodson, Evan Epstein, Jill Fisch, Jonathan Lipson, Jared Mayer, Naveen Thomas, Megan Wischmeier Shaner, and Emily Strauss for helpful comments.

   Elizabeth Pollman, The History and Revival of the Corporate Purpose Clause, 99 Tex. L. Rev. 1423, 1426 (2021) (“The legal and social technology of chartering has existed for over a millennium[.]”) ↑

2. See James Si Zeng, The Calculus of Shareholders’ Consent: A Constitutional Economics Theory of Corporate Charter Amendment Rules, 41 U. Pa. J. Int’l L. 429, 429 (2019). ↑

3. Jill E. Fisch, Stealth Governance: Shareholder Agreements and Private Ordering, 99 Wash. U. L. Rev. 913, 947 (2021). ↑

4. See Michael Klausner, Fact and Fiction in Corporate Law and Governance, 65 Stan. L. Rev. 1325, 1327–28 (2013) (summarizing the contractarian view as reflected in Frank H. Easterbrook & Daniel R. Fischel, The Economic Structure of Corporate Law 15 (1991)). ↑

5. See infra Part V.A (discussing the legislative response to W. Palm Beach Firefighters’ Pension Fund v. Moelis & Co., 311 A.3d 809 (Del. Ch. 2024)). ↑

6. See infra Part II (discussing the content of charters across the lifecycle of a corporation). ↑

7. See Del. Code Ann. tit. 8, § 103 (2025). ↑

8. See Alexander I. Platt, Unicorniphobia, 13 Harv. Bus. L. Rev. 115, 155–56, 166–68 (2023) (discussing the benefits to startups of operating privately). ↑

9. E.g., An Act to Incorporate the Delaware Railroad Company (1836)(chartering the Delaware Railroad Company); see also James Willard Hurst, The Legitimacy of the Business Corporation in the Law of the United States 1780–1970, at 13–55 (1970); S. Samuel Arsht, A History of Delaware Corporation Law, 1 Del. J. Corp. L. 1, 2–5 (1976). ↑

10. See Pollman, supra note 1, at 1426. ↑

11. See Arsht. supra note 9, at 4–8. ↑

12. See id. at 4; Brett H. McDonnell, Sticky Defaults and Altering Rules in Corporate Law, 60 SMU L. Rev. 383, 385 n.7 (2007). ↑

13. See Employer Identification Number, IRS, https://www.irs.gov/businesses/employer-identification-number#whoneedsanein [https://perma.cc/F89B-Y6ME] (discussing requirements to obtain an employer identification number with the Internal Revenue Service). ↑

14. See Del. Code Ann. tit. 8, § 224 (2025) (requiring a corporation to maintain a corporate minute book). ↑

15. See infra Parts III.B & C. ↑

16. See infra Part IV.D. ↑

17. See infra Part V.A (discussing W. Palm Beach Firefighters’ Pension Fund v. Moelis & Co., 311 A.3d 809 (Del. Ch. 2024) and the controversial legislative response). ↑

18. See infra Part V.B (discussing the “DExit” movement led by Elon Musk’s reincorporation of Tesla in Texas). ↑

19. See Russell C. Larcom, The Delaware Corporation 1–2 (1937); Pollman, supra note 1, at 1426–36; Charles M. Yablon, The Historical Race Competition for Corporate Charters and the Rise and Decline of New Jersey: 1880–1910, 32 J. Corp. L. 323, 331–335 (2007). ↑

20. See Larcom, supra note 19, at 4–7 (describing the “abuses and evils” of special chartering and the burdens of special incorporation requests on the legislature); Arsht, supra note 9, at 1–4 (indicating 144 special incorporations, amendments, and renewals by the Delaware legislature in the single year of 1873). ↑

21. See Arsht, supra note 9, at 4–5. ↑

22. See id. at 5–6; Yablon, supra note 19, at 359. ↑

23. See Larcom, supra note 19, at 7. ↑

24. See id. at 9; Arsht, supra note 9, at 6. ↑

25. See Larcom, supra note 19, at 9 (indicating that the Delaware promoters likely modeled their registered agent company on an example in New Jersey); Yablon, supra note 19, at 347–55. ↑

26. See Arsht, supra note 9, at 6–7. For statutory references from 1899 to 1967, I cite to Nate Emeritz’s extremely helpful A Compilation of Legislation Regarding the Delaware General Corporation Law From 1899 to 1967 (2025) [hereinafter DGCL Compilation]. This document is available at the Delaware Corporation Law Resource Center maintained by University of Pennsylvania Carey Law School at https://www.law.upenn.edu/delawarecorporatehistory/ [https://perma.cc/G668-PW2Q]. ↑

27. See Arsht, supra note 9, at 7; Yablon, supra note 19, at 359. ↑

28. See Hurst, supra note 9, at 70 (describing Delaware as a “liberal state”); Harwell Wells, The Modernization of Corporation Law, 1920–1940, 11 U. Pa. J. Bus. L. 573, 576, 584 n.62 (2009) (using the terms “liberal” and “chartermongering”). ↑

29. See Larcom, supra note 19, at 175 (describing Delaware’s lead in chartering large companies from 1915–1929); Sarath Sanga, The Origins of the Market for Corporate Law, 24 Am. L. & Econ. Rev. 369, 391–94 (2022) (indicating a sharp rise in Delaware corporations in the early twentieth century); Yablon, supra note 19, at 370 (indicating that Delaware became a “viable alternative” for large corporations by the early 1900s). But see Andrew Verstein, The Corporate Census (Feb. 25, 2025), https://ssrn.com/abstract=5154952 [https://perma.cc/9QCZ-X2QG] (arguing, based on a large data set of incorporations, that Delaware’s success in attracting charters before 1986 has been overstated). ↑

30. See Larcom, supra note 19, at 16 (describing the role of registered agent companies); Joel Seligman, A Brief History of Delaware’s General Corporation Law of 1899, 1 Del. J. Corp. L. 249, 271–72 (1976) (describing the formation of the Corporation Service Company in Delaware); Yablon, supra note 19, at 360 (describing early Delaware lawyers and registered agent companies). ↑

31. See Del. Code Ann. tit. 8, § 102(a)(1) & (2) (2025). ↑

32. See DGCL Compilation, supra note 26, at 12–13 (reproducing 21 Del. Laws 273, § 7 (1899)). ↑

33. See id. at 12 (reproducing 22 Del. Laws 166, § 1 (1901)). ↑

34. See id. at 2 (reproducing 56 Del. Laws 50, § 102 (1967)). ↑

35. See id. at 12 (reproducing 21 Del. Laws 273, § 7 (1899)). ↑

36. See Wells, supra note 28, at 603–610 (providing an overview of legal capital and its declining significance in corporate law); see also John C. Coffee, The Mandatory/Enabling Balance in Corporate Law: An Essay on the Judicial Role. 89 Colum. L. Rev. 1618, 1633–35 (1989). ↑

37. See Hurst, supra note 9, at 53 (explaining that the device of legal capital left “scope for evasion”); Wells, supra note 28, at 603–10; Coffee, supra note 36, at 1633–35 (suggesting that legal capital was easy to evade). ↑

38. See DGCL Compilation, supra note 26, at 11 (reproducing 29 Del. Laws 113, §§ 4, 5 (1917)). ↑

39. See id. at 2 (reproducing 56 Del. Laws 50, § 102 (1967)). ↑

40. See Robert Bartlett, Standardization and Innovation in Venture Capital Contracting: Evidence from Startup Company Charters, Sept. 2023, at 5, http://dx.doi.org/10.2139/ssrn.4568695 [https://perma.cc/8XEP-MXTL] (studying charters of VC-backed companies and finding that 85% are based on the National Venture Capital Association model legal documents). ↑

41. Larcom, supra note 19, at 14-16. ↑

42. For example, the law firm Orrick Herrington & Sutcliffe LLP provides a form of initial charter in its Orrick Tech Studio. See Forms, Orrick, https://www.orrick.com/en/tech-studio/resources/forms [https://perma.cc/BH35-9AUP] (on file with author) [hereinafter Orrick Charter]. The document is approximately two pages and 700 words. ↑

43. See Del. Code Ann. tit. 8, § 102(a)(4) (2024). ↑

44. See Frequently Asked Questions, Orrick, https://www.orrick.com/en/tech-studio/resources/faq/what-type-of-stock-should-my-corporation-have-how-much-stock-should-my-corporation-have#:~:text=The%20number%20and%20types%20of,or%20“voting%20stock [https://perma.cc/AY3M-LP5M]. ↑

45. See Del. Code Ann. tit. 8, § 102(a)(3) (2024). ↑

46. Pollman, supra note 1, at 1426–36. ↑

47. James B. Dill, The Statutory and Case Law Applicable to Private Companies Under the General Corporation Act of New Jersey 20 (1898) (“It is often sought to broaden the powers by inserting such words as, ‘to do any other business which the company from time to time determine.’”). ↑

48. See DGCL Compilation, supra note 26, at 2 (reproducing 56 Del. Laws 50, § 102 (1967)). ↑

49. See Del. Code Ann. tit. 8, § 145 (2024); Orrick Charter, supra note 42, at Article VI. ↑

50. See Del. Code Ann. tit. 8, § 102(b)(7) (2024); Orrick Charter, supra note 42, at Article VI. ↑

51. See Del. Code Ann. tit. 8, § 115 (2024); Orrick Charter, supra note 42, at Article VII. ↑

52. See Del. Code Ann. tit. 8, § 109 (2024) (providing a default rule that only shareholders can amend bylaws); id. at § 211(e) (providing a default rule of vote by written ballot); Orrick Charter, supra note 42, at Article V (providing for voice votes and permitting the board to amend bylaws). ↑

53. See, e.g., Jens Frankenreiter, Cathy Hwang, Yaron Nili & Eric Talley, Cleaning Corporate Governance, 170 U. Pa. L. Rev. 1, 37 (2021) (finding that 96% of publicly traded Delaware corporations include director exculpation clauses in their charters). ↑

54. See DGCL Compilation, supra note 26, at 9 (reproducing 35 Del. Laws 85, § 4 (1927)). ↑

55. See Nat’l Venture Capital Ass’n, NVCA Model Legal Documents, Certificate of Incorporation, 7–13 (2024) [hereinafter NVCA Charter]. ↑

56. See id. at 19–35. ↑

57. See id. at 15–16. ↑

58. See id. at 16–19. ↑

59. See Gabriel Rauterberg & Eric Talley, Contracting Out of the Fiduciary Duty of Loyalty: An Empirical Analysis of Corporate Opportunity Waivers, 117 Colum. L. Rev. 1075, 1079, 1126 (2017) (finding corporate opportunities waivers in public company charters); NVCA Charter, supra note 55, at 41. ↑

60. This is the word count excluding explanatory footnotes and appendices. ↑

61. Model Legal Documents, NVCA,https://nvca.org/model-legal-documents/ [https://perma.cc/YR6M-9Q9C]. ↑

62. See also Del. Code Ann. tit. 8, § 141 (2024). ↑

63. See NVCA Charter, supra note 55, at 15–16. ↑

64. See Nat’l Venture Capital Ass’n, NVCA Model Legal Documents, Voting Agreement 3–6 (2024). ↑

65. See id. ↑

66. Therese H. Maynard, Dana M. Warren & Shannon Trevino, Business Planning: Financing the Start-Up Business and Venture Capital Financing 520 (3d ed. 2018) (citing Model Business Corporation Act and California statutory requirements). ↑

67. See Colon v. Bumble, Inc., 305 A.3d 352 (Del. Ch. 2023) (permitting “identity-based voting”). ↑

68. See id. ↑

69. For example, the VC firm Andreessen Horowitz lists its portfolio at https://a16z.com/portfolio/ [https://perma.cc/T9ST-PXXE]. ↑

70. See NVCA Charter, supra note 55, at 31–32 (providing for conversion upon IPO). ↑

71. See Ofer Eldar, Dual-Class IPOs: A Solution to Unicorn Governance Failure, (Rsch. Handbook on the Structure of Private Equity & Venture Capital, Working Paper, Paper No. 741, 2023) (indicating that almost 30% of IPOs in 2017–2019 featured dual-class stock); Gladriel Shobe & Jarrod Shobe, The Dual-Class Spectrum, 39 Yale J. on Reg. 1286, 1294–1300 (2022) (discussing the history of dual-class stock). ↑

72. See Klausner, supra note 4, at 1333 (summarizing empirical studies indicating that most IPO charters in the late 90s and early 2000s featured staggered boards). ↑

73. See id. at 1352–55. ↑

74. See Del. Code Ann. tit. 8, § 102(b)(1) (2024). ↑

75. See Klausner, supra note 4, at 1337–39. ↑

76. See id. at 1359–62 (citing director independence and majority voting as examples); Megan Wischmeier Shaner, Interpreting Organizational “Contracts” and the Private Ordering of Public Company Governance, 60 Wm. & Mary L. Rev. 985, 1004 (2019) (discussing majority voting and proxy access). ↑

77. Klausner, supra note 4 at 1359 (discussing regulatory changes, shareholder proposals, and activist campaigns); Shaner, supra note 76, at 992–1005 (stating that “[t]he predominant venue for ex ante corporate governance in organizational documents has been the bylaws”). ↑

78. See supra Part I. ↑

79. See DGCL Compilation, supra note 26, at 2 (reproducing 56 Del. Laws 50, § 102 (1967)) (providing an example of the recording requirements as of the 1967 revisions to the statute). ↑

80. Ernest L. Folk, III, Review of the Delaware Corporation Law 1–2 (1967), https://libguides.law.widener.edu/c.php?g=772882&p=5544590 [https://perma.cc/L4LD-GLAZ]. ↑

81. See id. at 6–7. ↑

82. See id. at 1–2. ↑

83. Minutes of the Delaware Corporation Law Study Committee (July 14, 1964), Widener University, https://delawarelaw.widener.edu/about/library/research/delaware-corporation-law-documents/ [https://perma.cc/VLX4-X328]. ↑

84. Comment, Law for Sale: A Study of the Delaware Corporation Law of 1967, 117 U. Pa. L. Rev. 861, 869 (1969). ↑

85. See id. at 865 (identifying committee member David H. Jackman as President of the United States Corporation Company of New York). ↑

86. See 70 Del. Laws 349, § 1 (1996); 70 Del. Laws 587, §§ 2-6 (1996). ↑

87. See 83 Del. Laws 77, § 2 (2022). ↑

88. See Business Search, California Secretary of State, https://bizfileonline.sos.ca.gov/search/business [https://perma.cc/48PD-TKRR] (indicating that California provides free access to over 17 million business entity filings). California may be an outlier in the degree to which it provides free access to electronic images of filings. Another Delaware competitor, Nevada, appears to employ processes and charge fees that are similar to Delaware. See Commercial Recordings Copies and Certificate Services Fee Schedule, Nevada Secretary of State, https://www.nvsos.gov/sos/home/showpublisheddocument/1043/638345248868870000 [https://perma.cc/BEE4-WRPA]. Another competitor, Texas, lies somewhere in between by charging modest fees (compared to Delaware) for electronic copies. See Instructions for Ordering Copies and Certificates Using SOSDirect, Texas Secretary of State, https://www.sos.state.tx.us/corp/instructions-for-copies.shtml [https://perma.cc/N3BC-4PUE]. ↑

89. This information is available through the business entity search feature of the Delaware Division of Corporations website located at https://icis.corp.delaware.gov/Ecorp/EntitySearch/NameSearch.aspx. ↑

90. See id. ↑

91. See id. ↑

92. Delaware Division of Corporations, Fee Schedule Del. Dep’t of State (Revised Aug. 1, 2024), https://corpfiles.delaware.gov/AugustFee2024.pdf [https://perma.cc/ZXS7-RK5A]. ↑

93. Delaware Division of Corporations, Certification Memo Del. Dep’t of State, https://corpfiles.delaware.gov/Cover_Memo/updatedcertmemoandinstructionsSept24.pdf [https://perma.cc/UK3M-SPJD]. ↑

94. See id. ↑

95. See Bartlett, supra note 40, at 19 (describing a subscription service named VC Experts that provides access to charters of venture-backed companies); Evan Epstein, 157 Tom Callahan, CEO of NPM: Unlocking Liquidity in Private Markets, Boardroom Governance With Evan Epstein (Dec. 16, 2024), https://boardroom-governance.com/episodes/tom-callahan [https://perma.cc/WUZ2-AKHS ] (discussing how Nasdaq Private Markets provides institutional investors with access to charters of private companies). ↑

96. Amy Simmerman, William B. Chaldler & David Berger., Delaware’s Status as the Favored Corporate Home: Reflections and Considerations, Harv. L. Sch. F. on Corp. Gov. (May 8, 2024), https://corpgov.law.harvard.edu/2024/05/08/delawares-status-as-the-favored-corporate-home-reflections-and-considerations/ [https://perma.cc/AD4Z-6FZU ] (discussing the ability to obtain pre-clearance of corporate filings). ↑

97. Delaware corporate law has been characterized as catering to large corporations more generally. See Jonathan R. Macey & Geoffrey P. Miller, Towards an Interest-Group Theory of Delaware Corporate Law, 65 Tex. L. Rev. 469, 478 (1987) (discussing R. Posner & J. Scott, Economics of Corporation Law and Securities Regulation (1980)). ↑

98. See DGCL Compilation, supra note 6, at 12 (reproducing 21 Del. Laws 273, § 7 (1899)). ↑

99. Dill, supra note 48, at 18 (describing the charter as a contract between the corporation and the state, the individual stockholders and the corporation, and the stockholders themselves). ↑

100. See supra Part II.A. ↑

101. See supra text accompanying notes 21 & 22. ↑

102. See Simmerman, supra note 97 (“Unlike certain other states, Delaware reviews most filings only for the form of the filing and does not conduct substantive review, which allows filings to be processed quickly.”). ↑

103. See Jeffrey N. Gordon, The Mandatory Structure of Corporate Law, 89 Colum. L. Rev. 1549, 1553 (1989); Brett H. McDonnell, Sticky Defaults and Altering Rules in Corporate Law, 60 SMU L. Rev. 383, 402 (2007) (discussing charter amendment in terms of triggering a particular “altering rule” for opting out of a statutory default); Gabriel Rauterberg & Sarath Sanga, Altering Rules: The New Frontier for Corporate Governance, 42 Yale J. Reg. 291 (discussing a variety of altering rules in corporate law). ↑

104. See Gordon, supra note 103, at 1553. ↑

105. See e.g., Del. Code. Ann. tit. 8 § 102(b)(7) (2024) (permitting a corporation to adopt a provision exculpating directors or officers for breaches of the traditional duty of care). ↑

106. Bylaws can be changed unilaterally by shareholders or, in most corporations, the board. See Del. Code Ann., supra note 51. Other major corporate decisions are made by simple board resolution. See Rauterberg & Sanga, supra note 103, at 308. ↑

107. See McDonnell, supra note 103, at 402 (“[A]n altering rule that allows opting out of a default through either the charter or the bylaws is significantly less sticky than one that allows opting out only through the charter.”). ↑

108. See i.d. at 439 (identifying 36 matters that can only be altered in the charter); Gordon, supra note 103, at 1533. ↑

109. See e.g., Del. Code Ann. tit. 8 § 271(c) (2024) (allowing for sale of substantially all the corporation’s assets after approval by board and shareholders). ↑

110. See Rauterberg & Sanga, supra note 103, at 326–27. ↑

111. See Fisch, supra note 3, at 947. ↑

112. See id. at 948. ↑

113. See id. at 946. ↑

114. See Ann M. Lipton, Not Everything Is About Investors: The Case for Mandatory Stakeholder Disclosure, 37 Yale J. on Reg. 499 (2020). ↑

115. See supra Part II.B. ↑

116. See supra Part II.C. ↑

117. See Jennifer S. Fan, Regulating Unicorns: Disclosure and the New Private Economy, 57 B.C L. Rev. 583, 609 (2016) (arguing that unicorn startups should be required to post their certificates of incorporation on public websites and provide plain English summaries for employees). ↑

118. Yifat Aran, Beyond Covenants Not to Compete: Equilibrium in High-Tech Startup Labor Markets, 70 Stan. L. Rev. 1235, 1261–67 (2018) (discussing the role of equity compensation in Silicon Valley). ↑

119. See Abraham J.B. Cable, Fools Gold? Equity Compensation & the Mature Startup, 11 Va. L. & Bus. Rev. 615, 636–41 (2017). ↑

120. Fan, supra note 117, at 609. ↑

121. Yifat Aran & Raviv Murciano-Goroff, Equity Illusions, J. of Law & Econ. & Org. 196 (2021). ↑

122. See Abraham J.B. Cable, Stock Options of Adhesion, 50 J. Corp. L. (forthcoming 2025). ↑

123. See Fan, supra note 117, at 609. ↑

124. See Will Gornall & Ilya A. Strebulaev, Squaring Venture Capital Valuation with Reality, 135 J. Fin. Econ. 120 (2020). ↑

125. See e.g., See Yifat Aran, Making Disclosure Work for Start-Up Employees, 2019 Colum. Bus. L. Rev. 867, 952–55 (2019) (recommending that employees receive liquidation scenarios). ↑

126. See Steven Shavell, Economic Analysis of Property Law , ch. 9 p. 10 (Nat’l Bureau of Econ. Rsch,. Working Paper No. 9695) www.nber.org/system/files/working_papers/w9695/w9695.pdf (“An advantage of registration systems is that they may ease sale and resale of things by assuring buyers of the validity of sellers’ claims of ownership.”). To be precise, the U.S. recording system is distinguished from a pure real-estate registration system. See Shavell, supra note 65. ↑

127. In the property recording system, the legal advantage of recording is that a good faith purchaser for value can prevail over a competing transfer that was not recorded. See id. For an overview of the recording acts that produce this result, see Thomas W. Merrill, Henry E. Smith & Maureen E. Brady, Property: Principles and Policies 891–932 (4th ed. 2022). ↑

128. See Richard A. Posner, Economic Analysis of Law 89 (5th ed. 1998) (“Problems in the transfer of property rights are part of a larger problem, that of deciding who owns what property. A system of recorded titles is a great help.”). The filing system under Article 9 of the Uniform Commercial Code can be described in a similar fashion. See Douglas G. Baird, Notice Filing and the Problem of Ostensible Ownership, 12 J. Legal Stud. 53, 58 (1983). I focus primarily the real estate recording system, rather than Article 9’s filing requirement for perfecting a security interest in other forms of property, because the latter is a “notice filing system” with limited information in the actual filing, which complicates any analogy to corporate charters. ↑

129. See e.g., Sample VC Due Diligence Request List, Cooley GO, https://www.cooleygo.com/documents/sample-vc-due-diligence-request-list/ [https://perma.cc/NHZ9-UD4J]. ↑

130. See e.g., NVCA Model Legal Documents, Stock Purchase Agreement, Nat’l Venture Capital Ass’n, https://nvca.org/document/stock-purchase-agreement-updated-october-2024/ [https://perma.cc/BB9F-S87B] [hereinafter NVCA SPA] (representing that corporate documents have been provided to the purchaser). ↑

131. See e.g., id. at §14.13 (requiring delivery of a document certifying authenticity of corporate records). ↑

132. See Del. Code. Ann. tit. 8 § 109(b) (2024) (establishing that a charter has precedence over bylaws by stating that bylaws may contain any provision “not inconsistent with law or with the certificate of incorporation”). ↑

133. C.f., Baird, supra note 128, at 55 (arguing that Article 9 financing statements are not intended to broadly disseminate information to unsecured creditors but instead assist secured creditors with investigating potentially conflicting security interests). ↑

134. See Forcing Functions, Interaction Design Foundation, https://www.interaction-design.org/literature/book/the-glossary-of-human-computer-interaction/forcing-functions [https://perma.cc/72SM-3Q99 ](“A forcing function is an aspect of a design that prevents the user from taking an action without consciously considering information relevant to that action.”). ↑

135. See Abraham J.B. Cable, Time Enough for Counting: A Unicorn Retrospective, 39 Yale J. Reg. Bull. 23, 34–35 (2021) (describing the complex capital structure of Square); Elizabeth Pollman, Startup Governance, 168 U. Pa. L. Rev. 155, 200–16 (2019) (discussing complexity in the capital structure and governance of startups). ↑

136. See Cable, supra note 135, at 34–35 (discussing an IPO ratchet in Square’s senior preferred stock that caused dilution to junior stock). ↑

137. See supra text accompanying note 68. ↑

138. For example, the company could control a password for charter access. ↑

139. But see Platt, supra note 8, at 155–56, 166–68. ↑

140. See supra Part III (discussing the history of the charter recording requirement). ↑

141. See W. Palm Beach Firefighters’ Pension Fund v. Moelis & Co., 311 A.3d 809 (Del. Ch. 2024). For example, the court invalidated provisions requiring the board to place the controlling shareholder’s representatives on board committees. See id. at 818, 825. ↑

142. See id. at 822–23. ↑

143. See id. at 822. ↑

144. Letter in Opposition to the Proposed Amendment to the DGCL, Harv. L. Sch. F. on Corp. Gov. (June 7, 2024), https://corpgov.law.harvard.edu/2024/06/07/letter-in-opposition-to-the-proposed-amendment-to-the-dgcl/ [https://perma.cc/6SWP-WJ94] (summarizing arguments by proponents of the legislative action) [hereinafter Letter in Opposition]. ↑

145. Del. Code Ann. tit. 8 § 122(18) (2024). ↑

146. See Letter in Opposition, supra note 144 (criticizing the legislative process). ↑

147. See Fisch, supra note 3, at 947–48. ↑

148. See Sample Due Diligence Request List, supra note 129; NVCA SPA, supra note 130 (providing examples of how VCs verify facts in VC financings). ↑

149. Admittedly, the benefits of the court’s rule might be modest. But the cost of complying with the rule would also be modest judging by the reaction of the NVCA legal forms group to Moelis. See NVCA Charter, supra note 4, at Article VIII, n. 81 (relocating provisions to the charter in response to Moelis). ↑

150. See Larcom, supra note 19, at v–vi (providing an early description of competing perspectives on the effects of charter competition); Yablon, supra note 19, at 371–76 (summarizing competing views of charter competition by scholars such as Lucian Bebchuck, William Carey, Jonathan Macey, Roberta Romano, and Geoffrey Miller). ↑

151. Bartlett, supra note 40, at 27; Klausner, supra note 4, at 1329 (“Delaware is the only state in the race, and it dominates the market.”); Verstein, supra note 29, at 13-16 (finding that a majority of publicly traded companies incorporated in Delaware by 1986). ↑

152. See supra text accompanying notes 89–95. ↑

153. See supra text accompanying notes 93–98. ↑

154. As discussed in Part III, the specific interests are registered agent companies and (historically) county recording offices. See supra Part III. For a comprehensive analysis of how special interests within Delaware affect the corporate law system, see Macey & Miller, supra note 97, at 498–510. ↑

155. See Stephen Bainbridge, DEXIT News: Will Meta Follow Tesla to Texas?, ProfessorBainbridge.com (Jan. 31, 2025, 12:58 PM), https://www.professorbainbridge.com/professorbainbridgecom/2025/01/dexit-news-will-meta-follow-tesla-to-texas.html [https://perma.cc/6V8H-PFV8]. ↑

156. The seminal account of startup lawyers is Mark C. Suchman & Mia L. Cahill, The Hired Gun as Facilitator: Lawyers and the Suppression of Business Disputes in Silicon Valley, 21 Law & Soc. Inquiry 679 (1996). ↑

157. See Bartlett, supra note 40, at 2–3; Abraham J.B. Cable, Startup Lawyers at the Outskirts, 50 Willamette L. Rev. 163, 167–78 (2014) (citing the work of Lawrence Friedman, Robert Gordon, Linda Bernstein, Richard Painter, and others). ↑

158. See e.g., Gunderson Dettmer Launches ChatGD, a Homegrown Generative AI Chat App, to its Lawyers, Gunderson Dettmer (Aug. 9, 2023), https://www.gunder.com/en/news-insights/firm-news/gunderson-dettmer-launches-chatgd-a-homegrown-generative-ai-chat-app-to-its-lawyers [https://perma.cc/7AFL-XELV]. ↑

159. These improvements are warranted even if one thinks that it is beneficial for companies to control access to their filings. See supra text accompanying note 138. A properly designed system could provide both privacy (through a company-controlled password) and efficiency (through lower filing fees and faster access for the company and its permitted viewers). ↑